Professional Documents
Culture Documents
@WILEY*VCH
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Related titles from WILEY-VCH
Precursor-Derived Ceramics
Part I: ISBN 3-527-26830-8
Processing of Ceramics
ISBN 3-527-29814-2
ISBN 3-527-26824-3
R. J. Brook (Ed.)
M. Swain (Ed.)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Ceramic Hard Materials
@WILEY-VCH
Ralf Riedel (Editor)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Editor:
Prof. Dr. Ralf Riedel
Fachgebiet Disperse Feststoffe
Fachbereich Materialwissenschaft
Technische Universitit Darmstadt
PetersenstraRe 23
64287 Darmstadt
Germany
This book was carefully produced. Nevertheless, authors, editor and publisher do not warrant the
information contained therein to be free of errorb. Readers are advised to keep in mind that state-
ments, data, illustrations, procedural details or other items may inadvertently be inaccurate.
A catalogue record for this book is available from the British Library
All rights reserved (including those of translation in other languages). N o part of this book may be repro-
duced in any form by photoprinting, microfilm, or any other means nor transmitted or translated into
~ ~
machine language without written permission from the publishers. Registered names, trademarks, etc. used
in this book, even when not specifically maked as such, are not to be considered unprotected by law.
Composition: Alden Bookset, Oxford
Printing: betz-druck, Darmstadt
Bookbinding: Buchbinderei Osswald, NeustadtiWstr.
Printed in the Federal Republic of Germany
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Ute, Vincent, Lorenz and Marlene
This book is dedicate to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Preface
With increasing demand for improved efficiency of engines, plants and production
processes, ceramics have gained great importance as structural engineering
materials in recent years. Within the group of the so called advanced materials,
carbon in form of diamond or diamond-like structures, carbides, nitrides and bor-
ides have reached an outstanding position due to their excellent hardness and
thermo-chemical and thermo-mechanical properties. The distinct covalent bonding
of the aforementioned structures positively influences their hardness and their tribo-
logical behavior. Moreover, a series of oxides such as stishovite, a high pressure
modification of silica, or boron sub-oxides have been recently discovered to exhibit
high hardness apart from the well known alumina.
There is presently much effort in basic science and applied research to work on
novel ceramic hard materials denoted as super- or ultra-hard materials that can
compete with the hardness of conventional diamond. Aim and scope of the research
in this field is to develop hard materials with superior mechanical and chemical
properties and with similar hardness. Moreover, calculations of properties of
hypothetical carbon nitrides like C3N4 indicated that there might be compounds
exhibiting even higher hardness values than that of diamond. The low-temperature
synthesis of diamond and cubic boron nitride on the one hand as well as the success-
ful research on new carbon nitrides on the other hand have caused an enormous
impact around the world on both the basic science and the technological develop-
ment of these novel ultra-hard materials.
With the present book we wish to review comprehensively and concisely the state
of the art concerning the structure, synthesis, processing, properties and applica-
tions of ceramic hard materials in general. In particular, the synthesis, modeling
and properties of novel hard materials like binary carbon nitrides, ternary boron
carbonitrides and others are also addressed. It is the aim of this reference book
not only to reflect the state of the art and to give a sound review of the literature,
but to delineate the underlying concepts and bearing of this interdisciplinary field.
With the present edition we wish to show that the field of hard materials research
and development has to be recognized into the wider context of chemistry, physics
as well as materials science and engineering.
The book is organized in two volumes and three parts, covering the structure
and properties of ceramic hard materials (Volume 1, Part I), synthesis and pro-
cessing (Volume 1, Part 11) as well as the typical fields of applications (Volume 2,
Part 111).
Volume 1 starts with an introduction into novel ultra hard ceramics including
diamond and diamond-like carbon, carbon nitrides and silicon nitrides as well as
boron containing carbides, nitrides and carbonitrides. Here we wish to recognize
the great fundamental and technological challenge of developing new superhard
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
VIII Preface
materials which can compete with the hardest counterparts such as diamond and
cubic boron nitride.
In dealing with properties, the first Chapter in Part I is then devoted to the struc-
ture of crystalline and amorphous ceramic hard materials. The structural features
are responsible in particular for the intrinsic materials properties such as melting
point and hardness. It has been found that in many cases the hardness of a crystal-
line substance correlates with its melting point. Therefore, detailed knowledge of the
3dimensional arrangement of the atoms is required to understand the materials
behavior under certain conditions. More details of the individual crystal structures
with respect to a 3dimensional view can be found on our hard materials homepage
under the web address ‘www.hardmaterials.de’. Phase transitions and materials
synthesis under high pressure in laser heated diamond cells is the topic of the
continuing Chapter. The materials behavior under high pressure and temperature
is of fundamental interest for the synthesis of hard materials since many of the
ultra-hard substances like diamond, cubic boron nitride or stishovite are formed
naturally or synthetically under these harsh conditions. The next three Chapters
are concerned with the mechanical behavior and corrosion of ceramic hard
materials and their relation to microstructure. This correlation is an important
feature since hardness is not only governed by the intrinsic atomic structure of
the respective material but also to a great extend by its polycrystalline nature. There-
fore, the grain morphology and grain boundary chemistry play a decisive role in the
materials response under environmental or mechanical load. In the following
Chapter transition metal carbides, nitrides and carbonitrides are discussed with a
focus on their structure and bonding, thermodynamic behavior as well as on their
physical and mechanical properties. Part I is then completed by two Chapters
which deal with the theoretical design of novel sp2-bonded carbon allotropes and
novel superhard materials based on carbon and silicon nitrides. These Chapters
tribute to the fact that with proceeding computerization the number of calculated
novel solid structures that led to the prediction of new materials with hardness
comparable to or exceeding that of diamond has increased enormously in recent
years.
Part I1 continues with the synthesis and processing of ceramic hard materials.
Since the conventional powder technological synthesis and processing of ceramics
has been treated in a large number of published review articles here we concentrate
on novel synthetic routes that provide ceramic hard materials. Consequently, six
Chapters report on i) directed metal oxidation, ii) self-propagating high temperature
synthesis, iii) hydrothermal synthesis of diamond, chemical vapor deposition of
diamond (iv) and cubic boron nitride (v) films and finally vi) the polymer to ceramic
transformation. All these processes are particularly suitable for the formation of
refractories with high hardness. Part I1 is then closed by a Chapter on nano
structured superhard materials. In the course of this work high hardness is achieved
by microstructural control rather than by the synthesis of a distinct crystal structure.
In Volume 2 ceramic hard materials are highlighted in the light of their applica-
tions. Chapter 1 of Part I11 concisely reviews the history of diamond and diamond-
like super abrasive tools while Chapter 2 and 3 are concerned with the application of
chemical vapor deposited diamond and diamond-like carbon films. These sections
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Prejace IX
include the synthesis of optical grade CVD diamond windows and discuss their
physical and mechanical properties. The most important and wide-spread ceramic
hard materials are based on alumina. Chapter 4 reports on the processing develop-
ments to increase the hardness of alumina based ceramics for grinding and cutting
applications. Silicon carbide and silicon nitride materials are the most technologi-
cally important non-oxide compounds and have gained great significance in the
field of cutting ceramics and are treated in Chapters 5 and 6. Boron-based ceramics
are a further group of either established or candidate materials with extreme hard-
ness. Therefore, Chapter 7 deals with boron carbide or transition metal borides like
titanium diboride and their distinct properties and applications. In Chapter 8, clas-
sical hard metals comprised of tungsten carbide as the hard phase and cobalt as the
binder phase are discussed. Volume 2 is finally completed by a data base (Chapter 9)
containing approximately 130 hard materials including carbides, nitrides, borides,
silicides and oxides. The data base references the crystal structure, physical proper-
ties like melting point and density, mechanical properties (Youngs modulus, micro
hardness) and oxidation resistance of the respective compounds. Future develop-
ments of novel hard materials such as the recently discovered intermetallic phase
A1MgBl4will be updated on our internet homepage ‘www.hardmaterials.de’.
In closing these introductory remarks, I would like to emphasize that the special
chance to place a summary of the outstanding expertise on the field of present hard
materials research and development would not have been possible without the great
enthusiasm and commitment of all the colleagues who contributed in the writing of
this two volume set. I am grateful for their enormous efforts in compiling a fascinat-
ing series of articles imparting depth insight into the individual fields of modern hard
materials research. Finally, I wish to thank the Wiley-VCH Editors Peter Gregory
and Jorn Ritterbusch for encouraging me in the preparation of this book and for
their continuous support throughout the editorial process.
Ralf Riedel
March 2000
Darmstadt
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Foreword
part of the book accordingly gives close accounts of the performance of the different
classes of hard materials in the applications context.
The contributors to this text are to be congratulated on bringing their many dis-
ciplines to bear on this central theme. Materials science is well known to undergo
fashions as materials are developed and discarded and indeed as sectors of applica-
tion grow and decline. The one requirement which will remain is that the forming
and shaping of materials will always be necessary whatever the eventual sector of
application. We can accordingly be confident that the long history of hardness
studies, not least in the last two hundred years from the carbon tool steels, to
high speed steels, to stellite, to tungsten carbide, to cermets, to ceramics, and now
to diamond, boron, nitride and other special systems, will be continued with
informed imagination and with creative innovation. The present book is a splendid
platform on which to base such future development.
Richard Brook
January 2000
Oxford, UK
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Contents
1.6 Nitrides 23
1.6.1 Nitrides of Main Group Elements 24
1.6.2 Transition Metal Nitrides 25
1.6.3 Perspectives: Nitridosilicates 29
1.7 Oxide Ceramics 30
1.7.1 Hard Ceramics of Main Group Elements 30
1.7.2 Transition Metal Oxides 32
1.8 Amorphous Hard Materials 36
References 37
Index 997
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
List of Contributors
U. Anselmi-Tamburini R. Bohler
Dipartimento di Chimica Fisica Max-Planck-Institute for Chemistry
Universita di Pavia Saarstrasse 23
27100 Pavia D-55020 Mainz
Italy Germany
M. W. Bailey D. Brandon
De Beers Industrial Diamond Department of Materials Engineering
Division Pty Ltd Technion - Israel Institute of
Diamond Research Lab Technology
PO Box 1770 Haifa 32000
Southdale 21 35 Israel
South Africa
J. R. Brandon
G. Berg De Beers Industrial Diamond
Fachgebiet und Institut fur Division Pty Ltd
Werkstofiunde der TU Darmstadt Diamond Research Lab
und Staatliche PO Box 1770
Materialpru fungsanhalt Southdale 2 135
Grafenstrasse 2 South Africa
D-64283 Darmstadt
Germany E. Broszeit
Fachgebiet und Institut fur
Werkstofiunde der TU Darmstadt
C. Berger und Staatliche
Fachgebiet und Institut fur Materialprufungsanhalt
Werkstofiunde der TU Darmstadt Grafenstrasse 2
und Staatliche D-64283 Darmstadt
Materialpriifungsanhalt Germany
Grafenstrasse 2
D-64283 Darmstadt R. C. Burns
Germany De Beers Industrial Diamond
Division Pty Ltd
K. Bewilogua Diamond Research Lab
Fraunhofer Institut fur Schicht und PO Box 1770
Oberflachentechnik (IST)
Southdale 21 35
Bienroder Weg 54 E South Africa
D-38 108 Braunschweig
Germany
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
XXVIII List of Contributors
R. J. Caveney G. J. Davies
De Beers Industrial Diamond De Beers Industrial Diamond
Division Pty Ltd Division Pty Ltd
Diamond Research Lab Diamond Research Lab
PO Box 1770 PO Box 1770
Southdale 2135 Southdale 2135
South Africa South Africa
S. E. Coe B. R. Eggen
De Beers Industrial Diamond School of Chemistry, Physics and
Division Pty Ltd Environmental Sciences
Diamond Research Lab University of Sussex
PO Box 1770 Falmer
Southdale 2135 Brighton BNl 9QJ
South Africa UK
J. L. Collins D. Fister
De Beers Industrial Diamond HC Starck Gmbh
Division Pty Ltd Kraftwerkweg 3
Diamond Research Lab D-79725 Laufenburg
PO Box 1770 Germany
Southdale 2135
South Africa T. Frauenheim
Fachbereich Physik
P. Colombo Universitat/Gesamthochschule
Universita di Bologna Paderborn
Dipartimento di Chimica Applicata e D- 33095 Paderborn
Scienza dei Materiali Germany
viale Risorgimento 2
1-40I36 Bologna C. Friedrich
Italy Fachgebiet und Institut fur
Werkstoffkunde der TU Darmstadt
M. W. Cook und Staatliche
De Beers Industrial Diamond Materialprufungsanhalt
Division Pty Ltd Grafenstrasse 2
Diamond Research Lab D-64283 Darmstadt
PO Box 1770 Germany
Southdale 21 35
South Africa Y. G. Gogotsi
Institut fur Angewandte Mineralogie
C. R. Cousins Universitat Tubingen
Department of Physics Wilhelmstrasse 56
University of Exeter D-72074 Tubingen
Stocker Road Germany
Exeter EX4 4QL
UK
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
List of Contributors XXIX
J. 0. Hansen G. Jungnickel
De Beers Industrial Diamond Fachbereich Physik
Division Pty Ltd Universitat/Gesamthochschule
Diamond Research Lab Paderborn
PO Box 1770 D- 33095 Paderborn
Southdale 2 135 Germany
South Africa
C. P. Klages
M. Hoffmann Fraunhofer Institut fur Schicht und
Fakultat fur Maschinenbau Oberflachentechnik (IST)
Institut fur Werkstoffkunde I1 Bienroder Weg 54 E
Universitat Karlsruhe D-38108 Braunschweig
Kaiserstrasse 12 Germany
Postfach 6980
D-76 128 Karlsruhe T. Kraft
Germany lnstitut fur Angewandte Mineralogie
Universitat Tubingen
R. D. Hoffmann Wilhelmstrasse 56
Westfalische Wilhelms-Universitat D-72074 Tubingen
Miinster Germany
Anorganisch-Chemisches Institut
Wilhelm-Klemm-Strasse 8 A. Krell
D-48 149 Miinster Fraunhofer Institute for Ceramic
Germany Technologies and Sintered Materials
Winterbergstrasse 28
M. I. Heggie D-0 1277 Dresden
School of Chemistry, Physics and Germany
Environmental Sciences
University of Sussex C. D. Latham
Falmer Department of Physics
Brighton BN1 9QJ University of Exeter
UK Stocker Road
Exeter EX4 4QL
V. Jayaram UK
Department of Metallurgy
Indian Institute of Science W. Lengauer
Bangalore Institute for Chemical Technology of
India Inorganic Materials
Vienna University of Technology
W. Jeitschko Getreidemarkt 9/161
Westfalische Wilhelms-Universitat A-1060 Vienna
Miinster Austria
Anorganisch-Chemisches Institut
Wilhelm-Klemm-Strasse 8
D-48149 Miinster
Germany
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
XXX List of Contributors
J. E. Lowther R. Pottgen
Department of Physics Westfalische Wilhelms-Universitat
University of Witwatersrand Munster
Johannesburg Anorganisch-Chemisches Institut
South Africa Wilhelm-Klemm-Strasse 8
D-48149 Munster
S. Luyckx Germany
School of Process and Materials
Engineering F. Richter
University of the Witwatersrand Technische Universitat Chemnitz-
Johannesburg 2050 Zwickau
South Africa Institut fur Physik
D-09107 Chemnitz
Z. A. Munir Germany
Facility for Advanced Combustion
Synthesis R. Riedel
Department of Chemical Engineering Fachbereich Materialwissenschaft
and Materials Science Technical University of Darmstadt
University of California Petersenstrasse 23
Davis CA 95616 D-64287 Darmstadt
USA Germany
K. G. Nickel K. A. Schwetz
Universitat Tubingen Advanced Ceramics Lab
Applied Mineralogy Elektroschmelzwerk Gmbh
Wilhelmstrasse 56 Max-Schaidhauf-Strasse 25
D-72074 Tubingen D-87437 Kempten
Germany Germany
S. Ozbayraktar P. K. Sen
De Beers Industrial Diamond De Beers Industrial Diamond
Division Pty Ltd Division Pty Ltd
Diamond Research Lab Diamond Research Lab
PO Box 1770 PO Box 1770
Southdale 2135 Southdale 2135
South Africa South Africa
C. S. J. Pickles G. Serghiou
De Beers Industrial Diamond Max-Planck-Institute for Chemistry
Division Pty Ltd Saarstrasse 23
Diamond Research Lab Mainz
PO Box 1770 Germany
Southdale 2135
South Africa
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
List of Contributors XXXI
D. Sherman K. Takagi
Department of Materials Engineering Toyo Kohan Co. Ltd.
Technion - Israel Institute of Tokyo
Technology Japan
Haifa 32000
Israel R. Telle
Institut fur Gesteinshuttenkunde
M. Sibanda RWTH Aaachen
De Beers Industrial Diamond MauerstraBe 5
Division Pty Ltd D-52056 Aachen
Diamond Research Lab Germany
PO Box 1770
Southdale 2135 S. Veprek
South Africa Institute for Chemistry of Inorganic
Materials
I. Sigalas Technical University Munich
De Beers Industrial Diamond Lichtenbergstrasse 4
Division Pty Ltd D-85747 Garching b. Munich
Diamond Research Lab Germany
PO Box 1770
Southdale 2 135 C. J. H. Wort
South Africa De Beers Industrial Diamond
Division Pty Ltd
P. K. Sitch Diamond Research Lab
Fachbereich Physik PO Box 1770
Universitat/Gesamthochschule Southdale 2135
Paderborn South Africa
D- 33095 Paderbron
Germany A. Zerr
Fachgebiet Disperse Feststoffe
G. D. Soraru Technical University of Darmstadt
Universita di Trento Petersenstrasse 23
Dipartimiento di Ingegneria dei D-64287 Darmstadt
Materiali Germany
Via Mesiano 77
1-38050 Trento
Italy
R. S . Sussmann
De Beers Industrial Diamond
Division Pty Ltd
Diamond Research Lab
PO Box 1770
Southdale 2135
South Africa
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
List of Symbols
a! absorption coefficient
a atomic attraction constant
a! growth parameter
a power absorption coefficient
a?P, x polytypes or phases
P atomic repulsion constant
P geometrical factor
Y rake angle degrees
Yi secondary ion yield
r(n) orientational surface energy
"/S surface energy J
r width of X-ray reflection
s microplastic deformability
tan 6 dielectric loss factor
Ac concentratation difference
AG!98 Gibbs free energy J
ASint interfacial entropy
AT temperature change K or "C
AX change of size or mass
& elastic strain
& emissivity
i strain rate
rl degree of conversion to nitride
rl fraction of reaction completed
0 angle
8 constant relating tensile strength and hardness
20 X-ray scattering angle
6 entering angle degrees
6 thermal conductivity J cm-2 s-I K-'
61 thermal conductivity of rectants
62 thermal conductivity of products
x empirical parameter relating bulk modulus and bond length
x inclination angle degrees
x layer thickness, mean free path m
x polarity of bond
x thermal conductivity
x wavelength m
x X-ray wavelength
A mean free path m
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
XXXIV List of Symbols
P C/Ti ratio
P coefficient of friction
v Poisson ratio
7r complementary energy J
P density kg m-3
P dislocation density
P resistivity
Pm theoretical density of product
U conductivity
U electrical conductivity
ff stress Pa
UO median failure stress Pa
ffb fracture strength in bending, modulus of rupture
ffij local stress field
ffS Stefan-Boltzmann constant J cm-* sC1 KC4
UY yield stress GPa
7 annealing time
7- help time S
4 angle between crack and tensile stress
4, w, 7- ternary phases
4 azimuthal angle between polarization vector and substrate direction
4 constraint factor
X electron affinity
W wear coefficient
(4 7 average diffusion distance in time r
0 vacancy in crystal structure
liquid phase
lattice directions
Miller indices
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
List of Abbreviations
PP polymer pyrolisis
PVD physical vapor deposition
r rhombohedral
RBAO reaction bonded aluminum oxide
RBM reaction bonded mullite
rBN rhombohedral boron nitride
RBS Rutherford back-scattering
RBSN reaction bonded silicon nitride
RF radio frequency
RSF reduced spatial frequency
RSSC reaction sintered silicon carbide
S.C. simple cubic
s.c.cm. standard cubic centimeters
SAD small angle diffraction
SAXS small-angle X-ray scattering
scs Textron process Sic fibers with C core and C surface
SEM scanning electron microscope
SERR strain energy release rate
SHS self-propagating high-temperature synthesis
SIALON Si-A1-0-N (silicon aluminum oxynitride) fiber
SiCAlON SiC-AlN-Al20C composite fiber
Si-DLC Si-DLC hybrid
SIF stress intensity factor
SIMS secondary ion mass spectrometry
SNMS secondary neutron mass spectrometry
SP sintered powder
SRO short-range order
STM scanning tunnel microscopy
ta-C hydrogen-free amorphous carbon
taC tetrahedral amorphous carbon
TD theoretical density
TCNE tetracyanoethylene
TEM transmission electron microscope
TGA thermogravimetric analysis
TH triethoxysilane
TRS transverse rupture strength (= MOR)
TZP tetragonal zirconia polycrystals
UHP ultahigh purity
UPS ultraviolet photoelectron spectroscopy
uv ultaviolet
VAMAS Versailles Agreement on Materials and Standards
vc vapor phase formation and condensation process
VEC valence electron concentration
VLS vapor-liquid-solid process
vs vapor-solid reaction
wBN wurtzitic boron nitride
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
yttrium aluminium garnet, yttrium aluminate, Y2A15012
XLIII
List of Abbreviations
yttrium-lithium-fluorite
X-ray diffraction
XANES
XRD
YAG
YLF
XPS
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Introduction: Novel Ultrahard Materials
A. Zerr and *R. Riedel
Introduction
The synthesis of new materials with hardness comparable to or even harder than dia-
mond is of considerable fundamental and technological interest and is a great chal-
lenge to chemists, physicists, and materials scientists. Most of the known ultrahard
materials, including diamond and cubic boron nitride, were first synthesized in the
1950s and industrially manufactured using high pressure-high temperature processes
[I-31. Extensive research in this domain continues and recently a few new materials
have been synthesized or rediscovered as superhard ones (Si02-stishovite [4], cubic
Si3N4 [5]). New vapor deposition methods (CVD, PVD, laser ablation etc.) which
allow the deposition of diamond, cubic boron nitride and other hard materials
films at low temperature and low pressure (i.e. often under metastable conditions)
on a variety of substrates have been developed since the early 1980s [&lo].
Diamond and cBN (cubic boron nitride) combine excellent mechanical, chemical,
and physical properties. However, owing to its instability at high temperatures,
diamond cannot be used, for example, as a cutting tool for steel. Moreover, with
increasing temperatures diamond and cBN weaken due to the onset of the transforma-
tion to the graphite structure so that above 1100°C in a nonoxidizing atmosphere
boron carbide B4C(with a hardness of about 30 GPa) has been identified as the hardest
material [I I]. For this reason and because of the need to replace expensive diamond in
many other applications, new hard materials with comparable or even superior proper-
ties are required. Theoretical work on the carbon nitride C3N4predicted that this com-
pound could have a hardness comparable to or even greater than that of diamond
[12,13]. The synthesis of polycrystalline C3N4films was first reported in 1992 [I41
and in 1993 [ 151and has been under extensive investigationsince then, resulting in hun-
dreds of publications on this subject. However, in a few recent publications there has
been doubt the thermodynamic stability of theoretically predicted ultrahard C3N4
phases [ 161 and whether their hardness would be comparable to that of diamond
[ 171. In consequence, research on the low-temperature synthesis of diamond
and cubic boron nitride [18],the search for other possible candidate compounds as
well as microstructure design of known materials and their composites [ 19-21]
(heterostructures, whiskers, nanocomposites) remain the subject of experimental
and theoretical efforts. Microstructure design is addressed in this book by D. Sherman
and D. Brandon in Part I, by S. Vepfek in Part I1 and by A. Krell in Part 111.
This introduction deals with some of the latest experimental and theoretical
developments in the field of novel boron- and carbon-based ultrahard materials
as well as with new observations on a class of silicon-based compounds which
previously were not classified as ultrahard.
Hard Materials
Generally, hard materials are solids with high hardness in the range 8-10 on the
Mohs scale of hardness, given by the sequence of minerals which can be scratched
by the next (Table 1). Usually, common hard materials are subdivided into
compounds with metallic (like TIN or WC), ionic (Al2O3>,or covalent bounding
(diamond, Si3N4)[22]. The definition of ultrahard materials is that their hardness
values are comparable to that of diamond. Another definition often used in the
literature is that the hardness of such materials exceeds 40GPa. Among all
known single phase compounds diamond, cBN and probably boron carbides
(Bl3C2-Bl2C3)satisfy the latter definition. Accordingly one can expect that novel
ultrahard compounds will be found in the isothermal ternary phase diagram B-
C-N given schematically in Fig. 1. In particular, the carbon nitrides (e.g. C3N4)
and the boron carbonitrides (B,C,NZ) have been discussed as substitute materials
Table 1. Hardness ranking of minerals and some prominent synthetic ceramic materials according
to F. Mohs. In the case of the synthetic materials microhardness values are given in units of the
Knoop scale. The microhardness variations result from variations in the grain size, the load of
indentation, the phase composition and the used densification techniques. Modified after [ 1521.
Talcum Mg,[(OH)z/Si4O,oI 1
Hexagonal hBN 0.15-0.30
boron nitride*
Gypsum CaS04 2 H 2 0 2
Calcite CaCO, 3
Fluorite CaF2 4
Apatite Ca5[(F3OH)/(PO&I 5
Feldspar K[AlSi,O,] 6
Quartz Si02 7
Topaz A1z[F2/Si041 8
p-Silicon nitride* BSi3N4 17 14-26
Corundumt
Titanium nitride*
t g 3 } 9 21
14-29
' The microhardness values are taken from published sources [7] and [102].
The microhardness varations are taken mostly from previous work [25].
* Synthetic material.
Synthetic material or natural mineral.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Hard Materials XLVII
Figure 1. Schematic of the isothermal ternary B C - N phase diagram at temperatures below the
decomposition of the stoichiometric compounds and the regimes of composition of some solid solu-
tions discussed. For explanation of the compounds see text. First published in [I 521 and reproduced
with permission.
for diamond. One can also expect that metastable dense high pressure phases of the
compounds based on silicon or oxygen can belong to ultrahard materials. One of
these is the high pressure-high temperature stishovite phase of Si02 with a hardness
exceeding 33 GPa [4], which is almost an order of magnitude above that of quartz,
the ambient pressure phase of Si02 (Table 1). The second one is cubic Si3N4with the
spinel structure, whose hardness may be comparable to that of Si02-stishovite [5].
Hardness
Hardness is one of the quantitative parameters that describe resistance of a material
towards plastic (irreversible) deformation. Plastic deformations begin when the
shear component of the stress applied to a material exceeds some value called the
yield stress. There are many ways to create a plastic deformation and consequently
many ways to define and to measure resistance of a material towards such deforma-
tions. Hardness can be determined in several ways:
- Scratching methods (Mohs, Martens),
- Grinding methods (Rosiwal).
- Indenting methods (Vickers, Knoop, Brinell, Rockwell, Shore).
For example, the Vickers hardness, Hv, is defined as the applied load P divided by
the surface area of the impression, while the Knoop hardness H K is derived from the
load P divided by the projected area of the impression:
1854.4P
Hv =
d2 ’
14 229 P
HK=-
12
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
XLVIII Introduction: Novel Ultrahard Materials
H v = 3.2M3. (3)
However, this expression cannot be applied for superhard materials like diamond.
Each testing method yields different hardness values for one material. Thus, in
order to compare hardness values of different materials, the specific test method
and the test conditions have to be described carefully. Moreover, the hardness of
single crystals depends for many compounds on the crystallographic plane tested
and for the Knoop indentor additionally on the orientation of the indentor’s long
axis relative to the crystallographic axis of the examined crystal. For example, in
experiments on single-crystalaSi3N4the Vickers hardness measured on different crys-
tallographic planes varied by more than 34% and the Knoop hardness measured on
the same crystallographic plane by 7% [24]. Hardness of polycrystalline materials is
strongly influenced by: residual stresses, toughening phases, microstructural textures,
the grain size, the applied load and the porosity as well as by the structure and
composition of grain boundaries. As a consequence, the hardness of polycrystals
may significantly differ from that of the single crystal counterpart. For example,
due to the above reasons the measured microhardness of polycrystalline TIC varies
between 1200 and 3250Hv or by more than 270% and that of polycrystalline
Si3N4between 1500 and 4800 (Hv)or by more than 300% [25]. In Table 1, the micro-
hardness values and their variations for some prominent synthetic ceramic materials
are listed ranked according to the Mohs scale. The above topics are discussed in more
detail in Part I by D. Sherman and D. Brandon, and by A. Krell.
There are few theoretical attempts to describe hardness quantitatively in terms of
the elastic bulk or shear moduli of an ideal solid. The elastic moduli and the
corresponding hardness of a compound or element are strongly related to its
molar volume, to its chemical bonding, and to its crystal structure. Basically, it
can be stated that the higher the hardness the lower the molar volume and the
more covalent the bonding of the material.
Liu and Cohen suggested in their paper on hypothetical PC3N4 [12] that ‘on the
microscopic level, for ideal systems, hardness is determined by the bulk modulus’.
To estimate the bulk modulus they used an empirical model developed earlier
[26], where the bulk modulus scales as a homopolar energy gap divided by the
volume of the bond charge. The resulting relation gives theodependence of the
bulk modulus B (in GPa) as a function of the bond length d (in A) and the empirical
parameter A:
1971 - 220X
B= d3.5 ’ (4)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Hurd Materials XLlX
Table 2. Calculated equilibrium volumes, bulk moduli, and cohesive energies of some postulated
ultrahard materials. The calculated and/or experimental values of diamond, lonsdaleite, cubic
BN, and !3Si3N4are given for comparison. The experimental volume and density values are derived
from crystal structure data. CH.6 = all sp2-bonded hexagonal carbon phase 26,73,74]; Cb,c,t,.4= all
1
sp*-bonded body-centered tetragonal carbon [26,72,76]; BNh.c.t.-4= all sp -bonded body centered
tetragonal BN [119]. First published in [152] and reproduced with permission.
Cdiarnond cubic 5.67 (exp) 3.51 (exp) 444 (calc) 8.17/atom (calc)
435 (exp) 7.37/atom (exp)
Clonsdalelte hexagonal 6.47 (exp) 3.52 (exp) 440 (calc) 8.14/atom (calc)
cBN cubic 5.90 (exp) 3.45 (exp) 367 (calc) 13.2/BN pair (exp)
369 (exp)
hexagonal 10.42 (exp) 3.19-3.20 265 (calc) 74.3/unit cell (calc)
(exp) 256 (exp)
hexagonal 6.25 3.49 427 (calc) 81.5/unit cell
hexagonal 6.29 3.17 372 (calc) 9.06/atom
tetragonal 6.72 2.97 362 (calc) 8.47/atom
tetragonal 7.37 2.82 268 (calc) 13.36/BN pair
X is 0 for homopolar solids of group IV elements, and 1/2, 1, and 2 for heteropolar
solids of group 111-IV, 111-V, and 111-VI elements, respectively. From this, it is
evident that increasing ionicity results in a decrease in B. This expression has
been further improved to account for non-octet compounds such as Si3N4 by
using the average coordination number N, # 4 [26]:
(N,.) (1971 - 220X)
B=-
4 d3.5
Figure 2. Scattering of the Vickers hardness for a common set of hard materials when compared
with bulk modulus (gray region) and shear modulus (black region). Following this comparison
Teter suggested that ‘the shear modulus is a significantly better qualitative predictor of hardness
than the bulk modulus’ [ 171.
moduli from over thousand measurements and found that the shear modulus of
polycrystalline aggregates is a significantly better qualitative predictor of hardness
than the bulk modulus (Fig. 2).
Experimental bulk moduli can be obtained from the measurement of lattice
parameters and volumes as a function of pressure [30]. The single crystal elastic
moduli can be measured using the Brillouin spectroscopy, inelastic neutron scatter-
ing, ultrasonic measurements or the Schaefer-Bergmann method [311. Once the
single crystal moduli are known one can derive the bulk B and shear G moduli of
a polycrystalline material [32].
Figure 3. The different carbon modifications: (a) hexagonal graphite; (b) cubic diamond; and (c)
hexagonal lonsdaleite. First published in [ 1521 and reproduced with permission.
diamond, and lonsdaleite are, however, well investigated and their structures are
shown in Fig. 3. In lonsdaleite, comprised of the same tetrahedral configuration
as in diamond, the planes of six-membered carbon rings are in the chair and boat
conformation and are stacked in an ABAB sequence while in diamond, these
planes exhibit the chair conformation exclusively and are arranged in an
ABCABC sequence. The recent discovery of the c 6 0 molecule shown in Fig. 4,
which is the most prominent representative of the fullerene family, led to a new
type of carbon allotrope. In fullerenes, the carbon atoms form spherical clusters
comprised of six- and five-membered rings of the general composition C20+2m
(m = 0 , 2 , 3 . ..) where m is the number of hexagons. The c 6 0 fullerene has been
predicted by Kroto and Smalley since 1985 [36-38] and was isolated first in
macroscopic quantities by Kratschmer et al. in 1990 [39-41]. At present, the
spherical carbon clusters and their derivatives are discussed with respect to their
application in material science [42,43]. Furthermore, a variety of new polymeric
carbon networks have been postulated to exhibit interesting materials properties
[41,44]. Recently, carbon nitrides such as the hypothetical compound C3N4 have
been also considered for the synthesis of ultrahard materials [ 14,151.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LII Introduction: Novel Ultrahard Materials
Figure 4. Molecular model of the fullerene C60.The gray carbon atoms illustrate one of the 12
pentagons present in C6,,. First published in [152] and reproduced with permission.
Diamond
Besides the aesthetic appearance of diamond in the form of gemstones, diamond is
the hardest material known, and is, therefore, on the top of the Mohs scale and has
the lowest molar volume (3.4cm3) of any material. Owing to this unique property,
diamond is used for many technological applications such as abrasion, cutting, and
polishing. Diamond also has the highest thermal conductivity (2000 W m-' K-') of
any material at room temperature, four times as high as the value of Cu or Ag, is a
good electrical insulator, it has a small dielectric constant and exhibits a high
electron/hole mobility. Boron doped diamond (Typ IIb) exhibiting a p-type con-
ductivity has been known for a long time. Production of diamond with a reasonable
n-type conductivity was one of the most difficult tasks in diamond synthesis.
Koizumi and coworkers recently obtained diamond films exhibiting n-type conduc-
tivity by using microwave enhanced plasma CVD with phosphine PH3 as a donor
source for homoepitaxial deposition of n-type diamond 1451. Therefore, diamond
will be a most important candidate material for future applications in electronic
devices. In addition, diamond has low coefficients of friction and thermal expansion,
high chemical and corrosive resistance towards most acids and oxidizing substances,
it is transparent to visible and infrared light, withstands ionizing radiation and can,
therefore, also be used as window or lens material or as a protective coating for this
application.
Conventionally, diamond is synthesized at high pressure and high temperature
according to the method developed by the General Electric Company in the 1950s
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Carbon-based Hard Materials LIII
[l, 461. Synthetic diamond is now commercialized and is utilized for cutting, grinding,
and polishing. Recent developments in the high-pressure synthesis technique allow
growth of diamond single crystals up to 25 carats in weight [47]. The high-pressure
synthesis of diamond is described in more detail by M. W. Bailey et al. in Part 111,
while hydrothermal synthesis is discussed in Part I1 by K. G. Nickel et al.
Since the mid-1980s the vapor deposition synthesis of diamond films has
attracted increasing scientific and industrial interest. At present, vapor-grown
diamond products are commercially viable as thin-film-coated cutting tools and
freestanding thick-film cutting tools, as substrates for thermal management
application and radiation detectors, as optical windows and for production of
high-range audio-speaker diaphragms. Commercialization of the vapor-grown
diamond became possible due to the cost reduction below $5/carat as a result of
the dramatic increase in growth rate and of improvements in energy-use efficiency
[481.
Advanced diamond films are synthesized using the gas-phase decomposition of
volatile carbon sources such as methane (CH4), acetone (H3C-CO-CH3),
carbon monoxide (CO), acetylene (HC-CH), or adamantane (CI0Hl6)and is
conducted under temperature and pressure conditions where graphite is the stable
polymorph (Fig. 5). The most important methods for producing CVD-diamond
under metastable conditions are:
- Microwave Plasma Assisted Chemical Vapor Deposition (PACVD) [49,50],
- Heated Filament Assisted Chemical Vapor Deposition (HFCVD) [5 13,
Figure 5. Pressure and temperature conditions of the diamond synthesis: (a) shock wave production
of diamond; (b) high temperature, high pressure regime for the synthesis of diamond; (c) catalytic
region for diamond formation; (d) chemical vapor deposited diamond; and (e) transformation of
CG0into diamond. The most recent review of the P, T phase diagram of carbon can be found
elsewhere [151].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LIV Introduction: Novel Ultrahard Materials
-
kinetic reasons according to the simplified reaction:
H2
CH4 Cdiamond f 2H2.
In a typical process, the hydrocarbon precursor containing more than 95% H2-gasis
passed through a plasma (700-1000°C) or over a heating filament ( T M 2000") at
less than atmospheric pressure. The reaction product, solid carbon, is then deposited
on a substrate heated at 800-1000°C and contains both graphite and diamond. In
the presence of H2-gas, the formation of unsaturated carbon nuclei and hence the
growth of graphite is suppressed by atomic hydrogen formed under these conditions
and is due to the reaction of the solid carbon with hydrogen radicals (H') giving
volatile hydrocarbons. The reaction rate of Cgraphite with H' is about 20 times as
high as the rate of diamond. Therefore, the formation of diamond is promoted.
Growth rates of up to 0.9 mm h-' [55] or 20 carats per hour [48] have been achieved
on a variety of substrates making the CVD-diamond a highly interesting material
for technological applications.
Diamond films have also been deposited from hydrogen-free gas mixtures such
as C60/Ar using microwave apparatus. The deposition rates are, however,
significantly lower. Additionally, use of hydrogen-poor plasmas results in nanocrys-
talline (3-10nm) diamond films in contrast to micrometer sized crystals from the
hydrogen-rich plasmas [56].
There are continuous theoretical attempts to describe the mechanism of CVD-
diamond synthesis including mechanisms of surface reactions, diamond nucleation,
and film growth. To achieve this aim various phenomenological or first-principles
models, molecular dynamics and Monte Carlo simulations have been used [57,58].
While the dominant substrate for low pressure growth of diamond films is single
crystal silicon, several different materials such as Ta, Mo, W, Cu, Au, Ni, Sic, SO2,
and Si3N4have been coated with polycrystalline diamond or diamond-like deposits.
The nucleation rates and the adhesion of the generated films vary with the type of
substrate material, which is related to the formation of intermediate carbide layers
such as Sic or WC.
Identification of vapor grown diamond is accomplished by X-ray diffraction or
electron diffraction using a transmission electron microscope (TEM) and by
Raman spectroscopy. Diamond coatings show the characteristic first-order Raman
peak for diamond at 1332cm-' [59]. In many cases, an additional peak located at
~ 1 5 5cm-'
0 occurs in the Raman spectrum which is attributed to a highly disordered
carbon phase (diamond-likecarbon) [60]. This diamond-like carbon contains sp3 and
sp2 hybridized C-atoms which do not coincide with graphite.
Recently a new simple method for the synthesis of diamond, called metallic
reduction-pyrolysis-catalysis, was reported. In this method diamond powder was
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Carbon-based Hard Materials LV
Another class of novel carbon material is amorphic carbon which can be obtained
by laser ablation [70]. This amorphous phase contains sp3 -bonded nodules of pure
carbon embedded in a matrix of other types of carbon phases. The mean particle size
of the nodules is 20 nm, the density of the films is between 1.85 and 2.89 g crnp3.The
internal stresses derived by distortions of the bonding angles in random networks
are counterbalanced by the high surface to volume ratios of the nanoparticles. Sub-
strate discs 30mm in diameter were coated with uniform layers of amorphous
carbon by laser ablation of a graphite feedstock. Maximum growth rates of about
0.5 pm h-' and film thicknesses of up to 5 ym have been achieved on different sub-
strate materials such as Si, Ge, ZnS, Cu, stainless steel, quartz, glass, and plastics.
One outstanding property of laser ablation in comparison with the CVD process
for the preparation of carbon films is that the substrate temperature does not
exceed 35°C during the whole deposition procedure. In contrast to the CVD dia-
mond, which develops polycrystalline columnar structures, the laser-ablated amor-
phous carbon is deposited in the form of self-seeding nodules. Diamond-like carbon
films are discussed in more detail in Part I11 by C.-P. Klages and K. Bewilogua.
Figure 6. (a) Model of a section of the hypothetical polymeric network of the body centered
tetragonal structure (b.c.t.4) of carbon and BN suggested by R. Hoffmann et al. [72]. (b) Model
of a section of the hypothetical hexagonal structure (H-6) of carbon postulated by M. Tamor
and K. Hass [73,74]. In both structures, each carbon atom is considered to be trigonally coordinated
and sp2-bonded, exclusively. First published in [ 1521 and reproduced with permission.
(iv) The carbon skeleton of allene, 2HC=C=CH2, could also be utilized for the
formation of a three-dimensional carbon phase [41]. In this case, the oxidative
coupling of allene could provide a carbon structure with orthogonally arranged
chains of sp2 carbon atoms (Fig. 8). A similar carbon phase could be obtained
by the coupling of tetraethynylallene, (HCGC)~C=C=C(C=CH)~[75].Here,
the carbon chains are connected by butadiene units instead of single bonds.
Both the b.c.t.-4 and the H-6 phase of carbon are discussed in terms of intermediate
structures formed during the chemical vapor deposition of diamond. Furthermore,
these types of structures could play an important role in diamond-like phases. The
calculations of the structural properties of the b.c.t.-4 and H-6 carbon phase using
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LVIII Introduction: Novel Ultrahard Materials
Figure 7. Hypothetical super-diamond structure suggested earlier [41,75]. First published in [ I 521
and reproduced with permission.
Figure 8. Hypothetical allene structure of carbon postulated earlier [41,75]. First published in [I521
and reproduced with permission.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Carbon-based Hard Materials LIX
Fullerenes
It has been shown that fullerenes or their derivatives can exhibit very interesting
chemical, electrical, magnetic, and mechanical properties. Besides, for example,
the superconductivity, which has been experimentally verified for the alkali metal
fullerides M 3 C 6 0 with M = K or Rb [77], fullerenes can serve as a starting material
for diamond synthesis [78,79] and may exhibit high hardness themselves under high
pressure conditions [go].
In 1992 M. Regueiro et al. reported on the transformation of c 6 0 molecules into
polycrystalline diamond by nonhydrostatic compression at room temperature [78]:
P > 20 GPa/20"C
& c60 ' Cdiarnond.
In contrast c 6 0 withstands hydrostatic pressure up to 20 GPa [81]. However, the
football molecules seem to be unstable towards uniaxial or shear stresses whereas
they are stable under isotropic stress where the spherical molecules are homo-
geneously deformed. In a dense arrangement of c 6 0 spheroids, 48 of the 60
carbon atoms have a quasi-tetrahedral coordination which is required in the
diamond structure. Only small structural rearrangements are then necessary for
the transformation into diamond [78].
R. S. Ruoff and A. L. Ruoff proposed that c 6 0 is stiffer than diamond [go]. This
result has been estimated from the calculated bulk modulus of individual c 6 0
molecules. The calculations revealed a bulk modulus B = 843 GPa which is nearly
twice the experimental value of 441 GPa determined for diamond. Since solid c 6 0
forms a van der Waals crystal with f.c.c. lattice and the distance between the
single c 6 0 molecules is about l.Onm, B would be relatively small under normal
conditions. However, when the individual carbon spheres are compressed until
they touch each other, the bulk modulus of the crystal would become in the range
of that of the molecule. Taking into account a volume filling factor of 74% valid for
the f.c.c. structure, B is calculated to be 624 GPa using the volume fraction rule [80].
Therefore, diamond might be outranked by the hardness of c 6 0 crystals. This result
is of great fundamental interest, despite the fact that the extraordinary hardness
value will be only achieved at high pressures (20 GPa).
(P31c), pseudocubic (P-42m), and cubic (I-43d) [13]. Bulk moduli (and hardness) of
most of these compounds were predicted to be comparable with and the last one
even to exceed that of diamond.
The hardness of the hypothetical carbon nitride, pC3N4, has been comprehen-
sively discussed by A. Liu and M. Cohen [12,83]. In the case of &N4, the bulk
modulus is estimated to be higher than that of diamond (483 GPa) by semi-empirical
methods using Eq. (4) with a bond length d = 0.147 nm and an ionicity factor of
X = 0.5. From ab initio calculations of the total structural energy of C3N4 using
the first principles pseudopotential total energy method, a somewhat lower value
of 427GPa for B is derived. This finding can be attributed to the structural
difference between this material and the completely tetrahedrally coordinated zinc
blende structure from which the scaling law (Eq. (4)) was developed [12,83]. In
contrast, the compound pC3N4 is considered to have the pSi3N4 structure and is
only a partially tetrahedrally coordinated structure. The pSi3N4structure is related
to the well known hexagonal structure of the beryllium silicate Be2Si04.Therefore,
the unit cell of PC3N4 is hexagonal and contains two formula units (Fig. 9). The
carbon atoms are sp3 hybridized and tetrahedrally surrounded by N atoms, whereas
nitrogen has a planar trigonal coordination (sp3 hybridization). The overall struc-
ture can be described by a network of CN4 tetrahedra and NC3 sites.
Teter and Hemley suggested the cubic high pressure phase of Zn2Si04denoted as
willemite-I1 as another possible structure of a superhard C3N4[13]. This suggestion
was based on the fact that the structure of the zero-pressure phase of Zn2Si04
(willemite-I) is isomorphous with the structure of Be2Si04. Using first principles
pseudopotential total energy techniques they calculated energetics, structure, and
some physical properties of the cubic willemite-I1 structure as well as of four
other candidate structures of C3N4. They found that aC3N4 (the analogue of
aSi3N4) is energetically favorable at zero-pressure and the cubic willemite-I1
structure might be a high pressure phase of C3N4. The predicted transition pressure
from a-to the cubic willemite-I1phase was 68 GPa. The calculated bulk modulus of
this cubic C3N4was 496 GPa, which is significantly higher than the measured bulk
modulus of diamond (442 GPa) or predicted moduli of a- and pC3N4 (425 GPa and
450 GPa, respectively). This result suggested that the hardness of cubic C3N4with
willemite-I1 structure may exceed that of diamond [13]. However, in his later
work Teter revised this conclusion to the contrary according to the almost linear
relation he has found between shear moduli of polycrystalline compounds and
their hardness [17]. The calculated shear moduli of the polycrystalline p- or cubic
C3N4 are only 60% of the shear modulus of diamond and smaller than that of
cubic BN. Accordingly the hardness of these carbon nitrides (once synthesized)
will probably not exceed that of cubic BN [17]. Further discussion of possible struc-
tures of C3N4can be found in Part I in the chapter of J. E. Lowther.
Synthesis of carbon nitrides has been attempted for example by the pyrolysis of
organic nitrogen-containing compounds such as melamine-formaldehyde resin
and subsequent shock compression of the obtained residues [85]. However, no tetra-
hedral carbon nitride could be detected in the reaction products. In a similar
approach, L. Maya et al. obtained a carbon nitride with trigonally coordinated
carbon by the pyrolysis of organic compounds at 700°C and 225 MPa pressure
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Carbon-based Hard Materials LXI
a)
Figure 9. Theoretically suggested crystal structures of carbon nitride C3N4(filled circles represent
the carbon atoms): (a) Hexagonal structure corresponding to that of PSi3N4or phenacite, Be2Si04.
(b) Cubic structure corresponding to that of willemite-11, the high-pressure phase of Zn2Si04[ 131.
First published in [ 1521 and reproduced with permission.
cubic zinc blende structure of the crystallites was derived from TEM measurements
and confirmed by extended energy loss fine structure analysis [90].
Besides the preparation of polycrystalline C- and N-containing thin films by using
direct current magnetron sputtering with nitrogen as the sputtering gas [ 141, the
synthesis of C3N4 was experimentally verified in the work of C. Niu et a/. [15]. In
this paper, the preparation of the carbon nitride was realized by pulsed laser
ablation of graphite targets in the presence of a nitrogen source. The C-N films
deposited on Si (001) and on polycrystalline Ni were characterized by Rutherford
backscattering (RBS). photoelectron spectroscopy (XPS, ESCA), and electron
diffraction measurements in the TEM mode. The analysis of the diffraction data
(convergent beam electron diffraction) clearly showed that the reflections were in
accord with that expected for the PC3N4 structure. Furthermore, the spectroscopic
analyses revealed the presence of C and N in the deposited films. The bonding in
these solids is covalent. The analysis of the chemical composition of the films deter-
mined by RBS exhibited nitrogen contents of up to 45% which is the highest amount
measured in any binary C-N solids to that time. The nitrogen content did not
decrease after thermal annealing in flowing N2 up to 800°C indicating the relative
stability of the C-N material. Preliminary investigations of the hardness showed
that the films could not be scratched by hard metals [15].
Thin carbon nitride films containing small crystals were deposited by Yu e t a / . on
Si wafers in the (100) orientation using R F diode sputtering of graphite target with
pure N2 [91]. The films were about 1 pm thick with a nominal composition of
Co6N0.300.1 and contained polycrystalline grains of 0.5-1 pm in size with higher
nitrogen content. Many of the lattice spacings obtained from the ring patterns
measured with TEM matched those of the theoretically predicted PC3N4 [91].
Analogous results were reported by Fernandez et al. who used a dual ion beam
sputtering system to deposit C-N films with a nitrogen content of about 44% on
KCI substrates [92].
Preparation of crystalline carbon nitrides was reported to be possible by use of
chemical vapor deposition techniques. Films containing a-,PC3N4 and an unknown
C-N phase were obtained on nickel substrates from the gas mixture of nitrogen and
methane in a hot-filament CVD apparatus [93]. The composition of the grown films
close to the stoichiometric value N / C = I .33 was measured using energy-dispersive
X-ray analysis and Auger electron spectroscopy. The lattice constants of synthesized
SI- and (jC3N4agreed with the theoretically predicted values. The authors have not
observed bands of diamond or graphite in Raman spectra of the synthesized films.
Presence of amorphous C-N compounds was ruled out after analysis of the films’
morphology [93]. Plasma enhanced CDV process was applied by He and Chang for
synthesis of PC3N4 on a silicon wafer as a substrate [94]. As a precursor they used a
mixture of methane, hydrogen, and nitrogen. The nitrogen concentration was varied
between 0 and almost 75%. The strongest evidence for synthesis of BC3N4 at the
highest N2 content was obtained from the TEM measurements. The lattice spacings
calculated from the powder diffraction rings agreed with the theoretically predicted d-
spacings of PC3N4. However, contamination of the deposited carbon nitride by
iron carbide or nitride most probably in amorphous form and possible presence of
amorphous CH, double bounds C=C or C=N was also reported [94].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Carbon-based Hard Materials LXIII
(i) The samples contain many phases often with different chemical composition.
(ii) Crystalline carbon nitrides have been obtained in small amounts and mostly X-
ray or electron powder diffraction data were presented. Such data do not
always allow unambiguous determination of the structure especially when
other phases were not identified.
(iii) Small sample amounts and presence of other phases is a serious hindrance for
accurate determination of chemical composition of the obtained crystalline
carbon nitrides.
Improvements in the synthesis techniques and analytic equipment may help to solve
these problems. It also has to be mentioned here that the stability of some of the
predicted carbon nitrides was questioned in a number of studies based on chemical
and thermodynamical grounds. A recent discussion of this subject and further
references can be found elsewhere [ 171.
Boron Nitrides
Boron nitride is isoelectronic with elemental carbon and is, therefore, found in
similar crystal structures:
- the hexagonal (hBN) [lo41
- the cubic zinc blende (cBN) [2,3]
- the wurtzitic (wBN) [lo51
- the rhombohedral (rBN) [lo61 structure
In contrast to rhombohedral carbon, rBN has been synthesized in significant
amounts by several methods and is well investigated [107,108]. In contrast to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Boron-based Hard Materials LXV
Figure 10. P,Tphase diagram of boron nitride: (a) catalytic conversion of hBN into cBN; (b) meta-
stable region for the vapor phase formation of cBN. An alternative phase diagram of BN suggested
previously [I 1I] is shown by dashed lines.
graphite, the hexagons of hBN are arranged on top of each other, while the
structures of cBN and wBN completely correspond to diamond and lonsdaleite,
respectively. The zinc blende polymorph of BN is also known as sphalerite or
PBN. Cubic boron nitride is the second hardest material and is produced conven-
tionally at high pressures and temperatures, analogous to those methods developed
for the fabrication of synthetic diamond. It has been commercially available since
1969 [loo]. Hot-pressed monolithic and polycrystalline BN, comprised of both the
cubic and the wurtzitic polymorphs, is already used in cutting tool devices [log].
The generally accepted P,T phase diagram of BN shown in Fig. 10 is based on the
work of Wentorf and colleagues, in which melting temperatures of hBN and the
phase boundary h B N - + c B N at high pressures were measured. This phase
boundary extrapolated to room temperature by analogy with the carbon phase
diagram intersects the pressure axis [110], thus implying that hBN is the most
stable modification at ambient conditions. In contrast, it was suggested in some
recent publications [l 1 I] that cBN is stable at ambient pressure at temperatures
up to 1300°C. This suggestion is based on measurements of the thermodynamic
properties of all modifications of BN, where the difference of enthalpies of forma-
tion of cubic and hexagonal BN was found to be negative [l l l]:
AHFBN(298K)- AHfhBN(298K)= -16kJmol-'.
Using these thermodynamic data an alternative phase diagram of BN (also shown in
Fig. 10) was developed.
The technological applications of cBN are similar to those of diamond. However,
because of improved oxidation resistance, cBN allows high-speed cutting or
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXVI Introduction: Novel Ultrahard Materials
x>l
Figure 11. Model of the unit cell of the B4N structure. The atoms marked in black represent the
nitrogen atoms within the N-B-N-chain. First published in [152] and reproduced with permission.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXVIII Introduction: Novel Ultrahard Materials
structural feature, the new BN phase can be expected to exhibit interesting hard-
ness properties.
r! H \ I \ /
A
I c-c,
H
C N - i - H
H
-110 OC
giving boron carbonitrides, B,C,N2, with different carbon contents. The following
simplified equations represent the overall pyrolysis reactions providing BC4N and
BC2N (Fig. l), respectively:
1050"C/Ar
CN-B.3 BC4N + CH4 + 2H2,
n 1050"C/Ar
NH N-BH3 BC2N + 2CH4 + NH3 + H2.
U
Powder diffraction as well as photoelectron spectroscopy (XPS), electron energy
loss spectroscopy (EELS), and solid state NMR investigations of the synthesized
BC2N and BC4N revealed a graphite-like sheet structure with carbon substituted
for boron and nitrogen and with little sheet-to-sheet registry (turbostratic boron
carbonitride). The pyrolytic conversion of pyridine-borane to turbostratic B-C-N
via the formation of an intermediate polymeric compound is shown schematically
in Fig. 12. These boron carbonitride materials can be considered as solid solutions
of C and BN and are thermally stable up to 1800°C in an Ar atmosphere. At higher
temperatures, phase separation providing crystalline hBN, B4C, and elemental C is
observed (see Fig. 1) [ 1331. Microstructural investigation of the synthesized BC4N
by means of TEM showed also the presence of small amounts of isolated crystals
with cubic symmetry as determined by electron diffraction data [ 132,1341. The dif-
fraction data of the cubic crystals do not correspond to those of diamond or cBN.
Therefore, a cubic boron carbonitride phase is assumed for those single crystals
embedded in a turbostratic BC4N matrix. The most difficult issue in the synthesis
of hexagonal B,C,NZ compounds is proving of a single phase content. Indeed
there is no direct method of distinguishing between hexagonal B. C ,Nz, hBN and
..J
graphite, if all are turbostratically distorted. Andreev and Lundstrom pointed out
that in some of the above described work the available data are insufficient to
make unambiguous conclusions about single phase content [1351. They prepared
hexagonal B,C,N, materials by nitration of boron carbide B4C powder at 2100°C
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXX Introduction: Novel Ultrahard Materials
at different partial nitrogen pressures and found in all cases at least two coexisting
phases. For example, samples prepared at partial nitrogen pressures between 30 and
100% contained a carbon-poor three-dimensionally ordered phase and a carbon-
rich turbostratic phase [135]. Synthesis of ceramic materials from polymers is dis-
cussed in Part I1 by G. D. Soraru and P. Colombo.
The high-pressure and high-temperature formation of crystalline boron carbo-
nitride with a cubic lattice was reported in 1977 by various Russian authors using
hBN + C, hBN + B4C, or cBN + C as the reacting species [136,137]. High micro-
hardness values in the range between 60 and 80 GPa have been measured for these
cubic phase E C - N materials. High-pressure synthesis of the cubic B,C,N, phases
from the starting graphite-like (BN),C1-,x (with x = 0.15, 0.26 and 0.6) prepared
by CVD-method was performed at 14GPa and above 3300K [124]. Although
detailed chemical analysis was not performed, it was assumed that the obtained
material was a solid solution of cubic BN and diamond. X-ray powder diffraction
data suggested that this cubic material exhibits ideal mixing according to the Vegard’s
law and has the same composition as the starting hexagonal material. However, the
author also observed reflexes allowed only for cBN, which he explained by ordering of
the B-N pairs in the lattice of the synthesized cubic (BN),C1 -, materials. Addition-
ally, the author always observed the presence of graphite-like phases in a heat-treated
charge and small amounts of B4C in a few samples [124]. Ideal mixing in the cubic
(BN),C, compounds synthesized in this way was also reported 15 years later by
Kagi et al. [138]. In contrast, measurement of molar volume and of the bulk modulus
of cubic (BN),C1 --x phases synthesized in the YAG-laser heated diamond cell at pres-
sures of 30-50 GPa and temperatures of 2000-2500 K showed strong deviation from
ideal mixing [139]. According to these studies the molar volumes of cubic (BN),yCl
at ambient conditions are larger than that of the ideal solid solution and cubic
(BN),C1_, phases with x < 0.5 have a bulk modulus smaller than that of end-
member cBN [139]. If the first observation holds at pressures and temperatures of
synthesis then a mechanical mixture of diamond and cBN will appear at equilibrium
conditions. This is because high pressures favor formation of denser materials [ 1401.
It follows that this cubic (BN),C1 --x should be a metastable intermediate product of
transformation from hexagonal (BN),C, -, to the mixture of diamond and cBN.
The second observation suggests that the hardness of cubic (BN),C, phases with
x < 0.5 should be smaller than that of cBN, if the systematic approach of Liu
and Cohen [12] is applied. The conclusion of nonideal mixing made earlier [139]
can be called in question because hexagonal materials did not transform completely
into the cubic phases and quantitative chemical analysis of the synthesized cubic
phases was not performed. Nevertheless, metastability of cubic (BN),CI is
supported by the following results of Japanese authors. Sasaki et al. observed
crystallization of hexagonal BC2N in diamond and cBN in the presence of a Co
catalyst at relatively low pressures and temperatures (5.5 GPa and 1400-1600 K)
[141]. Nakano found that upon phase separation of metastable cubic BC2N to
diamond and cBN the composition of residual cubic (BN),C1 -, did not change
(x = 0.33). Accordingly, no other stable phases of type (BN),C, except diamond
and cBN exist at the temperature and pressure conditions of the work (7.7 GPa,
2300-2400°C) [142].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Silicon-based Materials LXXI
Boron Suboxides
In 1988, A. R. Badzian reported that boron suboxides exhibit exceptional high
hardness values [144]. The suboxides with compositions between B 2 0 and B220
were synthesized by the solid state reaction of elemental boron with B2O3 at
1600-2000°C. Subsequent melting of the reaction product under Ar provides
boron containing 4-5 at.% oxygen corresponding to B I 8 0and BZ20,respectively.
The crystal structure of the boron rich boron suboxides is related to that of a-
rhombohedra1 boron. The microhardness of boron suboxides was determined to
be between 31 GPa and 59 GPa, which is rather close to the hardness of cubic BN
[144, 1451.
Application of high pressure and high temperature allows synthesis of boron
suboxides with the high degree of crystallinity needed for studying the properties
of these compounds. For example, only fine powders with some secondary
amorphous products were obtained in the synthesis of boron suboxide with a
nominal composition B 6 0 at ambient pressure. Additionally, the synthesized
compounds were nonstoichiometric with compositions varying from B600,72to
B600.86.Hubert et al. showed that by reduction of B2O3 by boron at 1700°C and
5.5 GPa a highly crystalline material consisting of multiply twinned particles with
icosahedral symmetry can be obtained. The size of these icosahedral particles was
up to 20-30pm. The measured molar oxygen content was 0.96 f 0.08 (B600.96)
thus approximating the ideal value of 1 [146,147].
The a-rhombohedra1 structure of elemental boron has an electron deficiency that
can be reduced by the incorporation of interstitial atoms in structural voids. Filling
these voids with atoms can result in cross-linking of the boron icosahedra which is
realized for example in the structures of B4C (B12C3), B12P2, B4N, or boron rich
boron suboxides (see Fig. 11) and can contribute to a significant increase of the
rigidity of the solid state structure.
Silicon-based Materials
This section deals with high pressure phases of two silicon based compounds Si02
and Si3N4,which previously did not belong to the family of superhard materials.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXXII Introduction: Novel Ultrahard Materials
While the a- and p-phases of silicon nitride were known for a long time as hard
materials, the common phases of Si02 (e.g. quartz) lie much further down the
hardness scale (Table 1). The high pressure phase of Si02 with the rutile structure
(stishovite) was first reported in 1961 [148]. Si02-stishovite is stable at pressures
above 11 GPa and persists metastably at ambient conditions. Its density
(4.287gcmP3) is almost 62% higher than that of the ambient pressure quartz
phase (2.648 g ~ m - [149].
~ ) The bulk modulus of Si02-stishovite (298 GPa) exceeds
that of B4C (200 GPa) which hardness is comparable with that of cubic BN 141. Such
a strong increase in density and bulk modulus is due to an increase of the co-
ordination number of Si-atoms from four to six. Recently LCger et al. measured
the Knoop hardness of polycrystalline single phase Si02-stishovite to be between
31 and 35GPa, which exceeds the literature values of 30GPa for B4C and B 6 0
[4]. However, one has to bear in mind that the measured hardness values depend
strongly on the state of the sample and on the specific test method used. Thus to
make reliable comparisons hardnesses of the materials have to be measured using
the same method and taking into account the above mentioned peculiarities.
A change to six-fold coordination of Si-atoms occurs upon transformation of
hexagonal c1- and pSi3N4 to the high pressure cubic phase with spinel structure
observed in the laser heated diamond cell at 15GPa [5]. The technique of C02-
laser heating in a diamond cell is described in Part I by Zerr et al. The density of
the cubic Si3N4is 23% above that of a- or p-phases. According to first-principles
calculations, the elastic bulk and shear moduli of cSi3N4exceed that of Si02-stisho-
vite. Consequently, the hardness of the cubic silicon nitride can be comparable to
that of Si02-stishovite [5]. Detailed discussion of silicon nitride based materials is
given in Part 111 by M. Herrmann.
These results on the high pressure phases of two silicon based materials show a
strong potential of the high pressure techniques in synthesis of new materials with
promising mechanical properties. Furthermore, there are many unexplored paths
for future work, even if some of these ways seem on first glance to be completely
worked out, as for the case of Si02-stishovite.
Concluding Remarks
The experimental and theoretical research activities in the field of materials with hard-
ness comparable to or higher than diamond have been stimulated in recent years by the
discovery and postulation of novel carbon allotropes and novel carbon- and boron-
based materials. Comparison of the experimental and theoretical results obtained in
the past shows that first principles quantum mechanical and semi-empirical calcula-
tions are very useful for the screening of the structural and electronic properties of
new materials. The fascinating variety of new possible carbon allotropes which have
been calculated to be stable relative to diamond or can be stabilized for kinetic reasons
reflects the high potential and the topicality of modern carbon chemistry. Novel
carbon allotropes and carbon and silicon nitrides are subjects of discussion in Part I
in the chapter written by G. Jungnickel et al. and J. E. Lowther.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
However, the experimental evidence for the synthesis and structure of dense
superhard C3N4 phases reported in the literature is not unambiguous. Thus,
future research should be focused on the synthesis of sufficient amounts of these
new superhard materials and on their structural characterization. The structures
of all C3N4-polymorphsreported so far still remain to be verified by single crystal
X-ray or electron diffraction. Furthermore, the difficulties in the determination of
reliable hardness values of these new materials not only at room temperature but also
at elevated temperatures have to be overcome. Diamond will not lose its reputation
as the hardest material until the newly developed phases and compounds have been
characterized in detail and can be produced in reasonable amounts.
Acknowledgement
The authors thank J. Nicolich for preparation of crystal structure plots. This work
was supported by the Deutsche Forschungsgemeinschaft under the contract number
Ri 510/14-1.
References
1. F. P. Bundy, H. T. Hall, H. M. Strong, and R. J. Wentorf, Nature, 1955, 176, 51.
2. R. H. Wentorf, J . Chem. Phys. 1957, 26, 956.
3. R. H. Wentorf, J . Chem. Phys. 1961, 34, 809.
4. J. M. Leger, J. Haines, M. Schmidt, J. P. Petitet, A. S. Pereira, and J. A. H. da Jornada,
Nature, 1996, 383, 401.
5. A. Zerr, G. Miehe, G. Serghiou, M. Schwarz, E. Kroke, R. Riedel, H. FueR, P. Kroll, and R.
Boehler, Nature, 1999, 400, 340.
6. J. C. Angus and C. C. Hayman, Science, 1988, 241, 913.
7. K. E. Spear, J . Am. Ceram. SOC.1989, 72, 171.
8. W. A. Yarbrough and R. Messier, Science, 1990, 247, 688.
9. P. K. Bachmann, Adv. Muter. 1990, 2, 195.
10. P. K. Bachmann, Keramische Zeitschrift, 1992, 2, 93.
11. R. Telle, in Materials Science and Technology: A Comprehensive Treatment, R. W. Cahn, P.
Haasen, and E. J. Kramer (Series Eds), Vol. 11, N. V. Swain (Ed.), VCH, Weinheim, 1994,
p. 173.
12. A. Y. Liu and M. L. Cohen, Science, 1989, 245, 841.
13. D. M. Teter and R. J. Hemley, Nature, 1996, 271, 53.
14. J. Hoggin, Chem. Eng. News, May 25, 1992, 20.
15. C. Niu, Y. Z. Lu, and C. M. Lieber, Science, 1993, 261, 334.
16. A. J. Stevens, T. Koga, C. B. Agee, M. J. Aziz, and C. M. Lieber, J. Am. Chem. SOC.1996,118,
10900.
17. D. M. Teter, M R S Bull. 1998, 23, 22.
18. Y. Li, Y . Qian, H. Liao, Y. Ding, L. Yang, C. Xu, F. Li, and G. Zhou, Science, 1998,281,246.
19. U. Helmersson, S. Todorova, S. A. Barnett, J.-E. Sundgren, L. C. Markert, and J. E. Greene,
J. Appl. Phys. 1987, 62, 481.
20. I-W. Chen and A. Rosenflanz, Nature, 1997,389,701.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXXIV Introduction: Novel Ultrahard Materials
21. S. VepTek, P. Nesladek, A. Niederhofer, F. Glatz, M. Jilek, and M. Sima, Surf. Coat. Technol.
1998, 108-109, 138.
22. H. Holleck, J. Vac. Sci. Technol. 1986, A4, 2661.
23. K.KuBmaul, in Lexikon Werkstoftechnik, H. Grafen (Ed.), VDI, Dusseldorf, 1991, p. 409.
24. K. Niihara, and T. Hirai, J. Muter. Sci. Lett. 1978, 13, 2276.
25. R. W. Rice, C. Cm. Wu, and F. Borchelt, J . Am. Ceram. Soc. 1994,77, 2539.
26. M. L. Cohen, J. Hard Muter. 1991, 2, 13.
27. B. Farid and R. W. Godby, Phys. Rev. B, 1991,43, 14248.
28. R. W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, 3rd edn,
Wiley, New York, 1989.
29. A. Kelly and N. H. Macmillan, Strong Solids, 3rd edn, Clarendon Press, Oxford, 1986.
30. E. Knittle, R. M. Wentzcovitch, R. Jeanloz, and M. L. Cohen, Nature, 1989, 337, 349.
3 1. H. B. Huntington, in Solid State Physics: Advances in Research and Applications, F. Seitz and
D. Turnbull (Eds), Vol. 7, Academic Press, New York, 1958, p. 213.
32. 0. L. Anderson, in Physical Acoustics: Principles and Methods, W. P. Mason (Ed.), Vol. 111,
Part B, Academic Press, New York, 1965, p. 43.
33. R. E. Hanneman, H. M. Strong, and F. P. Bundy, Science, 1967, 155, 995.
34. H. Strunz, Mineralogische Tabellen, 7th edn, Akademische Verlagsgesellschaft Geest & Portig
K.-G., Leipzig, 1978.
35. A. Neuhaus, Proc. Intern. Mineralog. Assoc. 1960, 1, 33.
36. H. W. Kroto, J. R. Heath, S. C. O'Brien, R. F. Curl, and R. E. Smalley, Nature, 1985,318, 162.
37. H. W. Kroto, Science, 1988,242, 1139.
38. R. F. Curl and R. E. Smalley, Science, 1988, 242, 1017.
39. W. Kratschmer, L. D. Lamb, K. Fostiropoulos, and D. R. Huffman, Nature, 1990,347, 354.
40. H. W. Kroto, Angew. Chem. 1992, 104, 113.
41. F. Diederich and Y. Rubin, Angew. Chem. 1992, 104, 1123.
42. R. F. Curl and R. E. Smalley, Spektrum der Wissenschaft, 1991, 88.
43. W. Kratschmer and K. Fostiropoulos, Physik in unserer Zeit, 1992, 3, 105.
44. H. R. Karfunkel and T. Dressler, J . Am. Chem. Soc. 1992, 114, 2285.
45. S. Koizumi, M. Kamo, Y. Sato, H. Ozaki, and T. Inuzuka, Appl. Phys. Lett. 1997,71, 1065.
46. P. W. Bridgman, Sci. Am. 1955, 193, 42.
47. R. C. Burns, S. Kessler, M. Subanda, C. M. Welbourn, and D. L. Welch, in Proceedings of the
3rd NIRIM International Symposium on Advanced Materials ( I S A M '96), M. Akaishi et al.
(Eds), NIRIM, Tsukuba, 1996, p. 105.
48. J. E. Butler and H. Windischmann, M R S Bull. 1998, 23, 22.
49. M. Kamo, Y. Sato, K. S. Matsumoto, and N. Setaka, J . Cryst. Growth, 1983, 62, 642.
50. P. K. Bachmann, W. Drawl, D. Knight, R. Weimer, and R. F. Messier, in Diamond and
Diamond-Like Materials Synthesis, G. H. Johnson, A. R. Badzian, and M. W. Geis (Eds),
M R S Symp. Proc., Vol. EA-15, Materials Research Society 1988, 99.
51. S. Matsumoto, Y. Sato, M. Kamo, and N. Setaka, Jpn. J . Appl. Phys. 1982, 21, L183.
52. S. Matsumoto, J. Muter. Sci. Lett. 1985, 4, 600.
53. K. Kurihara, K. Sasaki, M. Kawarada, and N. Koshino, Appl. Phys. Lett. 1988, 6,437.
54. Y. Hirose and N. Kondo, in 35th Spring Meeting of the Japanese Appl. Phys. Society, 1988, p.
434.
55. N. Ohtake, H. Tokura, Y. Kuriyama, Y. Mashimo, and M. Yoshikawa, in Proc. 1st Int. Symp.
Diamond and Diamond-Like Films, J. P. Dismukes et a/. (Eds), The Electrochemical Society,
Pennington, NJ, Proceedings Vol. 89-12, 1989, p. 50.
56. D. M. Gruen, M R S Bull. 1998, 23, 32.
57. D. W. Brenner and J. H. Harrison, Am. Ceram. Soc. Bull. 1992,71, 1821.
58. A. M. Stoneham, I. J. Ford, and P. R. Chalker, MRS Bull. 1998, 23, 28.
59. S. A. Solin and A. K. Ramdas, Phys. Rev. B, 1970, 1, 1687.
60. C. Wild, N. Herres, J. Wagner, P. Koidl, and T. R. Anthony, in Proc. 1st Int. Symp. on
Diamond and Diamond-Like Films, J. P. Dismukes et al. (Eds), The Electrochemical Society,
Pennington, NJ, Proceedings Vol. 89-12, 1989, p. 283.
61. L. Holland, J . Vac. Sci. Technol. 1977, 14, 5.
62. L. Holland, Thin Solid Films, 1979, 58, 107.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References LXXV
63. 0. S. Selwyn and E. Kay, Plasma Chem. Plasma. Process, 1985, 5, 1983.
64. M. Ramsteiner and J. Wagner, Appl. Phys. Lett. 1987, 51, 1335.
65. K. Kobayashi, N. Mutsukara, and Y. Machi, J . Appl. Phys. 1986, 59, 910.
66. F. Jansen, M. Machonkin, S. Kaplan, and S. Hark, J. Vuc. Sci. Technol. 1985, A3, 605.
67. J. W. Rabelais and S. Kasi, Science, 1988, 239, 623.
68. N. Savvides, J . Appl. Phys. 1985, 58, 518.
69. J. C. Phillips, Phys. Rev. Lett. 1979, 42, 153.
70. C. B. Collins, F. Davanloo, T. J. Lee, J. H. You, and H. Park, Am. Cerum. Soc. Bull. 1992,71,
1535.
71. N. N. Matyusenko and V. E. Strel'nitzkii, J . Exp. Theor. Phys. Lett. 1979, 30, 199.
72. R. Hoffmann, T. Hughbanks, M. Kertesz, and P. H. Bird, J . Am. Chem. Soc. 1983,105,4831.
73. M. A. Tamor and K. C. Hass, J . Mater. Res. 1990, 5, 2273.
74. A. Y. Liu, M. L. Cohen. K. C. Hass, and M. A. Tamor, Phys. Rev. B, 1991,43, 6742.
75. A. H. Alberts and H. Wynberg, J. Chem. Soc., Chem. Commun. 1988, 748.
76. A. Y. Liu and M. L. Cohen, Phys. Rev. B, 1992,454579.
77. A. F. Hebard, M. J. Rosseinsky, R. C. Haddon, D. W. Murphy, S. T. Glarum, T. T. M. Palstra,
A. P. Ramirez, and A. R. Kortan, Nature, 1991, 350, 600.
78. M. N. Regueiro, P. Monceau, and J.-L. Hodeau. Nature, 1992, 355, 237.
79. M. N. Regueiro, Adv. Mafer. 1992, 4, 438.
80. R. S. Ruoff and A. L. Ruoff, Nature, 1991, 350, 663.
81. S. J. Duclos, K. Brister, R. C. Haddon, A. R. Kortan, and F. A. Thiel, Nature, 1991, 351,
380.
82. M. L. Cohen, Phys. Rev. B, 1985, 32, 7988.
83. A. Y. Liu and M. L. Cohen, Phys. Rev. B, 1990,41, 10727.
84. A. Y. Liu and R. M. Wetzcovitsch, Phys. Rev. B, 1994, 50, 10362.
85. M. R. Wixom, J . Am. Cerum. Soc. 1990, 73, 1973.
86. L. Maya, D. R. Cole, and E. W. Hagaman, J . Am. Ceram. Soc. 1991, 74, 1686.
87. H. X. Han and B. J. Feldman, Solid State Commun. 1988. 65, 921.
88. T. Sekine, H. Kanda, Y. Bando, M. Yokoyama, and K. Hojou, J . Muter. Sci. Lett. 1990, 9,
1376.
89. J. Kouvetakis, R. B. Kaner, M. L. Sattler, and N. Bartlett, J . Chem. Soc. Commun. 1986, 1758.
90. J. Martin-Gil, F. J. Martin-Gil, M. Sarikaya, M. Qian, M. Jose-Yacaman, and A. Rubio,
J. Appl. Phys. 1997,81, 2555.
91. K. M. Yu, M. L. Cohen, E. E. Haller, W. L. Hansen, A. Y. Liu, and I. C. Wu, Phys. Rev. B,
1994,49, 5034.
92. A. Fernandez, P. Prieto, C. Quiros, J. M. Sanz, J.-M. Martin, and B. Vacher, Appl. Phys. Lett.
1996, 69, 764.
93. Y. Chen, L. Guo, and E. G. Wang, Phil. Mag. Lett. 1997, 75, 155.
94. J. L. He and W. L. Chang, Surf. Coat. Technol. 1998,99, 184.
95. J. H. Nguyen and R. Jeanloz, Muter. Sci.Eng. A, 1996, 209, 23.
96. D. C. Nesting and J. V. Badding, Chem. Muter. 1996, 8, 1535.
97. D. C. Nesting, J. Kouvetakis, and J. V. Badding, in Proceedings ofthe 5th NIRIM International
Symposium on Advanced Materials (ISAM '98), M. Kamo et al. (Eds), NIRIM, Tsukuba,
Japan, 1998, 1.
98. A. J. Stevens, T. Koga, C. B. Agee, M. J. Aziz, and C. M. Lieber, J . Am. Chem. Soc. 1996,118,
10900.
99. V. P. Dymont, E. M. Nekrashevich, and I. M. Starchenko, JETP Lett. 1998,68,498.
100. L. Vel, G. Demazeau, and J. Etourneau, Mater. Sci. Eng. 1991, B10, 149.
101. B. P. Singh, Muter. Res. Bull. 1986, 21, 85.
102. F. Binder, Radex-Rundschau, 1975, 4, 531.
103. F. Thevenot, J . Eur. Cerum. Soc. 1990, 6, 205.
104. A. Lipp, K. A. Schwetz, and K. Hunold, J . Eur. Serum. Soc. 1989, 5, 3.
105. R. R. Wills, Int. J . High Technol. Serum. 1985, 1, 139.
106. T. Ishii, T. Sato, Y. Sekikawa, and J. M. Iwata, J . Cryst. Growth, 1981, 52, 285.
107. T. Matsuda, N. Uno, H. Nakae, and T. Hirai, J . Muter. Sci. 1986, 21, 649.
108. V. L. Solozhenko, V. A. Mukhanov, and N. V. Novikov, Dokludy Chem. 1990,312, 125.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LXXVI Introduction: Novel Ultrahard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
1.1 Introduction
Most hard materials are formed from high-melting elements in the middle of the
periods of the transition metals and the main group elements. The stability of these
elements and compounds arises from a maximum of filled bonding bands and a mini-
mum of filled antibonding bands. The hardness of many elements and materials is
frequently not well known, but usually hardness and high melting points correlate,
and melting points are better known. Figure 1 shows the melting points of the
elements and some well-known high-melting and hard materials. At this point it is
worthwhile to note that refractory carbides like HfC or TaC have significantly
higher melting points than the frequently used crucible ceramic alumina, A1203.
Most well-known hard materials have surprisingly simple crystal structures. Here
we give an overview of the crystal chemistry of hard materials. Different aspects of
the various crystal structures are discussed, such as close-packed structures with
filled octahedral voids like titanium carbide, the tetrahedral arrangements in Sic
or the three-dimensional inorganic networks in a- and P-Si3N4. Also we briefly
mention the synthesis and some applications of various hard materials.
The basic crystallographic data of the most important hard materials are listed in
Table 1. Further access to crystallographic data can be gained through various com-
pilations. The older literature has critically been evaluated in the Strukturberichte
[ 13, the Structure Reports [2] and in the Landolt-Bornstein [3]. The non-oxidic
materials are well covered in several editions of Pearson’s Handbook [MI.The
various structure types of these materials have been critically compiled by Parthe
and coworkers [7]. Access to the more recent literature is best achieved through
4000 4000
- TaC
. =- ZrC
-* 3000
Figure 1. Melting points of most elements and some selected high-melting inorganic materials.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 1 Structural Chemistry of Hard Materials
the ICSD data base [8]. The X-ray powder patterns of most solid state materials are
stored in the ICDD files [9]. Knowing the crystal structure they can also be
generated by the program LAZY PULVERIX [lo].
Figure 2. Crystal structures of cubic diamond (left-hand side, space group F d h ) and hexagonal
lonsdaleite (space group P6Jmmc). The tetrahedral environments of the carbon atoms and the
stacking sequences ABC, ABC and AB, AB, respectively, are indicated.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 1 Structural Chemistry qf Hard Materiuls
Figure 3. Crystal structure of c-BN. Two unit cells of the cubic structure are shown at the right-hand
side. In the middle the structure is projected along one of the threefold axes (cube diagonal). At the
left-hand side the tetrahedral environments around one kind of atoms is emphasized. Boron and
nitrogen atoms occupy equivalent positions, hence the structure may be thought to be composed
of NB4/4 as well as BN4/4tetrahedra.
c-Sic GH-SIC
Figure 4. Crystal structures of 3C- and 6H-Sic. The tetrahedral environments of the silicon atoms
and the stacking sequences are indicated.
and transparent, while technical Sic (carborundum) is black with a bluish tint, due to
a small excess of silicon. S i c is a very non-corrosive material with a remarkable
chemical and thermal stability.
The cubic modification of 3C-Sic (space group F43rn) is isotypic with c-BN. In this
modification the chair-like layers are stacked in ABC sequence (Fig. 4) as in cubic dia-
mond with a Si-C distance of 0.189nm in all polytypes. This is somewhat less than
the average distance of 0.195 nm, between the C-C (0.154nm) and Si-Si (0.235 nm)
single bond distances in the elements [ 1 11. For Sic a huge number of different stacking
sequences have been observed. In 1972 Parthe listed some 77 different polytypic forms
[19]. With the exception of the cubic 3C-Sic modification all these polytypes have
rhombohedral, trigonal, or hexagonal symmetry. The c dimensions of their respective
hexagonal cells are all multiples of 0.252 nm. They range from 0.505 nm to 149.5 nm,
all with the same a dimension of 0.308 nm. Growth by screw dislocations has been dis-
cussed as an explanation for the long-range order of the many polytypes [20,21]. As
an example the structure of 6H-Sic (space group P6gnc) is presented in Fig. 4. Here,
the stacking sequence is ABCACB.
Various symbols have been introduced to describe the many possible stacking
sequences of (not only) the modifications of Sic. As an example, the different
notations are presented in the following scheme for a structure with fourteen layers:
A B C B A B C B A B C A C B, A B C B A B C B A B C A C B A B C notation
14H Ramsdell
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 1 Structural Chemistry of Hard Materials
The simplest one is the ABC notation, where the letters A, B, C are used for the
three possible positions of the close packed layers, starting with the letter A for the
first layer. For more complex stacking sequences the Jagodzinski-Wyckoff notation
has been introduced. This notation is only applicable to structures which allow not
more than three stacking positions and where two subsequent layers cannot have the
same stacking position. B layers of a stacking sequence ABA, where the two
neighboring layers (A) are displaced by the same amount in the same direction
are designated with the letter h (hexagonal); and B layers of a sequence ABC
where the two neighboring layers (A and C) are displaced by opposite amounts
are assigned the letter c (cubic).
These sequences can be represented in a further abbreviated form by summing up
equal subsequences counting from one h to the next one, so the sequence ABC-
BABCBABCACB can be represented by ( h ~ ) ~ ( h c cUsing ) ~ . the Zhdanov notation
this can be further abbreviated by stating the number of layers and how often
they are repeated. Thus ( h ~ ) ~ ( h cis
c )represented
~ by 2432in the Zhdanov notation.
A more general notation was introduced by Ramsdell. This notation lists the
number of slabs in one unit cell and the respective symmetry with a capital letter
(C for cubic, H for hexagonal, T for trigonal, and R for rhombohedral). Since
there is sometimes more than one way to arrange a certain number of layers in a
trigonal or hexagonal cell this notation may be ambiguous.
Figure 5. The crystal structure of B13C2(‘B4C’)in the hexagonal setting. All boron atoms indicated
by filled circles form (empty) icosahedra. Other boron atoms (open circles) are located in the centers
of linear C-B-C units. At the left-hand side the stacking (ABC) of this rhombohedra1 structure is
indicated.
Figure 6. Crystal structure of B,O. On the right-hand side of the figure one layer of oxygen-bridged
(empty) B12icosahedra is outlined. The hexagonal cell of this rhombohedra1 structure and the stack-
ing of subsequent layers is presented at the left-hand side.
these diborides have a higher density and their production is more expensive. Never-
theless, some of them are used for special applications [22,32].
Apart from the diborides, also some monoborides are technically important.
Most of them are used as hard components in composite materials. One notable
compound is FeB. It has a remarkable adhesiveness on iron-based materials and
thus increases their hardness. Such hard coatings are produced via CVD techniques.
As an intermediate product elementary boron is obtained through the reaction
+ +
2BC13 3H2 -, 2B 6HC1 between 500 and 900°C, which subsequently reacts
with iron to form the FeB coating [32,34].
The transition metal monoborides crystallize with three different structure types.
The orthorhombic structures of CrB [35] and FeB [36] are presented in Fig. 8. The
boron atoms have a trigonal-prismatic metal environment and two near boron
neighbors, thus forming infinite boron zig-zag chains. The B-B distances are
comparable to the shorter B-B distances in the modifications of elemental boron.
The Cr-B and Fe-B distances within the trigonal prisms range from 0.212 to
0.223 nm. Each boron-centered trigonal prism shares two rectangular faces with
neighboring prisms forming infinite prism columns, which extend along the z
direction in CrB and along y in FeB. The difference between the two structures is
Figure 7. The AlB, type structure of TiB,. A projection along the hexagonal axis is shown on the
right-hand side and a perspective view is presented on the left. The two-dimensional boron network
is emphasized in both drawings.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.3 Crystal Chemistry of Borides and Boron Carbides 11
X C - 4V
CrB FeB
Figure 8. The crystal structures of CrB (space group Cmcm) and FeB (Pnrna). Each boron atom in
these orthorhombic structures is located in a trigonal prism of metal atoms. Each prism shares two
rectangular faces with adjacent prisms. Thus, the boron atoms form infinite zig-zag chains.
the way in which the infinite columns are connected with each other (Fig. 8). The
monoborides VB, NbB, TaB, WB, NIB, and the high-temperature beta modification
of MOB also crystallize with CrB type structure, while TiB, MnB, and COB are
isotypic with FeB [6].
The structure of (low-temperature) a-MOB[37] is closely related to that of CrB. It
is also composed of infinite columns of boron-centered molybdenum prisms (Fig. 9).
However, in contrast to the CrB type, the boron zig-zag chains are now rotated by
90" with respect to each other (as a consequence of the four-fold screw axis). The
Mo-B distances within the trigonal prismatic units range between 0.223 and
0.265 nm. The infinite boron zig-zag chains have B-B distances of 0.174 nm. WB
is isotypic with a-MOB [37].
The rare earth hexaborides, well known for their superior electron emitting
properties [38], are also quite hard materials. The hardness of YB6 is reported to
be comparable to that of TiB2 [32]; however, such borides are quite expensive.
YB6 is isotypic with CaB6 and LaB6 [3840], where the boron atoms form B6
Mo
B
X J-Y
Figure 10. Crystal structure of cubic YB6 (space group Pm3m). The three-dimensional network of
B, octahedra and the B24 coordination polyhedra around the yttrium atoms are emphasized.
OC
Figure 11. Crystal structure of WC. A projection along the hexagonal c axis is shown on the left-
hand side of the drawing and a view of the unit cell is presented on the right-hand side. The trigonal
W6C and WC6 prisms are outlined.
carbon atoms in WC are isolated (no C-C bonds) within the trigonal tungsten
prisms. This is a significant difference when compared with the various boride
structures discussed above. The filled prisms are stacked one upon each other by
sharing common triangular faces. Thus, filled and unfilled columns alternate. The
W-C and shortest W-W distances are 0.220 and 0.284 nm, respectively.
Besides the trigonal prismatic coordination of the carbon atoms, octahedral
coordination is observed, especially for the carbon atoms in the carbides of the
early transition metals [46]. Important examples are the NaCl type carbides TIC,
ZrC, HfC, VC, NbC, and TaC [42], where the metal atoms have a cubic close-
packed arrangement, and the tungsten carbide W2C, where the tungsten atoms
form a hexagonal close packed structure [47,48].
It is well known that close packed atoms form octahedral voids. The number of
these voids equals that of the number of close packed atoms. In the NaCl type struc-
ture of TIC all octahedral voids formed by the cubic close packed titanium atoms are
filled by carbon atoms, and the Ti& octahedra share only edges. In contrast, in
W2C the tungsten atoms form a hexagonal close packed arrangement and the
carbon atoms fill only half of the octahedral voids. The w6c octahedra are
linked not only via common edges, but also by common faces. The Parthe-Yvon
rule states that carbon atoms of transition metal carbides occupying octahedral
voids of face sharing transition metal octahedra, never occupy both of these adja-
cent octahedra [47]. This is demonstrated in Fig. 12 using the Jagodzinski-Wyckoff
notation for the close packed metal atoms. For every close packed transition metal
( T )layer with an h type packing only one of the adjacent face sharing octahedral
voids can be filled. For the carbide with the composition T2C this unoccupied
void is marked with an asterisk in Fig. 12. Examples for the carbide structures
shown in that figure are W2C with (anti-)Cd12 type structure Mo3C2 (hcc),
(Ta2V)C2 (hhc), V4C3, and TIC [47, 491. Perspective views of the TIC (space
group Fm3m) and W2C ( P h l ) structures are shown in Fig. 13 with an emphasis
on the octahedral carbon coordination. The Ti6C octahedra in TIC share all
edges, while edge-sharing only within the ab plane is observed for the W6C octa-
hedra in W2C.
It is interesting to compare the structures of WC and W2C. The main difference
between these two tungsten carbides is the carbon coordination: trigonal prisms of
tungsten atoms in WC and octahedra in W2C with W-C distances of 0.220 and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
14 1 Structural Chemistry of Hard Muteriuls
I
h'
I
h'
*
'h
0.209 nm, respectively. The difference in the W-C distances may be rationalized by
geometrical considerations: in a hard sphere model (equal W-W distances) trigonal
prismatic voids are larger than octahedral voids. However, there are also differences
in the coordination numbers: the tungsten atom in WC has six carbon neighbors, in
W2C it has only three. In addition to the W-C bonding both compounds have
strong W-W bonds. Each tungsten atom has eight (WC) or twelve (W2C) tungsten
neighbors with W-W distances ranging between 0.284 and 0.300 nm, comparable
to the W-W distances in the cubic body centered structure of the element
Tic w2c
Figure 13. Crystal structure of cubic T i c (NaC1 type) and hexagonal W2C (Cd12 type). Some
carbon-filled Ti6 and W6 octahedra are emphasized.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 The Structures of Transition Metal Carbides 15
(8 x 0.274 nm and 6 x 0.316 nm). W,C has a carbon content of only 3.2 weight-%.
It has a larger microhardness than WC, but it is also more brittle [42].
The transition metal carbides with the composition T2C are sometimes called
hemicarbides. In addition to tungsten, such carbides are also formed by molyb-
denum, vanadium, niobium, and tantalum. A large number of somewhat differing
crystal structures have been reported for these, all with very similar X-ray patterns,
corresponding to the hexagonal close packed arrangement of the metal atoms. The
differences arise through slight distortions and/or larger unit cells of lower sym-
metry, which result through differing order of occupied and empty carbon positions
[50]. Usually the Parthk-Yvon rule (of two face sharing metal octahedra only one
can be occupied by carbon atoms) is obeyed. The different structures may be
stable in differing ranges of temperature and they may also be stabilized by inter-
stitial impurities like nitrogen or oxygen.
Besides the systematic carbon defects due to the hexagonal stacking sequences,
carbon defects also occur by nonoccupancy of ‘allowed’ octahedral sites. Thus,
the compositions TC, T3C2, T4C3,and T2C discussed above are the upper limits
of the homogeneity ranges. Cubic titanium carbide for example has a homogeneity
range from TIC to about TiCo.5.At high temperature the defects are randomly
distributed, while an ordering of the defects is to be expected for thermodynamic
reasons at low temperature.
As examples, partial phase diagrams of the systems titanium-carbon and
vanadium<arbon are presented in Fig. 14. Most TCI-,y phases were not
intensively studied. The VC, system is by far the best investigated system. An
ordering of the carbon atoms was found for the compositions VsC7 [53-561 and
v6c5 [57,58]. The ordering of the defects influences the positions of the vanadium
atoms. The empty v6 octahedra are bigger than the filled ones [59]. In addition to
the long-range order established for the compositions V8C7 and v6c5 short-range
order has been studied by electron and neutron diffraction in the vicinity of the
ordered structures [52, 601. Long-range order has also been reported for Nb&5
[50,61], TIC,-, (x M 1/2) [50,62-641 and other defect carbides and nitrides
[50,65-691.
The transition metal carbides and nitrides have often been called interstitial
compounds [70]; however, this is somewhat misleading. The small boron, carbon,
or nitrogen atoms certainly occupy octahedral or trigonal prismatic voids of the
metal sublattice, but the arrangement of the metal atoms themselves is different
from that of the element. In the monocarbides the transition metal atoms show
cubic close packing. However, titanium, zirconium, and hafnium are packed
hexagonally and vanadium, niobium, and tantalum are body centered cubic [ 11.
Thus, these monocarbides are inorganic compounds with their individual crystal
structures and they should not be considered as an interstitial compound of a
transition metal host lattice.
Chemical bonding in several transition metal carbides was theoretically investi-
gated by quasi self consistent augmented-plane-wave (APW) calculations [71-73]
and by the extended Hiickel method [74]. These calculation indicated a charge
transfer from the early transition metal to the carbon atoms. A crystal orbital
overlap population analysis (COOP) revealed strong bonding T-T and T-C
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
16 1 Structural Chemistry of Hard Materials
-
9 3200
I
3200 4
-s
Y
3000
3
Liquid
2800 2800
2600 2600
2400 2400
2200 2200
2000 2000
1800 1800
1600 1600
1400 1400
1200 1200
1000 1000
-
800 800
600 600
0 10 20 30 40 50 60 70 0.70 0.75 0.80 0.85 0.90 0.95
Ti atomic % carbon C/V Ratio
42 43 44 45 46 47 48
atomic % carbon
Figure 14. Partial phase diagrams of the systems titanium-carbon [51] and vanadium-carbon [52].
In the latter diagram, regions with short-range (SRO) and long-range order (LRO) are indicated.
interactions. For these calculations usually full occupancy of all atomic sites of the
NaCl type structure in assumed, a supposition which may not be entirely correct. In
NbO with a defect-NaC1 type structure 25% of the niobium and oxygen positions
are vacant, and the vacancies are ordered [75].
Carbon plays an important role in steels, both in the form of solid solution and as
component of carbide phases. The cubic face centered modification of iron (y-Fe)
dissolves as much as 8 at-% (2 weight-%) carbon, which randomly occupies the
octahedral voids of the cubic close packed iron atoms. This cubic phase is called
austenite. On quenching, this phase distorts by a displacive (‘martensitic’) phase
transition to the corresponding tetragonal structure (martensite). The degree of
the distortion is proportional to the carbon content. For a carbon content of
zero, the distortion extrapolates to zero, that is, the lattice constants of the (tetra-
gonal) martensite extrapolate to those of pure y-Fe [76,77].
Cementite Fe3C [78] plays an important role in steels. This iron carbide is signi-
ficantly harder than pure iron. Its orthorhombic crystal structure is presented in
Fig. 15. The best structure refinement of a cementite type structure seems to have
been done for Fe2,7Mno.3C[79]. The carbon atoms occupy trigonal prismatic
voids of the iron atoms with Fe-C distances of 0.201 and 0.202nm. Three
additional iron neighbors at 0.237(2 x) and 0.28 1 nm complete the coordination
shell of the carbon atoms, resulting in tricapped trigonal prisms. The trigonal
prisms are condensed via common triangular faces. Columns of such trigonal
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 The Structures of Transition Metal Carbides 17
Figure 15. Crystal structure of cementite, Fe3C (space group Pnma). The carbon filled trigonal Fe6
prisms and one empty Fe6 octahedron are emphasized.
prisms are connected via Fe-Fe bonds forming empty Fe6 octahedra. The Fe-Fe
distances between the two crystallographically different iron atoms range from 0.250
to 0.268 nm. The shorter ones compare well with the Fe-Fe distance of 0.248 nm in
b.c.c. iron [l 11.
The various chromium carbides are relatively hard and brittle. They significantly
increase the hardness and pyrophoric stability of carbon rich hard materials. These
compounds are known as Stelliten. The hardness of alloyed steels [9] results from
several chromium-iron double carbides of compositions (Fe, Cr)3C2,(Cr, Fe)23C6r
and (Fe, Cr)7C3.These mixed carbides crystallize all in the structures of the respec-
tive pure chromium carbides with a mixed occupancy of the chromium positions by
chromium and iron atoms.
Cr3C2 [80] is the technically most important chromium carbide. It is used as a
cemented carbide with nickel as binder [41]. Its orthorhombic structure is shown
in Fig. 16. The striking structural motifs of Cr3C2 are carbon centered trigonal
prisms of the chromium atoms. These trigonal Cr6C prisms are connected by shar-
ing common faces as well as common edges. The two crystallographically different
carbon atoms in Cr3C2have Cr-C distances from the center to the corner of the
Figure 16. Crystal structure and coordination polyhedra of the chromium atoms of Cr3C2 (space
group Pnma). The carbon filled trigonal Cr6 prisms are also emphasized.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
18 1 Structural Chemistry of Hard Materials
Crl CR
Cr3 C r4
Figure 17. Projection of the crystal structure of cubic Cr23C6(space group Fm3m). Some carbon
filled square antiprisms are emphasized at the left-hand side. The coordination polyhedra of the
chromium atoms are also shown.
prisms ranging from 0.202 to 0.213 nm. In addition, the C1 atoms have two further
chromium neighbors outside the rectangular faces of the prisms at 0.229 and
0.262nm, and the C2 atoms have a seventh chromium neighbor at 0.226nm. The
shortest Cr-Cr distances in Cr3C2 of 0.250nm are exactly as long as in b.c.c.
chromium [l 11. Again, the stability of this peculiar carbide structure arises from
strongly bonding metalkarbon and metal-metal interactions.
The cubic structure of the chromium carbide with the highest metal content,
Cr23C6[81] is shown in Fig. 17. Here, the carbon atoms occupy square antiprisms
of the chromium atoms with Cr-C distances of 0.21 1 and 0.213 nm. The square
antiprisms are condensed via common edges and common corners. The Cr-Cr
distances for the four crystallographically different chromium atoms range from
0.242 to 0.293 nm. Each chromium atom has at least four chromium neighbors at
short distances. The shortest Cr-Cr contacts are significantly shorter than in
b.c.c. chromium (0.250 nm) [ 1 11, indicating strongly bonding Cr-Cr interactions.
In the structure of Cr13C6 the carbon filled square antiprisms are built by the Cr3
and Cr4 atoms. Crl and Cr2 have exclusively chromium neighbors in their coordi-
nation shell. The Crl atoms have twelve chromium neighbors in a cuboctahedral
environment, while the 16 chromium atoms around the Cr2 atoms form a Frank-
Kasper [82,83] polyhedron.
The structure of Cr7C3[84] is presented in Fig. 18. Structural elements are carbon
filled trigonal prisms of chromium atoms which are condensed via common edges
and corners. The Cr-C distances within the trigonal prisms range from 0.197 to
0.21 1 nm, essentially the same range as in Cr3C2.All chromium atoms in Cr7C3
have at least three carbon neighbors and the Cr-Cr distances range from 0.249
to 0.276nm. The predominant features of chemical bonding in Cr7C3 are similar
to CrZ3C6and Cr3C2,that is, strong Cr-C and Cr-Cr interactions.
Besides the chromium-iron carbides, the so-called q-carbides, like W3Fe3C
[85,86] and W3C03C [87] play an important role in steels [42]. They are not
separately produced. The q-carbides (and nitrides) comprise a wide variety of
different compositions T3T$X,T 6 T i x , and T4TiX, where T = Mo, W; T' = Fe,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 The Structures of Transition Metal Carbides 19
Figure 18. Crystal structure of orthorhombic Cr7C3(space group Pnma). The carbon filled trigonal
prisms are emphasized.
Co, Ni; X = C (and sometimes also nitrogen or even oxygen, e.g. Ti4Fe200.4[SS]).
In the cubic structure of the compounds T3T&Y and T6TLX the metal atoms
occupy the same atomic positions, but the structures differ in their X positions,
which are either sixteen- or eightfold [86,89]. In both cases, however, the C(N)
atoms occupy octahedral voids formed solely by the molybdenum or tungsten
atoms.
As an example, the structure of W3Fe3C is presented in Fig. 19. At first sight it
seems very complex; however, a description by two independent, interpenetrating
three-dimensional networks makes this structure type comprehensible. One network
consists of corner-shared stelh quadranguh [90] of iron atoms (right-hand part,
Fig. 19), and the second network is composed of corner-sharing w6/2c octahedra.
The latter network is different from that known for the Re03 structure. The two
Figure 19. Crystal structure of W3Fe3C.In the middle of the drawing all atoms of this cubic struc-
ture (space group Fd3rn) are shown. The carbon-filled octahedra of tungsten atoms are emphasized
on the left-hand side, only the condensed Fed tetrahedra are drawn on the right-hand side of the
figure.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
20 1 Structural Chemistry of Hard Materials
crystallographically different iron atoms have six iron neighbors each, at Fe-Fe
distances ranging from 0.238 to 0.250nm. In view of the Fe-Fe distance of
0.248 nm in b.c.c. iron [l 13, these contacts in W3Fe3C certainly indicate strong
Fe-Fe bonding. In addition, there is W-W and W-Fe bonding.
As it turns out, in most ternary phases the carbon atoms are surrounded by those
metal atoms forming the more stable (i.e. higher melting) carbide. The W-C
distances of 0.206nm within the W6& octahedra of W3Fe3C, are essentially the
same as in the W6/3Coctahedra of W2C(0.209 nm) [48]. The remarkable segregation
of the W3Fe3Cstructure into an iron and a W6/2Csublattice parallels the stability of
the binary carbides; that is, high-melting WC and metastable Fe3C.
MoSi,, AB
Figure 21. Crystal structures of tetragonal MoSi2, hexagonal CrSi,, and orthorhombic TiSi,. The
C N 10 polyhedra of silicon atoms around the transition metal atoms and parts of the two-
dimensionally close packed 7Si2 layers are outlined for each structure.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
22 1 Structural Chemistry of Hard Materials
neighbors in the form of a bicapped square prism. These polyhedra are condensed
via common edges and faces.
The unit cells of the orthorhombic TiSi2 (space group Fddd) and hexagonal CrSi2
(space group P6222) structures are also presented in Fig. 21. For both structures, the
7Sil0 polyhedra and parts of the T-centered Si6 rings are outlined. The Si-Si
distances range between 0.247 and 0.279 nm. These bonding Si-Si distances are
all significantly longer than the Si-Si bond length of 0.235 nm in elemental silicon
[I 11. This has to do with the fact, that the coordination number (CN) of the silicon
atoms in the transition metal disilicides is considerably higher (CN 10) than in the
diamond structure of elemental silicon (CN 4).
Besides Ti5Si3C1- with partially filled Mn5Si3type structure (see below), Ti3SiC2
is the only other phase in the ternary system Ti-Si-C [95-971. Ti3SiC2crystallizes in
the form of thin platelets which can be plastically deformed [98]. It combines refrac-
toriness (m.p. M 3200°C) and stability against aggressive media with thermal shock
resistance. A high potential as a component of ceramics and composite materials
was discussed [99,100]. The electronic structure of Ti3SiC2 was investigated by
self-consistent full-potential linear muffin-tin orbital calculations [ 1011.
The structure of Ti3SiC2is shown in Fig. 22. The titanium and silicon atoms are
close packed with the stacking sequence ( h h h ~ )Thus,
~ . Ti3SiC2adopts a structure
typical for a metal-rich carbide. The titanium octahedra are filled by carbon
atoms, and these Ti6C octahedra are condensed via common edges, forming
double layers which are separated by the silicon atoms. No siliconxarbon bonds
are observed. In compounds of similar composition Si-C bonds (0.193 nm) have
been found only in U3Si2C2[102].
Figure 22. Crystal structure of Ti3SiC2 (space group P6,lmmc). The close-packed titanium layers
containing the carbon atoms are shown on the left-hand side. The edge-sharing Ti& octahedra
are emphasized. On the right-hand side a (110) cut through the hexagonal cell is shown. The
titanium and silicon atoms form close-packed layers with the stacking sequence hhhc where the
silicon atoms correspond to the second h.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.6 Nitrides 23
Mol Si
Mo2 C
Figure 23. Crystal structure of Mo5Si3C(space group P6,lmcm). On the left-hand side of the draw-
ing a projection of the structure along the hexagonal z axis is presented. The carbon atoms (not
shown) are situated inside the emphasized Mo6 octahedra. These share faces, thus forming infinite
chains, which extend along the hexagonal axis. The near-neighbor coordinations of all atoms are
presented at the right-hand side of the figure.
1.6 Nitrides
In elemental form, carbon has many applications in both modifications as diamond
and as graphite. In contrast, the neighboring element nitrogen is a gas. Nevertheless,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
24 I Structural Chemistry of Hard Materials
a-Si3N4 P-Si3N4
Figure 24. Crystal structures of trigonal a-Si3N4 (space group P31c) and hexagonal P-Si3N4
(P63/m), emphasizing the corner-sharing SiN4 tetrahedra.
both elements form hard materials, which are structurally closely related, and both
form such refractory compounds with main-group elements (e.g. Sic, Si3N4)and
with transition metals (e.g. WC, TiN).
A B A B A B
Figure 25. Crystal structure of c-Fe2N. The iron atoms form hexagonal close-packed layers which
extend along the yz-plane of this orthorhombic structure (space group Pbcn). In the x direction these
planes have the stacking sequence AB, AB. Hence, the iron atoms form a distorted hexagonal close-
packed arrangement where the nitrogen atoms fill half of the octahedral voids.
Nitrogen also plays a role in steel, not only as a component of solid solutions in
carbides, but especially in the form of hard surface coatings. The crystal structures
of the respective hard materials are not all well known. In the present section we
discuss some representative structures that were determined reliably. The nomen-
clature of the iron nitrides derives from the phase diagram of the binary system
iron-nitrogen. Thus, the a- and y-phases are solid solutions of nitrogen in the a-
and y-modifications of iron.
The structure of E,-Fe2N [I211 was established recently from a combination of
neutron diffraction and high-resolution synchrotron powder data. c-Fe2N crystal-
lizes with anti-a-PbO, type structure as outlined in Fig. 25. The structure consists
of edge- and corner-sharing Fe6N octahedra with Fe-N distances ranging from
0.189 to 0.201 nm. The iron atoms form a hexagonally close-packed array and the
nitrogen atoms fill half of the octahedral voids in an ordered manner.
The structures of E-Fe3N and y'-Fe4N are also based on close-packed arrange-
ments of iron atoms [122]. In &-Fe3Nthe iron atoms are hexagonal close-packed
and nitrogen atoms occupy only corner-sharing octahedra. The ordered arrange-
ment of the nitrogen atoms results in a & times larger a axis. In this structure
the deviations of the iron atoms from the atomic positions in an ideal hexagonal
close-packed structure are small.
The iron atoms in y'-Fe4N have a face-centered cubic arrangement, and the nitro-
gen atoms occupy octahedral voids (Fig. 26). This results in a three-dimensionally
infinite arrangement of all-corner-sharing Fe6N octahedra with Fe-N distances
of 0.190nm, slightly shorter than the average Fe-N distance of 0.195nm in
E,-Fe2N. Since the nitrogen atoms occupy only one of the four octahedral voids
formed by the face-centered cubic iron atoms, the space group symmetry is lowered
from Fmjm to Pm%.
In addition to the congruently melting (m.p. x 3300°C) NaCl type phase TiNl --x
(x = 0.5 to 1.O) the phase diagram of the system titanium-nitrogen contains at least
three other nitrides, all of which are stable at temperatures of less than 1300°C[119].
Of these the nitride Ti2N crystallizing with (anti-)rutile type structure (Fig. 27) is
known for some time [123]. In this structure the titanium atoms form octahedra
which are filled by nitrogen atoms. The Ti6N octahedra are linked via common
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.6 Nitrides 27
Fe4N1-x li3AIN
Figure 26. The crystal structures of y'-Fe4N and Ti,AIN (space group P m h ) . In the binary nitride
the nitrogen atoms occupy one quarter of the octahedral voids formed by the cubic close-packed
iron atoms. In the ternary compound the nitrogen atoms fill those octahedral voids which are
formed solely by the titanium atoms.
edges, thus forming chains which extend along the c axis of the tetragonal cell
(Fig. 27). These chains are further condensed by sharing corners, and in this way
a three-dimensional network is formed. For the other two titanium nitrides the
structure determinations resulted in the compositions Ti3N2- and Ti4N3- x.
These nitrides are stable at only very limited temperature ranges between 1050
and 1300°C [119,124,125].
Considering composites and thin-film coatings containing TIN and AlN, one has
to be aware of the ternary compounds Ti3A1N [126], Ti2AlN [127], and Ti3A12N2
[126]. Of these the latter is thermodynamically stable only between 1200 and
1300°C. In all three structures the metal atoms are close packed. While the structure
of Ti3A12N2has a large cell with a ten-layer stacking sequence, the metal positions in
Ti3AlN and Ti2AlN correspond to cubic (ABC) and hexagonal (AB) close packed
structures (Fig. 28). In both of these the nitrogen atoms are situated in octahedral
voids formed solely by the titanium atoms. The cubic structure of Ti3A1N is
known as perovskite (CaTi03 type) structure and Ti2AlN is an H phase (Cr2A1C
type). There are many representatives of these carbides T3MX and T2MX where
T are early transition metals, mainly from groups 14 and 15, M are main-group
elements, especially Al, Ga, In, Sn, Pb, and X = C, N [5, 128-1311.
Layers of zirconium and aluminum atoms can also be discerned in the structure
of Zr3AlN [132]. In this complex nitride the nitrogen atoms are again situated in
Figure 27. Crystal structure of Ti2N with tetragonal (unti-)rutile type structure (space group
P42/mnm).The edge- and corner-sharing Ti6N octahedra are emphasized.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
28 I Structural Chemistry of Hard Materials
Y41lOfl YL X
Ti,AIN Ti,AIN
Figure 28. Crystal structure of Ti3AlN (CaTiO,, perovskite type) and Ti2AlN (Cr2AlC, H phase
type structure). In both structures the metal atoms are close packed. In the cubic perovskite type
structure the positions of the close-packed metal atoms correspond to those of the Cu3Au structure
(space group Pm3m) with the stacking sequence ABC, ABC. In the hexagonal Cr2A1Ctype structure
of Ti2AIN the stacking sequence of the close-packed metal atoms is AB, AB.
octahedral voids formed solely by the transition metal atoms. These Zr6N octahedra
share edges and corners, thus forming two-dimensionally infinite sheets with
aluminum layers in between them (Fig. 29). It is striking that the carbon and
nitrogen atoms in complex carbides and nitrides, containing transition metal and
main-group elements, are usually situated in octahedral voids formed by the
“I “I ”‘I “I “I
i W
Figure 29. Crystal structure of Zr,AlN (space group Cmcm). The layers of edge- and corner-sharing
Zr6N octahedra are emphasized.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.6 Nitrides 29
transition metals. Other examples already discussed above are the structures of
Ti3SiC2 [98] and Mo5Si3C [103]. Another example is the filled P-manganese type
structure of Nb3A12N and Mo3A12C [133]. Also, ,as already mentioned, in the q-
W3Fe3C type carbides and nitrides the carbon atoms are situated in octahedral
voids formed by those transition metal atoms which form the more stable (high-
melting) binary carbides. It can be assumed, that these octahedral Tgc and T6N
groups are very stable and that they possibly exist already in the melts.
e N
Ca2Si5N8 Ce,S,iN,
Figure 30. Crystal structures of monoclinic Ca2Si,N, (space group Cc) and tetragonal Ce3Si6NI
(Sm3Si6NII type, space group P4bm). The three-dimensional networks of comer-sharing SiN4
tetrahedra are emphasized.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
30 1 Structural Chemistry of Hard Materials
Figure 32. The wurtzite type structure of cL-BeO, space group P63nrc. The three-dimensional net-
work of condensed Be04-tetrahedra is outlined. The stacking of these tetrahedra has the sequence
AB, AB. The corresponding stacking ABC, ABC is known to occur for the zinc blende (sphalerite)
structure (Fig. 3 ) .
21- +y Brookite
Figure 34. The crystal structures of orthorhombic brookite (space group Pbca), tetragonal anatase
(14,lamd), and tetragonal rutile (P4*/mnrn).The edge- and corner-sharing Ti06 octahedra and the
more or less distorted trigonal-planar environments of the oxygen atoms are emphasized.
share three common edges with other octahedra in brookite and even four common
edges in anatase. According to the Pauling rules for ionic crystals, the stability of
such an atomic arrangement decreases as the number of common edges increases,
so the repulsion of the titanium atoms increases. The Ti-0 distances in anatase
are 0.193nm (4x) and 0.198nm (2x), nearly identical with rutile (4 x 0.195nm,
2 x 0.198nm). In brookite on the other hand, they vary from 0.187nm to
0.204nm. As a consequence, one would expect the average Ti-0 distance of
0.196 nm in brookite to be significantly greater than the average Ti-0 distances
in rutile (0.1959nm) and anatase (0.1949nm), which is not the case, probably
because the structure refinement of brookite, carried out from film data, was not
that accurate.
In this context it is worth mentioning that the stishovite form of SiOz [147] with
rutile type structure is a very hard material. Recent investigations on polycrystalline
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
34 I Structural Chemistry of Hard Materials
cubic Zr02
L--
. I
- - - _ _ _ _ _ I
monoclinic Zr02
Figure 35. The crystal structures of the cubic, tetragonal, and monoclinic forms of Zr02. The
different zirconium-oxygen polyhedra are outlined. The cubic fluorite type (space group, Fm3m)
high-temperature y-modification transforms on cooling by a displacive phase transition to
tetragonal P-ZrOz (P4,/nmc), which on further cooling transforms, again by a displacive phase
transition, to monoclinic a-Zr02 (P2,/c).
stishovite have shown that this material is harder than alumina and almost equals
the hardness of c-BN [148].It is thus the hardest known oxide.
The high-temperature (y-) forms of Zr02 and HfOz [149] as well as CeOz [150],
Thoz, and UOz [I511 crystallize with the cubic fluorite type structure. The metal
atoms form a close-packed f.c.c. arrangement and the oxygen atoms fill all tetra-
hedral voids. This way each metal atom obtains a cubic environment of eight
oxygen atoms (Fig. 35).
Thoz and UOz are used as nuclear fuels. Tho2 has the highest melting point of all
oxides (>3200"C) and an extremely low oxygen vapor pressure. It is therefore used
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.7 Oxide Ceramics 35
References
1. P. P. Ewald and C. Herrmann, Strukturberichte Vol. 1-7, in Z . Kristallogr. Akademische
Verlagsgesellschaft mbH, Leipzig, 1913-1939.
2. A. J. C. Wilson, W. B. Pearson, J. Trotter, and G. Ferguson (Eds), Structure Reports Vol. 8-
58, International Union for Crystallography, N. V. A. Oosthoek's Uitgevers MIJ, Utrecht,
1956-1993.
3. Landolt-Bornstein, Numerical Data and Functional Relationships in Science and Technology,
K.-H. Hellwege (Ed.), Group III. Crystal and Solid State Physics, Vol. 6, Structure Data of
Elements and Intermetallic Phases, P. Eckerlin, H. Kandler, and A. Stegher (Eds), Springer,
Berlin, 1971.
4. W. B. Pearson, The Crystal Chemistry and Physics of Metals and Alloys, Wiley-Interscience,
New York, 1972.
5. P. Villars and L. D. Calvert, Pearson's Handbook of Crystallographic Data for Intermetallic
Phases, ASM International, Materials Park, OH, 1991.
6. P. Villars, Pearson's Handbook, Desk Edition, Crystallographic Data for Intermetallic Phases,
Vol. 1 and 2, ASM International, Materials Park, OH, 1997.
7. E. ParthC, L. Gelato, B. Chabot, M. Penzo, K. Cenzual, and R. E. Gladyshevskii, TYPIX-
Standardized Data and Crystal Chemical Characterization of Inorganic Structure Types, in
Gmelin Handbook of Inorganic and Organometallic Chemistry, 8th edn., Springer, 1993.
8. Fachinformationszentrum FIZ Karlsruhe, Inorganic Crystal Structure Data Base - ICSD,
Karlsruhe, 1998.
9. International Centre for Diffraction Data - ICDD, Powder Diffraction File, Newtown Square,
Pennsylvania, 1997.
10. K. Yvon, W. Jeitschko, and E. Parthe, J. Appl. Crystallogr. 1977, 10, 73.
11. J. Donohue, The Structures ofthe Elements, Wiley, New York, 1974.
12. F. P. Bundy and J. S. Kasper, J. Chem. Phys. 1967,46, 3437.
13. V. L. Solozhenko, V. V. Chernyshev, G. V. Fetisov, V. B. Rybakov, and I. A. Petrusha,
J . Phys. Chem. Solids, 1990,8, 1011.
14. L. F. Vereshchagin, E. V. Zubova, L. N. Burenkova, and N. I. Revin, Dokl. Akad. Nauk SSR,
1968, 13, 25.
15. T. Soma, A. Sawaoka, and S. Saito, Muter. Res. Bull. 1974, 9, 755.
16. A. H. Comes de Mesquita, Acta Crystallogr. 1967, 23, 610.
17. Z. Li and R. C. Bradt, J . Muter. Sci. 1986, 21, 4366.
18. W. Buchner, R. Schliebs, G. Winter, and K. H. Buchel, Industrielle Anorganische Chemie,
Verlag Chemie, Weinheim, 1984.
19. E. Parthe, Crystallochimie des Structures Tetraedriques, Gordon & Breach, Paris, 1972.
20. P. T. B. Shaffer, Acta Crystallogr. Sect. B, 1969, 25, 477.
21. M. Dubey, U. S. Ram, and G. Singh, Acta Crystallogr. Sect. B, 1973, 29, 1548.
22. H. Briehl, Chemie der Werkstofle, B. G. Teubner, Stuttgart, 1995.
23. H. L. Yakel, Acta Crystallogr. Sect. B, 1975, 31, 1797.
24. R. Telle, Chemie in unserer Zeit, 1988, 22, 93.
25. A. Kirfel, A. Gupta, and G. Will, Acta Crystallogr. Sect. B, 1979,35, 1052.
26. H. Neidhard, R. Mattes, and H. J. Becher, Acta Crystallogr. Sect. B, 1970, 26, 315.
27. A. J. Perrotta, W. D. Townes, and J. A. Potenza, Acta Crystallogr. Sect. B, 1969, 25, 1223.
28. A. F. Wells, Structural Inorganic Chemistry, 5th edn., Clarendon Press, Oxford, 1990.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
38 I Structural Chemistry of Hard Materials
122. H. Jacobs, D. Rechenbach, and U. Zachwieja. J. Alloys Compd. 1995, 227, 10.
123. B. Holmberg, Acta Chem. Scand. 1962, 16, 1255.
124. W. Lengauer and P. Ettmayer, J. Less-Common Met. 1986, 120, 153.
125. W. Lengauer, J. Less-Common Met. 1986, 125, 127.
126. J. C. Schuster and J. Bauer, J . Solid State Chem. 1984, 53, 260.
127. W. Jeitschko, H. Nowotny, and F. Benesovsky, Monatsh. Chem. 1963,94, 1198.
128. H. H. Stadelmaier in Developments in the Structural Chemistry of Alloy Phases, B. C. Giessen
(Ed.), Plenum Press, New York, 1969.
129. L. E. Toth, Transition Metal Carbides and Nitrides, Academic Press, New York, 1971.
130. H. Nowotny, Angew. Chem. Intern. Edit. 1972, 11, 906.
131. Th. M. Gesing, K. H. Wachtmann, and W. Jeitschko, Z. Naturforsch. 1997, 52b, 176.
132. J. C. Schuster, Z . Kristallogr. 1986, 175, 211.
133. W. Jeitschko, H. Nowotny, and F. Benesovsky, Monatsh. Chem. 1964,95, 1212.
134. W. Schnick and H. Huppertz, Chem. Ew. J. 1997, 3, 679.
135. M. Woike and W. Jeitschko, Inorg. Chem. 1995, 34, 5105.
136. T. Schlieper and W. Schnick, Z . Anorg. Allg. Chem. 1995,621, 1037.
137. H. W. Hennicke, U. Hofmann, W. Lehnhauser, and K.-J. Leers, Keramik, in Ullmanns Enzyk-
lopadie der technischen Chemie, Vol. 13, VCH, Weinheim, 1974, pp. 711-735.
138. K. Heldt and A. Reckziegel, Oxidkeramik, in Ullmanns Enzyklopadie der technischen Chemie,
Vol. 17, VCH, Weinheim, 1974, pp. 515-529.
139. J. Lewis, D. Schwarzenbach, and H. D. Flack, Acta Crystallogr. Sect A , 1982, 38, 733.
140. T. M. Sabine and S. Hogg, Acta Crystallogr. Sect. B, 1969, 25, 2254.
141. D. K. Smith, C. F. Cline, and S . B. Austerman, Acta Crystallogr. 1965, 18, 393.
142. E. Schiebold, Z. Kristallogr. 1921, 56, 430.
143. N. G. Zorina and S . S . Kvitka, Kristallografiya, 1969, 13, 599.
144. R. Restori, D. Schwarzenbach, and J. R. Schneider, Acta Crystallogr. Sect. B, 1987,43,251.
145. M. Horn, C. F. Schwerdtfeger, and E. P. Meagher, Z . Kristallogr. 1972, 136, 273.
146. W. H. Baur, Acta Crystallogr. 1961, 14, 214.
147. M. Sugiyama, S. Endo, and K. Koto, Mineral. J. Jpn. 1987, 13, 455.
148. J. M. Leger, J. Haines, M. Schmidt, J. P. Petitet, A. S. Pereira, and J. A. H. de Jomada, Nature,
1996,383,401.
149. L. Passerini, Gazz. Chim. Ital. 1930, 60, 762.
150. G. Brauer and H. Gradinger, Z. Anorg. Allg. Chem. 1954, 277, 89.
151. K. Clausen, W. Hayes, J. E. Macdonald, P. Schnabel, M. T. Hutchings, and J. K. Kjems, High
Temp. High Press. 1983, 15, 383.
152. I. McGill, J. Helgorsky, A. LevCque, Th. Petzel, and K. Reinhardt, Rare Earth Elements, in
Ullmann’s Encyclopedia of Industrial Chemistry, 5th edn., Vol. A22, VCH, Weinheim, 1993,
pp. 607-649.
153. G. Teufer, Acta Crystallogr. 1962, 15, 1187.
154. D. K. Smith and H. W. Newkirk, Acta Crystallogr. 1965, 18, 983.
155. G. Routschka and K.-E. Granitzki, in Ullmann’s Encyclopedia of Industrial Chemistry, 5th
edn., Vol. A23, VCH, Weinheim, 1993, pp. 1 4 8 .
156. L. Michalowsky (Ed.), Neue keramische Werkstofle, Deutscher Verlag fur Grundstoffindustrie,
Leipzig, 1994.
157. R. Riedel, Advanced ceramics from inorganic polymers, in Materials Science and Technology,
R. W. Cahn, P. Haasen, and E. J. Kramer (Eds.), Vol. 17B, Chapter 11, VCH, Weinheim,
1996.
158. R. Riedel, J. Bill, and A. Kienzle, Appl. Organomet. Chem. 1996, 10, 241.
159. H.-P. Baldus and M. Jansen, Angew. Chem. 1997, 109, 338.
160. B. D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, Reading, Massachusetts, 1959.
161. L. V. Aziroff, Elements of X-Ray Crystallography, McGraw-Hill, New York, 1968.
162. B. E. Warren, X-Ray DifSraction, Addison-Wesley, Reading, Massachusetts, 1969.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
2.1 Introduction
At room temperature, pressures in excess of 500GPa can be attained using the
diamond anvil cell (DAC) technique [l]. This technique, described in a number of
comprehensive reviews [24], has found broad application in the high pressure
sciences, because diamond serves as an optical window from the far infrared to
the near ultraviolet wavelength regime and is transparent to X-rays [4,5].
Development of the convenient ruby pressure scale, where the red shift of the R1-
fluorescence line with pressure is used in situ [6,7], led to a much wider use of this
experimental technique.
A few high-pressure devices are used to achieve high temperatures at high static
pressures: Large-volume high-pressure devices (with characteristic sample dimen-
sions of 2-3mm) can be used at temperatures up to 3000K, and pressures of
about 25GPa (white field in Fig. 1). In an externally heated DAC temperature is
limited to about 1800 K. Above this temperature the diamond anvil transforms to
graphite, even in an inert atmosphere [2]. Moreover, these devices are limited to
pressures of about 30GPa due to weakening of the diamond anvils and of other
parts of the DAC upon heating. Additionally, pressure measurements are com-
plicated by the absence of accurate high-pressure-high-temperature equations of
state for any solid that could be used as a pressure gauge. Another possibility
involving use of an internally heated DAC, where the conducting sample is
heated by passing of an electric current through it [8,9], is constrained by its
range of applicability. The laser-heated DAC allows us to access much higher
static pressures and temperatures than the techniques listed above (gray field in
Fig. 1).
The technique of laser heating in a DAC is based on three main features: optical
transparency of diamond anviIs; the samples can be heated via the optical absorp-
tion of intense laser radiation, and the temperature can be determined from the
thermal radiation spectrum of the heated sample using the Planck formula [lo].
Laser radiation for heating of a sample in a DAC was first implemented by Ming
and Bassett [ll], who used a pulsed ruby laser, and a continuous-wave Nd-YAG
(yttrium-aluminum-garnet) laser to heat samples in a DAC above 3300 K, and up
to 2300K, respectively. Today two types of continuous wave infrared (IR) lasers
are extensively used in laser heating experiments: Solid state lasers (Nd-doped
YAG, or YLF (yttrium-lithium-fluorite) crystals with the most intense line at
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
42 2 Phase Transitions and Material Synthesis using the COI-Laser Heating Technique
s!? SUN
3
b 20 60
Pressure (GPa)
Figure 1. Pressure-temperature regime accessed to date using the C02-laser heating technique is
shown by the light gray field. The P-T values available for multianvil high pressure devices are
given by the white field. The temperature on the surface of the sun is shown for comparison.
1.06 pm), and C 0 2 gas lasers (most intense line at 10.6 pm). The Nd-YAG lasers
have been used for heating semiconductors [12], metals, and insulators containing
transition metals such as Fe, Ni, etc. [l 11. The absorption mechanism involves cou-
pling of the laser radiation with electronic excitations within the sample. The C02-
laser heating technique, first employed by Boehler and Chopelas [ 131, has been used
in experiments on numerous nonconducting inorganic (oxides, silicates, nitrides,
etc.) and organic materials (e.g. hydrocarbons C,H2, + 2 ) . Heating of these materials
with C02-laser radiation is possible due to the resonant absorption of the laser
radiation by lattice vibrations. The C02-laser heating technique is particularly
suitable for high-pressure-high-temperature experiments on hard materials (i.e.
cubic BN, Sic, Si3N4),since most of them are insulators.
In section two of this paper we describe the technical aspects of C02-laserheating
in a DAC. The third section focuses on the methods for measuring melting tem-
peratures at variable pressures, the fourth section on the determination of high
pressure and temperature phase diagrams, and in the fifth section some experiments
focusing on the synthesis of diamond and cubic BN from organic precursors will be
described.
COz-laser Ar-laser
Figure 2. Schematic drawing of a sample assemblage in the diamond cell. The sample is heated
with a C 0 2 laser. Nitrogen is used as a pressure medium in experiments on nitrides to preclude
decomposition. Argon is used as a pressure medium when chemically inert conditions are required.
Fluorescence of ruby crystals is excited using an Ar laser.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
44 2 Phase Transitions and Material Synthesis using the C02-LaserHeating Technique
after heating of the sample the pressure gradients are reduced to 3% of the average
value.
During laser heating the pressure in the sample chamber increases owing to
thermal expansion of the sample and of the surrounding pressure medium. To
reduce this effect in our C02-laser heating experiments, the sample to pressure
medium volume ratio was kept below 1 :10. We found that upon heating silicates
up to 2600K at around 20 GPa the thermal pressure does not exceed 0.3 GPa,
which is less than 2% of the average pressure value [13].
l4 I I
L
13
Figure 3. Schematic drawing of the experimental set-up for the C02-laser heating experiments in a
DAC: 1 = C 0 2 laser, 2 = Si mirror, 3 = ZnSe lens, 4 =diamond cell, 5 =reflecting objective,
6 = monochromator, 7 = CCD detector, 8 = PC, 9 and 12 = beam splitter, 10 = optical diode,
11 =regulating unit, 13 = ocular, 14 =power supply of the C 0 2 laser. The dashed and dotted line
shows the path of the C02-laser beam, the dashed lines show the visible light path, and the solid
ones depict electrical connections.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Technique of C02-Laser Heating in a Diamond Anvil Cell 45
radiation spectrum from this area is measured with a charge coupled device (CCD)
detector. To determine the temperature, the Planck formula is fitted to the measured
spectrum (see below). Approximately 50% of the incandescent radiation from the
sample is directed to an optical diode by a beam-splitter. The diode signal, after
amplification and conversion in the regulating unit, is used to stabilize the laser
power (see below). The second beam-splitter directs the sample radiation to the
ocular for visual observation.
where h is the Planck constant, k the Boltzmann constant, c the velocity of light in
vacuum, T the temperature of the BB, and X the wavelength.
The temperature and wavelength dependent emissivity E ( T ,A) is introduced
for description of a thermal radiation spectrum of a real body ZRB(T,X). Emis-
sivity is simply the ratio of the intensities of the real body to the BB at the
same X and T [5]:
In our experiments we assumed that the materials investigated are graybodies; that
is that their emissivities, which are smaller than unity, are wavelength independent
E ( T ,A) = e( T ) .The wavelength dependence of emissivity is only known for a few
materials in vacuum or at atmospheric pressure [18,19]. If we assume that E of
the investigated materials has the same wavelength dependence as tungsten [ 191
then the temperature correction would be about -100K at 3000K and about
-350K at 5000K. Figures 4a and 4b show the thermal radiation spectra (thick
solid lines) of S i c heated to about 6700K at 12GPa and of MgSi03 heated to
3000 K at 100 GPa, respectively. The temperature values correspond to the fits
(thin solid line) of the Planck formula with a wavelength-independent emissivity
to the measured spectra.
SIC
P=12 GPa
T=6700 K
.-
E
fn
C
9
-
c
550 650
I . 1 ’ I ’ I ’ I . I
0 600 700 750 800
a wavelength (nm)
E
Y,
C
d
-
c
5% of a chosen average value, when feedback stabilization was not employed. This
resulted in temperature fluctuations with a peak-to-peak amplitude exceeding 30%
of the average sample temperature. Figure 5a shows temperature fluctuations in a
sample heated for 60s with an unstabilized C 0 2 laser. This nonlinear relation
between laser power fluctuations and temperature fluctuations is most probably
due to the temperature dependent absorption of the sample material: For example,
experimental measurements of the absorption coefficient k(X) of MgO between 1
and 9 pm to about 1300K at atmospheric pressure showed that k(X) increases non-
linearly with temperature as well as with wavelength [20].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Technique ofC02-Las e r Heating in a Diamond Anvil Cell 47
2800 -
.- ...
2600 - c
2400- -..
L
-
.-. . -%\
,L :
,,--
..'
I
2200 - -9
*. %
.;-
1
, 1
-
T=2140 K
2ooo
1800). , . , . I . , . I . . , . I
0 10 20 30 40 50 60 70 80 90 100
time (x0.6 s)
(Mg,Fe)SiO,
3000 3010+50 K 29 GPa
28001
2600
2000
0 5 10 15 20 25 30 35 40 45 50
time (min)
Figure 5. (a) Temperature fluctuations in the (Mg,Fe)Si03 sample heated in a DAC with an
unstabilized COz laser: 100 temperature measurements within a time interval of about 60 s from
the center of the heated area (squares). During this time interval the temperature fluctuated between
2140 and 3000 K. (b) Temperature fluctuations in the (Mg, Fe)SiO, sample heated in a DAC with
laser stabilization to average temperatures of 2090 K (diamonds), 2420 K (circles) and 3010 K
(squares) were monitored during 35, 20, and 17 min, respectively. The laser stabilization suppressed
the peak-to-peak fluctuations to below 120 K, or to below 5% of the average temperature. After [21].
I . I . I . I , I . I , l
h
2600 1
2400-
!?
3
2200-
2
.l-i
8
E
0
0
I- 2ooo:
1800 -
1600!, . I . , . . , . I !
experiments the samples were excited with an argon ion laser (A = 457.9 nm) with
powers varying from a few milliwatts to about 100mW. The spatial resolution of
the measurements was typically 10-15pm. Spectra (see below) were taken from
the center of the laser-heated spot at room temperature using CCD, or photon
counting detection systems.
co2-laser \
Figure 8. Schematic of the cubic BN sample melted in a DAC at 10 GPa. Temperatures recorded for
this sample are shown in Fig. 7. The outgrowth in the center of the heated area is due to melting.
cubic BN only, no traces of the low pressure hexagonal BN phase were observed.
These results are consistent with the previous estimation of Corrigan and Bundy
~71.
5000 - (Mg,Fe)SiO,-pv
4500 - CaSi0,-pv ,
A,..
,*- multi-component 1
1500-.' multi-anvil -
~ - I - I ~ I - I - ~ ~
of the theoretical melting relation [40,41] to the data revealed that the Griineisen
parameter of MgO at 1 atm is 1.3, in good agreement with the literature value of
1.4 [43].
Melting temperatures of orthorhombic (Mg, Fe)Si03-perovskite and of cubic
CaSi03-perovskite were measured using the COJaser heating technique to
62 GPa [34] and 43 GPa [33], respectively (Fig. 9). Melting temperatures of both
perovskites (which are considered to be the most abundant lower mantle phases)
have similar strong pressure dependencies: The initial melting slope for both systems
is about 60 K GPa-', which is almost twice that of MgO. At the highest pressures
of these experiments the obtained melting temperatures exceeded 5000 K for
(Mg, Fe)Si03-perovskite and 4000 K for CaSi03-perovskite. Precise measurement
of such high melting temperatures is only possible if the radial and axial temperature
gradients are small, and if chemically inert pressure media (for example noble gases)
are used. These requirements were not fulfilled in earlier DAC experiments on
(Mg, Fe)Si03-perovskite [44-46]. In those experiments samples were heated with
YAG-laser radiation, which is only weakly absorbed by the sample, and no ther-
mally insulating pressure medium was used. In previous melting experiments on
CaSi03-perovskite [47], rhenium plates in contact with the sample were used to
absorb the Nd-YAG laser radiation, thus indirectly heating the sample. We have
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Determination of Melting Temperatures at High Pressures 53
Figure 10. (a) Back-scattering electron topographic image of the glass sample with the mantle
relevant composition heated just above the solidus to 2550 K at 20.6 GPa. Within a circular patch
of about 60 pm in diameter the starting material converted to a crystalline high pressure assemblage.
Cracks within this area became evident after pressure release. Another type of texture of about
lOpm in diameter in the center (arrow) is due to migration of molten sample material towards
the laser beam, as evident from quantitative measurements with an atomic force microscope in
the area indicated by dashed lines. (b) Three-dimensional surface topography of the 34 x 34pm
area, shown in Fig. 10(a) by dashed lines, measured with an atomic force microscope. Partial melting
of the sample within an area of about 10 pm in diameter caused height variations of up to 0.77 pm. In
contrast, height variations in the surrounding area are smaller than 0.2 pm. Reprinted with permis-
sion from [35]. Copyright [1998] American Association for the Advancement of Science.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
54 2 Phase Trunsitions and Material Synthesis using the C02-Laser Heating Technique
found, however, that at high temperatures CaSi03 reacts with rhenium [33], which
probably depresses the melting temperatures by up to 650 K [47].
The solidus temperature of a multicomponent system (for example, the Earth‘s
mantle material) cannot be predicted even if the melting temperatures of all end-
member components of the system are known. Previous solidus temperatures of
mantle relevant materials were measured to about 25 GPa using multianvil high
pressure devices [36,48,49]. The solidus of the (Mg, Fe)O-(Mg, Fe)Si03 assem-
blage, which can be considered as a simplified model for the Earth’s mantle material,
was measured to about 33 GPa in a DAC using a Nd-YLF laser [33]. Additionally,
melting of (Mg, FehSi04, which breaks down to the same phase assemblage, was
estimated in shock-wave experiments to be 4300 K at about 130 GPa [50]. For the
measurement of the solidus temperature of a multicomponent material with a
realistic Earth’s mantle composition we used [35] a different melting criterion
involving changes of the surface texture due to melting. Textural changes on the
sample surface due to melting were verified using a scanning electron microscope
(SEM) and an atomic force microscope (AFM). Figure 10a shows a back-scattered
electron topographic image of the sample which was partially molten at 2550K
and 20.6 GPa. Characteristic melting features (about 10 pm in diameter) can be
seen in the center of the heated area. In the molten area the height variations are
almost four times larger than in the surrounding subsolidus material, as evident
from quantitative topographic measurements with an AFM (Fig. lob). At pressures
between 20 and 30 GPa our data (Fig. 9) agree well with previous multianvil data and
with the solidus of the (Mg, Fe)-(Mg, Fe)Si03 assemblage measured using an YLF-
laser. Our results indicated a strong depression in the melting temperatures of the
multicomponent system as compared with the end-member phases: At 60 GPa the
solidus is about 900 K below the melting temperature of MgO. Our data [35] extrapo-
late to the melting point estimation from the shock wave experiment at 130GPa [50].
This supports the assumption that the solidus of the Earth’s mantle material is
dominated by the melting behavior of the (Mg, Fe)@(Mg, Fe)Si03 assemblage.
from the same parts of the samples that serve for phase identification, thus, avoiding
problems associated with temperature gradient; the samples are characterized in situ,
avoiding reverse transitions upon pressure release; higher temperatures decrease
possible kinetic problems on phase transitions.
We illustrate the use of the laser heated diamond anvil cell for synthesizing high
pressure phases and establishing phase boundaries of solids with the following three
studies.
12.0 GPa
"0°1 Z (1993)
Se(1995)
-I
:
. ,'Z(1993)
I- 2000 :
. ,
,>-
1
. .,
: 3
3-..z'Z(1994)
1500 R
1000- C = Coesite
- .
8 1'0 15 1'4
Pressure GPa
Figure 12. Measurements of the coesite-stishovite transformation [ 171: Se. The phase boundary
(dotted line) is an extension of Yagi and Akimoto's results [54]: Y through our data points. Previous
large volume press data from [56]: A, [57]: S, [55,58]: Z are also presented in the figure (see text).
After [ 171.
are strong discrepancies (of order 2 GPa) between existing estimates of this bound-
ary [17,53-551.
The experiments were carried out in a C02-laserheated diamond cell using argon
as pressure medium. The starting materials were thin discs of Si02 glass. Samples
were typically heated for about 1-2min close to their melting temperatures to
avoid kinetic effects on the transition, and then temperature quenched by switching
off the CO, laser. Phase identification of the samples was carried out using Raman
spectroscopy.
The phase boundary was determined with three different experiments [17]. In the
first experiment Si02 glass was converted to stishovite and the phase boundary was
approached in 1 GPa increments until coesite was formed (Fig. 11). Analogously, in
the second experiment, coesite was heated at increasing pressures until stishovite
was formed. In the third experiment coesite or stishovite was synthesized directly
from the starting glass material in the vicinity of the phase boundary. These results
are summarized in Fig. 12 together with previous work on the phase boundary at
lower temperatures. We obtained a phase boundary that is represented by
P(GPA) = 7.8 + 0.0010 T(K). Our estimated slope is less than one half that
obtained by Zhang [ 5 5 , 5 8 ] using multianvil high pressure devices.
primary host minerals for A1203at these depths [59], containing structural elements
(polyhedra) of low density solids (tetrahedra), as well as structural elements of high
density phases, such as end-member MgSi03 ilmenite (octahedra), or perovskite
(octahedra and dodecahedra) [60]. It is known that end-member MgSi03 ilmenite
is a low-temperature, high-pressure phase, and transforms to the garnet structure
above 2100 K [61]. Both phases transform to MgSi03-perovskite above about
23GPa [61]. In order to investigate if the addition of aluminum changes the
transition pressure to the perovskite structure, and whether magnesian perovskite
can accommodate the available A1203 in the perovskite structure, we performed
C02-laser heating experiments on Mg3A12Si3012garnet to 45 GPa, and tem-
peratures close to the melting point. This doubles the experimental P-T range of
previous investigations [62,63]. To identify the phases we used Raman as well as
fluorescence spectroscopy. The latter allowed us to identify small amounts of
A1203, which is hard to detect using Raman spectroscopy. To do this, we used
Mg3Al2Si3OI2 : Cr3+ garnet as a starting material because a free A1203: Cr3+ phase
has a strong fluorescence signal. Moreover, the spectrum of A1203:Cr3+ can be
distinguished easily from that of Mg3Al2Si3OI2 :Cr3+. Typical Raman spectra of
pyrope, ilmenite, and perovskite are shown in Fig. 13. Figure 14a shows fluorescence
spectra of Mg3A12Si3012 : Cr3+ after heating in the perovskite stability field.
f i
after heating at 2600 K
30 GPa
h
c
v)
26.4 GPa
C
3
.-
E
I after healing at 2730 K Ilrn
-
v)
C
0,
-
C
Ilrn a
22.1 GPa
P
P IT
&LA
P
PP
21.2 GPa
I . 1 . 1 . I I . 1 B 7 '
I - I - 1 -
45.3GPa
R,/R,ruby
43.3 GPa
0 GPa
710 I I
I
708 -
706 -
I I 1 I
0 10 20 30 40 50
b Pressure (GPa)
Figure 14. (a) Room temperature fluorescence spectra after heating Mg3Al2Si3OI2 : Cr3+ pyrope
crystals at about 2500 K at various pressures. R1 and R2 are the primary electronic peaks due to
Cr3+-A1203, the peaks labeled with N are N-lines (Nebenlinien) and those labeled with S are
vibrational sidebands of Al2O3 [64]; the feature labeled with an asterisk is residual fluorescence
from unconverted pyrope. (b) Comparison of the pressure dependencies of the R I fluorescence
lines of a ruby chip (located at the edge of the gasket) and A1203:Cr3+ produced after heating
Mg3A12Si3012 :Cr". Agreement of the R I peak positions of the ruby chips and the Cr3+-A1203
in the center of the sample shows that no pressure gradients are present in the sample chamber
after heating. After [16].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.4 Phase Diagrams, Decomposition Reactions. and Stability of Solids 59
P(GPa)
Figure 15. Mg3A12Si3012-pyrope-ilmenitcperovskitephase relations. The diagram also includes
data points from earlier multianvil work by [63]: K and [62]: I. After [16].
Pyrope transforms to the ilmenite structure above 21.5 GPa, to perovskite plus
ilmenite above 24GPa, and to perovskite above 29GPa (Fig. 15). The stability
field of the ilmenite structure is expanded to higher temperatures with respect to
end-member MgSi03-ilmenite. The transition pressures to aluminous perovskite
is about 2GPa higher with respect to pure MgSi03. Furthermore, we found that
some A1203 exists as a free phase together with perovskite up to 43 GPa, because
we detected A1203:Cr3+fluorescence up to this pressure (Fig. 14). Our results
show that the perovskite structure can accommodate at least 25mole% A1203
above 43 GPa.
Figure 16. (A) SEM-image of the diamond agglomerate recovered from a diamond cell after heating
of octadecane to above 3000K at 21 GPa. The area bounded by the dashed lines is enlarged in (B),
showing two spheres of molten diamond (indicated by arrows) on the agglomerate surface.
rn
600 800
I ' I . , . I . , .
400 1000 1200 1400 4200 4400
rel. wavenumber (cm-')
Figure 17. Raman spectra of liquid H2 and cubic BN at 0.6 GPa in a DAC after heating of borane-
dimethylamine to about 3000 K at 23 GPa.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
62 2 Phase Transitions and Material Synthesis using the C02-Laser Heating Technique
after heating at 23 GPa. No evidence for the presence of ternary BNC, in the heated
sample volume was found.
2.6 Conclusion
The studies described above show that the C02-laser heated diamond anvil cell can
be used for investigating nonconducting materials under controlled conditions in a
very broad pressure and temperature range. Samples can be analyzed in situ using
Raman spectroscopy, fluorescence spectroscopy, and X-ray diffraction. For
structural, chemical, and textural analysis of the recovered samples techniques
such as atomic force microscopy, scanning- and transmission electron microscopy
can be employed. All of the above techniques can be used for determining melting
curves, developing high P-high T phase diagrams, evaluating the stability of solids,
and monitoring of chemical reactions, which are essential for controllable synthesis
and characterization of new materials.
Acknowledgments
We thank 0. Tschauner for fruitful discussions and the Editor of the book for
encouraging us to undertake this work. A.Z. gratefully acknowledges the financial
support of the Deutsche Forschungsgemeinschaft.
References
1. A. L. Ruoff, H. Xia, and Q. Xia, The effect of a tapered aperture on x-ray diffraction from
sample with a pressure gradient: Studies on three samples with a maximum pressure of
560GPa, Rev. Sci. Znstrum. 1992, 63, 43424348.
2 . M. Eremets, High Pressure Experimental Methods, Oxford University Press, Oxford, 1996.
3. R. M. Hazen and L. W. Finger, Comparative Crystal Chemistry, John Wiley & Sons Ltd.,
Chichester, 1982.
4. A. Jayardman, Diamond anvil cell and high-pressure physical investigations, Rev. Mod Phys.
1983,55, 65-108.
5 . W. L. Wolfe and G. J. Zissis, The Infrared Handbook, Office of Naval Research, Department of
the Navy, Washington, DC, 1985.
6. R. A. Forman, G. J. Piermarini, J. D. Barnett, and S. Block, Pressure measurement made by the
utilization of ruby sharp-line luminescence, Science, 1972, 176, 284-285.
7. H. K. Mao, J. Xu, and P. M. Bell, Calibration of the ruby pressure gauge to 800 kbar under
quasi-hydrostatic conditions, J . Geophys. Res. 1986, 91, 46734676.
8. R. Boehler, M. Nicol, and M. L. Johnson, Internally heated diamond anvil cell: Phase diagram
and P-V-T of iron, in High Pressure Research in Mineral Physics, Geophys. Monogr. Ser., M. H.
Manghnani and Y. Syono (Eds), AGU, Washington, DC, 1987, pp. 173-176.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Reterences 63
9. L. G. Liu and W. A. Bassett, The melting of iron up to 200 kbar, J . Geophys. Res. 1975, 80,
3777.
10. M. Planck, Entropie und Temperatur strahlender Warme, Ann. P h j x 1900, 1, 719-737.
11. L. Ming and W. A. Bassett, Laser heating in the diamond anvil press up to 2000°C sustained
and 3000°C pulsed at pressures up to 260 kilobars, Rev. Sci. Instrum. 1974, 45, 11 15-1 118.
12. C.-S. Yoo, J. Akella, and M. Nicol, Chemistry at high pressures and temperatures: In-situ
synthesis and characterization of P-Si3N4 by DAC X-ray/laser heating studies, in The 3rd
NIRIM International Symposium on Advanced Materials (ISAM'96), M. Akaishi et al. (Eds),
NIRIM, Tsukuba, Japan, 1996, pp. 175-179.
13. R. Boehler and A. Chopelas, A new approach to laser heating in high pressure mineral physics,
Geophys. Res. Lett. 1991, 18, 1147-1 150.
14. M. Eremets, K. Takemura, H. Yusa, D. Goldberg, Y. Bando, and K. Kurashima, Melting of
cubic and hexagonal BN, in The 3rd NIRIM International Symposium on Advanced Materials
(ISAM'96). M. Akaishi et al. (Eds), NIRIM, Tsukuba, Japan, 1996, pp. 169-174.
15. J. H. Nguyen and R. Jeanloz, Initial description of a new carbon-nitride phase synthesized at
high pressures and temperatures, Muter. Sci.Eng. 1996, A209, 23-25.
16. G. Serghiou, A. Zerr, A. Chopelas, and R. Boehler, The transition of pyrope to perovskite,
Phj:F. Chem. Min. 1998, 25, 193-196.
17. G. Serghiou. A. Zerr, L. Chudinovskikh, and R. Boehler, The coesite-stishovite transition in a
laser-heated diamond cell, Geophys. Res. Lett. 1995, 22, 441444.
18. F. Cabannes, Facteurs de reflexion et d'emission des metaux, J. Physique, 1967, 28, 235-248.
19. J. C. De Vos, A new determination of the emissivity of tungsten ribbon, Physica, 1954, 20,69&
714.
20. U. P. Oppenheim and A. Goldman, Infrared spectral transmittance of MgO and BaF, crystals
between 27" and IOOO'C, J . Optic. Soc. Am. 1964,54, 127-128.
21. A. Zerr, Schmelztemperaturen von (Mg, Fe)S03-Perowskit und von MgO bei hohen Drucken,
PhD thesis, University of Mainz, Mainz, 1995.
22. S. R. Elliot. Physics of Amorphous Materials, Longman, London, 1990.
23. R. J. Hemley, P. M. Bell, and H. K. Mao, Laser techniques in high-pressure geophysics,
Science, 1987, 237, 605-612.
24. G. Serghiou, A. Zerr, and A. Chopelas, Phase transitions and disordering of CaSi03 upon com-
pression, Rev. High Press. Sci. Techno/. 1998, 7, 359-361.
25. A. Chopelas, The fluorescence sideband method for obtaining acoustic velocities at high com-
pressions: application to MgO and MgA1204, Phys. Chem. Min.1996, 23, 25-37.
26. A. Chopelas and R. Boehler, Thermal expansivity in the lower mantle, Geophys. Res. Lett. 1992,
19, 1983-1986.
27. F. R. Corrigan and F. P. Bundy, Direct transition among allotropic forms of boron nitride at
high pressures and temperatures, J . Chem. Phys. 1975, 63, 3812-3820.
28. D. A. Young, C . 4 . Zha, R. Boehler, J. Yen, M. Nicol, AS. Zinn, D. Schiferl, S. Kinkead, R. C.
Hanson, and D. A. Pinnick, Diatomic melting curves to very high pressure, Phys. Rev. B, 1987,
35, 5353-5356.
29. P. E. van Keken, D. A. Yuen, and A. P. van den Berg, Implications for mantle dynamics from
the high melting temp,eratures of perovskite, Science, 1994, 264, 1437-1439.
30. %-I. Karato, Importance of anelasticity in the interpretation of seismic tomography, Geophys.
Res. Lett. 1993, 20, 1623-1626.
31. A. Zerr and R. Boehler, Constraints on the melting temperature of the lower mantle from high
pressure experiments on MgO and magnesiowustite, Nature, 1994,371,506-508.
32. B. Riley, The determination of melting points at temperatures above 2000" Celcius, Rev. Hautes
Temper. Refract. 1966, 3, 207-2 11.
33. A. Zerr, G. Serghiou, and R. Boehler, Melting of CaSi03-perovskiteto 430 kbar and first in-situ
measurements of lower mantle eutectic temperatures, Geophys. Res. Lett. 1997, 24, 909-912.
34. A. Zerr and R. Boehler, Melting of (Mg, Fe)Si03-perovskite to 625 kbar: indication of a high
melting temperature in the lower mantle, Science, 1993, 262, 553-555.
35. A. Zerr, A. Diegeler, and R. Boehler, Solidus of Earth's deep mantle, Science, 1998,281,243-246.
36. J. Zhang and C. Herzberg, Melting experiments on anhydrous peridotite KLB-1 from 5.0 to
22.5 GPa, J . Geophys. Res. 1994,99, 17729-17742.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
64 2 Phase Transitions and Material Synthesis using the CO,-Laser Heating Technique
37. Z. Gong, R. E. Cohen, and L. L. Boyer, Molecular dynamics simulations of melting of MgO at
high pressures, in Annu. Rep. Director Geophys. Lab. Carnegie Instn. Wash., Washington, DC,
1991, pp. 129-134.
38. I. Jackson, Phase relations in the system LiF-MgF2 at elevated pressures: Implications for the
proposed mixed-oxide zone of the Earth’s mantle, Phys. Earth Planet. Inter. 1977, 14, 86-94.
39. E. Ohtani, Melting temperature distribution and fractionation in the lower mantle, Phys. Earth
Planet. Inter. 1983, 33, 12-25.
40. J. J. Gilvarry, The Lindemann and Gruneisen laws, Phys. Rev. 1956, 102, 308-316.
41. F. A. Lindemann, Uber die Berechnung molekularer Eigenfrequenzen, Phys. Z . 1910,11,609-
612.
42. E. A. Kraut and G. C. Kennedy, New melting law at high pressures, Phys. Rev. Lett. 1966,16,
608-609.
43. R. Boehler, Adiabats of quartz, coesite, olivine, and magnesium oxide to 50 kbar and 1000K,
and the adiabatic gradient in the Earth’s mantle, J . Geophys. Res. 1982,87, 5501-5506.
44. D. Heinz and R. Jeanloz, Measurement of the melting curve of Mgo 9Feo Si03 at lower mantle
conditions and geophysical implications, J. Geophys. Res. 1987,92, 11437-1 1444.
45. E. Knittle and R. Jeanloz, Melting curve of (Mg, Fe)Si03 perovskite to 96 GPa: evidence for a
structural transition in lower mantle melts, Geophys. Res. Lett. 1989, 16,421424.
46. J. S. Sweeney and D. L. Heinz, Melting of iron-magnesium-silicate perovskite, Geophys. Res.
Lett. 1993, 20, 855-858.
47. G. Shen and P. Lazor, Measurement of melting temperatures of some minerals under lower
mantle pressures, J . Geophys. Res. 1995, 100, 17699-17713.
48. E. lto and E. Takahashi, Melting of peridotite at uppermost lower-mantle conditions, Nature,
1987,328, 514-517.
49. E. Takahashi, Melting of a dry peridotite KLB-1 up to I4GPa: Implications on the origin of
peridotitic upper mantle, J. Geophys. Res. 1986, 91, 9367-9382.
50. K. G. Holland and T. J. Ahrens, Melting of (Mg, Fe)2Si04at the core-mantle boundary of the
Earth, Science, 1997, 275, 1623-1625.
51. J. Haines, J. M. Leger, and 0. Schulte, Pa3 modified fluorite-type structures in metal dioxides at
high pressure, Science, 1996, 271, 629-631.
52. J. M. Leger, J. Haines, M. Schmidt, J. P. Petitet, A. S. Pereira, and J. A. H. Jornada, Discovery
of hardest known oxide, Nature, 1996, 383, 401.
53. V. Swamy, S. K. Saxena, B. Sundman, and J. Zhang, A thermodynamic assessment of silica
phase diagram, J. Geophys. Res. 1994,99, 11787-1 1794.
54. T. Yagi and S. Akimoto, Direct determination of coesite-stishovite transition by in-situ X-ray
measurements, Tectonophysics, 1976,35, 259-270.
55. J. Zhang, B. Li, W. Utsumi, and R. Liebermann, Reversal of the coesite-stishovite phase trans-
formation, in AGU Spring Meeting, American Geophysical Union, Baltimore, 1994, p. 346.
56. S. Akimoto and Y. Syono, Coesite-stishovite transition J . Geophys. Res. 1969,74, 1653-1659.
57. K. Suito, Phase relations of pure Mg2Si04up to 200 kilobars, in High Pressure Research: Appli-
cation in Geophysics, M. H. Manghnani and S. Akimoto (Eds), Academic Press, Inc., New
York, 1977, pp. 255-266.
58. J. Zhang, R. C. Liebermann, T. Gasparik, and C. T. Herzberg, Melting and subsolidus relations
of Si02 at 9-14GPa, J. Geophys. Res., 1993,98, 19785-19793.
59. A. E. Ringwood, Composition and Petrology of the Earth’s Mantle, McGraw-Hill, New York,
1975.
60. P. McMillan, M. Akaogi, 1. Ohtani, Q. Williams, R. Nieman, and R. Sato, Cation disorder in
garnets along the Mg3A12Si3012-Mg4Si,0,2join: an infrared, Raman and NMR study, Phys.
Chem. Min. 1989, 16,428435.
61. H. Yusa, M. Akaogi, and 1. Eiji, Calorimetric study of MgSi03 garnet and pyroxene: heat
capacities, transition enthalpies, and equilibrium phase relations in MgSi03 at high pressures
and temperatures, J . Geophys. Res. 1993, 98, 64534460.
62. T. Irifune, T. Koizomi, and J. Ando, An experimental study of the garnet-perovskite transfor-
mation in the system MgSiO3-Mg3AI2Si3Ol2,Phys. Farth Planet. Inter. 1996, 15, 90-106.
63. M. Kanzaki, Ultra high-pressure phase relations in the system MgSiO3-Mg3AI2Si3Ol2,Phys.
Earth Planet. Inter. 1987, 49, 168-175.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 65
64. L. Zhang and A. Chopelas, Sound velocity of A1203 to 616kbar, Phys. Earth Planet. Inter.
1994,87, 77-83.
65. S. K. Saxena, L. S. Dubrovinsky, P. Lazor, Y. Cerenius, P. Haggkvist, M. Hanfland, and
H. Jingzhu, Stability of perovskite (MgSi03)in the Earth’s mantle, Science, 1996,274, 1357-1359.
66. G. Serghiou, A. Zerr, and R. Boehler, (MgFe)Si03-perovskite stability under lower mantle
conditions, Science, 1998,280, 2093-2095.
67. R. H. Wentorf, The behavior of some carbonaceous materials at very high pressures and high
temperatures, J. Phys. Chem. 1965,69, 3063-3069.
68. M. Ross and F. H. Ree, Repulsive forces of simple molecules and mixtures at high density and
temperature, J Chem. Phys. 1980, 73, 614M152.
69. F. Ancilotto, G. L. Chiarotti, S. Scandolo, and E. Tosatt, Dissociation of methane into hydro-
carbons at extreme (planetary) pressure and temperature, Science, 1997, 275, 1288-1290.
70. A. Zerr, G. Serghiou, and R. Boehler, C02-laser heating in a diamond cell to 80GPa
and 5000 K: phase transitions and synthesis under controlled conditions, in The 5th NIRIM
International Symposium on Advanced Materials (ISAM’98), M. Kamo et al. (Eds), NIRIM,
Tsukuba, Japan, 1998, pp. 5-8.
71. S. K. Sharma, H. K. Mao, and P. M. Bell, Raman measurements of hydrogen in the pressure
range 0.2-630 kbar at room temperature, Phys. Rev. Lett. 1980, 44,886-888.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
3.1 Introduction
In this contribution we review current knowledge of the mechanical properties of
hard materials and the relation between these properties and their microstructure.
However, a quantitative relationship is only meaningful to the extent that test
methods are clearly defined and understood, while the test data are only useful
within the context of the engineering requirements which determine material
selection for a specific application. It is also important to define what is meant by
a ‘hard material’.
We therefore discuss briefly some typical applications of hard materials and the
engineering requirements that must be satisfied in these applications. In Section
3.2 we define the mechanical properties which need to be evaluated, while in Section
3.3 we review the test method which have been developed to ensure accurate and
reproducible measurements. We then examine the microstructure-property rela-
tions and the size effects that dominate the mechanical strength of brittle materials,
as well as some important failure mechanisms.
may be beneficial, inhibiting subcritical crack growth and delaying the onset of
damage, but residual tensile stresses are always detrimental. Mismatch in the
thermal expansion coefficient and the elastic moduli of the coating and the sub-
strate are common sources of residual stress.
It is important to distinguish between elastic and plastic penetration of an
indentor. An elastomer has a low resistance to elastic deformation, but releasing
the load will generally leave no mark on the rubbery elastomer. By contrast, a
refractory metal may have a poor resistance to plastic penetration, but show rela-
tively little elastic relaxation when the indentor load is relaxed [3]. Throughout
this review we assume the elastic response to be small (but not negligible) compared
to the plastic response, so that the high hardness of a ‘hard material’ is also accom-
panied by high stiffness, corresponding to a large tensile modulus.
The mode of loading and the rate at which load is applied at the surface is
important [3]. There is a clear distinction between a sliding bearing, in which
contacts between the components are predominantly subjected to shear parallel
to the contact surface, and a roller (or ball) bearing in which the mechanical con-
tacts are made and broken in tension (perpendicular to the contact surface). Most
mechanical contacts can be approximated to ‘point’ contacts, since the ‘perfectly
flat’ surface has yet to be invented. More exactly, the ‘true’ contact area depends
on the load applied, the rigidity of the components, and their yield strength. A high
hardness implies a minimum contact area at a given normal load. It follows that
contact damage is, in the first instance, localized and restricted to the contact
regions. Mechanical wear occurs by removal of material from points of contact,
followed by transfer of the load to new points of contact. Areas worn down by
mechanical wear will establish renewed contact once neighboring areas are
similarly eroded, leaving the original contact areas once more proud of the nom-
inal contact surface.
Since surface damage is localized, events taking place at the surface often occur
rapidly. True contact areas are often on the scale of the surface finish, less than a
micron for a well-polished surface. Sliding velocities of the order of meters s-’
then correspond to contact times of less than a microsecond, so that the localized
damage events are commonly adiabatic, and accompanied by intense ‘frictional’
heating [ 7 , 8 ] . Hence the importance of cooling the workpiece in all grinding,
polishing, and mechanical working operations. Susceptibility to brittle failure is
the penalty accompanying the desirable properties of hard materials [9]. Premature
brittle failure can be prevented, either by suitable engineering design or by ensuring
that brittle crack propagation is inhibited. In successful engineering design tensile
loading is minimized, since brittle materials are commonly strong in compression.
This may be achieved either by pre-loading in compression (assembly of the compo-
nent in constraint) or by ensuring that the operating loads are compressive. Fracture
toughness is the common parameter used to measure the resistance to brittle crack
propagation [9-111, but many hard materials exhibit R-curve behavior, in which the
critical stress intensity factor increases with crack length [9,12]. The mode I fracture
toughness determines the resistance to macroscopic crack propagation, but the
short-crack resistance or ‘damage tolerance’ [131 often determines the performance
in service, when damage is commonly localized at the surface.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
68 3 Mechanical Properties and their Relation to Microstructure
Clearly, the primary engineering requirements for bearing components and cut-
ting tools include high stiffness and hardness, maintained to high temperatures
[14]. These are also the primary considerations in the selection of grit for grinding
wheels and polishing compounds. Erosion-resistant surface coating systems are
often based on hard and rigid materials, which may also employ elastomeric
materials to resist plastic but not elastic penetration. Armor systems are also
designed to resist penetration, at both sub- and supersonic velocities, and also
employ hard materials [15,161.
Finally, two other characteristics of hard materials are important: their density
and their chemical stability. The importance of these properties depends on the
application. Oxides are chemically more stable than nitrides, which are in turn
more stable than borides or carbides. The susceptibility to oxidation of boron
carbide prevents its application at high temperatures, but for ballistic protection,
where hardness, rigidity and low density take precedence, boron carbide is
unparalleled. Tungsten carbide, on the other hand, is the material of first choice
for cutting tools, because of its high hardness and stiffness, and high-temperature
oxidation resistance, but is far too dense for application in which weight or
inertial forces are important. Its use for ballistic protection is therefore out of
the question.
and the hard carbide phases are continuous. Since the ductile binder is continuous,
no crack can propagate without expending energy in ductile failure of the binder. On
the other hand, the three-dimensional continuity of the carbide skeleton maximizes
rigidity and stiffness of the composite for any given carbide content, so toughness is
obtained with a minimum loss of hardness.
Bulk cutting tools based on alternative materials, most notably alumina and
silicon nitride, have been developed [ 19,201, but, although niche applications have
been found, no substitute for tungsten carbide has yet been discovered. The closest
to a ‘breakthrough’ has been the development of polycrystalline diamond (which
lacks oxidation resistance) [21]. More recently, cubic boron nitride (cBN) has
found a large and growing market in high speed machining thanks to a combination
of hardness and oxidation resistance [22,23].
this application lack of toxicity is a primary concern, while the wear path is
limited, so that wear is not a major concern.
3.1.3 Coatings
Hard coatings can solve the problem of maximizing the hardness of the surface, to
optimize wear resistance, without sacrificing the toughness of the bulk component
and risking brittle fracture [26]. A successful hard coating must be compatible
with the substrate, exhibiting good adhesion and minimizing residual stress in the
coating. A wide spectrum of coating technologies is available, encompassing
electrolytic coatings, anodizing, diffusion coatings, plasma spraying, chemical
vapor deposition, laser deposition, and more. The various coating technologies
have one thing in common: the extraordinary difficulty in determining their mechan-
ical properties. The problem can be reduced to three separate factors: The mismatch
in elastic properties between the coating and the substrate, which leads to interfacial
shear stresses at any discontinuity in the coating; the dependence of the mechanical
properties on thickness, both because of thickness-dependent microstructural
changes and owing to a size dependence of the properties; difficulties in standardiz-
ing and interpreting mechanical tests on coatings.
These considerations are independent of other problems: surface roughness
and adhesion, which are usually avoided by assuming, or requiring, that the
adhesive strength of the coating-substrate interface exceeds the cohesive strength
of the coating, and residual stresses in the coating, typically biaxial, which also
affect the critical load for microcracking, peeling, and (eventually) spalling
(flaking).
where a and ,B are constants related to the attraction and repulsion components, r is
the distance between the atoms, and n > m. All four constants can be treated as
material properties. The net force between the atom pair is just
and is zero for r = ro, the equilibrium spacing in the absence of external forces. An
external driving force displaces the atoms by a distance u from the equilibrium
position, where u = r - ro, and the relationship between the external driving force
and the displacement is
F=-=-
dEp u. (3)
dr U=O
The second derivative, which determines the curvature, of the potential energy well
near y o , is constant for a given material and temperature, leading to Hooke's law:
*I
F = Ku. The constant K is proportional to the elastic, or Young's modulus
K o(
dU2 "=o
= f ( a ,p, n, m). (4)
Ec = cn
EiL,
whereh is the volume fraction of component i. The Reuss model gives a lower
bound for the elastic modulus, and assumes compatible loads or stresses:
1 J;:
E,=FE+
Most polyphase materials have elastic moduli which lie between these two bounds,
and more sophisticated models have been developed to narrow the gap. The
Hashin-Shtrikman model, in particular [27], has given better predictions for the
elastic modulus of a composite, while, more recently, Aboudi has used a cell
model [28] and finite element analysis has also been used to derive the relationship
between the volume fraction and effective modulus [29-321. These methods also
estimate the displacement and stress field in the composite.
3.2.2 Strength
The mechanical strength of hard materials is critical for load-bearing, structural
applications. These brittle materials only deform plastically at high temperatures,
or under severe hydrostatic constraint, since the Peierls stress for dislocation move-
ment is high. Failure is usually by unstable crack propagation under a tensile stress
that exceeds the tensile strength of the material. In terms of fracture mechanics, brit-
tle failure occurs when the Mode I stress intensity factor KI reaches the fracture
toughness of the material, K I c (see below).
An experimental uniaxial stress-strain relationship, determined in tension, would
provide the necessary design information for ceramics and hard materials: Young’s
modulus, Poisson’s ratio and the tensile fracture strength. However, tensile data for
brittle materials are often unreliable, due to parasitic bending stresses associated
with dimensional inaccuracy in machining tensile ‘dog-bone’ specimens and mis-
alignment of the specimen grips.
The test procedure most commonly used to validate the strength of ceramics and
hard materials is a flexure test, in either three-point (3PB) or four-point bending
(4PB). The specimens for bend tests are simple to machine and require less material,
while misalignment and parasitic bending can be avoided. The strength determined
in bending is variously termed the flexural strength, the transverse rupture strength,
TRS, or the modulus of rupture, MOR, and corresponds to the maximum tensile
stress developed in the surface ligaments of the sample at the point of fracture.
Several issues need to be addressed when a bend test is used to evaluate the
strength: The sources of errors that are responsible for statistical scatter in the
results; the statistical nature of the strength of a brittle material; the relationship
between the measured flexural strength and the strength derived from a valid
uniaxial tensile test, together with the effect of sample size on the strength; and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.2 Principal Mechanical Properties 73
Loadpointfriction u
. - Twisted specimen
orfixture
n n
Reducedmomentofinertia
Figure 1. Source of errors in bending tests. Reproduced with permission of . . . . . . . . . . (Ref. 33).
F = I -exp[-jv[g]"du], (7)
where uoand m denote the median failure stress [35,36] and the Weibull modulus, u,
is the volume of the test piece. The material parameter uomeasures the amplitude of
the failure strength, while the parameter rn measures the strength variability. If the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
74 3 Mechanical Properties and their Relation to Microstructure
failure probability function for N specimens is known, then the failure probability of
+
the ( N 1)th specimen can be predicted. The number of tests N required to ensure
predictive reliability is typically 20.
If a 3PB test piece is loaded, the volume element subjected to the maximum tensile
stress is very small. This volume is much larger for a 4PB specimen, but a uni-
directional tensile specimen will have the largest volume under maximum stress. It
follows that the median strength of a 3PB specimen may exceed that of a 4PB
specimen, which in turn should have a higher average strength than a tensile
specimen. The measured strength ratio between specimens of identical size subjected
to 3PB and to pure tension is predicted to be [34,37]:
which yields 1.73 for m = 10, a typical value for a structural ceramic. The predicted
strength ratio between a 4PB specimen (inner/outer span = 0.5) and a same-size
tensile specimen is:
(9)
ffT
which gives a strength ratio of 1.45 for same value of m = 10. Clearly, the loading
geometry must be considered when comparing the strength of ceramic components.
Actually, it was Leonard0 da Vinci (1452-1519) who first noticed that the
strength of steel wire increased for decreasing wire length. Indeed, assuming uniform
stresses in test pieces having different volume but the same failure probability,
Weibull statistics predicts a strength ratio:
which yields the value of 2 for V2 = 1000V1and m = 10, a typical situation for some
small test samples and the corresponding large engineering component.
The elastic modulus, the fracture toughness, the density, and the microstructure,
by contrast, are not usually sensitive to the size of the engineering component, and it
is principally the strength which is size-sensitive, and hence not a true material
parameter.
This has been demonstrated for a commercial glass ceramic [38], and for alumina
[39]. It has been suggested that the microstructure and the defect populations them-
selves may vary throughout the volume of a component, causing variations in
strength unpredictable by Weibull analysis [40,41]. An attempt to verify the minimal
specimen size required for strength determination was carried out lately [42].
Figure 2. The Griffith cracked body A, with free surfaces S, and crack length 2a.
mechanics. Griffith calculated the balance between mechanical energy and internal
energy in a solid body containing a crack:
u = -FL UE us, (11)
where FLand UE are the external work and elastic strain energy stored in the body
under stress, respectively and Us is the surface free energy, Fig. 2. For linear elastic
bodies, FL = 2uE. UE was calculated from Inglis' [45] expressions for the stress field
in a body containing an elliptical hole. An instability occurs when the change in the
mechanical energy exceeds the surface energy required for a unit increase in crack
area:
dU
-= 0.
dA
This critical point marks the onset of crack propagation. Griffith calculated the
critical stress far from the crack location for a plate of unit width, and showed
that for a homogeneous, isotropic material, a relationship exists between the critical
stress, uc and half the crack length, a:
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
76 3 Mechanical Properties and their Relation to Microstructure
J.
I
Figure 3. The three basic modes of fracture, distinguished by the displacement driving force: I, the
opening mode; 11, the shear mode; and 111, the out of plane shear mode.
where ”is is the surface energy, E* = E for loading under plane stress conditions, and
E* = E/(1 - v2) for loading under plane strain, E and v are the Young modulus
and Poisson ratio respectively, assuming homogeneous, isotropic materials.
Eq. (13) does not involve the local stress field in the vicinity of the crack tip, and
that the only energy characteristic of the material is the surface free energy.
Three possible deformation ‘modes’can be defined, Fig. 3. Mode I, is a tensile mode
characterized by an identical displacement normal to the crack plane for any two
adjacent points on the free surface of the crack. Mode I1 is an in-plane shear mode,
in which the two adjacent points have the same (but with opposite sign) displacement
in the plane of the crack and parallel to the direction of crack propagation. Mode I11
corresponds to an out-of-plane shear mode; with shear displacements in the plane of
the crack but perpendicular to the direction of crack propagation. Crack propagation
seldom occurs under a pure tensile stress, and shear stresses are commonly present in
the early stages of crack propagation, so that most crack propagation takes place, to
some extent, under mixed-mode loading conditions.
Irwin [461and Orowan [47] generalized the Griffith energy balance, using the term
‘strain energy release rate’ (SERR), 9,to characterize a general relationship between
crack length, sample geometry and the mode of loading. In mode I deformation,
assuming unit thickness of the cracked body, is defined by:
or
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.2 Principal Mechanical Properties 77
P P
P,6
Figure 4. Strain energy release during constant displacement and constant load.
where U is the strain energy stored in the cracked body, 7r the complementary
energy, a the projected crack length, and da is crack increment, as shown in
Fig. 4a. For an imposed displacement (displacement control) the strain energy is
reduced for increasing crack length, Fig. 4b, and the crack advances stably, while
for a fixed applied load (load control), Fig. 4c, the strain energy increases for
increasing crack length, and the crack advances unstably, if brittle materials is
considered.
Increasing the external applied driving force will increase the strain energy release
rate until the strain energy release rate reaches a critical value gIC, which defines a
failure criterion for a brittle, flaw-containing component. The material property
qIc,is variously referred to as the toughness, the fracture energy, or the work of
fracture. gICis determined by the rate at which the stored strain energy is relaxed
when a crack advances. The critical ‘strain energy release rate’ per unit area of crack
advance has units of N m-’ or J mP2.The lower bound for the work of fracture must
be twice the surface energy, T ~ since
, two new surfaces are created by the advancing
crack. The actual value of the fracture energy will always be much higher than that
predicted from the equilibrium surface energy, since fracture is irreversible (and
approximately adiabatic) and additional energy dissipating mechanisms are
active: localized friction between microfacets, mechanical interlocking of grains
which bridge the crack, and some possible plasticity at the crack root.
An alternative stress analysis approach, is based on the local stress field near a
crack tip, Fig. 5 . The solution of the boundary value problem for a semi-infinite,
linear-elastic cracked body was found by Williams [48], and yields the Williams’
expansion of the stress field in a cracked body. The first term determines the local
stress field in the vicinity of the crack tip:
where r is the distance from the crack tip, 6’ is the angle (Fig. 5), and M denotes the
mode of deformation. The parameter Kis the stress intensity factor (SIF), which is a
function of geometry and loading only, and not a material property. The nature of
the near-stress field is that of square root singularity in distance from the crack tip
[9,10]. K is the amplitude of the stress field singularity near the crack tip.
As the external load increases, the stress intensity approaches the fracture tough-
ness, KIc , and crack propagation is initiated. This alternative failure criterion
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
78 3 Mechanical Properties and their Relation to Microstructure
/”
crack tip
X
known, there is an exact relationship between the fracture toughness and the
remote stress:
An analogous relationship holds for the fracture toughness and the fracture
energy:
Excellent texts on fracture mechanics are available [9-111.Related topics are failure
criteria for the onset of cracking in laminated systems [55],dynamic fracture
mechanics [56],and nonlinear fracture mechanics [9,1I].
3.2.4 Hardness
Hardness is commonly defined as the resistance of a material to static penetration by
a harder material [3].The qualifier, ‘static’, excludes dynamic effects associated with
erosion or ballistic penetration. Hardness can be quantified by determining the
depth of penetration of a hard indentor, usually diamond, loaded perpendicular
to a planar surface of the material under test [57].The penetration depth is related
to the contact area over which the load is transferred between the indentor and the
sample. The ‘hardness’ is equal to the load divided by the area of plastic contact.
There is some ambiguity in ignoring recoverable, elastic penetration, but, apart
from elastomeric materials, the effects of elastic recovery during unloading are
usually small. The fixed load, P , applied to the indentor and the known indentor
geometry are used to calculate the hardness from the contact area, A , between the
indentor and the sample:
P
H --, (20)
- A
where H , is an average applied compressive stress, generally measured in GPa. In
polycrystalline solids, plastic deformation beneath the indentor is associated with
dislocation multiplication and glide [58].In metallic solids the critical flow stress
or yield stress of the material, oY,governs the hardness [59].There is therefore a
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
80 3 Mechanical Properties and their Relation to Microstructure
direct relationship between the hardness and the plastic flow stress, often expressed
as:
H, = Cgy, (21)
where C M 3 for a perfectly elastic, plastic solid [60]. The large proportionality
factor reflects the elastic constraint exerted on the plastic zone beneath the indentor
by the surrounding elastic material. No such constraint exists in the unidirectional,
tensile or compressive tests used to determine the uniaxial flow stress. When the
constitutive relationship describing the mechanical response is more complicated,
or for a brittle material (when the plastic response depends on the hydrostatic
component of the constraint), the proportionality constant may be appreciably
higher [61]. Amorphous, glassy materials posses a clearly defined hardness, related
to the Young modulus [62]. The same elastic constraint inhibits cracking in the
vicinity of the indentor, ensuring that some plastic flow occurs beneath the area
of elastic contact and precedes crack nucleation and propagation, even in very
brittle materials.
From the engineering perspective, a hardness test is an ideal method for monitor-
ing the mechanical properties of hard materials, since minimal sample preparation is
required and the test can be performed on actual components using simple
apparatus operated at low loads. The hardness test may also be considered ‘non-
destructive’, since components can often be put into service after testing. The
mechanical performance of hard materials is often ranked by their indentation
hardness, which in itself accounts for both the popularity and the technological
success of this simple, cost effective test.
The measured hardness of any material depends on parameters associated with
the test method and indentor geometry (Brinell, Vickers, Knoop, etc.), so that hard-
ness is not an intrinsic bulk property, comparable to elastic modulus, yield strength
or fracture toughness. In general, the measured hardness varies with the applied
load and the indentor shape and dimensions, but also with the microstructure
and prior history of the material, as well as with the loading rate, the environment
and the test temperature [3,63-671. In crystalline materials the resistance to disloca-
tion glide determines the hardness, and glide is activated by the shear component of
the compressive stress generated beneath the indentor. The maximum strain beneath
an indentor (elastic and plastic) may be some two orders of magnitude greater that
the tensile strain to failure of a brittle solid [68]. Such conditions are seldom encoun-
tered by hard materials operating in engineering systems. The strain rate in an
indentation test is also high when compared to a static tensile test, since the velocity
of the indentor is much greater than the usual crosshead speed in tensile testing. A
high strain rate increases the flow stress ((~y),the effect increasing with indentation
load and rate. By contrast, at high homologous temperatures thermally activated
creep is possible, reducing hardness with increasing temperature and decreasing
loading rate [69].
Indentation testing dominates the determination of hardness, but some alter-
native tests are available: The scratch test ranks materials in order of their ability
to scratch materials lower on the scale [70], and assigns a ‘Mho’ hardness (1822)
to each material, with diamond having the highest value of 10 on the Mho scale.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.3 Mechanical Testing of Hard Materials 81
by (i + 0.3)/(N - 0.4), i being the rank order of the ith test result. The Weibull
modulus m is estimated from the slope of the graph of In(ln[l/(l - F ) ] ) vs. lnc,
as shown for various geometries and processing routes, Fig. 6 . A detailed discussion
of Weibull strength analysis is given in [37]. The determination of m should always
be made by graphical plotting of the test results, since alternative failure mechan-
isms can give rise to different median strengths and Weibull moduli. These may
be associated with the probability of failure initiation at a surface as opposed to a
bulk defect. Surface flaws commonly have a wider size distribution, giving a
lower Weibull modulus, while bulk flaws, with a narrower distribution, correspond
to a higher value of m [71-741. In consequence, the slope of the Weibull plot changes,
increasing for the higher strength values, dominated by the size and distribution of
the bulk defects. A full understanding of the In In[I / ( 1 - F ) ] vs. In c curve requires
fractographic analysis of the failed samples and correlation of the observed failure
origins (surface cracks, bulk porosity, hard aggregates, inclusions) with the
measured strength.
Various internationally recognized standards exist for the determination of flex-
ure strength (3PB or 4PB):ASTM C1161 and MIL STD 1942 (USA), DIN 51-1 10-1
(Germany), B41-104 (France), JIS R 1601 (Japan), GB 8489 (China). The specimens
are prismatic rectangular bars. Commonly specified outer spans are 20, 40, or
80 mm, with corresponding 4PB inner spans of 10, 20, and 40 mm, respectively.
The specified cross section (height x width, all dimensions in mm) varies from
1.5 x 2 (American) and 2 x 4 (French) for small-span specimens, to 3 x 4 for the
medium span, and 6 x 8 for the large span.
Other test samples and loading geometries are sometimes required for specific
components or applications. The biaxial strength of thin plate is commonly
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.3 Mechanical Testing of Hard Materials 83
fracture toughness results obtained from such precracks have proved to be highly
reproducible [82-841.
Controlled precracking of a thin, notched specimen can be achieved by gluing the
sample to the side face of a ductile metal beam loaded in 3-point bending. Crack
arrest in the thin, brittle specimen occurs at the point where the stress in the thick
supporting beam becomes compressive [85]. Precracking of ceramic specimens by
cyclic fatigue of notched specimens at a slowly decreasing stress intensity factor
has also been reported [86]. Most applications of these test methods for
precracked-beams loaded by 3PB evaluate the plane-strain fracture toughness
using the methodology summarized in the ASTM standard E-399 [75].
The precise measurement of crack length is also difficult. Partial crack closure
after precracking, upon removal of the load, may preclude accurate crack length
measurement. The use of die penetrant as a visual crack marker is helpful but
inaccurate. Optical inspection at low magnifications is usually adequate, providing
the specimen finish is good. The fracture toughness of materials generally quoted in
the literature is the static or quasi-static fracture toughness, and is a poor guide to
dynamic behavior, slow crack growth, or mechanical fatigue strength.
Negotiations for an international fracture toughness standard are well advanced
within the framework of ASTM and VAMAS (Versailles Agreement on Materials
and Standards) [87]. The test geometry which is expected to become an ASTM
standard is the chevron-notched three-point bend beam. The chevron notch is cut
with a diamond saw, and the crack grows at low loads from the tip of the notch.
Stable crack growth continues until the crack front reaches the full cross-section
of the beam, when growth becomes unstable at the point of maximum load, the
critical load for fracture toughness evaluation. The fracture toughness determined
from a precracked beam may give values which differ significantly from those
determined from a chevron notched specimen. This difference is usually material-
dependent and it is not possible to predict which test will give the lower (more
dependable) value.
Alternative geometries and test methods to determine the fracture toughness of
ceramics and hard materials have been developed over the last decades: Herzian
indentation [88], the double torsion test [89], and compact tension tests [90-921.
3.3.4 Hardness
Microhardness tests in metals are performed under very low load [3], 0.01-2N,
which nevertheless activates dislocation glide. The higher values of microhardness
characteristic of hard materials require higher loads, typically 0.05-0.5 N, which
do not usually initiate microcracking (indentation cracking). At still higher loads,
2-100 N, microcrack initiation and propagation is expected in brittle materials,
resulting in higher values for the measured hardness, since microcracking increases
the compliance, and hence the elastic component of penetration, which is relaxed
(apart from frictional effects) when the load is removed. In addition to plastic
deformation, other viscous flow mechanisms may operate under high loads, such
as densification in both glasses and ceramics containing residual porosity [93]. It
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.3 Mechanical Testing of Hard Materials 85
follows that hardness tests should be performed over a range of load to ensure that a
load-independent value of this property can be defined [66],Fig. 7. In hard materials
a high-angle, pyramidal indentor, generates high residual tensile stresses on load
removal which may exceed the tensile strength, resulting in microcrack initiation
and propagation.
Valid hardness measurements require that the indentation be either very much
less than the grain size, in order to characterize the microhardness of individual
grains, or else appreciably greater than the grain size, to ensure that it is the hardness
of the polycrystal that is determined. Typical indent sizes, using standard micro-
hardness testing systems, are in the range 3-50pm, and up to 200pm in the low
load regime. A good optical microscope has a resolution of the order of 0.5pm
and should be capable of achieving 10% accuracy providing several measurements
are made to monitor the reproducibility.
The methods which have been developed to determine hardness vary in indentor
geometry and definition of the contact area. In Brinell’s method, a hard sphere of
diameter D is indented into the softer sample, and the Brine11 hardness is calculated
from the diameter of the circular indent which is left in the surface after removing
the load:
where d is the diameter of the impression. In the Rockwell method the depth of
penetration of the indentor, usually but not always a diamond cone, is used to
calculate the Rockwell hardness number on one of several scales (Rockwell A,
B, C , or D), depending on the applied load and the geometry of the indentor.
The Meyer hardness number is derived from the projected area of the residual
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
86 3 Mechanical Properties and their Relation to Microstructure
impression:
P
HM =4-.
?rd2
In the Knoop method an elongated diamond indentor with a specified ratio of the
two diagonals is used. Measurements taken from the long diagonal of the indent can
be used to assess hardness anisotropy by rotating the indentor about the loading
axis between measurements. The Berkovitz microhardness indentor employs a trian-
gular pyramid which avoids variations in hardness associated with anisotropy in the
plane of the surface and ensures that the indentor tip is truly a point (inaccuracy in
cutting a square pyramid results in a chisel edge at the apex). In the Vickers method
a square diamond pyramid with a standard apex angle of 136" is used and the
contact area is determined from the mean length of the two diagonals of the
indent, 2a, which is then translated into a Vickers hardness number:
P
Hv = 0.464-. (26)
a2
It is the Vickers test which is most commonly used to specify the hardness of hard
materials, with results often quoted in kgmm-2, but more correctly in SI units, GPa.
The repeated measurement of the size of indents, and the interpretation of indent
geometry for the purposes of calculation, may be tedious, and operator bias is
almost unavoidable. The edge of the impression is not always well defined, and mis-
leading edge effects may be associated with anisotropic plasticity or plastic recovery.
Faceted and elongated grains, or other microstructural features, together with the
limitations of contrast and resolution in the optical microscope, complicate the
interpretation, while the shape of the indent may differ in different materials: so-
called 'pin-cushion' or 'barreled' indents, associated with different constitutive
relations and frictional shear on the faces of the indentor in contact with the plastic
zone [3]. Mismeasurement of indent size is a major source of scatter in the experi-
mental data and the relative errors in the results of different operators.
Figure 8. Typical plastic deformation and cracking after indentation toughness test. Reproduced
with permission of . . . . . . . . . . (3).
useful test for monitoring damage tolerance and the resistance to short crack propa-
gation. The measured values should not be interpreted as ‘fracture toughness’.
Four microcrack morphologies may develop in a brittle solid after formation of
the plastic zone beneath the indent: median cracks, Fig. 9a, radial cracks (a half-
penny shaped elliptical crack Fig. 9b, much larger than the median crack), lateral
cracks (in a plane normal to the median and radial cracks), Fig. 9c, and Palmqvist
i I_2c‘ *I
I
I I
! I
a b
\- Median
mqvlst radial
vent
I d
C
Figure 9. Cracks formed during an indentation toughness test: (a) median cracks; (b) radial cracks;
(c) lateral cracks; and (d) Palmqvist cracks. Reproduced with permission of . . . . . . . . . . (3).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
88 3 Mechanical Properties and their Relation to Microstructure
radial cracks (originating near the edge of plastic zone beneath the indent), Fig. 9d.
In addition, Herzian cone cracks may develop, concentric with but outside the
plastic indent left by a hard sphere or cylindrical punch.
The median crack is a single, penny-shaped crack nucleated beneath the apex of
the plastic zone created by the indentor. The diameter of the median crack is com-
parable with the indent size, and a median crack is not visible in a polycrystalline,
opaque material. The driving force for nucleation of the median crack is the elastic
tensile stress developed normal to the indentation direction at the elastic-plastic
boundary when the external load is relaxed. Nucleation of a median crack depends
on the presence of a suitable flaw. Once nucleated, the median crack will propagate
spontaneously to a stable flaw size. The critical flaw size for growth is:
where Hv is the Vickers hardness, a, c, and R,, are the half-length of the diagonal
of the indent, the radius of the radial crack, and the radius of the plastic zone,
respectively, q5 the constraint factor, v and /-L are Poisson’s ratio and the coefficient
of friction between the indentor and the sample. Alternative’formulations exist!
The indent size, the surface crack morphology, and the extent of subsurface
damage are all influenced by the crystal structure, chemical composition, and micro-
structural morphology of the material.
that the influence of fracture toughness should exceed the effect of hardness in
improving erosion resistance.
Scratch tests have been used to evaluate the adhesion of thin films to a ceramic
substrate.
phases, and the presence of porosity, microcracking, and impurity phases [121,127].
The second aspect is the microstructural scale of the features, including their size,
size distribution, volume fraction, and distribution in space (microstructural homo-
geneity). The third factor is the orientation of microstructural features, both their
morphological anisotropy (alignment of platelets, whiskers, and fibers), and their
crystallographic anisotropy (preferred orientation of crystalline phases with respect
to the component geometry, commonly termed ‘texture’).
As in metals, the scale of the microstructure is related to the hardness (and,
through the hardness, to the plastic flow stress), with the Petch law applying well
into the submicron range of grain size. Increasing the hardness for a given composi-
tion is readily achieved by reducing grain size, typically by reducing powder particle
size for the raw material together with the sintering temperature used to densify
(sinter) the component. Unfortunately, processing to full density becomes more
difficult as the powder particle size decreases, while any residual porosity reduces
the hardness. Grain growth during sintering is limited by appropriate dopant
(trace) additions. Fine zirconia particles in alumina pin grain boundaries [ 1281,
while MgO additions inhibit grain boundary migration without affecting densifica-
tion [1291. Commercial aluminas with submicron grain sizes and near-theoretical
densities are available with a hardness of the order of 20 GPa (compare the tensile
modulus of 380 GPa).
Fracture toughness has been improved by providing microstructural toughening
mechanisms, discussed more fully below. Direct melt oxidation (DMO) leads to
alumina products which contain a network of some 15vol-% residual metal
[130,131]. The metal fibers intersect a propagating crack front and dissipate
energy by deforming plastically, reducing the stress intensity factor at the crack
tip. Platelets or whiskers in the microstructure, present as additions to the powder
mix or, preferably, grown in situ as a result of trace additions, also bridge the propa-
gating crack front, supporting some of the load and generating frictional forces as
they detach from the matrix. Tetragonal zirconia grains or particles transform to the
lower density monoclinic phase in the stress field at a crack tip, generating crack
closure forces which lead to ‘transformation toughening’.
The high temperature strength and creep properties of hard materials are sensitive
to the presence of glassy phases at grain boundaries [132], which promote viscous
sliding at the boundaries, lead to enhanced stress concentration at the boundary
junctions and nucleate cavities. The cavities grow under the hydrostatic tensile
component of applied load, coalescing to form ductile microcracks which initiate
the third stage of creep failure (see below), leading to stress rupture. In some
grades of silicon nitride a glassy oxynitride phase responsible for the excellent
sinterability of the material, can be crystallized after sintering, greatly improving
the creep resistance. Since creep failure is commonly initiated by cavitation, which
results in a volume increase, creep in hard materials depends on the sign of the
uniaxial load. Creep rates in compression are generally lower than those in tension
[ 1331. The viscous flow of the glassy phase in the grain boundary channels, a perco-
lation process, commonly determines the overall strain history.
The availability of nanopowders has led to the development of submicrometer,
fine-grained ceramics, which exhibit improved mechanical properties. Many
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.4 Microstructural Parameters and Mechanical Properties 93
glass-free oxide ceramics with a submicron grain size are superplastic under con-
ditions of moderate temperature (= 0.5 T,) and low strain rate to lop2s-’)
[134-1 361. The exact mechanism of superplastic deformation is controversial, but
it is clear that the grain boundaries play an important role, and that diffusion is a
dominant contributor to plastic deformation. Tensile strains to failure of 100-
1000% are routinely observed [ 134-1 391, and should enable complex shapes to be
formed from dense ceramic preforms. The strain rates required for ceramic super-
plasticity are too slow for most forming operations.
Through trial and error, it has been found that grain sizes of less than 1 pm are
required for superplasticity, and the finer the grain size, the larger the superplastic
strain to failure of the ceramic [140,141]. A second advantage offered by a fine
grain size is an increase in the strain rates observed during superplastic defor-
mation. The effect has been rationalized using the following phenomenological
relation:
as mobile slip dislocations are exhausted and boundary shear leads to grain
interlocking at the triple junctions. This primary, transient creep terminates at the
minimum creep rate, which may then persist for most of the creep life, correspond-
ing to the second, steady state, creep regime.
When an extended period of steady-state creep is present, this rate is denoted i.
and can often be described by the relation:
i = A 6 exp(-QIRT), (33)
where A and n are material parameters, and Q is an ‘activation energy’ for creep
[152]. The constant creep rate in the steady-state regime is usually associated with
strong work-hardening, arising from the blocking of secondary slip dislocations
by dislocation barriers on the primary slip system. The stress exponents close to 5
are due to dislocation climb and fully ductile behavior, and a stress exponent
close to 3 is due to dislocation climb [151, 1531.
The mechanisms responsible for fracture in structural ceramics at elevated tem-
peratures have been reviewed [ 1541. Sensitivity to flaws or microstructural inhomo-
geneities which nucleate microcracks are among the failure mechanisms. The flaws
which control failure under creep conditions are different from those responsible for
fast fracture at room temperature. A common feature is the development of cracks
through gradual damage accumulation, depend on the microstructure. The role
of cracks in the deformation and fracture behavior of polycrystalline structural
ceramics have been reviewed [1551.
FRONTAL-WAKE BRIDGED-INTERFACE
.--.
Figure 10. Schematic presentation of toughening mechanisms. Frontal-wake mechanisms: (a) dis-
location glide, (b) microcracking, (c) phase transformation, (d) ductile second phase. Bridging
mechanisms: (e) grains, (f) fibers, (g) whiskers, (h) ductile second phase. Reproduced with permission
of . . .. . . . . . . (9).
result of the debris and damage generated during cyclic loading, leading to unstable
crack propagation.
3.6 Conclusions
The design of components for applications requiring high stiffness, hardness and
wear resistance under extreme conditions of load and temperature leaves little
choice for materials selection outside the realm of ‘hard materials’. The spectrum
of materials options available to the design engineer is by no means as broad and
varied as could be wished, but a combination of advances in materials and process
development, together with improved methods of inspection and testing, has led to a
significant reduction in the uncertainties and risks associated with the estimation of
mechanical failure probabilities and component for brittle materials in service.
Coating technology remains the least understood area in this domain, and the
improvements in sophistication of production technology (the development of
diamond coating, for example) have not necessarily been accompanied by improved
understanding of the factors that determine the performance of hard coating in
service. The optimization of coating thickness, the development of graded
(‘functionally graded’) coatings, the effect of substrate surface preparation prior
to coating, and the compatibility of the elastic and thermal properties of substrate
and coating, are all complex topics. They need to be addressed.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Ref&wms 99
References
I . M. F. Ashby and D. R. H. Jones, Engineering Materials I and 2, Int. Series on Materials
Science and Technology, Vol. 39, Pergamon Press, Oxford, 1988.
2. M. F. Ashby, Materids Selection in Mechanical Engineering, Pergamon Press, Oxford, 1992.
3. I. J. McColm, Ceramic Hardness, Plenum Press, New York, 1990.
4. F. S. Petit, in Coating f o r High Temperature Applicutions, E. Lang (Ed.), Applied Science
Publications, Essex, 1983, pp. 1-32.
5. C. Puret and P. Pichoir, in Coating,/or Hight Tempertrture Applications, E. Lang (Ed.), Applied
Science Publications, Essex, 1983, pp. 33-78.
6. D. Satas, Coating Technology Handbook, Dekker, 1991.
7. Y. Kimura, K. Okada. and Y. Enomoto, Wear, 1989, 133, 147-161.
8. 0. 0. Ajayi and K. C. Ludema, Wear, 1988, 124, 237-257.
9. B. R. Lawn, Fracture of’ Brittle Solids, 2nd edn., Cambridge Solid State Science Series,
Cambridge University Press, 1993.
10. D. Broek, Elementnry Engineering Fracture Mechanics, 4th edn., Kluwer Academic Press,
Dordrecht, 1986.
11. M. F. Kanninen and C. H. Popelar, Advanced Fructure Mechanics, Oxford Engineering
Science Series 15, Oxford University Press, Oxford, 1985.
12. A. G. Evans, J . Am. Ceram. Soc. 1990, 73, 187-206.
13. J. Gallagher, Damage Tolerance Design Handbook, Revised edn., Metals and Ceramics Infor-
mation Center, Wright-Patterson Air Force Base, Columbus, OH, 1983.
14. T. Nishioka et al., Sumitomo Elec. Tech Rev. 1990. 29, 262-269.
15. R. C. Laible, in Ballistic Materials and Penetrution Mechanics, R. C. Laible (Ed.), Elsevier
Scientific Publishers, Amsterdam, 1980, pp. 133-143.
16. D. G. Brandon, Ceramic Armor, in Concise Encj~clopcJdiaof Advanced Ceramic Materials, R. J.
Brook (Ed.), Pergamon Press, Oxford, 1991, pp. 22-24.
17. D. W. Lazenby and P. Phillips, Cutting,for Construction, Architectural Press, London, 1978.
18. H. J. Swinheart (Ed.), Cutting Tools Materials Selections, Am. Soc. of Tools and Manufactury
Eng., Dearborn, Michigan, 1986.
19. T. Tatsuno and M. Miyamoto, Kobelco Technol. Rev. 1994, 117, 11-15.
20. G. Lutz, Int. Ceram. Rev. 1993, 42. 367-371.
21. R. T. Coelho, S. Yamada, D. K . Aspinwall, and M. L. H. Wise, Int. J . Machine Tools Manuf:
1995,35, 761-774.
22. K. Przyklenk, Indust. Diamond Rev. 1993, 53. 192-195.
23. T. Ikeda, Y. Kawate, and Y. Hiral, Kobelco Technol. Rev. 1989, 6, 1 4 .
24. J. Bruhn, S. Schicker, D. E. Garcia, R. Jenssen. F. Wagner, and N. Claussen, Ke.v Eng. Muter.
1997, 121, 73-80.
25. W. D. Kingery, H. K. Bowen, and D. R. Uhlmann, Introduction to Ceramics, 2nd edn., A
Wiley-Interscience Publication, John Wiley & Sons, Inc., New York, 1976.
26. S. Chattergee and S. Chandrashekhar, Key Engin. Muter. 1996, 114, 67-104.
27. Z. Hashin and S. Shtrikman. J. Mech. Phys. Solids, 1963, 11, 163.
28. M. Paley and J. Aboudi, Comp. Sci. Technol. 1991, 41,411429.
29. T Christman, A Needelman, and S. Suresh, Actu Metall. 1989, 37, 3029-3050.
30. G. Bao, J. W. Hutchinson, and R. M. McMeeking, Acta Metall. 1991, 39, 1871-1882.
31. A. Maewal, Comput.Meth. Appl. Mech. Eng. 1986, 57, 191-205.
32. C. T. Sun and W. C. Liao, J . Cotnp. Mater. 1990, 24, 977-993.
33. G . D. Quinn and R. Morrell, J . A m . Ceram. Soc. 1991, 74,2037-2066.
34. W. Weibull, Sived. Ins. Eng. Res. Proc., 151, 1-45,
35. E. J. Gumbel, Statistics of’E.utremes, Columbia University Press, New York, 1958, p. 136.
36. L. G. Johnson, Theory and Technique of’ Variation Research, Elsevier, Amsterdam, 1964, p. 137.
37. R. W. Davidge, Mechanical Behavior of Ceramics, Cambridge University Press, Cambridge,
1979.
38. D. Lewis, A m . Ceram. Soc. Bull. 1982, 61, 1208-1214.
39. M. J. Hanney and R. Morrell, Proc. Br. Ceram. Soc. 1982, 13, 277-290.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
100 3 Mechanical Properties and their Relation to Microstructure
40. D. Lewis 111 and S. M . Oyler, J. Am. Ceram. SOC.1976, 59, 289-307.
41. J. Seidel, N. Claussen, and J. Rodel, J. Europ. Ceram. SOC. 1995, 15, 395404.
42. T. Lube, M. Manner, Key Eng. Muter. 1997,132,488-491.
43. A. A. Griffith, Phil. Trans. Roy. SOC.Lond. 1920, A221, 163.
44. A. A. Griffith, Proc. First Znt. Cong. Appl. Mech., C. B. Biezeno, J. M. Burgers, and J. Waltman
Jr. (Eds), 1924, Delft, p. 35.
45. C. E. Inglis, Proc. Roy. SOC.Lond. 1913, A127,290.
46. G. R. Irwin, in Handbuch der Physik, Vol. 6, Springer-Verlag, Berlin, 1958, p. 551.
47. E. Orowan, Weld. Res. Supp. 1955, 34, 157s.
48. M. L. Williams, J. Appl. Mech. 1958, 24, 109.
49. T. T. Shih and C. Opoku, Eng. Frac. Mech. 1979, 12,479498.
50. D. P. Rooke and D. J. Cartwright, Compendium of Stress Intensity Factors, HMSO, London,
1976.
51. H. Tada, P. Paris, and G. Irwin, The Stress Analysis of Cracks Handbook, Del Research
Corporation, Hellertown, PA, USA, 1987.
52. Y. Murakami, Stress Intensity Factors Handbook, Vol. 1 and 2, Pergamon Press, Oxford, 1987.
53. D. P. Rooke, F. I. Baratta, and D. J. Cartwright, Eng. Fract. Mech. 1981, 14, 397426.
54. G. C. Sih, Handbook of Stress Intensity Factors, Lehigh University, 1973.
55. J. W. Hutchinson and Z . Suo, in Advances in Applied Mechanics, J. W. Hutchinson and T. Y.
Wu (Eds), Academic Press, New York, 1991.
56. L. B. Freund, Dynamic Fracture Mechanics, Cambridge University Press, Cambridge, 1990.
57. E. C. Combe and D. C. Smith, Br. Dent. J. 1964, 117, 237.
58. A. H. Cottrell, Dislocations and Plastic Flow in Crystals, Oxford University Press, Oxford, 1965.
59. D. Tabor, The Hardness of Metals, Clarendon Press, Oxford, 1951.
60. R. Hill, The Mathematical Theory of Plasticity, Clarendon Press, Oxford, 1950.
61. S. R. Witek, G. A. Miller, and M. P. Harmer, J. Am. Ceram. SOC.1989, 72, 469473.
62. L. A. Davis, c.-P. Chou, L. E. Tanner, and R. Ray, Scripta Metall. 1976, 10, 937.
63. H. Li and R. C. Bradt, J. Non-Cryst. Sol. 1992, 146, 197.
64. P. M. Sargent and T. F. Page, Proc. Br. Ceram. SOC.1987,26,209.
65. P. F. Becher, J. Am. Ceram. SOC.1974, 57, 107.
66. J. B. Quinn and G. D. Quinn, J. Muter. Sci. 1997, 32, 4331.
67. R. W. Rice, C. C. Wu, and F. Borchelt, J. Am. Ceram. SOC.1994.77. 2539-2553.
68. K. Zeng, E. Soderlung, A. E. Giannakopoulus, and D. J. Rowcliffe,’Acta Muter. 1996, 44,
1127-1141.
69. J. Lankford and D. L. Davidson, 3rd International Conference on Mechanical Behavior of
Materials, Pergamon, Oxford, 1980.
70. J. Valli, J. Vac. Sci. Technol. A , 1986, 4, 3007-3014.
71. S. Herada, N. Noda, T. Endo, and M. Nagano, J . SOC.Muter. Sci. Jpn. 1988, 37, 904-907.
72. Y. Le Petitcorps, M. Leheye, R. Pailer, and R. Naslain, Cornp. Sci. Technol. 1988, 32, 31-55.
73. B. Thomas, D. West, W. E. West, and K. Wyckoff, Powder Metall. Znt. 1988, 20, 39-44.
74. R. W. Davidge, Muter. Sci. Technol. 1986, 2, 902-909.
75. American Society for Testing and Materials, Standard Test for Plane Strain Fracture Tough-
ness of Metallic Materials, ASTM E-399, Philadelphia, Pennsylvania, 1987.
76. I. Hadley and M. G. Dawes, Fatigue Fract. Eng. Muter. Struct. 1996, 19, 963-973.
77. H. Awaji, T. Yamada, and H. Okuda, J. Jpn. Ceram. SOC.,Znt. Edn. 1991, 4, 403408.
78. G. D. Quinn, J. J. Kubler, and R. J. Gettings, VAMASTechnical Report No. 17, NIST,
Gaithersburg, Maryland, 1994.
79. T. Nishida, Y. Hanaki, and G. Pezzotti, J. Am. Ceram. SOC.1994,77, 606608.
80. D. Munz, in Fracture Toughness of Ceramics, Vol.6, R. C. Bradt, A. G. Evans, D. P. H. Has-
selman, and F. F. Lang (Eds), Plenum Press, New York, 1983.
81. B. R. Lawn and E. R. Fuller, J. Muter. Sci. 1975, 10,2016-2024.
82. R. Warren and B. Johnson, Powder Metall. 1984, 27, 25-29.
83. T. Sadahiro and S. Takatsu, in Modern Developments in Powder Metallurgy, Vol. 14, H. H.
Hauser, H. W. Antes, and G. D. Smith (Eds), Metal Powder Industries Federation, Princeton,
NJ, 1981, pp. 61-72.
84. T. Nose and T. Fujii, J. Am. Ceram. SOC.1988,71, 328-333.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 101
126. G. Petzow, R. Telle, and R. Danzer, Muter. Churact. 1991, 26, 289-302.
127. A. Zimmermann, M. Hoffman, B. D. Flinn, R. K. Bordia, T. J. Chuang, E. D. Fuller, and
J. Rodel, J . Am. Ceram. SOC.1998,81, 2449-2457.
128. L. A. Xue, K. Meyer, and I. Wei Chen, J. Am. Ceram. SOC.1992,75,822-829.
129. Yet Ming Chiang and W. D. Kingery, J. Am. Ceram. SOC.1989,72,271-277; 1990,73, 1153-
1158.
130. V. Jayaram, R. Manna, M. G. Kshetrapal, J. Sarker, and S. K. Biswas, J . Am. Ceram. SOC.
1996,79,710-772.
131. B. S. S. Daniel and V. S. R. Murthy, Muter. Des. 1995, 16, 155-161.
132. R. W. Evans, D. P. Thompson, and B. Wilshire, Br. Ceram. Trans. 1993,92, 28-30.
133. W. E. Luecke, S. M. Wiederhorn, B. J. Hockey, R. F. Krause Jr., and G. G. Long, J. Am.
Ceram. SOC.1995,78,2085-2096.
134. Y. Maehara and T. G. Langdon, J . Muter. Sci. 1990, 25, 2275-2286.
135. I.-W. Chen and L. A. Xue, J . Am. Cerum. SOC.1990,73,2585-2609.
136. T. G. Nieh, J. Wadsworth, and F. Wakai, Int. Muter. Rev. 1991,36, 146-161.
137. F. Wakai, S. Sakaguchi, and Y . Matsuno, Adv. Cerum. Muter. 1986, 1, 259-263.
138. T. G. Nieh, C. M. McNally, and J. Wadsworth, J. Muter. 1989, 41, 31-35.
139. K. Kajihara, Y. Yoshizawa, and T. Sakuma, Scripta Metall. Muter. 1993, 28, 559.
140. T. G. Nieh and J. Wadsworth, J. Muter. Res. 1990, 5, 2613-2615.
141. F. Wakai and T. Nagano, J. Muter. Sci. 1991, 26, 241-2447.
142. J. Karch, R. Birringer, and H. Gleiter, Nature, 1987, 330, 556558.
143. Z. Cui and H. Hahn, NanoStructured Muter. 1991, 1, 419-425.
144. A. H. Chokshi, A. Rosen, J. Karch, and H. Gleiter, Scripa Metall. 1989, 23, 1679-1684.
145. A. M. El-Sherik, U. Erb, G. Palumbo, and K. T. Aust, Scripta Metall. 1992, 27, 1185-1188.
146. H. J. Hofler and R. S. Averback, Scriptu Metall. 1990, 24, 2401-2406.
147. R. 0. Scattergood and C. C. Koch, Scripta Metall. Muter. 1992, 27, 1195-1200.
148. C. S. Pande, R. A. Masumura, and R. W. Armstrong, NanoStructuredMater. 1993,2,323-331.
149. V. G. Gryaznov, M. Yu. Gutkin, A. E. Romanov, and L. I. Trusov, J . Muter. Sci. 1993, 28,
43594365.
150. W. R. Cannon and T. G. Langdon, J . Muter. Sci. 1988,23, 1-20.
151. E. Yasuda, K. S. Ramesh, S . Kimura, Report of the Research Laboratory of Engineering
Materials, Tokyo Institute of Technology, 1985, 10, 115-134.
152. A. Hynes and R. Doremus, Crit. Rev. Solid State Muter. Sci. 1996,21, 129-187.
153. A. H. Chokshi and T. G. Langdon, Muter. Sci. Technol. 1991,7, 577-584.
154. D. S . Wilkinson, Fruct. Mech. Ceram. 1992, 10, 349-366.
155. D. P. H. Hasselman, K. V. Donaldsod, and A. Venkateswaran, Fract. Mech. Ceram. 1992,10,
493-507.
156. G. Grathwohl, Muterialwiss. Werkstofftech. 1988, 19, 113-124.
157. E. D. Whitney, Powder Metall. Znt. 1983, 15, 201-205.
158. A. G. Evans, Znt. J . Fract. 1980, 16,485498.
159. D. B. Marshal, M. C. Shaw, R. H. Dauskard, R. 0 Ritchie, M. J. Readey, and A. H. Heuer,
J . Am. Ceram. SOC.1990,73,2659-2666.
160. R. H. Dauskardt, B. J. Dalgleish, D. Yao, R. 0. Ritchie, andP. F. Becher, J. Muter. Sci. 1993,
28, 3258-3266.
161. C. J. Gilbert, J. J. Cao, W. J. MoberlyChan, L. C. DeJonghe, and R. 0. Ritchie, Actu Muter.
1996,44, 3199-3214.
162. C. J. Gilbert, R. N. Petrany, R. 0. Richie, R. H. Dauskardt, and R. W. Steinbrech, J . Muter.
Sci. 1995, 30, 643.
163. C. J. Gilbert and R. 0. Ritchie, Eng. Fract. Mech. 1998, 60, 303-313.
164. Jian-Ku-Shang and R. 0. Ritchie, Metall. Trans. A , 1989, 20, 897-908.
165. M. L. Wilkins, C. F. Cline, and C. A. Hondol, Tech. Rept. UCRL-50694, Lawrence Radiation
Laboratory, 1969.
166. M. L. Wilkins, in Ballistic Materials and Penetration Mechanics, R. C . Laible (Ed.), Elsevier,
Amsterdam, 1980, pp. 225-252.
167. H. Kolsky, in Ballistic Materials and Penetration Mechanics, R. C. Laible (Ed.), Elsevier,
Amsterdam, 1980, pp. 185-252.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 103
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
4.1 Introduction
Hardness of a material is a measure of its ability to resist deformation [l, 21. The
upper limit of the hardness is determined by the rigidity of the crystal structure
upon isostatic pressure. This ‘theoretical’ hardness is proportional to the reciprocal
value of the bulk modulus, B. From the Hooke’s law B = da/dE (CTis the applied
stress and E the resulting elastic strain) and from the relationship between the
stress (a force) and the first derivative of the binding energy, E, with bond distance,
(2)
a, at the equilibrium position ao, CT = (dEb/da)o, one obtains [3]
B= 0 .
Thus, the theoretical hardness increases with the increasing curvature of the inter-
atomic potential curve at the equilibrium bond distance, ao, that is, with increasing
bond energy and decreasing bond length. A high bond energy means a high electron
density between the atoms as found for a nonpolar, covalent bond between atoms of
small radii of the first period. In addition, a high coordination number is required in
order to maximize the value of B. For these reasons carbon in its metastable, four-
fold coordinated sp3-hybridization, diamond polymorph, is the hardest material
known, followed by cubic boron nitride, cBN [2].
Based on such general considerations C. M. Sung predicted in 1984 that C3N4
should have a hardness comparable to diamond [2]. Later on, Cohen [4] and
coworkers [5-71 derived Equation (2), which reproduced very well the bulk
moduli of many materials and confirmed the prediction of C. M. Sung regarding
the theoretical hardness of C3N4,
Here, ( N c ) is the average cpordination number and X is the polarity of the bond, B is
in GPa and a. is given in Angstroms [6]. For nonpolar, covalent bonds in diamond
X = 0, whereas in other compounds, such as cBN, Si3N4and C3N4 X > 0 which
decreases the value of the elastic modulus. The expected high theoretical hardness
of C3N4 is based on the small bond distance and relatively small polarity A.
However, like diamond and cBN, C3N4is a thermodynamically unstable com-
pound whose preparation requires a high degree of dissociation of nitrogen in
order to achieve the desired stoichiometry, a medium-energy ion bombardment to
promote the carbon into the metastable fourfold coordination and a relatively
high temperature of about >800°C in order to evaporate the paracyanogen,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.1 Introduction 105
(CN),, which is always formed during the deposition of the stoichiometric C3N4(see
[S] for further details). However, the Vickers hardness reported for stoichiometric
amorphous C3N4 films reached only about 30 GPa, far below that of diamond
(70-100 GPa [2,5,6]). Using similar synthetic principles, other authors succeeded
in the deposition of crystalline, stoichiometric C3N4 films [9-131 but they could
not measure the hardness because of an insufficient density of the polycrystalline
films [lo]. A large number of papers report on the ‘synthesis of carbon nitride’
but those films are highly substoichiometric with a large nitrogen deficiency even
below that of paracyanogen and have a lower hardness. Nevertheless, CN, may
find some ‘niche’ applications such as protective, low friction coatings on magnetic
storage discs [14-16].
In contrast to the C3N4 and substoichiometric CN, thin films of turbostratic
graphite with a relative low concentration of nitrogen, CN0,25p0,35 and three
dimensional cross-linking deposited at a relatively high temperature of 400°C
have a hardness of 40-50GPa [17-191. Also the hardness of diamond crystals can
be increased by nitrogen impurities 120,211, thus reaching the upper end of the
hardness scale in Fig. 1. These are two examples of the well known rule that the
microstructure determines the practically achievable strength (and hardness) of
engineering materials, which is orders of magnitude smaller than the theoretical
strength. The latter is calculated for the slip of crystal planes as c,(theor)
G/2n, where cT,(theor) is the theoretical critical fracture stress and G the shear
modulus [3,22]. The reason is that the deformation and fracture of materials
occurs due to the multiplication and movement of dislocations in crystalline
materials and growth of microcracks in glasses and ceramics, all of which require
a much smaller stress than the theoretical value a,(theor) mentioned above [3,22-
241. Therefore, the practical strength and hardness of materials is determined by the
4000
0
Figure 1. Vickers hardness of selected materials in comparison with that of the heterostructures and
nanocrystalline composites. Notice that the hardness of diamond depends on the quality and purity
of the crystal (see text).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
106 4 Nanostructured Superhard Materials
9000 I I I . I r n I . I I
8000 -
0
-
7000 - -
N
-
E 6000- 0 -
0 O . 0
E -
3 5000-
25
(I) 4000- 0 -
(I)
3000-
0
-
E -
nJ 2000-
0
r
1000 - -
0 I ' I ' I ' I ' I .
0 20 40 60 80 100
Figure 2. (a) An example of the indentation curve measured on a =5pm thick layer of
ncTiN/aSi3N4.L,,, and h,,, are the maximum applied load and indentation depth, hplastic is
the depth of the remanent, plastic deformation. The arrows -+ and t denote the loading
and unloading part. (b) Measured values of the hardness against the maximum applied load. (c)
Indentation depth at a constant applied load of 50mN against time which shows the absence of
creep.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.1 Introduction 107
c-
'ma, , I , I , I
I . I . I .
0.26
H, = 7529 [N/mm2]
0.24
- El(1-v2) = 520.28 [GPa]
.....................
sl
sf$0.22- -
-0 - -
C
0
.g 0.20
- -
c
E
9,
TI
E 0.18
- -
c
0.16z0
0
time [sec]
(C)
The ‘universal hardness’, H,, is obtained from the same formula if h,,, is inserted
instead of hplastic. The universal hardness includes both the elastic and plastic
deformation. The linear part of the unloading curve corresponds to the elastic
recovery when the diamond pyramid is in a constant area contact with the material.
Therefore it represents Hooke’s law and allows one to calculate the corresponding
‘elastic modulus’ E/(1 - v 2 )which is a complicated function of the bulk, shear, and
tensile moduli (v is the Poisson ratio). The details of the apparatus, the measuring
procedure and possible errors are given in the relevant papers to which we refer here
[25-281. If done correctly, the plastic hardness measured by the indentation agrees
within about 10-1 5% reasonably well with that from the classical Vickers method at
least in the range H 5 1500 kgmmp2 [25].
The so called ‘nanoindentation’, which is frequently used to measure < 1 pm thin
films, is subject to a number of possible errors [27,28]. For example, when
measuring soft materials, such as aluminum or pure iron with a small load and
indentation depth, the dislocations are pinned in the surface contaminant layer
(oxides, carbides) and, consequently, unrealistically high values of hardness are
found. The plastic deformation may also need a certain time to reach equilibrium
under the given load because of a finite velocity of dislocation movement. This
can be seen as creep (increase of the indentation depth) when a constant load is
applied for 10-30 s.
Therefore, in order to obtain reliable values of the hardness, one should always
use at least >4pm thick films, make sure that the maximum indentation depth,
h,,,, does not exceed 10% of the film thickness, check for possible dynamic
response of the material by sustaining the indentor at the maximum load L,,,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Concept for the Design of Superhard Materials 109
for at least 2&30 s and measure the hardness (always at a different position on the
layer) as a function of the maximum applied load. Figure 2b shows the typical
dependence of the measured plastic hardness on L,,, and the typical scattering
of individual measurements. In the case of superhard materials, one typically
finds a smaller hardness and large scattering of the individual measurements
when using small loads of SlOmN because the indentation depth of 50.1 pm is
comparable with the surface roughness. Figure 2c shows the time dependence of
A,, under the constantly applied maximum load L,,,. One notices that there is
no creep seen, at least on the time scale available with the given instrument. We
emphasize these problems here to make sure that in the measurements to be
reported hereafter, all these precautions have been taken in order to make the
results as reliable as possible. However, one should keep in mind that the Vickers
technique, which was originally developed for the conventional alloys and
materials with hardness not exceeding 2000-2500 kg mm-*, is not suitable for
the measurements of materials whose hardness approaches that of diamond.
Therefore we consider values of hardness of 7000-10 000 kg mmp2 (and even
higher) frequently found in our measurements only as an indication that these
materials approach the hardness of diamond, but the Vickers hardness scale is
probably not linear in that range. As it will be shown later, there is no dislocation
activity in the superhard nanocomposites and, therefore, no ‘plastic deformation’
in the classical sense. Strictly speaking one should use the term ‘remaining’, ‘per-
manent’ or ‘pseudoplastic’ deformation because it is a brittle fracture as found, for
example when indenting into a glass. Because the current literature does not differ-
entiate between these terms we adopt the term ‘plastic deformation’ keeping, how-
ever, in mind that it is in fact a brittle fracture whose microphysics and
micromechanics are not yet understood.
Last but not least, one should not forget that for practical applications the
hardness at room and at a high temperature is only one of many properties which
determine the applicability of a material. Further mechanical properties, such as
fracture toughness under static and dynamic load, coefficient of friction, corrosion
resistance, and reactivity with the material to be machined at room as well as at
elevated temperature are decisive for most applications.
known from the metallurgy, which are briefly summarized in Equation (4) (see e.g.
~31):
Here, the first term describes the Peierls-Nabarro stress which determines the dislo-
cation movement (slip within the crystallographic plane) in a crystal, the second
term corresponds to the solution hardening (C is the concentration of the impuri-
ties), the third term is the Hall-Petch relationship [3,30,3I], which describes grain
boundary strengthening (d is the crystallite size, see below), the fourth term
describes the strain hardening due to dislocation pile up ( a is a material constant,
G the shear modulus, b the Burgers vector, and p the dislocation density). The
last term describes the overall dependence of the crack growth on the size of
microcracks (voids) of the size a. ( E is the Young modulus, ys the surface
energy). Any of these processes allows the engineer to increase the strength of a
material by a different amount. For example, the hardness of TIN films deposited
in glow discharge plasmas under energetic ion bombardment which causes strength-
ening due to the formation of dislocations or due to a high biaxial compressive stress
is typically higher than that of a single crystal.
as well as ceramics (last term in Equation (4), Griffith theory [3]). There are several
different mechanisms and theories which describe Hall-Petch strengthening
[3,24,30-331: Pile-up models, work hardening and grain-grain boundary composite
models which yield different relationships for a. and kgb [32]. The work hardening
1
models and the core-mantle composite model of Koks 34 yield also a more com-
I .
plicated dependence of ofr on d . In ceramics, the d-'I dependence can originate
from the fact that the size a. of the possible flaws such as voids and microcracks,
which are formed during the processing of the material, also decreases with decreas-
ing grain size. For these reasons, the Hall-Petch relationship (5) should be consid-
ered as a semiempirical formula which is valid down to a certain minimum grain
size, for example of d > 450b = l00nm for the pile-up models and
d > 60b x 15 nm for a dislocation source model [32].
There are many examples which show that the strength and hardness of an alloy
can be increased by a factor of 4-7 if the grain size is decreased into the nanometer
range. One such example is the rapidly solidified A19&r3CelCo, (x = 1-2%) alloy
containing as a main component a nanosized quasicrystalline icosahedral phase [35].
A similar strengthening is found in many metals such as TiAl[36], Cu [37], and other
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Concept ,for the Design of Superhard Materials 111
I’
y1600-
.
E
m
251400-
cn
u)
g1200-
E
m
r
.g 1000-
-cnmn
800;,
€ I
d’
r I I I
a-Si,N,
Ol -
Figure 4. Idealized structure of the nanocrystalline-amorphous composite.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Concept ,for the Design OjSuperhard Materials 113
form such a bi-phase system, both materials must be immiscible (i.e. display a
thermodynamically driven segregation during the deposition) and the cohesive
energy at the interface between both phases must be high. Last but not least, the
amorphous phase must possess a high structural flexibility in order to accommodate
the coherency strain without forming dangling bonds and other flaws. In order to
ensure that such a composite will show no grain boundary sliding even at a high
temperature, both materials should be refractory in order to ensure that the opera-
tional temperature of the tools will be less than half the melting or decomposition
temperature, Tm,decomp 2 2500 K.
It is important to note that, although the strengthening caused by the decreasing
dislocation activity in the range of crystallite size of d 5 lOOnm finally reaches a
saturation due to the absence of dislocations in ultrasmall nanocrystals of
d 5 10 nm [38,39], the strength and hardness of such material can still increase
upon a further decrease of the crystallite size when another mechanism of
strengthening becomes dominant. In the absence of dislocations and grain boundary
sliding, the nanocomposites show a brittle behavior which means that the fracture
strength (and hardness) are proportional to the elastic modulus of the material. This
behavior has been confirmed in our nanocrystalline superhard composites [63-651.
The fracture stress of such material is then determined by the critical stress for the
growth of microcracks (see the last term in Eq. (4)). The size of the 'microcrack', ao,
that is, of a possible flaw and voids, is in a well compacted material always smaller
than the size of the nanocrystals, uo 5 d . Moreover, even if the critical stress for the
crack growth
should be exceeded by the applied (e.g. uniaxial) stress, the crack could hardly grow
to a size larger than d because it would have to bend to pass around the nano-
crystals. If p is the angle between the plane of such a bent crack and the direction
of the tensile stress gappl, the component ( T = ~ oapplsin p < rapplis acting in the
direction normal to the crack plane. Only this component can cause the crack to
grow. For randomly oriented nanocrystals embedded in a 51 nm thin matrix, as
found in our nanocomposites with the smallest crystallite size and highest hardness,
the cracks have to bend so strongly that cp + 0 and ( T <<~ oappl. Furthermore, upon
the bending also crack branching ('meandering') occurs in a three-dimensional iso-
tropic nanocomposite which further decreases the ability of such nanocracks to pro-
pagate [3, 1701. These are the reasons why the hardness of our composites increases
when the crystallite size decreases below 5-7 nm [64,66,67].
The obvious first (but not the only) choice of such an ncja system which meets the
above discussed criteria for achieving the superhardness is straight forward: Many
transition metals M, such as Ti, Zr, Hf, V, Nb, Ta, Cr form stable (refractory) and
hard nitrides which crystallize upon deposition even at relatively low temperature of
>lOO"C [68-701. Silicon nitride, on the other hand, grows amorphous even at
1100°C and it possesses the desirable structural flexibility due to the fourfold
coordination of silicon combined with the threefold coordination of nitrogen.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1 14 4 Nanostructured Superhard Materials
4.2.2 Heterostructures
In a theoretical paper published in 1970 [75] Koehler suggested a concept for the
design of strong solids which are now called ‘heterostructures’ or ‘superlattices’.
Originally, he suggested depositing epitaxial multilayers of two different metals
M(l) and M(2) having as different elastic constants as possible, EM(2) > EM(,),
but a similar thermal expansion and strong bonds. The thickness of the layers
should be so small that no dislocation source could operate within the layers. The
dislocations then would form within a layer M(l) with the smaller elastic modulus.
If, under applied stress, the dislocation would move towards the M( 1)/M(2)
interface, the elastic strain induced in the second layer M(2) with the higher elastic
modulus would cause a repelling force which will hinder the dislocation movement
across the interface. Thus, the strength of such multilayers should be much larger
than that expected from the rule of mixtures.
His prediction was further developed and experimentally confirmed by Lehoczky
who deposited AI-Cu and Al-Ag heterostructures and measured their mechanical
properties [76,77]. According to the rule of mixture, the applied stress, ca,which
causes elastic strain, E , is distributed between the layers proportionally to their
volume fractions VMcX)and elastic moduli EM(.)
ga = (VM(l)Eh4(l) + VM(2)EM(2))&. (7)
Lehoczky has shown that the tensile stress-strain characteristics measured on
multilayers consisting of two different metals displayed a much higher Young
modulus and tensile strength, which both increased with decreasing thickness of
the double layer (‘layer period’). For layers thickness of 570 nm the yield stress of
Al/Cu multilayers was 4.2 times larger and the tensile fracture stress was 2.4 to 3.4
times larger than the values given by the rule of mixture.
This work was followed by a number of researchers who confirmed the experi-
mental results on various metallic multilayer systems (e.g. Cu/Ni [78-801 as well
as on hard epitaxial and polycrystalline superlattices of nitrides, such as TiNjVN
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Concept f o r the Design of Superhard Materials 1 I5
[81], TiN/NbN [82,83], TiNV,Nbl -xN [84-861, Ti/TiN and WjWN [87], oxides
[88,89] and superlattices consisting of transition metal nitride and CN, [90]. In all
these cases, the hardness of the heterostructures increases with decreasing lattice
period, reaches a maximum of about 4000-5000kgmm-2 at 5-7nm and, after-
wards, decreases again.
The theoretical work on the heterostructures was summarized in several excellent
review papers [88,89,91]. Therefore, only a brief summary will be given here. For
a large lattice period, where the dislocation multiplication source can still operate,
the increase of the hardness and tensile strength (most researchers measured the
hardness because it is simpler than the measurement of tensile strength as done
by Lehoczky [76,77]) with decreasing layer thickness is due to the increase of the
critical stress needed to multiplicate dislocations, a, = Gb/lpp (Ipp is the distance
between the dislocation pinning sites [3]). Usually one finds strengthening depen-
dence similar to the Hall-Petch relationship, Eq. (9,but with a somewhat different
dependence on the layer period X-" (instead of the crystallite size, d , in Eq. ( 5 ) ) with
n = for layers with different slip systems and n = 1 for layers with a similar slip
systems [91]. In thin layers where no dislocation source can operate and the
Koehler's and Lehoczky's model applies, the maximum value of the critical shear
stress a,(X) which can cause a movement of a dislocation from the weaker layer 1
into the stronger layer 2 is given by [75-771
Here G M ( I ) are the shear moduli of the two materials, XI is the thickness of the softer
layer, b is the Burgers vector of the dislocation, and cp is the angle between the dislo-
cation slip plane in the layer M(1) and the interface M(l)/M(2). This theory predicts
that the strength (and hardness) depends mainly on the relative difference between the
+
shear moduli ( G M ( 2 ) - G M ( ~ ) ) / ( G M ( ~ ) G M M ( I ) )and the angle cp. For small period but
still X1 >> 46 the enhancement reaches an asymptotic value of
GM(2) - GM(l) GM(l) sincp
a,(max) = (9)
GM(2) + GM(l) 8.ir
The decrease of the hardness experimentally observed at small values of X < 5-
7nm is due to two effects: For XI smaller than about 15-20 times the Burgers
vector, b, the repelling mirror forces induced by the strain in the layer of stronger
materials on both sides of the layer M(l) begin to cancel out (these forces have
opposite sign with respect to each other) and, therefore, the enhancement of the
strength decreases. The other effect is a roughness of the interface due to the
interdiffusion during the preparation. These effects were discussed in a number of
papers ([82,92-951). In particular, Chu and Barnett [93] have shown that an
interface roughness of about l n m will cause a strong decrease of the hardness
and probably dominates the experimentally observed softening for X < 5-7 nm.
Recent experimental data on superlattices with a high immiscibility which form
sharper interfaces support this conclusion [94].
In a remark added in proofs, Koehler mentioned that the ideas described in his
paper should also be valid if one of the layers is amorphous. In spite of that note,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
116 4 Nanostructured Superhard Materials
the researchers studied only the epitaxial and, later on, polycrystalline heterostruc-
tures. Only recently, several papers appeared in which one of the layers consist of
amorphous CN,, the other of a transition metal nitride such as TIN [88,96,97] or
ZrN 1901. However, with decreasing layer thickness the layered structure vanished
and a nanocrystalline composite structure appeared [90]. Such films also show a
high hardness exceeding 4000 kg mm-2.
In summary, significant progress has been achieved in the preparation and under-
standing of the hardness and strength enhancement in heterostructures and it is best
described in the recent theoretical papers by Chu and Barnett [93] and Anderson and
Li [94]. The application of heterostructures for cutting tools has been well developed
[98] and cutting tools with superior properties as compared to single layer coatings
are available. However, the relatively complicated preparation technique, which
requires a periodic change of the source for the subsequent deposition of the multi-
layers, makes the investigation into the novel superhard nanocrystalline composites
very attractive because they can be deposited in a simple plasma chemical vapor
deposition (CVD), plasma vapor deposition (PVD) or a combined PVD and
CVD process. Therefore, the remaining part of this chapter will concentrate on
these nanocomposites.
phases to occur at such a low temperature which is compatible with these materials.
The problem is mainly the steel, because the maximum tolerable temperature is
about 550°C or even lower for high speed steel. Therefore, an activated chemical
or physical vapor deposition technique which allows one to coat simultaneously a
large number of nonplanar substrates (i.e. a large total area) at a sufficiently high
deposition rate of 2 3 pmh-' (about 1 nms-') such as plasma CVD and PVD
appear to be most appropriate. Because plasma CVD usually operates at a some-
what higher pressure of about 1-5 mbar whereas a lower pressure of <lo-* mbar
is used in plasma PVD (reactive sputtering or activated reactive evaporation), the
plasma CVD was used in our initial work. It provides also a somewhat better
flexibility regarding the choice of the ncM,N/aA,Nb system. However, plasma
CVD still suffers problems with scaling from a small experimental reactor to a
large scale production units. Also the use of volatile transition metal halides as
sources of those metals makes it less attractive for industrial applications because
of the problems associated with the corrosive nature of the chlorine to the
vacuum pumps and the hydrophilic nature of the deposits at the reactor walls
and in the vacuum tubings. For these reasons, a recently developed combined
plasma PVD and CVD technique [99] is more appropriate for future industrial
applications whereas plasma CVD will remain the preparative tool for basic studies
on a variety of nc/a and nc/nc systems.
Both, high frequency (HF) [63,64,66,67,100] and direct current (d.c.)
[65,1101,102] glow discharge were used for the deposition. The apparatus used
for the deposition in an H F discharge has been described [63]. Here we shall briefly
describe the plasma CVD apparatus which operates under conditions of an
abnormal d.c. glow discharge.
A schematic of this apparatus is shown in Fig. 5 [65,102]. It consists of a reactor
made of Inconel 600 which is heated by means of an oven whose temperature is
controlled via a thermocouple (T.C.) inserted into the substrate holder and a
thermoregulator (not shown). Because the thermocouple is at the potential of the
cathode, 350-700 V, the whole thermoregulation system is electrically insulated
from the ground. The details of the gas manifold and of the pumping system are
obvious from the figure. The delivery of the metal halides (in Fig. 5 TiC14 for the
deposition of ncTiN/aSi3N4 is shown) occurs via a thermostatized evaporator or
bubbler. Typically, the total pressure is maintained at 1-5 mbar and the flow rates
of HZ,N2 and TiC14 are in the ratio of about 20 : (1-5) :0.1 sccm. The SiH4 flow is
adjusted between 0 and about 51 sccm depending on the required Si3N4 content
in the films. The substrates are mounted at the cathode of the d.c. discharge
which is operated in the abnormal glow regime [103-1051. In this regime, the
whole surface of the cathode is covered by a uniform glow because of the high resis-
tivity of the discharge region between the cathode and the negative glow. This glow
also follows fairly uniformly the surface of nonplanar substrates. For industrial
r
ap lications, a pulsed d.c. discharge with a repetition frequency of the order of
10 Hz and variable duty time is more appropriate in order to obtain a better control
of the deposition conditions, the substrate heating, and to avoid arcing. Of course, a
uniform discharge over nonplanar substrates can be maintained also in the H F
regime of 10-80 MHz if a sufficiently high power density is applied [106,107].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
118 4 Nanostructured Superhard Materials
Pirani
%-
\themostate
' XCI*
Figure 5. Schematics of the apparatus for the deposition of ncTiN/aSi3N4 coatings in an abnormal
direct current glow discharge (see text).
TiN prepared from TiC14 usually suffers from a relatively large chlorine content
which, for thermal CVD films deposited at 1000°C, can approach almost 1 at.%,
and may be much higher in films prepared by means of plasma CVD at lower tem-
peratures. Following our earlier studies about the shifts of the chemical equilibrium
in intense glow discharge plasmas [log, 1091we have shown that the chlorine content
of TIN deposited at 500°C can be significantly reduced by increasing the discharge
current density at the cathode to 2 1 mA cm-2 [ 1101. Using this principle and a large
excess of nitrogen and hydrogen, ncTiN/aSi3N4 films deposited at 500-550°C
contain 50.5 at.% of chlorine, which is fairly tolerable.
Another question is how to achieve the desirable stoichiometry, that is, a
sufficiently high activity of nitrogen, a(N2),in order to achieve the segregation in
reaction (10). The phase diagram of the ternary Ti-Si-N system at 1273K shows
that this is achieved at loga(N2) 2 -6 which corresponds to a nitrogen pressure
+
of lop6atm [71,73]. At a lower value of a(N2) the TiN, Si phases coexist and
+
for loga(Nz) 5 -7.5TiNx TiSi2 appear. The phase diagram becomes more
complicated at an even lower nitrogen activity, but this is not of any interest here.
The important conclusions are: First, for a nitrogen pressure of 2 1 mbar as used in
our experiments, fully stoichiometric TIN + Si3N4 phases should be formed at
550°C if the activation energy for N2 reaction with Ti and Si (i.e. a sufficient
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.3 Preparation and Properties of Superhard Nunocrystalline Composites 119
1 I
very good agreement. The somewhat larger value of the crystallite size determined
from the HR-TEM is probably due to the fact that the experimentalist easily over-
looks small crystallites in the HR-TEM micrographs, in particular if these are in the
shadow of the larger ones.
XPS studies have shown that the silicon is bonded as Si3N4[63,101] and is always
amorphous [61,62]. The available EDX does not allow us to measure the nitrogen
content. Thus, the content of silicon nitride [Si3N4]corresponds to (7/3)-times the
measured content of silicon [Si]. This scale is used in Fig. 7 which shows a similar
dependence of the average TiN crystallite size on the Si3N4 content for films
deposited in an abnormal d.c. glow discharge (see apparatus in Fig. 5) where the
surface of the growing film is exposed to energetic ion bombardment. The measured
hardness of the films is shown as well. One notices that the TIN crystallite size
initially decreases with increasing content of Si3N4,passes a minimum, and increases
afterwards again. The hardness shows exactly the opposite behavior reaching a
maximum at the Si3N4content of about 20mol.% when the TiN crystallite size
approached its minimum. Such behavior has also been found for other systems
studied so far, ncW2N/aSi3N4[64] and ncVN/aSi3N4 [66,67] and it appears to be
a universal phenomenon in these systems.
For the understanding of the development and stability of the nanostructure,
which will be discussed in the following section, it is important to realize that all
other deposition parameters, such as temperature, discharge current density, partial
pressure of all gases including TiC14 but the SiH4 (which was the only variable
parameter) are kept constant in this series of experiments. Thus, all these ncM,N/
aSi3N4 systems choose the minimum crystallite size at the Si3N4content of about
20 mol.% because this minimizes their total Gibbs free enthalpy.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.3 Preparation and Properties of Superhard Nanocrystalline Composites 121
[ 2000
01 I
0 10 20 30
a - Si,N, content [mol %]
Figure 7. Dependence of the average TIN crystallite size and of the hardness on the Si3N4content in
composites deposited in a d.c. discharge.
One might speculate whether the initial decrease of the crystallite size is a result of
a decrease of the TiN (or more generally M,N) growth rate because of a decrease of
the supply of the Ti gaseous source when the SiH4 flow is increased. However, this
can be almost certainly ruled out because the nucleation rate remains high and
almost constant since the large supersaturation changes only by less than I % in
these experiments when the small flow rate of SiH4 is varied between 0 and
5 1 sccm as compared to 20-100 sccm. of the total gas flow. This is supported by
the fact that the deposition rate remains fairly constant even if the Si3N4 content
is varied from 0 to almost 40mol.% (see Fig. 3a in [63]). Moreover, even if this
speculative explanation should hold, it can never explain the increase of the
crystallite size with increasing the Si3N4content above about 20 mol.%, which is
close to the percolation threshold in three dimensional f.c.c. system [120]. The
crystallite size of TIN should continuously decrease with increasing content of
Si3N4if the development of the microstructure were kinetically controlled.
Therefore, we conclude that the minimum crystallite size together with the
concomitant maximum hardness of the ncTiN/aSi3N4 composites is a result of a
(relative?) thermodynamic stability of such a nanostructure at the percolation
threshold (see remark added in proof and ref. [170]). We recall that at this percola-
tion threshold the HR-TEM and XRD data show that the nanostructure consists of
isolated TiN nanocrystals with a nearly spherical shape and only very few nanocrys-
tals touching each other [61,62], that is, there is no indication of bicontinous or an
interwoven bi-phase systems. Such a nanostructure with the minimum crystallite
size has a maximum specific area of the interface. If, as experimentally observed,
the system adjusts the minimum crystallite size at the percolation threshold resulting
in the maximum specific area of the interface, this interface must possess an unusual
stability. As such a behavior was also found with the other systems (ncW2N/aSi3N4
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
122 4 Nanostructured Superhard Materials
4-1 F
1-
O0O0
-(u3
(I)
v)
4000
4
61 r 6000
-....
-5000
E E
-4000 $
0 Y
0 -3000 ;
o-, o o o o-,
5
a,
0 -2000
m
+
Tdep 09.07.97, [Si] 10.3%, [Cl] 0.9%
H = 2650 kg/rnrn2, d = 3.4prn
silicon - substrate
'-1000
5000 kgmm-* remained unchanged up to the highest temperature before the sample
peeled off. One example of such measurements is shown in Fig. 8a. For reasons
which are not fully understood yet [122] the hardness of the composites deposited
on Si-wafers is always somewhat lower. With this in mind we show in Fig. 8b
that for such samples both the crystallite size and the hardness do not change
upon isochronal annealing for 0.5h for each step up to 1100°C and only a small
coarsening is observed at 1150°C. Because Si3N4 decomposes under vacuum at
this temperature (saturation pressure of about lop6atm [74]), the absence of
coarsening (Ostwald ripening) is surprising. The lack of any change of the hardness
is just a consequence of the stability of the nanostructure. Nevertheless, this clearly
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
124 4 Nanostructured Superhard Materials
indicates that the grain boundaries and the ncTiN/aSi3N4interface possess a sur-
prising stability. In view of the fact that the maximum annealing temperature of
1100°Cwhere no coarsening occurred is about 0.7 Tdecat 1 atm of Si3N4,this finding
merits more detailed discussion (see below).
In contrast to these results, ncTiN/aSi3N4composites with a larger crystallite size
of 2 6 nm show recrystallization accompanied by a decrease of the hardness already
at 800°C. This suggests that the stabilizing effect of the ncTiN/aSi3N4 interface
increases with decreasing crystallite size.
The choice of boron nitride instead of Si3N4 is one of the most attractive
possibilities for several reasons: BN (in particular cBN) does not react with
ferrous-, aluminum-, and many other alloys and is fairly stable against oxidation
in air at elevated temperatures. Moreover, the equilibrium in the system described
by reaction (1 1) is shifted much more to the left hand side than in the reaction (10).
x+l
M,N + xBN 1 M,B, +-2
N2. (11)
This is due to the much higher stability of BN (AGf(1300K) = -68 kJ/mole and
atom) at high temperatures as compared with Si3N4 (AGf(1300K) = -44.9kJ/
mole and atom) [74]. The boron nitride matrix possesses also a high structural flexibil-
ity because of the coordinational flexibility of boron (threefold in hBN, fourfold in
cBN, and mixed in aBN). Furthermore, because TiB2 and many other transition
metal borides are refractory materials, also the ncM,N/(a and nc)BN interface
should have a high cohesive energy which is required for blocking the grain boundary
sliding and achieving the superhardness. These general considerations were recently
confirmed by experiments in which we have demonstrated that the hardness in the
ncTiN/(a and nc)BN coatings reached 7000 kgmm-2. This is important particularly
in comparison with the earlier results of other researchers to be briefly discussed here.
Using a conventional plasma PVD (e.g. sputtering) at a low pressure where such
conditions are not fulfilled leads to the formation of homogeneous phases (meta-
stable solutions), such as Til_,A1,N (see e.g. [123-1251) or Ti-B-N [126-1281.
Other authors have found phase segregation with the formation of binary phases
TIN and TiB2 in the Ti-B-N coatings with relatively little nitrogen [129-1361
corresponding to ‘region 4’ of the equilibrium phase diagram [73,131]. These appar-
ently contradictory results of different research groups can be probably attributed to
different plasma conditions at the surface of the substrate during the deposition, that
is, the above mentioned activation as well as to different film composition (mainly
nitrogen content). Although these data are not quite clear from the published papers
one can find there data supporting this hypothesis. Gissler, Mitterer, and coworkers
found that the deposited Ti-B-N films with a relatively low nitrogen content formed
two segregated binary phases TIN (or TiNl - J + TiB2 as expected from the equili-
brium phase diagrams [129,131-1361. The hardness of the films depended on the
composition showing a maximum for a composition of about Til B0.6p1N0,5 where
the coatings had a very fine, almost isotropic nanocrystalline microstructure [ 131-
1331. The value of the maximum hardness depended on the substrate temperature
and applied negative bias. For Tsubstrate of 3OW0O0C, ion energy of 126eV and
the ratio of the ion to atom flux to the substrate of 0.7, the hardness reached
about 50 GPa or slightly more for a nitrogen content of 5 15 at.% [134-1361. In gen-
eral, the hardness as a function of the composition did not follow the rule-of-mix-
ture but displayed a maximum at the given composition where the microstructure of
the films was very uniform (no columnar growth) and the crystallite size was 3-5 nm.
If the films were deposited at a nominal temperature of 20°C and post-annealed at
600-800°C, the hardness increased from 25-30 GPa to about 40 GPa after 30 min
[132]. Similar values of superhardness were obtained if the films were deposited at
about room temperature as ultrathin TiN/BN multilayers and post-annealed at
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
126 4 Nanostructured Superhard Materials
400°C for 20 h [131]. However, in spite of the significantly higher hardness as com-
pared with TiN, the performance of these films in cutting tests was slightly worse
than TIN and (TiA1)N coatings [ 131,132,1351. With increasing nitrogen content,
when the hexagonal BN (hBN) phase was formed, the hardness of the films strongly
decreased to 15-20 GPa for an overall composition of TiB2N2[135]. This decrease
was attributed to the formation of the soft hBN.
Andrievski has also found a strong hardness enhancement up to 52GPa in
TiBo.7No.5 thin films with a grain size of about 3 nm [137,138]. The microstructural
information regarding the possible segregation is unfortunately not available in his
papers. However, the development of the morphology of the films as a function of
the TiB2:TiN ratio (see Fig. 1 in [137]) is very similar to that found in our ncTiN/
aSi3N4and ncW2N/aSi3N4films (see Figs. 5 and 6 in [64]) thus suggesting that the
microstructure in his TiB2-TiN films corresponds to a similar two phase system.
In view of these results and the above discussed thermodynamic and kinetic con-
siderations we have chosen deposition conditions which promote both, the phase
segregation and the formation of cubic cBN which alone is ultrahard. This means
using a high plasma density, high deposition temperature and high ion bombard-
ment. The plasma CVD using halides, which can easily provide these conditions,
has the advantage of a large flexibility for fundamental research but is not the
best choice for large scale industrial processing (see the foregoing section). Therefore
a chlorine free, combined plasma CVD and PVD technique was developed for the
deposition of superhard ncTiN/BN coatings whose hardness reaches 70 GPa [99]
(see also remark added in proof and ref. [171]).
The list of possible superhard nitride nanocomposites can be further extended
including further immiscible systems. For example, the above mentioned metastable
solid solution Til -,Al,N coatings prepared by plasma PVD show a high stability
and corrosion oxidation at elevated temperature (see [123-125,1391 and references
therein). However, if deposited under conditions of a high activity of nitrogen, these
coatings will undergo segregation into the TIN and AlN phases which, under an
appropriate volume ratio, should also form a nanocrystalline superhard composite.
The TIN may act as template for the pseudomorphic growth of the metastable cubic
A1N [88,89,14&142] if the thickness of the TIN is sufficiently small. Similar effects
have been found also in the TiN/NbN, TiN/CrN, and TiN/CN, heterostructures
[88,89] and will surely be found in many others. Because CrN has similar tribo-
mechanical properties as TiN but a significantly higher oxidation resistance (see
e.g. [143,144] and references therein), nanocomposites based on this material in
combination with BN, AlN, Si3N4 or other immiscible nitrides have a large techno-
logical potential.
A very interesting template effect of TIN resulting in the development of a
superhard nanocomposite microstructure was reported by Sproul, Y.-W. Chung
and their co-workers for the TiN/CN, [96,97,145] and ZrN/CN, [90] systems.
The researchers were depositing heterostructures by unbalanced magnetron sputter-
ing from two opposing targets (Ti and graphite), between which the substrate was
mounted on a turntable. The substrate temperature was not controlled, but the
authors give 200°C as the upper limit. The thickness of the heterostructures was
varied by changing the rotation speed of the turntable. When the apparent ‘layer
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.3 Preparution and Properties of Superhard Nanocrystalline Composites 127
thickness’ (calculated from the deposition rate) decreased to 2 4 nm, the hardness
suddenly increased to almost 5000 kg mmP2 and the small angle X-ray diffraction
peaks from the layered structure vanished. The XRD showed a (1 11) texture of
the films and the HR-TEM micrographs revealed that the material consisted of
ncM,N/CN,x nanocomposite (M = Ti and Zr) with CN, pseudomorphic growth
on the M,,N template.
It is interesting to compare these TiN/CN, heterostructures and composites
prepared by reactive sputtering at a relatively low pressure of about 0.01 mbar
and temperature of 1200°C with the results obtained by plasma CVD at signi-
ficantly higher pressure of 1-5mbar and temperature of 560°C [67]. Because
carbon can substitute nitrogen in the TIN crystal lattice, TiC,N1-, forms a
homogeneous solid over the whole region of 0 < x < 1. The higher deposition
temperature and a high activation energy provided by the intense glow discharge
plasma during plasma CVD resulted in the formation of a well mixed titanium car-
bonitride TiC,Nl --I. The hardness of the films plotted against the composition
parameter, x, did not show any extremum, but followed monotonically the rule-
of-mixture increasing from the hardness of TIN of about 2000 kg mm-2 to that of
TIC of about 4000 kgmm-2 [67].
Knotek and coworkers [ 1151 reported on the preparation of ‘superstoichiometric
carbides’ MCI+x, which, in fact, consist of nanocrystals of stoichiometric carbides
imbedded in a aC:H matrix. With increasing excess, x,of the carbon, the hardness
of the films increases, reaches a maximum of about 4000 kgmmP2 and, afterwards,
decreases again upon a further increase of x in a similar manner as in our nano-
composites (Fig. 7). Similar results were recently reported by Voevodin and Zabinski
who used laser ablation for the deposition of ncTiC/aC films [I 16,1171. The maxi-
mum hardness of about 32GPa was found for the total carbon content of about
65at.% [116]. Although this is below the range of the ‘superhard’ materials, the
very high toughness of 70 N and low coefficient of friction of 0.15 [ 1171 are impressive.
The hardness of stoichiometric amorphous boron carbide of 25000 kg mm-2
[ 1461 exceeds significantly that of the crystalline B13C3 (often incorrectly written
as ‘B4C’ [147,148]) of about <3500kgmm-2. The fundamental unit of the crystal-
line BI3C3and of elemental boron is the B12boron icosahedron with a five-fold sym-
metry, which is incompatible with a closed packed three-dimensional crystalline
lattice. Therefore, the icosahedra are connected via bridging carbon (in BI3C3) or
boron (in cB) atoms [148]. The amorphous boron carbide consists also of the
boron icosahedra which are, however, randomly oriented and bridged. Thus,
aB,C (x M 4) can be considered as a composite consisting of randomly oriented
‘quasicrystals’ of the B-icosahedra which are imbedded in a carbon (or carbon
rich) amorphous matrix. Figure 9 shows the dependence of the hardness of aB,C
on the composition of the films [148,149].
The significant limitation regarding the applicability of these carbon containing
superhard nanostructured materials is the high solubility of carbon in ferrous
alloys (however, not in Al-alloys) and the low thermal stability of transition metal
and boron carbides against oxidation in air. The stability of BI3C3is limited by the
low melting point of the oxide of about 460”C, which allows a fast diffusion of
oxygen and resulting oxidation of the film. In the case of composites containing
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
128 4 Nanostructured Superhard Materials
x in B,C(s)
Figure 9. Hardness of amorphous boron carbide deposited from diborane and methane by plasma
CVD against the composition (upper scale) and the C : B ratio in the gas inlet.
aC and CN, one has to keep in mind that these materials decompose at a tempera-
ture of 30O-50O0C, which limits their possible applications.
These examples show that superhard nanocomposites of various composition can
be prepared and their properties tailored according to the desired applications. Many
of these nanocomposites may have limited applications due either to their chemical
and structural instability at elevated temperature or their solubility (or chemical
reactivity) with the machined material. However, the stable nanocomposites, such
as ncTiN/aSi3N4,ncTiN/BN or more generally ncM,N/A,Nb will surely find inter-
esting applications. Preliminary results from the dry, hard- and high-speed turning
which are currently being done at our institute in collaboration with the SHM
company are very promising (for more recent results see [171]).
during the deposition of the films. Consequently, the experimental result that
Ostwald ripening does not occur during the 30min annealing at 1100°C means
that the mean diffusion length of Ti in Si3N4must be much smaller than the thick-
ness of the aSi3N4tissue for the films at the percolation threshold, with the smallest
crystallite size and highest hardness. From the Einstein formula [153,154]
0 . 0 1 4 4 4
.-K 0.008-
z 0.006-
0.004-
0.002-
20 40 60 80 100 120
crystallite size [A]
Figure 10. Dependence of the measured tensile strain on the crystallite size for ncSi and ncTiN/
aSi3N4,see text.
energy strongly influences the mechanical and thermal properties (e.g. lattice
dilatation or contraction [159,160] and a decrease of the melting point [160]) and
the quantum localization phenomena dominate the optoelectronic properties of
solids [ 155,1601. In thin films of nanocrystalline materials lattice dilatation has
been found to increase with decreasing crystallite size for nanocrystalline silicon
[159] as well as for ncTiN [63] and ncW2N [64]. These data are reproduced in
Fig. 10. The lattice dilatation in films which do not have any significant biaxial
stress has been explained due to the density defect in the grain boundaries and
the anharmonicity of the interatomic potential [161]. (Our ncTiN/aSi3N4 films
have a very small compressive stress of 0.2-0.5 GPa [ 1621. Energetic ion bombard-
ment during the deposition can increase this stress to 5GPa [163].) Because the
elastic lattice energy increases with the second power of the strain [22] it
represents a destabilizing contribution to the total free energy of the system
regardless whether it is compressive or tensile. Thus, there has to be some stabilizing
contribution. Considering the Gibbs free enthalpy of the nanocomposite given by
Eq. (14), the only possibility seems to be to assume that the stabilizing effect is
associated with a high cohesive energy Eintof the ncM,N/aSi3N4 interface
A G j ( d ) = A G $ ( C O ) + + N ~ T E E-
* ~N~r d 2 ( E $ +A&,,). (14)
Here, A G j ( d ) and AG;(co) are the Gibbs free energy of the nanocrystalline and
coarse grained systems, respectively, the second term describes the destabilizing
effect of the lattice strain (see Fig. 10) and the last term the stabilizing effect of
the grain boundaries. N is the number of nanocrystals per mole, E$ > 0 is the
increase of the cohesive energy (dependent on d ) and is the entropy term of
the interface.
The decrease of the surface energy of the ionic transition metal nitride by ‘wetting’
it with aSi3N4 may account for the initial decrease of the crystallite size with
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
132 4 Nanostructured Superhard Materials
increasing silicon nitride content, but it cannot explain the increase of the crystallite
size at high Si3N4 content above the percolation threshold. Therefore, Eint must
reach a maximum stabilizing effect at the minimum crystallite size and the thickness
of the Si3N4interface at the percolation into separate ncTiN and aSi3N4 phases.
With a further increase of the Si3N4 content this stabilizing effect must decrease
and vanish, that is, E i t 5 E i y ) for d > d(min) M 3-3.5nm. Thus, for larger
crystallite size of d > 5-8nm the destabilizing effect of the grain boundaries
dominates and the thermodynamically driven Ostwald ripening (Eq. (12)) occurs.
We can not offer any complete theoretical explanation of this phenomenon now,
but only a hypothesis which may explain the experimentally observed results.
The possible explanation is associated with the changes of the band structure due
to quantum confinement phenomena in the nanocrystals and resultant increase of
the cohesive energy at the interface. Indeed, a strong increase of the band gap
was theoretically calculated [164,1651and experimentally observed in silicon [1661
as well as in compound semiconductors [160]. In metallic nanoclusters deposited
on a carbon substrate, the conduction band narrows and the Fermi level shows a
downshift [160,167]. This means, that the bonds which are formed at the nc/a
interface by a combination of the states near the Fermi level of the nanocrystalline
transition metal and the upper edge of the valence band of the 5 1 nm thin Si3N4
should also show downshift with decreasing crystallite size as long as the aSi3N4
tissue is thin. In such a case the system gains stability by increasing the specific
area of that interface. However, when the fraction of Si3N4 increases above the
value corresponding to the percolation threshold, its thickness increases and
the band gap decreases towards the bulk value. Thus, the stabilizing effect at the
nc/a-interface is lost and there is no driving force which would maximize its area.
Consequently, the crystallite size increases again. The observed minimum of the
crystallite size at the percolation threshold corresponds to a minimum of the total
free energy of the system, that is, to a balance between this stabilizing effect at the
interface and the destabilizing increase of the elastic lattice energy associated with
the strain.
This hypothesis, of course, needs further theoretical development and experi-
mental verification. One may imagine other effects which could explain the one or
the other observation. For example, the stability of the nanostructure might be
explained by a strong immiscibility of the TIN and Si3N4phases in a similar way
as it is believed to apply for heterostructures prepared of materials which show
high immiscibility [89,95,168,169]. However, such an explanation is in contra-
diction with the phase diagram of the TiN/Si3N4 at high temperature and low
pressure of nitrogen. Furthermore, it cannot explain why the ncTiN/aSi3N4 films
with a large crystallite size coarse at much lower temperature and why, during the
deposition, the crystallite size passes a minimum at the Si3N4 content
corresponding to the percolation threshold. For similar reasons also a strain
induced decrease of the diffusion in the ncTiN/aSi3N4 is unlikely to play any
important role.
Of course, one should keep in mind that the high structural stability of the nano-
structure at high temperature is not necessary condition for the occurrence of the
superhardness. Any phenomenon which avoids the grain boundary sliding will
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 Conclusions 133
4.4 Conclusions
The practical strength and hardness of materials can be increased by an appropriate
control of the microstructure, which hinders the multiplication and movement of
dislocations and growth of microcracks. One such possibility is the decrease of
the grain size which leads to a decrease of the dislocation activity and of the size
of microcracks and an increase of the strength and hardness by a factor of 3-8.
However, a new deformation mechanism, grain boundary sliding, becomes
important in nanocrystalline materials when the grain size decreases below about
10-20 nm. Grain boundary sliding results in a softening of nanocrystalline metals
typically for a grain size less than 6-10 nm.
Grain boundary sliding can be avoided by an appropriate design of nanocrystal-
line composites, such as ncTiN/aSi3N4, ncVN/aSi3N4, ncW2N/aSi3N4,ncTiN/BN,
and others (Si3N4and BN content of 17-23 mole.%), which have a small crystallite
size of 3-9 nm and a high cohesive energy of the ncja interface. These nanocompo-
sites show a high elastic recovery of 280% and Vickers microhardness exceeding
5000 kg mm-*. Presently available data indicate that these nanostructures are
thermodynamically stabilized and do not show Ostwald ripening up to a tempera-
ture of 1100°C. Films with either a larger or smaller content of Si3N4 than that
corresponding to the percolation threshold of about 20 mole.% have a larger
crystallite size, smaller hardness and show coarsening (Ostwald ripening) by
800°C. A hypothesis based on a strengthening of the interface due to quantum
confinement phenomena is suggested to explain these observations.
For many practical applications, such as tribological coatings of machine tools
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
134 4 Nanostructured Superhard Materials
for ferrous metals, aluminum alloys, superalloys, and others, nanocomposites which
do not contain silicon (which is soluble in such alloys), new ncM,N/a or ncA,Bb
nanocomposites were suggested and their preparation investigated. Among
others, ncTiN/a and ncBN nanocomposites, which were prepared by a combined
plasma PVD and CVD technique without any halides, showed a very high hardness
of 6000-7000 kg mm-2.
Some theoretical aspects were discussed to show the generality of the principle
for the design of superhard nanocomposites. The present review was complemented
by a brief summary of the recent development in the field of heterostructures
which represent another approach to the design of superhard nanostructured
materials.
Acknowledgments
I should like to thank the coworkers from my institute for their enthusiastic
collaborations, my colleagues and friends Profs. Li Shizhi, H. Warlimont, V.
Vitek, and many others for illuminating discussions, and my wife Dr. Maritza
Vepiek-Heijman for her valuable comments on the manuscript. I am very much
indebted to Drs. M. Jilek, M. Sima, and P. Holubar from the SHM Company for
a very fruitful collaboration towards the industrialization of our superhard nano-
composite coatings. Partial financial support of our work by the German Science
Foundation (DFG) and the Bavarian Ministry for Education and Science is
thankfully acknowledged.
References
1. D. Tabor, The Hardness of Metals, Clarendon Press, Oxford, 1951.
2. Ch.-M. Sung and M. Sung, Muter. Chem. Phys. 1996,43, 1-18.
3. R. W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, 3rd edn.,
Wiley, New York, 1989.
4. M. L. Cohen, Phys. Rev. B, 1984, 32, 7988-7991.
5. A. Y. Liu and M. L. Cohen, Science, 1989,245, 841-842.
6. M. L. Cohen, Solid State Commun. 1994, 92, 45-52.
7. M. L. Cohen, Mater. Sci. Eng. A , 1996, 209, 1 4 .
8. S. Vepfek, J. Weidemann, and F. Glatz, J . Vac. Sci. Technol. A , 1995, 13, 2914-2919.
9. Y. Zhang, Z. Zhou, and H. Li, Appl. Phys. Lett. 1996,68, 634-636.
10. Y. Zhang, private communication (1997).
11. Y . Chen, L. Guo, and E. G. Wang, Mod. Phys. Lett. B, 1996, 13, 615-617.
12. Y. Chen, L. Guo, and E. G. Wang, J . Phys.: Condens. Muter. 1996,8, L685SL690.
13. Y. Chen, L. Guo, and E. G. Wang, Philos. Mag. Lett. 1997, 75, 155-162.
14. E. C. Cutiongco, D. Li, Y.-W. Chung, and C. S. Bhatia, J . Tribol. 1996, 118, 543-548.
15. A. Khurshudov, K. Kato, and S. Daisuke, J . Vac. Sci. Technol. A , 1996, 14, 2935-2939.
16. X. Yun, R. C. Hsiao, and D. B. Bogy, IEEE Trans. Magn. 1997, 33, 938-943.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 135
17. H. Sjostrom, S. Stafstrom, M. Boman, and J.-E. Sundgren, Phys. Rev. Lett. 1995, 75, 1336-
1339; see also errata 1996, 76, 2205.
18. H. Sjostrom, L. Hultman, J.-E. Sundgren, S. V. Hainsworth, T. F. Page, and G. S. A. M.
Theunissen, J . Vac. Sci. Technol. A , 1996, 14, 56-62.
19. W. T. Zheng, H. Sjostrom, I. Ivanov, K. Z. Xing, E. Broitman, W. R. Salaneck, J. E. Greene,
and J.-E. Sundgren, J. Vac. Sci. Technol. A , 1996, 14, 2695-2701.
20. G. B. Bokii, N. F. Kirova, and V. I. Nepsha, Sov. Phys. Doklady, 1979, 24, 83-84.
21. A. F. Goncharov, Sov. Phys. Uspechy, 1987, 30, 525-534.
22. Ch. Kittel, Introduction in Solid State Physics, Wiley, New York, 1971.
23. E. Hornbogen and H. Warlimont, Metallkunde, Springer, Berlin, 1996.
24. A. Kelly and N. H. MacMillan, Strong Solids, 3rd edn., Clarendon Press, Oxford 1986.
25. H.-H. Behnke, Harterei Technische Mitteilungen, 1993,48, 3-10.
26. M. F. Doerner and W. D. Nix, J . Muter. Res. 1986, 1, 601-609.
27. T. F. Page and S. V. Hainsworth, Surf. Coat. Technol. 1993,61, 201-208.
28. W. C. Oliver and G. M. Pharr, J . Muter. Res. 1992, 7, 613-617 and 15641583.
29. K. G . Schmitt-Thomas, Metallkundejiir das Maschinenwesen, Vol. I & 11, Springer, Berlin, 1990.
30. E. 0. Hall, Proc. Phys. SOC.B, 1951, 64, 747-753.
31. N. J. Petch, J . Iron Steel Inst. 1953, 174, 25-28.
32. A. Lasalmonie and J. L. Strudel, J . Muter. Sci. 1986, 21, 1837-1852.
33. N. Hansen, Metall. Trans. A , 1985, 16A, 2167-2190.
34. U. F. Koks, Metall. Trans. 1970, 1, 1121-1143.
35. A. Inoue, H. M. Kimura, K. Sasamori, and T. Matsumoto, Muter. Trans. Jpn. Inst. Metals,
1994,35, 85-94.
36. H. Hahn and K. A. Padmanabhan, Philos. Mag. B, 1997,76, 559-571.
37. R. W. Siegel, Spektrum der Wissenschaft, 1997, 3, 62-67.
38. R. W. Siegel and G. E. Fougere, Muter. Res. Soc. Symp. Proc. 1995, 362, 219-229.
39. R. W. Siegel and G. E. Fougere, NanoStructured Muter. 1995, 6 , 205-216.
40. J. E. Carsley, J. Ning, W. W. Milligan, S. A. Hackney, and E. C. Aifantis, NanoStructured
Muter. 1995, 5, 441448.
41. A. H. Chokshi, A. Rosen, J. Karch, and H. Gleiter, Scripta Metall. 1989, 23, 1679-1684.
42. G. W. Nieman, J. R. Weertman, and R. W. Siegel, Scripta Metall. 1989, 23, 2013-2018.
43. J. S. C. Jang and C. C. Koch, Scripta Metall. Muter. 1990, 24, 1599-1604.
44. K. Lu, W. D. Wie, and J. T. Wang, Scripta Metall. Muter. 1990, 24, 2319-2323.
45. G. W. Nieman, J. R. Weertman, and R. W. Siegel, Scripta Metall. Muter. 1990, 24, 145-150.
46. G. W. Nieman, J. R. Weertman, and R. W. Siegel, J . Muter. Res. 1991, 6, 1012-1027.
47. G. E. Fougere, J. R. Weertman, R. W. Siegel, and S. Kim, Scripta Metall. Muter. 1992, 26,
1879-1 883.
48. R. Z. Valiev, F. Chmelik, F. Bordeaux, G. Kapelski, and B. Baudelet, Scripta Metall. Muter.
1992,27, 855-860.
49. A. M. El-Sherik, U. Erb, G. Palumbo, and K. T. Aust, Scripta Metall. Muter. 1992,27, 1185-
1188.
50. (a) P. G. Sanders, C. J. Youngdahl, and J. R. Weertman, Muter. Sci. Eng. 1997, A234, 77-82;
(b) P. G. Sanders, J. A. Eastman, and J. R. Weertman, Acta Muter. 1977, 45, 40194025;
(c) C. J. Youngdahl, P. G. Sanders, J. A. Eastman, and J. R. Weertman, Scripta Muter.
1997,37, 809-813.
51. J. Musil and F. Regent, J . Vac. Sci. Technol. 1998, A16, 3301-3305.
52. M. F. Ashby, Acta Metal. 1972, 20, 887-897.
53. H. Luthy, R. A. White, and 0. D. Sherby, Mafer. Sci. Eng. 1979, 39, 211-216.
54. F. A. Mohamed and T. G. Langdon, Metall. Trans. 1974,5,2339-2345.
55. J. Crampon and B. Escaig, J . Am. Ceramid Soc. 1980, 63, 686686.
56. R. C. Gifkins, Metall. Trans. 1976,7A, 1225-1232.
57. J. W. Edington, Metall. Trans. 1982, 13A, 703-715.
58. S . Yip, Nature, 1998, 391, 532-533.
59. J. Schiotz, F. D. Di Tolla, and K. W. Jacobsen, Nature, 1998, 391, 561-563.
60. W. W. Milligan, S. A. Hackney, M. Ke, and E. C . Aifantis, NanoStructured Muter. 1993, 2,
267-276.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
136 4 Nunostructured Superhard Materials
61. S. Vepiek, S. Christiansen, M. Albrecht, and H. P. Strunk, Muter. Res. SOC.Symp. Proc. 1997,
457,407412.
62. S. Christiansen, M. Albrecht, H. P. Strunk, and S. Vepfek, J. Vac. Sci. Technol. B, 1998, B16,
19-22.
63. S. Vepfek and S. Reiprich, Thin Solid Films, 1996, 268, 64-71.
64. S. Vepiek, M. Haussmann, and S. Reiprich, J. Vuc. Sci. Technol. 1996, A14, 4 6 5 1 .
65. S. Vepfek, Proc. of the 10th Znt. Conf. on Thin Films, Salamanca, Spain, September 1996, Thin
Solid Films, 1998, 317,449-454.
66. S. Vepiek, M. Haussmann, and Li Shizhi, Electrochem. SOC.Proc. 1996,96-5,619-628.
67. S. Vepfek, M. Haussmann, S. Reiprich, Li Shizhi, and J. Dian, Surf. Coat. Technol. 1996,86/
87, 394401.
68. G. W. Samsonow, Nitridi, Naukova Dumka, Kiev, 1969 (in Russian).
69. H. Holleck, J. Vuc. Sci. Technol. A , 1986, 4, 2661-2669.
70. H. Fischmeister and H. Jehn, Hurtstofschichten zur Verschleissminderung, DGM-Informa-
tionsgesellschaft, Stuttgart, 1987.
71. S. Sambasivan and W. T. Petuskey, J. Muter. Res. 1994, 9, 2362-2369.
72. T. Hirai and S. Hayashi, J. Muter. Sci. 1982, 17, 1320-1328; 1982, 17, 3336-3340; 1983, 18,
240 1-2406
73. P. Rogl and J. C. Schuster (Eds.), Phase Diagrams of Ternary Boron Nitride and Silicon Nitride
Systems, ASM International, Materials Park, OH, 1992.
74. M. W. Chase, C. A. Davies, J. R. Downey, D. J. Fruip, R. A. McDonald, and A. N. Syverud,
JANAF Thermochemical Tables, 3rd edn., Phys. Chem. Re$ Data, 1985, 14, No. 1.
75. J. S. Koehler, Phys. Rev. 1970, B2, 547-551.
76. S. L. Lehoczky, J. Appl. Phys. 1978,49, 5479-5485.
77. S. L. Lehoczky, Phys. Rev. Lett. 1978, 41, 1814-1818.
78. C. A. 0. Henning, F. W. Boswell, and J. M. Cobert, Acta Metall. 1975, 23, 193-197.
79. R. F. Bunshah, R. Nimmagadda, H. J. Doerr, B. A. Movchan, N. I. Grechanuk, and E. V.
Dabizha, Thin Solid Films, 1980, 72, 261-275.
80. S. Menezes and D. P. Anderson, J. Electrochem. SOC.1990, 137, 440-444.
81. U. Helmersson, S. Todorova, S. A. Barnett, J.-E. Sundgren, L. C. Markert, and J. E. Greene,
J. Appl. Phys. 1987, 62, 481-484.
82. M. Shinn, L. Hultman, and S. A. Barnett, J. Muter. Res. 1992, 7, 901-911.
83. X. Chu, M. S. Wong, W. D. Sproul, S. L. Rohde, and S. A. Barnett, J. Vac. Sci. Technol. 1992,
A10, 1604-1609.
84. K. M. Hubbard, T. R. Jervis, P. B. Mirkarimi, and S. C. Barnett, J. Appl. Phys. 1992,72,4466-
4478.
85. P. B. Mirkarimi, L. Hultman, and S. A. Barnett, Appl. Phys. Lett. 1990, 57, 2654-2656.
86. P. B. Mirkarimi, S. A. Barnett, K. M. Hubbard, T. R. Jervis, and L. Hultman, J .Muter. Res.
1994,9, 1456-1467.
87. K. K. Shih and D. B. Dove, Appl. Phys. Lett. 1992,61, 654-656.
88. W. D. Sproul, Surf. Coat. Technol. 1996,86/87, 170-178.
89. W. D. Sproul, Science, 1996, 273, 889-892.
90. M. L. Wu, X. W. Lin, V. P. Dravid, Y. W. Chung, M. S. Wong, and W. D. Sproul, J. Vac. Sci.
Technol. 1997, A15, 946950.
91. S. A. Barnett, in Physics of Thin Films, Vol. 17, Mechanics and Dielectric Properties, M. H.
Francombe and J. L. Vossen (Eds), Academic Press, Boston, 1993, p. 2.
92. J. E. Krzanowski, Scripta Metall. Muter. 1991,25, 1465-1470.
93. X. Chu and S. A. Barnett, J. Appl. Phys. 1995, 77,44034411.
94. S. A. Barnett, Invited Lecture presented at the TSM Annual Meeting, San Antonio, Texas,
February 1998 unpublished (see also [168, 1691).
95. P. M. Anderson and C. Li, Nunostructured Muter. 1995,5, 349-362.
96. D. Li, X. Chu, S. C. Cheng, X. W. Lin, V. P. Dravid, Y. W. Chung, M. S. Wong, and W. D.
Sproul, Appl. Phys. Lett. 1995,67,203-205.
97. D. Li, X. W. Lin, S. C. Cheng, V. P. Dravid, Y. W. Chung, M. S. Wong, and W. D. Sproul,
Appl. Phys. Lett. 1996, 68, 1211-1213:
98. T. L. Selinder, M. E. Sjostrand, A. Ostlund, M. Larsson, S. Hogmark, and P. Hedenqvist,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 131
Proc. 14th Int. Plansee Seminar, G. Kneringer, P. Rodhammer, and P. Wilharditz (Eds), Plan-
see AG, Reutte, Austria, 1997, p. 44.
99. S. Vepiek, P. Nesladek, A. Niederhofer, F. Glatz, M. Jilek, and M. Sima, 25th Int. Conf.
on Metallurgical Coatings and Thin Films, San Diego, April 1998, Surf. Coat. Technol. 1998,
108-109, 138-147.
100. S. Vepiek, S. Reiprich, and Li Shizhi, Appl. Phys. Lett. 1995, 66, 2540-2642.
101. Li Shizhi, Shi Yulong, and Peng Hongrui, Plasma Chem. Plasma Proc. 1992, 12, 287-297.
102. S. Vepiek, A. Niederhofer, P. Nesladek, and F. Glatz, Electrochem. SOC.Proc. 1997, 97-25,
3 17-324.
103. A. von Engel, Ionized Gases, 2nd edn., Clarendon Press, Oxford, 1965.
104. S . C. Brown, Introduction to Electrical Discharges in Gases, Wiley, New York, 1966.
105. G. Francis, Ionization Phenomena in Gases, Butterworth, London, 1960.
106. S. Vepfek, J. Th. Elmer, Ch. Eckmann, and M. Jurcik-Rajman, J. Electrochem. SOC.1987,136,
2398-2405.
107. S. Vepiek, Ch. Eckmann, and J. Th. Elmer, Plasma Chem. Plasma Proc. 1988,8,445466.
108. S. Vepi-ek, C. Brendel, and H. Schafer, J. Cryst. Growth, 1971, 9, 266-272.
109. S. Vepiek, in Current Contents in Materials Science, E. Kaldis (Ed.), North-Holland, Amster-
dam, 1979, pp. 151-236.
110. J. Patscheider, Li Shizhi, and S. Vepiek, Plasma Chem. Plasma Proc. 1996, 16, 341-363.
11 1. M. Diserens, J. Patscheider, and F. Levy, 25th Int. Conf. on Metallurgical Coatings and Thin
Films, San Diego, April 1998, Surf. Coatings Technol. 1998, 108-109, 241-246.
112. F. Vaz, L. Rebouta, S. Ramos, M. F. da Silva and J. C. Soares, 25th Int. Conf. on Metallurgical
Coatings and Thin Films,San Diego, April 1998, Surf. Coat. Technol. 1998, 108-109,236240.
113. H. Jensen, J. Sobota, and G. Sorensen, Surf. Coat. Technol. 1997, 94/95, 174178.
114. H. Jensen, J. Sobota, and G. Sorensen, 25th Int. Conf. on Metallurgical Coatings and Thin
Films, San Diego, April, 1998, Surf. Coat. Technol. unpublished.
115. 0 . Knotek, private communication, 1997.
116. A. A. Voevodin and J. S. Zabinski, J . Muter. Sci. 1998, 33, 319-327.
117. A. A. Voevodin and J. S . Zabinski, Diamond Rel. Mater. 1998, 7, 463467.
118. H. P. Klug and L. E. Alexander, X-Ray DzfSraction Procedures, Wiley, New York, 1974.
119. M. Riickschloss, PhD Thesis, Technical University, Munich, 1993.
120. R. Zallen, The Physics of Amorphous Solids, Wiley, New York, 1983. See chapter 4, in parti-
cular Figure 4.5for a two dimensional network and Table 4.4. Notice the following important
facts: ( I ) The structural percolation corresponds to the situation when the percolating phase
(here Si3N4)forms a continuous connection (‘cluster’) through the matrix phase (here TIN). It
occurs at a critical fraction of the percolating phase (here aSi3N4)Pc which depends on the
nature (site or bond percolation), structure see e.g. the f.c.c., b.c.c., and diamond lattices in
Table 4.4) and dimensionality of the network (here TIN). Pc = 1 for one dimensional
chain, 0.354.7 for two and 0.124.4 for three dimensional network. For the f.c.c lattice of
TiN Pc N 0.2 for the site percolation as observed in the present work. (2) Because of different
near-threshold scaling laws, the percolation phenomena of electrical and optical properties,
such as conductivity and optical reflection, commence only at a fraction of the percolating
phase which is significantly larger than P, (see e.g. Figure 4.5 and Table 4.3 in [120] and
the change of the reflectivity of the ncTiN/aSi3N4 films reported in [67]). (3) Even above the
percolation threshold (P(Si3N4)> Pc) there still may be some local connectivity within the
matrix (nc-TiN). In our particular case it means, that some of the TIN nanocrystals can be
in contact with each other.
121. T. E. Fischer, private communication, 1998.
122. So far, only films deposited on Si show a lower hardness. Films deposited on hard metal (Co
cemented tungsten carbide) or on such a substrate precoated with TiN or Ti, -.Al,N show the
same or somewhat larger hardness as those deposited on steel. For both steel and silicon sub-
strates the crystallite size and the Si3N4content are equal. The possible reasons of this effect
are under investigation.
123. 0. Knotek, M. Bohmer, and T. Leyendecker, J. Vac. Sci. Technol. 1986, A4, 2695-2700.
124. T. Leyendecker, 0. Lemmer, S. Esser, and J. Ebberink, Surf. Coat. Techno/. 1991, 48, 175-
178.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
138 4 Nanostructured Superhard Materials
161. S. Vepiek, F.-A. Sarott, and Z. Iqbal, Phys. Rev. B, 1987, 36, 3344-3350.
162. P. Nesladek and S. Vepiek, unpublished results, 1998. See also [170].
163. S . Vepfek, in Proc. of Muter. Res. Soc. Europe, Strasbourg, 1984, P. Pinnard and S. Kalbitzer
(Eds), Les Cditions de physique, Les Ullis, France, 1984, pp. 425444.
164. C. Delerue, M. Lannoo, and G. Allan, J . Lumin. 1993, 57, 249-256.
165. B. Delley and E. F. Steigmeier, Phys. Rev. B, 1993, 47, 1397-1400.
166. S. Vepiek and Th. Wirschem, in Handbook of Optical Properties Vol. ZZ: Optics of Small
Particles, Znterfaces and Surfaces, R. E. Hummel and P. Wissmann (Eds), CRC Press, Boca
Raton, 1997, pp. 129-143.
167. S. B. DiCenzo and G. K. Wertheim, in Clusters of’ Atoms and Molecules ZZ, H. Haberland
(Ed.), Springer Verlag, Berlin, 1994, pp. 361-383.
168. S. Barnett and A. Madan, Physics World, 1998, 11, 4548.
169. A. Madan, Yun-yu Wang, S. A. Barnett, C. Engstrom, H. J. Ljungcrantz, L. Hultman, and
M. Grimsditch, J. Appl. Phys. 1998, 84, 776-779.
170. A. Niederhofer, P. Neslidek, H. Mannling, S. Vepfek and M. Filek, Surf. Coatings Technol.
(1999) in press (Proc. Int. Conf. Metallurgical Coatings and Thin Films, San Diego, April
1999).
171. P. Holubai, M. Jilek and M. Sima, Suyf. Coatings Technol. (1999) in press (Proc. Int. Conf.
Metallurgical Coatings and Thin Films, San Diego, April 1999).
172. A. Voevodin, ibid.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
5.1 Introduction
Corrosion may be defined as the physical and chemical alteration of a material due to
its interaction with the environment of interest. It must be emphasized that corrosion
resistance is not a material property but a system property and real environments have
a high variability in both chemical and physical boundary conditions.
Corrosive alteration leads to a change in the desired original properties and so
corrosion usually, but not exclusively, has adverse effects and causes a decrease in
strength, toughness, hardness, and increases wear. There are many reasons for the
degradation of properties by corrosion. Decrease of the cross-section carrying the
load, pitting, stress induced by phase changes, and grain separation are some of
the common features.
Some hard materials, namely Sic and Si3N4,are designed to serve at high tem-
peratures, where corrosion resistance is an essential feature. This is the reason
why a large proportion of the ceramic literature on corrosion, including recent
textbooks [ 1-51, deals with these materials.
In order to react with the environment, the material must not be in equilibrium
with it. The hard materials covered in this chapter are non-oxides and hence this
condition is fulfilled for all of them in almost every application environment, includ-
ing ambient air. The applicability of hard materials thus rests on the kinetics of the
processes.
Before discussing individual materials we will introduce important terms and
concepts in corrosion like attack modes, media classes, location of attack, kinetics,
and their parameters. We feel that this is necessary for understanding of the dis-
cussion of the materials, because there is no generally accepted standard procedure
that allows us to compare the corrosion behavior of materials as easily as other
properties like mechanical strength or hardness.
The more common classification scheme is to divide the corrosive media into their
state of aggregation, that is to subdivide into corrosion by solids, liquids and gases.
While solid state corrosion is rarely dealt with, we have vast amount on literature on
hot gas corrosion. The case of corrosion by liquids is commonly further subdivided
into more specific cases, such as aqueous corrosion (e.g. acids and water), corrosion
by glasses, metal melts and salt melts. The last case is for historic reasons known in
the form of a rather misleading expression: hot corrosion. A special case, which
spans from the liquid into the gaseous state is given by the corrosion in hot water
systems: hydrothermal corrosion.
At first glance this may not make too much sense, but the experience is that the
different names mentioned above correspond to main forms of the attack mode and
often refer to different application temperature levels.
local corrosion with bubble and pit formation is the result, at least in the early stages
of this oxidation [7].
If the secondary phase is a grain-boundary phase, which allows a relatively fast
transport of the corroding agent into the material via the grain boundaries, we
will have an attack of the main material and/or the grain-boundary phase not
only on the surface, but in a zone stretching out into the bulk of the specimen.
This is known as internal corrosion.
Despite its localized character on a microscopic scale, intergranular corrosion
may be quite homogenous on the macroscopic scale. Additionally, the process of
internal corrosion is also capable of slowing down with time, so it may offer some
protective quality.
It must be emphasized that those processes do not exclude each other. Internal
corrosion may occur while a scale is formed on the outside. We will discuss the
problem in the context of the individual materials.
and
Si + O2 H Si02.
Calculations of the oxygen partial pressure P(0,) for the boundary condition both
by simply taking the values from the JANAF-tables [12] or by using a modern
thermochemical program [ 131 yield similar results with an equilibrium P(0,) as
low as bar at 1000°C. The prediction from the equilibrium would thus be,
that at conditions with P ( 0 2 ) > lop2*bar we should find the formation of a silica
layer on a bare Si surface. Experimental observation shows that this only takes
place at P(02)-levels more then 20 orders of magnitude higher.
Wagner’s solution was to transform the problem into a physically controlled one.
Silica is formed at conditions, which allow its condensation from the gas phase in
equilibrium with the substrate:
4Si + Si02 * SiO. (3)
If reaction (1) is fast, a concentration gradient of O2 inwards and SiO outwards must
exist. Then the interdiffusion coefficients and the effective thickness of the boundary
layer will control whether the critical P(Si0) of the condensation reaction (3) is
reached or not. From tables and estimations of those physical constants we can
relate back to the oxygen pressure of the bulk gas necessary to induce the critical
P(Si0) on the surface. The calculation yielded reasonable results for the active-
passive boundary of Si in streaming atmospheres with low oxygen contents and
accordingly the theory was later applied to other silica-formers [ 14,151.
This success story demonstrates that equilibrium calculations disregarding physi-
cal boundary conditions can be misleading, and that thermochemical calculation is
capable of giving good results if those boundaries are taken into account.
J= -D-dc
dl ’
where I = diffusion path length. With constant conditions, the gradient from the
surface to the bulk of the corrosive medium becomes constant, 1 stays constant
and accordingly the flux, J , becomes constant with time. Therefore the change in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
144 5 Corrosion of Hard Materials
10
- 5
h
f
Y
c
.-0
In
In
a
0
0
2
v
-8
-5
0 1 2 3 4 5 6 7 8
log (time [s])
Figure 1. Predicted behavior of a model case for active corrosion under different external conditions.
with the logarithmic rate constant klog.The same equation is valid for a more
complex situation, pertinent to ceramic systems. If the diffusion through a scale is
rate controlling and the scale crystallizes a phase, which is either completely block-
ing for the agent or has a diffusion coefficient much smaller than the original scale,
the effective cross section for the agent transport is likewise diminished. For a con-
stant crystallization rate we will end up with Eq. (1 1).
However, natural nucleation and crystal growth is complex, dependent on many
variables and usually not constant with time. The energy requirements are often so
strongly temperature dependent that we see either none or full crystallization. Thus
in pure systems we will often have kinetic breaks with time or temperature rather
-
h
u)
._
c
=I 300
P
2
c
e
2.200
ln
u)
Q)
c
Y
.-0
100
Jt>.
-m
0
0 50 100 150 200 250 3 I0
Time (arbitrary units)
Figure 2. Basic passive laws.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Corrosion Kinetics 147
then a nice curve following Eq. (11) exactly. This may be different for systems in
which phase formation or crystallization is linked to chemical changes of the
scale, induced by diffusion of components from the interior.
The shapes of curves with behavior according to the basic laws is illustrated in
Fig. 2. It is obvious from Fig. 2 that data from a single point (here: t = 28 and
x = 100) would not allow prediction of the behavior at any other time unless the
law is specified.
*Oo0 I
n 1500-
=!
cu
P
Y
Y 1000-
.-0
5
5
- 500-
0 20 40 60 80 100 10
Time (a.u.)
Figure 4. Growth change due to developing diffusion coefficients, D.
30
Time (a.u.)
Figure 5. Typical development of layer growth with increasing temperature in porous media with
pore closure due to passive corrosion.
- 4
?
m
Y
% 2
c
m
c
0 0
E
.-0)
a
3 -2
-4
0 20 40 60 80
Time (a.u.)
Figure 7. Weight change curves obtained from the simultaneous action of parabolic weight gain and
two different linear weight loss rates (‘para-linear behavior’).
easy ranking they are hardly capable of producing quantitative data outside the
narrow conditions set by the standardized experimental set-up [25].
For hot corrosion testing burner rigs can be applied [26].The agent is injected into
the flame of the burner and the sample is held or agitated at a distance, which allows
deposition of the corrosive agent from the gas phase at a constant rate.
In the ordinary furnace test for gas corrosion the samples are placed in a furnace
for a given time in the atmosphere of interest. The advantage of the furnace test is
that many different samples can be tested in one set of conditions, the drawback that
it gives only one data point in time per experiment. The pitfalls are apparatus
related: Temperature and atmosphere constancy have to be guaranteed, cross-
contamination must be avoided, and interaction with furnace linings and crucible
material minimized. The danger of those points is often underestimated; there are
reports of different results from a mere change in the furnace size [27].
Contamination by furnace linings is a source of error for all corrosion experi-
ments. There is ample evidence that the commonly used linings of A1203-ceramics
do contaminate samples with A1 and alkalis via gas phase transport, even if high-
purity A1,03 is used. The contamination is strongly enhanced in wet atmospheres
[281.
In situ measuring of corrosion is mostly done by thermogravimetric measure-
ment, that is the monitoring of a weight change during a isothermal period of
time. The advantage here is that the true shape of changes is recorded, which
allows proper kinetic modeling. The limitations of the method come firstly from
the apparatus: Resolution, accuracy, precision, and the often unchecked long-
term stability are major points of concern. The pitfalls of contamination by furnace
interior and crucibles, and local atmosphere changes are identical to those men-
tioned above.
For coarse estimations of danger levels in corrosion nonisothermal thermo-
gravimetry or differential thermal analysis (DTA) can be used.
layer thickness
corroded size
For homogenous surface attack by active corrosion the case simplifies to the
recession as there is no scale growing and the remaining cross section is of virgin
material.
The parameter of interest is then the recession rate, which is identical to kl of
Eq. (6) for linear kinetics. In corrosion science the SI-units ms-' or the more
vivid unit mmyear-' of engineering practice are useful. The relation to mass loss
is obtained by multiplying by the density of the lost material to yield kgmp2s-'
or equivalent units.
It should be emphasized again that this parameter is useless unless the physical
boundary conditions are specified. The quotation of partial pressures of outgasing
species is often more meaningful.
A problem arises when the attack is localized, because the attack is then different
for different phases and the outer surface is no longer identical to the virgin material.
The evaluation of penetration depth and rate may be performed in some cases using
dye tests. If penetration goes with formation of porosity, dyes like methylene-blue or
phosphorescing agents can be applied to show the depth of corrosion in cross
sections.
More complex situations, such as with chemical modification of the grain-
boundary phase, may be measured in optical thin sections, but they need usually
more careful evaluation by chemical profiling with microanalytical techniques.
Note that the kinetics of active corrosion may then deviate strongly from linear
kinetics because leaching by liquid media or evaporation out of channels often
involves diffusional problems.
A k1 is then not the parameter for ranking. A better choice would be the reference
to critical effects, such as the 'time needed for 0.1 mm penetration at 0.7 T,' [36].
Regrettably, such parameters are extremely scarce in the literature.
As the experimental determination and the investigations afterwards are highly
time-consuming and expensive, an often followed method of characterization is
by the effect of corrosion on mechanical properties, usually strength or wear proper-
ties [37,38], in addition to thermogravimetric data.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.5 Corrosion Measurement 153
Changes (a.u.)
0.0 -
\
\
\
\
Moving to the more complex cases with scale growth, Fig. 9 shows the range of
kinetic curves one may find. Figure 9 shows a parametric study for a single case of
para-linear behavior. Not even the sign of the curves is constant. We must therefore
discuss the individual contributions.
For linear scale growth the discussion above is valid. For processes of retarded
scale growth with time we describe the layer growth by k, (units of m2 s-l or equiva-
lents) for simple parabolic behavior (Eq. (8)).
The calculation of the appropriate mass change unit is not as straightforward as
in the linear case. This is because the mass change recorded by weighing or thermo-
balance is measuring the difference in masses and not the mass of the layer.
The latter may be inferred if the chemical reaction and the properties of the scale
material are known, introducing a stoichiometric factor u. As an example the
reaction
Sic + 1 4 0 2 = SiO2 + CO (13)
means that I mole of the condensed product Si02 (= 1Msio2= 60.09g with a
density psio2 of 2 . 2 g ~ m -in
~ the amorphous state) is produced from 1 mole of
S i c (= 40.09 g). Thus a recorded mass change of Ax = nMproduct- nkfmaterial
(= 19.99g) in Eq. (12) represents the production of a full mole of S O 2 . In general
it follows that the factor u is defined from the reaction equation as
(nM)condensed product
U = (14)
(nM)condensed product - (nM)material '
which is 3 for reaction (13). Because of the form of Eq. (8) the two k, are hence
[ y1 [$1
related by
3
Pproduct
2
k,(layer-thickness) - = kp(mass-change)
U2
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
154 5 Corrosion of Hard Materials
Neglecting the factor u has lead to occasional inconsistencies in the literature. The
discussion of it seems trivial but a close look at Eq. (14) shows that for other cases,
for example
SiBCN3 + 2.25 O2 H S O 2 + 4B203+ CO + 1.5N2, (16)
u can become very large with the result that substantial scale production may occur
almost invisible to thermogravimetry.
The layer thickness derived from any of the kinetic equations may be recalculated
to give a solution for homogeneous recession:
From this discussion one might conclude that layer thickness measurements are the
better choice for monitoring passive corrosion. While this is true from the point of
resolution (ellipsometry allows the determination of scales down to the nm scale) it
is difficult for high-temperature in situ recording and it implies dense transparent
layers, which may not be the case.
Thus a safe characterization will ask for both quantities to be measured indepen-
dently. Inconsistencies between measured and calculated values from Eq. (15) will
then immediately point to problems concerning scale structure and/or internal
corrosion.
Additional recording of mechanical property changes has the advantage of
including information on potentially dangerous new flaw populations derived by
corrosion. The overlay of several parameters (flaw geometry, cross-section changes,
stress field changes) will not allow us to use mechanical testing as the only criterion.
Thus, at the present time we prefer to compare and rank materials in terms of time
needed for penetration to a certain depth at the temperature of interest to make any
type of kinetic equation accessible. The fixed ‘critical condition’ is taken to be
0.1 mm, following Ashby and Jones [36], the time unit is in years to give a measure
for engineering practice. An engineer may then be able to decide that a material is
not useful for long-term applications if it drops to log t, < 0 at the temperature of
interest where tc is the critical time in years for a penetration of 0.1 mm. In the
following sections on the materials this is calculated but the uncertainties have to
be kept in mind.
5.6 Materials
5.6.1 Diamond and Diamond-like Carbons
Diamond is known to resist all acids and most other chemical substances in the
low-temperature range to 100°C. However, transformation of diamond into
other carbon phases may occur under local contact stress at room temperature
[39,401
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materiuls 155
Diamond will be etched by caustic alkalis and some oxidizing salts, for example
by NaClO and KOCl at 380°C [41], and K and Na nitrates above 400°C. Diamond
reacts with metals, which form carbides (e.g. W, Ta, Ti, and Zr) or which dissolve
carbon (e.g. Fe, Co, Mn, Ni, and Cr) [42]. Detailed knowledge of the temperatures
of reactions with metals is reviewed in the context of experiments with metals for the
catalysis of diamond synthesis [43].
Diamond corrosion is usually a problem of oxidation and graphitization. Graphi-
tization in inert atmospheres was observed at about 1500°C and becomes very fast at
2100°, where sizable diamonds are converted to graphite within minutes [44]. An
extrapolation of the data gives a T M 1450°C for t, = 0. It has been reported that
graphitization is prevented in pure H2 up to 2200°C [45].
Surface graphitization associated with concurrent oxidation, which may be found
at T > 600°C, is not thought to be a true conversion but an autocatalytic surface
reaction involving CO, species.
The oxidation of diamond is clearly an active corrosion process. At least up to
700°C diamond has a fast reacting { 11 1)-plane, an intermediate { 1 lo} and a slow
{ 100)-plane, which indicates reaction control. At higher temperatures and/or
lower oxygen pressures gas diffusion becomes rate determining in analogy with
graphite [46],and this is indicated by a more even attack [47]. Hence corrosion
rates are faster or start at lower temperatures for fine powders compared to films
and the corrosion in air is faster than in low-oxygen environments [48].
At 720°C diamond will burn in an oxygen jet. It is often stated in the literature
that oxidation starts at 600°C. The theoretical onset of oxidation should be well
below 600"C, because from the etching rates of Evans [44] for the { 111)-plane in
1 bar oxygen a t, well below zero ( ~ 0 . 0 2 )is calculated. Linear extrapolation
would give T M 340°C for tc = 0. However, in analogy with graphite there may
be an accumulation of C-0 complexes at the surface, which is faster then their
removal between 300°C and 500°C [49]. Surface graphitization at T > 600°C
points in this direction.
Also in analogy with other carbon forms, impurities may act as catalysts for the
oxidation, which can decrease oxidation temperatures by more then 100°C and
increase rates by orders of magnitude [50].
To a first approximation the temperature sensitivity of the oxidation rate at
T < 800°C is less dependent on the structure of diamond (films) then on crystallo-
graphic orientation, the data for the diamond faces bracket those of thin layers [47].
As described in more detail in the contribution on hydrothermal synthesis of
diamond in this book, hydrothermal reactions start for fine powders at 7OO0C,
but a significant etching and dissolution of large diamond grains is observed only
at T 2 800°C [51].
5.6.2 Carbides
The carbides considered as hard materials include B4C, WC, and Sic. In oxidizing
environments, in particular, their behavior is so different that they can not be com-
pared directly.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
156 5 Corrosion of Hard Materials
5.6.2.1.1 Corrosion by Liquid Media There is only limited information about the
behavior of pure Sic in low-temperature water, acids, and bases. A relatively
pure sintered S i c is applied in many pumping systems for the chemical industry
[%].This provides evidence that it is hardly affected by chemical attack.
Significant attacks can occur via a grain-boundary phase in other Sic types, and
this is usually accompanied by a decrease in strength and Weibull modulus. Thus it
is clear that liquid-phase sintered S i c will show less corrosion resistance with a
behavior depending on the chemistry of the additives. For Y/Al-sintered S i c signi-
ficant corrosion was reported for a treatment in boiling 20% HC1 and 50% NaOH;
concentrated acids (HZSO4, H3P04)and bases (KOH) had much less effect [56].
Sic with free Si (Si-Sic) is affected in particular by hot strong bases (10M
NaOH) and hot diluted acids (H2SO4 at pH = 3) [57].
5.6.2.1.4 Corrosion by Salt Melts (Hot Corrosion) Data and reviews of the
mechanisms of the hot corrosion behavior of Sic have been presented in a
number of papers by Jacobson and coworkers [65-681.
A representative member of the hot corrosion agents is Na2S04. Its generation is
explained in combustion environments by
2 NaCl + SO2 + 0.5 O2 + H 2 0 H Na2S04+ 2 HCl. (18)
Since the action in the liquid state is regarded as most effective, the first condition
for the window of hot corrosion is given by the melting and dewpoint of Na2S04.
The melting point of 884°C does not vary much, the dewpoint depends on pressure,
S-content of the fuel and NaC1-concentration in the atmosphere. For conditions
with NaCl > 1 p.p.m. and combustion pressures of 100 bar it is calculated to be
1100-1200°C from Eq. (18).
However, the effective corrosion action is generally believed not to be by the
sulfate itself. It stems from the activity of N a 2 0 arising from the decomposition
Na2S04 * N a 2 0 + SO3 (19)
because only the N a 2 0 is attacking the protective scale of Si02 by
-
x Si02 + N a 2 0 H N a 2 0 (SiO2).,. (20)
This is the reason why other compounds releasing alkali oxide on decomposition,
such as Na2C03,have similar hot corrosion effects. It should be noted that complex
salts, in particular those containing vanadium, can have an even stronger hot
corrosion effect, as is well known from metals and oxides [69,70].
From Eq. (19) it can be calculated that only for conditions with a low P(SO3)will we
have a dissociation sufficient to drive reaction (20) to the right side. Consequently the
window of hot corrosion has a third axis, the P ( S 0 3 )of the combustion, and it has been
computed [71] that the window should only open for low-S fuels (<0.5%, a condition
met by many modern fuels) and at comparatively high temperatures (FZ>IOOO"C).
In reality the attack already starts at the melting point of Na2S04. This is
attributed to internal controlling factors. Many Sic-based materials contain free
carbon, which promotes the dissociation of Na2S04 and makes the silicate melt
more basic. The dramatic attack of Na2S04 on S i c with free carbon has been
confirmed experimentally [72].
As is described in more detail for Si3N4,recent studies suggest that free carbon is
an enhancing factor but not a necessary condition [73,74]. S i c may act as the
reducing agent to promote sulfate dissociation either directly or via an oxygen
gradient in the silica scale consumed by reaction (20).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
158 5 Corrosion of Hard Materials
1 1
(ac)"', pm
500 200 100 50
2 400
z --0.75K,c
-
d (a~)"~
300
5E
c
;200 0 Type A
0 Type B
A Type C
3
c
0
m
t 100 0 Type D
Figure 10. Relation of fracture stress and pit dimension from a study of four different Sic materials
1781.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 159
Figure 10 also shows that the relation with pit size is not true for all Sic-materials:
Type D shows no correlation and has a much stronger decrease in strength
(= -50%). It was a material with free silicon. It follows that Si-Sic has less hot
corrosion resistance and that the attack and its consequences are due to the
distribution of the preferentially corroded Si.
To improve the hot corrosion resistance current developments aim for a protec-
tion by mullite-based coatings [79,80].
5.6.2.1.5 Corrosion by Hot Gases Halogens do not form condensed species with Si
at the temperatures of interest. Thus, if they are pure (no oxygen or water present)
they will automatically induce active corrosion behavior and calculated pressures of
species like SiF4 or SiCI4over Sic will be as high as the concentration of F and C1 in
the gas.
In the presence of enough oxygen to form silica, these pressures will be signi-
ficantly reduced for the case of attack by C12 and accordingly lower corrosion
rates are measured [81]. However, pressures of SiC14 over Si02 can exceed
lop6bar at temperatures above 700°C in high-C12 environments.
Calculated pressures for an attack by H2 are likewise significant even at low
temperatures [8]. The calculated main low-T species is CH4, but the kinetics are
so unfavorable for its formation that in reality the beginning of significant active
corrosion is at temperatures above 13OO0C,as is known from etching studies [82].
This is only true for pure Sic: grain boundaries and secondary phases in sintered
S i c are attacked at temperatures as low as 1000°C [83].
For the survival of S i c it is therefore of great importance whether silica is formed
or not. The boundary for the active-to-passive oxidation is described in Fig. 1 1. It
shows the literature data and a Wagner-type theoretical model [15] based on the
H i m 76
Gulbranran
66
Vaughn
*
so
I 1 I 1
-8
1,200 1,400 1,600 1,800 2,000
Temperature ("C)
Figure 11. Transition temperatures and oxygen pressures for active and passive modes of oxidation
of Sic after [15].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
160 5 Corrosion of Hard Materials
influence of carbon activity and the presence of condensed SiO. Even though the
model might be challenged in view of new thermodynamic data, arguing against
condensed SiO, the general position of the boundary between active (I) and passive
oxidation is known from the experiments.
The region denoted active (11) in Fig. 11 is for a special situation, when the
pressures of the produced gas species (CO) exceed 1 bar, which can lead to bubble
formation and spalling, and hence quasi-active behavior, even though nominally
silica is produced. Experimental evidence for this was obtained at temperatures of
about 1760°C [84].
It should be noted that the oxygen pressures for the transition in Fig. 1 I should
not be confused with those obtained by COX-equilibria.Despite low oxygen partial
pressures of, for example CO in equilibrium with carbon, we have passive oxidation
of S i c up to x 1400°Cin CO, because CO is a reactive and oxidizing gas species for
S i c [52,85,86]! Consequently active corrosion in COX-environmentsis measured
only at very high temperatures [87].
In passive oxidizing environments S i c forms silica (Eq. (13)). For pure S i c a
parabolic behavior is evident [88]. For SSiC there is evidence that oxygen does not
penetrate deeply into the interior [89], leading to a relatively homogeneous surface
oxidation.
Therefore SSiC has often parabolic oxidation rate constants nearly as good as
chemical vapor deposition (CVD)-Sic and powder. From a data collection [53]
plus some more recent data [88] the critical time t, for 0.1 mm recession is computed
and shown in Fig. 12. We have omitted data from experimental changes, which were
deliberately made with high additive contents to study their influence on oxidation.
It is evident from Fig. 12 that basically all cases meet a long-term stability criter-
ion up to very high temperatures.
Some cautionary remarks have to be added: At very high temperatures the
kinetics may not be adequately described by the parabolic law; at T > 1700°C the
problem of active (11) oxidation will become important; SSiC can show bubble
and pit formation at much lower temperatures, which is usually attributed to the
additives B and/or A1 [90].
This bubble-and-pit formation is sometimes only present after a ‘incubation
period’ of several lOOh at T > 1600°C. Most oxidation studies run only for times
< 100 h. Therefore long-term studies may be necessary for true qualification of Sic.
Industrial utility of Sic is not guaranteed by a sufficient resistance to oxidation
by O2 or air. The influence of other components of the atmosphere or transferred
impurities has to be evaluated. Often this is difficult because strong variations in
temperature and gas composition are encountered.
None the less it has been shown that SSiC can resist flue gases of aluminum
remelting furnaces for up to 1000h without much change in strength [91], although
this is certainly not true for all plants of this type [92]. Similarly nitrogen-based gases
of the system N2-H2-C0 did not decrease the strength of SSiC, while endothermic
gases had an adverse effect (-20-50%) at 1200-1300°C [93,94] in long-term
experiments. Typically B-doped Sic behaved better then Al-doped or those with
free Si. A good resistance of Sic-based materials to SO2 or H2S has also been
reported [95].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 161
I " 4
*
* *
* ssic
A powder&SC
A CVD
boron oxide, and 0.25-0.7% free boron. Dense and additive-free B4C specimens did
not interact with 10% sulfuric acid either at room temperature or on boiling. A
small mass loss (<O. 1 mg cm-2) in the initial period of boiling can be explained by
surface impurities. As has been shown [98],the corrosion resistance of boron carbide
powder in water, mineral acids and alkalis is largely dependent on its purity and the
ordering of the crystalline structure. The studies on the electrochemical behavior of
hot-pressed boron carbide in H2SO4 and NaOH solutions [99] have demonstrated
that the corrosion resistance of B4C depends to a large extent on the content of
additives and impurities.
However, during heating, strong oxidizing acids and oxidant mixtures such as
+
Cr03 H2SO4, KIO3 + +
H3P04, K2Cr207 KIO3, Cr03 + H2SO4 + H3P04,
+ + +
KMn04 H2SO4, H2SO4 HC104 K2Cr207[98] oxidize free carbon present
in many boron carbide materials. To remove free carbon from B4C, alkaline
solutions of hydrogen peroxide, bromine, and others were also used [98].
5.6.2.2.4 Corrosion by Hot Gases Boron carbide is mainly used at relatively low
temperatures [96], therefore its oxidation has not been studied as thoroughly as
that of the Si3N4and Sic ceramics. There are data on the oxidation of B4C powders
[lo21 and several papers on oxidation of boron carbide materials produced by
different methods [103-1071.
In [ 1081 different reactions in the system B4C-02 were calculated thermodynami-
cally at oxygen pressures of 105 and 1.3x lop3kPa. It was shown that the oxidation
of B4C may result both in condensed (B203,B, C) and gaseous (CO, COz, B202,
BO, BOz) products. At high oxygen partial pressures the reaction
B4C + 4 0 2 ++ 2 B 2 0 3+ C02 (24)
should dominate. This reaction was confirmed by the results of experimental
investigations on powders and sintered specimens [ 1091.
Oxidation of boron carbide powders starts at about 600°C. Studies on hot-pressed
B4C containing iron, aluminum, magnesium, calcium, titanium, silicon, and other
impurities at a level of up to 4% and oxidized in air [lo41 demonstrated that upon
nonisothermal heating of specimens exothermal effects due to oxidation can be
registered above 600°C. A detectable mass gain starts above 700°C. At 1000°C the
oxidation process intensifies and above 1200°C the material starts oxidizing consider-
ably, with a high thermal effect. At temperatures above 12OO0C,the specimen weight
decreases due to the B203vaporization, the rate of which increases significantly at this
temperature [ 1091. The X-ray diffraction (XRD) analysis of oxidized specimens did
not reveal any solid oxidation products other than B203.This is in agreement with
the thermodynamic calculations [98] and the experimental investigations showing
that reaction (24) dominates B4C oxidation.
The kinetic curves obtained under isothermal conditions (Fig. 13) follow the
paralinear law (Fig. 7) with competing B4C oxidation and B 2 0 3 vaporization.
3, 1
.
N
6 1
F o
@
E -10
-20
-300 < 1 2 3 4 5
t, h
Figure 13. Oxidation kinetics of hot-pressed B4C at: (1) 800°C; (2) 900°C; (3) 1000°C;(4) 1100°C;
and (5) 1250°C; reproduced from [2].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
164 5 Corrosion of Hard Materials
Thus, the oxidation kinetics due to the boron oxide vaporization at temperatures
above 1200°C result in weight loss of specimens (Fig. 13), which becomes linear
in time after the first 30min of oxidation. On oxidation at temperatures below
1100°C a liquid boron oxide layer (T, of amorphous B2O3 is 450°C [llO])
uniformly covers the specimen surface.
Under the conditions of active B203 vaporization the diffusion of impurities
apparently does not exert great influence on the oxidation process, unlike in the
case of oxidation of silica formers. One may suppose that up to 1200°C the
oxidation rate of B4C is controlled by the oxygen diffusion through the B203
layer.
Differencesin the thermal expansion coefficients of B203 and B4C lead to crack-
ing of the oxide layer after cooling from high temperatures. The XRD analysis
revealed B203 and H3B03 in the surface layer of oxidized specimens. However,
H3B03 is not formed during the oxidation of materials, but during the exposure
of specimens to air at room temperature for several days due to the hydration of
boron oxide according to reaction (23). The oxidized specimens kept in air were
covered with a white film due to the hydration of boron oxide.
It is important to note that the specimens of ceramics oxidized at 1400°C had a
graphite layer of up to 30 pm thickness under the B203film [107]. If the oxidation
process was accompanied by friction, the graphitized layer was formed at much
lower temperatures. The formation of free carbon was also observed upon oxidation
of B4C-Sic composites [l 111.
Active oxidation of B4C at a low partial oxygen pressure leads to the formation
of gaseous boron oxides and removal of them from the surface of the specimen.
Similar to other materials, porosity increases the reactive surface of boron carbide
specimens and weight gain on oxidation [1061.
All the above primarily refers to materials containing relatively small amounts of
impurities. However, boron carbide ceramics usually contain 7-20% Al or A1203to
increase hardness and 6-1 3% silicon to improve mechanical characteristics [98].
Aluminum- and silicon-doped materials often exhibit structural nonuniformity.
They have areas of pure boron carbide, and silicon- and aluminum-doped areas
[107].
In contrast to the oxidation of the additive-free material, weight gain instead of
its loss is observed above 1200"C, that is the rate of oxide formation is higher than
the rate of its vaporization. Electron microscopy investigations of specimen sur-
faces of Si- and Al-doped materials heated to 1430°C demonstrated glassy oxide
layers containing boron, oxygen, silicon, and aluminum. Thus, the oxide layer is
a borosilicate glass. Crystallization of aluminum borates forming needle-like
crystals also has been found in our work. The dissolution of silica and alumina
in boron oxide retards the rate of its vaporization. In general, addition of silicon
and aluminum to boron carbide facilitates the formation of borosilicate glass on
oxidation and decreases the rate of boron oxide vaporization, thus increasing
the oxidation resistance of the material [ll 11 [2]. However, to form a continuous
protective layer on the specimen surface, it is necessary to provide a uniform
distribution of additives in the material excluding the existence of pure boron
carbide areas.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 165
B4C reacts with C 0 2 to yield B203 and CO or free carbon [96]. Boron carbide
neither interacts with sulfur and phosphorus vapors, nor with nitrogen up to
1200°C. BN can be formed upon reaction with nitrogen at higher temperatures,
or when ammonia is added. With chlorine it reacts above 1000°C to form BC13
and graphite. Bromine and iodine do not react with B4C [98].
5.6.3 Nitrides
5.6.3.1 Silicon Nitride
In many cases the references quoted for Sic can be used to access corrosion data on
Si3N4-basedmaterials [2,3, 5, 35, 53, 54, 116, 1171. Si3N4ceramics are much more
complicated in their behavior because sintered types come with a significant amount
of grain-boundary phase, which determines the overall behavior.
+
Si3N4+ 6 H 2 0 H 3 Si02 4 NH3 (25)
starts at temperatures below 200°C [61], probably at about 130°C [121]. At 500°C
the powder is completely oxidized within a few hours.
Reaction (25) would suggest a passivating behavior but the solubility of silica is
favored even at low temperatures in the alkaline water present due to the dissolution
of NH3. Hence basic reaction kinetics are linear, it is a form of active corrosion.
The grain-boundary phase of Si3N4-ceramicsmay be even stronger affected by
hydrothermal corrosion. This leads to pit formation at conditions of 300°C and
8.6MPa [122]. The shape and depths of pit formation depend on the additive
type. It is interesting to note that a hot isostatically pressed (HIP) Si3N4 ‘without
sintering additives’ was attacked more strongly then those with additives, because
the grain boundaries with silica from the oxygen content of the powder were leached
and resulted in grain dislodgment [ 1231.
It should be noted that hydrothermal corrosion should not be confused with cor-
rosion by wet atmospheres and that the behavior changes at higher temperatures
when N2 + H2 is the product gas rather then NH3 [60].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 167
5.6.3.1.3 Corrosion by Metals Reactions of Si3N4 with metals and alloys are a
complex issue, because the atmosphere both for oxygen and nitrogen partial
pressures plays a role and silicides, nitrides, or ternary compounds can form. A
good compilation of the partly conflicting observations has been provided [1161.
For the interaction of the related SiAlONs and Fe-based alloys see published
work [124].
Predictions for reactions can be made on the basis of Ellingham-type diagrams. In
noble gas atmospheres or under vacuum, Si3N4reacts readily with most transition
elements below 1000°C and with many other metals at varying temperatures.
Whether this is true for a given Si3N4 ceramic under application conditions
depends on many parameters like a silica layer protection, diffusion coefficients,
wetting behavior etc. The reader is here referred to the reference mentioned above
[116].
5.6.3.1.4 Corrosion by Salt Melts Hot corrosion of Si3N4 has been extensively
studied and reviewed by researchers at NASA labs [67,68,77, 125, 1261. The basic
attack is that on the protective silica layer, which makes the behavior similar to
that of Sic and basic mechanisms can be taken from the chapter on S i c above.
Recent investigations [74] gave more insight into the details of the kinetics, which
is shown in Fig. 14. The surface tension differences between sulfate and silicate melts
result in a fractal pattern of hillocks of silicate melt beneath the sulfate, hereby
exposing unaltered silica with a strong oxygen potential gradient to the sulfate.
On consumption of this silica the decomposition of the sulfate is enhanced up to
a gas bubble formation. The exposed areas are the sites for stronger corrosion,
which can ultimately produce pits. However, the process does slow down, when
the sulfate is used up or a critical thickness of silicate layer is reached.
There is a strong indication that the grain-boundary phases modify the develop-
ment. Wide shallow pits and/or exposure of individual Si3N4-grainsare observed
[127,128], and Na advances into the grain-boundary phase.
The resulting degradation of mechanical properties is of similar order as for S i c
(20-50%) [26,127,129]. However, it seems that the degradation is not linear with
time but levels out with time [129]. Furthermore, in pitted samples there is an
increase in the Weibull modulus, with the result that the failure stress at low failure
probabilities is decreasing much less than the average strength [ 1301.
Taking into account the higher level of initial strength of Si3N4-basedceramics
the remaining strength after hot corrosion is usually higher compared to SIC.
5.6.3.1.5 Corrosion by Hot Gases Active corrosion has usually an even more
pronounced effect on strength, compared to Sic, as the phase between the grains
may vanish leading to grain dislodgment and inhomogeneous internal corrosion
[ 1311. All gases which produce high partial pressures of Si-bearing species (namely
halogens) are therefore dangerous, particular in low-oxygen environments [ 132-
1341.
In mixed and oxidizing atmospheres Si3N4 has a superior resistance to C1-
containing environments relative to Sic materials [ 135-1 371. A small water content
in H2S-H2 mixtures is also capable of producing a silica scale, which slows down
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 14. Schematic model for the hot corrosion kinetics of silica-formers after [74].
Ess
SiO,
3J
(not watersoluble)
Na,O * @i02
-
5 Corrosion of Hard Materials
immiscible at 9OOOC
1
(watersoluble)
n
of the Na-sulfate
Na,SO,
with amounts of
168
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 169
corrosion [ 1341. As for Sic the plain influence of a water vapor content in oxidizing
environments is probably moderate [ 1381 unless it transports impurities.
The active(1)-to-passive transition has been discussed with some controversy but
most models [14,139,140] and experimental data [141-1431 point to very similar
conditions as outlined in Fig. 11 for Sic.
The passive oxidation of pure Si3N4has been intensively studied. Thermodynamic
data suggest that a two-step oxidation
2 Si3N4+ 1; O2 H 3 Si2N20 N2 + (26)
and
Si2N20 14 O2 + 2 SiOz+ + N2 (27)
and accordingly a duplex layer in a oxygen potential gradient, developing in a scale,
is favored. However, the discussion of the correct kinetics, the presence or absence
of a sublayer of Si2N20,the activation energy, and other issues are still not settled
[14,144-1511. Extremely low oxidation rates up to high temperatures are, however,
clear.
This contrasts strongly with the behavior of sintered Si3N4. In Fig. 15 we have
calculated tc for pure and sintered qualities, assuming parabolic growth and using
the data compiled earlier [53].
-1
0
... 1 -0
0
.
::'"""": .: ,
~
0 .
0 .
-21
HPISSN
, I , I 0 , , ,,,
-3
800 1000 1200 1400 1600 1800
900 1100 1300 1500 1700
Temperature ("C)
Figure 15. Calculated critical times (years) t, for 0.1 mm homogeneous recession by parabolic silica
scale growth on Si3N4-basedmaterials using data from the collection of [53].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
170 5 Corrosion of Hard Materials
A number of things can be learned from Fig. 15. First, sintered Si3N4 never
reaches the level of oxidation resistance of pure Si3N4. Typically a temperature
difference of 250°C exists to have comparable behavior for even the best qualities.
Second, at low to moderate temperatures pure Si3N4oxidizes much slower then
S i c (Fig. 12), but sintered Si3N4is comparable to Sic. Third, there is a very wide
scattering of results for sintered Si3N4.And last, but not least, at about 1400°C
there is a rapid breakdown in log t, to values below and much below zero, indicating
very rapid oxidation of sintered Si3N4.
The reasons for these features are manifold and again related to the presence of
grain-boundary phases. We often have a very complex passive oxidation behavior
with leaching of grain-boundary constituents to the scale, crystallization, internal
oxidation, bubble formation, scale cracking, and other features. All these
phenomena are different for different systems and vary with the amount and
exact composition of the additives, For detailed information the reader is referred
to the textbooks and proceedings already mentioned and some reviews, which cite
papers from a vast amount of literature [140,152, 1531.
We have to point out some facts, which should make the reader cautious in inter-
preting Fig. 14 too strictly. At low temperatures some Si3N4 types (sintered with
pure Y or rare earth additives) contained secondary phases, which oxidize rapidly
and with large volume increases to yield the oxide phase. Since at low temperatures
(<lOOO°C) the outer scale becomes ridged, it leads to cracking of the specimen [154]
and breakaway or pseudolinear behavior. This is part of the reason why there is such
a large spread in data from Si3N4ceramics even at low temperatures.
Particularly at T > 1300"C, we commonly have deviations from the parabolic law
and all the forms described in Sections 4.3.2 and 4.4.3 can be observed. Since the
extrapolation to t, is done on the basis of a parabolic approximation it will not
give very reliable t,.
Perhaps the most disturbing characteristic of sintered Si3N4 oxidation is the
appearance of internal oxidation accompanied by outer scale growth. Leaching of
grain boundary constituents into the scale (Fig. 16) is observed for Y, Al, Mg,
REE, and others to varying degrees. The leaching is connected to internal oxidation
and is for some cases the rate limiting step [156]. In the leached zone, which extends
into the sample with the same order of magnitude as the outer scale thickness, we
often find increased porosity and hence decreased hardness [155]. At least this
zone must be incorporated into the value for penetration but can not be gained
from measuring layer thickness or mass change.
There is furthermore recent evidence from transmission electron microscopy
(TEM) studies for the dissolution of matrix grains in the grain-boundary phase
during oxidation at high temperatures, which widens the grain boundary films as
a function of depth into the material [157,158]. This type of penetration will be
even more difficult to detect and quantify.
Therefore, for most Si3N4 ceramics it will be necessary to have mechanical
property measurements along with corrosion to evaluate their applicability. None
the less there are rules of thumb, which may help in materials selection. The most
important one is that the level at which oxidative corrosion becomes the prime
problem, is connected to the eutectic of the system sintering additive(s) + Si02.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.6 Materials 171
Figure 16. EDX map for Y and SEM picture of SSN oxidized at 1400°Cfor 990 min [ 1551, showing
a leaching zone beneath the scale.
This may well apply also to the pure Si3N4 qualities, because here the eutectic
Si3N4-Si02 in the vicinity of 1600°C may be the ultimate temperature for long-
term use [1591.
The challenge for sintered Si3N4to be applicable at T > 1400°C has found first
success in studies which show that, for example, additions of MoSi, and surface
treatments are capable of enhancing the oxidation resistance [ 160, 1611.
5.6.3.2.1 Corrosion by Liquid Media The resistance of pure boron nitride to attack
by diluted acids is lower than that to attack by concentrated ones and decreases in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
172 5 Corrosion of Hard Materials
a
E -1 -
-2 -
I I
3O 1 2
t, h
Figure 17. Oxidation kinetics of wurzite BN Hexanite-R at: ( I ) 900°C; (2) 1000°C; (3) 1IOo"C, and
(4) 1200°C; reproduced from [2].
5.6.4 Carbonitrides
5.6.4.1 Silico-carbonitrides
The few investigations available for the new precursor-derived silico-carbonitrides
do not permit their corrosion behavior to be assessed in detail. Most studies of
interest here deal with thermal stability only.
The issue of thermal stability is important for these phases, because precursor-
derived ceramics are amorphous and have stoichiometries where detrimental
structural and chemical changes may occur on crystallization. The highest
temperatures recorded so far come from phases in the system Si-B-C-N
(2000°C [166]).
For the carbonitrides, the stoichiometry of the amorphous product is
Si3+.TN4Cs+p [167] and thus in terms of possible crystallization products there is
always C in addition to Si3N4 and Sic. In precursor-derived Si-C-N-ceramics,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
174 5 Corrosion of Hard Materials
wide stoichiometry range and almost never exist in the stoichiometric composition.
Due to their structural identity, they also form solid solutions in the Ti-C-N system.
Oxygen can be dissolved in the material until saturation of carbon or nitrogen
vacancies [179] with no changes in the f.c.c. (NaC1) crystal structure. In all cases,
whether we consider TIN, Tic, and TiC,N, materials, the oxidation behavior is
similar and these rutile-formers will be analyzed together.
The peculiarities of the defect structure of TIC and TiN allow for oxygen dissolu-
tion on the first stage of oxidation. Titanium oxides appear later, when the material
structure is saturated with oxygen. Anatase was observed at lower temperatures
and/or short oxidation times ( 5 min). With increasing time and temperature, rutile
is formed as the only oxidation product [108].
Oxidation of TIN powders [180], thin films [177], and dense bodies [I081 was
extensively investigated. It was shown that oxidation of TiN, following a logarith-
mic law, starts at room temperature. Above 500°C the oxidation rate increases and
the kinetics follow a parabolic law. Oxidation of a TiN powder prepared by plasma
vapor phase synthesis and having the specific surface area of 46 m2 g-' is noticeable
above e2250"C and results in the formation of Ti02 as anatase and brookite at lower
temperatures, and as rutile at higher temperatures [181].
The scales formed at oxidation of sintered TIN in air can be divided into two
major types with the following layer systems: At low temperatures ( 5800°C)
TiN-dense Ti02-air; at high temperatures (1000°C) TiN-porous Ti02-dense
Ti02-air.
The latter can develop into a multilayered scale during the long-term oxidation
above 1000°C [181]. Growth of such rutile scales was based on the simultaneous
diffusion of titanium and oxygen ions through the oxide layer in opposite directions.
We can suppose that the fine grain inner layer grows by inward diffusion of
oxygen, while the coarse grain outer layer grows by outward cation diffusion.
Oxidation of TIC is similar to that of TIN. Many carbides of transition metals
form carbon at the carbide-oxide interface. In the case of TIC, no carbon was
formed [182]. Oxidation of titanium and carbon at 1000°C occurs with the same
rate [183].
Both TiN and TIC are widely used in ceramic matrix composites for improving
electrical conductivity and mechanical properties [ 178, 184,1851. The formation of
a rutile scale was observed on the surfaces. The growth of such a scale should be
related to the diffusion of titanium to the surface of the composite and its oxidation
according to the reaction:
2TiN + 2 0 2 = 2Ti02 + N2. (36)
The protective properties of rutile are fairly low compared to those of alumina or
silica. Thus, the presence of TiN or TIC limits the high-temperature applications
of the above composites. At small amounts and particle size of TIN in silicon nitride
ceramics, a continuous silicate film can be formed, covering TIN particles and
protecting them from further oxidation [ 1781.
Thermodynamic analysis of the chemical equilibria in water-carbide systems
[loo] indicated that carbon, corresponding oxides, CH4, C 0 2 , CO, and H2 are the
main products of hydrothermal corrosion of metal carbides in the temperature
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
176 5 Corrosion of Hard Materials
range 300-1000°C. There exists a possibility for selective etching of T i c and carbon
layers formation. A low H 2 0 to carbide molar ratio results in the formation of
carbon layers. Lower pressure is beneficial here. A high H 2 0 to carbide molar
ratio means that under equilibrium conditions carbon can be oxidized to CO and
C02. Experimental studies showed that no protective oxide layer was formed
upon hydrothermal corrosion of carbides of transition metals. Raman spectra
obtained from the hydrothermally treated TIC powders demonstrated the presence
of graphite. Anatase crystals were also observed on the surface of TIC. With grow-
ing temperature, anatase was partly transformed to rutile [loo]. This is in agreement
with other published data on corrosion of TIC and TiN [186].
more significant and the oxidation kinetics can be described by a parabolic law with
an apparent activation energy of 472 kJ mol-' .
References
1. R. J. Fordham, D. J. Baxter, and T. Graziani (Eds), Corrosion of advanced ceramics. Key Eng.
Muter. 1995, 113, 256.
2. Y . G. Gogotsi and V. A. Lavrenko, Corrosion o f High-Performance Ceramics, Springer
Verlag, Berlin, 1992, p. 181.
3. K. G. Nickel (Ed.), Corrosion of Advanced Ceramics Measurement and Modelling. NATO
-
ment and Modelling, K. G. Nickel (Ed.), Kluwer, Dordrecht, 1994, pp. 205222.
69. H.J. Grabke, in Corrosion of Advanced Ceramics - Measurement and Modelling, K. G. Nickel
(Ed.), Kluwer, Dordrecht, 1994, pp. 223-234.
70. F. S. Pettit, G. H. Meier, and J. R. Blachere, in Corrosion of Advanced Ceramics - Measure-
ment and Modelling, K. G. Nickel (Ed.), Kluwer, Dordrecht, 1994, pp. 235-248.
71. N.S . Jacobson, Oxid. Metals, 1989, 31, 91-103.
72. N. S.Jacobson, J. Am. Ceram. SOC.1985,68,432-439.
73. C. Berthold and K. G. Nickel, Key Eng. Muter. 1997, 132-136, 1588-1591.
74. C. Berthold and K. G. Nickel, J . Europ. Ceram. SOC.1998, in press.
75. J. Schlichting, Werkstof Korros. 1975, 26, 753-758.
76. N. S. Jacobson, C. A. Stearns, and J. L. Smialek, Adv. Ceram. Muter. 1986, 1, 154-161.
77. D. S. Fox, N. S. Jacobson, and J. L. Smialek, Hot Corrosion of Ceramic Engine Materials,
Report DOE/NASA/50111-2, DOE/NASA, Cleveland, 1988.
78. D. S. Fox and J. L. Smialek, J . Am. Ceram. Soc. 1990, 73, 303-31 1.
79. K. N. Lee e f al., 1995.
80. K. N. Lee and R. A. Miller, J . Am. Ceram. SOC.1996, 79, 62G-626.
81. S. Y.Ip, M. J. McNallan, and D. S. Park, J . Am. Ceram. SOC.1992, 75, 1942-1948.
82. G. S. Fischman, A. Zangvil, and S. D. Brown, Adv. Ceram. Muter. 1988, 3, 66-70.
83. G. W. Hallum and T. P. Herbell, Adv. Ceram. Muter. 1988, 3, 171-175.
84. G. A. Schneider, K. G. Nickel, and G. Petzow, in Physics and Chemistry of Carbides, Nitrides
and Borides, NATO ASI Series E: 185, R. Freer (Ed.), Kluwer, Dordrecht, 1990, pp. 387401.
85. E. Fitzer and R. Ebi, in Silicon Carbide, K. L. Marshall, T. W. Faust, and C. E. Ryan (Eds),
South Carolina University Press, Columbia, SC, 1974, pp. 320-328.
86. K. Schwab, Korrosion poroser Sic-Keramiken unter Rujjilterbedingungen. Dissertation, Uni-
versity Tiibingen Mineralogy, 1996, p. 170.
87. T. Narushima et al., J . Am. Ceram. SOC.1993, 76, 2521-2524.
88. L. U. J. T. Ogbuji and E. J. Opila, J . Electrochem. SOC.1995, 142, 925-930.
89. J. A. Costello and R.E. Tressler, Ceram. Int. 1985, 11, 3944.
90. J. W. Fergus and W. L. Worell, in Corrosion and Corrosive Degradation of Ceramics, R. E.
Tressler and M. McNallan (Eds), Martinus Nijhoff, Boston, 1990, pp. 43-51.
91. R. M. Persoons, J. Luyten, and A. D. Stalios, in Corrosion of Advanced Ceramics, K. G. Nickel
(Ed.), Kluwer, Dordrecht, 1994, pp. 297-306.
92. M. McNallan, M. Van Roode, and J. R. Price, in Corrosion and Corrosive Degradation of
Ceramics, R. E. Tressler and M. McNallan (Eds), Martinus Nijhoff, Boston, 1990, pp. 445-
467.
93. D. P. Butt, R. E. Tressler, and K. E. Spear, J. Am. Ceram. Soc. 1992, 75, 3257-3267.
94. D. P. Butt, R. E. Tressler, and K. E. Spear, J. Am. Ceram. Soc. 1992, 75, 3268-3277.
95. H.Nickel and R. Forthmann, in Corrosion of Advanced Ceramics, K. G . Nickel (Ed.), Kluwer,
Dordrecht, 1994, pp. 99-1 16.
96. P. S. Kisly et al., Naukova Dumka, Kiev, 1988.
97. G. I. Postogvard, Y. G. Parkhomenko, and I. T. Ostapenko, Poroshkovaya Metall. 1983,8,
pp. 28-30.
98. T. Y.Kosolapova (Ed.), Nonmetallic Refractory Compounds, Metallurgiya, Moscow, 1985.
99. A. M.Pivak et at., in New Technological Processes in Powder Metallurgy, I. D. Radomyselsky
(Ed.), Institute for Problems of Materials Science, Kiev, 1986, pp. 31-35.
100. N. S. Jacobson, Y. G. Gogotsi, and M. Yoshimura, J . Muter. Chem. 1995.
101. F. Thevenot, in Proc. 1st ECERS, G. De With, R. A. Terpstra, and R. Metselaar (Eds),
Elsevier, Maastricht, 1989, pp. 2.1-2.25.
102. Y. L. Krutsky, G. V. Galevsky, and A. A. Kornilov, Poroshkovaya Metall. 1983,2, pp. 47-50.
103. V. A. Lavrenko et at., Oxid. Metals, 1976, 10, 85-95.
104. V.A. Lavrenko, Y. G. Gogotsi, and I. N. Frantsevich, Dokl. Akad. Nauk SSSR, 1984,27,114-
117.
105. V. A. Lavrenko and Y. G. Gogotsi, Oxid. Metals, 1988, 29, 193-202.
106. L. N. Yefimenko, E. V. Lifshits, and I. T. Ostapenko, Poroshkovaya Metall. 1987, 4,56-60.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
180 5 Corrosion of Hard Materials
190. A. Lebugle and G. Montel, Rev. Int. Hautes Temp. Refract. 1974, 11, 231.
191. R. J. Irving and I. G. Worsley, J . Less-Common Metals, 1968, 16, 103.
192. A. Munster and G. Schlamp, 2. P h p . Chem. 1960,25, 116.
193. V. B. Voitovich, V. A. Lavrenko, and A. M. Adejev, Oxid. Metals, 1994, 42, 145-161.
194. V. Schneider et al., in Corrosion of Advanced Ceramics, R. J. Fordham, D. J. Baxter, and
T. Graziani (Eds), TransTech, 1996, pp. 49-58.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
6.1 Introduction
This contribution describes the most important influences on the hardness of
materials from the viewpoints of the measuring approach and of probable chances
for further progress in the development of new materials. Therefore, the discussion
will be focused on questions that arise when precise measurement has to be achieved
in the investigation of hard materials.
- Which testing load should be applied? Is there a real size effect to be
expected from basic properties of the material (or is it an artifact of the measure-
ment?)?
- In investigations of new hard materials, the size of available samples is often
small, and many people prefer microhardness approaches. Can grain size effects
be recorded independent of the applied load?
- How does the hardness depend on the surface preparation? Do ground ceramics
(which are important in technical use) exhibit the same hardness as polished
samples (often preferred for tests in development procedures)?
- In ceramics, the dependence of the strength on flaws is a well-known
phenomenon associated with stress concentrations that cannot relax by plastic
deformation in these brittle materials. Does hardness testing and the development
of sintered materials with an increased hardness also have to consider flaw
populations?
- Is the intensively investigated activity of lattice processes inside the grains
ranges? Is there a different impact of the microstructure on the hardness at, for
example, 400°C or at 1400°C? Which elements of the microstructure are most
important for obtaining a high hardness in the different ranges of the
temperature?
It is not intended here to present new evidence for special micromechanisms.
Instead, the available knowledge about such processes is used to answer the
questions outlined above.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
184 6 Interrelations Between Influences
Figure 1. The Vickers hardness (testing load 1 kg) of fine-grained sintered alumina ceramics
prepared by a sol/gel approach starting with boehmite (y-seeding).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.3 Theoretical Considerations 185
are so small that they are invisible in the optical equipment of hardness testers,
clusters of pores, microcracks, and other defects located within the subregion of
the material affected by the indentation, all promote the penetration of an indenting
body in one way or another.
Therefore, the development of new hard materials requires the preparation of
phases with small microplastic deformability of the crystal lattice, and reduced
grain sizes in the sintered polycrystals, and highly perfect microstructures. For
example, sintered alumina ceramics which have a submicrometer microstructure
and associate this small grain size with a macrohardness of 20-24 GPa (compared
with 17-19 GPa in conventionally sintered corundum) were obtained only
when new processing approaches became available, which could reduce the flaw
population [ 11.
Table 1. Comparison of critical stresses for microplastic deformation of the crystal lattice and for
grain boundary fracture in AI2O3at room temperature [6].
I I
I---
grain boundary contrit,,,,. 1Id.nr :tri I?51 7e 6 necr
II
-
-
% -
,
Figure 2. The limitation of the free space available for dislocation or twin activity by the spatial size
of the plastic zone in a single crystal (upper curve/upper x-axis), and by the grain size of polycrystal-
line microstructures (lower curve/lower x-axis) [6].
Under this condition, the size effect (load effect) as measured in single crystals
simulates the grain size influence of the hardness in polycrystalline materials.
In the mean time, this simulation of the grain size influence in the hardness of poly-
crystals by the indentation size effect (load influence) in a single crystal applies under a
very interesting condition: it excludes, in the single crystal, all possible contributions
from grain boundaries. Therefore, the dzflerence between the plastic zone size effect
(load effect) measured in a single crystal (e.g. in sapphire) and the real grain size
effect observed in the sintered polycrystals (e.g. sintered corundum) gives a measure
that characterizes the amount of grain boundary contributions to the deformation
that governs the hardness at room temperature (experimental data c.p. ##6.5, Fig. 8).
To be able to compare grain size and indentation size influences, the size of the
plastically deformed zone must be known. This information comes from TEM
investigations where an approximately constant ratio 2RP1/2aM 4-5 between the
plastic zone size 2RP1and the length of the Vickers diagonals 2a can be derived for
single as well as polycrystalline ceramics and for hard materials with fundamentally
different bonding like (ionic) alumina [2] and (covalent) S i c [8].
Together with the measurements of the ISE in sapphire and of the grain size
dependence of the hardness in available sintered microstructures with grain sizes
20.2 l m , the theoretical extrapolation will give a forecast of the probable hardness of
nanostructured alumina ceramics. The model also explains the observed influence of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
188 6 Interrelations Between Influences
the grain size on the measured load effect and demonstrates why reliable information
about the grain size dependence of the hardness requires the use of larger testing loads.
Then, on increasing the load further, fundamental principles of solid state physics
suggest that the further extension of the plastically deformed zone in a single crystal
should occur much more easily by growth or by multiplication of pre-existing
elements of lattice deformation (dislocations, twins) than by the more difficult
generation of new dislocations in the virgin material around. Hence, on increasing
the load (the size of the indent), the microplastic deformability increases (the hard-
ness decreases) because the more easy growth of pre-existing dislocations becomes
more important and the role of the (difficult) generation of new dislocations
decreases.
To simplify the mathematical procedure, a characteristic area A e (2Rp1)2is
introduced. At some load PI, the microplastic deformability S can be defined as
an inverse local compressive yield stress:
S = (dP/dA)-' (la)
the smaller the increase in the load required to produce some incremental growth
dA = A - A l of the plastic zone, the larger is the deformability.
With different dislocation structures in the virgin ceramic and generated during
indentation as outlined above, the deformability may change with proceeding
penetration. Without hypothesizing any mechanism or specific character of this
change, two different symbols Si and 6,distinguish the following situations.
For the preceding penetration at loads P < P I , the average deformability Si is
simply expressed by
Note that Si characterizes an initial deformability only when PI is very small, and
that 6i gives dzflerent averages for increasing values of P I . With Eq. (la), the latter
integral equals Jt' dA = A , , and the expression for Si integrates to
Si = AI/P1. (1b)
On the other hand, with Eq. (la) the deformability 6,for the further extension of
the plastic zone at PI is
6, PI, A , ) (dP/dA)ibp,,A=Al,
= (lc)
where the considerations discussed above suggest 6, 2 Si.
~ = testing load, a = 1.8544 for the
In the hardness definition H = a P / ( 2 ~ 2 )( P
Vickers pyramid), 2a can be replaced by the size of the plastically deformed zone
2Rp1as far as the ratio n = 2Rp1/(2a)is approximately constant. With the character-
istic area A = (2Rp1)2,the hardness becomes
H = o'P/A, Pa)
where the parameter a' = n 2 0 depends on the value of n = RpI/a. In the associated
differential equation
dH -dP dA
H P A
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
190 6 Interrelations Between Influences
dA/A can be substituted by means of Eq. (lb) by an expression that depends on the
load P only:
dA--
-
A
-dA
A1
- Se(P1)dP
A1 = (;)p=p,(E)
=
(Se/Si> d P
P (2c)
+
For finite load ranges P I -+ PI AP, and Eq. (2b) can be integrated if a slowly
varying ratio Se/Si is regarded as a small perturbation within this load interval.
The result is
with H , = H(P,) and a scaling parameter P, that generally can be chosen arbi-
trarily. The choice of a large P,, however, offers the advantage that H , acquires
the character of an asymptotic (constant) level H , = Hasympt (P 4 00).
It is only here that the knowledge about the processes of microplastic deformation
becomes physically important for the model: if the deformability 6,for the further
extension of the plastic zone is larger than the average deformability Si for the
preceding penetration at any PI, the ratio Se/Si > 1 attributes a specific character
to the load influence in Eq. 3, and the hardness decreases with increasing load in
a way that agrees qualitatively with the measured data obtained for a Czochralski
grown sapphire crystal in the load range 3-100 N (Fig. 3 , interrupted line).
To obtain reliable data which are really characteristic for the single crystalline
material (independent of special features associated with one or another orientation),
I I I I I I I 1
1.o 2.0 4.0 7.0 10.0 20.0 40.0 70.0 100.0
indenter load (N)
Figure 3. Conventionally me2sured Vickers hardness on sapphire ((1510) surface, Vickers diagonals
parallel with the traces of (1012) and (1102) p 1 a n e s ) m T h e interrupted line is the fit accord-
ing to Eq. (3) with a constant (average) ratio (Se/Si) resulting in H , = 15.40GPa and
(Se/S,) = 1.119. An only slightly different fit results with the varying (S,/Si) of Eq. (3a)
( H , = 16.20GPa, ( & / c ? ~=) ~1.115).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.4 Influences of the Gruin Size and the State of the Surface 191
with = (he/Si)p+o. Since (S,/Si) as given by Eq. (3a) is a very slowly varying
function, fits according to (3a) or using an average (constant) value show only small
differences in practically interesting load ranges (Fig. 3), and it is obvious that use
and the form of Eq. (3a) has little practical meaning for the derived load effect
over the experimentally investigated range.
The expressions (3) and (3a) give a load influence that compares well with
measurements in single crystalline A120, (Fig. 3). However, it should be emphasized
again that with the phenomenological character of the model this agreement is not
an evidence for some specific micromechanism, it just means that there is no general
discrepancy between the dislocation idea and the observed indentation size effect.
30 Sintered Al 203
with different grain sizes
Sapphire
10 : I I I I I
1.o 3.0 6.0 10.0 30.0 60.0 100.0
indenter load (N)
Figure 4. Load dependence of the Vickers hardness of sapphire (same data as in Fig. 3) and of fine-
>
grained sintered alumina ceramics (relative density 99%, purity > 99.7% A120,) [16].
Table 2. Parameters of a least squares fit (Eqs (3) and (3a)) for the load effect in the hardness of
alumina ceramics with different grain sizes (Fig. 4).
6.4.2 The Effect of the Grain Size and the Surface State in
Ceramics when Recorded by Different Measuring Approaches
Figure 4 clearly demonstrates the reason of the declining grain size effect at decreas-
ing testing loads (smaller indentation size). A more conventional presentation is
given in Fig. 5. A similar grain size influence is recorded by Vickers indents at
10 kg (HVlO) and 1 kg (HV1; HV1 > HVlO due to the ISE), and close results are
obtained by HV1 and by a Knoop approach at 100 g (HKO. 1) because the longer
of the two Knoop diagonals exhibits a similar length at a load of 100 g to the Vickers
diagonal size at HVl. With further decreasing sizes of the plastic zones in the
0 0
22 -
-
0-
a
m s9!
g20- 4
m
m
al
C
T?
2 18-
-
2 HKO.l
.
t - - - ~
al
0
Y *--- HV1
5 HVlO
polished surfaces
high-purity Al (no doping additives)
14
0.1 0.2 0.4 0.6 0.8 1.0 2.0 4.0
D (w)
Figure 5. Investigations of the grain size effect in alumina ceramics by using different approaches [6].
Each data point indicates the average value of 6-15 indents.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
194 6 Interrelations Between Influences
24-1 ground
22 -
polished
i'
420-
0
2
$18-
:
C
16-
14 I I I I I 1 1 1 1 I I )
Figure 6. Influence of the surface preparation on the hardness of sintered alumina ceramics with
different grain sizes. In the submicrometer range, grinding results in a higher hardness than
measured on polished surfaces.
Vickers microhardness (400 g, HV0.4), however, the grain size effect becomes
smaller, and a similar hardness is derived for the single crystal (sapphire) as for
an even extremely fine-grained 0.2 pm polycrystal.
Figures 6 and 7 show that the hardness of ground surfaces may be higher than
measured on polished surfaces (unfortunately, no detailed investigation of this
effect is known in ceramics). The reason for this seems to be similar to the grain
size influence. Whereas no significant difference exists in alumina ceramics with
grain sizes >1 pm, an increasing advantage of the ground state is observed for
fine-grained materials with HV10 > 20 GPa (submicrometer A1203, A1203-
Ti(C, 0,N) composites). TEM has revealed extended dislocation activity in both
ground and in polished alumina surfaces [18], but the depth of the affected surface
layer was very different: the maximum depth of damage was about 20pm in
ground (40pm diamond) but less than 2pm in polished (0.25pm diamond)
surfaces.
Therefore, the most probable reason for the different hardness in ground and in
polished ceramics is that a high density of grinding-induced dislocation and twin
networks in the machined surfaces reduces the microplastic deformability (increase
the hardness) at an indentation site, in a similar way to small grain sizes.
Within the framework of the understanding outlined above it is clear that after
grinding, a subsequent further microplastic deformation (on indentation) must be
the more difficult the less mobile the grinding-induced dislocations are, that is the
closer the spacing of grain boundaries, the smaller the grain size is. As a
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Comparing the Grain S i x E j e c t and the Indentation Size Effect 1%
For indentation size effect of sapphire: 2a (pm) for n=l ,2R,, (pm) for n=4
0.1 1.o 10.0 100.0 1000.0
I I t I I I l l I I I l l I I I l l I I I l l L
I
10 ' I I I I I I I 1 I I I l l I I I l l I I I I [ I
alumina ceramics on the grain size. Note that the deformabilities 6, and Si are
phenomenological parameters that describe the microscopic deformability without
assuming specific mechanisms. For the use of these parameters within the frame-
work of the model introduced in Section 6.3.2 it is, therefore, unimportant whether
inelastic microdeformations are generated by lattice processes only (dislocations,
twins), or if additional grain-boundary deformation takes place.
If grain-boundary processes are active, they will increase the deformation and
reduce the hardness. Therefore, it is not a surprise to find in Fig. 8 a quite different
grain size dependence for polycrystalline alumina ceramics (HV 10) compared
with the (also experimentally measured!) size effect in the single crystal (sapphire,
represented by the curve derived under the assumption of a plastic zone size
2R,1 = 4 x 2a). The degree of this difference, however, is much larger than might
be expected: even microstructures with coarse grain sizes of the order of 10ym
exhibit a much higher microplastic deformation (lower hardness) than isolated
sapphire subregions of the same size.
It should be emphasized that this conclusion is independent of theoretical
speculations: all data points in Fig. 8 are measured values. Only the extrapolation
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Comparing the Grain Size Effectand the Indentation Size Effect 191
of the indentation size effect in sapphire towards smaller sizes depends on the model,
but this extrapolation was performed in a way that agrees well with published
nanoindentation results.
Comparing in Fig. 8 microstructures with average grain sizes of about 10 and
0.2 pm, respectively, the dzfeerence between the hardness of these plycrystalline
microstructures on the one hand, and of plastically deformed single crystalline
volumes that exhibit the same size as the grains in the polycrystals on the other
hand, is almost doubled when fine-grained materials are compared with coarser
sizes, and it will further increase for nanoceramics. Therefore, with the data in
Fig. 8 it seems rather improbable to find a sharp increase of the hardness in
future nanoscale corundum microstructures with grain sizes of 50nm or less.
Additionally to efforts intended to reduce the grain size, grain-boundary engineering
is, therefore, a second approach to the development of hard polycrystals, which
becomes the more important the more fine-grained the material is.
For example, the ratio between the contributions of lattice processes and of grain
boundaries to the inelastic deformation at an indentation site may be quite different
if different types of bonding (ionic, covalent) are considered. In fact, first
investigations indicate an improved resistance against inelastic grain boundary
microdeformation in submicrometer microstructures with a covalent bonding
character if these materials are produced without ionic or metallic binder phases.
Figure 9 compares highly-pure sintered alumina (>99.9% A1203) and binder-
free (!) ceramics based on WC and (Ti, W)C [20]. Whereas all these materials exhibit
a similar hardness when microstructures with average grain sizes 1 4 p m are
30
\v
WC\\
25 -
\? (Ti,W)C
\ *
\ *'
m
a \T\ ***
s?
0 pure AI,O,
0
\ *.
-
I 20-
>
(I)
acl
(I)
\ . - m
0 *=
\
E
c
\
\ *
+\
\
\
\
T \
15- \
blished surfaces] \
I I I l ~ l ~ l l I 1 1 1
I I I I I I I I
0 200 400 600 800 1000 1200 1400 1600
Temperature (OC)
Figure 10. Temperature dependence of the hardness in highly ionic (A1203)and highly covalent
(Sic) single crystals compared with polycrystalline (sintered) ceramics [21].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.7 Sunitnary 199
Figure 11. Intergranular damage at the edge of a 4 kg indent in a 2.5 pm alumina ceramic at 1600 C
P11.
the 9 pm material is less hard than the fine-grained ceramic due to the usual grain
size effect and less hard than the sapphire crystal (probably associated with some
defects). In the range 500-1 000°C the disadvantage of the coarser microstructure
compared with the 2.5 pm ceramic is offset because it is less susceptible to inelastic
grain boundary deformation due to its lower frequency of grain boundaries per
unit volume. Altogether, however, the influence of the microstructure on the
effect of the temperature is small at temperatures < 1000°C.
The situation changes when at some critical temperature > 1000°C weakening
grain boundaries cause a global instability of the microstructure at the indentation
site (Fig. 11). This temperature range is of special interest for ceramic tools, because
both cutting ceramics and grinding materials have to resist temperatures of more
than 1000°C at processing zones (cutting edges and tips). The consequence of the
general instability of the grain boundaries at high temperatures is the decrease of
the hardness to values which are much smaller than observed for the single crystal.
It is obvious that the critical temperature for this drop of the hardness depends
strongly on the grain boundary structure (amorphous interfaces, segregating
species) and on the grain size.
6.7 Summary
The measured size effect on the hardness of single and polycrystalline alumina can
be fitted assuming an increasing deformability, S,, when the extension of the plastic
zone proceeds by growth and multiplication of pre-existing elements of plasticity
instead of the more difficult generation of new dislocations, dislocation loops, or
twins. With increasing testing load (size of the indent), there is a decreasing hardness
due to the involvement an increasing number of pre-existing dislocations but, on the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
200 6 Interrelations Between Influences
other hand, with increasing size of the plastic zone the value of 6, approaches the
average deformability Si of the preceding stages of the penetration process, and
the degree of the load effect decreases. As a consequence, the hardness approaches
an asymptotic value H , at larger testing loads.
Smaller grain sizes restrict the microplastic deformation to smaller volumes, and
the process approaches H , at a smaller size of the indent. Therefore, the influence of
the load (size) on the hardness decreases in more fine-grained microstructures. The
grain size effect is larger when larger testing loads are applied, it disappears at small
loads. Hence, microhardness data should not be used to study grain size effects in the
hardness.
Comparing the hardness of polished and of ground surfaces, there is a tendency
for higher values in the latter. Whereas no significant difference exists in alumina
ceramics with grain sizes >1 pm, an increasing advantage of the ground state is
observed for fine-grained materials with HVlO > 20 GPa (submicrometer A1203,
A1203-Ti(C,0,N) composites). The most probable reason is that grinding-induced
dislocation and twin networks in the machined surfaces reduce the microplastic
deformability (increase the hardness) at an indentation site in a similar way to the
effect of small grain sizes.
A remarkably large discrepancy between the hardness of plastically deformed
single crystalline volumes and of polycrystalline microstructures with grains of
the same size, indicates a significant contribution of grain-boundary deformation
at the indentation site in polycrystalline alumina, even at room temperature and
even for coarse microstructures.
The increase of grain-boundary contributions to the deformation at small grain
sizes makes a sharp increase of the hardness in nanoceramics like alumina highly
improbable. Additionally to efforts that reduce the grain size, grain boundary
engineering is, therefore, a second approach to the development of hard poly-
crystals, which becomes more important the more fine-grained the material is.
First investigations indicate an improved resistance to inelastic grain-boundary
microdeformation in submicrometer microstructures with a covalent bonding
character if these materials are produced without ionic or metallic binder phases.
Under this condition, an increase of the hardness with decreasing grain sizes may
still exist at grain sizes of lOOnm or less.
Different mechanisms control the temperature effect in the hardness below
1000°C and at higher temperatures when weakening grain boundaries induce a
loss of the global stability of the microstructure. At lower temperature, thermally
activated processes result in a similar influence of the temperature in polycrystalline
ceramics to that in single crystals, and the influence of the microstructure on the
temperature effect is small. A steady decrease of the hardness with increasing
temperature is observed between room temperature and about 1000°C, with a
decreasing slope the higher the temperature is. The influence of the temperature
again increases when, beyond some critical temperature, weakening grain boun-
daries induce a drop of the hardness to much lower values than observed for the
single crystal at same testing load and temperature. The influence of the micro-
structure (e.g. of amorphous grain-boundary phases and of the grain size) on this
critical temperature is technically important because many ceramic tool materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Rejerences 20 1
run under conditions where the local temperature at the processing edge or tip
exceeds 1000°C.
References
1. A. Krell and P. Blank, A1203-Sinteringmaterial, process for the production of said material and
use thereof, European Patent EP-756 586 BI, Int. CL6 C04B 35/11 I , published 29 October 1997.
2. B. J. Hockey, Plastic deformation of aluminum oxide by indentation and abrasion, J . Am.
Ceram. SOC.1971,54, 223-231.
3. D. E. Munson and R. J. Lawrence, Dynamic deformation of polycrystalline alumina, J . Appl.
Phys. 1979,50,6272-6282.
4. R. A. Graham and W. P. Brooks, Shock-wave compression of sapphire from 15 to 420kbar,
J. Phys. Chem. Sol. 1971,32,2311-2330.
5. A. Krell, Gezielte Mikrostrukturentwicklung in Sinterkorund, PhD Thesis, Academy of
Sciences/Central Institute for Solid State Physics and Materials Science, Dresden, 1982.
6. A. Krell, Effects of load, grain size, and grain boundaries on the hardness of alumina, Proc.
22nd Cocoa Beach Conference (2&24 January 1998), American Ceramic Society, Westerville,
OH, 1998, pp. 159-168.
7. A. Krell and P. Blank, Inherent reinforcement of ceramic microstructures by grain boundary
engineering, J. Europ. Ceram. Soc. 1992, 9, 309-322.
8. J. Lankford and D. L. Davidson, The crack-initiation threshold in ceramic materials subjected
to elastic/plastic indentation, J . Muter. Sci. 1979, 14, 1662-1668.
9. G. P. Upit and S. A. Varchenya, The size effect in the hardness of single crystals, in The Science
of Hardness Testing and its Research Applications, J. H. Westbrook and H. Conrad (Eds), The
American Society for Metals, Metals Park, OH, 1973, pp. 135-146.
10. C. M. Perrott, Elastic-plastic indentation: Hardness and fracture, Wear, 1977,45,293-309.
I I . A. Krell, J. Woltersdorf, E. Pippel, and D. Schulze, On grain boundary strength in sintered
A120,, Philos. Mag. A , 1985, 51, 765-776.
12. A. Krell, Improved hardness and hierarchic influences on wear in submicron sintered alumina,
Muter. Sci.Eng. A , 1996, 209, 156163.
13. A. Krell, Vickers hardness and microfracture of single and polycrystalline AI2O3,Cryst. Sci.
Technol. 1980, 15, 16667-16664.
14. S . M. Wiederhorn, Fracture of sapphire, J . Am. Ceram. SOC.1969, 52, 485-491.
15. P. F. Becher, Fracture-strength anisotropy of sapphire, J . Am. Ceram. SOC. 1976, 59, 59-65.
16. A. Krell, A new look at grain size and load effects in the hardness of ceramics, Muter. Sci. Eng.
A245, 1998, 277-284.
17. A. Krell and P. Blank, Grain size dependence of hardness in dense submicrometer alumina,
J. Am. Ceram. SOC.1995,78, 1118-1120.
18. B. J. Hockey, Observations by transmission electron microscopy on the subsurface damage
produced in aluminum oxide by mechanical polishing and grinding, Proc. Br. Ceram. SOC.
1972,20,95-115.
19. W. C. Oliver and G. M. Pharr, An improved technique for determining hardness and elastic
modulus using load and displacement sensing indentation experiments, J . Muter. Res. 1992,
7, 15641583.
20. V. Richter, M. v. Ruthendorf, and J. Drobniewski, Perspektiven feinkorniger Hartmetalle, in
Pulvermetallurgie in Wissenschaft und Praxis, Vol. 13, Proc. Hagen Symposium Powder-
metallurgy, 13-14 November 1997, R. Ruthardt (Ed.), Werkstoff-Informationsgesellschaft,
Frankfurt, 1997, pp. 29-45.
21. A. Krell and 0. V. Bakun, High-temperature hardness of A1203-base ceramics, Acta Metall.
1986,34, 1315-1319.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
7.1 Introduction
Transition metal carbides, nitrides, and carbonitrides belong to a class of materials
that show an unusual combination of outstanding properties, amongst which are
exceptional hardness, high melting point (e.g. TaC and HfC have among the highest
known melting points of all materials), metallic lustre combined with vivid colors, and
simple metallic structures. In addition, they show high electrical and thermal conduc-
tivities. This combination of properties has attracted considerable attention and has
resulted in numerous technical applications [ 11 and theoretical investigations. Many
binary transition metal carbides and nitrides especially the face centered cubic 6
phases of the IVB group, exist over broad ranges of composition with an upper
limit of the nonmetal/metal ratio near 1. Practically all solid state properties show
a gradual change with this ratio. Other carbides with a different structure such as
tungsten carbide and chromium carbides have a very narrow homogeneity region.
Transition metal carbides and nitrides often form solid solutions resulting in
carbonitrides in which the solid-state properties also change with the [C]/([C] +
[Nl) ratio. This behavior offers a large field for tailoring carbonitrides.
The carbides, and to a minor extent also nitrides and carbonitrides, are easily
wetted by liquid metals, which has found application in hardmetals and cermets.
Since the basic invention of preparing composite materials by liquid phase sintering
of carbides with iron group metals, where a very successful combination of hardness
with toughness was achieved, these materials are applied as components in hard-
metals for a variety of cutting and drilling operations. In steels, carbide and carbo-
nitride precipitates cause hardening known as dispersion hardening.
The chemical stability of transition metal carbides, nitrides, and carbonitrides is
high, they are not readily attacked by dilute acids (except oxidizing acids and
hydrofluoric acid) or alkaline solutions. The thermal stability is connected with
their free energy of formation, the stability decreasing with increasing group
number. While the group IV transition metal nitrides TIN, ZrN, and HfN and
carbonitrides Ti(C, N), Zr(C, N), and Hf(C, N) can be melted without decomposi-
tion, the nitrides and carbonitrides of the other groups decompose before melting
points are reached. As a result nitrogen is released.
Tables 1 and 2 give a short overview of the properties important for technical
applications of the most well characterized and important binary transition metal
carbides and nitrides. Some specific properties which cannot easily be listed in a
table, as well as the properties of transition metals carbonitrides) especially those
of the most commonly used group IVB carbonitrides) are described in detail later.
Table 1. Bulk properties at room temperature of some important transition metal carbides (disordered state). For f.c.c. carbides (Pearson symbol
cF8) the composition is near 50 at-% C, except for VC0.88.The melting points do not apply to this composition.
Compound Pearson Lattice Density Micro- Melting Young's Heat Linear thermal Electrical
symbol parameter(s) hardness point modulus conductivity expansion coefficient resistivity
(nm) (gcm-') (GPa) ("C) (GPa) (Wm-' K- (IO-~)K-' (FQ cm)
6'
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
h,
0
P
u
3
a
z.g.
3
Table 2. Bulk properties of important transition metal nitrides. Values for room temperature and for stoichiometric composition.
5k
?I
Phase Pearson Lattice Color Density Micro- Melting Young’s Heat Linear thermal Electrical T, $.
symbol parameter( s) hardness point* modulus conductivity expansion coefficient resistivity 5
2
(nm) ( g ~ m - ~ (GPa)
) (“C) (GPa) (Wm-’ K- ) (10-6K-’) (PQcm) (K) .
3
?
TiN cF8 0.4242 golden yellow 5.39 17 3050 420 29 9.9 27 5.8
ZrN cF8 0.4570 green yellow 7.32 15 3000 460 11 7.8 24 10.47 $
HfN cF8 0.4516 dark yellow 13.83 18 3330 380 11 8.5 27 6.92 3
VN cF8 0.4138 brown 6.04 5.7 2350 380 11 10.8 65 8.9
NbN cF8 0.4392 light yellow 8.16 11 d 360 3.8 10.2 60 17.2 ?I
~ ~
TaN cF8 0.4238 gray yellow 15.9 32 d ~ 8.0 - 8.9
~ ~ -
TaN hP3 a = 0.2936 dark gray 14.3 11 d 9.5 128 2.
-
c = 0.2885 2.
CrN cF8 0.4148 gray brown 6.14 13 d 450 11.7 - 640 -
* d = decomposes
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.2 General Features of Structure and Bonding 205
E h
c
Figure 1. Hexagonal (h) and cubic (c) stacking sequence of metal atoms of close-packed h.c.p. and
f.c.c. transition metal carbides and nitrides. Open large circles designate metal atoms, filled small
circles designate interstitial nonmetal. In h sequences only 50% of the interstices can be occupied.
Nitrogen and carbon atoms are randomly distributed among the nonmetal sublattice.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
206 7 Transition Metal Curbides, Nitrides, and Carbonitrides
FTiNl --x, GZrN1--x, and 6HfN1px, crystallize in the sodium chloride structure and
no low-temperature modifications of these stoichiometric nitrides are known. In
the VB group, also all three metals form stoichiometric f.c.c. nitrides, but GVNl.oo
transforms at 205K into a tetragonal modification [6], FNbNl,ooat around 1310-
1320°C into hexagonal qNbN [7] and GTaNI.ooat around 1800°C into hexagonal
ETaN [8], the transformation temperatures depending on the nitrogen equilibrium
pressure.
In the VIB group these f.c.c. nitrides are increasingly less stable. GCrN1--x requires
a relatively high nitrogen pressure of preparation and transforms at below ca. 280 K
into a tetragonal antiferromagnetic compound [9] and yMoNl --x occurs only at up
to 38 at-% N using flowing ammonia, a value that could be increased if the nitriding
potential is additionally increased. This stability feature continues further to the
Mn-N system, where the f.c.c. phase has the stoichiometry T4N and to the Fe-N
system. In the latter only the f.c.c. solid solution of nitrogen in yFe is observed
[lo] (the Fe4N phase has f.c.c. structure with respect to the metal positions only).
Figure 2. Density of states (decomposed into the different contributions and total) of Tic and TIN [12].
The ionic contribution to the binding mechanism in these compounds can be esti-
mated from the charge transfer which occurs from the metal atom to the nonmetal
atom. It is about half an electron, which contributes to the electrostatic interaction
of metal and nonmetal.
7.3 Preparation
Transition metal carbides can be prepared by direct combination of the elements,
preferably as powder mixtures, at elevated temperatures:
Ti + C --+ GTiC.
If heat is generated by the reaction such as with Ti and C, the reaction is exothermic
enough to keep up the reaction temperature. The process can be carried out in form
of a self-sustaining high-temperature synthesis (SHS). The reaction front, once
ignited, travels through the powder bed producing the nitride and dissipating the
heat over a certain period of time. Also nitrides can be produced by SHS techniques.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
208 7 Transition Metal Carbides, Nitrides, and Carbonitrides
This reaction usually proceeds over several intermediate products and the resulting
nitride can contain oxygen and carbon.
For the preparation of large crystals (1 cm3) zone-annealing techniques were
developed [14],which are similar to those zone melting already known from material
purification processes. The starting materials are either solid-state sintered nitride
powders or metal powders pressed into rods. The applied temperatures must be
very high (around 3000°C) in order to overcome the low diffusivities of nitrogen
in nitrides, but they must not exceed the melting temperatures of the compounds
in order to avoid incongruent melting and decomposition.
Carbonitride powders for industrial applications are prepared by mixing carbide
and nitride powders intimately together and heating the mixtures to high temperatures,
TIC + TIN + 2 Ti(Co,5No.5),
whereas for the case of the Ti-C-N system any proportion leads to a single phase
product Ti(C,N1 -J provided that sufficient time or a sufficient high temperature
is provided for equilibration. For this reaction the nitrogen equilibrium pressure
is important either for preparation of Ti(C,N) or if this compound is heated at
high temperatures such as in sintering procedures. Figure 3 gives experimental as
1800 1400
-300 otm
-3 0 otm .x’ .
-I otm/
Figure 3. Nitrogen equilibrium pressure of Ti(C, N) as a function of composition from [16] with lit-
+
erature data. The composition is x in Tic, -,yN,, which is 1 - [C]/([C] [Nl).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
210 7 Transition Metal Carbides, Nitrides, and Carbonitrides
well as calculated results [15, 161 for the nitrogen equilibrium pressure as a function
of the carbon/nitrogen ratio in Ti(C,N, -~r).
A great variety of methods exists for the preparation of nitride and carbonitride
layers. These can be divided into the groups CVD, PVD [17], and diffusion layer and
thermal surface modification techniques.
(i) Chemical vapor deposition (CVD): a volatile transition metal compound such
as a chloride is brought to reaction with ammonia, nitrogen/hydrogen, or
ammonia/hydrogen/nitrogen mixtures. For carbonitrides a carbon-containing
gas such as CH4 or other hydrocarbon is used [18]. Also single metal-organic
precursors can be used which make the construction of the furnaces more
simple. An example is Bis(benzene)vanadium, V(C6H6). which is solid at
room temperature, and can be used for deposition of V(C, N ) layers [ 191. The
nitride or carbonitride that forms as a reaction product in the gas phase
nucleates on solids, like hardmetals, and forms solid, strongly adhering
layers at temperatures around 1000°C. Thus hardened steels cannot be
coated unless a hardness decrease is accepted.
(ii) Physical vapor deposition (PVD): The transition metal is evaporated or
sputtered by glow discharge or arc heating, and reacted with low-pressure.
Numerous modifications of PVD methods have been developed [20-231. The
nitride layers that are deposited on the substrate surfaces often have high lattice
defect concentrations [24] and hence exhibit significant deviations from gener-
ally accepted bulk values (lattice parameter, hardness, composition) [25]. PVD
processes applied to coat hardmetals generally result in a smoother interface
than the CVD process because of the lower temperatures.
(iii) Diffusional techniques: Carbide, nitride, and carbonitride formation, which is
used for surface hardening, can be achieved by subjecting the material surface
to the action of various salt baths, ammonia, methane, molecular or atomic
nitrogen, or nitrogen ions [26]. Reactive plasma techniques and high-pressure
techniques were developed to keep the workpiece temperature low and to
avoid thermally induced changes in the workpiece. Laser irradiation together
with an N2 jet directed on the heated spot is employed for limited local heating
of the sample [27].
7.4 Characterization
7.4.1 Chemical Analysis
Carbon analysis of transition metal carbides is nearly exclusively performed by
combustion of the carbide in oxygen and measurement of the reaction products
C 0 2 or CO.
A common technique is vacuum or carrier-gas hot extraction where CO is
formed and measured with an IR spectrometer or with a thermal conductivity
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Characterization 21 1
detector. The sample is heated to more than 2000°C in the presence of a bath
metal such as platinum in a stream of a carrier gas such as He. Due to the high
temperatures involved, the method is suitable for powder samples with larger
grain size or even for small solid pieces. The accuracy of the method is on the
order of 1-2 rel-% C.
In another technique measurement of the change in electrical conductivity of an
alkaline solution resulting from the C 0 2 absorption is performed. An oxidizing aid
such as V 2 0 5 or copper oxide is used to facilitate the oxidation.
Several procedures have been applied for the analysis of nitrogen in nitrides and
carbonitrides.
The Kjeldahl analysis involves the dissolution of nitrides in acids whereby the
nitrogen is converted into ammonium ions. The dissolution process is time con-
suming since some nitrides, such as GTiN,-,, are chemically very stable against
acids and must be dissolved under rigorous conditions by the action of hydrofluoric
acid in PTFE containers at 100-200°C.
Also vacuum or carrier-gas hot extraction such as for the carbides can be applied.
If carbon is present the CO is absorbed after IR measurement and the remaining
nitrogen is measured usually in a thermoconductivity cell. Temperature-controlled
furnaces can yield a fractional nitrogen analysis which resolves nitrogen differently
bonded or different nitride phases.
In another combustion analysis the sample is flash-heated at temperatures of up
to 1800°C.The evolved N2 is either measured volumetrically in an azotometer or via
a gas chromatograph [28,29] equipped with a thermoconductivity detector. N2,
COz, and H 2 0 can also be determined simultaneously and hence the C and the H
contents. The accuracy of the method is better than 1 rel-% N for finely powdered
samples.
7.5 Thermodynamics
7.5.1 Stability of Carbides
The thermodynamic stability of the carbides is different between heavier and lighter
metals. Stoichiometric composition of the carbides can more easily be attained with
the heavier elements of the IVB and VB groups the than with the carbides of the
metals of the first long period. Vanadium carbide shows a maximum carbon content
of VCo.89.Also chromium does not form a carbide with 50 at-% C, instead a carbide
of stoichiometry Cr3Cz.While the monocarbides of the IVB and VB groups can be
melted congruently the carbides of the VIB group metals decompose peritectically
before their melting points are reached. The monocarbide MoC decomposes in a
peritectoid manner already at about 1230°C into Mo2C C. +
2000
T/OC / "b I 2200
1800 T/ K
2000
1600
1800
1400
1600
1200 40 50
0 10 20 30
at% C
2000
2200
T/OC
T/ K
1800
2000
1600
1800
1400
0 10 20 30 40 50
Nb at% c
(b)
Figure 4. Recently reinvestigated portions of phase diagrams of the group VB carbide systems
(a) V-C, (b) N b C , (c) Ta-C [8].
Each carbon atom has eight metal neighbors in the form of a quadratic antiprism. In
that respect the structure differs basically from all other carbides.
2300
T/OC 2500
2200
T/ K
2400
21 00
2300
2000
2200
1900
2100
1800
2000
1700
I 1'0 20 30 40 50
Ta at% C
(C)
such as (Ti, W)C and the q phases [35] such as W3C03Cof which the occurrence in
decarburized compacts (e.g. near interfaces of CVD coated cutting tools) is
unwanted causing embrittlement.
F.c.c. transition metal carbides often form complete solid solutions with each
other particularly at high temperatures. They undergo phase separation at lower
temperatures [36,37]. Miscibility gaps have been observed in the quasibinary
systems Tic-ZrC, Tic-HfC, VC-NbC, and VC-TaC. At high temperatures also
33
T 31
m
a
(3
29 i
Q)
c
P
27 r"0
L
0
2
25
23
60 0 70 0.80 0 90
Composition [ C ] / M
Figure 5. Microhardness of ordered (B) and disordered (A) VC, --I as a function of composition
1341.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
216 7 Transition Metal Carbides, Nitrides, and Carbonitrides
complete miscibility is observed in the systems where the f.c.c. phase is a high-
temperature phase at least for one binary constituent such as in the group IVB
carbide (f.c.c. WC and group VB carbide) f.c.c. MoC quasibinary systems. Even at
low temperatures the solid solubility of MoC and WC in the f.c.c. carbides of the IVB
and VB groups is substantial and is exploited for use in cutting tools.
3500 I I
T/OC 3500
3000 T/K
3000
2500
2500
2000
2000
1500
1500
1000
1000
500
0 10 20 30 40 50
Ti at% N
Figure 6. Phase diagram of the Ti-N system [38].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.5 Thermodynamics 2 17
2500
T/OC 2500
2000 T/ K
2000
1500
1500
1000
1000
500
500
0
0 10 20 30 40 50
v at% N
Figure 7. Phase diagram of the V-N system [lo] with reinvestigated homogeneity ranges in the
temperature range 1100-1700°C (thick lines) [46].
of the subnitride phases which exist in the neighboring systems Ti-N and Hf-N
occur here. The Zr3N4phase [40] is obviously not of technical importance.
Besides the f.c.c. 6HfNl the Hf-N system contains two subnitride phases [41],
qHf3Nz-, and <Hf4N3-x.Only a few phase equilibria studies have been performed
for this system and they are cited in the ASM Handbook [lo]. The decomposition
temperatures of the subnitride phases need to be re-investigated.
2000 j j
1400/ P-Nb2N
lzyyl
1000
0 10 20 30 40 50 60
Nb at% N
Figure 8. Phase diagram of the Nb-N system with reinvestigated homogeneity ranges and phase
reactions (thick lines and data points from electron-probe micoanalysis and thick broken line for
the yNb4N3+tx+ GNbNI --x transition [7]). The thin solid lines are previous data.
2000
- 2200
T/OC
1800
T/ K
- 2000
1600
- 1800
1400
0 10 20 30 40 50
Ta at% N
(a)
3500 *
3000
2500
s!3
E 2000
c
a
0.
+E, 1500
E-TaN t
,
1
1000
Figure 10. Phase diagram of the Cr-N system [lo] with recent results on the homogeneity ranges of
Cr,N and CrN, - X . The homogeneity range of the latter broadens substantially upon increasing
temperature. The nitrogen-rich phase boundary of CrN,-, is for 30 bar N2 and extends most
probably at least up to 50 at-% N at higher pressures.
homogeneity range of CrNl --x is interesting and it is not completely clear whether
the line compound occurring below about 1200°Chas the same structure as CrNl - x
at higher temperature.
Cr2N is readily dissolved by dilute acids, whereas CrN appears to withstand acid
attack for prolonged periods. CrN is paramagnetic at room temperature and under-
goes a paramagnetic-antiferromagnetic first-order transition at temperatures of
276-286 K, depending on the composition [9].
In the Mo-N system three nitride phases have been reported 1551. An f.c.c. phase,
originally named yMo2N because it has a composition around 33 at-% N, should
instead be written as yMoNl - x , to be consistent with the notation for other f.c.c.
nitrides. yMoNlPx has been reported to also exist at nitrogen contents above
33 at-% N. It can be assumed that it is a phase with a very broad homogeneity
range. This is also indicated by the results obtained from high-pressure experiments.
yMoNl - x undergoes an ordering transition between 400 and 850"C, depending on
the composition, into the so-called pMo2N. Another phase, GMoN, with a structure
similar to that of WC, has been prepared by various methods, including nitridation
of Mo powder with flowing ammonia, or reaction of MoC15 with NH,. It has been
structurally characterized to have several types of ordering arrangements of nitro-
gen in the centers of trigonal prismatic structure elements [56].
Molybdenum nitrides are studied for their interesting properties in catalytic
reactions such as for the desulfurization and denitridation of oil products which
is of increased importance for use in cars where sulfur can destroy the activity
of the catalyzer. These powders can be produced with a high surface area by
reaction of MOO, or MoS2 with NH3. Also layers of molybdenum nitrides, pre-
pared by various techniques such as reactive sputtering or ion beam deposition,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.5 Thermodynamics 22 1
were studied for application of wear-resistant coatings as well as for diffusion bar-
riers [57,58].
Ti
Tic TIN
at% N
(a)
Zr
ZrC 10 20 30 40
ZrN
at%N
(b)
Figure 11. Phase diagrams of (a) the Ti-C-N and (b) Zr-C-N systems for T = 1120°C [63].
I I I I I I
I \ I
0.3 X
LL
a 0.2
0.1
fcc
I
01 I I I I I I
0 0.1 0.2 0.3 0.4 0.5 0.6
Tic Site Fraction Mo
(a)
TIN G+graphite/G+WC+graphite WN
wc
Figure 12. Calculated phase diagrams of (a) the Ti-Mo-C-N [64] system for 1600°C with experi-
mental data of [67] and of (b) the Ti-W-C-N [65] system for 1423°C with experimental data of
[68]. Both for 50 at-% nonmetal in the representation of a reciprocal salt pair.
hard constituent in cermets [66]. These systems are technically most important for
50 at-% carbon + nitrogen and were described in terms of a section TiN-Tic-
MoN-MoC and TiN-Tic-WN-WC. Figure 12a, b presents the calculated version
of these phase diagrams for a nonmetal/metal ratio = 1 together with experimental
data. A homogeneous f.c.c. phase (Ti, Mo)(C, N) separates into two f.c.c. phases, of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
224 7 Transition Metal Carbides, Nitrides, and Carbonitrides
which one is richer in Ti and N than the other (Fig. 12a). The two-phase field
f.c.c. + f.c.c.2 becomes smaller at higher temperatures.
A more complicated situation exists for the Ti-W-C-N system because of the pre-
sence of a hexagonal WC phase. Figure 12b shows the calculated phase diagram of
this system, also for a nonmetal/metal ratio of 1 [65] together with experimental data
[68]. Also two f.c.c. phases form and the two phase field is even larger than in the
corresponding system with Mo because of the higher nitrogen equilibrium
pressure. Upon increasing carbon concentration graphite and WC form together
with the two f.c.c. phases.
The phase equilibria in the Ti-Mo-C-N and Ti-W-C-N systems have influence
on the microstructures of cermets (see 7.7.1), at least partially responsible for a
favorable core-rim type microstructure of the hard particles [69].
7.6.2 Color
Transition metal carbides generally show a gray color, the exceptions are TaC and
NbC, which have dark-brown/bronze color.
Many mononitrides exhibit an intense color of bright metallic appearance which
depends on the nonmetal/metal ratio. Table2 lists the colors of the mononitrides
near the composition indicated by their formula. These colors can be observed
not only for bulk samples but also play an important role when thin films [71] are
produced for ornamental purposes (see 7.7.2).
Very interesting is the color change of the stoichiometric group IVB carbonitrides
+
as a function of the [C]/([C] [“I) ratio (Fig. 13). The higher number of the period
the more the vivid violet color extends into the ternary region starting from the
yellow colored nitrides.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Curbonitrides 225
+
Figure 13. Colors of the group IVB carbonitrides as a function of the [C]/([C] [N])ratio (which is x
in MeC,NI -.J. Note that the colored region extends wider into the carbon-rich region the higher is
the period number [73].
50
45
F
5 40
-0,79
iij 25
a,
L - 0,4
20
-..- 0,oi
15 1 , 1 , 1 , ( , 1 , 1
50
45
_----------
0.61
0,43
20 0,25
_ ,1.1.
0
15 I I I I I , , , , , , ,
resistivity of group IVB carbonitrides (Fig. 15) which shows a general increase of resis-
tivity with increasing [C]/([C] + [q) ratio. Only at around [C]/([C] [N]) = 0.5 the +
resistivity increase is much less than at high and low carbon contents, respectively.
Some transition metal carbides are superconducting. The highest T, value has
been reported for NbC (1 1.1 K) and TaC (10.3 K) and hexagonal WC (10.OK).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Carbonitrides 227
50
45
z-
i.$ 40
35
.-
c
U
$ 30
8
g 25
c
20
15
0 200 400 600 800 1000 1200 1400
temperature ["C]
(C)
Tic, VC, HfC, and ZrC are not superconducting above 1 K. The electrical con-
.
ductivities of transition metal carbides and nitrides are greatly influenced by the
nonmetal/metal ratio; as this ratio approaches unity the electrical conductivity
reaches a maximum. This is shown for FTiNl --x in Fig. 16 [74]. A similar behavior
can be observed for other carbides and nitrides. The electrical conductivities of these
compounds decrease linearly with increasing temperature [75].
Superconductivity is common among all f.c.c. transition metal nitrides of the
IVB and VB groups as well as yMoN, -.x and GMoN. GNbN, -.x shows the highest
7
0.06
0.05
E 0.04
u
R
F
b
m
0.03
0
0.02
O.O10
0 0.2 0.4 0.6 0.8 1
COMPOSITION, [C]I([Cl+[NI)
+
Figure 15. Electrical conductivity of group IVB carbonitrides as a function of the [C]/([C] [N])
+
ratio [74]: Ti(C,N), Zr(C, N), 0 Hf(C, N).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
228 7 Transition Metal Carbides, Nitrides, and Carbonitrides
0,4
0,50 0,60 0,70 0,80 0,90 1 ,oo
composition [N]/p]
Figure 16. Electrical conductivity of TiN, --x as a function of composition [74].
25
20
.-'Y
h
15
% v
v
d lo
0
100 1100 2100 3100 4100
T (K)
Figure 17. Thermal expansion data of WC for ambient and higher pressures [77]. Reprinted by per-
mission of The American Ceramic Society.
GTi(C,N1_,y): LY,,
= (9.9 - 1.4 x [C]/([C]+ [N]))x 10-6K-1
GZr(C,NI -.y): a,, = (7.8 - 0.3 x + [N])) x lop6K-I
[C]/([C]
GHf(C,NI Q,, = (8.5 - 2.4 x [C]/([C]+ [N])) x loP6 K-',
where [C] and [N] are the carbon and nitrogen contents, respectively. These
+
expressions are valid for 50at-% [C] [N]. It can be realized that Zr(C,N) has
the smallest thermal expansion coefficient at nitrogen-rich compositions, and also
the smallest increase as a function of increasing carbon content. For the (nonaver-
aged) linear thermal expansion fitted by polynomial expressions see [go].
7.6.5 Diffusivities
In transition metal carbides, nitrides, and carbonitrides, nitrogen diffusion occurs via
a vacancy mechanism in the nonmetal sublattice. Because the vacancy concentration
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
230 7 Transition Metal Carbides, Nitrides, and Carbonitrides
12.0 -
1
11.0 -
-
h
id
L 10.0 -
v
8
e
.-
9.0 -
-
2
0
8.0-
-3
.-
L
-
0
C
I! 1.0 -
f
8
U
6.0 -
5.0 -
, 1 1 1 1 1 1 1 1 1 1 1 , 1 1 1
Figure 18. High-temperature thermal expansion of various transition metal nitrides [79].
is a function of composition within the homogeneity range of the f.c.c. phases, the
nitrogen diffusivity is often composition-dependent as well. This has only been
recently started to be investigated in transition metal carbides and nitrides for
which a detailed analysis of the nitrogen diffusion profile is a necessary pre-requisite.
Many diffusivity studies were conducted by investigating nitride or carbide layer
growth as a function of time, which gives the average nitrogen diffusivity in the
various phases [81]. A parabolic relationship, d = kt’/*, holds for this layer
growth, where d is the layer thickness, k is a constant, and t is the diffusion time.
The parabolic behavior indicates that the nitride layer growth is indeed diffusion-
controlled, provided that the sample is semi-infinite. Recent efforts for investigation
this layer growth has shown that the thickness of the sample can have crucial
influence on the layer thickness [82,83]. Methods were developed using wedge-
shaped diffusion couples in order to exploit this behavior for diffusivity
measurements [84]. A wedge shaped diffusion couple is presented in Fig. 19,
which shows the thickness enhancement of diffusion layers upon decreasing
sample thickness.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Carbonitrides 23 1
Figure 19. Diffusion layers of hafnium nitrides formed in a wedge-type sample. Note the thickness
enhancement upon decreasing sample thickness [82].
Table 3. (a) Concentration dependency of the nonmetal diffusivity of transition metal carbides and
nitrides (for compounds with a = 0 no significant concentration dependency is present). If two
values are listed the first value corresponds to the nonmetal rich and the second to the nonmetal
poor composition.
2.9 3.95
6.5
20 5.0
1.28 3.2
0.718 2.9
0.0121 2.5
0.02 2.70
0.14 2.14
0.54 2.73
12.7 2.93
13.6 2.92
3.4 3.6
120 4.0
3.51 2.68
nitride phases were investigated for their elastic properties, mainly the Young’s
modulus E, and the Poisson’s ratio v. Both, bulk samples as well as thin film samples
were measured and generally polycrystalline material was studied [911.
Among these studies, many of the data suffer from lacking or inappropriate
characterization of composition and microstructure. The porosity of bulk samples
as well as the substrate of thin films can influence the results substantially. Correc-
tion procedures for accounting for such influences are sometimes inappropriate [92].
Depending on the correction function, 5% porosity can cause a drop in E of about
50 GPa.
For T i c most values for the Young’s modulus group around 450-460 GPa for a
composition around 50 at-% carbon, the Poisson ratio is 0.17-0.19. The Young’s
modulus of the important hexagonal phase WC is very high at room temperature
[77] and drops by about 20% upon reaching 2000 K (Fig. 20).
Although the measured Young’s modulus of TIN1-, show, quite a scatter
because of the above-mentioned problems, it can be stated that E increases with
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Carbonitrides 233
L
0 Kauk &Kockelmanna
a
28
0 Kreimer
24
0 Lee 8 Gilmre (polyuys.)
v1 0 Koster L Rauscher ”
3
n
3 0 Bukatov et al.
W
0 + Brown et al.
25
v) 0
X Neshpor 8 Sarnsonov
C
ET)
c + C
3
0
600 X
550
0 1000 2000 3000 4000
increasing nitrogen content. Many of the thin film samples show a higher Young’s
modulus than bulk samples (see Fig. 21). Poisson’s ratios of 0.18-0.321 were
reported, the most probable value is likely to be around 0.22. Similar statements
can be made for ZrNl-, and H f N - , with values reaching up to 500 and
459 GPa, respectively (both values are for different kind of bulk samples: diffusion
annealed or hot-pressed) and with Y = 0.25 for both compounds. Young’s modulus
data on group VB transition metal nitrides are scarce but were measured for single
crystals of 6VN and 6NbN [93], yielding the maximum values in the (100) direction
of 487 and 490 GPa, respectively. For CrN, values up to 520 GPa were reported.
For Ti(C,N) an increase of E with carbon content was measured on several
samples (see citations in [92]), the porosity of which was, however, rather larger
and different in different samples. A well characterized sample Ti(C0.20N0.80) with
a porosity of only 3.5% [63] yielded the highest value of 451 GPa (and v = 0.212),
which probably comes very near to the value for an ideal zero-porosity samples. An
investigation of the E and G behavior as a function of temperature was recently
conducted on polycrystalline samples of TIN [94], where a linear relationship was
found for both quantities (see Fig. 22).
For an extensive review on the elastic properties ( E and Y) of transition metal
nitrides, carbonitrides (and carbides) as well as on the porosity correction routines
the reader is referred to the tables given in [92]. Upon application and development
of high-temperature investigation techniques, such as Brillouin scattering [95,96] a
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
234 7 Transition Metal Carbides, Nitrides, and Carbonitrides
800 I I I I I -
0
a 600 -
(3
Y
Y,
3
-
3 0
400 -
0
tn
c3)
c
13
0
>- a
n
"
0.0 0.2 0.4 0.6 0.8 1 .o
Figure 21. Room temperature Young's modulus of TiN, - as a function of composition and for
different sample types [92].
7.6.7 Microhardness
Sometimes large differences are encountered between the microhardness of bulk
material and thin films. This is most probably due to the different dislocation density
and the grain size of both types of materials. Also the difference in testing (micro- or
nano-indentation for films, heavier load up to several N for bulk samples) most
probably yield differences. In addition, the composition (as for any of the properties
discussed here) influences the microhardness. Because of these peculiarities the
values are difficult to compare.
Many of the transition metal carbides such as T i c and ordered VCI --x are very
hard compounds. In GTiC1- the microhardness increases with increasing carbon
content [97], a phenomenon that is probably closely related to the valence electron
concentration (VEC) with a maximum stability at VEC = 8 at the composition TIC.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Carbonitrides 235
435 I I I I 180
430 - 175
425 - 170 F
- 9
420 165 0
rn
41 5 - 160 53
U
41 0 - 155
Q
405 - 150 f
0)
400 - 145
395 I I I I
140
Figure 22. Young’s modulus ( E ) and shear modulus (G) of TIN as a function of temperature for a
composition of 50 at-% N [94].
Hardness data for transition metal carbides near the composition indicated by the
formula are contained in Table 1.
Single crystal investigations and orientation-dependent measurements on poly-
crystalline material of transition metal carbides show that the microhardness is
orientation dependent [98] (e.g. for WC and NbC [99]). For T i c the Knoop hardness
in the (1 10) plane is 27 GPa whereas that in (100) plane is 31 GPa, for example.
Upon ordering the microhardness can change significantly (compare Fig. 5). Gen-
erally the hardness values of annealed crystals are lower than as-grown crystals
due to their lower dislocation density. Polycrystalline material has higher values
by about 1-2GPa than single crystals due to the grain boundary influence on
hardness (Hall-Petch relation).
Single-crystal and polycrystalline transition metal carbides have also been
investigated with respect to creep, plasticity and slip systems. The f.c.c. carbides
show slip upon mechanical load within the (111) plane in the (110) direction
[98,100]. The ductile-to-brittle transformation temperature of TIC is about 800°C
and is dependent on the grain size. The yield stress in TIC obeys a Hall-Petch
type of relation, that is, the yield stress is inversely proportional to the square
root of the grain size. TIC and ZrC show plastic deformation at surprisingly low
temperatures around 1000°C.
Figures 23a-e show the hardness vs. composition for a variety of transition metal
nitrides [43,74,79,101] and carbonitrides [88]. The hardness decreases with
increasing nitrogen content for the group VB nitrides [43,79,101]. This behavior
can probably be explained by an electronically induced lattice softening as a
result of the generation of antibonding states. It is known from band structure
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
236 7 Transition Metal Carbides, Nitrides, and Carbonitrides
26
25
24
23
19
18
17
16
0,50 0,60 0,70 0,80 0,90 1,oo
composition [N]/[Ti]
(a)
35
30
0
0,70 0,80 0,90 1,oo
composition "]/[Me]
(b)
Figure 23. Microhardness of nitrides and carbonitrides as a function of composition. (a) TiN [74],
(b) VN [43], NbN and TaN [loll, (c) Cr-N [54], (d) Ti(C, N) +, and Ti(C, B, (e) Zr(C, N) +
and Hf(C, N) [73].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
231
7.6 Properties of Important Transition Metal Carbides, Nitrides, and Carbonitrides
1
1
50
0,8
03
composition [C]/([C]+[N])
40
composition [C]/([C]+[N])
Composition [at% N]
0,6
0,6
1
30
p-Cr,N
I
0,4
0,4
20
1
I0
02
02
I
0
0
0
12
18
16
14
-m1 4 1
28
26
20
E 18
16
24
22
14
24
22
g 20
a 12
(e)
(d)
E
2
E
(C)
P
E
.-0
gc
rn
ln
ln
.-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
238 7 Transition Metal Carbides, Nitrides, and Carbonitrides
calculations that in f.c.c. transition metal compounds with more than eight valence
electrons the antibonding states are occupied. An increasing valence electron con-
centration due to increasing nitrogen content would then lead to a decrease in
hardness when the number of valence electrons exceeds eight. Indeed in GTiN, --x
[74] a maximum in microhardness was found for the compound STiN0.67
corresponding to 7.3 valence electrons. For 6HfNl --x, however, a smooth increase
in microhardness with increasing nitrogen content was observed. This cannot be
explained in the same manner, and is probably a result of the difference in electronic
states due to 5f electrons. The microhardness of chromium nitrides [54] is shown in
Fig. 23c as obtained from diffusion couples. Upon use of diffusion couples, which
exhibit thick diffusion layers, bulk values of the microhardnesses can be measured
as a function of composition, provided that the nitrogen diffusion profile is
measured.
The microhardness of the IVB group carbonitrides (Fig. 23d, e) increases with
increasing [C]/([C]+ [Nl) ratio. For Zr(C, N) and Hf(C, N) the increase is steep on
the nitrogen-rich side and flattens on the carbon-rich side, so that a significant
positive deviation from linearity can be observed. Contrary, for Ti(C, N) this devia-
tion is only very small for 50 at-% nonmetal but significant for a substoichiometric
composition of 45 at-% nonmetal.
favorable properties such as high hardness and wear resistance, good electrical and
heat conductivities and is readily wetted by the binder metals Co and Ni.
The ternary system W-C-Co is shown in Fig. 24 at 1423°C [102]. WC coexists with
the Co phase and forms a pseudobinary lower melting eutectic between WC and Co.
This eutectic facilitates full densification during liquid phase sintering. The carbon
content must be kept close to the theoretical value because a decrease in total
carbon content will lead to the formation of -q carbides (W, Co)& which are brittle
and harmful to the performance of these materials. In modern hardmetals, also
f.c.c. carbides such as TaC, NbC and T i c are admixed which increase the high-tem-
perature performance. Figure 25 shows a microstructure of a modern WC/Co-based
hardmetal containing about 10weight-% of f.c.c. carbides Tic, TaC, and NbC.
Shortly after the invention of WC-Co hardmetals, T i c based hardmetals have
been proposed as cutting tools but have not found extended use because of their
brittleness. Only in the last two decades Tic-based hardmetals have found wider
use when titanium nitride, GTiN, had been introduced as a further constituent.
GTiN is used as one of the hard components in sintered carbonitride compositions
containing TIC, Mo2C, WC, TaC, NbC, and other compounds [69,103]. Carboni-
trides for such materials can be produced either by blending the various powders
before liquid phase sintering with Ni and/or Co as a binder phase, or directly
within the liquid phase sintering cycle. The modern titanium carbonitride cermets
[ 103,1041 are characterized microstructurally by a structured hard phase featuring
core-rim structures embedded in a tough Co-Ni binder. These carbonitrides are
often called cermets or Ti(C, N) hardmetals, a microstructure of which is given in
Fig. 26. The hard particles a roundly shaped as compared to the faceted WC
grains (Fig. 25). In the microstructure of Fig. 26, which is a SEM image, core-
rim type grains with a dark core and a gray rim as well as grains with a white
core and a gray rim can be seen.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
240 7 Transition Metal Carbides, Nitrides, trnd Carbonitrides
Figure 26. Microstructure (SEM image) of a cemented carbonitride (WIDIA TTI 25) showing
round hard particles with core-rim type structure.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.7 Industrial Applications 241
The boundary between hardmetals and cermets is not strict because many of these
compacts resemble microstructure features of both type of materials [ 1061 faceted
WC crystals together with round-shaped titanium carbonitride-based hard particles.
Generally, these titanium carbonitride hardmetals are comparable with respect to
properties and microstructure to WC-based hardmetals. The powders of these
materials are liquid phase sintered with Ni or Ni-Co binder metal alloys. The
core-and-rim structure of the hard phase usually exhibit a molybdenum- and
carbon-rich (Ti, Mo)C rim and a titanium- and nitrogen-rich Ti(C, N) but can
also be inverted (compare Fig. 26). The metallurgy of the phase reactions is (because
of the complexity of the multicomponent system) not yet fully understood [69].
According to the theoretical background of the phase reactions within the
system Ti-Mo-C-N [67] the quaternary compound (Ti, Mo)(C, N) with a random
distribution of the metals Ti and Mo on the metal sites and the nonmetals C and
N on the nonmetal sites undergoes a decomposition with the formation of isotypic
f.c.c. phases but with a preferential arrangement of N around the Ti atoms and Mo
around the C atoms. This decomposition is of a spinodal type, which is a decompo-
sition induced by local composition fluctuations without necessary activation energy
rather than a classical nucleation and growth process. The question of whether the
spinodal decomposition is a necessary prerequisite of core-and-rim formation is still
unsettled because core-and-rim type structures may develop via reactions of the
hard constituents with the liquid binder metal as well. The core-rim structure has
turned out to be most probably stable [64, 1051. It has been claimed to be favorable
because the molybdenum-containing rim is much better wetted by the binder alloy
than the core. Modern Ti(C, N) hardmetals have been successfully applied for finish-
ing operations of high-strength steel grades and ductile cast irons at high cutting
speeds but with moderate cross section of the chips. The main advantages of
Ti(C, N) hardmetals are the high chemical resistance and the low frictional welding
tendency giving longer life of the cutting edge and good workpiece surface.
Hardmetal-like microstructures are also produced by spray coating processes in
which material is deposited by high velocity. Sprayed parts are used in a variety
of processes where sliding wear occurs (aerospace industry, stone industry), advan-
tage being that the parts can be refurbished. Various methods have been developed,
such as detonation gun spraying and high-velocity oxygen fuel spraying. With these
techniques layers of WC-Co, Cr3C2-Ni/Cr, and TiC/Ni mixtures are deposited with
a thickness up to the mm range [107].
Figure 27. Milling and turning inserts with various coatings. The top coating is TiN (golden yellow),
Zr(C, N) (violet) and A1203 (black). From WIDIA, Germany.
Figure 28. Microstructure of a coated WC/Co-based cutting tool for turning steel (WIDIA TN250)
where the layer is composed of three sublayers.
(Ti,Al)N gold + dark blue, depending on the Ti/AI ratio, Ti(C,N), Zr(C,N)
(colors: see Fig. 13), CrN (metallic to brown) are used [l 1I, 1121 for example on
watch cases, lighters, frames for glasses, and bathroom furnishings where the
combination of color with increased scratch and corrosion resistance is exploited.
The color of these materials stems from the absorption behavior, which is the
result of the response of the material to the incident light, in turn a function of
the electron density of states. The CIE system is widely applied for defining colors.
This is a plot of in the coordinates L* (lightness), a* (red-green value), and b*
(yellow-blue). The color can be also dependent on the surface roughness, the
layer thickness and the structural disordering. For example, in TiN the color
appearance is compositionally dominated whereas in ZrN it is microstructure domi-
nated. Thus, a certain variability exists for establishing colors apart from composi-
tional point of view and related materials can behave differently. However, this
behavior brings some difficulties with respect to the reproducibility of colors
which in turn is influenced by the color perception of the human eye [113].
The deposition of TIN and ZrN on glass for windows represents a combination of
ornamental application with the economic constraint of energy saving in buildings.
ZrN layers, especially, have been found to exhibit favorable properties for transparent
heat mirrors, i.e. windows which transmit visible light but reflect heat radiation [114].
Chromium nitride layers (fabricated by, e.g. cathodic arc plasma deposition) are
interesting because of their corrosion properties as well as because of their excellent
adhesion properties and fine-grained structure. They are applied for die-casting
moulds where excellent edge properties are necessary [115,116]; some of these
layer can have a multiphase character composed of Cr(N), Cr2N, and CrN1-,
[117]. Sputter deposited ternary chromium nitrides such as Cr,Me, -,N with
Me =Ti, Nb, Mo, and W additions and with grain sizes of up to 25 nm have been
found [118] to show either a hardness minimum (Me = Mo, Ti) or a maximum of up
to 27 GPa (Me = W, Nb).
The Ti-A1-N system contains three ternary nitride phases but the most important
phase for industrial applications is the thermodynamically unstable f.c.c. phase
(Ti, A1)N [119,120], which can be prepared by magnetron sputtering and has
found application for hardmetal tools [121-1231. (Ti,Al)N was found to exhibit a
better oxidation resistance than TiN, depending on the Ti/Al ratio which can
vary over a broad range. At high A1 contents a hexagonal phase forms probably
related to or identical with the wurtzite-type AIN. Drill bits coated with (Ti, A1)N
show a better performance than coated with TIN. The best results (about two-
fold lifetime) were obtained with (Tio,5Alo.5)N. Film hardnesses are on the order
of 20-26 GPa, depending on the composition.
Layers composed of (Ti, Zr)N show a higher hardness (up to 33.5 GPa) but the
performance is not as good as of (Ti,Al)N [121]. In (Ti,Zr)N layers a spinodal
decomposition around the composition (Tio.sZro.7)Nwas found [ 1241 which was
also described for the phase equilibria of bulk samples [36,125].
Zr(C,N) layers together with a variety of other layers were successfully intro-
duced for cutting tools for turning steel. A microstructure of such a multilayer is
shown in Fig. 30 where Zr(C, N) represents a thick top layer. The favorable perfor-
mance of such Zr(C, N) containing layers stems for the thermal expansion behavior
of Zr(C,N), which is the lowest amongst the group IV carbonitrides, and which
reduces tensile stress upon thermal load. Such a behavior enhances the performance
of the coatings because of increased crack resistance and adhesion.
While the first and second generation of coatings contained monolayers of TiN or
T i c and multilayers composed of TiN, Ti(C, N), TIC, TiB2, and (Ti, A1)N (together
with A1203)there is a general trend in research towards multicomponent/multiphase
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.7 Industrial Applications 245
+ Zr(C,N)
+ AI,O,
+ Ti(C,N)
6 TIN
f- Substrate
10 pm
H
Figure 30. Microstructure of a Zr(C, N) top-layer coated steel turning hardmetal (WIDIA TN7015).
due to the mismatch) can also be ruled out at least for a part of the superlattices. The
presence of dislocation line energy effects in which the shear modulus is important to
move a dislocation (line energy= Gb2/2,where G is the shear modulus and b is the
Burgers vector) can also contribute to the hardness increase but the observed increase
is much higher than calculated from this theory (there is, however, some uncertainty
in these calculation due to the lack of data for G ) .
While a Hall-Petch type hardening is probably not the most prevailing
mechanism in the above cited multilayers, it has been made responsible in nano-
crystalline layers composed of two phases rather different in nature (crystalline,
amorphous or different crystal structure). Although extremely fine-sized powder
can show a negative Hall-Petch effect (nanosoftening), films composed of two
such different phases having strong interphase boundaries suppress such effect
resulting in a drastic increase in hardness.
Films of the Ti-B-N system [133,134] consisting of a thermodynamically
unstable f.c.c. phase Ti(B, N), where the B/N ratio can be varied, were also found
to show increased hardness of up to 50 GPa. These films can be made single crystal-
line, the hardness of which is dependent on the ion bombardment intensity. Thus a
dislocation density mechanism seems to be responsible for the hardening effect.
Nitrides in the form of thin layers have an important application in electronic
devices such as VLSI (very large scale integrated circuits) as Ohmic interface
materials to prevent the formation of intermediate phases between Si and A1 or,
in the most recent developments, where A1 is replaced by Cu because of the better
electromigration properties upon continuing miniaturization, between Si and Cu.
This application uses the chemical inertness, their high electrical conductivities
and the ease with which nitrides can be deposited as layers. Mainly TiN deposited
by PVD techniques is used in these circuits but also tungsten and molybdenum
nitrides [58]. Also Tio.8A10,2N was studied and has shown a better performance
than TIN [135]. For the new Si/Cu interconnection, TaN has proved to be of super-
ior performance [136]. Films of tantalum nitrides are also used or of potential use for
a variety of electronic applications ranging from thin-film resistors as well as for
thermal printing heads and for wear and corrosion resistant coatings [137]. These
films can be deposited by CVD with and without plasma [I 381 and PVD methods
[136,137]. For CVD methods organic precursors [139,140], TaC1, [141], or TaBr5
[138] can be used. If a low resistivity of the tantalum nitride films is wanted, then
the f.c.c. TaN,_, is desired without presence of the insulation Ta3NS phase,
+
which is difficult if TaCl, NH3 as starting substances are used. TaNl --x films
(deposited by both by PVD and CVD) have a resistivity of 130-250pRcm [142].
Niobium nitride films show good performance for vacuum microelectronic devices
for displays because of a high work function and resistance to sputtering [143].
its alloys (e.g. TiA16V4) are of interest in order to circumvent the strong cladding
behavior of these materials [26]. Also plasma nitriding techniques are used for the
preparation of surface layers on the order of several microns thickness. The
microstructure of these diffusion layers depend strongly of the alloy composition,
the temperature and the nitrogen pressure [ 1441.
A technique which is related to this surface refining method, is the surface laser
heating of alloys. In this method a strong laser is used to melt titanium alloys at
the surface with simultaneous supply of nitrogen via a jet or by operating in nitrogen
atmosphere [27,145]. By such a method nitrides (though high-melting) form by
surface melting together with an intermediate diffusion layer form in the solid
state. Nitride layer thicknesses of the order of millimeters can be prepared. These
layers are, for example, applied in large turbine blades in order to prepare a sur-
face withstanding the intensive cavitation of water droplets reaching supersonic
velocity.
One of the most interesting developments in the field of hardmetals and cermets
for cutting tools is the creation of a graded surface which makes use of the favorable
properties of functional-gradient materials by dissipating the thermal and mechan-
ical load onto a larger part of the cutting insert in order to increase the lifetime or to
withstand extreme conditions of cutting [146]. Figure 31 shows surface-near
microstructures of such functional-gradient cermets composed of (Ti, W)(C, N)
with additions of TaC and NbC and a Co binder metal. Upon sintering, the
hexagonal phase WC forms (which can also be present in the starting formulation).
Depending on the composition and the sintering cycle a smooth variation of the
phase content as a function of the distance from the surface can be seen. This
morphology is caused by the action of reactive gases such as nitrogen and carbon
monoxide introduced at specific points of the sintering cycle. The action of reactive
gases can be such modified that different gradients can be fabricated and even layers
of different chemical and phase composition can be prepared together with gradi-
ents. This functional-gradient hardmetal has a Ti(C, N) top layer, an intermediate
WC-Co layer and a gradient in which the (Ti, W)(C, N) phase is enriched at the
outer side and continuously reaches bulk composition. In view of the rather high
production costs of coated materials (the coating process is about 15-20% of the
total production costs of a cutting insert) such modified cemented carbonitrides
could be lower the production costs significantly because the diffusion layers and
gradients are established during the sintering cycle and (provided similar perfor-
mance can be obtained) an additional coating cycle is then unnecessary.
Acknowledgments
The author would like to thank Dr. K. Dreyer, WIDIA GesmbH, Essen (Germany) for
support with figures and microstructures of tools parts. The help of Mr. W. Prohaska
with the preparation of the phase diagram drawings is gratefully acknowledged.
References
1. P. Ettmayer and W. Lengauer, Nitrides in Ullmann’s Encyclopedia of Industrial Chemistry,
Verlag Chemie, Weinheim, 1991, A17, p. 341.
2. E. Parthe and K. Yvon, Acta Crystallogr. B, 1970, 26, 153.
3. K. Yvon and E. Parthe, Acta Crystallogr. B, 1970, 26, 149.
4. T. Epicier, in The Physics and Chemistry of Carbides, Nitrides and Borides, R. Freer (Ed.),
Kluwer, Dordrecht, 1990, p. 215.
5 . C. H. de Novion, B. Beuneu, T. Priem, N. Lorenzelli, and A. Finel, in The Physics and
Chemistry of Carbides, Nitrides and Borides, R. Freer (Ed.), Kluwer, Dordrecht, 1990, p. 329.
6. F. Kubel, W. Lengauer, K. Yvon, A. Junod, and K. Knorr, Phys. Rev. B, 1988,38, 12908.
7. W. Lengauer, M. Bohn, B. Wollein, and K. Lisak, Acta Muter. 1999, in press.
8. H. Wiesenberger, W. Lengauer, and P. Ettmayer, Acta Muter. 1998,46,651.
9. J. D. Browne, P. R. Lidell, R. Street, and T. Mills, Phys. Stat. Solid, 1970, Al, 715.
10. T. B. Massalski (Ed.), Binary Alloy Phase Diagrams, 2, ASM International, Metals Park, OH,
1990.
11. K. Schwarz, CRC Crit. Rev. Solid State Muter. Sci. 1987, 13, 211.
12. A. Neckel, in The Physics and Chemistry of Carbides, Nitrides and Borides, R. Freer (Ed.),
Kluwer, Dordrecht, 1990, p. 485.
13. P. Blaha and K. Schwarz, Int. J. Quantum Chem. 1983,23, 1535.
14. A. N. Christensen and S. Fregerslev, Acta Chem. Scand. A , 1978, A31, 861.
15. R. Kieffer, H. Nowotny, P. Ettmayer, and M. Freudhofmeier, Monatsh. Chem. 1970,101,65.
16. H. Pastor, Muter. Sci. Eng. 1988, A105/106, 401.
17. J.-E. Sundgren and L. Hultman in: Materials and Processes f o r Surface and Interface Engineer-
ing, Y .Pauleau (Ed.), Kluwer, Dordrecht, 1995, p. 453.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 249
18. W. Schintlmeister and 0. Pacher, J . Vac. Sci. Technol. 1975, 12/4, 743.
19. S. Abisset and F. Maury, Surf. Coat. Technol. 1998, 108-109, 200.
20. W. D. Sproul, J . Vac. Sci. Technol. A , 1994, 12, 1594.
21. J. Smolik and K. Zdunek, Vacuum, 1998, 51/3, 441.
22. V. Craciun, D. Craciun, C. Ghica, L. Trupina, C. Flueraru, and N. Nastase, Appl. Surf: Sci.
1999, 138-139, 593.
23. K. Obata, K. Sugioka, H. Takai, and K. Midorikawa, Applied Surf. Sci. 1999, 138-139,
335.
24. A. J. Perry and J. R. Treglio, J . Vac. Sci. Technol. A , 1995, 13/3, 1067.
25. S. Xiao, C. P. Lungu, and 0. Takai, Thin Solid Films, 1998, 334, 173.
26. J. Magnan, G. C. Weatherly, and M.-C. Cheynet, Metall. Muter. Trans., A , 1999, 30A, 19.
27. C. Hu, H. Xin, L. M. Watson, and T. N. Baker, Acta Muter. 1997, 45/10, 431 1.
28. W. Lengauer, Talanta, 1991, 38, 659.
29. R. Tiubler, S. Binder, M. Groschner, W. Lengauer, and P. Ettmayer, Mikrochim. Acta I, 1992,
107, 337.
30. G. F. Bastin and H. J. M. Heijligers, J . Microsc. Spectrosc. Electron. 1986, 11, 215.
31. G. F. Bastin and H. J. M. Heijligers, Quantitative Electron Probe Microanalysis of Nitrogen,
University of Technology, Eindhoven, 1988.
32. W. Lengauer, J. Bauer, M. Bohn, H. Wiesenberger, and P. Ettmayer, Mikrochim. Acta, 1997,
126, 279.
33. K. Frisk, A Thermochemical Evaluation of the Cr-N, Fe-N, M e N and Cr-Mo-N Systems,
Report TRITA-MAC-0393, Materials Research Center, Royal Institute of Technology,
Stockholm, 1989.
34. V. N. Lipatnikov, W. Lengauer, P. Ettmayer, E. Keil, G. Groboth, and E. Kny, J . Alloys
Comp. 1997,261, 192.
35. R. Ward in M T P International Review of Science, Inorganic Chemistry Series One, D. W. A.
Sharp (Ed.), Butterworths, London 1972, p. 5.
36. H. Holleck, Binare und ternare Carbid- und Nitridsysteme der Ubergangsmetalle, Borntrager,
Berlin 1984.
37. H. Holleck, J . Vac. Sci. Technol. A , 1986, A6, 2661.
38. W. Lengauer, Acta Metall. Muter. 1991, 39, 2985.
39. K. Zheng and R. Schmid-Fetzer, Z . Metallk. 1996, 87, 540.
40. R. Juza, A. Rabenau, and I. Nitschke, Z. Anorg. Allgem. Chem. 1964,332, 1.
41. E. Rudy, Metall. Trans. 1970, 1, 1249.
42. 0. N. Carlsson, J. F. Smith, and R. H. Nafziger, Metall. Trans. A , 1986, A17, 1647.
43. W. Lengauer and P. Ettmayer, J . Less-Common Met. 1985, 109, 351.
44. W. Lengauer, J . Phys. Chem. Solids, 1988,49, 59.
45. N. Kieda, N. Mizutani, and M. Kato, J . Less-Common. Met. 1988, 144, 293.
46. C. Teichmann, W. Lengauer, P. Ettmayer, J. Bauer, and M. Bohn, Metall. Muter. Trans A ,
1997,28A, 837.
47. G. Brauer and W. Kern, Z. Anorg. Allgem. Chem. 1983,507, 127.
48. G. Brauer and W. Kern, Z. Anorg. Allgem. Chem. 1984,512, 7.
49. R. Berger, W. Lengauer, and P. Ettmayer, J . Alloys Comp. 1997, 259, L9.
50. M. Joguet, W. Lengauer, M. Bohn, and J. Bauer, J . Alloys Comp. 1998,269,233.
51. J. Gatterer, G. Dufek, P. Ettmayer, and R. Kieffer, Monatsh. Chem. 1975, 106, 1137.
52. G. Brauer, E. Mohr, A. Neuhaus, and A. Skokan, Monatsh. Chem. 1972, 103,794.
53. K. Frisk, J. Alloys. Comp. 1996, 278, 216.
54. W. Mayr, W. Lengauer, D. Rafaja, J. Bauer, and M. Bohn, J . Phase Equil. 1999, 20, 35.
55. P. Ettmayer and A. Vendl, in Science of Hard Materials, R. K. Viswanadham, D. J. Rowcliffe,
and J. Gurland (Eds.), Plenum, New York, 1983, p. 47.
56. A. Bezinge, K. Yvon, J. Muller, W. Lengauer, and P. Ettmayer, Solidstate Commun. 1987,63,
141.
57. S. L. Roberson, D. Finello, and R. F. Davis, Muter. Sci. Eng. A , 1998, 248, 198.
58. V. P. Anitha, A. Bhattacharya, N. G. Patil, and S. Major, Thin Solid Films, 1993, 236, 306.
59. W. Mayr, W. Lengauer, V. Buscaglia, J. Bauer, M. Bohn, and M. Fialin, J. Alloys Comp. 1997,
262-263, 521.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
250 7 Transition Metal Carbides, Nitrides, and Carbonitrides
143. Y. Gotoh, M. Nagao, T. Ura, H. Tsuji, and J. Ishikawa, Nucl. Znstr. Meth. Phys. Res. B, 1999,
B148, 925.
144. J. K. Gregory, W. Lengauer, and W. Mayr, in Surface Performance of Titanium, J. K.
Gregory, H. J. Rack, and D. Eylon (Eds.), TMS, Warrendale, PA, 1996, p. 65.
145. H. Xin, S. Mridha, and T. N. Baker, J . Muter. Sci. 1996, 31, 22.
146. W. Lengauer, L. Chen, J. Garcia, V. Ucakar, K. Dreyer, D. Kassel, and H. W. Daub,
Proc. PM2tec 99, Znt. Congress on Powder Metallurgy and Particulate Materials, Vancouver
1999.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
8.1 Introduction
Novel procedures continuously being introduced into the synthesis of new
materials have lead to the speculation that it could be possibie to produce a
material with hardness exceeding that of diamond, the hardest material presently
known to man. As the chemistry of diamond is based upon understanding the
structural bonding of carbon, a modification of known carbon structures through
application of high pressure or temperature would seem a promising route. Indeed
there has been much speculation about the hardness of carbon found in fullerene
complexes, yet production of ‘post-diamond’ phases has not met with any success,
at least through known synthesis procedures. Even so the commercial synthesis of
diamond (and more recently cubic boron-nitride) has lead to much speculation in
the field and with it enormous excitement as there is potential for new superhard
materials.
Properties of diamond, its hardness, unique electrical and thermal properties,
arise from a three-dimensional structure of very strong covalent bonds. Similar
properties are also associated with cubic boron-nitride. That other similar highly
covalent three-dimensional structures may exhibit similar properties to that of
diamond is highly likely. A possible material that first stimulated interest in such
structures was PSi3N4 [I], a hard material with quite interesting properties. A
possible replacement of one (or several) of the Si atoms with C would tentatively
suggest the synthesis of a material that is harder than PSi3N4 and even compete
with diamond. There is therefore a rich possibility of structures in the tertiary
Si,C,,N, system that afford unique potential for synthesis. But problems are
always encountered when dealing with nitrogen, the N2 dimer has a very strong
covalent bond and does not react with carbon readily when it competes with
either carbon or silicon in forming stable Si,C,,Nz structures. Graphite is very
likely more stable than diamond, and in a tertiary material like Si,C,,N, the forma-
tion of a SirC, structures is also possible. Quite clearly sophisticated techniques are
needed for a successful synthesis [2,3].
Given the cost and complexity of a commercial synthesis procedure, even when
identified, an unknown synthesis route requires considerable investment. For this
reason computer modeling procedures have been introduced that attempt to
predict new structures and examine phase transitions connecting them. Such
procedures are continuously being applied [4] and, bearing in mind that funda-
mentals of the modeling procedures are also being debated, indications in the
last few years have shown that reliable calculations of material properties can be
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
254 8 New Superhard Materials: Carbon and Silicon Nitrides
achieved at little expense. Liu and Cohen [5] applied computer modeling to
investigate a hypothetical phase of PC3N4 and suggested that such a material
would have a large binding energy and may be metastable under ambient condi-
tions. More importantly, the calculations implied that this material would have
a hardness comparable with (or even greater than) diamond. The computational
technique used had already been applied to diamond with considerable success
[6] and its reliability simulated enormous interest and effort in investigating
other materials.
This chapter examines some potential structures that have been proposed for
different phases of carbon-nitride and also the tertiary material silicoti-carbon-
nitride. First we briefly examine some basic modeling procedures that have been
employed to date when investigating these materials and then discuss properties
of the predicted structures.
D,a, re,, k, and Be, fitted to small molecular systems was applied to structures of
carbon-nitride by Guo and Goodard [7]. However, these results were in conflict
with many other calculations for carbon-nitrides highlighting that use of such a
simple interatomic interaction has danger when the potential is transferred to
materials with a far higher density. In the case of highly covalent materials like
diamond and (most likely) related superhard materials the potential is likely to be
far more complicated than used by Guo and Goodard. In order to study fracture
and related degradation processes in silicon nitride a similar effective potential
has been obtained [8] representing an average of the interactions. This allows for
simulations of up to lo6 effective atoms. However, the drawback of such an
approach is that it apparently disregards the possibility of atomic segregation.
One of the most promising analytic potentials that has been introduced in this
regard is the Tersoff potential [9] that has the form V ( r ) = f c ( r ) ( f R ( r )+ / f A ( v ) )
where f ( r ) and b are complicated functions containing angular as well as a
radial (v) dependences. For a typical binary solid some 24 empirical parameters
are now needed to adequately describe the potential, with considerable success
being reported for C and Si structures. Suitable parameters for the Tersoff
potential have now been obtained for cubic BN [lo] and recently for PSi3N4
[l 11 but earlier results [12] had indicated that parameters in the Tersoff potential
may not be unique. Further work is now being undertaken to try and find a
transferable Tersoff potential for the Si,C,,N, system, but in all likelihood such
a potential may have a limited range of transferability amongst closely related
structures.
Quite clearly the semi-empirical procedure is very dependent upon the choice of
parameters, although it has been possible to obtain sensible parameters from fixed
systems. The shortcoming with the approach is that the parameters may not be
transferable to other (especially unknown) situations. At the same time once a reli-
able potential is obtained there are many excellent molecular dynamics algorithms
available to compute the subsequent motion of the system quite efficiently and with
low computer cost.
rather than solid state concepts for computing electronic interactions it has proven
quite useful in examining various stoichiometric forms of carbon-nitride.
involved. Thus there are two ways of deducing the pressure for a phase transforma-
tion. Either from the slope at the common energy of E / V ,or by the intersection of
E/P.The former is specific in that it refers to distinct phases whereas the latter
makes no assumption of the path followed.
For example, if we look at the transformation from graphite into diamond we
calculate from E/V that a pressure of P = 80 GPa is needed to transform graphite
into diamond by overcoming the energy barrier. The commercial synthesis process
involves pressures far less than this, although high pressure treatment of graphite (or
CG0type material) leading to diamond is consistent with application of such large
pressures. However, the calculated relative energies of graphite and diamond are
very close to each other. This means that the pressure for the transformation
found from the E/P intersection is calculated to be very small, and this would be
the situation if the graphite to diamond transformation itself was continuous. Of
course the transformation process of graphite into diamond itself is not continuous
but rather involves changes from graphite into other (rhombohedral) graphitic
forms through to hexagonal diamond and then finally diamond. Chemical
influences (such as metal flux) as well as temperature factors are important in the
synthesis. Collectively these latter features are influencing the relative energetics
of the various phases.
Thus we must conclude that whereas the calculated pressure from E/P has some
guiding significance, the pressure obtained from E / V identifies a chosen synthesis
route. The pressure calculated in this way will be an upper value needed for the
transformation between the two phases. Accordingly, an in depth modeling of the
synthesis procedure needs an investigation of the stability of the intermediate
phases if they are known. Yet the important physical properties of the system are
essentially contained in relative values of the energies of the two phases and how
the energy changes under compression.
missing C atom in the C3N4layer suggests there could be several ways of stacking in
much the same way as, for example, graphitic BN [36]. A rhombohedra1 [37] form of
the material in graphitic form is predicted to be stable at a material density lower
than the 0-phase and more recently other graphitic forms with still lower density
have been suggested as being possible precursors to the formation of the superhard
phase [34,35]. High energy amorphous structures have also been investigated [14,38]
and these have shown competition between the C-C, C-N, and N-N bonds in
forming such a structure. Possible crystalline structures with stoichiometries different
from C3N4[6,39] have also been examined.
Figure :1 Some superhard crystalline phases of C3N4:(a) P-phase (P63/m),(b) cubic (P43m),(c)
cubic (1434 structures. Black spheres represent C atoms.
layers. This structure we shall name p-C3N4(odd-phase). Probably there are many
possibilities between. There is also the possibility that one of the planes may have no
missing C atoms with the adjacent plane having two missing C atoms: a C2N4-C4N4
structure. The three possibilities are shown in Fig. 3.
The other type of graphitic structure considered recently by Teter and Helmley [35]
also contains 14 atoms in the unit cell but is somewhat different to the p-C3N4 struc-
tures as here two possible C-N bond lengths are in the hexagonal plane. This structure
is shown in Fig. 4 and emphasizes C-N bonding between layers.
As the stability of the graphitic phases probably relates to the intralayer structure
than the interlayer spacing we investigate the layer charge density for the phases. In
Fig. 5 we show the charge density of the P63/m phase; rather surprisingly most of
the charge on N is directed away from the C-N bond toward the vacant site, as is the
case for some superhard structures.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8.3 Carbon Nitride 26 1
Figure 3. p-structures of hexagonal graphite C3N4: (a) odd-phase structure, (b) the even-phase
structure, and (c) the CZN4-C4N4planar structure. Carbon atoms are shown in black.
The low energy P63/m graphite phase appears to change to one of the higher
energy p-C3N4 graphite phases following compression well before the superhard
(3-phase is reached. However this p-C3N4 graphite phase seems slightly softer
than the P63/m phase and as such requires less pressure to surmount the energy
barrier that leads to the p-phase. If a graphitic phase of C3N4 can be established
it would therefore appear quite fruitful to use pressure (especially uniaxial) tech-
niques to change between the graphitic phases and possibly attain a superhard
phase. Of the graphite structures stacking involving C4N4-C2N4 sheets was
unlikely but each of the p-graphite phases were found quite stable and with relative
energies lying very close to each other. Forces between adjacent planes in the graphi-
tic structures are very small and so graphitic carbon nitride is extremely soft along
the c-axis. This weak interaction probably leads to dispersed layers of graphitic
sheets in the material as recently observed using high resolution spectroscopy of
carbon nitride thin films [43].
standard Tersoff parameters for C and for N those that give best results specifically
for the cubic form of C3N4,bearing in mind (as pointed out by Kroll [12]) that the
Tersoff potential has limitations in describing N-N bonds. The subsequent compu-
tational procedure was quite time consuming for convergence, this indicating that
the initial starting point was indeed a highly metastable configuration. Relaxation
was monitored throughout and we noted importance of the various bonding config-
urations, in particular the build up of C-N bonds and competition with C-C and N-
N bonds. Although the number of atoms in the simulation was small we did observe
that N-N bonds were often not connected into the complete network as much as C-
C bonds were. Instead the N atoms tended to group into relatively loosely bonded
structures often isolated from the main continuous network. This can be seen in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8.3 Carbon Nitride 263
Fig. 6 which shows the distribution of atoms in the amorphous 56-atom cell at a
volume very near the stable volume for such a structure.
This equilibrium volume corresponds to a material density of p = 2.9 g ~ m - still
~,
relatively high but far lower than the crystalline hard material p = 3.8 gcmP3.
Specifically at p = 2.9 gcmP3 there are a larger number of C-N bonds than either
C-C or N-N bonds and this is typical of most volumes considered up to a material
density of 4.5 g ~ m - Thus
~ . the modeling has shown that C-N bond formation is
preferred over others.
-220.4t 1
-221.0'. I
40 50
, ' '
60
'
vo~ume(a.u.~)
Figure 7. Energy-volume relationships for various phases of C3N4.
transition pressure necessary to transfer between the various graphite phases to be less
than 5 GPa: a uniaxial compression is far less than this. Hydrostatic pressures that lead
to the formation of the superhard P-phase calculated in the same way for graphite
phases however are much higher. For example again from the slope at the common
energy intersection we calculate that a pressure of about 80GPa is needed to go
from the P63/m phase to the superhard P-phase and a little less at about 55 GPa to
go from the odd-phase p-C3N4 phase and only slightly more for other graphitic
phases. Such transition pressures are well above the value of 12 GPa suggested recently
by Teter and Helmley [35] to go from P63/m graphite to the superhard j3-phase,
although the way in which this value was estimated is not clear. Very likely this
value was based upon a upon common energy/pressure arguments that neglect the
importance of intermediate changes in volume of the structures.
The importance of the local tetrahedral structure has also been noted in the
analysis of a lower density form of SiC2N4 compound recently produced by
Riedel and coworkers [45]. Here X-ray diffraction showed the possibility of an
almost cubic (Pn3m) phase with a unit cell of measured dimension a = 0.61885 nm.
In such a material the local Si related tetrahedral structure is realized as SiN4 units
that are connected through N=C=N chains. The flexibility of the N=C=N bond
has been considered an essential feature for more extended SiC2N4structures [45,51].
Following earlier theoretical results on fJC3N4,one of the hardest forms of
SiC2N4is expected to be fJSiC2N4.This is obtained by replacing two C atoms in
the 14-atom unit cell of fJC3N4with Si atoms maintaining lowest C-N bond lengths
[52]. Possibly this form can be attained through a synthesis starting from one of
structures discussed above.
8.4.1 PSiC2N4
In Fig. 8 we show the structure of fJSiC2N4as viewed along the c-axis of the
hexagonal unit cell with details of the structure being presented in Table 3 .
We see that the bulk modulus of SiC2N4is about $ that of C3N4.
In Fig. 9 we show the calculated charge density in one of the hexagonal planes.
There are two important features to be noted on comparing this with a similar
charge distribution for the fJC3N4shown in Fig. 1. First very little charge is located
about the Si atom, suggesting that Si bonding be insignificant compared with
the C=N bond. Second is the direction of the C=N bonding; if we look at the
N atom we see that the main charge is not directed along the C=N bond itself
but with N(2p) orbitals being directed perpendicular to the bond. Surprisingly the
C-N bond is very similar in both silicon-carbon-nitride and carbon nitride again
suggesting the importance of the C=N bond.
Si N C N Si
a
N
b Si C Si
N
N N
C Si C Si
Figure 10. Cubic phase of SiC2N4.Large open circles are Si, small black circles C and small open
circles N. The nature of the Si-N=C=N-Si bond is shown for each case.
essential building block and a linear Si-N=C=N-Si bridging unit. The measured
unit cell constant is 0.618nm, and Riedel et al. noted that this size of cell would
have a short interatomic spacing between the C and N atoms of 0.11 9 nm. They
therefore suggested that the measured cell constant may correspond to an effec-
tive value relating to a random distribution of N atoms about the Si-C-Si unit.
The linearity of the N=C=N bond can also be maintained in a similar symmetry
in the manner suggested in Fig. 10b and more recently Kroll et af. [51] have sug-
gested another structure for the cubic phase involving N=C=N bond with overall
P4n2 symmetry shown in Fig. 1Oc. The nonlinear C=N=C bond is expected to
lead to a softening of the structure [53]. As yet the several cubic structures
have not been related to the potentially superhard BSiC2N4 phase. We have
calculated the structure of the three new possible phases and the results are
given in Table 3.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
268 8 New Superhard Materials: Carbon and Silicon Nitrides
60 120
volurne(a.u.3)
Figure 11. Energy-volume behavior of different phases of SiC2N4:(a), (b), and (c) identify the
different phases shown in Figure 10. The pressure shown is obtained from the intersection point
of the c-phase and the P-phase.
8.5 Conclusions
There is understandably some pessimism regarding a commercial synthesis of
superhard carbon nitride especially in view of the intense effort that has been put
into its synthesis by many laboratories in the last few years with no firm result to
date. Yet the wisdom of the computational arguments upon which the various
structures of the material have been predicted seems firm and they have been
shown to be capable of giving us valuable insight in many other areas of materials
science. Experiences gained regarding the commercial synthesis of diamond and
cubic BN could probably lead ways that carbon nitride can be synthesized, yet
other more novel routes also cannot be excluded.
Acknowledgements
The National Research Foundation (South Africa) is thanked for its support.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 269
References
1. 0. Borgen and H. M. Seip, Actu. Chem. Scand. 1961, 15, 1789.
2. C. Niu, Y. Z. Lu, and C. M. Lieber, Science, 1993, 261, 334.
3. K. M. Yu, M. L. Cohen, E. E. Haller, W. L. Hansen, A. Y. Liu, and I. C. Wu, Phys. Rev. B,
1994,49, 5034.
4. M. L. Cohen and J. R. Chebkowsky, Electronic Structure and Optical Properties of Semiconduc-
tors, Springer Verlag, Germany, 1988.
5. A. Y. Liu and M. L. Cohen, Phys. Rev. B, 1990,41, 10727.
6. Y. Miyamoto, M. L. Cohen, and S. G. Louie, Solid State Commun. 1997, 102, 605.
7. Y. Guo and W. A. Goddard 111, Chem. Phys. Lett. 1995, 237, 72.
8. R. K. Kalia, A. Nakano, A. Omeltchenko, K. Tsurta, and P. Vashishta, Phys. Rev. Lett. 1997,
78, 2144.
9. J. Tersoff, Phys. Rev. Lett. 1988, 61, 632.
10. W. Sckkal, B. Bouhafs, H. Aourag, and M. Certier, J . Phys.: Condens. Mutter, 1988,10,4975.
11. F. de Brito Mota, J. F. Justo, and A. Fazzio, Phys. Rev. B, 1998, 58, 8323.
12. P. M. Kroll, PhD Thesis, University of Paderborn, 1996.
13. D. Porezag, Th. Frauenheim, Th. Kohler, G. Seifert, and R. Kaschner, Phys. Rev. B, 1995,51,
12947.
14. F. Weich, J. Widany, and Th. Frauenheim, Phys. Rev. Lett. 1997, 78, 3326.
15. R. M. Dreizler and E. K. U. Gross, Density Functional Theory, Springer Verlag, Berlin, 1990.
16. M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. Joanopoulos, Rev. Mod. Phys.
1992, 64, 1045.
17. D. M. Ceperley and B. J. Adler, Phys. Rev. B, 1980, 45, 566.
18. J. P. Perdew, Phys. Rev. B, 1986, 33, 8822.
19. M. Bockstede, A. Kley, J. Neugebauer, and M. Schemer, Comp. Phys. Comm. 1997,107, 187.
20. N. Troullier and J. L. Martins, Phys. Rev. B, 1991, 43, 1993.
21. F. D. Murnaghan, Proc. Nat. Acud. Sci. USA, 1994,3, 24.
22. F. Birch, J . Geophys. Res. 1952, 57, 227.
23. A. R. Merchant, D. G. McCulloch, D. R. McKenzie, Y. Yin, L. Hall, and E. G. Gerstner,
J. Appl. Phys. 1996, 79, 6914
24. M. Y. Ching, D. Li, X. Lin, V. P. Dravid, Y. Chung, M. Wong, and W. D. Sproul, J. Vuc.Sci.
Technol. 1993, 11, 521.
25. D. M. Bhusari, C. K. Chen, K. H. Chen, T. J. Chuang, L. C. Chen, and M. C. Lin, J. Muter.
Res. 1997, 12, 322.
26. J. Martin-Gill, F. Martin-Gill, M. Sarikaya, M. Qian, M. Jose-Yacaman, and A. Rubio, J . Appl.
Phys. 1997, 81,2555.
27. K. Yamamoto, Y. Koga, K. Yase, S. Fijiara, and M. Kubota, Jpn. J . Appl. Phys. 1997, 36,
L230.
28. Z. J. Zhang, S. Fan, and C. M. Lieber, Appl. Rev. Lett. 1995,66, 3582.
29. M. Tabbal, P. Merel, S. Moisa, M. Chaker, A. Ricard, and M. Moisan, Appl. Rev. Lett. 1996,
69, 1698.
30. Y. Chen, L. P. Guo, and E. G. Wang, Philos. Mag. Lett. 1997, 75, 155.
31. Kehui Wu, E. G. Wang, Jian Qing, and Guichang, Xu, J . Appl. Phys. 1998, 83, 1702.
32. C. Ronning, H. Feldermann, R. Merk, H. Hofsass, P. Reinke, and J. U. Thiele, Phys. Rev. B,
1998,58, 2207.
33. H. W. Lu, X. R. Zou, and J. Y. Feng, J . Phys. D: Appl. Phys. 1998,31,363.
34. J. Ortega and 0. F. Sankey, Phys. Rev. B, 1995, 51, 2624.
35. D. M. Teter and R. J. Hemley, Science, 1996, 271, 53.
36. J. Furthmuller, J. Hafner, and G. Kresse, Phys. Rev. B, 1994, 50, 15606.
37. A. M Liu and R. M. Wentzcovitch, Phys. Rev. B, 1994, 50, 10362.
38. J. E. Lowther, Phys. Rev. B, 1998, 57, 5724.
39. C. Cote and M. L. Cohen, Phys. Rev. B, 1997, 55, 5684.
40. A. Y. Liu and M. L. Cohen, Science, 1989, 245, 841.
41. J. V. Badding, Adv. Muter. 1997, 11, 877.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
270 8 New Superhard Materials: Carbon and Silicon Nitrides
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
9.1 Introduction
It has been speculated that there exists a multitude of metastable forms of carbon.
This arises from the preparation of macroscopic quantities of fullerenes [l, 21, the
observation of carbon nanotubes [3] and onions [4]. and the successes in carbon
network chemistry [5,6]. This is based on the well-known fact that carbon atoms
can frequently occur in different local environments.
In particular, carbon can form hybrid orbitals and is found to be either sp3
bonded and four-fold coordinated as in diamond, sp2 bonded and, hence, three-
fold coordinated as in graphite, or sp bonded and two-fold coordinated as in
acetylene. Whereas in diamond one finds o bonds as in silicon, in the latter two
forms the carbon atoms can lower their energy by creating additional 7c bonds to
their neighbors, that is they form double and triple bonds. In contrast to diamond,
materials based on these under-coordinated types of carbon tend to form layered or
chain-like structures which are usually less dense than graphite. Diamond forms
very strong and short bonds and is, therefore, a promising material for applications
in harsh environments. The intraplane bonding in graphite is very strong but the
weaker interplanar forces prevent graphite from being well protected against
mechanical or chemical attack. Carbon atoms can develop intermediate hybridiza-
tion states dependent on their local neighborhood as in the various amorphous
phases that have been prepared during the past decades.
Semiconductor devices based upon carbon would be very attractive when taking
advantage of the short and stiff carbon-carbon bonds in combination with the light-
ness of the carbon atoms since this would result in a material with a thermal conduc-
tivity almost as good as in diamond, which is stable against thermal degradation. A
large band gap and a high saturation velocity for electrons as in diamond [7,8]
would make it a superior material for superfast, high-power, high-temperature,
and wide band gap applications.
Unfortunately, diamond itself, though intensively studied, has been found to be
less suitable for device applications. Although it can be doped by boron to make it
a p-type semiconductor it is almost impossible to obtain the n-type counterpart
for a p-n junction. Diamond cannot be doped by the most likely candidate, nitrogen,
as a substitutional impurity. Instead nitrogen moves off its lattice site causing a dan-
gling bond on one of its carbon neighbors which in turn results in a deep level [9-121.
Other impurities such as phosphorus [16-181 or lithium [19] have been tried for n-
type doping in diamond but their size reduces the solubility in diamond to near
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
272 9 Effective Doping in Novel sp2 Bonded Carbon Allotropes
zero [9]. The most promising alternative, P, results in an electrical activity much less
than would be expected from its concentration so that the doping efficiency is rather
low [17]. Moreover, it has been found that the phosphorus-vacancy complex [20]
is more stable than substitutional P. This complex traps injected electrons and
deactivates the impurities.
In amorphous carbon (a-C) substitutional impurities can relax off their sites as in
diamond or cause substantial rearrangements in the carbon host. Injected holes or
electrons migrate in a-C to occupied high energy n: or low energy unoccupied n*
levels, respectively. Associated n-bonded clusters relax to screen the additional
charge [12-141 and the conductivity can only occur through variable range hopping
between such clusters.
In the light of these findings, in a recent paper [ 151 the doping problem in carbon
has been attacked from a different point of view. Instead of searching for other atom
types or complexes that effectively would act as shallow donors metastable carbon
modifications which accept both boron and nitrogen as substitutional impurities
have been proposed to exist.
Among the hypothetical forms of carbon predicted prior to this work are stacked
layers made of double and acetylene triple bonds [21] yielding structures like
‘graphyne’ and related polymeric networks. These planar configurations were
predicted to have quite large electronic gaps but also found to have a lower
number of atoms per unit layer area than graphite. Equivalent building blocks
were used to create polyines in quasi-one-dimensional frameworks [22] or ‘super-
diamond’ and related structures [5] in three-dimensionally connected networks.
Another class of networks has been proposed by combining sp2 atoms in a non-
layered arrangement. The idea was to find a material based upon the thermo-
dynamically most stable short-range order but with properties close to diamond.
In order to make use of the short bonding between two graphitic carbon atoms to
obtain the highest bulk moduli, Hoffman et al. [23] predicted the particular structure
which they named BCT4 to be metastable. Tamor and Hass proposed that a very
similar structure called H6 possibly exists [24]. Though the latter is topologically
directly related to diamond it has been found to be rather unstable unlike the metal-
lic phase BCT4 [25,26] and converts without a pronounced barrier back to
diamond [27]. The most remarkable feature of these structures, as well as another
energetically unfavorable structure [28] (BCC4 in the following), is that they do
not contain any six-membered rings as is typical for graphite.
Other purely three-fold coordinated networks have been proposed inspired by
tesselations of infinite triply periodic minimal surfaces [28-321. All of these
structures are composed of five-, six-, and seven-membered rings which are the
major components in amorphous structures too. They establish the whole family
of negatively curved graphitic sheets. The density of these structures is usually
lower than that of graphite but they have been predicted to be more stable than
fullerene c 6 0 [32,33]. Electronically the proposed structures behave like either
metals or insulators with gaps smaller than that of f.c.c. c 6 0 [34]. An exception
with respect to this is polybenzene [28] which has been found to be a low
energy form of carbon that opens an indirect band gap of 2.96 eV [34]. This excep-
tionally large gap has been attributed to the very special topology which consists of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
9.1 Introduction 273
(a,
centered tetragonal symmetry (I4lmd) with the generating atom at the position
i, 0). For the idealized structure with equal bond lengths and angles the lattice
constants are given by a = b = 4d and c = J8d, where d is the common bond
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
9.2 Lattice Description 275
Figure 1. Perspective views of the wide band gap models showing the three-dimensionally connected
networks.
bonded. Here, however, the projection of the axis onto a plane perpendicular to it
is a square, whereas in the former case it is a triangle. SC24 does not contain such
a feature.
BCT8 and R6 both contain a rigid multichannel structure which may be promis-
ing for applications where anisotropic transport at a microscopic level is of interest.
Such properties have been predicted for hypothetical crystals made of carbon
nanotubes [57]. However, synthesizing such crystals requires one to prepare,
purify, and align carbon tubes of equal size [58].
Finally, we investigate the doping behaviour of a prototype zeolite structure
named melanophlogite, which has been detected to be one of the lowest energy
forms of this kind [46]. The space group of this structure is P m n . The simple
cubic system contains 46 atoms in its primitive cell and we will refer to it as
SC46. The model used throughout these studies is shown in Fig. l(d).
Table 1. DF-TB ground state properties of hypothetical model structures compared with the results
for natural carbon allotropes and recent calculations with more sophisticated methods.
H6 and BCT4 both have a bulk modulus which, within the tight-binding model is
comparable in magnitude to that of diamond. In accord with an empirical TB
scheme [24] H6 seems to be even harder than diamond while SCF plane wave
calculations [25,26] predict that the almost equal bulk moduli of BCT4 and
H6 are 17% below the diamond value. The bulk moduli of the models have been
determined by calculating the elastic compliances after applying suitable strains
to the crystals and inversion of the volume compressibility [68].
A four-fold coordinated structure obtained by decoration of melanophlogite with
carbon has been determined to be one of the two lowest energy forms of zeolite type
hypothetical carbon allotropes [46]. Within a converged SCF plane wave calculation
the total energy of this structure was determined to exceed that of diamond by only
0.09 eV/atom. The band gap was found to be almost 4 eV and only 6% lower than
that of diamond utilizing the local density approximation. We find the structure to
be 0.12 eV/atom less stable than diamond with a band gap which is about 4.4 eV and
68% that of diamond.
Table 2. Electronic properties of the hypothetical structures compared with diamond. sp indicates
minimal, spd extended basis set, see text. B-C and N - C mark the HOMO participation of the atomic
orbitals at the impurity and the nearest carbon neighbors.
Structure Gap width [eV] B acceptor level [ev] N donor level [eV] A 0 contribution to HOMO
1.0
0.8
0.6
0.4
0.2
0.0
0.8
0.6
8n
0.4
0.2
"1" 0.0
0.8
0.6
0.4
0.2
0.0
-25 -20 -15 -10 -5 0 5 -20 -15 -10 -5 0
Figure 2. The (broadened) electronic densities of states for BCTB, R6, SC24, and SC46 show the
wide band gap character of these materials. Diamond and graphite DOS are plotted for comparison.
The most remarkable feature of the structures presented is that they can take up
boron and nitrogen at substitutional lattice sites without a strong relaxation. In
recent DF-TB studies of the doping effect in diamond [12] we could reproduce
the experimental findings that boron yields a shallow acceptor level and nitrogen
a deep donor level which is impractical for electronic applications. Whereas
boron relaxes only a little, nitrogen moves substantially off the substitutional lattice
site. A lone pair electron and a dangling state at one of the carbon neighbors is
created. The carbon defect lies deep in the gap at 3 eV below the conduction band
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
9.5 Electronic Properties 28 1
and pins the Fermi level. A Boron atom in contrast stays close to its lattice site. The
associated acceptor level is about 0.1 eV above the valence band and pins the Fermi
level. Both values are close to the experimental activation energies of about 2.00 eV
for the single nitrogen donor and 0.37 eV for the boron acceptor [8].
Testing zeolite structures as possible hosts for boron and nitrogen is more
complicated because the impurity can be placed on different lattice sites. We have
checked the three unique places in the melanophlogite prototype structure and
present the results for the best bonding situation found for either boron or nitrogen.
We only observed shallow acceptor (activation energy = 0.32 eV) and donor levels
(activation energy = 0.17eV) in the pure sp3 environment for SC46 when the local
symmetries forced all four carbon neighbors of boron or nitrogen to relax away by
exactly the same distance such that the B-N or C-N bonds had the same length.
This caused degenerate levels at the Fermi energy to appear. Jahn-Teller distortions
removed these degeneracies and yielded lower energy structures.
Generally, all the ground states found were characterized by a off lattice site
relaxation of the impurity. A bond to one of the carbon neighbors was stretched
and became weaker than the others. In the nitrogen case this bond finally was
broken due to the electrostatic interaction between the nitrogen lone pair orbital
and the carbon dangling bond leaving an occupied deep carbon state. Therefore,
it seems to be fairly general that n-type doping by nitrogen is inhibited in a
carbon sp3 environment. Significant lattice relaxations to screen the injected
charge plus the creation of too deep a carbon level seem to be always favored
over establishing a real donor state.
This is in sharp contrast to the hypothetical carbon crystals BCT8 and R6. In
these structures the nitrogen atom is naturally incorporated at a lattice site in a
favorable local geometry. It must take part in the two single and the isolated
double bond associated with each site. The lattice relaxation is extremely weak
and N stays at the lattice site as does a substitutional boron atom. The highest
occupied molecular orbital (HOMO) is occupied by just one electron and acts as
a donor and acceptor, respectively. The activation energies determined within
DF-TB are summarized in Table 2. The HOMO is primarily located at the
substitutional atom and spread over its neighbors as can be deduced from the
participation analysis, see Table 2. Note, that the numbers given are the largest
contributions found. The remaining parts are almost equally distributed over all
the other atomic orbitals in the system. In the case of diamond and the zeolites
the donor level participation ratio is very large indicating that the highest molecular
orbital is strongly localized in the neighborhood of the impurity.
Amorphous carbon materials consist of a mixture of sp3- and sp2-bonded carbon
atoms and the latter form more or less extended n-bonded subclusters in a rigid
single-bond environment [63]. Considering this structure a similar doping effect as
discussed above might be expected. However, the size distribution of x-
clusters [69] and the associated strain in the systems yield broad bond length distri-
bution functions [70]. As a result, the n- and rc*-bandsare broadened and the tails of
the bands overlap with potential acceptor and donor levels. This causes electrons to
be either trapped from the highest energy x-levels or injected into lowest energy x*-
bands, repectively. It is accompanied with relaxation processes of the carbon
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
282 9 Effective Doping in Novel sp2 Bonded Carbon Allotropes
network in order to saturate carbon defect states which are always present in
amorphous carbon. Hence, doping is very inefficient in full contrast to the situation
in the proposed crystalline allotropes where the injected hole or electron cannot
migrate to distant n-states.
Finally, we carefully checked the doping results obtained within the DF-TB
scheme against a parameter-free self-consistent-field method employing norm-
conserving pseudopotentials to model atom cores and a localized real-space basis
set of Gaussian orbitals to expand the molecular wave functions and the charge
density [7 11. This computational scheme invokes the local spin density approxima-
tion for exchange and correlation terms within the density functional theory to solve
the many-body Schrodinger equation. The technique has been successfully applied
to study diamond growth [72] reactions, graphitization tendencies at diamond
surfaces [73] and the auto-catalysis of Stone-Wales transformations in fullerene
C60[74]. However, the computational efforts within this code forced us to represent
the new crystals by relatively small clusters. We cut out representative pieces of the
structures including two screw axes and three double bonds in a row interconnecting
them. The cut was done to ensure that only (J dangling bonds occured on the outside
of the clusters which were saturated by hydrogen. The hydrogen atoms then were
kept fixed in order to simulate the crystal boundary. The carbon atoms and the
substitutional impurity could fully relax within this hydrogen cage by a conjugate
gradient algorithm.
The cluster representing BCT8 was found to have an electronic gap of 3.9 eV and
the one representing R6 had a gap of 2.7 eV. The energy differences of the acceptor
and donor levels to the respective band edges calculated within this technique are
0.63eV and 0.67eV in the case of BCT8 and 1.37eV and 0.38eV in the case of
R6. The activation energies tend to be a little higher than determined within DF-
TB. However, this is not surprising taking into consideration that a hydrocarbon
molecule with an artificially strained boundary is used to model the extended crystal.
In the molecules, the central double bond is elongated compared to the outer
two already without any impurity. Nevertheless, the calculations prove that the
relaxation of the impurities is very small so that they do not remarkably move off
the substitutional lattice site. Except for the boron acceptor in R6 we find shallow
levels close to states representing the valence or conduction band edges, respectively.
The differences compared to the periodic DF-TB calculation are enhanced owing to
charge transfer effects to the surface of the molecule in particular into constraint
bonds.
9.6 Conclusions
We could show that it is possible to get n-type doping effects in carbon based
materials using nitrogen as a dopant. The structures where this works, however,
are purely sp2 bonded low energy forms of carbon. The semiconducting allotropes
are characterized by unique screw axes of single bonded sp2 carbon atoms held
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 283
Acknowledgments
We gratefully acknowledge support from the Deutsche Forschungsgemeinschaft
and the EPSRC.
References
1. H. W. Kroto, J . R. Heath, S. C. O’Brien, R. F. Curl, and R. E. Smalley, Nature, 1985,318, 162.
2. W. Kratschmer, L. D. Lamb, K. Fostiropoulos, and D. R. Huffman, Nature, 1990, 347, 6291.
3. S. Iijima, Nature, 1991, 354, 56.
4. D. Ugarte, Europhys. Lett. 1993, 22, 45.
5. F. Diederich and Y. Rubin, Angew. Chem.: Internat. Edn. 1992, 104, 1123.
6. Z. Xu and J. S. Moore, Angew. Chem.: Internat. Edn. 1993, 105, 1394.
7. A. T. Collins, Muter. Sci. Eng. B, 1992, 11, 257.
8. 0.Madelung (Ed.), Semiconductors: Group IV Elements and III- V Compounds, Springer, Berlin,
1991.
9. S. A. Kajihara, A. Antonelli, J. Bernholc, and R. Car, Phys. Rev. Lett. 1991, 66, 2010.
10. C. Cunha, S. Canuto, and A. Fazzio, Phys. Rev. B, 1993,48, 17806.
11. S. J. Breuer and P. R. Briddon, Phys. Rev. B, 1996, 53, 7819.
12. P. K. Sitch, Th. Kohler, G. Jungnickel, D. Porezag, and Th. Frauenheim, Solid State Commun.
1996, 100, 549.
13. P. Stumm, D. A. Drabold, and P. A. Fedders, J . Appl. Phys. 1997, 81, 773.
14. S. Poykko, M. Kaukonen, M. J. Puska, and R. M. Nieminen, Comp. Muter. Sci. 1998,10,351.
15. G. Jungnickel, P. K. Sitch, Th. Frauenheim, B. R. Eggen, M. I. Heggie, C . D. Latham, and
C. S. G. Cousins, Phys. Rev. B, 1998,57, R661.
16. J. F. Prins, Diamond Rel. Muter. 1995, 4, 580.
17. H. Spicka, M. Griesser, H. Hutter, M. Grasserbauer, S. Bohr, R. Haubner, and B. Lux,
Diamond Rel. Muter. 1996, 5, 383.
18. S. Koizumi, M. Kamo, Y. Sato, S. Mitd, A. Sawabe, A. Reznik, C. Uzan-Saguy, and R. Kalish,
Diamond Rel. Muter. 1998, 7, 540.
19. S. Prawer, Diamond Re/. Muter. 1995, 4, 684.
20. R. Jones, J. E. Lowther, and J. Goss, Appl. Phys. Lett. 1996, 69, 2489.
21. R. H. Baughman, H. Eckhardt, and M. Kertesz, J . Chem. Phys. 1987, 87, 6687.
22. U. H. F. Bunz, Angew. Chem.: Internat. Edn. 1994, 106, 1127.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
284 9 Eflective Doping in Novel sp2 Bonded Carbon Allotropes
66. H.-D. Beckhaus, S. Verevkin, C. Ruchardt, F. Diederich, C. Thilgen, H.-U. ter Meer, H. Mohn,
and W. Miiller, Angew. Chem.: Internat. Edn. 1994, 33, 996.
67. M. T. Yin and M. L. Cohen, Phys. Rev. B, 1984, 29, 6996.
68. J. F. Nye, Physical Properties ojCrystu/s, Clarendon, Oxford, 1985.
69. U. Stephan, Th. Frauenheim, P. Blaudeck, and G. Jungnickel, Phys. Rev. B, 1994, 50, 1489.
70. G. Jungnickel, Th. Kohler, M. Haase, S. Deutschmann, U. Stephan, P. Blaudeck, and
Th. Frauenheim, J . Non-Cryst. Sol. 1995, 183, 161.
71. R. Jone, and A. Sayyash, J. Phys. C: Solid State Phys. 1986, 19, L653.
72. C. D. Latham, M. I. Heggie, R. Jone, and P. R. Briddon, Diamond Rel. Muter. 1994, 3, 1370.
73. C. D. Latham and M. I. Hegaie. Diamond Re/. Muter. 1995. 4, 528.
74. B. R. Eggen, M. I. Heggie, EyJungnickel, C. D. Latham, R. Jone, and P. R. Briddon, Science,
1996. 272. 87.
75. A. Y. Liu; private communication.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Synthesis and Processing
Part I1
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
and geometry of the preform), and the termination of the reaction (by a calcium
containing ‘poisoning’ agent, used to coat the external surface of the preform), as
well as the actual DMO process conditions (temperature and oxygen partial
pressure).
Subsequent work extended the DMO concept to the gaseous nitridation of
liquid aluminum to form AlN, as well as to a number of boride and carbide products
(especially titanium and zirconium) formed by preform infiltration [2,6]. At the time
of writing (1998) it is not at all clear what the commercial future of these processes
will be. It should be noted that there is a wide range of ‘ceramic’processing technol-
ogies under development, all of which are based on chemical reactions occurring in
combinations of solid, liquid, and gaseous phases. Concrete is an unconventional
example, not usually classed as a ceramic. More conventionally, reaction bonded
silicon nitride (RBSN - gaseous nitridation of a silicon preform), reaction sintered
silicon carbide (the REFEL process - molten silicon infiltrated into a carbon-
containing silicon carbide preform), reaction bonded aluminum oxide (RBAO -
gaseous oxidation of an aluminum-containing alumina preform), and boron carbide
reaction bonded by infiltrating a molten aluminum alloy, all rely on a chemical
reaction to bond the ceramic product into a shaped component.
The DMO process produces a near net-shape product without the need for any
externally applied pressure, and is (in principle) capable of being used for the
economic manufacture of very large ceramic components. It is also capable of yield-
ing a ceramic composite product with some very desirable engineering properties.
Despite the high expectations attached to the process and the extensive investment,
especially by the American government but also by private industry, it is still not
clear which areas of application are going to prove economically viable in the
long term.
of the oxidation process at elevated temperatures has been attributed [9] to the
build-up of biaxial residual stresses which inhibit further diffusion. The presence
of moisture initially increases the rate of oxidation due to the development of hydro-
gen blisters. Only when the metal is in the liquid state (700-1 100°C) does a-alumina
form directly [lo-121. The influence of magnesium additions has been summarized
by Lea and Ball [13] and Field [14]. The first authors [13] showed that y-alumina for-
mation is accelerated at the Al-amorphous alumina interface, as expected since
magnesium cations are incorporated in the spinel structure.
Subsequently, magnesium diffuses rapidly to the surface along grain boundaries
to form MgO, with concurrent crystallization of the amorphous phase to y-alumina.
Finally, the MgO layer becomes continuous so that the rapidly diffusing magnesium
no longer has direct access to external oxygen and reacts instead with y-alumina to
form MgA1204, releasing aluminum metal particles which are embedded in the
spinel. By contrast, Field [14] detected no spinel but rather direct oxidation of
magnesium to primary MgO below a cracked, amorphous surface oxide layer.
This was followed by secondary precipitation of MgO within the amorphous
alumina, which was ascribed to reduction of alumina by magnesium.
The formation of MgO on alloys containing as little as 0.5% Mg [15] at =40O0C
has been confirmed, while spinel and reduced, aluminum metal-rich particles are
frequently observed in the oxide film [16]. Additions of copper and silicon inhibit
crystallization [ 171, possibly by keeping magnesium in solution. Strong segregation
of magnesium at grain boundaries [18-201 and surfaces [20] was cited as the
probable cause for formation of MgO in dilute alloys [18,19,21]. In situ microscopy
[22] has revealed MgO nucleating at grain boundaries in Al-Mg-Zn, suggesting that
segregation could stabilize MgO and rapid grain boundary diffusion could maintain
a flux of magnesium for further growth.
Experiments on liquid alloys of A1-Mg [23] in the range 650-900°C have
demonstrated that a duplex layer of MgO and spinel is formed initially, the
former by direct oxidation at the free surface and the latter by reaction of MgO
with the liquid alloy. The proportion of MgO in the duplex layer peaked (after
2000 s at 800°C) and was gradually replaced by MgA1204,with reducing magnesium
content in the alloy. Measurements of the strength of the oxide film have implied [24]
that additions of Ca, Li, and Na initially increase the oxidation rate, while Be, in
quantities as small as 0.1 YO,considerably reduces the rate of oxidation. Zinc
additions to A1-Mg alloys [25] were found to greatly increase the oxidation rate
at very low concentrations (0.2%). In one elegant study of diffusion couples between
Al-Mg and alumina/spinel/magnesia [26] it was found that reduction of spinel to
alumina by aluminum was slow, while an A1-3.6Mg alloy readily reacted with
MgO and alumina to form spinel. In particular, silicon contamination was
found to promote alloy wetting of the oxide and to render the interfacial oxide
less protective.
The relevance of these results for the DMO process can be summarized by three
points: The importance of MgO as an ubiquitous surface layer that promotes rapid
oxidations; the importance of limiting the amount of magnesium in the alloy so that
the oxide layer, through which the melt migrates, is not too thick; and the effects of
Si, Zn, and Li in promoting the DMO process.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.3 Related Ceramic Processing Routes 293
with effectively zero residual porosity. This process (patented as the REFEL process
in the western world) has been somewhat neglected in recent years, and the heat-
treatments developed to tailor high performance, high toughness microstructures
in Sic have yet to be applied to RSSC.
Infiltration of alumina preforms by aluminum alloy melts has also been achieved,
both by applying external pressure (in the absence of adequate wetting and capillary
absorption), and by ensuring the presence of wetting aids in the melt (most notably
magnesium, as in the DMO process). This infiltration processing route has not yet
been exploited commercially. Reactive wetting by an aluminum alloy melt is also the
basis of the process developed for boron carbide and marketed by the Dow
Corporation. In this case a B4C powder preform is reacted with an aluminum
melt at elevated temperatures. The molten aluminum is sucked into the pore
channels in the preform by capillary forces and reacts to form a range of mixed
aluminum borocarbides. The final ceramic product is essentially pore-free, but
contains a network of residual aluminum alloy. Since the ceramic phases are contin-
uous, the product is hard, but the residual aluminum phase is also interconnected
(while the volume fraction of residual alloy is to some extent controllable) so that
crack propagation is significantly inhibited. As we shall see below, both the micro-
structure and the mechanical behavior of DMO products are similar to those
obtained by reactive infiltration.
The combination of infiltration and reaction that characterizes DMO has been
exploited to make a number of composites. As long ago as 1953, it was shown
that silica containing refractories were reduced by molten aluminum to form
alumina and silicon [4]. Subsequently [27], the displacement reaction was extended
to the formation of composites of alumina with residual Al-Si. More recently, the
A1-Si02 displacement reaction has been used in the infiltration of dense preforms of
silica [28] and mullite [29,30] by molten aluminum. Extension of the reactive
infiltration process to porous silica-containing preforms [31,321 has resulted in
the fabrication of metal-matrix composites in which the silica was replaced by a
mixture of about 65% alumina and 35% metal, while the pores were infiltrated
by molten alloy. In contrast to DMO, the displacement reaction appears to proceed
at a critical temperature of X I 100-1200°C and without the need for a volatile solute
element or oxygen. Borosilicate glass has also been used as an initiator to enable the
infiltration of Al-Si alloys into alumina preforms [33].
Closer in concept to the DMO process is the infiltration of aluminum alloys in
nitrogen to yield A1-A1N composites. Low temperatures (< 1000°C) and high
magnesium or strontium content promote the spontaneous infiltration of liquid
metal with a small concurrent nitridation to yield dispersions of A1N [34]. Similar
alloys may be infiltrated at higher temperatures, resulting in nitride contents that
increase with temperature to yield A1N-matrix composites [35-381. This process
can result in particulate loadings of up to 75%. The mechanical properties of
these aluminum nitride composites have been extensively characterized [39].
A further extension of the DMO process has been in the fabrication of silicon
nitride and titanium nitride matrix composites by nitridation of the appropriate
molten alloys [6,40], but it is unclear whether these processes will have commercial
applications. Other examples of reactive infiltration include aluminum into Si3N4to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 Directed Metal Oxidation Incubation 295
yield an A1N-Si3N4-Al-Si composite [41] and zirconium into B4C to yield ZrB2-
ZrC-Zr [4244].
T=l100"C
A'2°3
\
\
\ iquid)
4 4 6 (0.40wt.%)
~ ~ ~ 0 . at.'IoMg
Figure 1. Ternary isothermal section for the A1-Mg-0 system at 1100°C [47].
T = 1100°C
.23
I I I I I I I I I 1-36
0.2 0.4 0.6 0.8
Al Mg
Figure 2. Stability fields for the alloy and corresponding oxides as a function of oxygen partial
pressure at 1100°C in the system A1-Mg-0 [47].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
I .4 Directed Metal Oxidation Incubation 297
I
I
‘I
Time
Figure 3. Schematic of weight gain (TG) and corresponding rate of weight gain (DTG) curves for
oxidation of a binary A1-Mg alloy in the temperature range 1050-1250°C[51].
oxygen, the volume of spinel formed should be less than the equivalent volume of
periclase, suggesting that the remaining volume could persist as liquid metal-filled
channels in the spinel which maintain contact with the MgO. It has also been
shown that mechanical disruption of the surface [48] can induce preferential for-
mation of spinel, although this does not appear to be a prerequisite for the DMO
process [5 11. A classical incubation period, characterized by extremely low rates
of weight gain, is then observed as the temperature increases (Fig. 3). The rate of
spinel formation drops as the spinel zone develops a duplex structure characterized
by a dense, metal-free layer at the interface with MgO but containing microchannels
of alloy near the melt [48,51]. It has been established that the true external surface is
composed of MgO, by using low voltage X-ray spectroscopy in the scanning elec-
tron microscope [51], and this has been attributed to de-mixing of cations in
spinel in the presence of an oxygen potential gradient [53]. These microstructural
changes have been attributed to the diffusion-limited supply of magnesium from
the melt reservoir which prevents the continued formation of MgO. Direct oxidation
of the alloy melt then leads to sealing of the metal channels by spinel, preventing
contact between the melt and the surface MgO.
It is worth emphasizing the change in the oxidation maxima, and the subsequent
drop in oxidation rate between the ‘low’ and ‘high’ temperature regimes. In the low
temperature regime, below 950°C for A1-3 Mg, magnesium is completely depleted
from the alloy reservoir and the oxidation product shifts from MgO to spinel,
though never to alumina [23,51]. In the high temperature regime, the incubation
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
298 I Directed Metal Oxidation
period is characterized by bulk magnesium levels that are only marginally reduced,
together with a duplex surface oxide as described above. The temperature for the
onset of the spinel ‘peak’ increases with both magnesium content and heating rate
[46,51], as would be expected from both the phase equilibria and the activation
energy for spinel formation.
The end of the incubation period [49,5 11 is marked by a reversal of the phenom-
ena that led to the low incubation rate: Alloy channels that were isolated from the
surface, reconnect and advance through the spinel layer. The mechanism for regen-
eration of the liquid metal channels seems to be associated with an increase in melt
oxygen activity in equilibrium with spinel when the magnesium concentration
decreases. Hence magnesium depletion at the surface, due to the rapid formation
of MgO, should promote spinel dissolution and promote capillary wicking of
molten alloy through the spinel to the free surface. This hypothesis is supported
by the reduced incubation time for ternary alloys that contain silicon, since silicon
is known to reduce the activity of magnesium in the melt and should promote spinel
dissolution [54]. Cracking of the surface oxide film could also lead to direct exposure
of the melt to the gas phase [48], but this does not appear to be a prerequisite for
terminating the incubation stage. A simple thermodynamic criterion for the
molten alloy to repenetrate the channels is that the liquid-spinel interfacial energy
should be less than one half the grain-boundary energy, and this criterion may only
be satisfied above some minimum temperature [ 1,491. This could explain why alloy
repenetration does not begin at a temperature of 900°C for an A1-3Mg alloy,
although this temperature is past the peak rate of spinel formation and well into
the incubation regime.
When the alloy in the channels penetrates to the spinel-magnesia interface near
the free surface, their composition should be in equilibrium with spinel, leading to
the formation of spinel nodules [49,51,55]. Spinel-alloy composite growth should
then continue until limited by further magnesium solute depletion to below the
+
concentration defined by the alloy + spinel alumina ternary equilibrium. At this
juncture, the spinel nodules should be replaced by alumina nodules, and the alloy
composition in the advancing composite should be given by that determined by
the spinel-alumina buffer layer at the original melt surface. The extent of spinel
growth which occurs before the A1203-A1 composite begins to form increases
with the magnesium content of the alloy and, for a fixed magnesium composition,
decreases as the temperature increases, as predicted from the ternary Al-Mg-0
equilibrium. For example, an A1-8 Mg alloy has been observed to grow a spinel-
based composite to a thickness of hundreds of microns at 1100°C until the alloy
magnesium concentration reaches =0.5%, at which point alumina is the stable
oxide. In contrast, an A1-3 Mg alloy was depleted to a similar critical magnesium
concentration after ~ 1 0 pm0 of spinel growth [51].
The oxide at the external surface of the composite is predominantly a thin
(0.1 pm) layer of MgO with an intermittent, underlying spinel layer: The microstruc-
ture is discussed below. It could be argued that the entire process of incubation and
reinitiation is caused by excess magnesium, and that a starting composition near the
three-phase equilibrium should trigger immediate growth of alumina. However,
controlled growth of a DMO composite has never been reported from a bulk
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 Directed Metal Oxidation Incubation 299
The mean intercept lengths of the ceramic and alloy channels, measured perpen-
dicular to the growth direction, are about 5 pm and 2 pm, respectively. These
intercept lengths are not related to the grain sizes, and both phases are mono-
crystalline over distances hundreds of times larger than the measured intercept
lengths. In effect, the liquid metal in the composite develops a grain size upon
solidification which is similar to that seen in a slowly cooled, cast aluminum alloy.
A typical growth history is shown schematically in Fig. 3. Following the
incubation period (when present) the growth rate rises to a maximum. During
this period, the oxide nodules nucleated during incubation (initially spinel and
subsequently alumina) grow radially from their point of nucleation, coalesce and
cover the surface of the melt. Within each nodule the orientation of the alumina
crystal lattice is approximately constant, with the c-axis aligned along the
growth direction [61]. The alumina grain size perpendicular to the growth direction
is therefore determined by the nodule spacing. As the nodules coalesce to form a
planar growth front the growth rate begins to drop monotonically with time
and (equivalently) composite thickness [46,62]. Occasionally, at high temperatures
or with small samples, growth acceleration has been observed and correlated,
through differential thermal analysis, with adiabatic heating due to the exothermic
nature of the oxidation reaction [46]. Frequently, oscillations in the growth rate
are superimposed upon the average, decreasing rate [47,63] and have been
associated with the passivation of old nodules and nucleation of new growth
centers.
Microstructural variations that accompany growth are of two kinds. Firstly, the
scale of the microstructure is finer near the oxidation front and coarser at the
original initiating surface [64,65], corresponding to the nucleation of new nodules
during the growth process. Secondly, oscillations in the growth rate are often
reflected in microstructural banding in which the residual metal content of the
layers differs slightly. At higher growth temperatures, above 1300°C for A1-3 Mg,
growth may cease abruptly after 1-2 mm of composite has formed.
The incubation period is terminated by re-penetration of molten alloy through the
passivating spinel layer and the nucleation of spinel nodules. Selective oxidation
of the volatile magnesium leads to depletion of the solute at the surface in order
+ +
to reach the three-phase equilibrium: (Al, Mg) alumina spinel. Alumina then
becomes the preferred oxide product. The top (1-2pm) of the surface at the
growth front has a complex chemistry whose details are still a matter for debate.
In general terms, the surface consists of a layer of MgO separated from the growing
composite by a thin wetting layer of molten alloy. This model borrows schematically
(Fig. 4) from the microstructure found in liquid-solid reactive diffusion couples,
such as Al-MgO or A1-Si02, in which the alloy wets its own oxidation product
and penetrates the solid-solid interface, which in the present case is A1203-Mg0.
Such microstructural observations were the basis for the steady state growth
hypothesis [45,48,59],which assumed anionic oxygen transport through the MgO
layer, controlled by either electronic or cationic vacancy defects, followed by
dissolution of MgO into the molten alloy, diffusion of dissolved oxygen through
the thin intervening alloy layer and, finally, an oxidation reaction to form alumina
at the interface with the preformed, columnar alumina crystals. As indicated in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
302 1 Directed Metal Oxidation
Figure 4. Schematic of surface microstructure during DMO growth. A thin alloy film is formed
between a surface MgO layer and the growing alumina skeleton and is continuous with underlying
alloy channels. The reactions at A, B, and C are described in the text.
DMO composite growth from aluminum alloys that contained no Mg but other,
alternative volatile elements (Na, Li, Zn). In each case DMO growth occurs at
temperatures for which the vapor pressure of the volatile solute is an appreciable
fraction of the partial pressure of oxygen in the gas phase.
The mechanism of oxygen transfer across the surface oxide layer is also problem-
atic. Observed growth rates require that diffusion occur through the surface oxide
at rates far greater than those measured for the fastest diffusing species (cation
vacancies) in polycrystalline MgO [66].In addition, the decrease in growth rate
with increasing thickness implies that viscous flow of liquid through the capillary
channels of the composite might also be rate determining. The results of interrupted
experiments [63] suggest that microstructural banding might be related to the
formation of alumina by alternate dissolution of both spinel and magnesia into
the molten alloy. The volume fraction of the alumina phase in these experiments
also corresponded approximately to that expected from a comparison of the
molar volume of oxygen in the alumina-spinel and alumina-magnesia phase
pairs. This hypothesis also predicts correctly the volume fraction of spinel formed
from MgO during the spinel growth peak in the incubation stage, as well as that
of alumina formed from a silica nucleant.
Alternating bands of magnesia and spinel at the growth interface with the molten
alloy could be associated with periodic fluctuations in the magnesium concentration
of the melt. Calculation of the partial pressure of oxygen (PO,) in equilibrium with
different alloy-oxide combinations suggests that a low magnesium content in the
melt maximizes the driving force for the formation of alumina by dissolution of
MgO, while the resulting enrichment of magnesium then favors epitaxial spinel
formation on the MgO crystals [63].Subsequent growth of alumina then occurs
by dissolution of spinel, leading to a decrease in the magnesium concentration by
diffusion into the bulk, and the cycle is then repeated. Isolation of molten metal
from the MgO decreases the rate of magnesium oxidation and hence the rate of
weight gain during this period of spinel dissolution. Subsequent re-penetration of
spinel by the molten alloy once more initiates direct dissolution of MgO. The
repeated formation of MgO required by the above sequence is most likely due to
the oxidation of magnesium in the vapor phase, although cation de-mixing may
also be possible if the spinel layer is sufficiently thick.
In summary, composite formation by DMO of Al-Mg alloys is considered to be
made possible by the following sequence of events.
0) A buffer layer of spinel at the original melt surface ensures a supply of molten
alloy, containing =0.2-0.5% of magnesium, to the free surface, close to the
three-phase equilibrium composition of the alloy with spinel and alumina.
(ii) The high vapor pressure of magnesium leads to the formation of MgO at the
surface of the composite.
(iii) A displacement reaction with aluminum leads to the formation of alumina, not
by direct nucleation but rather by transport of dissolved oxygen to the Al-
A1203 interface. This reaction is periodically mediated by spinel, due to local
variations in the magnesium activity that are not immediately compensated
by bulk diffusion from the melt reservoir.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
304 1 Directed Metal Oxidation
The high-temperature limit for oxidative growth has been attributed to excessive
loss of magnesium vaporizing from the surface [50,62,65]. Unless the magnesium
can be replenished by diffusion through the liquid, a passivating layer of alumina
or spinel will form. In principle, providing the bulk reservoir contains sufficient
magnesium, oxidative growth is expected to resume eventually, the times involved
being comparable to the incubation period. Related experiments [65] have shown
that sufficient magnesium may be available, either in the alloy channels of a
previously grown composite or in the spinel layer at the interface with the melt,
to continue DMO growth when placed in contact with a pure aluminum melt.
This demonstrates the very low levels of magnesium actually necessary for com-
posite formation, but it is likely that, in the absence of a magnesium reservoir,
growth will eventually cease due to magnesium losses by spinel precipitation or
magnesium evaporation.
than alumina. During growth in this system, however, the principal surface oxide
appears to be ZnO [56,57], although it should be noted that the two cations
magnesium and zinc readily substitute for each other in the periclase lattice. When
magnesium-rich alloys are used, the surface oxide has been found to revert to
MgO, as found in binary Al-Mg [65]. Additions of silicon improve the uniformity
of growth in Al-Mg [48] and Al-Zn-Mg alloys [56], even though binary Al-Si
alloys cannot initiate composite formation. For example, binary Al-Mg alloys may
nucleate only a few nodules per square centimeter, which grow irregularly outwards
from the isolated nucleation sites and never develop a planar growth front. Silicon
additions greatly increase the surface density of nodules and lead to a very much
more uniform and smooth growth front. Improved wetting by reduction of the
alloy-oxide interfacial energy has been suggested as the primary reason for more
regular growth, although prolonged incubation and microstructural banding, due
to periodic magnesium depletion, are also observed in Al-Mg-Si alloys.
It has already been suggested that oscillations in the growth rate are probably
linked to the periodic surface depletion of magnesium. The observation that
DMO composites grow initially with a uniform microstructure and a smoothly
varying growth rate, before the onset of the growth instability [62,63], is consistent
with the hypothesis that the instability occurs when the distance from the growth
surface to the magnesium reservoir in the alloy melt is sufficient for solute diffusion
through the liquid channels to become rate limiting. Furthermore, uniform micro-
structures are commonly found in ternary alloys which contain additions such as
silicon, palladium, or indium [70], and for these solutes thermodynamic calculations
have shown that the stability of alumina is extended to higher magnesium con-
centrations. This improved alumina stability should make spinel formation less
likely and so ensure a single alumina growth mechanism via MgO dissolution
into molten aluminum. However, an important feature of oscillatory growth that
still has no satisfactory explanation is the lateral uniformity of alloy composition
across the growth surface that should be necessary to achieve coordinated cyclic
behavior over the whole of the growth front.
the later stages. Nevertheless, there is still no clear explanation relating the micro-
structural length scale to the alloy composition and the DMO growth temperature.
It is possible to produce composites with an alumina content in excess of 85 vol-% at
high temperatures. However, this appears to be achieved either at the expense of a
reduced growth rate or with premature termination of the growth process before
complete conversion of the alloy to composite [60,74]. Excessive evaporative loss
of magnesium and sealing of alloy channels by alumina are the probable combined
causes of this high temperature behavior, which prevails above about 1300°C
(typically 1300°C for Al-Cu-Mg and 1200°C for Al-Ni-Mg [74], 1300°C for Al-
3 Mg [62], and 1400°C for Al-Zn-Mg [65] and Al-Si-Mg [46]). A further compli-
cation in complex alloys is the enrichment of nonoxidizing elements, such as silicon,
nickel, or copper, which fail to diffuse away from the interface and instead pre-
cipitate refractory intermetallic compounds [74].
A controversial area in the growth of DMO A1203-A1 are the temperature and
oxygen partial pressure dependencies of the oxidation rate. One difficulty in
reconciling experimental results obtained by different authors and assigning an
unambiguous growth mechanism has been the absence of a clear regime of constant
growth rate [65]. The available evidence suggests that the rate of weight gain starts to
fall after little more than 1-3 mm of composite growth. This decrease in growth rate
approximates a parabolic time dependence in some instances [46,75] and has been
attributed variously to enrichment of nonoxidizing elements near the surface,
depletion of magnesium in the melt, or percolation restrictions on the supply of
metal through the capillary channels [62,71].
Absence of data on melt viscosity, alloy channel tortuosity, alloy-alumina
interfacial energies, and melt-oxide contact angles precludes a meaningful analysis
of viscous flow of the melt through the percolation channels and its possible control
of the growth rate. Some thermogravimetric studies have been conducted over a
sufficient range of temperature and time to permit the determination of average
growth rates and an analysis of activation energies which are valid for a limited
range of temperature and composite thickness. Variations in alumina content
across the composite and uneven growth across the surface make it difficult to
define a linear interface mobility from gravimetric data. Measured activation
energies range from 90 kJ mol-' for a complex Al-Si-Zn-Mg-Cu-Fe alloy [45],
to 213 kJ mol-' for Al-1 Mg [76], 360 kJ mol-' for A1-5 Mg [77], and 400 kJ mol-'
for Al-Mg-Si [71], with other values typically around 270-300 kJ mol-' [46].
Attempts to relate these values to specific mechanisms, such as ionic diffusion in
MgO [59], have largely been abandoned because of the wide energy spread and,
in some cases, the absence of well-defined Arrhenius behavior [50].Similar inconsis-
tencies have been observed in the oxygen partial pressure dependence of DMO
growth. Power law dependencies of 1/4 and 1/6 for the oxygen partial pressure
were reported for Al-Mg-Si [71], but growth from Al-Si-Zn-Mg [45] was found
to be insensitive to PO2. In contrast, an exponent of unity was proposed for Al-
5Mg with 2 and 5% Si, while growth from an A1-5Mg-10% alloy was again
found to be insensitive to PO2 [78].
It would appear that the growth of the composite is inherently cyclic in the short
term, involving the formation and dissolution of oxides such as MgO and ZnO, and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Directed Metal Oxidation Growth 307
is self-limiting in the long term, due to surface depletion of the volatile element,
enrichment of the melt in the channels by nonoxidizing elements, or to an
insufficient flux of molten alloy through the capillary channels.
Here, it was argued that the oxygen supply through the pores of the alumina
preform was the rate-limiting step, and that growth became progressively easier
as the alloy percolation channels approached the free surface. The oxidation rate
was also found to decrease with increase in the silicon content of the melt.
In the above work, the presence of alumina preform particles appeared to favor
infiltration but no residual metal was found between the preform alumina particles
and the matrix alumina and, surprisingly, neither was any epitaxy identified [72]. If
the particles aid infiltration, then the rate of infiltration would be expected to
increase with the particle surface area, and this was indeed the case for coarse
alumina particles DMO infiltrated using Al-Zn, Al-Zn-Mg, and Al-Mg alloys
[54,75,79,80]. A maximum in the growth rate was observed for particle sizes of
about 20-50pm [57] and 5-8pm [80]. The decrease in growth rate at smaller
particle sizes was attributed either to the difficulty of developing the surface
oxide layer (MgO/ZnO) or to a diminished supply of liquid metal owing to the
restriction in viscous flow due to the reduced size of the pore channels. Interest-
ingly, in contrast to earlier work on Al-Mg based alloys, composites grown
from Al-Zn-Mg alloys, in which copious nucleation of ZnO was observed on alu-
mina particles ahead of the main infiltration front, also exhibited rapid wetting of
the Zn0-A1203 interface, as well as epitaxy between the matrix alumina and the
particulate alumina preform. In one paper the microstructure of the preform-
infiltrated matrix has been reported to be refined when compared to that obtained
in the absence of an alumina preform, but the magnitude of the effect was not
specified [72]. The importance of wetting of the alumina preform by the liquid
alloy and of the preform-ZnO interface was demonstrated by the absence of
infiltration of Al-Zn-Mg into Zr02 [57].
Quantitative assessment of the role of pore size and particle packing fraction (the
‘green’ density of the preform) in the growth process remains elusive, largely owing
to the absence of steady-state growth over a sufficiently wide range of preform
parameters. Watari et al. [75] have reported a growth rate that was proportional
to the square root of the pore size. However, the rate of infiltration into the alumina
preform observed in this study was nonlinear and the data only appear to obey the
given pore size relationship for composites of 1-2 mm in thickness and at a single
temperature.
Silicon carbide has been a popular choice of particulate for preforms in the
DMO growth of A1203-A1 composites, promising both hardness and chemical
stability in an oxidizing atmosphere. A distinctive feature of DMO growth into
S i c preforms is the reactivity of Sic with both the alloy melt and with oxygen.
Sufficient silicon must be present in the alloy to prevent formation of aluminum
carbide during growth, since aluminum carbide is hygroscopic and leads to time-
dependent disintegration of the composite under ambient conditions [8 11. Super-
ficial oxidation of S i c leads to reactive infiltration of liquid aluminum into the
silica-coated Sic. Initial published work on the A1203-SiC-A1 formulations
showed that growth rates increased with decreasing particle size 1821, and attribu-
ted the increase to wetting of the silica layer by the melt. It was subsequently
pointed out [83] that oxidation of Sic, either during pre-sintering of the preform
or during infiltration, would be expected to increase the amount of silica available
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Directed Metul Oxidation Growth 309
with decreasing Sic particle size in the preform. In effect, the volume fraction of
Sic in the composite is reduced and the silica product acts as an additional
source of oxygen that promotes reactive infiltration. This effect was found to
increase the residual alloy content of the DMO composite up to 30 vol-%, using
a 5pm particle size S i c particulate preform, as compared with 10-15 vol-% of
residual alloy with 50-300 pm S i c particles.
The reactive infiltration of a porous silica preform to yield a metal-rich A1203-A1
DMO composite appears to be similar [31]. In this case, the individual silica particles
are completely reacted and replaced by Al2O3-Al, while the pore spaces are
filled with the resulting silicon-rich alloy. Similarly, it was shown that additions of
1-20% Si02 and Sn02 to Sic preforms could increase the final alloy content of the
resulting composite [84]. In summary, optimizing the S i c preform particle size can
promote the production of DMO composites with a higher alloy content. However,
this is often at the expense of significant residual porosity in the final product,
particularly at higher temperatures (> 1000°C). Residual porosity may be avoided
by coating the Sic particles with reactive alumina (by the precipitation of aluminum
hydroxide followed by calcination). Composites made from preforms with these
alumina-coated S i c particles contain less porosity and residual metal, and have
higher hardness [85].
When excessive reactive wetting of the preform leads to metal migration over
particle surfaces ahead of the main reaction front, residual porosity becomes a
ubiquitous feature of DMO infiltration. The surface of a partially-infiltrated
region then becomes sealed before the underlying pores are filled with the advancing
composite. In the case of oxidation in air, a similar phenomenon can lead to the
entrapment of nitrogen-containing pores.
The crystallinity of alumina formed by DMO infiltration into S i c particulate pre-
forms follows the expected behavior of heterogeneous nucleation on S i c particles
and large monocrystalline regions with [00011 parallel to the growth direction
within the larger voids [83,86]. Manor et al. [82] found that the infiltration rate
increased as the Sic particle size decreased. However, no systematic trend was
found in another work [83], most probably due to widely different silicon concentra-
tions in the percolation channels resulting from reduction of the silica layer during
infiltration.
strength of the fiber interface with the matrix. In the latter coating system, the role of
alumina is to reduce the thermal mismatch with the fiber, while the middle zirconia
layer both promotes fiber pull-out and protects the fiber against attack by the melt.
The outer silica layer promotes wetting by the molten metal. The alloy used in the
last example was A1-5 Mg, but a quaternary alloy, with silicon and zinc additions
and less magnesium, would probably have minimized degradation of the fiber and
reduced the incubation time, possibly obviating the need for the silica coating. An
attempt to infiltrate aluminosilicate (mullite-based) fibers resulted in extensive
attack associated with the reduction of silicon from the oxide and its dissolution
in the melt [90].
strength of the alloy and fractographic analysis of the composite, gave an estimated
fracture energy contribution of 150-200 J mP2.This value corresponds to a tough-
ness increment of about 7MPam'/2 (assuming a Young's modulus of 250GPa)
over that for polycrystalline alumina, which is consistent with the maximum value
reported by Aghajanian and coworkers. The alloy in this case, though not specified,
was reported to contain Cu, Fe, and Si, implying a higher fraction of intermetallics
than in the Al-Mg alloy studied by Aghajanian and coworkers. Pickard et al. [91]
found R-curve behavior in four-point bending of A1203-Al composite samples
made from Al-Mg and Al-M Si Zn, with an initial toughness of 3 MPam'I2
P-
rising to a plateau of 5 MPa m 12. However,
-
a complex commercial alloy (A-380)
yielded DMO composites which contained intermetallic phases and also gave a
constant toughness of 3MPam'I2. The toughness appears to be increased by
additions of Sic [91]; in particular, the increment in the plateau of the crack
resistance curve increased with S i c particle size. It was suggested that particle
bridging and frictional sliding at the interface could contribute significantly to
energy dissipation. A model developed to explain the increase in KIc accounted
satisfactorily for the toughness of composites containing 67 and 165 pm Sic
particles, but underestimated the value for 35 pm Sic. It was concluded that
particulate contributions to toughening exceeded those due to plastic dissipation,
at least for the composites derived by DMO from this alloy (A-380), which con-
tained a high fraction of brittle intermetallics and displayed a flat R-curve. In situ
measurements of crack opening displacement in a A1203-SiC-A1 DMO composite
[95] confirmed that the intrinsic (crack tip) toughness was about 2 MPa similar
to many other alumina matrix composites) and that the observed increases of up to
7.5 MPa ml/' could be attributed entirely to crack wake effects, including frictional
sliding, S i c particle bridging and plastic deformation of alloy bridges, rather than to
changes in the crack-tip profile. Anderson [96] measured KI in four-point bending
for both a base composite and a Sic-reinforced composite and obtained 7.8 and
5.6 MPa m1/2,respectively. The decrease observed after growth into the particulate
preform is inconsistent with the results of Pickard et al. but it is unlikely that the
alloys used for the two composites were the same.
In additional results [97] for the fracture toughness, KIc, determined from a single
edge-notched beam test, values of 3.6MParn'l2 (for a 100pm Sic particulate
preform) up to 6.2MPam'12 (13 pm Sic) were reported. This trend is the opposite
of that reported by Pickard et al. [91]. Values of the elastic modulus reported in this
later work ranged from 261 to 328 GPa.
Most mechanical properties data report only rupture strength, and, since such
data are sensitive to the size and distribution of preexisting bulk defects and surface
finish, they are not easily interpreted without some help from microstructural
analysis. Flexural strengths in the range 250-300MPa were reported for two
alumina-aluminum composites, the higher values for composites that contained
more metal and less porosity, and which also had a higher fracture toughness
[60]. Values around 150-300 MPa have been published for alumina and silicon
carbide-reinforced compositions, but the particle sizes were not disclosed [92].
Pickard et al. [91] studied Sic reinforced materials and found that the finest
(35 pm) Sic particle size gave the highest strength (400 MPa) but the minimum
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.6 Mechanical Properties 3 13
toughness. The coarsest preform particles, 165 pm Sic, yielded strengths only of the
order of 150 MPa. In this work microcracking at the Sic particles due to thermal
expansion mismatch with the alumina matrix was held responsible for the strength
degradation. Raman peak shift measurements have shown [98] that residual stresses
in the silicon carbide particulate preform are a minimum for large particles, so it is
possible that unfilled pore space in the preform (whose dimensions should scale with
the particle size) could be responsible for the strength loss.
1.9 Applications
The engineering advantages of DMO composite ceramics determine the possible
areas of application, although actual penetration into these markets depends on
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
I .9 Applications 3 17
ceramic plate into a composite armor system can significantly reduce the weight of
armor which is required to defeat a given threat. Two examples are the boron
carbide armor used to protect the pilot’s seat in combat helicopters and the
alumina plate inserts used to improve the personal protection provided by a polymer
reinforced armored vest. The accepted criterion for judging the performance of an
armor system is the V5,test, which is the minimum weight per unit area of armor
required to defeat a given projectile under standard conditions (a 50% chance of
penetrating the armor module at a specified projectile velocity V ) .
Sic particulate-reinforced DMO tiles are known to have attractive ballistic
properties against light and medium threats (up to and including medium caliber,
high velocity rounds and fragments). The DMO materials are candidates for
vehicular add-on armor [ 1171 and are reported to have been used to provide
added protection to armored personnel carriers by the US Army in the Persian
Gulf war against Iraq. There is no indication that DMO armor has ever been
considered seriously as a candidate ballistic material to defeat the heavy threat
faced by a main-line battle tank. In terms of cost, DMO armor should be competi-
tive with steel because of the reduced weight [118]; however, it is by no means clear
that the DMO product is superior to other grades of alumina-based ceramic armor
tile, either in terms of cost or performance.
References
1. M. S. Newkirk, A. W. Urquhart, H. R. Zwicker, and E. Breval, J . Muter. Res. 1986,1,81-89.
2. M. S. Newkirk, H. D. Lesher, D. R. White, C. R. Kennedy, A. W. Urquhart, and T. D. Claar,
Ceram. Eng. Sci. Proc. 1987, 8, 879-885.
3. E. T. Thews, Met. Ind. (Lond.) 1927, 31, 597.
4. K. J. Brondyke, J . Am. Ceram. SOC.1953, 36, 171-174.
5. J. G. Lindsay, W. T. Bakker, and E. W. Dewing, J . Am. Cerum. Soc. 1964, 47, 90-94.
6. A. W. Urquhart, Muter. Sci. Eng. A , 1991, A144, 75-82.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 3 19
51. 0. Salas, V. Jayaram, K. C. Vlach, C. G. Levi, and R. Mehrablan, J . Am. Ceram. SOC.1995,
78, 609-62 I.
52. C. D. Qin, M. 0.Lai, B. Cotterell, and S. Shi, Muter. Res. SOC.Symp. Proc. 1995,365, 133-138.
53. H. Schmalzried and W. Laqua, Oxid. Metals, 1981, 15, 339-353.
54. K. C. Vlach, C. G. Levi, and R. Mehrabian, in Processing and Manufacturing of Advanced
Materials for High Temperature Applications, V. A. Ravi and T. S. Srivatsan (Eds), The
Minerals, Metals and Materials Society, Warrendalc, USA, 1992, pp. 157-171.
55. V. S. R. Murthy and B. S. Rao, J . Muter. Sci. 1995, 30, 3091-3097.
56. M. Sindel, N. A. Travitzky, and N. Claussen, J . Am. Ceram. SOC.1990, 73, 2615-2618.
57. M. Hanabe, V. Jayaram, and T. A. Bhaskaran, Acta Muter. 1996,44, 819-829.
58. P. Xiao and B. Derby, J. Am. Ceram. Soc. 1994,77, 1771-1776.
59. A. S. Nagelberg, S. Antolin, and A. W. Urquhart, J. Am. Ceram. Soc. 1992, 75, 455462.
60. M. K. Aghdjanian, N. H. Macmillan, C. R. Kennedy, S. J. Luszcz, and R. Roy, J . Muter. Sci.
1989,24, 658-670.
61. E. Breval, M. K. Aghajanian, and S. J. Luszcz, J . Am. Ceram. SOC.1990, 73, 2610-2614.
62. 0.Salas, A Fundamental Study of the Formation of A1,03/Al Composites from the Oxidation
of Liquid A1 Alloys, PhD thesis, University of California at Santa Barbara, USA, 1992.
63. 0. Salas, V. Jayaram, K. C. Vlach, C. G. Levi, and R. Mehrabian, in Processing and
Manufacturing of Advanced Materials for High Temperature Applications, V. A. Ravi and
T. S. Srivatsan (Eds), The Minerals, Metals and Materials Society, Warrendale, USA, 1992,
pp. 143-156.
64. P. Padmapriya, T. A. Abinandanan, and V. Jayaram, Scripta Muter. 1997,37, 1883-1888.
65. P. Xiao and B. Derby, J. Europ. Ceram. SOC.1997, 12, 185-195.
66. V. Jayaram, J . Muter. Sci. 1996,31,45914598.
67. F. J. A. H. Guillard, R. J. Hand, and W. E. Lee, Br. Cerum. Trans. 1994, 93, 129-136.
68. X. Gu and R. J. Hand, J . Muter. Sci. 1995,30,6005-6012.
69. X. Gu and R. J. Hand, J . Europ. Ceram. SOC.1996, 16, 929-935.
70. M. Pinkas, A. Venkert, and E. Manor, Scripta Muter. 1997,36, 377-383.
71. A. S. Nagelberg, Muter. Res. Soc. Symp. Proc. 1989, 155, 275-282.
72. E. Breval and A. S. Nagelberg, Mater. Res. Soc. Symp. Proc. 1989, 132,93-98.
73. A. Guha and V. S. R. Murthy, Bull. Muter. Sci. 1996, 19, 1117-1124.
74. S. C. Khatri, M. J. Koczak, T. Chou, and Y. Kagawa, Ceram. Eng. Sci. Proc. 1992,13,494-502.
75. T. Watari, K. Mori, T. Torikai, and 0. Matsuda, J . Am. Ceram. Soc. 1994,77,2599-2602.
76. S. C. Koh, D. K. Kim, and C. H. Kim, J . Muter. Synth. Proc. 1995,3, 105-110.
77. H. Venugopalan, K. Tankala, and T. Debroy, Muter. Sci. Eng. 1996, A210, 6 4 7 5 .
78. S. C. Khatri, M. J. Koczak, T. Chou, and Y. Kagawa, Ceram. Eng. Sci. Proc. 1992,13,485493.
79. D. D. Upadhyay, R. Bhat, S. Ramanathan, and S. K. Roy, J . Alloys Compd. 1994, 205,
274279.
80. H. Venugopalan and T. Debroy, Muter. Sci. Eng. 1997, A232, 39-46.
81. S. Kumar, V. Jayaram, T. V. Mani, and K. G. K. Warrier, J . Muter. Sci. 1997,32,47194722.
82. E. Manor, H. Ni, C. G. Levi, and R. Mehrabian, J . Am. Ceram. Soc. 1993,76, 1777-1787.
83. S. P. Dhandapani, V. Jayaram, and M. K. Surappa, Acta Metall. Muter. 1994, 42, 649-656.
84. S. Lee and D. K. Kim, Ceram. Eng. Sci. Proc. 1990,11,795-805.
85. V. Jayaram, S. Kumar, T. V. Mani, M. S. M. Saifullah, J. Sarkar, and K. G. Warrier, in
Inorganic Matrix Composites, M. K. Surappa (Ed.), The Minerals, Metals and Materials
Society, Warrendale, USA, 1996, pp. 193-205.
86. V. S. R. Murthy and A. Deepak, Br. Ceram. Trans. 1996, 95, 173-176.
87. A. S. Fareed, G. H. Schiroky, and C. R. Kennedy, Ceram. Eng. Sci. Proc. 1993,14,794-801.
88. A. S. Fareed and G. H. Schiroky, Ceram. Eng. Sci. Proc. 1994, 15, 344352.
89. X. Gu and E. G. Butler, Ceram. Eng. Sci. Proc. 1997, 18, 371-378.
90. I. S. Kim and H. W. Hennicke, Znterceram. 1992,41, 75-76.
91. S. M. Pickard, E. Manor, H. Ni, A. G. Evans, and R. Mehrabian, Acta Metall. Mater. 1992,
40, 177-184.
92. D. J. Landini, H. D. Lesher, and J. Gesing, in Proc. Third International Symp. Ceramic
Materials und Components for Engines, V. J. Tennery (Ed.), American Ceramic Society,
Westerville, USA, 1988, pp. 1536-1551.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Rejerences 32 1
93. J. Weinstein and B. Rossing, in High Performance Composites for the 1990s, S. K. Das, C. P.
Ballard, and F. Marikar (Eds.), The Minerals, Metals and Materials Society, Warrendale,
USA, 1991, pp. 339-357.
94. B. D. Flinn, M. Ruehle, and A. G. Evans, Acta Metall. 1989, 37, 3001-3006.
95. J. Roedel, M. Sindel, M. Dransmann, R. W. Steinbrech, and N. Claussen, J . Europ. Ceram.
SOC.1994, 14, 153-161.
96. C. A. Andersson and M. K. Aghajanian, Ceram. Eng. Sci. Proc. 1988, 9, 621-626.
97. I-S. Kim and I-G. Kim, J . Muter. Sci.Lett. 1997, 16, 772-775.
98. R. Arvind Singh, A. K. Sood, V. Jayaram, and S. K. Biswas, Scripta Muter. 1998,38,617-622.
99. H-T. Lin and K. Breder, J . Am. Ceram. Soc. 1996,79, 2218-2220.
100. Reported in Product Brochures, ALANX Products, 101 Lake Drive, Newark, Delaware
19702, USA.
101. R. K. Dwivedi, Cerum. Eng. Sci. Proc. 1991, 12, 2203-2221.
102. V. Jayaram, R. Manna, M. G. Kshetrapal, J. Sarkar, and S. K. Biswas, J. Am. Ceram. Soc.
1996,79, 770-772.
103. A. Ravikiran, V. Jayaram, and S. K. Biswas, J . Am. Ceram. SOC.1997, 80, 219-224.
104. J. Zhang, D. G. Zhu, and L. Yang, Wear, 1998,215, 34-39.
105. R. Arvind Singh, V. Jayaram, and S. K . Biswas, J . Muter. Res. 1999, 14, 6 6 6 7 .
106. C. A. Anderson, P. Barron-Antolin, A. S. Fareed, and G. H. Schiroky, in Proc. Int. Con$
Whisker and Fiber Toughened Ceramics, American Society for Metals, 1988, pp. 209-21 5.
107. A. S. Fareed, B. Sonurpalak, C. T. Lee, A. J. Fortini, and G. H. Schiroky, Ceram. Eng. Sci.
Proc. 1990, 11, 782-794.
108. A. Fareed, B. Sonurpalak, P. A. Craig, and J. E. Garnier, Ceram. Eng. Sci. Proc. 1992, 13,
804818.
109. F. Heredia, A. G. Evans, and C. A. Andersson, J . Am. Ceram. SOC.1995,78,2790-2800.
110. D. G. Kolman and D. P. Butt, J . Electrochem. SOC.1997, 144, 3785-3791.
111. W. A. Kem, M. J. McNallan, and R. E. Tressler, J . Am. Ceram. Soc. 1996, 79, 1881-1891.
112. T. M. Strobel, J. P. Hurley, K. Breder, and J. E. Holowczak, Ceram. Eng. Sci.Proc. 1994, 15,
579-586.
113. H. Venugopalan, K . Tankala, and T. Debroy, J . Am. Ceram. Soc. 1994,77, 3045-3047.
114. S. L. Swartz, D. R. White, and L. E. Cross, J . Muter. Sci. 1992, 27, 3932-3938.
115. Reported in Futuretech, Technical Insights Inc., Englewood, USA, 1990, 101, 1-19.
116. B. W. Sorenson, G. H. Schiroky, and A. W. Urquhart, Turbomachinery, 1990,31,20-26.
117. S. Ashley, Mech. Eng. 1991, 113, 44-49.
1 18. Reported in Structural CeramicslCerumic Matrix Composites, Techmonitoring, SRI Inter-
national, 1995, 44-45.
119. L. P. Zawada and S. S. Lee, Ceram. Eng. Sci. Proc. 1995, 16, 337-339.
120. Reported in: Society and Industrial News, Am. Ceram. Soc. Bull. 1995, 73, 28.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
2.1 Introduction
Of the many advantages of self-propagating high-temperature synthesis (SHS),
most can be realized in the synthesis of hard (primarily refractory) materials.
Among the most important and generally applicable of these advantages are: the
energy efficiency provided by this process, its tendency to produce relatively pure
products, and its suitability for the simultaneous synthesis and densification of
monolithic and composite materials. The synthesis and densification of composites
can be accomplished in one step through the use of this method. In recent years, the
utilization of the SHS process for the preparation of hard and other types of
materials has been extensive with a variety of experimental approaches and modifi-
cations, including chemical, thermal, mechanical, and electromagnetic activations.
The basis of the SHS method is the propensity of highly exothermic reactions to
become self-sustaining after they have been initiated by a brief energy input. A
requirement for such a phenomenon is that a portion of the heat generated by the
chemical reaction is conducted forward to raise the temperature of the next
reactants layer to a temperature at which the reaction rate becomes significant, a
temperature commonly referred to as the 'ignition' temperature. Part of the heat
generated in this layer will be conducted forward and thus the process is repeated
and the reaction moves along the reactants within a usually narrow zone, typically
referred to as the combustion zone or combustion wave. The process is demon-
strated by the sequence of photographs in Fig 1 .
As will be seen in more detail below, the SHS process is governed by the chemical
and physical parameters of the reacting system. And under real experimental condi-
tions, the process is also controlled by the ambient conditions. To understand the
SHS process and its governing parameters, we consider the general Fourier heat
transfer relationship accounting for a heat source,
+
pC,(dT/dt) = fi(d2T/dx2) p Q ( d q / d t ) - h / w ( T - To) - u s / w ( T 4- T t ) , ( 1 )
where p is the density of the product (g ~ m - ~Cp
) , is its heat capacity (J g-' K-I), K. is
the thermal conductivity (JcmP2s-' K-*), Q is the heat of the reaction (Jg-I), q is
the fraction of the reaction completed, h is the convective heat transfer coefficient
(J cm-2 sC1 K-' , us is the Stefan-Boltzmann constant (J cmP2s-' KP4), T is
temperature (K), To is the ambient temperature, w is the sample thickness (cm), t
is time (s), and x (cm) is the coordinate along which the combustion wave is
advancing. The term on the left-hand side of Eq. (1) is the algebraic total rate of
heat accumulated in the combustion zone and is the sum of the net rate of conducted
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.1 Introduction 323
heat, the rate of chemical heat evolution, the rate of heat loss due to convection,
and the rate of heat loss due to radiation. The reaction rate is assumed to have an
Arrhenius dependence on temperature and has the form
dq/dt = ko(l - q)"exp(-E/RT), (2)
where E is the activation energy for the process, n is the order of the reaction, R is
the gas constant, and ko is a constant.
Under ideal (adiabatic) conditions, the last two terms on the right-hand side of Eq. (1)
are neglected and the calculated maximum temperature in the combustion zone is the
adiabatic combustion temperature, Tad.For most of the important refractory materials
(including many hard materials) Tad is high, exceeding 4000 K in some cases. The high
value of this parameter is a thermodynamic consequence of the high exothermic heat of
formation of these materials. Table 1 lists the values for Tad and the enthalpies of
formation for selected hard materials, including carbides, nitrides, oxides, and borides.
With few exceptions, the adiabatic temperatures of the listed materials exceed the
empirically established minimum limit for SHS, about 2000 K [I]. Generally the excep-
tions are compounds of tungsten (W2C,WC, WB, W2B),molybdenum (Mo2C,MoN,
MOB),and chromium (Cr3C2).It should be emphasized that the temperatures in Table
1 are based on adiabatic conditions and thus are upper limits which are seldom
achieved in practice. In some cases, notably in the case of nitrides, the adiabatic tem-
perature is so high as to exceed the dissociation temperature of the nitride phase.
Examining Table 1 shows a general trend between the enthalpy of formation of the
material and its adiabatic temperature. However, a more accurate relationship is
between the ratio of the enthalpy to the heat capacity (H/Cp)29X and the adiabatic
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
324 2 Self-propagating High-Temperature Synthesis of Hard Materials
temperature (Tad),as has been shown previously [2]. Such a relationship is shown in
Fig. 2 for the materials listed in Table 1 for which reliable thermodynamic data are
available. With two notable exceptions, the data fit a straight line and a more useful
limit for SHS can now be stated as a (H/Cp)298 RZ 3000 K.
In SHS reactions, the maximum temperature is ideally achieved within the narrow
combustion zone, indicating that the reaction is complete within the combustion
wave. However, for kinetic reasons, some reactions continue beyond the zone and
consequently the maximum temperature is reached after the passage of the wave.
Schematic representations of these two cases are shown in Figs 3(a) and (b), respec-
tively. The figures show profiles of temperature and the conversion parameter. In the
latter case, diffusion of one or both reactants through a product layer can provide a
kinetic barrier for the process and hence its lack of completion within the com-
bustion zone. Such a behavior can be seen in systems where the particle size of
the reactants is relatively large, as was reported earlier for the synthesis of titanium
carbide [3]. Since in this case the conversion is incomplete in the wave, the tem-
perature of the wave is relatively low and hence its rate of propagation is low. In
general the nature of the product is not affected by a change in velocity as long as
the propagation is in the steady-state mode [4-6]. When the propagation is in the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.1 Introduction 325
E
2 0
1 5
1 0
5
F TiCh
0
0 2 4 6 8 1 0 1 2 1 4
Tadx1 0- 3( K )
Figure 2. The relationship between (AH,,,/C,) and the adiabatic temperature for hard materials.
non-steady-state mode (e.g. spin or pulsating wave propagation), the reaction may
be incomplete or the products are not the same as those obtained under steady-state
propagation [6]. An example of spin propagation is shown in Fig. 4 for the synthesis
of nickel silicides [7].
Wave
Propagation I I&
XrdL
Wave propagation
t-------
~
k 4b
T,
Vf
k i . After-bum 4 *
Figure 3. Schematic representation of the temperature, T , degree of conversion, 7,and rate of heat
release, 4, during wave propagation for: (a) a reaction completed within the wave, (b) a reaction
which extends beyond the wave.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 5. Temperature profiles during the combustion synthesis of: (a) Ti + C + 25 weight-% Ni,
Figure 4. Evidence of spin combustion during the synthesis of Ni silicides [7].
2 Self-propagating High-Temperuture Synthesis of Hard Materials
60
45 52.5
22.5 30 37.5
time (s)
+
0
2500
2500
326
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Mechanistic Characterization of the Process 327
2500
2000
G 1500
0
G
I000
500
O L J I 1 I I I I 1
0 15 30 45 60
t 6)
2500 -
2000 -
-Y 1500 ~
G -
1000
500 -
O r . I I I
Figure 6. Temperature profile of a 52% dense Zr + 1.5B + 15 weight YOTiB2: (a) complete profile,
(b) enlarged portion of the initial segment of the profile [lo].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
328 2 Self-Propagating High-Temperature Synthesis of Hard Materials
v
1000
500
- 0
3.00 3.04 3.08 3.12 3.16 3.20
- 150000
50000
s
b 0
-50000
3.00 3.04 3.08 3.12 3.16 3.20
-.
PI
-I.E+7
" ! -1.5E+7
+ +
Figure 7. Thermal profile characterizations of the wave in the reaction of Zr 1.5B 15 weight-%
TiB2: (a) smoothed temperature profile, (b) first derivative of the profile, (c) second derivative of the
profile [lo].
respectively. The former is the temperature profile for the synthesis of Tic-Ni
composite and the latter is for the synthesis of TIN [12]. In the latter case, both
visual observations and thermogravimetric analyses confirmed that the reaction
between Ti and nitrogen continues well after the passage of the wave [8,13].
As indicated above, determination of the kinetic parameters of the SHS process
can be made through a mathematical analysis of temperature profiles [14]. Integra-
tion of an adiabatic form of Eq. (1) along with appropriate substitutions gives the
following expression for the conversion parameter [9],
V(X) = [Cpu(T- TO)- 61( d T / d x ) I / [ ( - (3) ) QWI,
~ ~ l ) ( d T / d x+
where u is the wave velocity, K~ and K~ are the thermal conductivity of the reac-
tants and products, respectively, To is the ambient temperature, and the other
symbols are as defined above. Equation (3), which gives the spatial distribution
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Mechanistic Characterization of the Process 329
UL 20
10
0
3.00 3.04 3.08 3.12 3.16 3.20
t (s)
":E
(a)
0.6
0.2
0
-0.2
3.00 3.05 3.10 3.15 3.20
+ +
Figure 8. Thermal profile characterization of the reaction Zr 1.5B 15 weight-% TiB,: (a) the
time dependence of the reaction rate, aq/dt, (b) the time dependence of the conversion factor, q [lo].
of the conversion parameter, can be differentiated with respect to time to give the
reaction rate, &/at, which is related to x by d v / d t = u ( a q / a x ) .Utilization of this
(the Boddington-Laye) method of kinetic analysis has been made in investigations
on the combustion synthesis of zirconium boride, molybdenum silicide, and tita-
nium silicide [lo, 111. Figure 6(a) shows a typical profile during the reaction of
+ +
Zr 1.5B 15weight-% TiB2 and Fig. 6(b) shows an enlarged section of the
initial portion of the profile. A smoothed version of the latter as well as its first and
second derivatives are shown in Figs 7(a) to (c), respectively. From Fig. 7(b), the maxi-
mum heating rate at the leading edge of the wave is calculated as 1.6 x lo5K s-'.
Using Eq. (3), the reaction rate and the temporal distribution of the conversion
parameter can be calculated, as shown in Figs 8(a) and (b), respectively. It can be
seen from Fig. 7(b) that the conversion is complete in about 8 x lO-*s and with
the knowledge of the wave velocity, the width of the combustion zone is calculated
as about 0.4 mm.
Using the calculated reaction rates and their dependence on temperature, activa-
tion energies for the formation of zirconium boride can be calculated, as shown in
Fig. 9. The lines represent data for different levels of conversion. Those representing
7 values of 0.6, 0.7, and 0.8 have similar slopes. The lack of conformity of the lines
for q of 0.5 and 0.9 is related to the uncertainties in the process of obtaining
derivatives at these ends of the conversion profile, as explained elsewhere [lo].
The average value for the activation energy for the combustion synthesis reaction
Zr + 1.5B + XTiB2 (with X between 0.15 and 0.22) was calculated as 144.5 x
15 kJ molF' . The activation energy was also calculated from measurements of the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
330 2 Self Propagating High-Temperature Synthesis of Hard Materials
4 -
g3 -
5 * F \ ,
5 2 -
I -
t 0.9
0 1 I I I
3 4 5 6 I
1 0 ' (IK)
~
Figure 9. Arrhenius plot of the temperature dependence of the reaction rate for Zr + 1.SB+XTiB,
for 7 values from 0.5 to 0.9.
temperature dependence of the wave velocity, shown in Fig. 10. The resulting value,
148.6 x 12kJmol-', is in good agreement with that determined from the profile
analysis, indicating that the reaction was complete within the relatively narrow
reaction zone.
Similar analyses were made for the synthesis of MoSi2 and Ti2Si3 [l 11. In these
cases, particularly for MoSi2, the reaction is not complete within a narrow zone
as can be seen from Fig. 11. Here the width of the zone is calculated as
approximately 1.3mm. The wide reaction zone for molybdenum silicide is, in
part, the consequence of a relatively low enthalpy of formation as manifested by
an adiabatic temperature near the empirically established lower limit. This is the
reason why some composites of MoSi2 cannot be directly synthesized by SHS
without some form of activation.
-7.8
E = 134 W/mol (Qconstant) 1
-8.0
r.
F
-3
e
-8.2
d
C
-8.4
-8.6
0 E = 148 Wimol (Qvarying)
I I I I
-II
3.4 3.5 3.6 3.7 3.8 3.9
1 0 ' (~1 1 ~ )
Figure 10. Arrhenius plot of the temperature dependence of the wave velocity in the reaction
Zr = 1.5B+XTiB2 [lo]: 0, from velocity measurements; 0 , from temperature profiles.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Efect of Experimental Parameters 33 1
0.9
0.8
0.7
0.6
0.5
Er
0.4
0.3
0.2
0.1
0.0
-0.1
2.4 2.5 2.6 2.7 2.8 2.9
t 6)
Figure 11. The time dependence of the conversion profile, 7, for the synthesis of MoSi,
H298 +
TO
1298
(a + bT)d T = / Tad
TO
(up + b p )dT, (4)
where H298is the enthalpy of formation of the product at 298 K, A u and Ab are the
algebraic sums of the Cp coefficients of the product and reactants (assuming a
Cp = a + bT form), ap and b, are the coefficients for the heat capacity equation
of the product. In the case of TIC, shown in Fig. 12(b), the adiabatic temperature
(for To = 298 K) is at the melting point of the product and increasing To initially
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
332 2 Self-propagating High-Temperature Synthesis of Hard Materials
3340 -
-
3320
3300 -
8 3280 -
2 3260 -
3240 -
3220 -
A
3200 -
I
400 800 1200 1600 2000 2400 2800
To(K)
(b)
Figure 12. (a) Schematic representation of the influence of the initial temperature on the adiabatic
temperature: His the enthalpy and T is temperature. (b) The dependence of the adiabatic tempera-
ture on the initial temperature for TIC [15].
has no effect on Taduntil all the product is in the molten state. Further increase in To
will result in an increase in the adiabatic temperature. The preheating of reactants
before ignition is a method employed in the synthesis of materials whose adiabatic
temperatures are relatively low. This is to be distinguished from the practice of heat-
ing the reactants (e.g. in a furnace) until the entire sample reacts. This latter form of
combustion synthesis has been erroneously referred to as SHS, but, in fact, it has no
discernible combustion wave and it is more appropriately referred to as the simul-
taneous (or volume) combustion form of combustion synthesis. For this form of
combustion synthesis, the rate of heating to the ignition temperature plays a
major role in the thermal behavior of the process and the nature of the resulting
products [16]. Lower heating rates result in the formation of precombustion
phases through interfacial diffusion with the consequence of a reduced driving
force in the main (combustion) reaction [17,18].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
---
2.3 Effect of Experimental Parameters 333
h
Y
Y
1800
e3
-
c
E i
-
1700
E
I-
1600 -
.-
c
m
n
m
*e I500 -
U
In principle, the roles of stoichiometry and dilution are similar for product phases
with a narrow compositional stability range. For example, in the synthesis of MoSi2
+ +
and Sic, changing x in Mo (1 x)Si and Si (1 x)C has the same effect as + +
adding a diluent, as seen in Fig. 13, which shows the effect of stoichiometry on
the adiabatic temperature. This, however, is not the case when the product phase
is thermodynamically stable over a range of composition, as is the case of many
hard compounds (e.g. TIC, ZrC, NbC, TIN, ZrN, and others). For such materials,
the enthalpy of formation is a function of composition and thus changing the
stoichiometry of the reactants has the effect of changing the heat of the reaction
as well as the stoichiometry of the product. Using a modeling approach in which
the product is presumed to have a fixed nominal stoichiometry [19], the influence
of the mixture ratio of the nonmetal to the metal on the combustion wave velocity
was calculated and compared to experimental results, as shown in Figs 14(a) and (b)
for synthesis in the Ti-B and Zr-B systems [20-251. In these cases, increasing the
value of p from about 0.3 to 1.0 has the effect of reducing the degree of dilution
by the excess metal in the starting mixture. An experimental investigation on the
effect of stoichiometry on the combustion synthesis of TIC was recently carried
out by Makino et al. [26]. Figure 15(a) shows the dependence of the existence and
nature of combustion waves on the C/Ti ratio (p) for samples (1Omm diameter
and 35mm length) with initial relative densities ranging from about 55 to 95%
and a carbon particle (diameter) size of 20pm. Steady-state combustion was
observed in the range 0.47 5 p 5 1.4 and a pulsating mode of combustion was
observed for 1.4 5 p 5 1.9. For p > 1.9 and p < 0.47, the samples either were
ignited but the wave became extinguished, or the samples could not be ignited at
all. As the figure shows, the relative density had little influence on these boundaries,
The observed lower limit for steady-state propagation, p M 0.47, coincides with the
lower limit of phase stability in the Ticl p.y or about 32 at% C. On the other hand,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
334 2 Self-propagating High-Temperature Synthesis of Hard Materials
Ti - B
1 Zr- B
0 0.5 1 .o
P
(b)
Figure 14. Modeling results of the influence of the mixture ratio of nonmetal to the metal, p, on the
combustion wave velocity in the: (a) Ti-B system, (b) Zr-B system [19]. Tois the initial temperature:
data from [20-251.
the corresponding upper limit, p = 1.4, falls beyond the stoichiometric limit of TIC
phase and thus the product must contain excess carbon. The dependence of the
velocity on the stoichiometric ratio is shown in Fig. 15(b). The rate of propagation
increases to a maximum corresponding to the limit of steady-state propagation, that
is, the highest velocity (reaction rate) is obtained under a condition of excess carbon.
It has been suggested that excess carbon enhances the kinetics of the reaction
between liquid titanium and solid carbon [27].
The effect of dilution on the synthesis reaction of hard (refractory) materials has
been examined experimentally and through modeling. Figure 16 shows the effect of
adding TIC as a diluent on the adiabatic temperature and the fraction melted of the
product [ 151. The adiabatic temperature remains constant with the addition of TIC
up to 20 mol%, at this point the product is entirely solid and further addition of the
diluent decreases the adiabatic temperature as shown in the figure. The decrease in
the combustion wave temperature is expected to result in a decrease in the wave
velocity. Experimental confirmation of this is provided in Fig. 17 for the case of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Effect of’E.xperimcnta1 Parameters 335
I Specimen diameter=lO mm
0 Steady
0 Pulsating
OS
I
t I
A Extinction
X No ignition
0 1 2 3
15
- P*I
A 0.55-0.65
00.65-0.75
-f
fl
3 0 0 0.75-0.85
10 - 0 OM-0.95
0
.-
-
0 0 .
m *
a&
.E 5 -
.*
mE Open: steady
Filled: pulsating
0 1 2 3
Mixture ratio, p
(b)
Figure 15. The effect of the mixture ratio C/Ti (= p) on: (a) the relative density and nature of the
combustion mode, (b) the velocity of the wave for various relative densities of the reactants [26].
3300 160
3200 50
k
40
s. 3100
h
h
3000
i-”
20
2700 0
0 10 20 30 40 50 60 70
TIC (mol%)
Figure 16. The influence of dilution (by TIC) on the adiabatic temperature and fraction of product
melted (v) in the synthesis of TIC [15].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
336 2 Self-propagating High-Temperature Synthesis of Hard Materials
p = 1.0 d,. W)
h
Borovinskaya
E Present
.- 0
&el= 0.56
0
.-
0
m v
0.0 0.5
Degree of dilution
Figure 17. The effect of dilution on the wave velocity in the synthesis of TIC [26]. Experimental
results from [27].
where vmaxis the maximum (adiabatic) velocity and wmin is the minimum velocity
before the wave becomes extinguished. Equation (5) was used with the relationship
between velocity and temperature,
w2 = ( K RT2/QE)exp(-E/RT), (6)
where Q is the heat of the reaction, E is the activation energy of the process, T is
temperature, R is the gas constant, and K is a constant. The resultant relationship,
which provides boundaries for the existence of SHS, is derived as
where T,, and Tminare the maximum (adiabatic) temperature and the minimum, that
is, the temperature of the wave just before it becomes extinguished, and d is the degree
of dilution [4,5]. Using this relationship, SHS diagrams for TiB,, H B 2 ,ZrB2, and TIC
have been constructed. An example is shown in Figs 18(a) and (b) for TiB2 and Tic,
respectively. Details of the experimental results and the theoretical derivations of this
model as well as others are given in the literature [4,5].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Ejfect of Experimental Parameters 331
1400
TiB,
0.
400
200
0 20 40 60 80 100
-
(E=398KJ/Mol) -/ / \
3
1200
g 1000 - This Work
(E=398KJ/MoI)
800 -
600 -
400 - (E=l38KJ/Mol)
200 '
0
I
0.2
I
0.4
I
0.6
I
0.8 1.0
d
(b)
Figure 18. Theoretical limits of SHS (SHS diagram) and experimental results for: (a) TiB2, (b) Tic
[4,51.
The effect of particle size of one or more of the reactants on the SHS process has
been examined in experimental and modeling studies. In general, an increase in
particle size results in an increase in the diffusion length and hence a decrease in
the reaction rate, with the rate expected to be inversely proportional to particle
size. Figure 19 shows experimental observations on the relationship between the
particle size of titanium and the wave velocity in the synthesis of TIC and TiB2
[21]. This trend is also predicted by modeling studies of Makino and Law, showing
the anticipated inverse dependence of the velocity on particle size [29]. Their
prediction of this dependence is compared to results of experimental observations
in Fig. 20 for the synthesis of TIC [26]. The wave velocity decreases by a factor of
about four as the carbon particle size increased from 1 to 25pm. However, the
particle size of Ti had no apparent influence on the velocity within the range of
10-75 pm. The fact that titanium is molten in the combustion zone is believed to
provide the explanation for the lack of a dependence. The experimental results of
the trend of velocity with carbon particle size provide an unexpected observation
at the small particle size end. When small (<5 pm) carbon particles were used, the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
338 2 Self-Propagating High-Temperature Synthesis of Hard Muterials
24 1 I
i
.
h
:
v 12
>
n
”
0 0.1 0.5 1.0 1.5
r, (mm)
Figure 19. Dependence of the combustion rate of Ti + 2B and Ti i-C mixtures on the Ti particle size
for several initial temperatures (To):0 , To = 20°C; 0, To = 200°C; 0,
TO= 400°C [21].
wave propagated in a pulsating mode, albeit at a much higher average velocity. The
significance of this, as was pointed out earlier, is that the products of non-steady-
state combustion are inhomogeneous. In another investigation, the size of the Ti
particles was shown to have little effect on the density of the TIC product, but it
had an effect on the resulting microstructure [30].
Another important experimental parameter in combustion synthesis is the relative
density of the reactants. Figure 21 shows the dependence of the wave velocity on
density in the combustion synthesis of TiNi [31]. These and similar experimental
p = 1.0
30 d&mm) P..,
0 Deevi et al. 1.6 0.56
1.0 0.52-0.60
2”E
h
25
-
”
25 ”
20
B
-
0
d)
Open: steady
Filled pulsating
I I I I
0 10 20 30 40 50
Diameter of C particles, d, (pm)
Figure 20. The effect of the particle size of carbon on the combustion wave velocity in the synthesis
of TIC [26].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Effect of E.uperimentul Purameters 339
1.5
1.o
0.5
0 20 40 60
Porosity ( % )
Figure 21. The effect of porosity on the wave velocity in the synthesis of TiNi [31].
results show a relatively common pattern in which the wave velocity and com-
bustion temperature have maximum values at an optimum relative density [213.
The relative density has three roles in the SHS reactions. It influences: the energy
density (e.g. J ~ m - of~ )the reaction; the kinetics of the reaction through changes
in the particle-particle interfacial contact; and the wave propagation dynamics
through changes in the effective thermal conductivity of the reactants. At low rela-
tive densities, the velocity and temperature are expected to be low because of all the
three factors listed above. On the other hand, at high relative densities the effective
thermal conductivity is so high that heat is conducted at such a high rate to distant
regions ahead of the wave making it impossible to reach the ignition temperature in
the layer immediately ahead of the wave. Although the existence of a maximum is
not always observed experimentally, the concept of wave extinction at high densities
is always verified. Highly dense reactant powders cannot be ignited. The role of
particle size in this behavior is shown by study on the synthesis of TIC, Fig. 22
[26]. Here the particle size of carbon is changed while holding the size of the Ti
particles constant at 25 pm. With small carbon particles (1 pm), a maximum in
40
4(w)p , ( w
0 Shkiro and
3 f AA Borovinskaya 0. I 20
h
A A Present 1 25
30 8 0 10 25
v
v " 25 25
0
.-
0
I OA 0
20
>
M
'gm 10
"
n
0.4 0.5 0.6 0.7 0.8 0.9 1.0
Relative density
Figure 22. The effect of relative density of the reactants on the wave velocity in the synthesis of TIC
P61.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
340 2 Self-Propagating High-Temperature Synthesis of' Hard Materials
Figure 23. The dependence of the degree of completion of conversion to TiN on the pressure of
nitrogen gas in preexisting pores: solid lines, 1000 K; dashed lines, 1500 K [32].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Effectof Experimental Parameters 341
10' ' I
4.0
I
6.0
I
8.0
I
10.0
I
12.0
I
14.0
V,"(cm')
Figure 24. A generalized relationship between the pressure of nitrogen and the molar volume of
elements with useful nitrides at complete conversions. T = 1500 K, porosity = 0.5. Nitride
phases: A = M3N4,B = MN, C = M,N,, D = M,N [32].
a pressure, since Eq. (8) does not account for the kinetics of the process. It only
provides an indication of the significance of permeation in the synthesis of the
nitrides. This can be better demonstrated by Fig. 24 in which the pressure for com-
plete conversion is plotted as a function of the molar volume of the metal and
shown for homologous series of nitrides. It can be seen from this figure, for
example, that the nitriding of Hf and Zr (to form HfN and ZrN) is less affected
by permeation than B (to form BN).
Experimentally, the role of porosity (relative density) in nitride formation by
SHS is complex and depends on phase transformation (as it affects permeation)
and the diffusion of nitrogen atoms through the metal or the nitride. The influence
of the former relates to the melting of the metal. If the combustion temperature
exceeds the melting point of the metal, then permeation is affected by the
concomitant change in the continuous porosity. Such a behavior is shown in
Fig. 25(a) where the degree of conversion to TIN is plotted against the relative
density of the starting Ti powder for synthesis in one atmosphere nitrogen pressure
[33]. Maximum conversion occurs at a relative density of about 55% and the
decrease with subsequent increase in density is due to a decrease in the local avail-
ability of nitrogen resulting from the decrease in permeability. The decrease on the
other side, that is, as the density decreases from 55%, is due to the melting of Ti
and its influence on the open porosity in the Ti reactant. For this end of the curve,
the surface area of the Ti in contact with gaseous nitrogen is high, leading to a
higher reaction temperature and more extensive melting of the titanium [34]. In
systems where the combustion temperature does not exceed the melting point of
the metal, the dependence of conversion on density does not show a maximum
but, as predicted, decreases with an increase in the relative density, as shown in
Fig. 25(b) for the synthesis of NbN [35]. As can be seen from Figs 25(a) and
(b), complete conversion to the nitride phases is not accomplished under the
experimental conditions of one atmosphere of nitrogen. In the case of TiN,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
342 2 Self-Propuguting High-Tetnpeniture Synthesis of Hurt1 Muterid\
+
8
a 40
I I I I I I
45 50 55 60 65 I0 15
Relative density of Ti pellets (%)
(a)
1
48
44
3
-
0
z
2
0
40
- 0 0
* 0
-
3 36
.-
v1
0
32 - 0
28 -
0
,,LLL
0
2056 57 58 59 60 61
62
Relative Density (%)
Figure 25. The effect of relative density of the metal compact on the combustion wave velocity: (a)
for the synthesis of TIN, (b) for the synthesis of NbN [33,35].
complete conversion can be achieved with the appropriate control of the melting of
Ti by the addition of a diluent.
45
5
h
,x 40
.-
A
.
+
5
*
35
rt;
30 - I I I
45 50 55 60 65
Uncombusted reactants density (YO)
Figure 26. The relationship between the relative densities of the reactants and product in the
synthesis of TIC [30].
this problem, cannot overcome the primary intrinsic source of porosity (see below).
Thus the need to densify the product through simultaneous or subsequent applica-
tion of pressure is of major concern in SHS synthesis.
The sources of porosity in SHS products have been analyzed and classified as
intrinsic and extrinsic 1381. The extrinsic sources of porosity are the initial porosity
of the reactants and the porosity generated by the expulsion of volatile impurities
during synthesis. The contribution of the latter is difficult to calculate exactly, but
an approximate relationship has been derived which shows the dependence of the
level of porosity on the amount of impurity as
up + CipmRTIMiPc, (9)
where wp is the volume of pores per unit volume of product, Ci is the concentration
of the impurity in the reactants, pm is the theoretical density of the product material,
Mi is the atomic or molecular mass of the impurity, P, is the pressure to overcome
the confining forces in the powder, R is the gas constant, and T is the absolute
temperature. Results of such an analysis for the synthesis of TIC are plotted in
Fig. 27 for selected (assumed) impurities. The calculations were made assuming a
combustion temperature of 3000 K (the adiabatic temperature of TIC is 3210 K)
and the confining pressure, P, (the pressure needed to push the compacted particles
apart), was assumed to be 200 atm. The figure shows that at an impurity level of, for
example, lOOOppm the volume percentage porosity can range from about 4.5% for
Ba as an impurity to about 26% for Mg.
Intrinsic sources of porosity are the molar volume change (due to the formation of
the product) and the porosity generated by the thermal migration. Figure 28 shows
the relationship between the molar volume of the reactants and the product for
selected refractory materials. The lines represent a value of x in the relationship
V , = xVR, where Vp and VR are the molar volumes of the product and reactants,
respectively. The figure shows that the formation of TaC, B4C, for example results
in a porosity of about lo%, while the formation of Sic, MoSi2, results in a porosity
of about 30%. The other intrinsic source porosity is thermal migration (the Soret
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
344 2 SelflPropagating High-Temperature Synthesis of' Hard Muteriuls
100
Impurity:
.-
VI
a
.s 1.o
Tic : T = 3000K
2. 0.1
1.o 10 1o2 10' 10'
Impurity level, ppm
Figure 27. The contribution of impurity volatilization to pore formation in the synthesis of TIC [38].
34
v, = xv, x= I .o
30 -
26 -
10 14 18 22 26 30 34 38
v, ( c d
Figure 28. The relationship between reactants and product molar volume for selected compounds
1381.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Synthesis of' Dense Materirrls 345
30 r-- 50 I
40
50 rl
i
10
4 6 8
log (dT/dX) (Wcm) log (dl'idX) (Wcm) log (dT/dX) (Wcm)
Figure 29. The effect of temperature gradient on the pore/grain size ratio for: (a) Ti + Ni; (b)
Zr + 2B; (c) Ta + C [40].
diameter to particle diameter was calculated and the results are shown in Fig. 29 for
the three selected examples of TIN, ZrBz, and TIC [40]. For each case, curves are
shown as a function of combustion wave velocity, Slower wave propagation and
higher temperature gradients give rise to higher porosity from thermal migration.
Fi ure 29 shows that under realistic conditions of temperature gradients (lo5-
10f Kcm - ') [lo] and wave velocities (l-lOcms-'), the ratio d p / l ranges from 2 to
10%. However, it should be noted that while thermal migration can give rise to
pore formation, experimentally it is difficult to isolate its contribution. All of the
sources discussed above, in some combination, contribute to the overall porosity
of SHS products, which can be as high as 50%.
With few exceptions, the general desire is to prepare dense materials by the SHS
process and thus several simultaneous or sequential steps are taken to densify these
materials. In the former, the synthesis and densification is achieved in one step, and
in the latter the combustion process is followed, within a short time, by a densi-
fication process. The idea is to take advantage of the plasticity of the hot product.
Examples of the sequential densification include the application of uniaxial pressure
[4 1-44], hot-isostatic pressing (HIP) [45,46], pseudo-HIPing [47], dynamic
(explosive) compaction [48-5 11, high speed forging [52-541, and pressureless and
microwave sintering [55-571. Although the relatively rapid nature of the SHS
process was expected to give rise to products with high defect concentrations,
experimental observations on the sintering of SHS produced materials have not
substantiated this expectation [58,59]. In a related investigation, it was found that
shock compressed Ti + C powders could be reaction-sintered to give a highly refined
microstructure (grain size of 6 pm) [60].
The simplest form of densification, uniaxial pressing, has been used to densify a
variety of materials such as TIC + NiAl, TIC + Ni3Al [41], T i c + Ni, TIC + Fe [42],
and others [43,44]. Two parameters play an important role in this type of post-
reaction densification: the magnitude of the applied pressure and the time of its
application relative to the completion of the combustion process. Figure 30 shows
the dependence of the residual porosity in Tic-Ni/Al cermets on the applied
pressure. As the figure shows, the relative density increases from about 50 to 99%
as the pressure is increased from ambient (0.1 MPa) to about 28 MPa [41]. The
timing of the application of the pressure relative to the synthesis reaction has also
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
346 2 SelJ Propagating High-Temperature Synthesis of Hard Materials
Pressure (MPa)
Figure 30. The effect of the applied pressure on the residual porosity in the synthesis of: 0,
TiC-
NiAI, and W TiC-Ni3A1[41].
been investigated [42,61]. Fu et al. examined the various parameters related to pres-
sure application [42]. Figure 31 shows a schematic diagram of the pressure-time
parameters. The time for wave propagation, tw, is followed by a delay time, td,
before the pressure is changed from the ambient, Po, to the desired (high) pressure,
Ph. The time duration through which the pressure is maintained at P h is designated
as t,, as shown in the figure. The effect of the delay time on the (relative) density of
the composites 75 weight-% TiB2-25 weight-% Fe and 75 weight-% TIC-25
weight-% Ni is shown in Fig. 32. The results show that for these two cases, the
highest densities are obtained with little or no delay time (td 5 1 s). In both cases,
the time at pressure, t,, was 10s but the values for P h were 95 and 80MPa for
the Fe and Ni-containing cermets, respectively. The effect of the time of application
of the pressure is shown in Fig. 33. For both cermets, the densification approaches
an asymptotic value after a hold of about one minute. Furthermore, and in quali-
tative agreement with the work of Dunmead et al. [41], the final density depended
on the level of the applied pressure, Fig. 34. The difference in behavior of the two
Time
Figure 31. Schematic representation of the sequence of pressure application during SHS synthesis
and densification [42].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Synthesis of' Dense Materials 341
1 00
75TiB1-25Fe
t,= I0 scc
98'
5
h
92
.c
Q
88
84 75TiC-25Ni
t,=l 0 sec
P = 80 MPa
80
2 4 6 8 10
td(set)
Figure 32. Effect of the delay time in pressure application on the relative density of TiB2-Fe and
Tic-Ni composites [42].
100
99
5
h
98
6 0 -
1
Q 4 0
75TiBJ25Fe
i
97 t,= 1 sec
P = YSMPa
96
I
0 20 40 60 80 100 120
t, (set)
Figure 33. Effect of the time of pressure application on the relative density of Tic-Ni and TiB2-Fe
composites [42].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
348 2 Self Propagating High-Temperature Synthesis of Hard Materials
I00
99 tp= 10 sec
98
g 97
," 96 75TiBJ25Fe
0.
t"= 1 sec
95 tp= 10 sec
93
94 I
U
40 60 80 100
L
120
Pressure (MPa)
Figure 34. The effect of the applied pressure on the final density of Tic-Ni and TiB,-Fe composites
~421.
part of this chapter. Kinetic limitations of SHS reactions are dominant when the
rate of the reaction, &1/8r, is too low or when the thermal conductivity, K , is not
optimum. Too low a K value leads to inadequate heat build up in the reactant
layer just ahead of the front and hence the absence of ignition conditions. Similar
end results are obtained when, for too high a K , the heat is diffused to regions far
ahead of the front. Because of a low reaction enthalpy (thermodynamic limitation),
materials such as B4C, Sic, WC, Cr3C2,Mo2C, and others cannot be synthesized by
a normal SHS process. Kinetic limitations have been given as reasons for the
inability to synthesize such materials as TaC. AlN-SIC, and others. Many com-
posites of hard materials where the formation of one of the involved phases is
energetically lethargic cannot be synthesized by SHS because of these limitations.
These include such materials as MoSi,-SiC, .xB4C-TiB2 (for x > 0.5), and others.
The common approach to activate these reactions is to preheat the reactants in
order to either raise the adiabatic temperature for a subsequent SHS reaction or
to cause a reaction in a simultaneous combustion mode, as discussed above. This
is done through furnace heating or by a ‘chemical oven’ [67]. The latter involves
the placement of the reactants within a powder mixture of other more exothermic
reactants. However, in both of these methods diffusional reactions can occur
before the main combustion reaction and can result in the formation of extraneous
(undesirable) phases in the product [68,69]. To overcome the limitation of the SHS
process and thus facilitate the synthesis of a large number of materials systems, a
new process was recently developed in which the activation is effected by the
imposition of an electric field [70]. A schematic representation of this method of
activation is shown in Fig. 35. Two electrodes are placed in contact with opposite
faces of the compact containing the reactants and a voltage is applied across
them. In most cases, the application of a field (within limits) does not result in the
initiation of a combustion wave. Such a wave, however, can be initiated and
propagated after the addition of energy from an ignition source, in the usual way.
In all cases investigated, the initiation of a combustion wave is not accomplished
below a minimum (threshold) value of voltage (or voltage per sample thickness,
Product Reactants
0 5 10 15 20 25
E (V/cm)
Figure 36. The effect of the applied field on the combustion wave velocity in the synthesis of p-Sic.
field). Figure 36 shows that in the case of the synthesis of Sic, the threshold is about
6Vcm-' [71]. The figure further shows that there are three regions in the depen-
dence of wave velocity on field strength. As indicated, below the threshold no
wave propagates, and above the threshold the wave velocity increases nearly linearly
with field. At relatively high field values, above about 20Vcm-', the reaction
between Si and C does not require an ignition source but takes place in a simulta-
neous mode. In this case the Joule heating by the current is sufficient to ignite the
reaction. As will be discussed in more detail, the role of the applied field is believed
to be primarily thermal in nature, providing additional (Joule) energy. In most
cases, this energy is localized to the narrow reaction zone, as will be shown later.
Real-time measurements of the changes in voltage and current during wave
propagation provide the results shown in Fig. 37 for the synthesis of Sic [71].
Both the measured parameters of current and voltage and the calculated parameter
of resistance remain reasonably constant during wave propagation, as indicated by
the interval between the start (S) and end (E) of the process. In view of the fact that
the electrical conductivities of the reactants and product are significantly different,
these observations are taken as indicative of a current that is confined to the com-
bustion wave. Modeling and other experimental studies have provided support for
this conclusion, as will be discussed below.
The modeling study was based on a modification of the applicable Fourier rela-
tionship, Eq. (1). The modification includes the addition of another heat source
from the Joule heating by the current. The rate of heat generation from this
source is uE2, where u is the electrical conductivity and E is the applied field. The
modeling was initially done on the synthesis of Sic, using a two-dimensional
approach and accounting for heat loss by radiation, to bring the results closer to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
’7.4 Syirtliesis h!, Field- Activuted Self+vpagating High-temperature Synthesis 35 1
20 , 1
15
d
‘c)
10
>
-
5
0
0 5 10 15
time (s)
Figure 37. Variations of real-time voltage and current and calculated resistance during wave
propagation in the synthesis of P-SiC [71]: S and E signify the start and end of wave propagation,
respectively.
0
x (cm)
Figure 38. Calculated conversion profiles for the synthesis of S i c with applied voltage 20V [72].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
352 2 Self-Propagating High-Temperature Synthesis of Hard Materials
(E=14.2 V/cm)
confined to a very narrow zone; and the confinement zone is located in the leading
edge of the combustion wave. The confinement of the current to the reaction zone is
an obvious advantage when the energy efficiency of this form of activation is consid-
ered. However, this situation is not universally true but is dependent on the electrical
conductivity of the product relative to the reactants. The implication of this is that in
systems where the product is more conducting, the current is not confined to the
reaction zone but is, at least partially, carried by the product. Both experimental
and modeling studies have provided support for this expectation [6].
The application of a field has been shown to have the following effects in combus-
tion synthesis: it makes possible the synthesis by SHS of materials which, previously,
could not be prepared by this method [7&74]; influences the nature and the relative
amounts of product phases [75,76]; makes possible the synthesis in relatively dense
reactants [77]; and influences the microstructure of the product [78]. A few examples
of these will be discussed here and more details on the others are provided in the refer-
enced papers. An example of the benefit of the field is demonstrated by the synthesis of
B4C-TiB2 composites. With a goal of producing hard and light materials for armor
application, the desire is to increase the relative content of the carbide phase in this
composite. With normal SHS, composites with a B4C/TiB2molar ratio, y , of 50.5
can be synthesized without preheating the reactants [79]. While composites with
higher values y can be synthesized by preheating the reactants, the product contained
other extraneous phases (e.g. TIC). With application of a field, composites with y
values up to 5 could be synthesized [80]. Figure 40 shows the effect of the field on
the wave velocity in the synthesis of B4C-TiB2 composites with 1 5 y 5 5. As in
other field-activated combustion synthesis, a minimum (threshold) field value was
necessary to propagate the combustion wave. The threshold value depended on the
value of y , as seen in Fig. 41. Higher threshold values were required for higher
molar ratio 01) values. The products in all cases contained the two desired phases
only. Since the imposition of the field is a requirement for the synthesis of composites
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.4 Synthesis by Field-Activated Selfpropagating High-temperature Synthesis 3 53
H B,C/TiB,=I
0 B,CmiB,=2
0 B,C/TiB,=3
0 B,C/TiB,=5
0 10 20 30 40 50 60
Voltage ( V )
Figure 40. Effect of field on the wave velocity for the synthesis of yB4C-TIC composites with
1 5 y 5 5 [SO].
with y > 0.5, the removal of the field during wave propagation affords the oppor-
tunity to investigate the mechanism of formation of the two phases from elemental
reactants. Figure 42 shows the X-ray diffraction (XRD) patterns obtained from a
B4C-TiB2 composite with y = 1 in which the wave was frozen by the removal of
the field [80]. The upper pattern was obtained from the region behind the combustion
zone, and as can be seen it contains only the peaks for B4C and TiB2, the product of
synthesis. The middle pattern is for the combustion zone itself and shows the presence
of TiB2 and carbon (no peak for boron is expected since it was used as an amorphous
powder). Finally, the lower pattern is for the area ahead of the wave and shows the
presence of the reactant elements only. These observations indicate that the
formation of TiB2 is the first step in the composite formation, an expected
conclusion in view of the large difference in the thermodynamic driving force for
the formation of the two compounds. It should be recalled that B4C cannot be
formed without field activation.
2 2
F
Y
2 0
1 8
m
-o
C
1 6
>
8
0 1 2 3 4 5 6
B,C/Ti B, Molar Ratio
Figure 41. Effect of the molar ratio B,C/TiC on the threshold voltage in the combustion synthesis of
B,C-Tic composites.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
354 2 SelfiPrapagating High-Ternperuture Synthesis of Hurd Materials
reacted region
300 TiR,
0
25 30 35 40 45 50 55 60 65
2 theta
Figure 42. X-ray diffraction patterns of frozen reaction zone with analyses for the reacted region,
combustion zone, and unreacted region [80].
Another example of the role of the field in synthesis is that related to the prepara-
tion of tungsten silicides [75]. The two stable silicides, WSi2 and W5Si3, have
relatively low adiabatic combustion temperatures (1459 and 954 K, respectively)
and thus cannot be synthesized by SHS. However, both can be synthesized by
field activation. For WSi2, a combustion wave can be initiated above a threshold
value of about 15Vcm-' and the product is the desired phase. On the other
hand, the case of W5Si3 is more complex. Although a combustion wave can be
initiated above a threshold field of about 14Vcm-', the product does not contain
the desired phase unless the field is above about 19 Vcm-I. Microstructural obser-
vations on the products for the case of W5Si3showed that under the higher field the
product contained a liquid phase and it was proposed that the formation of the
desired phase was aided by the higher diffusion rates in the liquid phase.
The formation of composites and solid solutions in the binary A1N-SIC offers
another important example of the role of the field in the synthesis of materials.
The two compounds form extensive solid solutions with the hexagonal 2H structure
above about 1960°C [81]. The spinodal decomposition of these, at lower tempera-
tures, gives rise to the formation of composites modulated at the nanoscale [82].
The common method to synthesize the solid solutions is to heat mixed powders
of AlN and Sic (aor p) at 2100°C for up to 16h. As will be seen here, using
+
field-activated synthesis for the reaction S3N4 4A1+ 3C, the formation of solid
solutions can be effected in less than a minute [76,83]. Although this reaction is
highly exothermic (with an adiabatic temperature of 2504 K), it does not sustain a
self-propagating combustion wave without the application of a field above a
threshold value of 8.0 V cm-' . At this field, however, the reaction is not complete,
with the product containing SiC-rich and A1N-rich 2H solid solutions and Si and
C . The reaction becomes complete at a field of 12.5Vcm-' but the product is a com-
posite of two solid solutions. When the applied field is 25.0 V cm-I, the product is a
single phase solid solution. Careful analysis of the product phases showed that their
compositions changed with the strength of the applied field, as shown in Fig. 43. As
the field increased, the compositions of the two phases moved closer to each other
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.4 Synthesis hy Field-Activated SelJlpropagating High-temperature Synthesis 3 55
0.8
0.6
zd 0.4
0 WDS-AIN
o WDS-Sic
0.2 A XRD-AIN
x XRD-Sic
n I I
12.5 20 25 30
Field (V/cm)
Figure 43. Effect of the applied field on the phase composition of combustion-synthesized AIN-SIC
composites and solid solutions [76].
Figure 44. X-ray dot maps of A1 and Si in AIN-SIC solid solutions prepared under fields of 25 (left)
and 30Vcm-' (right) [76].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
356 2 SeIf-Propagating High-Temperature Synthesis of Hard Materials
and at 25 V cm-' , the compositions are the same, indicating the formation of the solid
solution. Increasing the field to 30 V cm-', had no effect on the XRD generated com-
position, but a more significant effect on the elemental distribution, as seen in Fig. 44.
Although the XRD results indicate that the product with 25 V cm-' is single phase,
the elements (A1 and Si) are not uniformly distributed in the sample. However, a
nearly complete uniform phase is obtained when the field was 30 V cm-' . These obser-
vations are further evidence to the role of the field in the synthesis of materials.
that of TIC (3140°C) is not likely to be reached during real (i.e. nonadiabatic)
conditions. Moreover, it was observed that the grain boundary strength of the
TIC and TiB2 prepared by the SHS-dynamic compaction method were relatively
weak. This conclusion was based on observations of grain boundary failure
during hardness measurements and was explained by a formation of microcracks
along grain boundaries during the rapid cooling. Support for this conclusion was
provided by significant improvements in mechanical properties after annealing for
several hours at 1500 and 2000°C. As pointed out above, the existence of cracks
is related to the rate of cooling subsequent to synthesis and densification. The use
of a pressure medium in the 'pseudo-HIPing' approach is believed to ameliorate
this problem by decreasing the rate of cooling of the densified sample [65].
In order to improve the mechanical properties of such monolithic hard materials,
a second phase is usually added. In many cases the second phase is a metal and thus
the formation of a cermet is accomplished. The addition of a ductile phase acts to
reduce the residual stress and to increase the fracture toughness of the material.
The characteristics of the combustion processes aimed at the formation of metal-
ceramic composites based on T i c and TiB2 have been investigated in several
recent studies. Capaldi and Wood [88] investigated the SHS processes of syn-
thesizing TIC cermets with the metallic component being iron, nickel, copper, and
manganese. Notable differences in the process characteristics were observed,
depending on the nature of the metal additive (binder). Figure 45 shows the
dependence of wave propagation velocity on combustion temperature for all four
metals. These results imply differences in the activation energy of the synthesis
process, see Eq. (6). Contrary to what may be attributed to the role of the metal
as an inert component, these results suggest that the metal plays a more direct
role in the synthesis. Thus differences between the metals in terms of influencing
the kinetics of the process may arise from possible differences, for example, in the
formation of intermediate intermetallic phases, or differences in the kinetics of a
dissolution-precipitation process responsible for the formation of the ceramic
component [88]. The possibility of formation of low melting phases, represented
-8.0
-8.2
-8.4
5
9
-8.6
-2 -8.8
3 -9.0
-2
u
-9.2
-9.4
-9.6
-9.8
- 10.0
3.5~10' 4.0~10' 4 . 5 ~ 1 0 5.0~10'
~ 5.5~10.~
I/T, (I/K)
Figure 45. Apparent activation energy determination for Tic, Fe-Tic, Ni-Tic, and Cu-Tic [88].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
358 2 Self-propagating High-Temperature Synthesis of Hard Materials
by eutectics between Ti and the binding metal, can alter the kinetics of the dis-
solution-precipitation process involved in the combustion synthesis of carbides.
Another difference in the synthesis of these cermets is represented by the variation
in the ‘ignition temperature’. In some cases, for example in the system Ti-C-Fe, the
ignition temperature is well below the melting point of the pure Ti metal, an
occurrence which has been subject to much debate in the early SHS literature.
The prevalent point of view has been that ignition occurs after the melting on the
metal. In the work of Capaldi and Wood [88], however, when iron was used as
the binder, titanium and iron react in the solid-state producing FeTiz which is a
eutectic compound with a melting point of 1085°C. In the case of the use of Mn,
the T i c that formed was carbon-deficient relative to that formed with other
metals, probably because of the formation of manganese carbide.
The synthesis of cermets in dense form have been reported by several authors. La
Salvia et al. reported the results of high-speed forging of TiC-Ni [54] and TiC-Ni-
Mo [89] cermets. In both cases the high speed load was applied 5-10 s after the end
of the combustion process, when the samples’ temperature was still above the
ductile-to-brittle transition. In the first case cermets containing 5 , 25, and
30 mol% of the metal were prepared. The microstructure was generally character-
ized by spheroidal grains of T i c surrounded by a discontinuous Ni layer probably
due to poor wetting of melted Ni on Tic. A substantial fraction of TIC particles
exhibited bonding and necking, forming a continuous carbide skeleton. The samples
with the lower metal content exhibited a certain porosity due to partial vaporization
of Ni due to the high reaction temperature. The hardness and compressive strength
of these samples were in the range of values of conventionally processed cermets.
This indicates that the bonding between Ni and T i c is excellent. The addition up
to 10 weight-% of Mo [89] was based on the suggestion that its presence improves
the wettability of the metallic phase on Tic. In traditional processing this
improvement in wettability decreases drastically the tendency of T i c grain
growth, reducing the possibility of carbide-carbide grain contacts. In the
combustion-synthesized/dynamically-compacted materials this effect was not so
evident. Figure 46 shows that the influence of Mo content on the size of the TIC
particles is not significant. On the other hand, evidence for incomplete solubility
of Mo within the matrix is observed. However, a decrease in the number of micro-
structural flaws such as voids, interphase debonding, and binder microcracks,
attributed to the improved wetting of the Ni-alloy binder on the carbide phase, is
observed with an increase in Mo content. As a result, the compressive strength,
bend strength, fracture toughness, and Young’s moduli, were slightly improved,
while the Vickers microhardness seemed to be independent of the Mo content.
Another example of synthesis of Tic-based compact cermets is reported by Xing
et al. [90]. These authors describe the synthesis of dense NiA1-20 vol-% T i c com-
posites through a reaction of a mixture of Ni, Al, Ti, and C powders in a hot
press. The reaction was performed using the volume combustion mode: heating
the green mixture until a reaction is observed in the entire sample. The samples
were heated up to 1500°C under the influence of an applied pressure. The products
were 98.9% dense but the T i c particles (0.2-1 pm) were not uniformly dispersed in
the matrix. The combustion process had similar ignition characteristics to the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Selected Recent Examples of Synthesis of Hard Materials 359
E
I4.5
I
,
al
.*v1
-.U
0
z I I
4.0 -
.g 3.5 -
3.0 i
0 2 4 6 8 10 12
Mo (wt%)
Figure 46. Influence of the amount of Mo on carbide particle size for combustion synthesized/
impacted forged Tic-Ni based materials [SS].
system Ni-A1, with an ignition temperature of 588°C. This temperature is below the
lowest eutectic temperature in the Ni-A1 system, suggesting that the combustion
process commences through a solid state reaction.
A different approach to the synthesis of dense hard materials is reported by
Degnan and Wood [91]. They describe a process in which Fe-(W,Ti)C cermet
powders, produced by the SHS method, are used as additives to molten iron for
the production of a particulate-reinforced material with improved wear properties.
The presence of the additive provided an enhancement of wetting between the metal
matrix and the added particles. The product was a fairly uniform material which
showed a considerably improved wear resistance over iron-based alloys. Sharivker
et al. [92] followed a somewhat similar route by using SHS-produced TIC or
Tio.97M00.03C as additives to produce metal-based hard materials.
An interesting exception to the synthesis of dense materials through external
pressure application is the work of Ohyanagi et al. [93] in which the formation of a
surface dense TIC by a pressureless process is reported. Such a material was obtained
using reactants with high green densities (6549%). The formation of a fully dense
surface layer appears to be the result of the solidification of a supercooled solid solu-
tion of carbon in molten Ti. The presence of a nonequilibrium solution is suggested
by the existence of an exothermal peak on the cooling curve which became evident in
samples with green densities greater than 76%. The temperature corresponding to
the maximum of this peak decreased when the cooling rate was increased. The sur-
face became completely dense with a texture typical of solids obtained from solidifi-
cation of supercooled liquid. The presence of the liquid only at the surface was
explained by a combination of capillary and gas expulsion effects [93].
In the case of borides, the production of pure dense materials is more difficult
because TiB2 does not deform plastically even at very high temperatures due to
its intrinsically high concentration of Peierls barriers to dislocation movement.
Recent investigations on borides have been totally devoted to the synthesis of
composites either metal-ceramic or ceramicseramic. Woodger et al. [94,95]
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
360 2 Self-Propagating High-Temperature Synthesis of Hard Materials
2000,
h
5 1900
:1800
*
$
g2 1700
c
.gs 1600
D
u
3 1500
1400- I I
20 30 40 50 60 70 80
Percent relative green density
Figure 47. Influence of green density on the combustion temperature for the synthesis of Al-HfB2
composites for samples with 60 vol-% Al [94].
h*
$
v 0.35
.-3
-0
0
.>”
0.30 -
s
0.25 --
20 30 40 50 60
Percent relative green density
70 80
Figure 48. Influence of green density on the wave velocity in the synthesis of Al-HfB2 composites
for samples with 60 vol-% Al [94].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Selected Recent Examples of’Synthesis of Hard Materials 361
1600 1 T
92 1550 -
f 1500 -
2 1450
2 -
C
2
.3
1400 -
3E
P
-
1350
I , I I I
35 40 45 50 5 5 60 65 70 75
Percent relative green density
Figure 49. The influence of green density on the combustion temperature for the synthesis of Ni,Ti-
TiB2 composites. Samples containing 70 vol-% Ni,Ti [95].
effects, due to the large density difference between A1 and HfB2, were observed
together with the presence of small amounts of Hf-A1 intermetallic phases.
In the case of TiB2,cermets with up to 90 vol-% Ni3Ti have been synthesized from
elemental reactant powders. An unusual dependence of the combustion temperature
on relative green density was also observed, Fig. 49. The bimodal nature of the
combustion temperature curve was explained by a two-mechanism process. The
first peak in combustion temperature is likely to be related to the highly exothermic
TiB2 reaction in the presence of a small amount of diluent metals. The second peak
can be associated with the improved reaction kinetics resulting from the enhanced
capillary spreading of the molten phase. The presence of the molten phase makes
possible the synthesis of samples with very high relative densities. As shown in
Fig. 50 a rapid increase in relative density with metal content is observed up to
85% relative density, then this is followed by a marked decrease, probably related
to the decrease in the reaction temperature as a consequence of increased dilution
of the reacting mixture. The synthesis of a similar cermet has been reported by
Yang and Petric [96] and Yi and Petric [97] who used elemental mixtures to syn-
thesize Ti3A1(Nb)-TiB cermets. In this case the combustion process was reported
to be initiated by the reaction between Ti and Al, which subsequently ignites the
combustion between Ti and B.
A further example of production of boride-based fully-dense materials is recently
reported by Olevsky et al. 199). The product, a TiB,-TiN-Ni cermet, a material
with a unique combination of high temperature hardness and stability and with a
reasonable ductility/fracture toughness.
Other composite materials based on TIC and TiB2 have been produced recently
through the SHS process. Attempting to develop a process for the production of
ceramic armor, Lis et al. [55] considered the combustion synthesis of TiB2-Tic
composites. Such composites have a lower hardness than either TiB2 and T i c
single phases, but a fracture toughness and wear resistance which are significantly
higher than the corresponding values of the single phase materials. In order to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
362 2 SeEf-Propagating High-Temperature Synthesis of Hard Materials
40 ' '
50
I
60
I
70
I
80
I
90
Volume percent Ni,Ti
Figure 50. The effect of Ni3Ti content on final density of Ni,Ti-TiB, composite. All samples were
60% relative green density [95].
reduce the cost of the raw materials, these composites were obtained through the use
of Ti and B4Cpowders as reactants with the anticipated product containing a TiB2/
T i c molar ratio of 2. Under such conditions, the adiabatic temperature is high
enough to reach the lowest eutectic temperature (T, = 2520°C) making it possible
to attain high product density. The combustion process was found to be strongly
dependent on the reactants' particle size. A stable combustion process could be
obtained only for B4C and Ti powders with particle sizes below 10 and 49pm,
respectively. The synthesized composite was then densified by pressureless sintering
at temperatures between 2000 and 2100°C or by hot-pressing at temperatures
between 1800 and 2000°C using a ram pressure of 200 atm. The sintering time for
the combustion synthesized samples was considerably shorter than for mixtures
of commercial T i c and TiB2 powders, and the microhardness of the product
(Hv = 29.5 GPa) was higher than values reported in the literature.
The addition of alumina to TiB2 or TIC has become an attractive approach in
recent years. The addition of alumina to titanium carbide or boride improves the
oxidation resistance and results in composites which possess superior mechanical
strength and fracture toughness, without a significant drop in their overall micro-
hardness. Moreover, these composites can be obtained through SHS using, as
reactant, the relatively inexpensive titanium oxide instead of the metal. In a
recent example, a ceramic-ceramic composite to be used as a grinding material
has been prepared by Vershinnikov et al. [98] using the reaction 3TiO2 + 3Bi2O3+
+
lOAl = 3TiB2 5A1203.Kecskes et al. [50] showed that the dynamic consolidation
technique can be used to produce dense TiB2-A1203 composites through SHS
using the previous thermite reaction. Sample densities of 90% can be obtained by
varying the reaction ignition-consolidation delay time. The TiB2 grain size was
found to be related to that of the precursor Ti02. A detailed study on the influence
of several experimental parameters on product microstructure was reported by
Bowen and Derby [loo] for TiC-A1203 composite always obtained through a reaction
between TiOz and A1 plus C. The combustion kinetics and the product microstructure
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Selected Recent Examples of Synthesis of Hard Materials 363
O h
0 10 20 30 40 50
Aluminium particle size (pm)
Figure 51. Combustion wave velocity against A1 particle size for the combustion synthesis of TIC-
A1203 composites [100].
result is strongly dependent on the reactant particle size. The wave propagation rate,
for example, was strongly dependent on A1 particle size, but less significantly on Ti02
particle size. In both cases a decrease of propagation rate is observed when the particle
size is increased, as seen in Figs 5 1 and 52. In contrast, when the carbon particle size
was increased from 0.02 to 7.2 pm, the wave propagation increased but then decreased
for larger particle size, Fig. 53. This behavior was explained by an assumed barrier
role played by the extremely fine carbon, decreasing the titanium-aluminum particle
contact, a necessary first step in the reaction, Fig. 54. The effect of diluents on the
reaction products was also considered in the cited work [loo]. When A1203 was
added as a diluent it tended to induce cracking in the final product, probably because
the combustion temperature decreased below the melting point of alumina. On the
other hand, the addition of A1 as a diluent produced an unexpected increase in
wave velocity with increasing diluent content, an observation that can be explained
2 '
0 5 10 15 20
Titania particle size (Fm)
Figure 52. Combustion wave velocity against TiOz particle size for the combustion synthesis of
TiC-A1203 composites [ 1001.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
364 2 Self-propagating High-Temperature Synthesis of Hard Materials
0 10 20 30 40 50
Carbon particle size (pm)
Figure 53. Combustion wave velocity against C particle size for the combustion synthesis of TiCi-
A120, composites [loo].
Figure 54. Schematic diagram showing how small carbon black particles act as a barrier to TiO2-A1
particles contact [loo].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Selected Recent Examples of Synthesis of’ Hard Materials 365
I 2 3 4 5 6 7
vent
N,
Figure 55. Experimental set-up for SHS Cr7C3coating preparation: 1 =pressure gauge, 2 = reactor,
3 = heater, 4 = container, 5 =igniter, 6 =reactant charge, 7 = substrate [106].
(A) (B)
EI13 Non of any product is formed on the substrate
U Only some droplets are formed on the substrate
IT3 Only some islands are formed on the substrate
About 60% area is covered
111 Covered completely, but not smootly
Covered completely and smootly
Figure 56. Effect of preheating temperature and ambient pressure on the formation of SHS Cr7C3
coating on Cr: (A) single layer of reactants; (B) triple layer of reactants [106].
show that the composition of the layer is not homogeneous: the Cr content increases
and the C content decreases with distance from the top of the coating. Concomitantly,
the mechanical properties change with distance, with the maximum hardness being
measured on the top of the surface. The adhesion of the coating to the substrate
was found to be exceptionally good, even after several thermal cycles.
Considerable interest has been generated during the past few years in the synthesis
of diamond containing materials through SHS. In this regard the advantage of
SHS in avoiding the drawback of the traditional methods of production of
diamond-containing tools has been highlighted. These drawbacks relate to the
necessity to use long reaction times at very high pressures. In the SHS process,
reaction times can be reduced (from several hours in the traditional methods) to a
few seconds. Furthermore, the suitability of the SHS process for the production
of functionally-graded materials is an additional attractive feature. However,
there is one possible drawback in the use of the SHS process: the extremely high
reaction temperatures. Diamond transforms into graphite above about 1400-
1500°C under an inert gas atmosphere or in vacuum. As a result, it is of concern
that the mechanical characteristics of diamond might degrade during a typical
SHS thermal cycle. Theoretical analysis of the combustion process in diamond-
containing mixtures have been proposed by Padyukov et al. [107]. The extent of
diamond grain preservation was shown to depend more on the time the diamonds
stayed in the combustion zone than on the absolute value of the combustion tem-
perature. Surface graphitization, on the other hand, increases with the combustion
temperature. Three different approaches have been proposed in order to reduce the
effect of the combustion temperature on the diamond degradation [108]. In the first
approach two-layer samples are used, as shown in Fig. 57. One contains the pure
reacting mixture, for example Ti + B or Ni + Al, while the second contains the
same mixture diluted with diamond powder. The first layer, without diamond
acting as an inert diluent, has a more rapid combustion rate and acts to increase
the rate of combustion in the second layer. This results in a reduction of the dwell
time of the diamond in the high-temperature region. Moreover, such a procedure
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Selected Recent Examples of Synthesis of Hard Materials 367
Figure 57. Schematic of the bilayer pellet used for the SHS synthesis of diamond containing
materials [108].
z
a"
15
14 I I I I I
0 10 20 30 40 50
Mass % ,TIH,
included in the matrix in order to improve the mechanical properties of the product.
Intermetallics (Ni-A1 or Ni-Al-Cu-Ti), TiB2, and TiB2/Si represent examples of a
matrix of the first kind, while Tic-Si, Tic-Ti-A1, and Tic-TiN-Ni-Mo-A1 repre-
sent a matrix of the second kind. The densification of the products has been
obtained using the typical approaches already used for other SHS processes such
as hot-pressing, hot-isostatic pressing, pseudo-HIP, explosive consolidation, and
high-velocity forging. Ohyanagi et al. [47] obtained good results with a diamond
content up to 70 mol% (34 weight-%) using dynamic pseudo-isostatic compaction.
Similar results are reported by Padyukov and Levashov [109].
SHS reactions have also been used in the metal brazing of diamonds on cutting
tools [110]. In this case an SHS process is used to supply the heat necessary for
brazing a single particle of diamond on a copper-tungsten plate through a silver,
copper, and titanium alloy. A pellet of Ti + C was placed under the tungsten-
copper plate and then ignited. The speed of the thermal cycle ensured that no
damage to the diamond occurs. It has been suggested that a supercooled liquid
forms as a result of the rapid rate of cooling under such experimental conditions.
The presence of the liquid brazing alloy even at relatively low temperatures ensures
the process of stress relief.
Acknowledgment
This work was supported by a grant from the (US) Army Research Office, ARO.
References
1. Z. A. Munir and U. Anselmi-Tamburini, Self-propagating exothermic reactions: The synthesis
of high temperature materials by combustion, Mater. Sci. Rep. 1989, 3, 277-365.
2. Z. A. Munir, Synthesis of high temperature materials by self-propagating combustion
methods, Ceramic Bull. 1988, 67, 342-349.
3. V. M. Shkiro, V. N. Doronin, and I. P. Borovinskaya, Study of the concentration structure of
the titanium-carbon combustion wave, Combust. Explos. Shock Waves, 1981, 16,270-376.
4. Z. A. Munir and N. Sata, SSH diagrams: theoretical analysis and experimental observations,
Int. J. SHS, 1992, 1, 355-365.
5. Z. A. Munir and W. Lai, The SHS diagram of Tic, Combust. Sci. Technol. 1992,88,201-209.
6. A. Feng and Z. A. Munir, The effect of product conductivity on field-activated combustion
synthesis, J . Am. Ceram. SOC.1997,80, 1222-1230.
7. S. Zhang and Z. A. Munir, Spin combustion in the system nickel-silicon, J. Mater. Sci. 1992,
27,4789-5794.
8. M. E. Grami and Z. A. Munir, The effect of nitrogen pressure and diluent content on the
combustion synthesis of TiN, J. Am. Ceram. SOC.1990,73, 2222-2227.
9. S. D. Dunmead, Z. A. Munir, and J. B. Holt, Temperature profile analysis in combustion
synthesis: I, theory and background, J . Am. Ceram. SOC.1992, 75, 175-179.
10. S. D. Dunmead, Z. A. Munir, and J. B. Holt, Temperature profile analysis in combustion
synthesis: 11, experimental observations, J. Am. Ceram. SOC.1992, 75, 180-188.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 369
1 1. L. L. Wang and Z. A. Munir, Kinetic analysis of the combustion synthesis of molybdenum and
titanium silicides, Metall. Trans. 1995, 26B, 591-601.
12. S. D. Dunmead, Z. A. Munir, and J. B. Holt, Gas-solid reactions under a self-propagating
mode, Solid State Zonics, 1989, 32/33, 47&480.
13. S. Deevi and Z. A. Munir, The mechanism of synthesis of titanium nitride by self-sustaining
reactions, J. Mater. Res. 1990, 5, 2177-2183.
14. T. Boddington, P. G. Laye, J. Tipping, and D. Whalley, Kinetic analysis of temperature
profiles for pyrotechnic systems, Combust. Flame, 1986, 24, 359-368.
15. J. B. Holt and Z. A. Munir, Combustion synthesis of titanium carbide: theory and experiment,
J. Mater. Sci. 1986, 21, 251-259.
16. K. A. Philpot and Z . A. Munir, The influence of pre-ignition processes on the combustion
synthesis of nickel aluminides, in Materials Processing by Self-Propagating High- Temperature
Synthesis, K. A. Gabriel, S. G. Wax, and J. W. McCauley (Eds), 1987, p. 105.
17. L. L. Wang, Z. A. Munir, and J. B. Holt, Synthesis of Al20,-B4C composites via a thermite-
based combustion reaction, J. Mater. Synth. Proc. 1994, 2, 227-237.
18. L. L. Wang, Z. A. Munir, and J. B. Holt, Formation of MgO-B4C composites via a thermite-
based combustion reaction, J. Am. Ceram. SOC.1995, 78, 756-764.
19. A. Makino and C. K. Law, SHS combustion characteristics of several ceramics and inter-
metallic compounds, J. Am. Ceram. Soc. 1994, 77, 778-786.
20. A. P. Hardt and R. W. Holsinger, Propagation of gasless reactions in solids - 11. Experimental
study of exothermic intermetallic reaction rates, Combust. Flame, 1973, 21, 91-97.
21. A. I. Kirdyashkin, Y. M. Maksimov, and A. G. Merzhanov, Effects of capillary flow on
combustion in a gas-free system, Combust. Explos. Shock Waves, 1981, 17, 591-595.
22. A. A. Zenin, A. G. Merzhanov, and G. A. Nersisyan, Thermal wave structure in SHS
processes, Combust. Explos. Shock Waves, 1981, 17, 63-71.
23. I. P. Borovinskaya, A. G . Merzhanov, N. P. Novikov, and A. K. Filonenko, Gasless combus-
tion of mixtures of transition metals with boron, Combust. Explos. Shock Waves, 1974,10,2-10.
24. A. G. Merzhanov and I. P. Borovinskaya, A new class of combustion processes, Combust. Sci.
Technol. 1975, 10, 195-201.
25. N. P. Novikov, I. P. Borovinskaya, and A. G. Merzhanov, Dependence of the composition of
the products and the combustion rates in metal-boron systems on the ratio of the reagents,
Combust. Explos. Shock Waves, 1974, 10, 177-178.
26. A. Makino, N. Araki, and T. Kuwabara, Flammability limits, dilution limits, and effect of par-
ticle size on burning velocity in combustion synthesis of TIC, J S M E Znt. J. B: Fluids Therm.
Eng. 1994, 37, 576582.
27. V. M. Shkiro and I. P. Borovinskya, Study of characteristics of the combustion of mixtures of
titanium with carbon, in Combustion Processes in Chemical Technologyand Metallurgy, A. G.
Merzhanov (Ed.), Akad. Nauk SSSR, OIKhF, Chernogolovka, 1975, p. 253.
28. Y. B. Zeldovich, Theory for the limit of the quiet flame propagation, J. Exp. Theor. Phys.
1941, 11, 159-169.
29. A. Makino and C. K. Law, Heterogeneous flame propagation in the self-propagating
high-temperature synthesis (SHS) process: theory and experimental comparisons, Twenty-
fourth Symposium (International) on Combustion, The Combustion Institute, 1992,
pp. 1883-1 89 1.
30. D. C. Halverson, K. H. Ewald, and Z. A. Munir, Influence of reactant characteristics on the
microstructure of combustion-synthesized titanium carbide, J. Mater. Sci. 1993, 28, 4583-
4594.
31. A. D. Bratchikov, A. G. Merzhanov, V. I. Itin, V. N. Khachin, E. F. Dudarev, V. E. Gyunter,
V. M. Maslov, and D. B. Chernov, Self-propagating high-temperature synthesis of titanium
nickelide, Sov. Powder Metall. Met. Ceram. 1980, 19, 5-8.
32. Z. A. Munir and J. B. Holt, The combustion synthesis of refractory nitrides, Part I: theoretical
analysis, J. Mater. Sci. 1987, 22, 710-714.
33. M. E. Grami and Z. A. Munir, The effect of porosity on the combustion synthesis of titanium
nitride, J. A m . Ceram. Soc. 1990, 73, 1235-1239.
34. Z. A. Munir, The mechanism of synthesis of high temperature materials by combustion
processes, High Temp. Sci. 1990, 27, 279-293.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
310 2 Self-Propagating High-Temperature Synthesis of Hard Materials
35. S. Zhang and Z. A. Munir, The combustion synthesis of refractory nitrides: Part 11, the
synthesis of niobium nitride, J. Muter. Sci. 1991, 26, 3380-3385.
36. V. A. Gorshkov, G. N. Komratov, and V. I. Yukhvid, Production of cast higher chromium
carbide using self-propagating high-temperature synthesis, Sov. Powder Metall. Met. Ceram.
1992.31.952-955.
37. U. Anselmi-Tamburini, F. Maglia, G. Spinolo, and Z. A. Munir, NijYSZ cermets from
combustion synthesis: effect of process parameter on product microstructure, J . Am. Ceram.
SOC.1998,81, 1765-1772.
38. Z. A. Munir, Analysis of the origin of porosity in combustion synthesized materials, J. Muter.
Synth. Proc. 1993, 1, 387-394.
39. J. H. Matlock and J. P. Stark, Thermal diffusion ofvacancies in aluminum, Acta Metall. 1971,
14,923-929.
40. Z. A. Munir and L. L. Wang, The contribution of thermal migration to pore formation during
SHS reactions, in Proceedings of the First US-Japan Workshop on Combustion Synthesis,
Y. Kaieda and J. B. Holt (Eds), National Research Institute for Metals, Tokyo, Japan, 11-
12 January 1990, pp. 123-137.
41. S. D. Dunmead, Z. A. Munir, J. B. Holt, and D. D. Kingman, Simultaneous synthesis and
densification of TiC/Ni4-A1 composites, J. Mater. Sci. 1990, 26, 241c-2416.
42. Z. Y. Fu, W. M. Wang, R. Z. Yuan, and Z. A. Munir, Fabrication of cermets by SHS-QP
method, Int. J. SHS, 1993, 2, 307-313.
43. S. Adachi, T. Wada, T. Mihara, Y. Miyamoto, M. Koizumi, and 0. Yamada, Fabrication of
titanium carbide cermets by high-pressure self-combustion sintering of titanium powder and
carbon fiber, J. Am. Ceram. SOC.1989,72, 805-809.
44. P. D. Zavitsanos, J. J. Gebhardt, and A. Gatti, The use of self-propagating high-temperature
synthesis of high density titanium diboride, in Combustion and Plasma Synthesis of High-
Temperature Materials, Z. A. Munir and J. B. Holt (Eds), VCH Publishers, New York,
1990, pp. 170-178.
45. Y. Miyamoto, SHS/HIP compaction using inorganic fuels, Int. J. SHS, 1993, 1,479489.
46. Y. Miyamoto, Z. Li, Y. S. Kang, and K. Tanihata, Development of ceramic/metal hetero-
geneous composites by SHS/HIP compaction, Int. J. SHS, 1996, 5, 69-75.
47. M. Ohyanagi, T. Yoshikawa, M. Koizumi, S. Hosomi, E. A. Levashov, and I. P. Borovin-
skaya, Fabrication of diamond-dispersed cermets by SHSidynamic pseudo isostatic com-
paction (DPIC), Znt. J. SHS, 1995, 4, 387-394.
48. H. A. Grebe, A. Advani, N. N. Thadhani, and T. Kottke, Combustion synthesis and
subsequent explosive densification of titanium carbide ceramics, Metall. Trans. 1992, 23A,
2365-2372.
49. L. J. Kecskes, T. Kottke, and A. Niiler, Microstructural properties of combustion-synthesized
and dynamically consolidated titanium boride and titanium carbide, J. Am. Ceram. SOC.1990,
73, 1274-1282.
50. L. J. Kecskes, A. Niiler, T. Kottke, VK. V. Logan, and G. R. Villalobos, Dynamic consolida-
tion of combustion-synthesized alumina-titanium diboride composita ceramics, J. Am. Ceram.
SOC.1996,79, 2687-2695.
51. B. H. Rabin, G. E. Kroth, and R. L. Williams, Fabrication of TiC-AI2O3 composites by
combustion synthesis and subsequent dynamic consolidification, Int. J. SHS, 1992, 1, 336341.
52. J. C. La Salvia, L. W. Meyer, and M. A. Meyers, Densification of reaction-synthesized
titanium carbide by high-velocity forging, J. Am. Ceram. SOC.1992, 75, 592-602.
53. D. A. Hoke, M. A. Meyers, L. W. Meyer, and G. T. Gray, Reaction synthesis/dynamic
compaction of titanium diboride, Metall. Trans. 1992, 23A, 77-78.
54. J. C. La Salvia, M. A. Meyer, and D. K. Kim, Combustion synthesis/dynamic consolidation of
Tic-Ni cermets, J. Muter. Synth. Proc. 1995, 2, 255-274.
55. J. Lis, S. Majorowski, V. Hlavacek, and J. A. Puszynski, Combustion synthesis and densifica-
tion of TiB2-TiC composite powders, Int. J. SHS, 1995, 4, 275-285.
56. H. Zhang, S. Yin, J. Zhong, and S. Guo, TiC-TiB2-AI2O3 ceramics made by combustion
synthesis and pressureless sintering, Int. J. SHS, 1996, 5, 185-190.
57. G. Xu, H. Zhuang, J. Cai, and S. Xu, Microwave sintering of SHS-produced AIN powders,
J. Muter. Synth. Proc. 1996, 4, 307-314.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 371
82. M. Miura, T. Yoyo, and S. Hirano, Phase separation and toughening of SIC-AIN solid-
solution ceramics, J . Muter. Sci. 1993, 28, 4188-
83. H. Xue and Z. A. Munir, The synthesis of composites and solid solutions of SIC-AIN by field-
activated combustion, Scripta Muter. 1996, 35, 979-982.
84. C. He and G. C. Stangle, The mechanism and kinetics of the niobium-carbon reaction under
self-propagating synthesis-like conditions, J . Muter. Res. 1995, 10, 2829-2841.
85. H. Xue and Z. A. Munir, Field-activated combustion synthesis of TaC, Int. J . SHS, 1996, 5,
229-239.
86. J. H. Lee, N. N. Thadhani, and H. A. Grebe, Reaction sintering of shock-compressed Ti + C
powder mixtures, Metall. Muter. Trans. 1996, 21A, 1749-1 759.
87. L. Wang, M. R. Wixom, and L. T. Thompson, Structural and mechanical properties of TiB,
and TIC prepared by self-propagating high-temperature synthesis/dynamic compaction,
J. Muter. Sci. 1994, 29, 524-543.
88. M. J. Capaldi and J. V. Wood, Production and characterization of Tic-containing
materials by self-propagating high-temperature synthesis, J. Muter. Synth. Proc. 1996, 4,
245-253.
89. J. C. La Salvia and M. A. Meyers, Microstructure, properties, and mechanisms of Tic-Mo-Ni
cermets produced by SHS, Int. J . SHS, 1995, 4, 43-57.
90. Z. P. Xing, J. T. Guo, G. Y. An, and Z. Q. Hu, Hot pressing aided exothermic synthesis and
densification of NiAl and NiAI-TIC composites, Int. J . SHS, 1996, 5, 51-56.
91. C. C. Degnan and J. V. Wood, Processing and wear properties of a (W,TiC) particulate
reinforced ferrous-based metal matrix composite, A. Poursartip (Ed.), Proc. 10th Int. Conf.
Compos. Muter. Vancouver, Canada, 1995, pp. 77 1-778.
92. S. Y. Sharivker, I. P. Golovinskaya, I. N. Vorob’eva, and V. M. Bunin, Wear of tungsten-
less hard alloys based on SHS-titanium carbide, Powder Metall. Metal Ceram. 1996, 35,
55-57.
93. M. Ohyanagi, M. Kanno, and M. Koizumi, T i c combustion synthesis and fabrication of a
body with a densified surface layer by the pressureless method, J . Muter. Synth. Proc. 1993,
1, 31 1-322.
94. T. C. Woodger, J. J. Moore, H. C. Yi, and J. Y. Guigne, Combustion synthesis of A1-HfB,
metal matrix composites, T. S. Srivatsan and J. J. Moore (Eds), Proceedings of Processing
and Fabrication of Advanced Materials V ,TMS, 1996, pp. 223-233.
95. T. C. Woodger, J. J. Moore, H. C. Yi, and J. Y. Guigne, Combustion synthesis of Ni3Ti-TiBz
intermetallic matrix composites, T. S. Srivatsan and J. J. Moore (Eds), Proceedings of Pro-
cessing and Fabrication of Advanced Materials V ,TMS, 1996, pp. 399408.
96. W. Y. Yang and A. Petric, Microstructure of the Ti3A1(Nb)/TiB composite produced by
combustion synthesis, Metall. Muter. Trans. 1995, 26A, 3037-3043.
97. H. C. Yi and A. Petric, Combustion synthesis of Ti-Al-Nb matrix composites reinforced by
titanium borides, J. Muter. Synth. Proc. 1994, 2, 161-167.
98. Y. I. Vershinnikov, N. S. Pesotskaya, S. S. Mamyan, and I. P. Borovinskaya, SHS of com-
posite ceramic powder TiB,-AI,03 and its abrasive characteristics, Muter. Powder Technol.
1993, PTM93, 749-753.
99. F. Olevsky, P. Mogilevsky, E. Y. Gutmanas, and I. Gotman, Synthesis of in situ TiB2/TiN
ceramic matrix composites from dense BN-Ti and BN-Ti-Ni powder blends, Metall.
Muter. Trans. 1996,27A, 2071-2079.
100. C. R. Bowen and B. Derby, The formation of TiC/Alz03 microstructures by a self-
propagating high-temperature synthesis reaction, J. Muter. Sci. 1996, 31, 3791-3803.
101. H. C. Yi, H. J. Feng, J. J. Moore, A. Petric, and J. Y. Guigne, Combustion synthesis and
mechanical properties of TiC/Al2O3-AI composites, Int. J . SHS, 1996, 5, 39-50.
102. C. W. Kim, C. S. Kim, S. C. Koh, and D. K. Kim, Combustion synthesis and densification of
the TiC-AIzO3-AI composites, Int. J . SHS, 1996, 5, 159-168.
103. Z. Li, K. Tanihata, and Y. Miyamoto, Fabrication of multifunctional gradient materials in the
system of A1~03/TiC/Ni/TiC/Alz03 by SHS/HIP, Advanced Materials ’93, N. Mizutani (Ed.),
Elsevier Science, Amsterdam, 1994, pp. 677-680.
104. 0. Odawara, Long ceramic-lined pipes produced by a centrifugal thermal process, J. Am.
Ceram. SOC.1990,73, 629-633.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 373
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
3.1 Introduction
Cubic diamond and hexagonal graphite are the thermodynamically stable con-
densed phases in the P-Tdiagram of carbon with diamond and graphite dominating
at very high and low pressures respectively. Direct transformation between the
forms is kinetically impeded due to the necessity of a complete reconstruction of
strongly bonded covalent networks.
Natural diamonds are supposed to have grown in the depth of the mantle 121
under high-pressure high-temperature (HPHT) conditions above the Berman-
Simon line [ 11, shown in Fig. 1. Although possible mechanisms of natural diamond
formation are still debated, C-H-0 fluids seem to play an important role [3].
Reports on man-made diamond obtained by HPHT synthesis were first published
in 1955 by General Electric [4]. Usually, metals able to dissolve carbon under HPHT
conditions are used as catalysts and increase growth rates. Diamond crystals of
several millimeters in size can be obtained in this way, but usually small grains
for abrasives are produced. Direct conversion of graphite to diamond without
catalyst in HPHT apparatus is possible, but uneconomical for industrial applica-
tion. Direct transformation can be done by the detonation method and produces
nanosized powders of diamond and diamond-like carbon [5].
Diamond films and coatings can be obtained via various chemical vapor deposi-
tion (CVD) methods at and below atmospheric pressure [6]. This highly metastable
and kinetically driven diamond growth process requires the presence of atomic
hydrogen, which selectively etches graphitic and amorphous carbon while diamond
is unaffected.
While HPHT, shock wave, and CVD synthesis of diamond have been commer-
cialized, these methods have serious limitations. None of them allows economical
growth of large single crystals and the diamond quality does not satisfy the require-
ments for electronic or gem applications. Since the quality and size of the best
natural diamonds have not yet been reproduced in the lab, the search for a better
method of diamond synthesis continues. Recent work demonstrates that diamond
can be formed metastably under a variety of conditions.
Related to the CVD process is the low-pressure solid-state source (LPSSS) pro-
cess, where preformed mixtures of carbon and diamond seeds [7] or carbon and
metal [8] are treated in a H2 plasma, thus forming compact diamond ceramics.
Diamond particles have been found in combustion products [9], synthesized from
fullerenes by pressurization at room temperature [ 101 or sputtering at high tempera-
tures [ 111, obtained by decomposition of polymers with an appropriate structure
[ 121, by heating carbon ‘onions’ with electron beam in a TEM [ 131, by chlorination
of S i c [14], or by deposition from supercritical fluids [15-171.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.1 Introduction 375
0 6 0
Temperature [“C]
Figure 1. P-Tdiagram of the low-pressure, low-temperature labile equilibriums of carbon solution:
1 = graphite-diamond equilibrium line, 2 = glassy carbon-diamond transition line, 3 =range of
pneumatolytic hydrothermal processes, 4 = oxidative corrosion of diamond, 5 = anticipated area
of diamond hydrosynthesis, 6 and 7 =diamond synthesis from glassy carbon precursors, 8 = low-
pressure, low-temperature hydrothermal homoepitaxy of diamond. Reproduced from [ 151 with
permission from A. Szymanski.
The presence of small diamond particles has been confirmed independently in all
cases using several methods of analysis. These recent findings became primarily
possible due to the development of new analytical techniques that allow the
detection of diamond in small quantities and with particle sizes down to a few nan-
ometers. The formation of diamond particles in the range of the thermodynamic
stability of graphite can be explained either by kinetic factors, or by a higher stabi-
lity of small diamond particles compared to graphite, which has been controversially
discussed [ 181.
While some of the metastable routes to diamond mentioned above will probably
remain a scientific curiosity, hydrothermal synthesis has attracted more attention
due to its commercial potential. Synthesis from a supercritical water-based fluid is
an excellent method for preparing powders, coatings, and single crystals of many
different materials [ 191, which offers many advantages over other preparation meth-
ods because it is inexpensive, environmentally benign, reduces free energies for var-
ious equilibria, enables the synthesis of phases that may not be stable under other
conditions, and permits production of composites from incompatible phases that
can not be obtained by other means. It is the only commercial method of growing
large quartz single crystals and synthetic emeralds. Thus, if applied to diamond, it
may potentially yield diamond crystals, which might be competitive with the Cullinan
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
376 3 Hydrothermal Synthesis of Diamond
diamond. Rustum Roy calls the hydrothermal synthesis ‘The fourth revolution in
diamond synthesis’ [20]. Although we think that it is too early to consider it to be a
‘revolution’, it is certainly worthy of attention and further evaluation as a potential
method of diamond synthesis. The goal of the ongoing research is to make diamonds
from a liquid phase which is several orders of magnitude denser than vapor, in the
hope of increasing growth kinetics compared to CVD synthesis.
In this chapter, we explain why hydrothermal synthesis deserves special consid-
eration and present a review of the current state of the art in hydrothermal synthesis
of diamond as a potential future method of making diamond or hard diamond-like
sp3-carbon.
that diamond grows from a nucleus in the postmagmatic fluid, the growth rate
should be controlled by the diffusion process of sp3-structural elements to the
seed crystal surface resulting in thin colloidal cluster layers which may in turn
form lonsdaleite or fibrous diamond. Due to coalescence, slow transformation to
thick polycrystalline layers or monocrystals with cubic structure might occur.
Coexistence, intergrowth and thus presumable paragenesis of crystals of dia-
mond, quartz and a-and 0-Sic have been found in a diamondiferous kimberlite
[34] where crystallization of quartz instead of coesite indicates maximum pressures
of <2.0-2.8 GPa at temperatures of 1000-1200°C. This might be important in the
view of diamond hydrosynthesis from silicon carbides in the metastable range.
Figure 2. Calculated temperature dependence of the major species in the carbon-water system for a
C : H 2 0 molar ratio of 1 : 1 at 100 MPa.
Figure 3. SEM micrographs of: (a) aggregate after 12 h hydrothermal treatment of nanocrystalline
diamond at 740°C and 300 MPa, (b) plate-like crystals on the surface of a diamond seed crystal after
150 h hydrothermal treatment of graphite and diamond single crystals at 800°C and 300 MPa, (c)
liquid-phase hydrothermal homoepitaxy of diamond on diamond seed at 170MPa and 400°C in
a specially prepared water solution. Reproduced from [I51 with permission from A. Szymanski.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
382 3 Hydrothermal Synthesis of Diamond
p-Sic -
I
01 I
0 1 2 3 4 5
amount of water I mol
Figure 4. Calculated equilibrium activity of Si02, S i c and C for compositions of 1 mol S i c plus
variable amounts of H 2 0 at 500°C and 200 MPa.
This computation is typical for the whole supercritical T-P range. It shows that Si02
is stable in the whole range. Sic becomes unstable at a H 2 0 :Sic molar ratio of 2 : 1,
which is a critical point, whereas carbon remains stable until a slightly larger
H 2 0 :Sic molar ratio than 2: 1 is reached. This indicates the possibility of direct
carbon formation in the course of hydrothermal leaching of S i c by selective
oxidation of Si according to reaction (8). Reaction (8) is restricted to low
H20:SiC molar ratios whereas reaction (6) and (7) are impeded at H20:SiC
molar ratio 5 2 : 1. The equilibrium amount of carbon increases with increasing
temperature and decreasing pressure.
I I
1100 1200 1300 1400 1500 1600
Wavenumber / cm-I
Figure 5. Raman spectra of carbon on a-Sic crystals, X = 632.8 nm, after 4 h hydrothermal treat-
ment at 600°C and 200 MPa: (a) homogeneous carbon film, (b) diamond particle at seed crystal
edge, (c) example of an inhomogeneous carbon film after 4 h hydrothermal treatment at 700°C
and 200 MPa.
(Fig. 5c) on the Sic-crystals as well as on the gold capsules which indicates also
carbon precipitation from the fluid.
Carbon formation via hydrothermal treatment of P-Sic was intensively studied at
400-800°C and 100-500 MPa [54]. Carbon formed at all temperatures, but
reasonable amounts in reasonable time were found only at and above 550°C.
Above 600°C carbon deposits on the capsule walls indicated carbon precipitation
from the fluid. In effort to reveal the presence of diamond after hydrothermal
treatment, the residues were leached in 20% H F solution to remove Si02 and in
hot HC104 to remove graphitic carbon because large amounts of silica and non-
diamond carbon obscured the picture.
XRD investigations showed graphite reflections and in some cases a very weak
reflection at 0.206nm, attributed to diamond. In the Raman spectra, shifts of the
graphite D band, and additional bands and broad features in the range 1100-
1340 cm-', confirmed the presence of amorphous and nanocrystalline carbon
and indicate the possible presence of disordered nanocrystalline sp3-carbon.
Small particles with a size below 0.3pm were found, and sometimes intergrown
with several larger P-SIC particles after wet oxidation, and exhibited predomi-
nantly hexagonal (0001) and face-centered cubic (1 11) and (100) faces that were
attributed rather to diamond than to graphite. The morphology of the particles
was very similar to that of gas phase diamond powders. Reflections corresponding
to lattice spacing of 0.206 and 0.192 nm in small angle diffraction (SAD) patterns
suggested the presence of cubic diamond as well as of lonsdaleite or other diamond
polytypes.
Hydrothermal experiments with P-Sic powders and watery solutions of 10%
Na2C03, 10% NaHC03 and 50% NaHCOO yielded results which were very
close to each other, because sodium hydrogen carbonate and sodium formiate can
be converted to disodium carbonate at high temperatures according to
2NaHC03 = Na2C03 H 2 0 + (9)
and
2NaHCOO = Na2C03+ C H 2 0 = Na2C03 C + + H20. (10)
Carbon formation was observed at and above 400°C and became pronounced at
600°C. Raman spectra strongly suggest the formation of larger quantities of sp3-
bonded carbon than in the experiments with distilled water.
Small amounts of well crystallized diamond were found after hydrothermal
treatment of p-Sic powder at 700-750°C in the presence of diamond seed [55].
After removing silica and nondiamond carbon, small (<3 pm) carbon particles
of predominantly octahedral shape, thus being probably diamond, were found.
They were attached to the surfaces of the single crystal seed but could be removed
by intense ultrasonic treatment. Tetrahedral hillocks of <1 pm size as well as
tetrahedral etch pits appeared on the seed diamond surfaces. The hillocks always
showed a faceted structure and sometimes common orientation, as expected for
homoepitaxial diamond growth.
The microcrystalline diamond seed did not show any increase in size after hydro-
thermal treatment, but particles of <50 nm size, which typically show a regular
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
386 3 Hydrothermal Synthesis of Diamond
3.4 Outlook
Taking into account the fact that hydrothermal formation of carbon takes place at
relatively low temperatures (200-800°C) and pressures (<100 MPa), this method
seems to be very attractive. It has obvious advantages over the HPHT synthesis
due to less extreme temperature-pressure conditions and the possibility of much
larger pressure vessels, which would make the hydrothermal method attractive for
industrial use. However, the temperature of synthesis should be decreased prefer-
ably to 200-250°C to make commercialization and up-scaling easier.
A better understanding of the mechanisms of diamond growth under hydrother-
mal conditions should make it possible to optimize the conditions for the growth of
diamond coatings or large single crystals. High growth rates compared to CVD
synthesis may be achievable.
Current understanding of the diamond synthesis under hydrothermal conditions
and quantity of diamond produced are comparable with those of CVD synthesis of
diamond at the end of the 1950s or the beginning of the 1960s. Thus, we are at the
beginning of the road and further research is strongly recommended.
Acknowledgments
We thank Professor Dr. Andrzej M. Szymanski, Faculty of Chemistry of the
Warsaw University of Technology for supply of materials for publication, and
the German Research Society for financial support (grant DFG Ni 299/4).
References
I . R. Berman and S. F. Simon, Z . Elektrochem. 1955, 59, 333-338.
2. R. C. DeVries, R. Roy, S. Somiya, and S. Yamada, Trans. Mater. Res. Soc. Jpn. 1994, 19B,
64 1-663.
3. R. C. DeVries, in Advanced Ceramics IIZ, S. Somiya (Ed.), Elsevier, Amsterdam, 1990, pp. 181-
205.
4. F. P. Bundy, H. T. Hall, H. M. Strong, and R. H. Wentorf, Nature, 1955, 176, 51-55.
5. K. Yamada and A. B. Sawaoka, J . Am. Ceram. Soc. 1994, 77, 1104-1106.
6. P. K. Bachmann, D. Leers, and H. Lydtin, Diamond Rel. Muter. 1991, 1, 1-12.
7. R. Roy, H. S. Dewan and P. Ravindranathan, Mater. Res. Bull. 1993, 28, 861-866.
8. R. Roy, H. S. Dewan, and P. Ravindranathan, J. Mater. Chem. 1993,3,685-686.
9. B. V. Derjaguin and D. V. Fedoseev, in New Diamond Science and Technology,M R S Znt. Conf.
Proc. 1991, pp. 3-10.
10. M. N. Regueiro, P. Monceau, and J.-L. Hodeau, Nature, 1992,355, 237-239.
11. A. R. Krauss, D. M. Gruen, D. Zhou, T. G. McCauley, L. C. Qin, T. Corrigan, 0. Auciello,
S. Pimenov, V. Konov, and A. Karabutov, in M R S Fall Meeting Abstracts, Paper W10.5,
MRS, 1997, p. 499.
12. G. T. Visscher, D. C. Nesting, J. V. Badding, and P. A. Bianconi, Science, 1993,260, 1496-1499.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
388 3 Hydrothermal Synthesis of Diamond
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
4.1 Introduction
The development of the scientific knowledge of diamond, of its properties and com-
position, and the various attempts to synthesize this much sought-after material
form an exciting chapter in the general history of science. Several review articles
covering historical aspects of diamond in general or of its synthesis under conditions
of metastability in particular are available. A recent short history of diamond
synthesis, in which older sources back to the year 1711 are quoted, was written by
Angus [l]. DeVries et al. [2], in an article containing numerous references to earlier
publications, laid their emphasis on the contributions of Russian researchers.
The modern era of diamond growth can be traced back to the days between 26
November 1952 and 7 January 1953, when Eversole at the laboratories of Union
Carbide Corporation in the USA achieved the first overgrowth of diamond from
low-pressure gas phases on the surface of diamond powder particles. Thus the
growth of diamond under pressure-temperature conditions at which graphite is
the thermodynamically stable phase of carbon, obviously predates the first success-
ful high-pressure, high-temperature syntheses of diamond at ASEA in Sweden in
February 1953 and at the General Electric Company in the USA in December
1954 [l]. Eversole was the first to synthesize diamond by any method, the procedure
used by him, however, was far from economic viability, owing to a very low yield
and the necessity of repeated removal of simultaneously deposited graphite by a
treatment in an atmosphere of pure hydrogen. Additionally, there was much
doubt in the scientific community that diamond could be grown at all under condi-
tions of thermodynamic metastability. Therefore the work of Eversole did not
attract enough attention to be a starting point of much more intensified efforts.
The results of Eversole were confirmed about 15 years later independently by the
group of Angus at Case Western Reserve University in the USA and by researchers
of the Physical Chemistry Institute of Moscow.
The important role of hydrogen in the low-pressure synthesis of diamond was
recognized in the work of Angus. Molecular hydrogen suppresses the spontaneous
nucleation and growth of graphite during the formation of diamond, but it does not
completely prevent the formation of the stable carbon allotrope, which has to be
removed from the deposit in a separate etching step. Therefore it was a very
important breakthrough in the efforts to grow diamond by a chemical vapor
deposition process, when in the seventies the Soviet group found out (stimulated
by results of Case Western Reserve University group on the action of atomic
hydrogen as an efficient remover of nondiamond carbon phases), that in the
presence of a superequilibrium concentration of atomic hydrogen diamond could
be deposited continuously, with much higher growth rates, largely free from
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Preparation Methodsfor Diamond Films 391
graphite and not only epitaxially on diamond crystallites but also on nondiamond
substrates.
This Russian achievement can most probably be dated back to the early 1970s.
Nevertheless, it took another decade, until (thanks to a series of papers from the
Japanese NIRIM (National Institute for Research in Inorganic Materials), reveal-
ing detailed experimental information) the possibility and the potential of diamond
growth from the gas phase was recognized by the general scientific community. The
Japanese work, connected with the names of Kamo, Matsumoto, Sato, and Setaka,
had started in 1974, based on results in the Russian literature. Among various
diamond CVD processes, developed over about 15 years, the hot-filament CVD
process (HF-CVD) and the microwave-plasma CVD method (MWP-CVD), for
the first time described in detail by the Japanese group, have continued to be the
most used diamond deposition processes.
1000 -
DC arcjet
100 *
10.
10
V oxy-acetylene
flames
r I pmlh
1 low-power density MW plasmas
hot filaments
0.1
non-isothermal DC, RF or ECR plasmas
thermal decomposition
0.01
1000 2000 3000 4000 5000 6000 71 DO
gas phase temperature I K
Figure 1. Correlation of typical gas temperatures and growth rates r (in pm h-') of diamond CVD
processes (after [3]). See text for the explanation of abbreviations.
processes, running typically at pressures beyond 10 mbar and allowing growth rates
of several pm h-' .
In the following, the HF-CVD method will be presented in some detail because of
its simplicity and scientific as well as technical importance (thanks to its up-scaling
potential). Several results of studies on HF-CVD processes are of fundamental
importance for an understanding of diamond deposition in general. The presenta-
tion of plasma-based methods will focus on microwave-activated processes because
they are probably the most often used diamond CVD growth methods for virtually
impurity-free material.
To pump
Figure 2. Schematic view of the HF-CVD deposition chamber as used by Matsumoto et al.
(adapted from [4,5]).
citly mentioned in these papers; it is however probable that this or a similar kind of
pretreatment was used in order to achieve a dense nucleation and early coalescence
of the crystallites to closed films. Figure 3 shows the surface of a closed diamond film
with well-shaped facets grown by HF-CVD in the author’s laboratory. The typical
appearance of isolated, non-twinned particles, is shown schematically in Fig. 5.
There can be no reasonable doubt that the diamond phase of carbon was actually
(RBS) a tungsten concentration around 0.1 at-% was measured by Gheerhaert et al.
in films grown at filament temperatures of about 2250°C [Ill. For tantalum
filaments at 2300°C maximum Ta concentrations of 0.03 at-% were detected
using secondary mass spectrometry (SIMS) [121. For carbide-forming filament
materials the degree of filament material incorporation strongly depends on the
carburization state of the filaments. This is caused by the fact that at the relevant
temperatures the carbides have lower vapor pressures than the corresponding
uncarburized materials [ 121. Therefore, in several studies relatively high amounts
of filament impurities were found at the diamond-substrate interface when the
deposition process was started with uncarburized filaments.
The mostly used gas atmospheres for hot-filament deposition of diamond gener-
ally consist of mixtures of 0.1 to a few vol-% methane and hydrogen. In contrast to
plasma-activated deposition methods, additions of oxygen-containing source gases
in HF-CVD of diamond are possible only to a limited extent, due to the tendency of
oxidation and volatilization of the filament material. Nevertheless, additions of
oxygen, water vapor, or oxygen-bearing organic compounds to the HF-CVD gas
feed can have several advantages if the quantitative gas phase composition is
carefully chosen. This was first demonstrated by Hirose and Terasawa, who used
a range of oxygen- and nitrogen-containing organic compounds [13]. Diamond
films with good crystallinity and high phase purity (according to Raman and
electron diffraction characterization) were obtained at growth rates about an
order of magnitude higher than typical for oxygen-free gas feeds. Compared with
methane-hydrogen mixtures, much higher gas pressures, up to nearly one atmos-
phere, and gas phase concentrations of the carbon-bearing species were possible
in these experiments. These results were later essentially confirmed by other authors,
achieving growth rates up to 10 pm h-' by using ethanol and acetone (1-8 vol-%) in
hydrogen at pressures up to 13 kPa [14].
Kawato and Kondo carried out a comparison of CH4-H2 and CH4-H2-02
mixtures and showed that an increased growth rate and extended pressure and con-
centration ranges (40 kPa, 4 vol-%) are also obtainable with additions of molecular
oxygen [15]. Numerous investigations have been performed in order to reveal the
reasons for the beneficial effects of oxygen-containing gases observed in diamond
HF-CVD (and MWP-CVD) [15-201. Results of these studies can only be sum-
marized here without further comments. The effects of oxygen have generally
been attributed to
- promotion of hydrogen dissociation and methane decomposition,
- a reduction of the acetylene concentration, leading to less nondiamond carbon
phases,
- formation of active growth sites by improved hydrogen abstraction, and
- an efficient removal of non-diamond carbon phases by hydroxy radicals.
As shown by Bachmann, there are more general limits to the C,H,O-concentrations
which can be used in diamond HF-CVD and similarly in most other diamond CVD
processes. A compilation of diamond-forming gas phase compositions from a large
number of publications shows that (within a ternary C,H,O-concentration diagram)
those concentrations leading to diamond growth are restricted to an essentially
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Preparution Methods for Diamond Films 397
wedge-shaped area extending from the H-corner along the CO line of the triangle
[21]. Smaller carbon concentrations do not lead to the formation of deposits at
all, due to the dominance of etching processes over carbon deposition. At con-
centrations higher than those situated in the diamond wedge sp2-coordinated
carbon phases tend to dominate.
As a key to the growth of fiber-textured diamond films, additions of molecular
nitrogen or nitrogen-containing compounds have in the recent years gained
increased interest, see below. Homoepitaxial growth experiments of Cao et al. [22]
showed, that nitrogen strongly enhances the growth rate of { 100) facets. This effect
can be attributed to nitrogen compounds breaking the dimers of the (2x1)
reconstructed {loo} surface, thus favoring the (100) growth direction.
electrical field strength can be generated within a section of the tube, just beyond a
substrate table. Owing to the geometric limitations of suitable waveguides at
2.45GHz, quartz tubes are limited in diameter to around 5cm. Therefore the
coatable area on a substrate is usually between 1 and 10cm2. The experimental
conditions leading to successful growth of diamond are very similar to those in
HF-CVD: Pressures are typically between 10 and loombar, the feed gas in the
first NIRIM publication consisted of 1 vol-YOmethane in hydrogen and the sub-
strate was heated to 800-1000°C in a plasma generated at several 100W to 1 kW.
Under these conditions, growth rates between 0.5 and 3 pm h-’ can be achieved,
the lower value being more realistic for well faceted, high phase purity films.
In early MWP-CVD experiments oxygen-free gas feeds, generally 0.1-1 vol-%
methane in hydrogen, were commonly in use. However, since it was shown by
Hirose and Terasawa and others, that higher growth rates and better film qualities
at larger pressures and carbon concentrations could be achieved using oxygen com-
pounds in HF-CVD [13], it became popular to use oxygen-containing compounds
like 02,H 2 0 , CO, C02, alcohols, ethers, and others in microwave plasma CVD
of diamond. Independent of the gas composition actually used, hydrogen, methane,
and acetylene are the prevailing components of the in situ and exhaust gas of typical
MWP-CVD processes in C,H gas phases. Oxygen-containing components are
predominantly converted to carbon monoxide. Oxygen additions suppress the
acetylene formation and lead eventually to a practically complete disappearance
of hydrocarbons in favor of CO. The spectrum of stable products is fairly similar
to that of HF-CVD processes, pointing to a basic similarity of both types of
processes [24]. In fact the main role of the plasma in MWP-CVD (analogous to
the filament in HF-CVD) seems to be the dissociation of molecular hydrogen into
atoms, which initiate a reaction chain leading to the methyl radical as the most
important diamond growth species, and help to gasify unwanted sp2-bonded
carbon deposits (see below).
While the tubular reactor design was used in most laboratories working on MWP-
CVD of diamond in the mid-80s, commercial reactors appeared on the market after
a few years. Major drawbacks of the NIRIM-type reactors are the limited substrate
size and difficulties in preventing the plasma from interactions with the tube
material, leading to etching and redeposition of silicon contaminations in the
films, localized heating, deformation, or even puncturing of the tube. Also, MW
power independent adjustment of the substrate temperature by external heating is
difficult. The development of MWP-CVD sources at ASTeX, Applied Science and
Technologies, at Woburn, MA, USA was outlined by P. K. Bachmann [25], who
also contributed to the design of the so called ‘bell jar’ reactor which was the first
microwave plasma CVD apparatus permitting growth of diamond on substrates
up to 7.5cm in diameter and became the first commercially available diamond
CVD reactor at all. This reactor contains a dome-shaped silica bell jar of lOcm
diameter in which a plasma is generated at a location of maximum electrical field
strength well apart from the wall material.
Today the ASTeX high-pressure microwave source reactor (HPMS) is in use in
many laboratories. A schematic cross-section of this apparatus is shown in Fig. 4
[25]. Microwave power is transmitted by a coupling device from a rectangular wave-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Preparution Metkods,for Diamond Films 399
MW
=i inlet
gas ,- substrate
t o pump
+----
Figure 4. Schematic cross-section through a modern 2.45 GHz ASTeX HPMS (‘High-pressure
Microwave Source’) MWP-CVD reactor. Microwave radiation is coupled through a silica
window separating the atmospheric pressure region within the waveguide and around the antenna
from the plasma region under reduced pressure. The reactor is equipped with a R F induction-
heatable graphite stage and can be used for substrates up to lOcm diameter. Modified by a
water-cooled graphite stage, reactors of this type can be used for up to 5 kW MW power. Inspection
windows are not shown here (adapted from [25]).
PBCs or the lack of parallel PBCs. Growth on these faces involves the addition of
growth units to pre-existing steps or kinks, without an activation barrier. These
interrelations between surface energy and growth rates (in spite of often being
able to describe general features of crystal morphology) do not necessarily hold
always and exactly: Growth rates of different faces of crystals are often observed
to depend differently on growth conditions like temperature, supersaturation
and kind of growth species, impurities acting as inhibitors or catalysts and so on.
Diamond CVD is a nice example illustrating the influence of growth conditions
on crystal shape and film texture.
A PBC analysis of the diamond lattice leads one to expect ( 11 1) faces (containing
at least two different types of periodic bond chains) as the only well-defined flat
faces of diamond crystallites. (001) faces should be kinked (K-faces) and appear
(if at all) as curved surfaces owing to the lack of pronounced cusp-like growth
rate or surface energy minima in (001) directions. This expectation is, however, in
conflict with experimental evidence, showing that (001) faces of diamond and
other crystals with a diamond lattice are generally slow-growing well-defined flat
faces of similar or even larger morphological importance than ( 111) faces. Giling
and Van Enckevort explained this prominence of (001) faces in terms of a surface
reconstruction whereby surface atoms are able to lower their energy by dimerization
with neighboring atoms, thus reducing the number of dangling bonds at the surface.
The surface reconstruction process leads to F-face-like behavior of (001) faces, that
is, growth by two-dimensional nucleation of new layers [311.
A cubic crystal bounded by (001) and (111) faces generally forms a cubo-
octahedron which degenerates into a cube or an octahedron as the limiting shape
for vanishing portions of ( 111) and (001) faces. The relative share of these two
types of faces depends on the ratio of growth rates Vool/Vlll. In Fig. 5 crystal
shapes are shown for a number of ‘growth parameters’ a, defined by
31/2VooI/ V11 [32]. The arrows indicate the largest diameter, which is the direction
of fastest growth of an individual crystal. In order to realize a cubic diamond crys-
tallite, the growth rate ratio has to become Vool/V111I 3-‘/2 = 0.58; for a pure
octahedron a ratio 2 31/2= 1.73 is required.
As shown by Clawing et al. [33], all forms of cubo-octahedrons, including the
limiting forms of cubes and octahedrons, can be produced as single isolated crystal-
lites by a suitable selection of growth conditions, that is, growth rate ratios
0.58 V0ol/Vl11 2 1.73 are accessible. Thick films grown under the extreme
condition of a growth rate ratio near 0.58 (a = 1) exhibited a (1 11) texture with
1 1.5 2 2.5 3
Figure 5. Morphologies of cubo-octahedral crystals, exhibiting (001) and { 11I } facets, grown at
different growth parameters, a. Arrows indicate the directions of fastest growth (Reprinted from
Diamond and Related Materials, 3, C. Wild et al., Oriented CVD diamond films: twin formation,
structure and morphology, pp. 373-38 1, Copyright 1994, with permission from Elsevier Science).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Preparation Merhodsfor Diamond Films 403
Figure 6. SEM micrographs of a (001)-textured diamond film with {OOl) morphology (by courtesy
of M. Fryda).
(001) facets as to be expected from the considerations made above. Similarly, (001)-
textured films with { 11 1) facets were obtained in the opposite extreme (a! = 3),
where pure octahedra were grown as single crystallites. Upon a reduction of a! to
about 1.9, corresponding to single crystallites appearing as octahedra with truncated
corners, a material dominated by (001) facets, with (001) about 10-15" from the
substrate normal, was formed. Figure 6 shows the surface of a film of this kind,
grown at the author's institute.
Growth of diamond films is frequently accompanied, however, by the formation
of twins stabilizing a pronounced (011) texture. An analysis of the formation
mechanism of this type of texture was presented by Wild et al. According to the
principle of evolutionary selection, a (01 1) texture should only occur at cy = 1.5,
with protruding corners formed by {OOl} and { 11 1) faces. However, roof-shaped
structures bounded only by { 11 1) faces were observed in thick (01 1)-textured
films [34]. Taking into account the effect of twinning on (111) faces the authors
presented arguments for the destabilization of growth of (001) tips in favor of a
preferential (01 1) growth direction, forming the observed roof-shaped appearance.
As shown in [34] and [35], high values of a are favored by high methane concen-
trations and low deposition temperatures. These findings are in agreement with
results of direct growth rate measurements performed on single cubo-octahedral
crystallites and homoepitaxial films, showing a stronger methane-concentration
dependence of the [OOl] growth rate, however a larger activation energy (tem-
perature dependence) for [11 11 growth [36]. Knowledge of the growth parameter
dependence of a permits a large scope in tailoring film texture and morphology
independently. Thus, using a two step procedure, the growth of smooth, perfectly
(001)-textured films with (001) morphology is made possible [37].
The concentration of molecular nitrogen in the source gas was recognized as a
very important third parameter determining the growth morphology and texture:
Small amounts (10-100 p.p.m.) of nitrogen were found to improve the crystalline
perfection and favor the formation of smooth (001 }-faceted (001)-textured films
considerably [22]. An analysis of HF-CVD homoepitaxial film growth rates
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
404 4 Chemical Vapor Deposition of Diamond Films
showed, that the addition of 100 p.p.m. nitrogen accelerated the [ l l l ] growth by a
factor of 2.2, however, a 4.4-fold increase was found for the [OOl] growth rate.
Therefore, (Y is multiplied by a factor of 2. A probable explanation of this phenom-
enon was given in terms of the surface reconstruction by dimerization on (001)
faces, leading to the F-face type behavior (see above): The adsorption of nitro-
gen-containing species could partially break this surface reconstruction, leading to
the formation of kink sites and an enhancement of the growth rate on this face.
An alternative mechanism leading to the growth of textured diamond films is
related to the bombardment with low-energy ions [38]. By using a carefully
controlled substrate bias-voltage during film growth, films with (001) texture and
(001) morphology could be obtained on diamond (film) substrates of any orienta-
tion. Seemingly the ion flux leads to a renucleation and a preferential growth of only
those crystallites which are oriented with one [OOl] direction parallel to the ion flux
direction, possibly due to their higher stability with respect to a flow of etching H+
ions.
Figure 7. SEM micrograph of a heteroepitaxial diamond film obtained on (001) silicon by bias-
induced nucleation, film thickness =6.5 pm. Vertical and horizontal edges of the photograph are
orientated parallel to the Si [I 101 and [LlO] directions, respectively (by courtesy of X. Jiang).
in this direction has already been made; tilt angle as well as pole figure FWHM
values of less than 5” were observed in films of 10-20pm thickness; in thicker
films (100 pm), about 2” have been reached.
on 3C-Sic films on silicon, exceeding probably the quality of other films grown on Si
or S i c before [51]. The ultimate goal of diamond heteroepitaxy is doubtlessly the
achievement of deposition of very high quality films on silicon, owing to its cheap.-
ness and ready availability of large-diameter wafers.
H2 +2H AG$,OO
= +75 kJ mole-', (2)
C, + 4H +CH4 AGy500 = -470 kJ mole-'. (3)
Methane reacts with H forming a series of CH, radicals (x = 3.0)
CH4
CH3 + H --
+ H +CH3 + H2
CH2 + H2
CH2 + H +CHI + H2
AGy5oo = -42 kJ mole- 1 ,
AGySoo= +17 kJ mole-',
AGySoo= -38 kJmole- 1 ,
(4)
(5)
(6)
CHI+H+C+H2 AGy5oo = -84 kJ mole-', (7)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
408 4 Chernicd Vapor Deposition of Diurnond Films
and carbon atoms are finally added to diamond ( c d is a carbon atom in the diamond
lattice):
c + Cd --f 2cd AGy5oo = -480 kJ mole-'. (8)
Summarizing reactions (3) to (8), one obtains a thermodynamically strongly favored
net reaction (9) for the graphite-to-diamond transformation, driven by the input of
atomic hydrogen:
C, + 8H -+ Cd + 4H2 AGy5oo = - 1097 kJ mole-', (9)
The mole fraction of methyl in the growth chamber was then calculated by doubling
the sum of the ethane and ethylene mole fractions. From the methyl mole fractions
H atom concentrations in the chamber were estimated by assuming a partial
equilibrium for the reaction (12):
CH4+H*CH3+H2. (12)
The methyl mole fraction increased from 2 x lop4 to 7 x lop4, when the methane
concentration was raised from 0.5 to 3.2 vol-YO. The H fraction amounted to
about 0.01 at 2vol-% methane.
The most direct experimental approach to the species involved in diamond
growth, molecular beam mass spectrometry of stable and reactive species extracted
through a 300 pm orifice in the substrate plane, was used by Hsu to investigate the
HF-CVD as well as the MWP-CVD process [57,58]. The most interesting results of
these measurements are as follows.
(i) In HF-CVD, acetylene is the predominant carbonaceous species near the
substrate, if a multi-filament arrangement is used, while methane prevails
under a single filament.
(ii) Methyl and acetylene mole fractions saturate at lo-’ and 7 x lo-’, if the
methane concentration is increased beyond 2%, accompanied by a decrease
in the H atom fraction from 2 x lop’ at methane fractions below 1% to
about 1 Op4 at 7 % methane (attributable to filament poisoning).
(iii) Owing to a rapid removal of H atoms by surface recombination, reaction (12) is
not in partial equilibrium, as assumed [56], but shifted to the methane side.
Based on these relatively early investigations, no decision could be made as to which
of the various species detected was actually responsible for diamond growth,
although in many of these studies only CH3 and C2H2 were left as possible
candidates. Very sophisticated diagnostic techniques involving nonlinear laser
spectroscopy have, in recent years. been applied to study growth atmospheres in
diamond CVD processes. However, the most conclusive information about which
species are responsible for bringing carbon to the growing diamond film came
from a number of suitably designed growth experiments. Several of these experi-
ments point to CH3 as the principal species in commonly used CVD environments,
although minor contributions of other species like acetylene cannot completely be
excluded.
Isotopic labeling experiments using 13CH4/’2C2Hz feed gases in the hot-filament
process were performed by Chu et ul. and D’Evelyn et al., first for polycrystalline
film growth [59,60], later for homoepitaxial growth on diamond {OOl}, {Oll},
and { 111) surfaces [61], and reanalyzed in order to account for the nonlinearity
of the first-order Raman shift frequency versus I3C mole fraction relation, which
was used to determine the isotopic composition of the diamond films grown [62].
Irrespective of the crystallographic orientation, corrected ”C mole fractions in
the films were identical, within experimental errors, to the corresponding mole
fractions in methane sampled directly above the growth surface. Owing to the fast
reaction (12), the isotopic composition of methyl and methane can be assumed to be
virtually equal. Isotopic scrambling between methane and acetylene, on the other
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
410 4 Chiwiccrl Vripor. Deposition of Dicltnoncl Films
hand, was suppressed by injecting the gas sufficiently close to the diamond surface.
Similar experiments were performed by MWP-CVD, also demonstrating that the
methyl radical is about an order of magnitude more efficient in diamond formation
than acetylene [63].
Prevention of the scrambling between C, and C2 species is also the essential idea
behind the flow tube studies of Martin and Hill [64], experiments with forced
convection of Schiifer et ul. [65],and experiments with supersonic jets performed
by Lee et ul. [66],all demonstrating the importance of methyl radicals as diamond
precursors. Martin and Hill used a microwave plasma in order to generate a stream
of hydrogen atoms in a flow-tube system. After injection of a small amount of
methane into the flowing gas the formation of diamond could be achieved in a
furnace at 8SO’C a few centimeters downstream, corresponding to a reaction time
of the order of 1 ms. This time is sufficiently long for reaction (12) to come to
equilibrium, but too short for major Cl-species formation. Diamond formation
was also observed after acetylene injections, however the rates were about an
order of magnitude smaller than for methane.
The investigations of Schiifer et ul. were stimulated by observations made during
HF-CVD experiments using arrays of four parallel uncoiled Ta wires, 1 1 .S mm from
each other. Under conditions of free convection of the gas phase a strong decrease of
the diamond growth rate from the edges of a wafer towards its center was noted. The
authors concluded that the reason for this observation was a virtually complete
conversion of carbon species to acetylene, which does not act as a growth species.
In order to test these assumptions, they investigated the effects of vertical gas jets
(hydrogen with 0.5% methane and acetylene) blown through the filament array with
a velocity of about 1000 cm s-’ and striking the substrate vertically. With methane,
maximum growth rates of several pm h-’ were found at the stagiiation point of the
jet while in the case of acetylene a growth rate depression was observed at this point.
A practically complete prevention of the interconversion between CH, and C2H,
was achieved in the experiments of Lee rt al., using supersonic free jets of methyl
radicals, generated by the pyrolysis of azomethane, and atomic hydrogen. Gas
phase chemistry can thus be virtually eliminated. Good quality diamond crystallites
could be grown on submicron diamond seeds, exposed to these jets at 650-8SO’C.
Diamond deposition ceased, when either the azomethane pyrolysis or the glow
discharge generating atomic hydrogen was turned of. Graphitic carbon was the
main product, when an acetylene jet was used instead of a methyl jet.
and man-made materials is not only based on its mechanical properties but also its
optical transparency over a broad wavelength range, its unsurpassed thermal con-
ductivity at ambient temperatures, several outstanding electrical properties (band
gap, breakdown field strength, charge carrier mobilities, saturated carrier velocities)
and chemical inertness. Many prospective applications make use of one or more of
these physical properties. Several handbooks have meanwhile appeared which give
very exhaustive accounts of the properties and applications of natural or man-made
diamond single crystals [67] and diamond films [68,69]. The subsequent sections will
therefore be restricted to a concise introduction of only the most important applica-
tions (which often are still in a state of infancy or even only prospective) and the
properties which are relevant for these applications.
piece materials, significant progress has been made in tool lifetime and reliability.
Therefore, diamond-coated tools are now just beginning to play a role in the
machining of nonferrous and nonmetallic materials [70].
K. = f(CvA). (13)
Using v w 1O6cmC' (of the order of the velocity of sound in diamond) and the
literature value of C (1.82 J K-' cmP3), a mean free path length of 0.4 pm can be
calculated, roughly two orders of magnitude larger than for NaCl or silica single
crystals. Theoretically, the large value of A can be traced back to the strong
bonds between light atoms forming the diamond lattice. This consideration also
demonstrates that a diamond crystal has to be perfect over dimensions of at least
a few micrometers in order to prevent phonon scattering by point, line, or planar
defects like impurities, dislocations, or stacking faults, reducing K. below the best
single crystal values. The thermal conductivity is therefore critically dependent on
crystalline quality and purity of a diamond crystal or film.
Early thermal-conductivity measurements on CVD diamond were performed on
relatively thin films (less than 30 pm) with average lateral grain sizes of only a few
micrometers, grown by MWP-CVD [71] or HF-CVD [72]. For both growth
methods the in-plane thermal conductivity K~~ decreased from about 10 to less
than 2 Wcm-' K-', if the methane concentration in the source gas was increased
to several percent, accompanied by a deterioration of the film quality as to be
judged from Raman spectra and (in the case of the HF-CVD samples) an increase
of the H-content from 0.1 to 1.0%.
Significantly larger thermal conductivities were achieved, when films of several
l00pm thickness were grown by MWP-CVD or HF-CVD [73,74]. Owing to
grain dimensions attaining 50-100 ym in the top regions of films of 350-600 pm
thickness, grain boundary scattering was strongly reduced and room temperature
thermal conductivities of 17 Wcm-' K-'could be achieved. The average thermal
conductivity of a thick diamond film measured in-plane or vertically to the film
plane can be considered as an integral over increments from local contributions
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 Properties and Applications of' Chemical Vapor Deposited Diamond 4 13
of differential slices parallel to the film plane. Graebner and coworkers analyzed
thick films with respect to the local values of in-plane and vertical thermal con-
ductivities and discovered, that both quantities increased with the distance from
the substrate surface up to values of 21 and 23.5 W cm-' K-' in the top region
of a 350 pm thick film [75,77]. The latter value is already close to the best single
crystal data.
The high thermal conductivity of diamond forms the basis for the application of
CVD diamond in heat sinks which is expected to play an increasing role in the near
future: Semiconductor-based laser diodes or IMPATT microwave diodes generate a
large amount of heat which has to be removed efficiently in order to prevent an early
failure of the device due to excessive heating. The efficiency of water-cooled heat
sinks is limited, due to the relatively small area across which the heat flux is entering
into the sink. Therefore heat spreaders are used in order to enlarge the area, over
which the heat is conducted to the heat sink. These spreaders must not present a
large thermal resistance themselves. Single-crystalline diamonds synthesized by
the high-pressure, high-temperature method are mainly used for this purpose
today [67]; however, the price of these crystals increases very rapidly with size.
Therefore CVD films can gain importance where larger heat sinks with dimensions
reaching 1 cm or more are necessary [69].
Thermal management of computers could possibly open a much larger future
market for thick diamond films of maximum thermal conductivity. Modern multi-
chip modules involve power density levels which are difficult to cope with using
conventional substrate materials. The physical properties of diamond, aside from
thermal conductivity, the highly insulating nature, and a very low coefficient of
thermal expansion, make it an optimal choice for managing the thermal problems
encountered in present days two-dimensional packaging and in prospective three-
dimensional arrangements of integrated circuits [69].
Table 1. Comparison of physical properties of diamond with those of other semiconductors (after
[781)
Si GaAs 3C-Sic Diamond
where c in KFM is the speed of light, the other symbols are defined in the table.
JFM was designed to measure the upper limits of a material's performance in a
transistor handling simultaneously high frequency and high power. The superior
position of diamond is mainly due to its very large breakdown field strength. In
the derivation of KFM the thermal limits on the speed of a semiconductor
device are considered; it is mainly the unsurpassed thermal conductivity of
diamond which gives it the highest KFM value. Apart from these advantages of
diamond as a semiconductor material for high-power, high-frequency applications,
its very large band gap, making undoped pure diamond very highly insulating at
room temperature and up to very high temperatures, is the prerequisite for
prospective applications in active devices working at elevated temperatures up to
500°C or more.
So far, only p-type doping with boron (incorporated from a suitable volatile
source during CVD growth) has become a standard practice. In homoepitaxial
films, mobilities reaching 1600 cm2V-' s-' have been reported, indicating a good
crystalline quality similar to natural IIb-type diamond single crystals. Doping of
the n-type by phosphorus during CVD growth is still in its infancy. Correspond-
ingly, the list of 24 publications on active diamond devices in [69] mainly refers to
unipolar field-effect transistors, mostly prepared from in situ p-doped homoepitaxial
films or B ion implanted layers on single crystals. In several cases transistor function
and operation at elevated temperatures up to 500°C were demonstrated. In order to
develop an economically viable diamond device technology, however, diamond
single crystals are hardly suitable substrates. Owing to grain-boundary effects in
randomly oriented polycrystalline films grown on cheap silicon substrates, device
performance in this material is inferior by orders of magnitude. Therefore, the
growth of heteroepitaxial films on Si or Sic substrates might gain importance for
future electronic applications. Even if the films are still polycrystalline, the detri-
mental effect of grain boundaries on physical properties is reduced owing to
much smaller intergrain angles. Electrical properties (in-plane hole mobility,
through-plane leakage currents) as well as possibly the thermal conductivity are
therefore superior to those of randomly oriented material [79]. For a metal-oxide
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 Properties and Applications of Chemical Vapor Deposited Diamond 4 15
E"AC
----xt- 5.5 eV
clean H-terminated
Figure 8. Alignments of valence and conduction bands of clean and H-terminated (1 11) surfaces of
p-doped diamond, EF = Fermi level. The figures have been aligned at the vacuum level (adapted
from [SO]).
+
Xads X-
X- -Xads
+X2
+ e- ,
+ e-.
(16)
(17)
It is inferred that generally electrochemical reactions involving surface-absorbed
species will proceed comparably slowly on pure diamond surfaces. On the other
hand, outer sphere electron transfer reactions not requiring a strong surface
interaction of the species involved, will in principle not be affected by the inert
nature of the diamond surface, as to be demonstrated, for example, by the reversible
behavior of the Ru(III)/Ru(II) redox couple [85].
The actual technical potential of diamond electrodes for analytical and prepara-
tive electrochemical applications can so far not be fully assessed. From various
prospective applications mentioned in the literature, making use of the exceptional
stability, wide potential window and low background currents of diamond
electrodes, only one example shall be mentioned here, the reduction of nitrate to
ammonium according to the net Eq. (18)
NO, + 7H20 + 8e- +N H 4 0 H + 90H-, Eo = -0.15 V vs. SHE. (IS)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 4 17
Thanks to the special properties of diamond electrode surfaces, this reaction can be
run with high efficiency and stability [86]. A potential application is the removal of
nitrates from the effluents of combustion processes or from drinking water.
4.5 Summary
Only a small selection of possible applications of CVD diamond could be touched
on here. Many of the expectations which were placed on the applications of CVD
diamond in the early years of the ‘new diamond era’ have so far not come true,
due to reasons like still too high deposition costs or insufficient film qualities for
certain applications, major advances in competitive materials and technologies, or
just the lack of a market. Nevertheless, thanks to the tremendous progress which
has been made in the recent years and will undoubtly be made in the future, CVD
diamond can be expected to play an ever increasing role as the ultimate material for
a variety of technical purposes.
References
1. J. C. Angus, in The Physics of Diamond, A. Paoletti and A. Tucciarone (Eds), 10s Press,
Amsterdam, Oxford, Tokyo, Washington DC, 1997, p. 1.
2. R. C . DeVries, A. Badzian, and R. Roy, M R S Bulletin, 1996, 65-75.
3. P. K. Bachmann and H. Lydtin, in Diumondand Diamond-like Films and Coatings, R. E. Clausing,
L. L. Horton, J. C. Angus, and P. Koidl (Eds), Plenum, New York, 1991, p. 829.
4. S. Matsumoto, Y. Sato, M. Kamo, and N. Setaka, Jpn. J. Appl. Phys. 1982, 21, L183-Ll85.
5. S. Matsumoto, Y. Sato, M. Tsutsumi, and N. Setaka, J. Muter. Sci. 1982, 17, 3106-31 12.
6. S. Okoli, R. Haubner, and B. Lux, J. PhJjsique ZV 1991, C2, 923-930.
7. M. Sommer and F. W. Smith, J . Muter. Res. 1990, 5, 2433-2440.
8. B. Singh, Y. Arie, A. W. Levine, and 0. R. Mesker, Appl. Phys. Lett. 1988, 52, 4511152.
9. L. Schafer, U. Bringmann, C.-P. Klages, U. Meier, and K. Kohse-Hoinghaus, in Diamond and
Diamond-Like Films and Coatings, R. E. Clausing et al. (Eds), Plenum Press, New York, 1991,
p. 643..
10. F. Jansen, M. A. Machonkin, and D. E. Kuhman, J. Vac. Sci. Technol. 1990, AS, 3785-3790.
11. E. Gheeraert, A. Deneuville, M. Brunel, and J. C. Oberlin, Diamond Rel. Muter. 1992,1,504-507.
12. M. Griesser, G. Stingeder, M. Grasserbauer, H. Baumann, F. Link, P. Wurzinger, H. Lux,
R. Haubner, and B. Lux, Diamond Rel. Muter. 1994, 3, 638-644.
13. Y. Hirose and Y. Terasawa, Jpn. J. Appl. Phys. 1986, 25, L519-L521.
14. M. Murakawa, S. Takeuchi, H. Miyazawa, andY. Hirose, Surf: Coat. Technol. 1988,36,303-310.
15. T. Kawato and K. Kondo, Jpn. J. Appl. Phys. 1987,26, 1429-1432.
16. Y. Saito, K. Sato, H. Tanaka, K. Fujita, and S. Matsuda, J. Muter. Sci. 1988, 23, 842-846.
17. C . Chen, Y. C. Huang, S. Hosomi, and I. Yoshida, Muter. Res. Bull. 1989, 24, 87-94.
18. J. A. Mucha, D. L. Flamm, and D. E. Ibbotson, J. Appl. Phys. 1989, 65, 3448-3452.
19. S. J. Harris and A. M. Weiner, Appl. Phys. Lett. 1989, 55, 2179-2181.
20. D. N. Belton and S. J. Schmieg, J . Appl. Phys. 1991,69, 3032-3036.
21. P. K. Bachmann, D. Leers, and H. Lydtin, Diamond Rel. Muter. 1991, 1, 1-12.
22. G. Z. Cao, J. J. Schermer, W. J. P. van Enckevort, W. A. L. M. Elst, and L. J. Giling, J. Appl.
Phys. 1996, 79, 1357-1364.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
418 4 Chemical Vapor Deposition of Diamond Films
23. M. Kamo, Y. Sato, S. Matsumoto, and N. Setaka, J. Cryst. Growth, 1983, 62, 642-644.
24. C.-P. Klages, Appl. Phys. A 1993, 56, 513-526.
25. P. K. Bachmann, in Handbook of Industrial Diamond and Diamond Films, M. A. Prelas,
G. Popovici, and L. K. Bigelow (Eds), Marcel Dekker, New York, Basel, Hong Kong, 1998,
Chapter 21.
26. H. Kawarada, K. S. Mar, and A. Hiraki, Jpn. J . Appl. Phys. 1987, 26, L1032-LIO34.
27. J. Suzuki, Jpn. J . Appl. Phys. 1989, 28, L281LL283.
28. A. A. Chernov, Modern Crystallography ZIZ, Springer Verlag, Berlin, Heidelberg, New York,
1984, p. 283.
29. A. Van der Drift, Philips Res. Repts. 1967, 22, 267-288.
30. P. Hartman, in Crystal Growth; An Introduction, P. Hartmann (Ed.), North-Holland, Amsterdam,
1973, p. 367.
31. L. J. Giling and W. J. P. Van Enckevort, Surface Science, 1985, 161, 567-583.
32. C. Wild, P. Koidl, W. Miiller-Sebert, H. Walcher, R. Kohl, N. Herres, R. Locher, R. Samlenski,
and R. Brenn, Diamond Rel. Mater. 1993, 2, 158-168.
33. R. E. Clausing, L. Heatherley, L. L. Horton, E. D. Specht, G. M. Begun, and Z. L. Wang,
Diamond Rel. Muter. 1992, 1, 411415.
34. Ch. Wild, N. Herres, and P. Koidl, J. Appl. Phys. 1990, 68, 973-978.
35. Ch. Wild, R. Kohl, N. Herres, W. Muller-Sebert, and P. Koidl, Diamond Rel. Muter. 1994, 3,
373-381.
36. H. Maeda, K. Ohtsubo, M. Irie, N. Ohya, K. Kusakabe, and S. Morooka, J . Muter. Res. 1995,
10, 3115-3123.
37. R. Locher, C. Wild, N. Herres, D. Behr, and P. Koidl, Appl. Phys. Lett. 1994,65, 34-36.
38. X. Jiang, W. J. Zhang, M. Paul, and C.-P. Klages, Appl. Phys. Lett. 1996, 68, 1927-1929.
39. B. R. Stoner and J.T. Glass, Appl. Phys. Lett, 1992, 60, 698-700.
40. X. Jiang and C.-P. Klages, lecture at Diamond 92, Heidelberg 1992, published in: Diamond Rel.
Muter. 1993,2, 1112-1 113.
41. S. Yugo, T. Kanai, T. Kimura, and T. Muto, Appl. Phys. Lett. 1991,58, 10361038.
42. B. R. Stoner, G.-H. M. Ma, and S. D. Wolter, J.T. Glass, Phvs. Rev. B 1992.45. 11067-11084.
43. X. Jiang, C.-P. Klages, M. Rosler, R. Zachai, M. Hartweg, and H.-J. Fusser, Appl. Phys. 1993,
A51,483489.
44. K. Schiffmann and X. Jiang, Appl. Phys. 1994, A59, 17.
45. J. Gerber, S. Sattel, K. Jung, H. Ehrhardt, and J. Robertson, Diamond Rel. Muter. 1995, 4,
559-562.
46. S. Sattel, J. Gerber, and H. Ehrhardt, Phys. Stat. Sol. ( A ) 1996, 154, 141-153.
47. Y. Lifshitz, S. R. Kasi, and J. W. Rabalais, Phys. Rev. Lett. 1989, 62, 1290-1293.
48. J. L. Robertson, S. C. Moss, Y. Lifshitz, S. R. Kasi, J. W. Rabalais, G. D. Lempert, and
E. Rapoport, Science, 1989, 243, 1047-1050.
49. X. Jiang, K. Schiffmann, and C.-P. Klages, Phys. Rev. B 1994,50,8402-8410.
50. T. Suzuki and A. Argoitia, Phys. Stat. Sol. ( A ) 1996, 154, 239-254.
51. H. Kawarada, T. Suesada, and Nagasawa, Appl. Phys. Lett. 1995, 66, 583-585.
52. J. C. Angus, in The Physics of Diamond, A. Paoletti and A. Tucciarone (Eds), 10s Press,
Amsterdam, Oxford, Tokyo, Washington DC, 1997, p. 9.
53. D. G. Goodwin and J. E. Butler, in Handbook of Industrial Diamond and Diamond Films, M. A.
Prelas, G. Popovici, and L. K. Bigelow (Eds), Marcel Dekker, New York, Basel, Hong Kong,
1998, Chapter 11.
54. F. G. Celii, P. E. Pehrsson, H.-T. Wang, and J. E. Butler, Appl. Phys. Lett. 1988,52,2043-2045.
5 5 . J. E. Butler and F. G. Celii, in Proc. 1st Znt’l Symp. on Diamond and Diamondlike Films, J. P.
Dismukes et al. (Eds), The Electrochemical Society, Pennington, NJ, 1989, p. 317.
56. S. J. Harris and A. M. Weiner, J . Appl. Phys. 1990,67,6520-6526.
57. W. L. HSU,Appl. P h y ~Lett.
. 1991,59, 142771429,
58. W. L. Hsu, in Proc. 2ndZnt’l Symp. on DiamondMaterials, A. J. Purdes et al. (Eds), The Electro-
chemical Society, Pennington, NJ, USA, 1991, p. 217.
59. C. J. Chu, M. P. D’Evelyn, R. H. Hauge, and J. L. Margrave, Muter. Res. SOC.Symp. Proc.
1990, 162, 85-90.
60. C. J. Chu, M. P. D’Evelyn, R. H. Hauge, and J. L. Margrave, J . Muter. Res. 1990,5,2405-2413.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 419
61. C. J. Chu, M . P. D’Evelyn, R. H. Hauge, and J. L. Margrave, J . Appl. Phys. 1991,70, 1695-1705.
62. M. P. D’Evelyn, C. J. Chu, R. H. Hauge, and J. L. Margrave, J . Appl. Phys. 1992,71, 1528-1530.
63. C. E. Johnson, W. A. Weimer, and F. M. Cerio, J. Muter. Res. 1992, 7, 1427--1431.
64. L. R. Martin and M. W. Hill, J. Muter. Sci. Lett. 1990, 9, 621-623.
65. L. Schafer, M. Sattler, and C.-P. Klages, in Applications of Diamond and Related Materials,
Y. Tzeng e f al. (Eds), Elsevier, Amsterdam, 1991, p. 453.
66. S. S. Lee, D. W. Minsek, D. J. Vestyck, and P. Chen, Science 1994, 263, 1596-1598.
67. J. E. Field (Ed.), The Properties qf Diamond, Academic Press, London, 1997.
68. G. Davis (Ed.), Properties and Growth qf Diamond, EMIS Datareview Series, INSPEC,
London, UK, 1994.
69. M. A. Prelas, G. Popovici, and L. K. Bigelow (Eds.) Handbook of Industrial Diamond and
Diamond Films,Marcel Dekker, New York, Basel, Hong Kong, 1998.
70. C.-P. Klages, M. Fryda, T. Matthee, L. Schafer, and H. Dimigen, in Proceedings ofthe 14th
International Plansee Seminar, G. Kneringer, P. Rodhammer und P. Wilhartitz (Eds), Plansee
AG, Reutte, Austria, 1997, Vol. 3, 1-15.
71. A. Ono, T. Baba. H. Funamoto, and A. Nishikawa, Jpn. J . Appl. Phys. 1986, 25, L808-LSlO.
72. D. T. Morelli, C. P. Beetz, and T. A. Perry, J . Appl. Phys. 1988, 64,3063-3066.
73. D. T. Morelli, T. M. Hartnett, and C. J. Robinson, Appl. Phys. Letf. 1991, 59, 2112-2114.
74. J. E. Graebner and J. A. Herb, Diamond Films Technol. 1992, 1, 155-163.
75. J. E. Graebner, S. Jin, G. W. Kammlott, J. A. Herb, and C. F. Gardinier, Appl. Phys. Lett. 1992,
30, 1576-1 578.
76. J. E. Graebner, S. Jin. G. W. Kammlott, B. Bacon, L. Seibles, and W. Banholzer, J. Appl. Phys.
1992,71, 5353-5356.
77. J. E. Graebner, S. Jin, G. W. Kammlott, J. A. Herb, and C. F. Gardinier, Appl. Phys. Lett. 1992,
359,401403.
78. J. F. Prins, in The Physics of Diamond, A. Paoletti, and A. Tucciarone (Eds), 10s Press,
Amsterdam, Oxford, Tokyo, Washington DC, 1997, p. 41 1.
79. C.-P. Klages in The Physics of Diamond, A. Paoletti, and A. Tucciarone (Eds), 10s Press,
Amsterdam, Oxford, Tokyo, Washington DC, 1997, p. 73.
80. R. J. Nemanich, P. K. Baumann, M. C. Benjamin, S. P. Bozeman, and B. L. Ward, in The
Physics of Diamond, A. Paoletti and A. Tucciarone (Eds), 10s Press, Amsterdam, Oxford,
Tokyo, Washington DC, 1997, p. 537.
81. P. M. Natishan and A. Morrish, Materials Letters, 1989, 8. 269-272.
82. M . Werner, R. Job, A. Zaitzev, W. R. Fahrner, W. Seifert, C. Johnston, and P. R. Chalker,
Phys. Stat. Sol. (‘A), 1996, 154, 385-393.
83. G. M . Swain, J. Electrochem. Soc. 1994, 141, 3382-3393.
84. H. B. Martin, A. Argoitma, U. Landau, A. B. Anderson, and J. C. Angus, J. Electrochem. Soc.
1996, 143, L133-LI36.
85. N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J . Electrochem. Soc. 1996, 143, L238-L240.
86. R. Tenne, K. Patel. K. Hashimoto. and A. Fujishima, J . Elecfroanal. Chem. 1993,347,409415.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
5.1 Introduction
Connected with the advancement of synthetic diamond as a technical material, cubic
boron nitride (cBN) has attracted increasing interest during the past decades.
Boron nitride exists in high density diamond-like and in low density graphite-like
modifications. High density cubic boron nitride has a zinc-blende structure with
eight atoms in the elementary cell. Each B atom and each N atom is surrounded
by four N or B atoms, respectively. Like in the diamond structure the neighboring
atoms are tetrahedrally arranged with fourfold coordination. Diamond and cBN
have similar lattice constants. However, diamond has a pure covalent bonding
and cBN a covalent-ionic mixed bonding. Hexagonal boron nitride (hBN) and
graphite have similar structures, both with a threefold coordination of nearest
neighbors. Beside hBN and cBN two other BN structures are known, the rhombo-
hedral (rBN) and the metastable high-density wurtzite-type BN (wBN). Apparently
the wBN phase only can be synthesized by shock compression. On the one hand
hBN and rBN and on the other hand cBN and wBN have similar structures and
the same short range order but different stacking sequences of atomic planes (see
[l]). In Table 1 the structural data as well as some properties of cBN, diamond,
hBN, and graphite are summarized.
Cubic BN is the second hardest material after diamond. It does not react with iron
in tribological contact [2] and can therefore (in contrast to diamond) be used to coat
tools for machining of ferrous materials.
Because of its wide band gap (Eg= 6.4eV [ 3 ] ) , cBN transmits electromagnetic
radiation in a broad range including the near ultraviolet as well as the visible and
infrared region. Moreover, cBN can be doped both n- and p-type [4] and has
good thermal conductivity. Therefore, additional potential applications are as pro-
tective coatings for optical devices or in electronics, either as a high temperature
semiconductor or as heat dissipating material.
Analogous to graphite in the carbon case, boron nitride exhibits a low-density
relatively soft hexagonal phase hBN. Formation of boron nitride by chemical
reaction at atmospheric pressure yields the hexagonal phase. However, by using
the high pressure, high temperature (HPHT) synthesis established by Wentorf [5]
it is possible to transform hBN into cBN. In this method, ( p , T ) values in the
range 1200-2000°C and 4.5-7.5 GPa, respectively [6], were applied corresponding
to conditions where cBN is thermodynamically stable [7] and formed comparatively
readily. Today, small crystallites of cBN having a typical size in the several 100 mm
range [S] are produced in considerable quantities with this method and are processed
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.2 Empirical Results 42 1
0.333 0.335
Density (gcm-') 3.48 3.52 2.28 2.27
Hardness (GPa) 5&70 9&100 approx. 10 <10
Young's modulus (GPa) 700-900 1050 a 87 a 19
c 34 c 5
Band structure Indirect Indirect Direct Metallic
semiconductor semiconductor semiconductor
Optical band gap (eV) 6.1 5.5 5.2 -
Refractive index 2.12 2.42 1.72 -
(A = 589.3 nm)
Resistivity (0 cm) 10'O 1013 a 10'' a lo-'
c lo1* c 1
Dopability p and n P (n ?)
Thermal conductivity <750 2000 1 6.4-10.8
(W m-I K - I )
Oxidation at air ("C) >loo0 >500 >I000 -
Reactivity with iron at No Yes No Yes
elevated temperatures
mainly to sintered polycrystalline boron nitride (pcBN) pieces as cutting edges for
tools.
However, by growing thin BN films from the vapor phase it is possible to get cubic
BN as a thin film at moderately increased temperatures as well as subatmospheric
pressure. Starting in the SOs, mainly by Japanese researchers [9], great efforts have
been made in this field up to now.
The aim of this chapter is to present an overview of some important recent
developments in cBN thin film deposition. The physical mechanisms which are
causal for cBN formation in ion-assisted thin film deposition will be discussed in
some detail.
did predominate which use electron beam evaporation of boron. The necessary ion
impact to the growing film was performed either by special activation and ionization
of the nitrogen containing gas atmosphere together with a suitable substrate bias to
accelerate the ions [10-12] or by means of ion beam sources [13,14]. The use of ion
beam sources to realize ion beam impact to the growing film has also been combined
with boron vaporization by ion beam sputtering [15,16] or laser ablation [17] of
either a boron or BN target. Mass-selected ion beam deposition (MSIB) [18] and
cathodic arc deposition (CAD) [19] have also been successfully used to deposit
cBN thin films. With the last mentioned two methods the film is formed essentially
by deposition of ions.
Because of its potential for application, R F sputtering [20,21] and R F magnetron
sputtering [22,23] has attracted increasing attention. Deposition by d.c. magnetron
sputtering (which is even more interesting from an application point of view) has
also been successfully applied to cBN thin films, either using a B4C target [24]
(section 5.4.1) or a boron target which is operated at elevated temperature where
boron gets conductive [25] (section 5.4.2).
Different variants of CVD techniques were applied, for example R F glow dis-
charge using the gases diborane B2H6, N2, Ar, He, and H2 [26] as well as electron
cyclotron resonance (ECR) plasma CVD processes with B2H6and N2 [27] or with
the vapor of N-trimethylborazine [28]. The authors reported that cBN could only be
generated if the negative substrate bias was sufficiently high. BN films deposited by
such CVD methods seem to have in general somewhat lower cBN contents than
those prepared by PVD techniques. Probably this is due to the incorporation of
hydrogen into the films [29].
Zhang et ul. [30,31] deposited films on nickel in a hot filament assisted R F glow
discharge with the reactant gases being B2H6, NH3, and H2. Obviously the sub-
strates were not biased. Scanning electron microscopic images revealed crystallites
with dimensions of about 2 pm. As an essential reason for the rather large crystal-
lites, a catalytic effect of Ni was assumed. Unfortunately, for the evidence of the
cubic phase only X-ray diffraction patterns, but no infrared spectra, were published.
Therefore considerable doubt that these crystallites consist of cBN remains.
Konyashin et al. [32] deposited BN films using plasma assisted CVD with gas mix-
tures of argon, hydrogen, and borane-ammonia (BH3-NH3) without substrate bias
at temperatures up 1000°C. Regrettably no data on hydrogen content and B and N
concentrations in the films were given. From IR spectra and X-ray diffraction
patterns of films deposited with high hydrogen concentrations in the process gas,
the existence of single phase cBN was concluded. The authors assumed a deposition
mechanism which includes a selective etching of sp2 bonded BN and the formation
of metastable BNH, species.
Reinke et al. [33] analyzed numerous deposition methods with respect to film
structure (cBN or hBN) on the one hand and ion bombardment on the other
hand. To describe the ion bombardment quantitatively they used the ion energy
and the ratio of fluxes of impinging ions (i) and deposited boron atoms (a).
Figure 1 shows the results for the ratio i/u against the ion energy. The total
number of deposited atoms can be estimated taking into account the deposition
rate and the mass densities of the BN phases. As can be seen from that presentation,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.2 Empirical Results 423
0
0
0
0 oo 0
0 0 .
0 0 0
10
I
100
0
.
1000
Ion energy [eV]
Figure 1. Data compilation according to Reinke et al. [33] showing the phase of BN thin films in
dependence on ion energy, Ei, and the ratio of bombarding ions and deposited atoms, i/a, for
numerous deposition methods. Filled and open symbols represent cBN and hBN films, respectively.
cBN has been grown in a broad range of ion energies Ei,beginning with quite low
energies connected with a large i/u (these are mostly CVD methods) up to higher E;
together with small i/u. The limiting case is represented by mass selected ion deposi-
tion where the film grows solely by deposition of Bf and Nf ions. In this case i/u is
one and cBN was formed at ion energies down to 125 eV [34].
In their successful attempts to grow cBN thin films, most authors have applied
moderately increased substrate temperatures, T,, mostly in the range between 200
and 500°C. However, the growth of cBN thin film is also possible at temperatures
as high as 1000°C [35]. The low temperature limit for cBN film deposition is at
about 150'C, as has been found by Mirkarimi [36] for ion-assisted pulsed laser
deposition and more recently by Hofsass et ul. [34] for MSIB deposition.
Figure 2. Cross-section TEM micrograph of a cBN film on single crystal silicon with the typical
interlayers of amorphous material (probably a mixture of B, N, and Si), turbostratic BN.
finally yield cBN growth (i.e. under proper substrate temperature and ion bombard-
ment as discussed above) the tBN exhibits a preferred orientation such that the c-
axis of the tBN clusters are predominantly perpendicular to the substrate surface
normal. After a certain amount of tBN deposition the film structure changes into
cBN. In Fig. 2 a cross-section transmission electron micrograph of this typical
phase sequence is shown. The thickness of the amorphous and tBN interlayers
depends on the deposition conditions. Especially the tBN thickness may be very
large, however, under optimum conditions the transition to cBN growth takes
place very soon and the thickness of the hexagonal BN layer is only about 1 nm
or even not detectable in cross-section TEM.
The cBN films are polycrystalline with columnar grains about 10-20nm in
diameter extended perpendicular to the surface.
The cBN crystallites have a preferential orientation. Some authors report on a
(1 10) texture [38,39], others [40] found a so-called double-fiber texture. Both
textures are characterized by an alignment of { 11l} planes perpendicular to the
surface. The difference is that in the case of the (110) texture the (111) planes
have a preferred azimuthal orientation (i.e. the (1 10) direction is parallel to the
substrate normal), whereas in the case of the double-fiber texture the orientation
about the in-plane [l 1 11 axis is random. However, both textures do support the
suggestion that the structural transition from hBN to cBN growth is characterized
by three (1 11) spaces of cBN (dl11 = 0.209 nm) matching two (0002) spaces of tBN
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.2 Empirical Results 425
(do,,, M 0.34 nm), thus forming a semicoherent interface. In the meantime this
model has been confirmed using cross-section TEM by us and others (see Fig. 3).
A necessary condition for cBN phase formation in thin films is that boron and
nitrogen are incorporated in a nearly 1: 1-ratio. Based on highly accurate com-
position measurement using neutron depth profiling, Hackenberger et al. [41]
investigated the relationship between film stoichiometry and the phases present in
the film. From these measurements, together with an analysis of data from the
literature, they found evidence that film stoichiometry is one of the factors that
stabilize the cubic phase.
Boron nitride films containing both the hBN and cBN phase cause strong infrared
absorption bands at about 1050cm-' owing to the presence of cBN (transversal
optical mode) [42] as well as at about 800 and 1400cm-' corresponding to B-N
bending (out-of-plane vibration) and stretching (in-plane vibrations) modes, respec-
tively [43].
Actually, the IR absorption features depend on the chemical short-range order of
the lattice, and their appearance is only a necessary (not a sufficient) condition for
the existence of the respective phases. There exist some more BN phases with either
sp2 or sp3 short-range order which show essentially the same IR absorption as hBN
and cBN, respectively. However, if used in combination with more direct methods
such as electron diffraction, IR absorption is a fast and reliable method for determi-
nation of the phase content of BN thin films and has found widespread application.
In Fig. 4 the IR absorption spectrum of a BN film is given as an example which
shows a distinct absorption near 1050cm-' and only weak absorption at about
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
426 5 Vupor Phusr Depositiori of Cubic Boron Nitride Films
1 .o
0.9
2
T-
Y
Wavenumber [cm']
Figure 4. Infrared absorption spectrum of a 130 nm thick BN film on silicon consisting mainly of
nanocrystallline cBN
800 and 1400 cm-I, respectively. This film consists practically of phase pure cBN.
The slight sp2-like absorption is mainly due to the non-cubic (amorphous and turbo-
stratic) interlayer mentioned above. Probably the not perfectly sp3 bound grain
boundary regions of the nanocrystalline cBN film contribute to the sp2 signal, too.
There were some attempts to reveal quantitative data on phase contents in mixed
BN films from IR spectra. Jager et al. [44] determined the cBN content from the
absorption coefficients a h (at a wave number near v = 1400cm-' for hBN) and
a, (near v = 1050cm-' for cBN). The ratio of the volume fractions,f,, and,fvhof
both phases was calculated as fvc/fih = 0.6ac/ah. However, in practice it is much
more simple to consider the peak heights 1;(i = h, c) in the spectra than to determine
the absorption coefficients ai. Schiitze et al. [24] derived the rough approximation
Z; z Q; (i = h,c) which leads to,jvc/fvh= 0.61,/Zh.
Considering the mass fractions,f,, it follows that,fi,,,/f;,,h = ( p , B N / p h B N ) *jvc/fvh
and,fm,/fn,h z Z,/Z, and the cBN content simply can be calculated as Zc/(Zc + Zh).
Using this relation one gets for the film with the spectrum shown in Fig. 4 a cBN
content of SO-90% which is also in good agreement with the impression gained
from cross-section TEM images.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3Morlel.~of c B N Forrnution 427
100
a,
20 c m I
8
0 2 4 6 8 10 12
Compressive Stress [GPa]
Figure 5. Cubic phase content in RN thin film deposited by IBAD against the compressive film
stress (adapted from [48]).
stress level in the films. An analysis of the respective cBN content of the films (Fig. 5)
showed that a certain threshold of compressive stress exists above which cBN for-
mation in the film takes place. They also reported on a typical dependence of the
film stress on the deposition time (zero stress + increase up to maximum + decrease
to saturation value) which at least qualitatively agrees with the above mentioned
sequence aBN + hBN cBN. An important conclusion from this model is that
---f
the compressive stress not only exists in conjunction with the cBN phase in the
layer, but is the cause of the cBN formation. An analogous model has already
been applied by the same authors to explain the formation of tetrahedral amor-
phous carbon (taC) [49].
Several authors conducted experiments under well-controlled conditions and
varied both the energy and mass of the ions in order to discriminate between the
specific role of ion energy and momentum. Kester and Messier [ 141 found a distinct
increase of the cBN content of the films at a certain threshold value of the trans-
ferred momentum per arriving atom. For the momentum transferred from a
single impinging ion they used an expression given in [50] for a single, binary, elastic,
head-on collision betwcen the ion and an atom of the film. These results underline
the relevance of momentum transfer processes in cBN formation. However, because
of the quite restricting suppositions as well as an error existing in the formula used,
their quantitative accuracy is limited.
Mirkarimi rt al. [17] reported on similar BN deposition experiments performed by
pulsed laser evaporation of an hBN target together with an ion beam bombardment
of the growing film performed by a special ion source. They varied the ion energy
(Eion= 500-1200 eV) and used ions of different mass (mion = ???Ar, m K r , nzxe).The
authors found that the cBN content of the films scaled best with ( M Z ~ ~ ~ E ~ ~ ~ )
(i.e. with the total momentum of impinging ions per deposited atom, ptoi/ii).Despite
the quite different deposition parameters in the particular experimental series all
dependences of the cBN percentage on those parameters could be traced back to
practically one and the same curve when ptoi/a as parameter was used. As an
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Spirfter-Deposirion of cBN Films 429
100 ,
80
c
--
E
6 60
s
2 4 0
41
20
0
0 100 200 300 400
P,,,/a (eV x amu)’
Figure 6. Dependence of the cBN percentage in BN thin films on p t o t / ufor different ion masses
(adapted from [ 171).
example, in Fig. 6 the cBN content dependence on ptot/u for different ion masses is
+
shown. To get practically phase-pure cBN films (i.e. Zc/(Zc Zh)2 0.8),p,,,/a has to
exceed a critical value of about 20(1-250 (eV amu)’I2 (amu: atomic mass units).
In addition, based on their ion impact approach, Mirkarimi et (11. [ 171 established
a quantitative model of defect generation during ion-assisted film growth. They
found that the maximum height of the defect production distribution exhibits
the same (w~lol,Eion)”2 dependence as cBN formation. Moreover, as the authors
emphasize, this is the very same dependence as that found by Windischmann [51]
for the formation of intrinsic compressive stress in thin films deposited under ion
bombardment.
Therefore, we conclude that the ion impact and stress models approaching the
subject from different directions using different ‘pictures’, however, have a great
deal in common. The ion impact during film growth causes subplantation [52] of
particles into the subsurface region of the film thus modifying its physical state:
The film gets more defect-rich, densified and compressively stressed. These are
all different characteristic of the same physical phenomenon. From that point
of view it is not surprising that critical values of the compressive stress or ptot/u
can be found which mark the onset of cBN growth. However. those critical
values represent ‘projections’ of the cBN formation mechanism onto the p l o , / a
and film stress scales. respectively. They d o not deliver information on the physical
mechanism of cBN formation itself.
R F diode sputtering:
- target diameter: 150 mm (hBN and B4C)
- target-substrate distance: 70 mm
1.0 1000
-*- intensity ratio
0.8 - - 800
n
4 -
+
3u 0.6 - 5?
v
1
3 0 - -1500 3r
0
3 0.4-
2
8
-400 2
0
.M
c
3
cn
8
U
0.2-
-
- 200
E
Y
Y
0.0 - *-*'
l ~ l ' l ' l ~ l ' l I ~' l0 ' ~ '
0 -50 -100 -150 -200 -250 -300 -350 -400 -450
Substrate bias voltage [V]
Figure 7. cBN content together with the deposition rate of BN films deposited at different substrate
bias voltages by d.c. magnetron sputtering of a boron target.
and 0.2Pa. The sputter gas was argon plus 10% nitrogen and polished silicon
wafers have been used as substrates. For d.c. sputtering of the boron target a
power of 200 W was utilized.
Varying the substrate bias voltage leads to the evolution of BN layers containing
various phase compositions (Fig. 7): From noncubic modifications obtained with-
out external substrate bias and at UBs = -lOOV to films with the highest cBN
content at UBs = -350 V. A further increase of the substrate bias to UBs = -400 V
causes growth conditions resulting in BN films with a lower cBN/hBN ratio. The
growth rate decreases if - UBs is increased mainly due to resputtering. In contrast
to that, for R F magnetron sputtering with either an hBN or a boron target
performed in the same vacuum chamber maximum cBN content was obtained at
UBs = -15OV [25]. However, it has been revealed by Langmuir probe measure-
ments [61] that this reduction of substrate bias voltage for optimum cBN formation
is connected with an about twice as high plasma density in front the substrate.
Therefore, qualitatively the decreased plasma density has to be compensated by a
higher ion energy.
A quantitative evaluation of the three magnetron sputter processes with respect to
the film-forming particle fluxes (for the analysis procedure see section 5.5.2) revealed
that within the error of measurement the d.c. process with boron target had the same
criticalptotlabetween 180 and 265 (eV amu)'I2 as the two R F processes [62]. This is
in good agreement with the values found for pulsed laser deposition by Mirkarimi
et al. [17].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.5 Discrimination between Nucleation and Growth Phase 433
80
b
A1
20
0
Deposition time [a. u.]
0.o001 0.001 0.01 0.1 1
AUI
Figure 8. (a) Scheme of the observed substrate current increase during BN film deposition where f
means the current during hBN growth, (b) cBN content against relative current increase A f/f, the
parameters of the deposition experiments were: R F powered boron carbide target (750W), 10%
nitrogen in Ar/N2 and UBs = -200 V.
increased (typical values of AZ/Isub z 0.1) the measured ion current Z,stayed nearly
constant. Furthermore a clear correlation between AI and the cBN content in the
films was found (Fig. 8b). With increasing AZ the cBN content in the films increased,
too. An additional hint to higher secondary electron yields of cBN compared to
hBN was the brightness difference in scanning electron microscope images taken
in the secondary electron mode under identical conditions for both samples [55].
The method described allows control of the cBN deposition process. For example
in the cBN nucleation range (left part of interval B in Fig. 8a) the deposition
parameters can be modified to ‘softer’ conditions. One possibility is to change the
sputter gas composition from high Ar contents, which are related to high stress
and cause cBN nucleation [48], to pure nitrogen. Figure 9 shows a SIMS profile
of a film grown under changed conditions. The deposition was started under typical
‘cBN conditions’ with 10% N2 in the sputter gas and after nucleation the gas
composition was changed to 100% NZ.The diagram reveals that the higher nitrogen
’
r_l
60
50
s40
Y
+g
,g* 30
20
10
u 0
0.00 0.05 0.10 0.15 0.20
Depth [PI
Figure 9. SIMS depth profile of boron, nitrogen, and carbon in a cBN film sputter deposited under
different conditions on a silicon substrate.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.5 Discriniination between Nucleation and Growth Phase 435
Table 2. Deposition conditions for separate investigations of nucleation and growth phase
1 1000 3 -15 - -
formed after a deposition time of 1 min. In step 4 the deposition was continued in a
pure nitrogen atmosphere. The whole multistep process has been repeated several
times and the substrate bias voltage during step 4 was varied in each case within
the range from -2OOV and the floating potential of the substrate holder (about
+10 V). The total deposition time was chosen so that all films got comparable thick-
ness values of 100-150nm.
The film structure has been characterized by infrared absorption, electron diffrac-
tion, and cross-section transmission electron microscopy. It was found that by
changing the gas atmosphere to pure nitrogen during step 4, but maintaining the
substrate bias voltage at -150V, the IR peak intensity ratio Zc/(Zc + Zh) was
about 0.84.9, which is the same value as in the case of cBN growth under optimum
conditions in an argon-nitrogen mixture (97% Ar/3 YON2). By transmission elec-
tron microscopy and electron diffraction it was confirmed that the BN films were
composed of small (10-50nm) grains of the cubic phase. The film composition
was measured using elastic recoil detection analysis (ERDA) and it was found
that the films were (within the error of measurement) stoichiometric BN
(B/N = 0.95 f 0.05). Practically 100% cubic film structure was obtained within a
wide range of substrate bias voltage during step 4, ranging from -200 V to values
as low as -6OV (Fig. 10). The deposition rate was about 200nmh-' for
UBs = -200 V and increased gradually with decreasing bias voltage due to reduced
resputtering. The increase of the deposition rate between -40 and -60 V (Fig. lo),
however, is mainly due to the different mass density of the cubic and hexagonal BN
phase. From that it was concluded that for voltage values below - 100 V, resputter-
ing could be neglected and the rate of mass deposition was practically constant.
0.8 - - 0.8
n
cf
+ 0
u* 0.6 - -0.6 3
--.
W
U"
3.
v
.-0 3
E 0.4 - -0.4 8
."0
8 0.2 - I
: E-0.2
U
-*- intensity ratio
0.0 - a 0 0 - 0.0
I I I I I I I I
0 -40 -80 -120 -160 -200
Substrate bias voltage [V]
Figure 10. Cubic BN phase content together with the deposition rate of BN films deposited at
different substrate bias voltages during step 4.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.5 Discrimination between Nucleation m d Growth Phase 437
Wavenumber [cm']
Figure 11. FT-IR transmission spectra demonstrating the phase inhomogeneity of BN films depos-
ited at U,, = -60 V. The IR beam with a diameter of about 1 mm was moved in 1 mm steps radially
from the outer region towards the center of the substrate (spectra a-g). Within the lateral resolution
limit of the method a sharp transition from noncubic (outer part) to cubic BN (inner part) is found.
For UBs= -60V and -8OV, only a circular central part of the film (having a
diameter of about 3cm) was cubic BN. Outside this central part, the film was
purely sp2 bonded (hBN). Considering the lateral resolution of the method, the
boundary between the cBN and hBN region was very sharp. This can be seen
from Fig. 11: the infrared spectra shown in this figure were measured in distances
of 1 mm using a beam diameter of 1 mm. Within a distance of only about 3 mm
the absorption features change significantly between the typical pictures of the
cubic and hexagonal BN phase.
The charge carrier density in the magnetron plasma was measured by Langmuir
probe measurements [61]. As an example, in Fig. 12 the distribution of the charge
carrier density is presented which had been measured lOmm in front of the sub-
strate. Because the charge carrier density at the plasma sheath is directly related
to the ion current towards the substrate (see below), the inhomogeneity visible in
Fig. 12 will be reproduced in the ion flux towards the substrate, that is, the ion
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
438 5 Vapor Phase Deposition qf Cubic Boron Nitride Filnis
flux will be about 15% higher in the center of the substrate holder than at r = 2cm.
This is obviously a relevant fact for the BN phase formation and will be discussed
later.
In order to describe the growth conditions in a quantitative manner the particle
fluxes towards the substrate have been estimated. The total flux of incorporated
atoms (both boron and nitrogen), Fa, was estimated from film stoichiometry,
growth rate, and mass density. The flux of ions cannot be directly measured for
several reasons, above all because of the high resistivity of the film which necessitates
an RF substrate bias. Therefore an estimation based on the Bohm sheath criterion
was made [25]. The Bohm criterion [67] yields an estimation of the ion flux, F,,
towards the substrate when ion density, ni, electron temperature, T,, and ion
mass, mi, are known. These quantities, however, can be determined by Langmuir
probe measurements (ni, T,) or estimated from gas composition (mi). Finally, the
kinetic ion energy, Ei, was roughly estimated (neglecting the plasma potential) to
be Ei = elUBsl.
Using the values of Fa, Fi, and Ei, the total incorporated momentum per
deposited atom, ptot/a, for the experimental series was calculated (Fig. 13). Under
optimum conditions for cBN growth with the given method used during step 3
(97% Ar/3% N2, Uss = -150V) ptot/a is about 200(eVamu)'/2, which is in good
P '
agreement with Mirkarimis results. Changing to ure nitrogen as the sputtering
gas, ptot/a reduces to approximately 150 (eV amu)' *. With an additional reduction
of UBs it was finally found that the continuation of cBN growth withptot/avalues as
low as about 50 (eV amu)'/2 was possible. These results lead to the conclusion that
the continuation of cBN growth after a certain initial cBN layer has been formed is
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.5 Discrimination between Nucleation and Growth Phase 439
800
H
h-BN C-BN
N
-
-
n
600 -
s
*2
2cd
v
u
---
400-
. .
ArM2=9713
a” 200 -
0
. I
.
I
. a
1
H
I I I
governed by an epitaxial mechanism. It has been found that the upper few mono-
layers of a growing cBN film are sp2 bonded [68,69]. The film forming particles
penetrate this layer and reach somewhat deeper regions of the film where they finally
become incorporated into the already existing cBN. The certain minimum energy
and/or momentum of the ions which was found to be still necessary to maintain
the cBN growth is obviously due to the fact that the film-forming particles have
to penetrate the sp2 bonded layer.
In the meantime, Litvinov and Clarke [70] have confirmed the results of Hahn
et al. [64,65]. They deposited boron nitride films in an UHV chamber by ion
beam sputtering of a boron target together with a nitrogen ion bombardment of
the growing film. The substrate temperature was 1000°C. The authors found that
after cBN nucleation at UBs = -96 V the substrate bias voltage can be significantly
reduced without interruption of the cBN growth. At UBs = -56V the crystalline
quality of the layers is even better that at -96 V. The reduction of the bias voltage
below approximately -50 V, however, leads to the formation of predominantly
hexagonal phase material.
Interestingly, a similar nucleation effect as found for the ion impact seems to exist
in the case of the substrate temperature. As has been found by several authors
[34,36] a minimum substrate temperature of 150°C is necessary in order to nucleate
cubic BN in ion-assisted deposition. However, very recently Feldermann et al. [71]
have found for mass selected ion beam deposition that switching off the substrate
heating after cBN nucleation and allowing the substrate to cool down to room
temperature, does not interrupt cBN growth.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
440 5 Vapor Phase Deposition of Cubic Boron Nitride Films
For films sputter deposited from a boron carbide target, k markedly increased at
X < 400nm [55]. This indicates an effect of the incorporated carbon (5-10 at-%,
see section 5.5.1) in the films. The cathodoluminescence behavior of hBN and
cBN films was investigated by Taylor et al. [80] with respect to possible applications
of the large band gap material BN in the deep blue and UV range.
With the Faraday cup arrangement in the substrate holder used for the experiments
on cBN nucleation which were discussed in section 5.5.1 additionally secondary elec-
tron yields, yi, both of hBN and cBN films could be determined. Here yi is the yield for
ion-induced electron emission. For both phases in the ion energy range between 100
and 600 eV for an argon-nitrogen gas mixture, "ii near to 1 were estimated. According
to the observed substrate current increase (section 5.5.1) the values of cBN were
calculated to be about 20% higher than those of hBN [86].
The plasma etching behavior of BN films is interesting, for example with respect
to deposition process modifications. However, there are still many open questions
concerning the effect of additional gas components in the process gases of cBN
deposition processes. For ion etching in pure Ar the ratio of the sputter yields of
hBN and cBN was revealed to be YhBN/YcBN5 1 [55,64]. On the other hand,
Harris et al. [87] observed, for an etching gas consisting of 1% CH4 and 99% HZ,
a considerably higher etch rate for the hBN than for the cBN phase.
The surface tension of sputter deposited cBN films was determined by contact
angle measurements with different well known liquids [88]. The value of
40 mN m-l is nearly the same as that of DLC (diamond like carbon) [89].
References
1. J. H. Edgar, in Properties ojGroup III Nitrides, J. H. Edgar (Ed.), INSPEC, the Institution of
Electrical Engineers, London, 1994, pp. 3-21,
2. A. Inspektor, C. E. Bauer, and E. J. Oles, Surf. Coat. Technol. 1994,68/69, 359-368.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 443
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
6.1 Introduction
Polymer pyrolysis to form advanced ceramics allows the production of highly
covalent refractory components (fibers, films, membranes, foams, joints, monolithic
bodies, ceramic matrix composites) that are difficult to fabricate via the traditional
powder processing route [14]. Yajima was the first to demonstrate the feasibility of
producing high-strength SiC-based fibers from pyrolysis of polycarbosilane [5]. In
this process, a thermoplastic pre-ceramic polymer is first shaped into the desired
form, cross-linked into a pre-ceramic network and finally converted into a ceramic
material by a pyrolysis process in a controlled atmosphere (Fig. 1). A common
feature of the polymer route is the formation of intermediates called amorphous
covalent ceramics (ACC) [6]. These are formed after removal of the organic
components and before crystallization that occurs at higher temperatures.
Many experimental techniques have been applied to characterize the polymer-to-
ceramic transformations. They range from differential thermal analysis/thermo-
gravimetrical analysis/mass spectometry (DTA/TGA/MS) [7,8] to spectroscopies
which are especially helpful for the low temperature range when the ceramic residue
is still amorphous such as: magic angle spinning nuclear magnetic resonance
(MAS NMR) [9,10], XPS [11, 121, and X-ray absorption near edge structure-
extended X-ray absorption fine structure (XANES-EXAFS) [I 3,141, to X-ray
diffraction (XRD) and transmission electron microscopy (TEM) to characterize
the crystallization process [15,16]. Among the various systems already synthesized,
S i c obtained from polycarbosilane is certainly the most widely studied and the
proposed structure of the amorphous silicon carbide phase obtained after pyrolysis
in Ar at 800°C is shown in Fig. 2 [17].
One of the most attractive and challenging features of this process is the possibi-
lity of designing the composition, the microstructure, and the properties of the final
ceramic material by controlling each processing step. The processing steps are:
- the polymer composition and/or its architecture,
~ the cross-linking step,
- the pyrolytic conversion process, and
- the final high temperature annealing to induce densification and/or crystal-
lization.
Indeed, many Si-containing ceramic systems have been already prepared from poly-
mers of different composition such as: polysiloxanes (-R, R2Si-O-), to give SiOC
[ 18,191, polysilanes (-R1 R2-Si-), [20,2 I], and polycarbosilanes (-R2Si-CH2-),
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.1 Introduction 447
f Pb'O'ecular
recursor I
1
Inert or
Reactive
Atmosphere
Fibers
A m o r p hou s
Films Covalent
Foams C e ra m ic
Joints
lnfiltralion
for CMC
N a n o - c' r y s t a I Iin e
Bulks Ceramic
Figure 1. Flow chart for the preparation of ceramic components viu polymer pyrolysis.
[22-241 to give S i c and polysilazane (-R2Si-NH-), to give SiCN [25] and SiBCN
ceramics [26].
Moreover, the various polymeric precursors can be easily modified by chemical
reactions to obtain new systems. A good example is the preparation of polytitanocar-
bosilane (PTC) from polycarbosilane and titanium alkoxide [27]. This precursor
transforms into SiC/TiC during pyrolysis. The fibers produced from this modified
precursors seem to exhibit better mechanical properties than those of S i c [28]. The
same reaction used to produce PTC gives the opportunity of modifying the polycar-
bosilane with different elements just by changing the nature of the metal alkoxide and
accordingly a large variety of new metal organic polymers, precursors for S-M-C4
(M = Al, Zr, Ta) systems have been obtained [29-311. Another approach to control
the composition and the microstructure of the final product is to use ceramic or metal-
lic powders as filling agents or reagents in the starting polymers [32,33]. The selectivity
toward defined ceramic phases can also be controlled by varying the polymer archi-
tecture. In this case the pyrolysis in inert atmosphere of-[MeHSiNH],- leads to SiCN
ceramics at temperatures >8OO0C whereas pyrolysis of isostructural -[H2SiNMeIx-
gives Si3N4and carbon [34] . The precursor molecular structure can also lead to dif-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
448 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
, Si, I -
\
. I -Si-C, I ’S .i .
CH+~, / “i:
si-Cqc/
’ \ I
Figure 2. Proposed structure of the amorphous silicon carbide phase obtained from polycarbo-
silane.
ferent microstructures of the cross-linked pre-ceramic networks that in turn have a sig-
nificant effect on both the pyrolysis chemistry and the composition and microstructure
of the ceramic products. Indeed, a study performed on polycarbosilane/siloxane pre-
cursors having similar chemical composition but different molecular structures
pointed out the formation of different preceramic microstructures that survive up
to high temperatures (1000°C) leading to SiOC glasses of different microporosity
and thermal stability [35].
The curing step can play a vital role in defining the structure and mechanical
properties of polymer derived ceramics. Indeed, the tensile strength of Sic-based
Nicalon fibers drops quickly at high temperatures due to the oxygen introduced
during the thermal cross-linking stage. It is believed that the oxygen gives rise,
upon pyrolysis, to a SiOC phase that decomposes above 1300°C with evolution of
SiO and CO [36]. This problem has been overcome by using a radiation (electron
beam or y-ray irradiation) curing process in helium gas which reduces the oxygen
content down to 0.4weight % and improves the fiber resistance up to 1750°C [37].
A different approach to controlling the composition and structure of the
ceramic products is through the control of the transformation process itself. The
heating rate and the maximum pyrolysis temperature can have a strong influence
on the ceramic yield and the ceramic composition [38]. In Fig. 3 thermogravimetry
curves obtained on a commercial polycarbosilane at different heating rates are
reported. It can be noticed that the ceramic yield is strongly related to the
pyrolysis rate, and it increases by using low heating rates. The pyrolysis atmosphere
can be varied from inert (Ar, He, N2) to reactive (NH,, 0,)and accordingly silicon
nitride and carbonitride-based materials, such as Si-N, Si-C-N, Si-0-N, and Si-
M-0-N (M = Al, Zr, Ti) have been produced from various polysiloxanes, poly-
carbosilane and polysilanes precursors [3942]. More recently synthetic silicates
have been fabricated by thermooxidation of polysiloxanes filled with different
oxide or metallic powders [43].
As far as the final heat treatment step is concerned, it is well known that the
maximum pyrolysis temperature governs the densification of the amorphous
ceramic phase and its crystallization process [44]. The densification of polycarbosi-
lane-derived amorphous silicon carbide was studied [ 17,451 in the temperature
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.1 Introduction 449
-1 0
-$
u1
- -20
Em
.-
0)
-30
range 900-1000°C (Fig. 4). An activation energy of 82 kcal mole-' was obtained
from the initial slope of the densification curves (Fig. 5). The observed increase of
density during the isothermal firing treatments has been primarily ascribed to reduc-
tion of the free volume of the amorphous phase. Two main processes may account
for this effect: the progress of the condensation reactions between residual CH
groups in the structure with the elimination of H2 and CH4 and the formation of
new Si-C-Si bridges with a consequent increase of the cross-linking of the network;
and a rearrangement of the open amorphous covalent structure toward a more com-
pact configuration with no change in chemical composition. For both of these
0 2 4 6 8 10 12
Time (h)
Figure 4. Evolution of the density, p, of polycarbosilane-derived amorphous silicon carbide with the
firing temperature. 900°C; +
1000°C; 950°C.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
450 6 Polymer to Ceramic Trunsftirrnution: Processing of' Ceramic Bodies and Thin Films
IO~IRT
Figure 5. Arrhenius plot for the densification rate of the polycarbosilane-derived amorphous silicon
carbide.
have been used as active binders in the traditional powders processing [61-631; and
polymer-derived amorphous ceramic powders have been sintered to give dense,
monolithic ceramics [64-671. However, in all these hybrid processes, the potential
advantages of the polymer route are lost, at least to some extent.
Recently, Riedel and coworkers [56,68-721, demonstrated the possibility of
producing a dense, crack-free, monolithic ceramic in the Si-C-N system by direct
pyrolysis of a polymeric green compact. In this method, a pre-ceramic polymer is
first converted into a pre-ceramic network with a controlled cross-linking process.
The powders obtained are then sieved and pressed to get a polymeric pre-ceramic
component, that can be pyrolyzed without cracking, into the final dense ceramic
body. These innovative ceramic components display negligible creep rates up to
very high temperatures [73,74]. The key point of this process is the partial cross-
linking of the polymer precursor that leads to an infusible network preventing the
component from melting and bloating during the first stages of the pyrolysis pro-
cess. Moreover, the open porosity of the green body guarantees the efficient removal
of the gaseous products during pyrolysis preventing the development of cracks.
Thus, the significant volume shrinkage usually associated with the polymer-to-
ceramic transformation no longer prevents the fabrication of bulk, crack-free
ceramic components. However, in order to optimize the microstructure of the cera-
mic product, the influence of many polymer properties and processing parameters
still must be assessed. As an example, the cross-linking degree and the weight loss
during pyrolysis of the polymeric precursor, the powder size and powder size
distribution, the amount of porosity and the pore size distribution of the green
compact, the pyrolysis rate and atmosphere are just some of the parameters that
could strongly influence the microstructure and the related physical, chemical,
and mechanical properties of the final ceramic product.
Here we report on the fabrication and characterization of dense ceramic materials in
the SiAlOC system from pre-ceramic green bodies obtained by pressing partially cross-
linked polyaluminocarbosilane (PAlC) powders with different aluminum loads. The
synthesis of PAlC is carried out reacting a commercial polycarbosilane (PCS) with
an aluminum alkoxide, following a published procedure [27]. PAlC has been already
investigated in our laboratory both with the aim of getting insights into the chemical
reaction between the polycarbosilane chains and the aluminum alkoxide and to clarify
the structure of the resulting amorphous Sic-based ceramic [29,40,47].
The objective of the present work is to investigate the role of some polymer prop-
erties (cross-linking degree, weight loss during pyrolysis, etc.) on the microstructure
and mechanical properties of the ceramic materials obtained by pyrolysis in inert
atmosphere at 1000°C. Indeed, the forementioned parameters can be easily modified
by varying the amount of aluminum alkoxide reacted with the polycarbosilane.
Accordingly, several PAlC compositions have been prepared with a nominal Al/
Si ratio between 0 and 0.25. The corresponding ceramic materials have been
characterized mainly by measuring their density, porosity and mechanical proper-
ties through four point bending tests both at room temperature and at high tempera-
ture up to 14OO0C,in air. The results will be discussed with the aim of disclosing the
key parameters that allow the processing of high strength and high modulus ceramic
components.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
452 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
Finally, we will show that this innovative processing route can be extended to the
production of bulk ceramic components by pyrolysis in reactive atmosphere, such as
for example, production of SiAlON ceramics by nitridation of PAlC in ammonia
flow [75].
Figure 6. Green and corresponding ceramic components obtained from PAlC with AljSi = 0.05
ratio.
measuring their open porosity (P)by mercury infiltration (Carlo Erba equipment)
and their bulk density, Pb. The structure of the partially cross-linked polymers
was investigated by recording 29Si CP MAS NMR spectra on a MSL 400 spec-
trometer operating at 79.6MHz. Details of the experimental set up are reported
elsewhere [47].
Monolithic, dense and crack-free SiAlOC ceramic components were obtained (see
Fig. 6) after a pyrolysis process (1 "C min-' , Ar flow, 100 cm3min-') at 1000°C for
1 h. The weight and the length of the samples were recorded before and after the heat
treatment in order to measure the associated weight loss and the linear shrinkage.
The microstructure of the ceramic samples was investigated measuring the open
porosity, P,and the bulk density, &,.
Modulus of rupture, MOR. was obtained on polished samples by four point bend
tests (4 x 20 mm) with 1 mm min-' deformation rate using Instron equipment. The
elastic modulus, E , was measured at room temperature, by the acoustic resonance
method using a self-assembled apparatus and, at high temperature, from the
stress-strain curve obtained in the four point bend tests.
6.4 Results
6.4.1 Characterization of the Pre-ceramic Precursors
The cross-linking mechanism of polycarbosilane with various metallic alkoxides has
been already published [45]. The Si-H bonds of the PCS react with the M-OR
groups (M=Ti, Al, Zr, Ta) of the metallic alkoxide to give Si-0 bonds that
probably form siloxane bridges between the polycarbosilane chains. The 29Si CP
MAS NMR spectra recorded on the starting PCS and on two PAlCs with an AljSi
ratio of 0.25 and 1 are presented in Fig. 7. The spectra show two peaks at 0p.p.m.
(Sic4 sites) and at - 17.6p.p.m. (SiC3H sites), characteristic of the polycarbosilane
backbone and a third peak, present only in the Al-containing compositions, at
10 p.p.m. that is assigned to SiC30 sites. By increasing the A1 content, a decrease of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
454 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
AVSi = 0.25
50 0 -50 -100
Chemical Shift (ppm)
Figure 7. 29Si CP MAS NMR spectra of the partially cross-linked pre-ceramic polymers.
18
A
8
W
s
14
c,
.d
v1
g 10
0
pc
6 t. 1 .. I . . I .
0.3
# . . . I . . I
0 0.1 0.2
Al/Si (Atomic ratio)
Figure 8. Evolution of the porosity of the pre-ceramic samples as a function of the AljSi ratio.
the highest ones for the samples with Al/Si=0.25 (1% 5 S 5 1.5%). The S value
measured for the green compact derived from pure PCS is 1.3%, close to the
values measured on PAlCs with the maximum aluminum load. These results suggest
that, by increasing the amount of aluminum alkoxide used for the preparation of the
modified polycarbosilanes, the cross-linking degree advances, in agreement with the
NMR results, and correspondingly increases the elastic recovery of the pre-ceramic
components. The high S value measured for the PCS green bodies indicates that the
thermal cross-linking process used in this case leads to a polymer with a high cross-
linking decree which exhibits a highly elastic behavior during pressing.
0 0.1
Al/Si (Atomic ratio)
0.2
i
1
2
3
4
T
0.3
Figure 9. Evolution of the elastic recovery of the pre-ceramic samples as a function of the AljSi
ratio.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
456 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
.cd
200
5
W
d 100
0
E
0 0.1 0.2 0.3
Al/Si (Atomic ratio)
Figure 10. Evolution of the flexural strength, modulus of rupture (MOR), of the ceramic samples
pyrolized at 1000"C, 1h, with the AljSi ratio.
independent of the A1 content. Values of density, porosity, flexural strength, and elastic
modulus measured on the various ceramic components are reported in Table 1.
The open porosity of the PAlCs-derived ceramic samples shows a weak depen-
dence from the amount of aluminum and falls in the range ~ 8 - 1 3 % .It increases
up to ~ 2 0 %for the PCS-derived component. Such low values of open porosity
have been already reported in the literature for similar polymer-derived ceramics
[56] and indicate that this processing route is suitable for the fabrication of dense
(up to 92% of theoretical density) covalent ceramics at low temperature. The
porosity values measured for the ceramic products are comparable with the starting
pre-ceramic samples. Indeed, results of the mercury infiltration experiments indicate
that the pyrolysis process leads to the removal of the finer pore fraction (below
10 nm) leaving almost unchanged the amount and the size of the coarser pores.
The evolution of the MOR and E with the Al/Si ratio is reported in Figs 10 and
11. Surprisingly, contrary to the behavior of all the other studied parameters, these
Table 1. Results of the microstructural and room temperature mechanical characterization per-
formed on the ceramic samples pyrolyzed at 1000°C for 1 h.
160
9 batch 2
2 120 A batch 3
k
W
u
W 80
40
0 0.1 0.2 0.3
Al/Si (Atomic ratio)
Figure 11. Evolution of the elastic modulus, E , of the ceramic samples pyrolized at lOOO"C, 1 h, with
the Al/Si ratio.
properties show a strong dependence from the composition of the starting polymers
with up to 100% of variation. The highest values of MOR (up to ~ 2 5 MPa)
0 and of
E (up to ~ 1 4 0 G P ahave
) been recorded for the samples with AljSi ratio of 0.05,
then the increase of the A1 load leads to a decrease of both mechanical parameters.
PCS-derived samples display very low values of MOR and E similar to the samples
with the highest AljSi ratio.
200 1 ' ' ' ' 1 ' ' ' ' ' " 1 ' 1 ' ' ' ' ' ' ' ' ' ' ~ ' '
1 5 0-
-
n
Q i I T
0
100-
T +
* r
-
m -
W
50- -
0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1100°C. Above this temperature the slight decrease of E is probably related to the
viscous deformation of the amorphous silicon oxycarbide phase that shows a glass
transition temperature of ~ ~ 1 3 0 0 [76].
°C
6.5 Discussion
In this work, crack-free, dense (up to 92% theoretical density), amorphous SiAlOC
ceramic components with good mechanical properties have been obtained at low
temperature ( T = 1000°C) by direct pyrolysis of pre-ceramic green bodies.
Elastic modulus and flexural strength show a clear dependence from the A1
content and the maximum values were recorded for the composition low in Al.
These behaviors cannot be ascribed to a different weight loss or linear shrinkage
during pyrolysis due to the constancy of these parameter with the polymer com-
position. Thus, in order to explain this experimental data we tried to correlate the
elastic modulus with the amount of porosity measured by mercury infiltration.
The dependence of E on the porosity shows a high level of scatter (see Fig. 13).
As an example, ceramic components with a porosity level of about 10% display
elastic modulus values from z60GPa to =120GPa, with a 100% variation. These
findings suggest that other microstructure parameters, apart from the amount of
open porosity must play a crucial role in governing the Young modulus of the
studied polymer-derived ceramics. Recently some studies have been published
that take into account the role of the contact area on which mechanical stresses
are really acting between the particles [77]. According to these models, samples
with the same P but with different E values must have a different amount of contact
area between the ceramic grains.
In our samples, the interface between the grains (derived by the decomposition of
the initial pre-ceramic particles) in the ceramic component can be high only if the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Discussion 459
t
5 10 15 20 25
Porosity, P (YO)
Figure 13. Elastic modulus of the samples pyrolyzed at 1000°Cfor 1 h plotted against the porosity of
the ceramic components.
contact area between the pre-ceramic powders in the pressed green body is also high.
In fact, during the pyrolysis, the development of the ceramic bond between the indi-
vidual particles takes place with the formation of primary chemical bonds via con-
densation reactions such as:
-Si-CH3 + H-Si- +S - C H 2 - S i s + H2
ESi-CH, + H3C-Si- +=Si-CH2-Si- + SH4.
These reactions can occur only if the reacting groups are close enough for the
formation of -Si-CH2-SiE bridges, that is, if they belong to the interface between
two pre-ceramic grains. The extension of the contact surface between the partially
cross-linked polymeric grains should be high for the compositions that can be
easily plastically deformed under pressure, with a low cross-linking degree. In
order to test this hypothesis, the elastic modulus has been plotted against the elastic
recovery, that is a parameter directly related to the cross-linking degree of the pre-
ceramic polymers (see Fig. 14). In this case a very good relation between E and S can
be observed suggesting that, actually, the cross-linking degree is a vital parameter
for tailoring the mechanical properties of polymer derived ceramics. In particular,
the cross-linking degree must be high enough to prevent bloating but also low
0 0.5 1 1.5 2
Elastic Recovery, S (YO)
Figure 14. Elastic modulus of the samples pyrolyzed at 1000°C for 1 h plotted against the elastic
recovery of the green components.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
460 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
/I
0
- I \ I
-10 *
2.
z
(P
-20 -
-.'
Adensity
0
-E
fn
fn
/.
-30
-
-40
0 400 800 1200 1600
T ("C)
Figure 15. Processing of SiAlON ceramics by nitridation of PAlC precursors: evolution of the
density and weight loss with the firing temperature.
I -> N2
600
NH3 <-
N
.ol
E
Y
m
E!
m
.-
0
c0
al
n
a
T ("C)
Figure 16. Processing of SiAlON ceramics by nitridation of PAlC precursors: evolution of the
specific surface area with the firing temperature.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
462 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
0.25 4
.
0
-I
-I
-0.05
LIL" 1
-0.15 1 I
T ("C)
Figure 17. Processing of SiAlON ceramics by nitridation of PAlC precursors: differential thermal
analysis (DTA) and dilatometry of a monolithic amorphous sample obtained after pyrolysis in
ammonia at 1000°C.
Table 3. Measured values of hardness (H,), elastic modulus ( E ) and fracture toughness (KIc) for
amorphous (1300°C) and microcrystalline (1500°C) DSiAlON pellets.
6.9 Experimental
S i c and SiOC films were produced by spinning a solution containing the precursor
on (100) silicon waters or silica slides. Detailed experimental and analysis procedure
is reported elsewhere [59,60]. The pre-ceramic polymers used were: for S i c films,
polycarbosilane (PCS, Dow Corning), or allylhydridopolycarbosilane (HPCS,
Starfire Systems, Watervliet, NY). For SiOC films, methylhydroxylsiloxane or
phenylmethylhydroxylsiloxane (SR350 and SR355, respectively, General Electric
Silicone Products, Waterford, NY). In addition, SiOC films were produced also
from different alkoxides: methyltriethoxysilane or phenyltriethoxysilane (MTES
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
464 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
SR350
I , . . l . . .
the structure of the PCS films remained amorphous, possibly due to the excess
carbon inhibiting crystallization or growth of Sic nanocrystals.
Performing the pyrolysis of PCS films at 1000°Cin vacuum, a structure character-
istic of nanocrystalline silicon carbide was obtained, but the compositional data
show that some oxygen contamination was present. Increasing the processing
temperature to 1200°C results in the increase of crystallization of PSiC (see later).
The SR350 films heat treated at 1200°C in vacuum exhibits a composition and
structure that can best be described as an oxycarbide glass, that is an amorphous
silicate material in which some of the Si-0 bonds are replaced by Si-C bonds. In
fact, the FTIR spectrum shows the persistence of Si-C bonds after heating, even
if this finding is less evident when heating at temperatures 11000°C.
It is thus clear that the processing atmosphere strongly affects the resulting
composition and microstructure of both SiOC and S i c coatings, and that vacuum
is the only environment which allows the synthesis of both silicon carbide and silicon
oxycarbide thin films [51].
In Fig. 19 is shown the influence of the processing temperature on the com-
positional and microstructural development of PCS films, heat treated in vacuum
for 1h. The peak at about 2100cm-' (see Fig. 19a) is attributable to the Si-H stretch-
ing; the peaks at 1400cmp' are due to the CH3 asymmetric in Si-CH3 whereas the
peak at 1250cm-' is due to the Si-CH, symmetric deformation. The peaks at 1355
and 1020cm-' are respectively assigned to the CH2 deformation and wagging of
Si-CH,-Si, and the peaks at 820 and 790cm-' are respectively assigned to the Si-
CH3 rocking and S i x stretching. The spectra show that, for heat treatments above
600°C, the absorption bands that are associated with organic species are eliminated,
indicating the conversion to a ceramic material of the pre-ceramic polymer.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
466 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
Figure 19. (a) Fourier transform transmission spectra and (b) glancing angle X-ray diffraction
(XRD) patterns for PCS films heat treated for 1 h in vacuum at different temperatures.
Figure 20. (a) Fourier transform transmission spectra and (b) glancing angle X-ray diffraction
(XRD) patterns for PCS films heat treated for different times in vacuum at 1000°C.
at 1200°C yielding a Sic coating 200 nm thick. No major defects (cracks, pits) are
visible in the films.
The effect of heating time on PCS films, pyrolyzed at 1000°C in vacuum, is
reported in Fig. 20. We can see that the crystallinity of the S i c coatings clearly
increases with increasing the isothermal treatment time. Accordingly, the composi-
tional analysis revealed that for heat treatments shorter than 6 h (at 1000°C), the
films are carbon-rich, whereas for longer dwelling times they become silicon-rich,
with a significant reduction in hydrogen and a lesser reduction in carbon and
oxygen. At the same time, Raman spectroscopy shows a strong reduction of the
free turbostratic carbon phase [52]. In Fig. 21 is reported the variation of the
average dimension of the PSiC crystallites in S i c films from polycarbosilane, as
a function of temperature and dwelling time (at 1000°C). The grain size seems
to be enhanced with respect to PCS-derived Nicalon fibers and PCS pyrolysis
residues that have been treated at similar temperatures. This is most probably
due to differences in oxygen contamination among the different PCS-derived
materials, and/or to their different processing histories (especially the kind of
annealing atmosphere). Again, concurrent with the transformations happening
in the material during the isothermal treatment, the film thickness decreases as a
function of the dwelling time.
Figure 21. Average PSiC crystallite dimension for PCS-derlved thin films as a function of: (a) pro-
cessing temperature and (b) dwelling time (heating in vacuum at lO0OT).
During this travel, the ions transfer energy to the atoms, which constitute the
coating material, by nuclear collisions and electronic excitations. As a result, a
change in composition, structure and density occurs, which depends on the product
of the fluence (which is proportional to the number of irradiating ions) by the linear
density of energy deposited by each impinging ion. In particular, besides merely
promoting densification, ion irradiation has a much higher efficiency than UV or
electrons for breaking all bonds and creating three-dimensional cross-links in
polymers (organic or inorganic), and especially for radiolyzing hydrogen and
alkyl groups. One has to bear in mind that the number of irradiated ions amounts
to just a few monolayers, and thus the properties of the substrates are not strongly
affected by this processing method (Si is amorphized within the ion range).
In Fig. 22 are reported the FTIR spectra for SR350 and PCS films, submitted to
irradiation with different doses of 3 MeV Au ions, or annealed, or irradiated and
then annealed (1 h at 1000°C in vacuum). We can easily see that with increasing
ion dose the bands relative to the organic moieties (see previous discussion) show
a progressive smearing out, and are replaced by broad bands of conjugated bonds
when the deposited energy reaches values of ~ 4 eV.0 Similar behavior and threshold
of collapse into a three-dimensional network are generally found for organic poly-
mers, and in our case indicate the actual occurrence of the polymer-to-ceramic
transformation. With respect to spectra of annealed films, the bands of conjugated
bonds are significantly broader and shifted toward lower wavenumbers (of
=2Ocm-' for the main Si-C and Si-0-Si modes) in spectra of films converted by
irradiation. An oxygen sub-stoichiometry in silica films or a compressive state of
stress can induce such a shift. In our case, this is more probably due to a higher
degree of bond distortion because of the combination of free radicals at random,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.10 Results and Discussion 469
Irradiated
+ Annealed
Annealed
=
5x10''
5x10''
As deposited
Figure 22. Fourier transform spectra of SR350 and PCS films irradiated with different doses of
3 MeV Au ions, or annealed (1 h lOOO"C), or irradiated with 5 x lOI5 Au' at cm-2 and then annealed
(1 h lO0OT).
and to some atomic disorder in Au-irradiated films. Associated with the structural
transformation in the specimens, there is a compositional change, which depends on
the dose and the type of irradiating ion (see Table 4). At high doses, the irradiation
process allows to obtain ceramic films whose composition is very similar to that of
the starting precursor, apart from a large decrease in hydrogen content. No major
contamination (oxygen) is introduced, and carbon is retained in the structure,
whereas conventional annealing is shown to strongly alter the Si:C ratio (and to
introduce some oxygen in the PCS films) [58,59,60]. Under irradiation, the films
also are subjected, without cracking, to a compaction that can be as high as 65%.
After irradiation, all the films are X-ray amorphous, while PSiC nanocrystallites
develop in PCS films, even after they have been irradiated, during the annealing
treatment.
SiOC and Sic films obtained by ion irradiation contain some carbon clusters (free
C). Ion irradiation, besides provoking structural and compositional chances in the
films, affects the hybridization state of the carbon clusters, as can be seen in Fig. 23,
which reports some Raman spectra collected on irradiated, annealed, irradiated and
then annealed SR350 and PCS samples. The presence of a single Raman peak in the
case of both SR350 and PCS irradiated films, indicates that the carbon clusters
possess a high degree of sp3 hybridization, which is typical of diamond-like films.
The hybridization state of the carbon clusters becomes a little more graphite-like
in films irradiated and then annealed (according to the degree of splitting in two
components of the Raman carbon peak), and is absolutely graphitic (sp2 hybridiza-
tion, turbostratic carbon) in precipitates formed during annealing of as-deposited
films [58].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
470 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
PCS As-deposited 1.90 f 0.1 0.08 f 0.02 4.50 f 0.2 1.10 0.28
He 1 x 10l6 1.90 0.92 1.oo 2.14 17.1
c 1 x 10I6 1.85 0.06 0.60 2.15 29.0
Au 1 x 10" 1.75 0.06 0.62 2.01 19.4
Annealed 0.80 0.20 0.25 2.45 19.0
Irradiated Au + annealed 1.65 0.05 0.04 2.41 34.6
SR350 As-deposited 1.00 4~0.05 1.55 f 0.05 2.9 f 0.1 1.25 0.23
He 5 x lot6 1.oo 1.40 0.29 2.05 21.0
c 1 x 10'6 1.oo ISO 0.37 2.16 18.5
Au 1 x 10'' 1.oo 1S O 0.55 2.06 15.5
Annealed 0.60 1.28 0.52 2.08 7.0
Irradiated Au + annealed 0.98 1.45 0.02 2.40 18.5
The carbon clusters afford some interesting optical properties to the SiOC and
Sic films [79], but are also responsible for the high hardness measured on irradiated
films (see Table 4). This is at least two times larger than that of conventionally
annealed specimens, mainly because of the previously mentioned diamond-like
nature of the C precipitates in the amorphous SiOC or Sic ceramic matrix of
irradiated specimens. The effect depends on the type of irradiating ion, but also
on the nature of the side groups of the polymeric chain (CH3 vs. C6H5) [59,60].
The annealing of irradiated films does not seem to affect much their hardness,
I . . , I . . . I . . .
Figure 24. Profilometer trace and AFM image of a -250 pm-wide strip realized by ion irradiation
(1 x loL5500 keV C+ at cm-’) in a PCS film.
despite the partial change in hybridization state of the C clusters, probably because
this effect is compensated by the further hydrogen removal and compaction.
Another interesting consequence of the ion irradiation of polycarbosilanes and
polysiloxanes is that the resulting films are thermochemically very stable towards
high temperature annealing, as shown by compositional analysis (see Table 4).
Indeed, carbon atoms having already lost their hydrogen neighbors can no longer
evolve (as CO or C,H,) and leave Si atoms with dangling bonds able to react
with O2 or H 2 0 of the atmosphere, as instead happens during annealing of non-
irradiated films [79]. Thus, the postirradiation annealing treatment results mainly
in the nearly complete elimination of hydrogen from the ceramic films. Finally,
ion irradiation promotes a good adherence of the coatings to the substrate, due
to some interfacial mixing.
The capability of patterning the films using ion irradiation, which can be quite
important when fabricating actual devices, is illustrated in Figs 24 and 25. Figure 24
shows the profilometer trace and the AFM image of a strip realized in a PCS film
by ion irradiation (1 x lOI5 500 keVC+ at cm-*) using a 250 pm mask, while Fig. 25
Figure 25. Profilometer trace and AFM image of a -250 pm-wide strip realized by ion irradiation
(1 x 500 keV C+ at cm-2) in a PCS film, after dissolving the non-irradiated part of the film.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
472 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
shows the same specimen after the nonirradiated part of the film was removed by dis-
solving it in monochlorobenzene. It is evident the ~ 4 8 % compaction of the film due to
the irradiation, as well as the rather good lateral resolution of the strip profile. The high
edges present in the strip after dissolving the non-irradiated part of the film (Fig. 25) are
probably due to an irradiation-induced cross-linking of the PCS polymer at the inter-
face between the irradiated and non-irradiated areas, which makes it insoluble, but is
not accompanied by any compaction.
6.11 Conclusions
Polymer pyrolysis appears to be a very promising processing route to advanced
covalent ceramics. This process allows the formation of new amorphous covalent
ceramic phases in the general system Si-M-C-N-0 (with M = B, Ti, Al, Zr . . .)
which show exceptional oxidation and creep resistance at high temperature up to
1500°C.
Another advantage of this process is its high flexibility. Indeed, the composition,
structure (amorphous or crystalline), microstructure, and properties of the ceramic
material can be controlled and adjusted by varying many different parameters such
as: the composition and architecture of the pre-ceramic polymer, the amount and
nature of the filler, the cross-linking step, and the pyrolysis parameters (atmosphere,
heating rate, final temperature). Also, nonconventional heating systems such as
laser or microwave heating or even athermal conversion processes like ion bombard-
ment can be efficiently used.
Polymer pyrolysis is also very flexible regarding the type of component that can be
obtained: fibers, films, membranes, foams, ceramic/ceramic joints, ceramic matrix
composites, and monolithic bodies can all be fabricated. The present study has
been focused on the formation of monolithic component and thin ceramic films.
Dense Sic-based monolithic components with MOR values up to ~ 2 5 MPa 0
and E values up to ~ 1 4 0 G P ahave been fabricated from partially cross-linked
polyaluminocarbosilanes, PAlC, by pyrolysis at 1000°Cin a controlled atmosphere.
The cross-linking degree of the pre-ceramic polymers was controlled by varying the
amount of aluminum alkoxide reacted with the PCS. Among the many investigated
parameters (weight loss and linear shrinkage during pyrolysis, porosity of the pre-
ceramic and ceramic components) only the cross-linking degree was shown to really
play a key role in the development of ceramic materials with elevated mechanical
properties. This result has been related to the ability of the pre-ceramic powders
to be permanently deformed under pressure leading to an extended contact surface
both between the green particles and between the ceramic grains.
Defect-free Sic and SiOC ceramic thin films were obtained by heating in inert
atmosphere or by ion irradiation. The composition and microstructure of the
coatings can be varied by changing processing conditions (heating atmosphere,
maximum heating temperature, dwelling time, type and dose of irradiating ions).
The segregation of C clusters during ion irradiation affords interesting optical
(luminescence) and mechanical (high hardness) properties to the films.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 473
Acknowledgments
Gian Domenico Soraru would like to thank Prof. J. D. Mackenzie, who introduced
him to the field of polymer-derived ceramics, Dr. Florence Babonneau for the MAS
NMR study performed in the framework of a very effective, 10-year long collabora-
tion, and the contribution of his students and colleagues: Dr. Alberto Ravagni, Ing.
Massimo Mercadini, Ing. Fabiano Vulcan, Ing. Alberto Chiari, Ing. Marcello
Beber, Ing. Paolo Bosetti, Dr. Vincenzo Sglavo, Dr. Antonella Glisenti. MURST
is also acknowledged for the 40% financial support.
Paolo Colombo would like to thank Prof. C. G. Pantano for the constant support
and encouragement in pursuing research in the field of polymer-derived ceramics.
He also gratefully acknowledges the fundamental contribution of Dr. J. C. Pivin,
who performed the ion irradiation experiments as well as many characterizations
on the specimens.
References
1. J. Lipowitz, Cerum. Bull., 1991, 70, 1888-1894.
2. Y. D. Blum, R. M. Platz, and E. J. Crawford, J . Am. Cerurn. Soc. 1990,73, 170-172.
3. R. Riedel, Advanced ceramics from inorganic polymers, in Materials Science and Technology,
Vol. 176, Processing of Ceramics, R. J. Brook (Ed.), VCH, Weinheim, 1996, pp. 1-50.
4. J. Bill and F. Aldinger, Precursor-derived covalent ceramics, in Precursor-Derived Ceramics,
J. Bill, F. Wakai, and F. Aldinger (Eds.), Wiley-VCH, Weinheim, 1999, pp. 33-51.
5. S. Yajima, J. Hayashi, and M. Omori, Chem. Lett. 1975, 931-934.
6. G. D. Sorard, F. Babonneau, and J. D. Mackenzie, Structural concepts on new amorphous
covalent solids, J . Non-Cryst. Solids, 1988, 106, 25G261.
7. V. Belot, R. J. P. Corriu, D. Leclercq, P. H. Mutin, and A. Vioux, Thermal redistribution
reactions in crosslinked polysiloxanes, J . Polym. Sci.: Part A: Polym. Chem. 1992,30,613-623.
8. R. Campostrini, G. D’Andrea, G. Carturan, R. Ceccato, and G. D. Sorard, Pyrolysis study of
methyl-substituted Si-H containing gels as precursors for oxycarbide glasses, by combined ther-
mogravimetric, gas chromatographic and mass spectrometric analysis, J . Muter. Chem. 1996,6,
585-594.
9. F. Babonneau, J. Livage, G. D. Sorard, G. Carturan, and J. D. Mackenzie, SiC/TiC ceramics
via polymer route: a structural investigation. New J . Chem. 1990, 14, 539-544.
10. F. Babonneau, G. D. Sorard, and J. D. Mackenzie, 29Si MAS-NMR investigation of the
conversion process of a polytitanocarbosilane into SiC/TiC ceramics, J. Muter. Sci. 1990, 25,
36643670,
11. G. D. Sorard, A. Glisenti, G. Granozzi, F. Babonneau, and J. D. Mackenzie, The pyrolysis
process of a polytitanocarbosilane into SiC/TiC ceramics: an XPS study, J . Muter. Res.
1990, 5, 1958-1962.
12. G. Granozzi, A. Glisenti, and G. D. Sorard, XPS characterization of mixed carbides obtained
from polymer precursors, Muter. Rex Soc. Symp. Proc., 1990, 180, 81 1-814.
13. C. Laffon, A. M. Flank, P. Lagarde, M. Laridjani, R. Hagege, P. Orly, J. Cotteret, J. Dixmier,
J. L. Miquel, H. Hommel, and A. P. Legrand, Study of Nicalon-based ceramic fibres and
powders by EXAFS spectrometry, X-ray diffractometry and some additional methods,
J . Muter. Sci. 1989, 24, 1503-1512.
14. F. Babonneau and G. D. Sorard, Spectroscopic characterization of the nitridation process of
polymeric precursors to Si-M-N-0 systems (M = Ti, Zr, Al), Muter. Res. SOC.Symp. Proc.
1993, 287, 239-244.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
414 6 Polymer to Ceramic Transformation: Processing of Ceramic Bodies and Thin Films
39. K. Okamura, M. Sato, and Y. Hasegawa, Silicon nitride fibers and silicon oxynitride fibers by
the nitridation of polycarbosilane, Cerum. Int. 1987, 13, 55-61.
40. G . D. Soraru, A. Ravagni, R. Campostrini, and F. Babonneau, Synthesis and characterization
of p-sialon ceramics from organosilicon polymers, J . Am. Ceram. Soc. 1991, 74, 2220-2223.
41. G. D. Soraru, A. Ravagni, R. Dal Maschio, G. Carturan, and F. Babonneau, Polymer-derived
Si3N4-Zr02 nanocomposite powders, J. Muter. Res. 1992, 7, 1266-1270.
42. G. Granozzi, A. Glisenti, R. Bertoncello, and G. D. Soraru, XPS study of the nitridation
process of a polytitanocarbosilane into Si-Ti-N-0 ceramics, Muter. Res. Soc. Symp. Proc.
1992,271, 899-904.
43. D. Suttor, H.-J. Kleebe, and G. Ziegler, Formation of mullite from filled siloxanes, J . Am.
Ceram. Soc. 1997,80,2541-2548.
44. M. Monthioux and 0. Delverdier, Thermal behavior of (organosilicon) polymer-derived
ceramics. V: main facts and trends, J . Eur. Ceram. Soc. 1996, 16, 721-737.
45. G. D. Soraru, F. Babonneau, and J. D. Mackenzie, Mixed carbides via polymer route, Muter.
Res. Soc. Symp. Proc. 1990, 180, 815-818.
46. M. Narisawa and K. Okamura, Polymer pyrolysis, in Chemical Processing ofcerumics, B. I.
Lee and E. J. A. Pope (Eds), Marcel Dekker, New York, 1994, pp. 375-394.
47. G . D. Soraru, M. Mercadini, R. Dal Maschio, F. Taulelle, and F. Babonneau, Si-AI-0-N
fibers from polymer precursor: Synthesis, structural and mechanical characterization, J . Am.
Ceram. Soc. 1993, 76, 2595-2600.
48. L. V. Interrante, C. W. Whitmarsh, C-Y. Yang, W. Sherwood, W. Schmidt, P. S. Marchetti,
and G. E. Maciel, Processing of Si-based ceramics and ceramic composites using hydridopoly-
carbosilane (HPCS), in Silicon-BusedStructural Ceramics, Ceramic Transactions, Vol. 42, 1994,
pp. 57-69.
49. S. T. Gonczy, E. P. Butler, N. R. Khasgiwale, L. Tsakalakos, W. R. Cannon, and S. C.
Danforth, Blackglas-Nicalon composites with CVD boron nitride fiber interface coatings,
Ceram. Eng. Sci. Proc. 1995, 16, 433441.
50. P. Carri and G. D. Soraru, Sol-gel processing of continuous fiber reinforced composites by
the liquid infiltration and pyrolysis (LIP) method, in Innovative Processing and Synthesis of
Ceramics, Glasses and Composites, N. P. Bansal, K. V. Logan, and J. P. Singh (Eds), Ceramic
Transactions, Vol. 85, 1997, pp. 405416. The American Ceramic Society, Westerville, Ohio,
USA.
51. P. Colombo, T. E. Paulson, and C. G. Pantano, Atmosphere effects in the processing of silicon
carbide and silicon oxycarbide thin films and coatings, J . Sol-Gel. Sci. Technol. 1994,2,601-604.
52. P. Colombo, T. E. Paulson, and C. G. Pantano, Synthesis of S i c thin films with polycarbosilane
(PCS), J . Am. Ceram. Soc. 1997, 80, 2333-2340.
53. P. Colombo, G. A. Danko, and R. Silberglitt, Production of SIC and SiOC from preceramic
polymers: comparison of microwave and conventional heating, in Ceramics ~- Getting into the
2000s, Part C (Section F), Techna Srl, Faenza, 1998. pp. 353-360.
54. P. Colombo and M. Modesti, Silicon oxycarbide ceramic foams from a preceramic polymer,
J . Am. Ceram. Soc. 1999,82, 573-578.
55. P. Colombo, V. Sglavo, E. Pippel, and J. Woltersdorf, Joining of Reaction-bonded silicon
carbide using a preceramic polymer, J. Muter. Sci. 1998, 33, 2409-2416.
56. R. Riedel, G. Passing, H. Schonfelder, and R. J. Brook, Synthesis of Dense Silicon-based Cera-
mic at Low Temperatures, Nature, 1992, 355, 714717.
57. K. Jakubenas and H. L. Marcus, Silicon carbide laser pyrolysis of polycarbosilane, J . Am.
Ceram. Soc. 1995. 78, 2263-2266.
58. J. C. Pivin and P. Colombo, Conversion of inorganic-organic polymers to ceramics by ion
implantation, Nucl. Instr. Meth. Phys. Res. B, 1996, 120, 262-265.
59. J. C. Pivin and P. Colombo, Ceramic coatings by ion irradiation of polycarbosilanes and
polysiloxanes. Part I: Conversion mechanism, J . Muter. Sci. 1997, 32, 61634173.
60. J. C. Pivin and P. Colombo, Ceramic coatings by ion irradiation of polycarbosilanes and
polysiloxanes. Part 11: Hardness and thermochemical stability, J . Muter. Sci. 1997, 32, 6175-
6182.
61. S. Yajima, T. Shishido, and K. Okamura, SIC bodies sintered with three-dimensional cross-
linked polycarbosilane, Ceram. Bull. 1977, 56, 1060-1063.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
476 6 Polymer tn Ceramic Transformation: Processing of Ceramic Bodies and Thin Film&
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
1.1.1 Introduction
Tooling systems containing diamond and cubic boron nitride materials are used
today across the breadth of modern industry, which could not function effectively
without them. Their range of application covers petroleum exploration, mining,
stone and concrete sawing, the grinding of tungsten carbides, glass, and ceramics,
the machining of a wide range of engineering materials including aluminum
alloys, modern composite materials, hard ferrous materials, and also in such
unlikely fields as the cutting of frozen foods. Most applications take advantage of
the extremely high hardness and wear resistance of these so-called ‘ultrahard’ or
‘superabrasive’ materials. However, more recently, the outstanding thermal and
optical properties of diamond are being more widely exploited, giving rise to further
(and new) areas of application.
Figure 1. Diamond trade routes from the 1st century BC to the 3rd century AD.
Other developments were taking place during this period. A diamond tipped
lathe tool was used by J. Ramsden in 1773, the beginning of precision turning.
The use of diamond as a wire drawing die was patented in England in 1819. By
the end of the 19th century, large saw blades with diamond set around their
periphery were in use in France for cutting stone. The cutting of marble and lime-
stone used during the construction of large buildings in Paris in the early 1900s
utilized diamond-containing saw blades up to 2-3 m in diameter.
In 1842, the first mention was made of a diamond tool utilizing relatively small
pieces of diamond when Pritchard reported grinding and polishing microscope
lenses using a tool consisting of diamond grit hammered into the surface of a cast
iron tool. The first patent for a diamond wheel was granted to A. L. Caverdon in
France, in 1878, when he produced a wheel containing fine diamond particles
held onto a metal wheel form possibly by jamming them into indentations made
in the surface of the wheel. This method was described by Carl Zeiss in 1906.
Most modern diamond tools, including the majority of saw blades and grinding
wheels, consist of small particles of diamond held in a matrix or ‘bond’ which
wears down during use to expose new diamond particles, which carry out the
cutting. This is in contrast to the very early tools which used relatively large
single stones or slivers from single stones. A patent for such a bond system was
granted to the Western Electric Company in 1927. During the early 1930s, in
order to grind the then-new material, cemented tungsten carbide, which had
been invented by Krupp in Germany in 1928, grinding wheels containing diamond
held in a resin bond were developed and several patents were granted between 1932
and 1933.
was developed. Whole synthetic crystals or cut pieces of the order of several milli-
meters in size are used for these high precision industrial applications.
Figure 6. pcD machining a cast metal-matrix composite automotive brake motor in A1 20% Sic,.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystullizution qf Diamond 485
h
(3
W
cc
I3
v)
v)
W
cc
a
natural diamond genesis information is to be found elsewhere [7, 81. The conditions
under which natural diamonds are formed will, however, be touched upon when
the phase transformation diagram for carbon is discussed later in this section on
crystallization.
To gain a balanced general view of the many ways in which man has crystallized
diamond, the thermodynamically stable and metastable phases of elemental carbon
and reaction dynamics between them, over obtainable pressures and temperatures,
should be considered. The most up to date phase and transformation diagrams are
to be found in a review by Bundy et al. [9]. Figure 7 is an adaptation of the pressure-
temperature phase transition diagram taken from this review article, together with
further information gleaned from the extensive diamond- and carbon-related scien-
tific literature.
The dominant thermodynamically stable forms of carbon are graphite, diamond,
liquid, and vapor. The phase diagram presented by Bundy [9] and adapted in Fig. 7
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallization of‘ Diamond 487
has some new notable features over many of the earlier diagrams. The slope of the
diamond melting line is positive, and molten carbon is said to be metallic with little
evidence for transformations from electrically conducting to nonconducting forms.
For consideration of diamond crystallization, one of the most important features is
the diamond-graphite equilibrium line which stretches from 1.7GPa at OK to
12 GPa at 5000 K, the diamond-graphite-liquid carbon triple point. The activation
energies required to initiate transformations from one form of carbon to another are
very largely due to the high cohesive energy of carbon in its crystalline structures.
Davies and Evans [lo] showed that the activation energy for the graphitization of
the { 1 lo} surface of pure diamond in vacuo is 728 f 50 kJ mol-’, which is about
twice the bond energy of the C-C bond in aliphatic hydrocarbons (3.7eV per CJ
C-C bond). This is also about the same energy required for the vaporization of gra-
phite. These very high activation energies result in very large areas of pressure and
temperature over which a metastable form of carbon can subsist. Probably the
most important example of this is the indefinite existence of diamond at room tem-
perature and pressure even though, at these conditions, it is very ‘deep’within the gra-
phite stability field. To overcome the activation energy for conversion to graphite in
vacuo, diamond must be heated to a temperature-of about 2000 K before graphitiza-
tion becomes rapid [ 101. In addition to the allotropes of carbon, there are polytypes of
carbon, that is structural variations based upon lattice plane stacking difference.
Important examples are hexagonal diamond, which occurs in methods of very
rapid and hot direct transformation into diamond, and rhombohedra1 graphite, pro-
duced by cold working normal hexagonal graphite crystals. When the nanoscale
groupings of pure carbon atoms are considered, other solid forms with very distinctive
properties are known and are probably best described as ‘metastable’ over the whole
range of pressures and temperatures in Fig. 7. These are amorphous forms such as
carbon black and glassy carbon, condensate of very long linear carbon molecules col-
lectively called carbynes and, of course, the now very extensive variations of fullerenes
(C60, C70, etc.) and the derived nanotubes [ 1 I].
Importantly for diamond crystallization, all these forms of carbon, including gra-
phite, very stubbornly persist at pressures far into the diamond stable region. The
practical synthesis of diamond could thus be considered as the search for means
of finding practical convenient kinetic pathways to avoid the high activation barriers
between the multifarious nondiamond carbon structures and diamond. The two
very general approaches are to use solvents for carbon and to precipitate diamond
primarily in the diamond stable part of the phase diagram or to ‘quench’ carbonac-
eous species from very high energy states such as vapors, liquid, or plasmas. Some of
the more recently reported means could well be considered to be combinations of
these [12, 131. Some of the most important diamond crystallization techniques in
the recent literature will now be itemized using Fig. 7 as a guide.
2175 K at this pressure even after treatments lasting 8 h or more. These findings are
consistent with the current theories of natural diamond genesis.
The two regions both marked N in Fig. 7 are intended to illustrate the conditions
under which natural diamond formed in the earth’s crust, probably over geological
time scales (about 3.3 billion years). There are two environments linked to natural
diamond formation, the peridotitic (depicted by the lower pressure N region on
Fig. 7) and the eclogitic (at higher pressures and about 200 K hotter, also marked N).
The pressure given by the N positions are estimated from the accepted
correspondence of 0. 1 GPa pressure to 3 km depth into the earth crust. The lower
pressure limit of N is given by the boundary zone between the upper region litho-
sphere and the mobile asthenosphere at about 150 km in the earth’s mantle.
It is believed that natural diamonds, crystallized from C 0 2 or CH4 volatiles below
the lithosphere and after existing for great periods of time at these depths, are
brought to the surface by subsequent eruptions of magma. For more detailed discus-
sions of the geology, mineral inclusions in natural diamond and general matters
related to diamond genesis, the reader is referred to Harris [7], Kesson [8], Haggerty
[29], and Meyer [21].
The position marked B at 4 GPa and temperatures between 1775 and 2075 K,
gives the condition under which well faceted diamonds of about 200 pm diameter
have been crystallized using specially treated phenolic resins as the source of
carbon and molten cobalt as solvent [30]. The notable point of this particular crys-
tallization is that it occurred well into the graphite stable region of Fig. 7 and is thus
an example of the ‘metastable growth of diamond’. An explanation for this may be
found in consideration of the relative solubilities of phenolic resins and diamond in
molten cobalt allowing dissolution of one metastable form and precipitation of
another, namely diamond. This is an example where rules governing the transitions
between metastable states, such as the Ostwald and Ostwald-Volmer rules, can be
applied [3I].
The curved dashed line CI-Cz in Fig. 7 depicts the proposed boundary between
the formation of the hexagonal (below the line) or cubic (above the line) polytypes of
diamond, after explosive shock compression or fast heating, mainly using laser
heated diamond anvil cells, of graphite structures in the diamond stable region,
described by Bundy et al. [9]. This reference is a detailed review article and is a
major source of reference for work concerned with the direct conversion of graphite
and other carbon structures to diamond. The diamonds that can be retrieved from
these various methods, either of the hexagonal or cubic form, are always small and
crystallographically defective. This is because of the very short times at high tem-
perature and pressure, and the need to satisfy fast thermal quench requirements.
Despite this, diamond powders have been commercially produced by the Du Pont
Company of the USA and others using explosive shock of metal containing small
crystals of graphite as a carbon source [32].
The direct conversion of hydrocarbons to diamond has also been demonstrated.
C 3 is the region where organic compounds such as anthracene, camphene, fluorene,
pyrene, sucrose, polyethylene, adamantane, and paraffin wax have been converted
to very fine diamonds under high pressure (about lpm in size) (Wentorf [33]).
Aliphatic compounds seem to work but compounds with aromatic rings and/or
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
490 1 Diamond Materials and their Applications
large amounts of nitrogen only formed graphite of excellent crystallinity under high
pressure pyrolysis and did not form diamond.
There is a growing body of literature describing the crystallization of diamond
from fullerenes as a starting material. One of the latest [34] claims that c 6 0 can be
transformed to diamond at 5GPa and 1675K using a solvent/catalyst without
forming graphite first. The generation of ‘bulk’ polycrystalline diamond at room
temperature has been reported by the crushing of c 6 0 at about 20 GPa [35], (position
C4 in Fig. 7). The conditions used in the polycrystalline diamond (cobalt-matrix)
anvil cell were deliberately nonhydrostatic with respect to pressure. It was suggested,
from the high efficiency of conversion and fast kinetics involved, that, if cheap bulk
C60became available at some time in the future, a feasible method for the commer-
cial production of polycrystalline diamond might result.
The metastable crystallization of diamond in the appropriate pressure range of
0-0.5 GPa and temperature range 675-1300K has now been accomplished using a
‘galaxy’ of techniques in region D in Fig. 7. These include chemical vapor deposi-
tion from carbonaceous gases, mainly methane, hydrogen mixtures using many
varied means of generating plasma [36, 371 (see also Chapter 4 of Part 11), the
use of a combination of lasers and carbon dioxide as source (the so called QQC
deposition process) [13], hydrothermal synthesis in the C-H-0 and C-H-O-
halogen systems [38], low pressure solid source processes (the so called LPSSS
methods) [12], and many more. The literature in this general area of metastable
diamond growth has now become extensive, as discussed in Section 1.2.6.
From a commercial point of view, the production of diamond abrasives is, at this
time, dominated by the high-pressure, high-temperature, transition metal solvent/cat-
alyst techniques where the crystallization takes place in the diamond stable region.
The gasket allows an even distribution of stress in the die taper and along the
tapered anvil cone. In this way, both the die and anvil are supported by the resulting
compressive stresses.
The essential difference between the belt and girdle devices is the shape of the die
bore/taper and the matching anvils. For the belt apparatus, the die bore is curved
continuously from center to outer surface whereas, for the girdle, the die bore is
straight and proceeds out on both sides in a linear progression.
1700 , I I
1600
1300
1240
0 1.0 2.0 3.0 4.0 Wt.% C 99 100
4.7 9.0 13.15 16.7 At.% C
product types. The diamond industry has spent many years and significant amounts
of money developing and refining the large scale high pressure equipment appro-
priate to generate and maintain economically the required crystallization condi-
tions. The high pressure equipment used by the two largest manufacturing
companies, namely De Beers Industrial Diamond Division and the General Electric
Corporation are proprietary, highly evolved designs based upon the belt and/or
girdle apparatus described in Section 1.2.3.1. The Chinese producers mainly use
cubic multi-anvil high pressure designs, and the Russians and Ukrainians use
various modifications of the toroidal equipment.
Possibly the first question a crystal grower must ask when investigating or devel-
oping a crystal growth method is, ‘how is an appropriate driving force for nucleation
and growth generated and controlled’?
The central part of Fig. 8 is considered as representative of a general transition
metal/carbon phase diagram at a pertinent pressure (5.7 GPa) and temperature
range. The driving force for crystallization of diamond using ‘ideal’ graphite as
source carbon would be given by the difference in solubility between graphite and
diamond (the supersaturation) at the chosen conditions. For example, in Fig. 8 at
1450°C and 5.7 GPa this would be the solubility difference AB, (A’B’). Note that
this supersaturation for diamond crystallization (let it be Ac) would decrease as tem-
perature is increased and would essentially be zero at the diamond/graphite equili-
brium temperature, E(eq.)in Fig. 8. The slopes of the diamond/metal liquidus line
and graphite/metal liquidus line at constant temperature will change with pressure.
The difference between these solubility lines, Ac, will thus also change with pressure
and similarly become zero at the diamond/graphite equilibrium condition. Thus the
driving force for crystallization, Ac, is a function of the ‘distance’ from the diamond
graphite equilibrium line in pressure/temperature space, Fig. 9. The super-
saturation, Ac, can of course be related to the nucleation rate and growth rate of
the diamond in the classical way as shown in Fig. 10.
Note the extremely strong dependence of nucleation rate upon supersaturation
and that there is generally a critical level of supersaturation, Accrit,which must be
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallization of Diamond 493
DIAMOND STABLE
a
GRAPHITE STABLE
TEMPERATURE
Figure 10. The classical relationships of the rates of nucleation and growth to supersaturation, Ac,
for crystallization from a solution.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
494 1 Diamond Materials and their Applications
exceeded before the probability for nucleation becomes large enough for an obser-
vable nucleation rate. The growth rate is an increasing function of supersaturation.
The relative growth rates of different crystallographic growth zones is also a func-
tion of Ac, hence the morphology shift with pressure, temperature condition as
depicted in Fig. 9.
The relationships between solubility, supersaturation, pressure, and temperature
depend upon the details of the microscopic mechanisms operative in the context of
the detailed active impurity chemistry and the specific geometric relationship set up
between the source carbon, molten metal solution, and nucleation site, and or grow-
ing diamond. For an introduction into these extensive subjects the reader is referred
to Burns and Davies [16] where the issues such as nitrogen and boron active impur-
ity chemistries, diamond surface reconstruction, morphology and solvent/catalyst
are discussed.
An important conclusion to be drawn from the above fundamental explanation is
that the structure of the source carbon as it pertains to solubility in the chosen
molten metal is very important. Will and Graf [46] compare six graphite types
using an iron-nickel solvent and show a clear dependence of rate of diamond forma-
tion and diamond size upon the starting crystallinity of the graphites. Moreover
these authors include consideration of the influence of gaseous impurities in the gra-
phites and the cluster theories of Sunagawa [47] concerning the carbon species in
solution in the metal. As the source carbon became more crystalline, the pressure
needed for diamond nucleation increased by up to 0.5GPa. This could be inter-
preted as the need to increase the pressure in order to exceed the critical supersatura-
tion for spontaneous nucleation, Accrit,owing to the much lower solubility of the
recrystallized and thus more crystalline graphite.
Another striking example of the importance of the structure of source carbon
leading to crystallization consequences via the supersaturation driving force is
described in Wang et al. [48]. Here, the pressure needed for spontaneous nucleation
was shown to increase with time as the graphitic source material recrystallized in the
high pressure reaction chamber.
In addition to the explained (at least in general terms) dependence of crystalliza-
tion on the structure of the source carbon, the character of the transport of solute
from dissolving source to growing diamond crystals needs to be considered.
When a diamond synthesis reaction chamber from a process for making grits is
broken open after completion of the crystallization cycle, it is observed that each
grown crystal is separated by a metal alloy film from the undissolved carbon
source material [16]. The metal film, which of course was molten during the
growth of the diamonds, is usually about 50-200 pm in thickness. Figure 11 is a
schematic drawing of this growth geometry. The carbon source material is dissolved
at the molten metal interface, D,and the resultant carbon species in solution trans-
ported by diffusion to the growing diamond surface. The dimensions of these molten
metal films are such that it is unlikely that convection plays any significant role in
carbon transport for this type of crystallization. The growth of diamond grits of
good morphology occurs through lateral propagation of steps over the surface.
The detailed kinetics of surface adsorption, surface diffusion and incorporation
of carbon species into the step structures needs to be studied to understand the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallizatioiz of Diamond 495
and properties that affect the strengths of the diamond particles. Not surprisingly,
the most important macroscopic structures of the diamond particles turn out to be
size, shape, crack, and inclusion content. It was also demonstrated that particular
nitrogen related lattice defects such as the H3 and N-V center affect the strength
of the crystals. For a comprehensive discussion of the observed impurity lattice
defect centres such as H3 and N-V, the reader is referred to Davies [51]. An impor-
tant observation in [50] was that some of the microscopic defects may well
strengthen the diamond by impeding dislocation motion and possibly acting as
crack arresters.
The very complex nature of the macroscopic and microscopic structures as they
affect strength and the behavior of diamond abrasive/particles still requires
extensive work to elucidate fully. However, from a crystallization point of view,
gaining control over the crystallization behavior is the key to the production'of
optimal diamond abrasives. This, of course, may be achieved by choice and
manipulation of the pressure/temperature conditions, source carbon structure
and solvent/catalyst metal type, leading to control over nucleation and growth
rates.
Large diamonds are best grown using the temperature gradient technique, where
diamond is used as the source of carbon. In this technique, as described by Wentorf
[52], the driving force for reconstitution, under diamond stable conditions, is pro-
vided by the higher solubility of diamond in a hot zone of the solvent/catalyst
and the consequent crystallization of diamond in a cooler zone at constant pressure
(Fig. 9).
The design of the high pressure cell also influences the rate at which this carbon
transport mechanism proceeds. A schematic diagram of a high pressure cell used to
grow reconstituted diamond is shown in Fig. 12. A useful axial temperature gradient
is established by manipulating the distance between the diamond source and seed
pad, which (at constant cell pressure) determines the growth rate and hence, quality
and size of the diamonds grown on the seeds in a fixed time. The small, well formed
diamond seeds, which act as a template for new growth, are mechanically attached
to the seed pad. For a reaction cell pressure of about 5.5 GPa, the temperature at the
diamond source is about 1450°C. The solvent/ catalyst bath is agitated by convec-
tion and soon saturates with carbon. The solvent/catalyst in the vicinity of the seeds,
at a temperature of 1420°C, is therefore supersaturated with respect to diamond,
providing the driving force for diamond growth on the seeds, which are selected
to be about 0.5 mm in size [53, 541.
The nucleated diamond is restrained by the seed pad, resulting in truncated cubo-
octahedral growth, whereas the orientation of the grown diamonds depends on the
orientation of the seed. The habit of the reconstituted diamond depends largely on
the solvent/catalyst, internal pressure and temperature of the reaction cell.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallization of Diamond 497
t
DIAMOND POWDE
CARBON HEATER
SOLVENT-CATALYS AT
GROWING CRYSTAL
SEED CRYSTAL T2 f
Figure 12. Schematic diagram of a high pressure cell used in the reconstitution method.
Figure 13. Diamonds grown by De Beers Diamond Research Laboratory from seeds using the tem-
perature gradient technique.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
498 1 Diamond Materials and their Applications
1.2.5.1.2 Type IIa and IIb Diamonds Aluminum, titanium, niobium, and zirconium
are used to remove nitrogen, in the growth of near colorless synthetic diamonds
(type IIa) [62, 66, 671. These nitrogen-getter elements have a high affinity for nitro-
gen and contribute to nitrogen removal by forming stable nitrides. As mentioned
earlier in the chapter, there is a relationship between nitrogen in the grown diamond
and nitrogen solubility in the solvent/catalyst.
The dependence of nitrogen content on the alloy composition is shown in Fig. 14,
where it can be seen that an increase in cobalt content (which is accompanied by
decreased nitrogen solubility in the iron-cobalt solvent/catalyst alloy) results in
increased nitrogen concentration in the grown diamond [67].
Nitrogen concentration in the crystals decrease with increasing amounts of alumi-
num or titanium added to the solvent/catalyst (Fig. 15). Even when large amounts of
aluminum are added, it is difficult to achieve purity levels of less than 0.1 p.p.m.
nitrogen. This is to be expected as the aluminum nitrides are believed to readily
decompose, whilst titanium nitrides are not readily decomposed, making titanium
Table 1. Summary of large synthetic diamonds grown for research purposes by some of the leading producers of synthetic diamonds
Max. size Growth Growth Nitrogen
Solvent ( 4 time (h) rate Color content Morphology
(mg h-') (p.p.m.1
~~ ~
Co-Ti [62] 3.4 300 2.3 near colorless <0.01-2 ( l l l ) , (100) > {113), {115), {llO}
Fe-A1 [65] 4.6 500 1.8 near colorless 0.4-1 1 { l l l ) > (113) > (100) > (110)
Fe-Co [62] 25 1000 5 yellow 100-300 { l l l } , {loo} >> {IlO), j113)
Fe-AI-B (0.02 wt YOB) [65] 5.1 160 1.3 blue - { l l l ) , {110), (100) {113), {115} x
- - - yellow 50-100 { I l l } , (110) >> {110), (113) 2
Ni [43]
Ni -Fe 3.5 - - yellow 50--400 (1111, {loo) 0
Fe-Co-Ti [65] 4 - - near colorless <O.Ol {Ill}, (100) > {113}, {115}, {llO) $CI
-
=:
b
8'
;
R
P
W
W
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
500 1 Diamond Materials and their Applications
Figure 14. Nitrogen concentration in diamond versus the cobalt content of iron + cobalt solvent.
The diamonds were grown with 1.5 weight-% titanium addition using (a) high-purity graphite
and (b) synthetic diamond powder [67].
,
\
,
', Al
\
Ti
--_
Figure 15. Nitrogen concentration of diamond grown with A1 or Ti added as a nitrogen getter [67].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystullization of' Diamond 501
impurities. These impurities fall into two classes. There are impurities on an atomic
scale (point defects) and these include nitrogen, boron, and metallic elements such as
nickel. The second class of inclusions are present in the form of metallic or nonme-
tallic particles, both microscopic and macroscopic. The occurrence of these inclu-
sions is controlled by crystal growth kinetics.
1.2.5.2.2 Point Defects Nitrogen and boron are ubiquitous atomic-scale impurities
in synthetic diamonds. The free energy of formation of nitrides and borides in the
solvent/catalyst melt will determine the extent to which these impurities are
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
502 I Diamond Materials and their Applications
Figure 16. Dependence of critical growth rate of diamond crystals on { 100) faces of seed crystals, on
varying seed size [70].
Figure 17. Cross section of a cobalt-titanium grown crystal with inclusions resulting from hoppered
growth.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystullizution of Diamond 503
ever, small numbers of cut and polished synthetic diamonds, mostly produced in
the former Soviet Union using the so-called BARS equipment, have appeared in
the gem trade in recent years. The majority have been yellow in color but a few
near-colorless synthetic diamonds have also appeared. As uncut crystals, it is
easy to distinguish synthetic diamonds from natural diamonds by their distinctive
morphology and the presence of a seed. But cut and polished synthetic diamonds
are also readily identifiable as such by standard gemmological techniques such as
microscopy of inclusions and of features on polished faces (due to sector-
dependent differential hardness) and zonation of color or fluorescence (due to
sector-dependent differential uptake of optically active impurities) [79-841. Near-
colorless and boron-doped, blue in color, synthetic diamonds tend to exhibit
strong, long-lived phosphorescence. To enable large numbers of diamonds to be
screened rapidly for the presence of synthetic diamonds, instruments have been
developed. The first of these instruments detects an optical absorption line at
415 nm present in the majority of natural diamonds. Those diamonds not exhibit-
ing this feature may be examined using a second instrument which, by exciting
near-surface fluorescence with very short wavelength ultraviolet light, produces a
fluorescence pattern from which the growth structure of the diamond may be
inferred. These patterns are very different for synthetic as compared to natural dia-
mond [85].
lysis by Berman and Simon [ 141, which has been reviewed recently [90] in the light of
new data, particularly the direct electrochemical determination of the free energy dif-
ference for the diamond to graphite transformation [91]. From the equilibrium expres-
sion of Berman and Simon, Eq. (1) in Section 1.2.2, it may be possible to include
additional terms which offer an explanation for some cases of novel non-CVD synth-
esis, or to expand the existing terms such as AH', the enthalpy change at zero pressure.
Figure 19. Micrograph of a section of a diamond synthesis capsule in polarized light at 200x mag-
nification. Recrystallization of coarse graphite flakes of high crystallinity is a process which
competes with diamond growth even in the diamond stable region. The tabular shape suggests
that surface energy is a significant driving force for recrystallisation even in this size range.
(interfacial) energy.
In practice, these solid-vacuum interfacial energies are not likely to be seen, since
the diamond surface and the graphite prism-face surfaces will reconstruct or react
with adatoms such as hydrogen and oxygen to reduce their surface energy. Under
atmospheric conditions, diamond is terminated by a range of hydride and oxide
groups and is a low energy surface [loo].
During high pressure solvent/catalyst synthesis, the interfacial energies of metal
melt-diamond and metal melt-graphite have values lying between the metal surface
tension, 1800mJ mP2,and the solid-vacuum surface energy. Some values have been
estimated from contact angle data [loll although the complexities of contact angle
work in the diamond stable region are formidable.
By including a surface energy (interfacial energy) term in the equilibrium expres-
sion it is possible to make some interesting predictions.
1
AG', = A H ! - T L I S ; , ~ A VTd P + A(yA) = 0
where y is the interfacial (surface) energy of any phase and A is its molar area, which
is not an intensive variable like pressure or temperature but varies with both the size
of the particles and the density of the phase, diamond having a lower molar area for
the same particle size than graphite.
For instance, using this approach, Nuth [ 1021 has predicted that diamonds (in
vacuo) are more stable than graphite in the size range 1-5nm, a result which was
supported by Badziag et al. [lo31 using a different approach.
There is experimental support for this in the finding that nanometre diamonds are
abundant in primitive meteorites and possibly in the interstellar dust clouds [ 1041.
Similarly, growth of graphite in the diamond-stable region during high pressure,
high temperature synthesis, and the corollary, an overpressure for nucleation of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallization of Diamond 507
diamond, are explainable using the additional interfacial energy term without the
need for kinetic arguments [105, 1061. Once the growing particles are in the size
range of microns, the molar surface area becomes too small for this term to have
any effect and the classical phase equilibrium line is applicable.
There is another way of looking at the same effect, since a surface energy per unit
area is equivalent to a line force per unit length, or surface tension, hence there is an
excess pressure in a spherical droplet given by the Laplace equation:
A P = -2Y (3)
Y
where y is the surface energy of any phase and r is the radius of curvature. For diamond
with sigma at 5 J mP2and Y at 1 nm, the excess pressure in a nanometre sized sphere is
5 GPa, which would be sufficient to stabilize the diamond phase to at least 1600K.
This could explain processes such as laser pyrolysis of hydrocarbons [ 1071, where
small quantities of nanometre sized diamond have been detected in the carbon-
aceous products. A recent result suggests that it is even possible to generate and
stabilize a diamond core in the centre of a graphitic ‘onion’ a giant fullerene, by
electron irradiation and annealing of the outer graphite layers. In this process,
‘the sputtering-induced loss of atoms and the closure of the shells around vacancies
lead to extreme surface tension, which is finally responsible for the spherical shape of
the onions’ [108].
This publication is a good example of the development of analytical techniques
such as high resolution electron microscopy lattice imaging [1081 and micrometre-
area Raman spectroscopy (microRaman) [ 1091 which permit the researcher to
determine that diamond is present in crystallites of submicrometre, perhaps even
nanometre size. Early scientists like Moissan [I 101 and Hannay [I 113 performed
experiments in the pressure and temperature ranges now being cited as feasible
for hydrothermal synthesis of diamond [38, 112, 1 131. Had they been equipped to
use the analytical techniques now available, they may have been proven correct in
their claims, but in a particle size regime much finer than they were able to inspect.
HPHT: nongraphitic LPSSS (low pressure solid state Triple laser process using
carbon [30], [92] source) as by Roy et al. [ 1 16, 1 171 CO, (QQC process) [ 131
Hydrothermal synthesis - Hydrothermal flow synthesis with PVD using ion beams -
nonflow [113, 381 flame [116, 1171 reviewed by Prins [123]
Laser in absence of Laser in presence of hydrogen [ 1181 Precipitation from
hydrogen [105, 1151 nickel-sodium hydroxide
melts [ 1241
Diamonds in interstellar Fluidized bed deposition on
space [102-1041 particles from acetylene flame [ 1191
Electron irradiation of Discharge electrolysis from
carbon ‘onions’; nanometre water:organics mixtures [120-1221
scale HPHT [ 1081
Bundy and Kaspers [I271 Modified belt 1967 13 GPa and1000"C Hexagonal diamond
Boehler [I281 DAC 1994 10 GPa and 1700°C
Xu and Huang [ 1291 DAC 1994 6GPa and 1700°C Raman at 1333 cm-'
Cubic boron nitride (cBN) is presently the second hardest material known to man
after diamond. The compound boron nitride is unknown in nature. The combina-
tion of atoms of boron and nitrogen in stoichiometric arrangements is such that
B-N is isoelectronic with C-C. It is not surprising that, by analogy with carbon
structures, similar crystalline structures of boron nitride occur. By analogy with gra-
phite, hexagonal boron nitride is a soft, slippery material with an anisotropic, hex-
agonal structure made up of sheets containing interlinked six-membered rings
(3 x B-N). There are some differences however, notably, the n-orbital overlap of
graphite is absent, so hBN is an electrical insulator and is white in color. For kinetic
reasons, all conventional chemical synthetic routes for making boron nitride result
in the hexagonal structure. Again, by analogy with carbon, boron nitride can take
up the diamond structure with alternating atoms being boron and nitrogen. This is
the common zinc blende or sphalerite structure. This phase of boron nitride has a
hardness about half that of diamond; and a bulk modulus of about 367 GPa com-
pared to the 435 GPa of diamond.
The value of cBN as an abrasive lies in its much higher oxidation stability com-
pared to that of diamond (1200°C compared to 600°C) and in its reduced chemical
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.2 The Crystallization of Diamond 5 11
Figure 21. Phase diagram of boron nitride showing the latest stability fields suggested by Soloz-
henko [ 1341 compared to the earlier Bundy-Wentorf diagram [ 1381.
interaction with ferrous materials when compared to diamond. It is, therefore, the
abrasive of first choice for the grinding of hard, ferrous materials.
When the formation of cBN was first reported [133], the pressure and temperature
phase diagram was believed to be very similar in form and stability field positions to
that of carbon, with the equilibrium line between the hexagonal and cubic phases
intercepting the pressure axis when temperature is at absolute zero [134]. The
most recent phase diagram for boron nitride is that of Solozhenko, shown in
Fig. 21 [135]. This diagram was initially calculated and subsequently partially
experimentally substantiated. The hexagonal-cubic-liquid triple point is at about
7 GPa and 3800 K. The hexagonal-cubic equilibrium line intercepts the temperature
axis at just above 1500K. However, the exact temperature of the intercept is not
well-established [136], although it is now generally agreed that cBN is the stable
form of boron nitride at room temperature and pressure. This is where boron nitride
differs significantly from diamond. It does not transform to the hexagonal structure
until temperatures of at least 1200°C are exceeded.
cBN is synthesized using the same general technology as that used for commercial
diamond synthesis. A mixture of hBN as source and various catalyst/solvents are
subjected to pressures in the range 4.5-5.5 GPa and temperatures between about
1500 K and 2000 K. The elements initially investigated and demonstrated to
function as solvent/catalysts were the alkali and alkali earth metals, as well as
antimony, tin, and lead [137]. Subsequently, the nitrides of lithium, calcium. and
magnesium were shown to be more effective. Single crystals of cBN, in the size
ranges appropriate to grinding applications, both black and amber in color, are
now readily synthesized by several companies on a commercial basis.
The recent developments in the boron nitride phase diagram, particularly with
respect to the room temperature and pressure thermodynamic stability of cBN,
have stimulated work aimed at producing cBN over a wider range of conditions
[136]. In conjunction with this, it is now known that a wide range of nitrogenous
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5 12 1 Diamond Materials and their Applications
compounds are capable of aiding the transformation of the hexagonal to the cubic
form [138].
In 1971, Katzman and Libby, from the University of California at Los Angeles
[144], described the formation of sintered diamond compacts in the presence of 20
vol% of cobalt binder. The sintering was carried out for about 20 min at 6.2 GPa in
a piston and cylinder type high pressure-high temperature apparatus, and at a
temperature (159OOC) which was above the cobalt-diamond eutectic, 1570°C at
62 kBar. They claimed their results suggested that the cobalt cleans the diamond
surfaces of any adsorbed gases that might prohibit or retard grain growth or
fusion, thereby making unnecessary the surface cleaning and degassing procedures
described by Stromberg and Stephens. They also claim that the optimum cobalt
content was 20% by volume. An attempt to substitute cobalt with nickel under iden-
tical sintering conditions also led to a somewhat softer material. Compacts made
with diamond grit sizes of 1-5 pm and 0-2 pm were harder than those made with
grit sizes of 10-20 pm.
Wentorf and Rocco in 1973 [145] described in situ diamond sintering on top of a
layer of cemented tungsten carbide (WC). The idea here was that any solvent/catalyst
metal (cobalt) which was present in the metal carbide would act as a source for liquid
phase sintering. The sintering method involves placing in a cylindrical mold, a 0.5-
1 mm layer of diamond powder on top of a thicker layer (lomm) of WC-lOwt%
cobalt powder. The mold is then subjected to high pressures (>5 GPa) and tempera-
tures (1400-1600°C) for at least 1Omin to 1 h. Under these conditions, the cobalt
melts and liquid phase sintering of both the carbide and the diamond occur.
Veraschagin et al. [146], in 1975 showed that, if the metallic binder was initially in
the form of a disc (rather than a powder), better packing densities were achieved and
the molten binder would thus be found mainly in the voids rather than in between
diamond grains, thus resulting in strong and uniform diamond-to-diamond bonding
throughout the compact.
Figure 22. Schematic description of processes taking place during cold compaction.
larger at a given pressure. But the difference between coarse and fine grains gets
smaller as pressure is increased.
Crushing of diamond particles during this stage dramatically shifts the particle
size distribution.
The powder, which has a lower number of sharp edges per particle, is more stable
to compression forces than the one with many spallings, sharp edges, etc. Since their
weakest planes are the planes where they contain inclusions; the density, size and
shape of the inclusions may be a critical factor in the crushing of diamond particles
during the cold compaction stage.
4
pressure
temper fure
Figure 23. Deformation of diamond grains under high pressure and temperature
Once the heat is turned on at high pressures, the densification of the diamond
compact proceeds mainly by further crushing and rearrangement of the crushed par-
ticles up to 700°C [149]. After 700°C, densification proceeds with plastic deforma-
tion [150-152]. The extent of plastic deformation is a function of temperature and
pressure. At higher temperatures, the diamond grains become more round, their
sharp edges disappear, they start to deform in the zones of contact with each
other. Pores get smaller, even diminish at a rate initially determined by their size
at the end of the cold compaction stage. Pore size reduction rate is faster for coarser
pores than for smaller pores [149]. But the time required for a pore to disappear is
constant at a given pressure and temperature to a first approximation, regardless of
size [153, 1541.
Also at higher temperatures, graphitization of the diamond surfaces, which are
not in contact with another diamond surface, takes place. The amount of graphiti-
zation is a function of the temperature at a given pressure [150]. The degree of
graphitization in diamond compacts can be related to the specific resistivity of the
compacts since graphite is a relatively conductive material compared to diamond
which is an insulator. That is, the more graphitization takes place, the more conduc-
tive the material becomes.
At even higher temperatures, bulk graphitization of diamond grains starts to take
place and the hardness and wear properties of pcD produced (binderless) start to
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5 16 1 Diamond Materials and their Applications
As the pores in the diamond layer become filled, the magnitude of the pressure
gradient is decreased. Accordingly, the rate of mass transfer is decreased,
although its direction is maintained. As this pressure gradient decreases, the
effect of the pressure gradient in the reaction volume of the capsule, which was
found to be about 1 GPa in the axial direction by Russian researchers in their
own capsule [ 1571, starts dominating the infiltration gradient. Whatever the direc-
tion of the pressure gradient is, this stage of mass transfer is completed when
there is no pressure gradient at the interface between diamond layer and the
substrate and when the pores are completely filled. Thus, the rate of mass transfer
approximates zero and the binder content in the diamond layer becomes a
maximum.
The graphitization of the diamond surface, described in Section 1.3.3.2, is impor-
tant in terms of infiltration of the solvent/catalyst binder. Metals such as cobalt dis-
solve graphite more readily than diamond at the sintering pressures (Fig. 9b).
Hence, the more graphitization that has taken place, the more cobalt is expected
to be in the final material.
A characteristic feature of the fourth mass transfer stage is a change in its direc-
tion: it has been observed [ 155-1 571 that the quantity of liquid phase in the diamond
layer is decreased and increased in the substrate. The beginning of mass transfer
indicates that a pressure gradient arises in the system, directed from the diamond
layer to the substrate. The origin of this phenomenon was explained by Shulzhenko
et al. [ 155, 1561 in terms of temperature gradients and their changes resulting from
the surface tension of the molten binder-diamond interface.
Also, this effect can be explained by the fact that, as the sintering of diamond
compact proceeds, the area of the contacts between the diamond grains increases.
This accelerates shrinkage and the densification process. This in turn according to
Shulzhenko et al. squeezes the excess cobalt, which cannot be accommodated in
the ever shrinking pores and cavities in between diamond grains, out of the diamond
layer. One destination of this stage of mass transfer is the diamond/substrate
interface.
As a result of the change in direction of binder mass transfer in the fourth
stage, the binder content of the diamond layer is decreased by an amount of the
order of couple of percent [155, 1581. The binder content of the diamond layer
can be controlled by lowering the sintering temperature and extending the sintering
time.
Almost complete exclusion of cobalt from large areas where two grains have
grown together is observed by TEM studies on polycrystalline diamonds sintered
with cobalt binder [159]. No thin, residual layer of cobalt was detected at the
grain boundaries. The thickness of regrown diamonds is usually much lower than
the original diamond grains it is precipitated onto, and it joins with the regrown
diamond precipitated onto a neighboring diamond grain forming a high angle
grain boundary. So it can be said that only the surfaces of the diamonds in contact
with a pore just before infiltration takes place are potential candidates for
reprecipitation of new dislocation free diamond. The areas next to the diamond
to diamond contact points or surfaces are the preferred locations for diamond
reprecipitation due to lesser surface energy requirements.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5 18 1 Diamond Materials and their Applications
1.3.3.5 Conclusions
During cold compaction of diamond powders, coarser powders densify more and
have larger size pores than finer powders at a given pressure.
During hot compaction, the densification of the diamond compact proceeds
mainly by crushing and rearrangement of the crushed particles up to 700°C.
Above that temperature, the densification proceeds with plastic deformation. The
extent of plastic deformation is a function of temperature and pressure. Pore size
reduction rate is faster for coarser pores than for smaller pores.
The degree of plastic deformation and of graphitization taking place during hot
compaction seems to influence the wear resistance and hardness of the sintered dia-
mond compacts. When both plastic deformation and graphitization are taking
place, properties are more sensitive to plastic deformation until bulk graphitization
of diamond grains starts.
lnfiltration rate of the metallic binder from the substrate is influenced by pressure
gradients created by the applied external load, pressure gradients inside the capsule
and the pore size of the diamond compact at the onset of infiltration, temperature,
and the substrate binder mean free path and chemistry.
Table 4. Properties of the pcBN products and other cutting tool materials.
tantalum, chromium, molybdenum, and tungsten (which are, in fact, the elements of
groups IVb, Vb and Vlb).
In Sumitomo's patent, high heat conductivity together with high hardness are
mentioned as paramount properties when selecting a suitable binder material. A
rapid rise and fall in temperature in the tool during interrupted cutting often
leads to the formation of cracks. Therefore, a binder material that couples high ther-
mal stability and hardness with high thermal conductivity at high temperatures is
suitable for this kind of application.
The manufacturing process (especially the cold and hot compaction stages) of
pcBN is very similar to that of pcD, summarized in Section 1.3.3. However, the
types of binders used and liquid phase sintering mechanisms can be quite different.
In pcBN production, 'reactive sintering' plays a major role in terms of driving the
densification process, whereas in pcD, dissolution and precipitation and/or
adhesion/coalescence are the main driving mechanisms. A typical example for
pcBN would be the following reaction [162].
3A1+ 2BN 4 A1B2 + 2A1N (4)
where the reaction products A1B2 and AlN act as binders [ 1631. However, the chemi-
cal reaction to achieve these products is the driving force for sintering this particular
pcBN product.
The properties of the pcBN products are not only determined by the final micro-
structure, but also by the phases formed during sintering. The pcBN properties are
listed in Table 4 and the properties of the main phases are given in Table 5.
A direct conversion process can produce translucent pcBN which has almost
theoretical density and very high thermal conductivities. In 1972 and later in
1974, Wakatsuki et al. [164, 1651 sintered a polycrystalline cBN directly from
hBN at pressures as low as 55kBar and temperatures of 110&1400°C. Corrigan
[1661 observed that thermal conductivity of directly converted compacts increases
with increasing grain size. Fukunaga and Akaishi [167] used a small amount of
Mg3BN3 to promote conversion from hBN to pcBN at pressures of 5-6GPa and
at 1500°C.
ul
N
0
b
6'
$
a
6'
c
Table 5. Properties of the main phases found in pcBN products after sintering. D
%
a
cBN AIN TiB2 TIN TIC A1203 Diamond
(for reference) $
L
L
Crystal sructure cubic hexagonal hexagonal cubic cubic hexagonal cubic %2
Density [g ~ m - ~ ] 3.48 3.26 4.52 5.22-5.44 4.92 3.96 3.51
~
Melting point ["C] 2700 2300 3197 2930 3065 2050 s.
Knoop hardness [GPa] 41 12 26.5 19 28-35 23 57-104 2
Young modulus [GPa] 700-800 318 434540 370 450 421 1141
Thermal expansion coefficient [10-6K-'] 3.2 3.9 8.1 8 8.6 8.1 1.54.8
Thermal conductivity [W.m-'K-'] 150-700 200 80 25 3343 14 500-2000
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.4 New Ultrahard Materials 521
1.4.1 Introduction
Diamond’s combination of properties make it a unique material. Although hardness
is its primary characteristic, thermal conductivity, compressive strength, refractive
index, spectral transmittance, and chemical stability are either the highest or among
the highest found in nature.
This combination of properties can be traced to the same structural characteris-
tics of diamond that give rise to its high hardness. It is therefore reasonable to expect
that other ultrahard materials would also exhibit such a suite of properties. This
would make them also desirable for a number of industrial applications.
New ultrahard materials might possess new attributes other than a higher hard-
ness that would make them more attractive than diamond in some cases. A different
chemical composition would give rise to different interaction of cutting tool and
workpiece as is the case for cBN, as would a different crystal structure and ultrahard
particle morphology.
The search for new ultrahard materials is motivated largely by these considera-
tions, but also from purely economic ones. It might be possible to synthesize such
materials through routes cheaper than those involved in the case of diamond,
thus accessing new applications and new markets. In this paragraph we briefly
review the basic ideas behind the search for new ultrahard materials, as well as
the latest developments in the search for some of the ‘identified’ cases.
1.4.2 Hardness
Hardness is a measure of a material’s ability to resist elastic and plastic deformation.
The hardness of non-ideal material is determined by the intrinsic stiffness of the
material, as well as by the nature of its defects, be they point defects, dislocations,
or macroscopic defects such as microcracks etc. For ideal systems, the hardness of a
material will scale with its bulk modulus.
Figure 24 shows the Knoop hardness as a function of the bulk modulus for a
number of representative materials [168]. Table 6 shows the hardness as well as
the bulk modulus of a number of representative materials.
The bulk modulus of a solid can be calculated by means of ab initio calculations.
A review of this field can be found elsewhere [169]. Such calculations require exten-
sive computer time. Liu and Cohen in 1989 [ 1701 proposed an empirical formula for
the bulk modulus of solids with the zinc blende structure:
B = (19.71 - 2 . 2 0 h ) p (5)
where B is the bulk modulus in GPa, d is the bond length in Angstroms, and h is a
measure of the ionicity of the compound. For purely covalent compounds (group
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
522 1 Diamond Materials and their Applications
80oo
diamond
7000
6ooo
n
5ooo
N cubic BN
E
E
m 4000
?5
z
I 3Mx) ZrC
Sic
Y
0
2ooo oAL03
rn-ZrO,
Si
1000
a 0"
OMgO
NaCl %a0
a 0
I I I I I I I I
0 50 loo 150 200 250 300 350 400 450
6,( G W
Figure 24. Knoop hardness (KHN) as a function of bulk modulus (B,) for representative materials.
IV), h = 0 while, for compounds of groups 111-V and 11-VI, h is equal to 1 and 2
respectively. This expression can be modified to allow for nontetrahedrally coordi-
nated compounds as follows [ 171, 1721:
B = (Nc/4)(1971 - 220qd-3.5 (6)
Where Nc is the average coordination number.
From Eq. (6) it follows that in order for the bulk modulus to be high, the crystal
must have: a large coordination number; a high degree of covalency; and short
interatomic distances, meaning small atoms. Although a high coordination
number may be attainable in metallic materials, where it can be as high as 12, metal-
lic bonds are much weaker than covalent bonds which can allow a maximum coor-
dination number of 4 in naturally occurring materials. Because of these
considerations the search for new materials is confined to tetrahedrally coordinated
compounds with high degree of covalency and with small atoms.
On the basis of this model, it was suggested that, for h = 0.5 and d = 1.47-1.49, a
hypothetical tetrahedral compound between carbon and nitrogen would have a bulk
modulus of 461-483 GPa, which would exceed that of diamond (443 GPa).
However, not all potentially hard or ultrahard materials can be found within the
two top periods of the periodic table, nor are they confined to tetrahedral coordi-
nation. A number of sp2 coordinated structures have been proposed which,
mainly due to their high density, would posses a high bulk modulus and would
therefore potentially exhibit high hardness [ 173-1 751. Superdiamond structures
have been proposed by Diedrich et al. [176] and Alberts et al. [177], based on poly-
merized derivatives of tetraethylmethane C(C2H)4and the carbon skeleton of allene
CH2=C=CH2 respectively. Leger [ 1681 investigated the feasibility of generating
ultrahard materials by producing the high pressure phases of a number of oxides.
Such phases would have a higher degree of coordination and bonds with a higher
degree of covalency than their ambient pressure allotropes.
1.4.3 C3N4
In 1989, Liu and Cohen reported local density approximation pseudopotential cal-
culations on a hypothetical carbon nitride phase, PC3N4 which suggested that it
could be metastable under ambient conditions [170, 1781 and would have a short
c-N bond length (0.147 nm), low ionicity ( ~ 7 % and ) a bulk modulus higher than
that of diamond. Figure 25 shows the crystal structure of hexagonal PC3N4derived
from pSi3N4 [ 1791.
Alternative structures have been proposed by Liu and Wentzcovitch [ 1SO], Wentz-
covitch and Martins [181] and Wentzcovitch [182], proposing a cubic zinc blende
structure with one carbon vacancy per unit cell, and a structure resembling graphitic
CN with one carbon vacancy per four nitrogen sites. This study was based on an ab
initio molecular dynamics scheme with a variable cell shape algorithm.
More recently Teter and Henley [ 1831, using first principles pseudopotential total
energy techniques, predicted a cubic form of C3N4with a zero pressure bulk mod-
ulus of 495GPa, higher than that of diamond. The same authors predicted that
aC3N4and graphite-structure C3N4are energetically more favorable than PC3N4.
Figures 26 and 27 show the structures of the proposed cubic C3N4 and aC3N4
compounds respectively.
The above theoretical predictions have generated a great deal of experimental
activity aimed at synthesising a new ultrahard compound.
Several attempts have been made to synthesize thin film C3N4,be it cubic or hex-
agonal. These include reactive magnetron sputtering [184187], laser ablation [188]
ion beam assisted deposition (IBAD) [189], plasma [191, 1921 and plasma-enhanced
chemical vapour deposition [190].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
524 1 Diamond Materials and their Applications
Figure 25. Hexagonal crystal structure of PC3N4 derived from pSi3N4. Two types of sp2’ bonded
N atoms are present. The first type is in a trigonal planar configuration with the trigonal plane
perpendicular to the c-axis. The second type forms, together with the carbon atoms, the tunnel at
the center [179].
In most efforts, the films produced were either amorphous, or their nitrogen con-
tent was less than the target of 57%. D. Li et al. [184] using a magnetron sputtering
system produced crystalline carbon nitride/titanium nitride composite coatings with
a reported hardness of 55 GPa.
Figure 26. Two adjacent unit cells of cubic C3N4carbon atoms are shown schematically as the black
atoms [192].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
I .4 New Ultrahard Materials 525
P’C3N4
Figure 27. Predicted crystal structure of aC,N4 [192], compared with that of PC3N4.
Zhang et al. [ 1941 at Harvard were able to synthesize an amorphous CN thin film
with composition C2N. This material had exceptional hardness [195-1971. In some
cases, tiny crystallises of covalent C-N were observed in the amorphous films [19&
2021. Bhusari et af. in February 1997 [193] were able to synthesise large crystals
(750 pm) of Si-containing carbon nitride consisting of a predominantly C-N
network, by microwave CVD.
More recent efforts involving nitrogen ion beam assisted deposition [203, 206,
2111, hot filament CVD with or without rf plasma [204, 205, 2091, or microwave
plasma enhanced CVD [213, 2141, with bias assistance [207, 208, 2121 were able to
produce thin films containing C3N4,some with the BC3N4structure and some with
the cubic or a-structure.
Yan et al. [205] using hot filament CVD were able to synthesize fully crystalline
films combining both aC3N4and PC3N4.
Xu et al. [210] using a reactive magnetron plasma source claim to have deposited
polycrystalline BC3N4films with crystallites as large as 20 pm.
Xiao-Ming et al. [211] using argon arc nitrogen-assisted bombardment claim to
have deposited completely polycrystalline BC3N4 films. They quote a measured
hardness of 52.6 GPa.
Similar results were obtained by Wu et al. [212] who used RF plasma enhanced
CVD.
Finally Chen et al. [213], using microwave plasma enhanced CVD, claim to have
grown crystals larger than 10 pm, and propose that the incorporation of silicon from
the substrate into the C-N structure promotes BC3N4crystal growth.
Efforts to synthesize C3N4at high pressures have been much less extensive than
those made to synthesize this material in the thin film form. A number of trials [214]
indicate that it may be possible, through the use of pressure to incorporate nitrogen
into sp’-bonded carbon nitride.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
526 1 Diamond Materials and their Applications
1.4.7 Stishovite
The search for new ultrahard materials has also been extended into nonboron based
oxides. The expectation was that high pressure allotropes would exhibit a high
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Industrial Applications of Diamond and cBN 527
coordination of the metallic ion, as well as more ionic bonding, thus hopefully pos-
sessing higher stiffness and, therefore, higher hardness.
Leger et al. [166] in 1994 measured the bulk modulus of the high pressure phases
of HfOz and RuOz, which they obtained at pressures above 42 and 12GPa. The
resulting bulk moduli place the two compounds above diamond in the case of
Hf02 and between cBN and diamond for Ru 0 2 .
In 1996 Leger et al. [228] reported the synthesis of stishovite from c1 quartz at
20 GPa and 1100°C and measured a hardness of 33 GPa for the polycrystalline com-
pact produced. As the samples obtained were very small, a maximum load of 1.9 N
during hardness testing was possible. Synthesis of this material in larger quantities
has not yet been possible.
1.5.1 Introduction
As discussed earlier, the modern industrial diamond business began to expand sig-
nificantly following the commercial availability of synthetic diamond in the late
1950s. Before then, only natural diamond produced as a by-product of gem dia-
mond mining was available and was, by its nature, limited in volume. Demand
was increasing to the point where new application development was a largely self-
defeating exercise since, the more successful they were in terms of using large
volumes of diamond, the less feasible they became in terms of total world supply.
With the introduction of synthetic diamond, this situation changed dramatically
on two important fronts. Firstly, the only constraint remaining on the development
of new applications was the cost and technology of producing a suitable product;
and secondly the constraints on product characteristics imposed by the nature of
a natural raw material were removed: diamond could now be specifically engineered
for its intended use.
Natural and synthetic diamond are complementary, not competing products.
While today a minor player, natural industrial diamond still plays an important
role in a variety of special applications where one or other of its particular
characteristics will make it the preferred material (refer to Section 1.5.5 for some
examples).
Industrial applications of diamond have developed over the years as a result of
developments by tool makers, machinery manufacturers and the advert of new
materials. In addition, new diamond (and cBN) products have evolved, either as
a result of technological advances in synthesis or in response to the requirements
of a new application, and it is this multi-partnership relationship within the
industry which has resulted in the dramatic growth and diversification since the
early 1960s.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
528 I Diamond Materials and their Applications
- to release worn abrasive particles as they reach the end of their useful lives, thus
exposing new ones.
It is therefore a general requirement of a bonded tool that the abrasive and the bond
material wear at similar rates.
The most common types of bond used are as follows:
depends primarily upon the workpiece material, the machining conditions and the
primary role of the machining operation (e.g. to produce a high surface finish or
high material removal rates). The role of the EP bond layer is to support the abrasive
and not wear significantly during the life of the tool. When new, EP tools have a high
degree of protrusion of the abrasive particles above the bond layer (much higher
than the other types of bond) and hence can machine at high material removal
rates while generating low cutting forces. The term often used to describe such a
typical application for each of the four bond types and some of their particular
advantages. The characteristics of the abrasive used can be selected from a wide
range available; size, resistance to impact and fracture characteristics being
among the more important and Fig. 36 summarizes the main differences and their
effects.
Grinding operations can take place with the application of a coolant or lubricant
(wet grinding) or with no such application (dry grinding).
~ ~ ~ oblockv ~ -r ,
n ~orits
withstand higher forces - withstand lower forces -
greater exposure - less exposure -
higher rates of stock removal - lower rates of stock removal -
lower degree of surface finish - higher degree of surface finish -
suited to more ductile, suited to hard, brittle,
long chipping workpiece materials short chipping workpiece materials
Figure 37. A typical friable diamond abrasive PDA321 used (normally in metal clad form) in resin
bonded grinding wheels.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
534 1 Diamond Materials and their Applications
Resin bond tools are used to machine a wide variety of non-ferrous, usually
abrasive, workpiece materials. A very small number of ferrous materials can be
ground economically with diamond, for example, some carbon rich cast irons
which, providing machining temperatures are maintained relatively low, do not
chemically react with the diamond, and other difficult-to-machine materials such as
some grades of stainless steel which have a tendency to work harden during machin-
ing. The major application area for resin bond diamond wheels is the grinding of
cemented tungsten carbide (which was also its initial application area in the 1940s).
The list below gives the main areas of application in descending order of relative size:
- grinding cemented tungsten carbide
- grinding ceramics and cermets
- grinding semiconductors
- grinding polycrystalline diamond (pcD)
- machining stone/glass
- others
Although the machining of cemented tungsten carbide represents approximately
half the resin bond tools used, other areas are growing, notably semiconductors,
ceramics and cermets, pcD and the machining of stone.
Figure 40. A typical high strength diamond abrasive PDA999 used in grinding wheels.
Metal bond diamond wheels are predominantly used in the machining of glass
including windows for the automotive and construction industries, decorative
glass such as lead crystal and also optical and electronic components. The diamond
types most commonly used in metal bond tools are relatively high in impact strength
with regular crystal shapes. Figure 40 shows a typical high strength diamond abra-
sive. Sintered metal bonds commonly based on bronze are much more abrasion
resistant than resin bonds, and therefore better suited to machining workpiece
material which produce highly abrasive swarf, glass being one example. The list
below summarizes the main application areas for metal bond diamond wheels.
- grinding glass
~ machining refractory materials
- machining semi-precious stones
- grinding pcD
- grinding ceramics
bond systems is that a degree of porosity can be built into the bond. Resin and metal
bonds are effectively 100% dense and clearance for the swarf produced during
grinding is provided by the protrusion of the abrasive particle above the surface
of the bond. The pores, or voids, in porous vitrified bonds provide extra clearance
volume for the removal of swarf out of the grinding zone and also the transport of
coolant into it.
In general, all diamond grinding wheels operate at average peripheral speeds in
the range 2&30ms-'. If a material is being ground dry, without coolant, the
wheel speeds tend to be lower, and also the difficult materials such as cermets and
pcD tend to be ground at lower speeds to avoid the excess generation of heat. A
small number of specialized applications use much higher wheel speeds, one example
being the sawing of glass tubes which, providing the machine is specially designed
for the purpose, with particular attention to the coolant application, can be cut
much more quickly with a wheel rotating at speeds up to 60-80 ms-' .
1.5.2.2.3 Grinding with Cubic Boron Nitride Abrasives Cubic boron nitride abra-
sives are used to machine hard ferrous materials. As with diamond, all four of the
main bond types are used and, during the early days of its use, in the 1970s and early
1980s, resin bond tools were the majority. Today, vitrified bond tools are the most
commonly used type, particularly in high volume production operations, with resin,
electroplated, and metal bonds failing in second, third and fourth place respectively.
A reason for the growth in the use of vitrified bond cBN tools is that this type of
bond system offers a good compromise between high material removal rates, ease
of use, and low wear rate resulting in the ability to hold tight tolerances.
The growth in the use of cBN since the early 1970s has been due to its penetration
into volume production operations, for example in the automotive and aerospace
industries. Prior to that time, cBN wheels were used primarily in the toolroom for
sharpening and re-sharpening high speed steel drills, milling cutters and other tool-
ing and also in small wheels for precision jig grinding (Fig. 43).
The economics of cBN grinding have been improved by the adoption of creep feed
grinding techniques and also the use of higher wheel peripheral speeds. Both of these
developments required the development of suitable grinding machines.
Creep feed (or deep) grinding techniques were originally developed using con-
ventional wheels and the principle of the technique is shown in Fig. 44. In 'normal'
or reciprocating grinding mode, a shallow depth of cut is used together with a relatively
high table speed to achieve the required material removal rate. Under these conditions,
the arc of cut, and hence the area of wheel in contact with the workpiece at any one
time, is very small. In creep feed grinding, a significantly deeper cut is made at a
slow table speed. Under these conditions, the arc of cut is much longer and hence
the number of particles instantaneously in contact with the workpiece is much greater.
The specific rate of material removal is a function of the depth of cut and the feed rate
of the table. In general, when comparing the two modes of grinding, and assuming
equal material removal rates, the main differences for creep feed grinding are that
- the number of particles in contact is higher
- the load on each particle is lower
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
538 I Diuniontl Matrrials mu' their. App1iuition.s
~ the total normal load in the machine is greater (due to the greater number of
particles)
- the life of the grinding wheel is longer
~ surface finish produced is normally improved.
Wheel Speed
Figure 45. Effect of wheel speed on grinding wheel performance.
T
cost
ever increasing range of size and also other characteristics such as strength. Devel-
opments by tool manufacturers and machine builders improved the economics of
machining hard natural stone types such as the granites, making their processing
more economical. There was also a movement towards the use of granite from
marble in the construction industry. One of the reasons for this was that it was
recognized that granite was far more durable than marble in respect of both wear
if used on floors, and also more resistant to chemical attack and discoloration in
urban environments if used as a cladding on the exterior of prestige buildings. Its
use, therefore, became more widespread and the quantity of diamond abrasives
used to process it has continued to increase.
The types of diamond abrasive used for cutting stone and concrcte are normally
those with high impact strength and of a relatively coarse size.
The normal production route for stone products includes:
- extraction from the quarry in the form of large blocks
- squaring and stabbing the rough blocks in the stone yard
- cutting the blocks into either cladding or tiles
- surfacing and polishing the finished items.
In all these operations, diamond tools are widely used. For quarrying, diamond wire
sawing is established as the standard method of extraction for soft rock (e.g. marble)
and the technology for hard rock is being established. Development of diamond
wire began in the early 1970s but it came into common use only in the late 1980s.
It consists of a high tensile steel wire fitted with beads containing the diamond abra-
sive, typically 40 beads per metre. The overall diameter of the beads is 6-12mm
(Fig. 50). Two intersecting holes drilled in the quarry face and the wire, which
can be up to 500m long, is fed through them, joined to form an endless loop and
then driven at speeds of 20-50msC' while tension is applied, thus cutting a slot
through the rock and enabling a block to be cut (Fig. 51). The resulting block suffers
much less damage compared with using explosive techniques and this, in turn, leads
to less waste in subsequent processing. A major problem encountered with early
designs was erosion of the wire by the fine particles of stone produced by the
sawing, leading to breakage. This is now alleviated by protecting the wire with
rubber or plastic sleeving bonded to the wire between the beads.
When quarrying granite, a relatively low wire speed is typically used, 20-30 ms-'
and, for a medium grade of material, a block measuring 30m long x 10m
wide x 5 m deep can be separated from the surrounding material in less than 12 h.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Industrial Applications of Diamond and cBN 543
slabs in turn are ground to final thickness and sawn to final size, both operations
using diamond tools. The final stage in the production of tiles is polishing which
is normally achieved using automatic multistage machines (Fig. 54). In this final
process, many of the stages are now completed using diamond tools, the early
stages in polishing using metal bond tools with high strength diamond to achieve the
high material removal rate required and the later stages using resin bond tools incor-
porating lower strength, more friable, grades of diamond particle. Figure 55 shows a
selection of polishing tools.
Another major application area for diamond abrasive tools is in the construction,
refurbishment and demolition industries. In many instances, concrete, masonry and
also roadways need to be cut or drilled and diamond tools provide an economic
solution, their advantages being speed, accuracy and also reduced damage to the
surrounding structure compared to other methods using percussive hammers or
thermal lances. Circular saw blades are used in a wide variety of operations and
metal bond tools are most commonly used. On roadways, both asphalt and con-
crete, diamond saws are used to cut trenches during repairs and also to provide
slots to embed sensor wires for traffic management systems (e.g. traffic signals)
(Fig. 56). Concrete roadways, runways and pedestrian walkways can, under wet
conditions, become slippery and diamond blades, up to 100 spaced along a spindle,
are used to cut a series of parallel grooves in the concrete to aid water drainage and
improve skid resistance and hence safety in critical areas (Fig. 57).
Saw blades are used to cut openings in walls and floors during both building and
refurbishment and can also be used during controlled demolition. The advantage of
a diamond saw, compared to a saw of conventional abrasive, is its ability to cut both
concrete, masonry and steel reinforcement (Fig. 58).
The majority of all diamond sawing of concrete is carried out with use of water as
a coolant. This has the combined effect of controlling the temperatures at the cutting
edge and also removing the swarf produced by the saw. Some modern blades are
designed to run dry (Fig. 59). These are now widely used on construction sites for
cutting tiles and the minor correction work which needs to be carried out.
Wire saw machines, similar in principle to those used for cutting stone, are also
used widely for large demolition operations. When using a circular saw, there is a
limit on the depth of cut which can be made due to the presence of the drive spindle
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
in the centre of the blade. This is a problem if deep sections have to be cut. The wire
saw does not have any such limitation and concrete, heavily reinforced with steel as
used in large structures such as road bridges, can be cut with modern wire saw
machines. Figure 60 shows such an operation.
Drilling stone with diamond bits is now widely used and concrete, both plain and
steel reinforced, brickwork and other masonry have been drilled for many years
using diamond core bits (Fig. 61). These tools can range in size from approximately
20mm to in excess of a meter. They are all metal bonded tools and were until
recently designed to operate with water as a coolant. In some areas, the use of
water can be inconvenient and a relatively recent development has been the
design of core bits which do not need to be cooled with water. These dry drill bits
are widely used on construction sites to produce accurate holes in walls for electrical,
plumbing and other services (Fig. 62).
element and also as a wear resistant material. As a cutting tool material, the general
rules which apply to abrasives also apply to pcD and pcBN.
~ pcD is for use in non-ferrous applications.
~ pcBN is for use on ferrous workpieces.
Cemented
-
100
-
10
Carbon tool
- steel
0.1
I I I I I I
- mineral materials
The automotive and aerospace industries provide a large number of applications for
pcD, particularly with the movement to more widespread use of advanced alloys
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
I .5 Industrial Applications of Diamond and cBN 55 1
I PCD
Ideal cutting
tool material
Engine components
Gearbox components
Transmission components
Braking systems
Interior fittings
Body panels
which, dependent upon the alloy, can be very abrasive. Similar classes of material
are used in the manufacture of gearbox casings and these can also be machined
more efficiently with pcD than with any other cutting tool materials (Fig. 67).
The cylinder bores of aluminum-silicon alloy engine blocks are also typically
machined with pcD. Figure 68 shows such a component from a V-12 engine
being finish bored with pcD.
A material originally used in racing cars and motorcycles is a metal matrix com-
posite (MMC) made of aluminum reinforced with silicon carbide particles. Its light
weight and high strength made it ideal for use as a disc brake rotor and also other
components. The presence of the silicon carbide makes it a difficult material to
machine, pcD being the best cutting tool material. Figure 69 shows a typical
MMC machining operation on a brake disc rotor. Machining of reinforced plastics
is also a major application for pcD tooling. Figure 70 shows part of a glass
reinforced plastic (GFRP) truck body section which is machined with a pcD router.
A relatively recent application area for pcD has been in the woodworking
industry. Wood composites such as medium density fibreboard (MDF), chipboard,
laminated boards with plastic coatings, for use in furniture and on floorings, and
also composites, used in wall claddings and in ceilings, all tend to be highly abrasive
Figure 72. pcD is typically used on medium density fibreboard (MDF), chipboard, and abrasive
hardwoods.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Industrial Applications of Diamond und cBN 555
Speed = 1000 d m i n
Depth = 10mm
Feed ~ 0 1 mdtooth
0.15 Infeed = 2 rnmlpass
f
c_tc;;3p 3';
Figure 73. Wear rates of HSS (high speed steel), WC (tungsten carbide) and pcD routers machining
chipboard.
turn, means that pcD drill bits have the potential to drill much faster than conven-
tional bits and, in addition, the high abrasion resistance of pcD results in the cutters
remaining sharp and hence extending considerably the life of the drill bit.
Drilling for oil and gas is a highly specialized subject: pcD cutters were introduced
to this application area in the early 1980s and, since that time, the design of the drill
bits and the pcD cutting elements themselves have resulted in ever-increasing perfor-
mance.
A factor which limited the early pcD oil and gas exploration bits was premature
failure, either of the tungsten carbide supporting the pcD or of the braze attaching
the cutter to the bit body. Today, most of these difficulties have been overcome, and
pcD bits offer high penetration rates and long bit lives when drilling medium hard
abrasive rock formations.
pcD drill bits are also used in mining, although the volume is considerably less
than that in the oil and gas exploration industry. One example is in methane drai-
nage, where holes are drilled into the rock surrounding the coal seam and the
methane pumped out to prevent it seeping into the mining area and hence causing
a fire hazard. Also, in underground mining, holes have to be drilled in large quan-
tities for roof bolts to be inserted to provide roof support, and blast holes have to be
drilled into which explosives are packed. Figure 74 shows a selection of typical pcD
mining bits.
used as a die to reduce fine wire (Fig. 75). pcD wire drawing dies are most widely
used to draw non-ferrous wires such as copper where the life, compared to cemented
tungsten carbide, is up to 1000 times longer.
A more recent application field has been the use of pcD as a wear part, such as
a bearing, a workpiece support rest in a machining operation, a nozzle for
abrasive fluids and in high precision gauges. Figure 76 shows a pcD used as a
workpiece support rest for an automotive engine component and Fig. 77 shows
precision automatic gauging equipment fitted with pcD tipped measuring fingers.
In such cases, the high wear resistance of pcD leads to improved accuracy being
achieved.
Figure 76. SYNDITE pcD wear parts support a crankshaft during machining
hardness at these high temperatures, efficient machining can be conducted. The high
temperatures generated are limited to the cutting zone only, providing the machin-
ing conditions are selected correctly. The bulk temperature of the workpiece remains
low and coolant is only required in a minority of cases. pcBN tool materials fall lar-
gely into two classes:
- high cBN content (>go% by volume cBN) in a ceramic binder
- low cBN content ( ~ 6 0 %by volume cBN) in a ceramic binder
High cBN content materials are used in rough machining applications, whereas
lower cBN content tools tend to perform better in finishing applications where shal-
low depths of cut are being taken and a good surface finish is required.
Abrasion resistant cast-irons, for example Ni-hard, are used to manufacture
pumps for pumping slurry and gravel in the quarrying, mining and associated
heavy industries. These materials are difficult to machine with conventional tooling
materials but can be machined easily with a high cBN content tool material such an
Amborite. Steel rolling mill-rolls, which are normally ground to final size, can also
be machined with pcBN, the advantage being higher production rates and easier
swarf disposal since most rough machining operations using pcBN are conducted
dry, without coolant (Fig. 79). pcBN can also be used in milling, and Fig. 80
shows an example of a milling operation on a machine tool slideway.
In the automotive industry, high content pcBN is used to machine a wide range of
cast iron components such as brake discs (Fig. 81) and cast iron cylinder blocks
(Fig. 82).
Using low-content pcBN tools for finishing previous operations often replaces a
grinding operation. In many cases, a hardened steel component was finish ground
because its hardness meant that it was not possible to machine it in any other
way. Many such components may now be turned to final tolerance and surface
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1.5 Industrial Applications of Diamond and cBN 559
finish using suitable pcBN tooling. Figure 83 shows a gear component being finished
in this way.
Low content pcBN can also be used in fine milling. Figure 84 shows a n automo-
tive gearbox component being finish milled with a low content pcBN (DBCSO).
hardness and abrasion resistance were recognized, but it was only possible to use
relatively large pieces of diamond, several millimeters in size. More recently, other
outstanding properties have been recognized, in particular the high degree of optical
transparency of type I1 diamonds and the very high thermal conductivity of (the
relatively rare) type HA diamond. In recent years, the technology to synthesize
economically synthetic single crystal diamond of several millimeters in size has been
developed and this has contributed to the availability of relatively large diamonds
for industrial use.
Single diamond crystals are currently used in a wide variety of industrial appli-
cations, albeit some highly specialized, and these range from rock drilling to
sophisticated applications in micro-surgery and electronics.
One advantage of synthetic crystals is that they can be synthesized and processed
to specific required sizes and also crystallographic orientations chosen according to
the application (Fig. 85).
i/--g
MONODITE \-Facet
MONODITE Triangle
Cube
Whole
Stone
4
-
c
MONODRESS
MONODITE Triangle
In the medical field, diamond blades are used in surgical scalpels (Fig. 88) and in
microtome knives (Fig. 89) to prepare tissue samples for microscopic examination.
Diamond knives are also used for cleaving fibre optics.
The types of equipment in which diamond crystals are used include anvils for
high-pressure research (in excess of a million atmospheres or Megabar) (Fig. 90),
heatsinks, bearings for precision chronometers and other ultra-sensitive electronic
meters, styli for audio equipment and surface measuring instruments, hardness-
testing indenters (Fig. 91) and distance stops on machine tools.
In addition, diamond is used for metal-forming applications, such as burnishing,
engine turning, and the ruling of diffraction gratings, where little or no material is
actually removed. The most important of such forming operations is wire drawing.
In top-quality wire-drawing dies, the ultimate wear resistance and surface polish rely
on the particular qualities of diamond crystals. Such dies enable large quantities of
very fine wire (down to 10pm in diameter) to be produced with the required
precision (Fig. 92).
Acknowledgments
The authors are indebted to Dr I. Sigalas and Dr R. J. Caveney for the overall
editing of the article, and to Mrs Chancellor Teffo for her very competent secretarial
help.
References
1. S. Tennant, Phil Trans. R. Soc. 1797,97, 591.
2. H . Liander and E. Lundblad, Arkiv Kemi, 1969, 16, 139-149.
3. F. P. Bundy, H. T. Hall, H. M. Strong, and R. H. Wentorf, Nature, 1955,176 (No. 4471), 51-55.
4. F. P. Bundy, Nature (London), 1973, 241, 116-1 18.
5. G. Davies, Diamond, Adam Hilger, Bristol, 1984, Chapter 4.
6. H. M. Strong, Am. J. Phys. 1989, 57, 794-802.
7. J. W. Harris, Diamond geology, in The Properties of Natural and Synthetic Diamond, J. E.
Field (Ed.), Academic Press, London, 1992, pp. 342-393.
8. S. E. Kesson and A. E. Ringwood, Chemical Geology, 1989, 78, 97-1 18.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 567
81. J. E. Shigley, E. Fritsch, C. M. Stockton, J. I. Koivula, C. W. Fryer, and R. E. Kane, Gems and
Gemology, 1986,22, 192-208.
82. J. E. Shigley, E. Fritsch, C. M. Stockton, J. I. Koivula, C. W. Fryer, R. E. Kane, D. R.
Hargett, and C. W. Welch, Gems and Gemology, 1987, 23, 187-206.
83. M. L. T. Rooney, C. M . Welbourn, J. E. Shigley, E. Fritsch, and I. Reinitz, Gems and
Gemology, 1993, 29. 3 8 4 5 .
84a. J. E. Shigley and E. Fritsch, J . Cryst. Growth, 1993, 128, 425428.
84b. J. E. Shigley, E. Fritsch, I. Reinitz, and T. M. Moses, Gems andGemology, 1995,31, 256-264.
85. C. M. Welbourn, M. Cooper, and P. M. Spear, Gems and Gemology, 1996, 32(3), 3 156-169.
86. M. Kamo, Y. Sato, S. Matsumoto, and N. Setaka, J . Cryst. Growth, 1983, 62(3), 642-644.
87. J. C. Angus, H. A. Will, and H. S. Stanko, J . Appl. Phys., 1968, 39, 2915.
88. W. G. Eversole, US Patent 3, 030, 187 and US Patent 3, 030, 188 1962.
89. K. V. Ravi, Mater. Sci. Eng., 1993 B19. 203-227.
90. R. Bennan, Solid State Commun., 1996, 99(1), 35-31.
91. K. T. Jacob, Solid State Commun., 1995, 94(9), 763-765.
92. C. J. Brannon and S. L. McCollum, in New Diamond Science and Technology, R. Messier et al.
(Eds), Materials Research Society, Pittsburgh, PA, USA, 1991, p. 117.
93. A. Tsuzuki, S.-I. Hirano, and S. Naka, J . Muter. Sci., 1985, 20, 2260-2264.
94. H. Uchikawa, H. Hagiwara, and K Nakamura, in Science and Technology of New Diamond, S.
Saito et al. (Eds), Terra Scientific Publishing Co., 1990. 227.
95. Hawtin, J. B, Lewis, N. Moul, and R. H. Philips, Roy Soc. Phil. Trans. Ser A , , 1966,261(116),
67-95.
96. L. F. Vereschagin, E. N. Yakoviev, L. M. Buchnev, and B. K. Dymov, Teplofizika Vysokikh
Temperafur, 1977, 1S(2), 316-321.
97. Y. Bar-Yam and T. D. Moustakas, Nature, 1989, 32, 786787.
98. A. A. Abramzon and A. A. Novozhenets, Russ. J . Phys. Chem., 1992, 66(8), 122c1222.
99. J. Abrahamson, Carbon, 1973, 11, 337-362.
100. J. 0. Hansen, R. G. Copperthwaite, T. E. Derry, and J. M. Pratt, J. Colloid Interface Sci.,
1989, 130(2), 347-358.
101. R. A. Munson, Carbon, 1967, 5, 471474.
102. J. A. Nuth, Astrophys. Space Sci., 1987, 139(1), 103-109.
103. P. Badziag, W. S. Verwoerd, W. P. Ellis, and N. R. Greiner, Nature, 1990,343(6255), 244245.
104. E. Anders and E. Zimmer, M eteoritics, 1993, 28, 49c514.
105. A. N. Nesterov, Y. I. Merezhko, and V. V. Chernikov, Run. J . Phys. Chem., 1983,57(2), 307-308.
106. D. V. Fedoseev, ColloidJ. USSR, 1978,40(2), 341-342.
107. D. V. Fedoseev, B. V. Deryagin I. G. Varshavskaya, and A. V. Laverentev, Prog. Surf. Sci.,
1994,4S, 84-87.
108. F. Banhart and P. M. Ajayan, Nature, 1996. 382,433435.
109. B. Wei, J. Zhang, J. Liang W. Liu Z Gao, and D. Wu, J . Muter. Sci. Lett., 1997, 16,402403.
110. H. Moissan, C. R. Acad. Sci. Paris, 1894, 118, 32G32.5.
111. J. B. Hannay, Nature, 1880, 22, 255-256.
112. Y. G . Gogotsi K. G . Nickel, and P. Kofstad, J . Muter. Chern. 1995, S(12), 2313-2314.
113. M. A. Cappeli, US Patent 5,417,953, 1995.
114. J-Y Wang, Y-Z Wan, D-W Zhang, and Z-J Liu, J . Muter. Res., 1997, 12(12), 3250-3253.
115. F. Davanloo, E. M. Juengerman, D. R. Jander, T. J. Lee, and C. B Collins, J . Muter. Res.,
1990, S(l l), 2398-2404.
116. R. Roy, H. S. Dewan, and P. Ravindranathan, Muter. Res. Bull., . 1993, 28, 861-866.
117. R. Roy, K. A. Cherian: J. P Cheng, A. Badzian, C. Langlade, H. Dewan, and W. Drawl, lnnov.
Muter. Res., 1996, 1(1), 165-87.
118. P. A. Molian and A. Waschek, J . Muter. Sci.,1993, 28, 1773.
119. M . Horio, A. Saito, K. Unou, H. Nakazono, N . Shibuya, S. Shima, and A. Kosaka, Chem.
Eng. Sci., 1996, Sl(1 l), 3033-~3038.
120. Y. Namba, J . Vac. Sci. Technol. A , , 1992, 10(5), 3368-3370.
121. M. Takaya, Japanese Patent Hei 6-25896, 1994.
122. T. Suzuki, T. Ishihara, T. Yamazaki, and S. Wada, Jpn J . Appl. Phys., 1997,36 part 2, No. 4B,
L50kL506.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
570 1 Diamond Materials and their Applications
165. M. Wakatsuki and K . Ichinose, 4th Internutionul Conference on High Pressure, Kyoto, 1974,
441445.
166. F. R. Corrigan, 6th AIRAPT Conference on High Pressure Science and Technology,' 1979, 1,
994-999.
167. 0. Fukunaga and M. Akaishi, High Press. Res., 1990, 5, 911-913.
168. J. M. Leger, J . Haines, and B. Blanzat, J . Muter. Sci. Lett., 1994, 13, 1688-1690.
169. M. C. Payne, M. T. Teter, D. C. Alan, T. A. Arias, and J. D. Ioannopoulos, Rev. Mod. Phys.,
1992,64, 1045.
170. A. Y. Liu and M. L. Cohen, Science, 1989, 245, 1981.
171. M. L. Cohen, J . Hard Mater., 1991, 2, 13.
172. P. K. Lou, M. L. Cohen, and G. Martinez, A m . Phys. Soc., 1987, 35, 919&9194.
173. R. Hoffmann, T. Hughbanks, M. Kerttesz, andP. H. Bird,J. A m . Chem. Soc., 1983,105,4831.
174. M. A. Tamor and K . C. Hass, J . Muter. Res., 1990, 5, 2273.
175. A. Y. Liu, M. L. Cohen, K. C. Hass and M. A. Tamor, Phys. Rev. B, 1991,43, 6742.
176. F. Diedrich and Y. Rubin, Angew. Chem. Int. Ed. Engl., 1992, 31, 1101.
177. H. Alberts. T. Sekine, H. Kanda, Y. Bando, and K. Hojou, J . Mater. Sci.Lett., 1990,9, 1376.
178. A . Y. Liu and M. L. Cohen, Phys. Rev. B., 1990,41, 10727-10734.
179. J. V. Badding, Adv. Mater., 1997, 9, 877-886.
180. A. Y. Liu and R. M. Wentzcovitch, Phys. Rev., 1994, 1350, 10362.
181. R. M. Wentzcovitch and J. L. Martins, SolidState Commun., 1991, 78, 831.
182. R. M. Wentzcovitch, Phys. Rev., 1991, B44, 2358.
183. D. M. Teter and R. J. Henley, Science, 1996, 271, 53-55.
184. D. Li, S. Lopez, Y. W. Chung, M. S. Wong. and W. D. Sproul, J . Vuc.Sci. TecAnol., 1995,13,
1063.
185. T. A. Yeh, C. L. Lin, J. M. Sivertsen, and J. H. Judy, IEEE Trans. Laga., 1991, 27, 5163.
186. J. J. Cuomo, P. A. Leary, D. Yu, W. Reuter, and M. Frisch, J . Vac. Sci. Technol., 1979, 16,
299.
187. H. Sjostrom, W. Lanford, B. Hjovarso, K. Xing, and J. E. Sundgren, J . Muter. Res., 1996, 11,
981.
188. C. Niu, Y. Z. Lu. and C. M. Lieber, Science, 1993, 261, 334.
189. K. Ogata, J. F. D. Chubaci, and F. Fujimoto, J . Appl. Phys., 1994, 76, 3791.
190. J. Schwan, W. Dworschak, K . Jung, and H. Erhardt, Diamond Rel. Muter.. 1994,3, 1034.
191. T. Y. Yen and C. P. Chon, Appl. Phys. Lett., 1995, 67, 2801.
192. Y. Guo and W. Goddard 111, Chem. Phys. Lett., 1995,237,12-76; J. E. Lowther, Phys. Rev. B ,
1998,57, 57245727.
193. D. M. Bhusari, C. K. Chen, K. H. Chen, T. J. Chuang, L. C. Chen, and M. C. Liu, J . Muter.
Res., 1997, 12, 322-325.
194. Z. J. Zhang, S. Fan, J. Huang, and C. M. Lieber, Appl. Phys. Lett., 1996, 68(19), 2639-2641.
195. F. Fujimoto and K. Ogata, Jpn. J . Appl. Phys., 1993, 32, L. 420.
196. D. Liu, Y. W. Chung, M. S. Wong, and W. D. Sproul, J . Appl. Phys., 1993 74, 219.
197. Z. M. Ran, Y. C. Du, Z. F. Ying, Y. X. Qiu, X. X. Xiong, J. D. Wu, and F. M. Li, Appl. Phys.
Lett., 1994, 65, 1361.
198. C. Niu, Y. Z. Lu, and C. M. Lieber, Science, 1993, 261, 334.
199. S. Kumai and T. L. Tansley, Solid State Commun., 1993, 88, 803.
200. D. Norton, K. J. Boyd, A/H. Al-Bayari, S. S. Todorov, and J. W. Rabalais, Phys. Rev. Lett.,
1994, 73, 118.
201. K. M. Yu, M. L. Cohen, E. E. Hailer, W. L. Hausen, A. Y. Liu, andI. C. Wu, Phys. Rev., 1994,
B49, 5034.
202. J. P. Riviera, D. Texier, J. Delafond, M. Jaouen, E. L. Mathe, and J. Chanmond, Muter. Lett.,
1995, 22, 115.
203. X. W. Su, H. W. Song, F. Z. Cui, W. Z. Li, and H. D. Li, Surj: Coat. Technol., 1996,84,388-391.
204. Y. Zhang, H. Li, and Q. Xue, Muter. Res. Soc., 1996, 317-322.
205. C. Yan, Liping Guo, and E. G. Wang, Phil. Mag. Lett., 1997, 75, 155-162.
206. K. Yanamoto, Y. Koga, K. Yase, S. Fuziwara, and M. Kubota, Jpn. J . Appl. Phys., 1997,36,
L23CL233.
207. E. G. Wang, Chen Yan, and Guo Liping, Physica Scripta, 1997, T69, 108-114
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
572 1 Diamond Materials and their Applications
208. M. J. Yacaman, J. J. M. Gil, and M. Sarikaya, Muter. Chem. Phys., 1997,47, 109-117.
209. D. W. Wu, W. Fan, H. X. Guo, M. B. He, X. 0. Meng, and X. J. Fan, Solid State Commun.,
1997, 103, 193-196.
210. S. Xu, Han-Shi Ki, S. Lee, and Yin-Au Li, IEEE, 1997,213-214.
21 1. Xiao-Ming He, Li Shu, Wen-Zhi Li, and Heng-De-Li, J . Muter. Res., 1997, 12, 1595-1602.
212. Wu Dawei, Fu Dejan, Guo Huaixi, Zhang Zhihong, Meng Xianquan, and Fan Xiangiun,
Phys. Rev. B, 1997,56,49494954.
213. Chen, L. L., D. M. Bhusari, C. Y. Yang, K. H. Chen, T. J. Chuang, M. C. Lin, C. K. Chen,
and Y. F. Huang, Thin Solid Films,1997, 303, 66-75.
214. J. Badding, Adv. Mater., 1997, 9, 877.
215. K. Ploog, P. Rauh, W. Stoeger, and H. Schmidt, J . Cryst. Growth, 1972, 13/14. 350.
216. K. Ploog, H. Schmidt, E. Amberger, G. Will, and K. H. Kossobutzki, J . Less-Common Met.,
1972, 29, 161.
217. H. Saitoh, K. Yoshida, and W. A. Yarbrough, J . Muter. Res., 1993, 8, 8.
218. L. Vel, G. Demazeau and J. Etournea, Mat. Sci. Eng., 1991, BID, 149.
219. A. R. Badzian, T. Niemyski, and E. Ockusmik, Proc. 3rd lnt. Con$ on CVD, F. A. Glaski
(Ed.), The Electrochemical Society, Penington, NJ, 1972, p. 747.
220. K. Montasser, S. Hattori, and S. Morita, Thin Solid Films,1984, 117, 311.
221. J. Loeffier, F. Steinbach, J. Bill, J. Mayer, and F. Aidinger, Z. Metailkol, 1996, 87, 170.
222. D. Hegermann R. Riedel, W. Dresier, C. Oehr, B. Schindler, and H. Brunnen, Chem. Vap.
Depos., 1997, 3, 257
223. A. Badzian, Appl. Phys. Lett., 1988, 53, 2494.
224. P. F. McMillan, W. T. Petuskey, H. Hubert, K. J. Kingma, L. A. Garvie, A. Grzechnik, and A.
Chizmeshya, Adriutico Research Conference, Miramare, Trieste, Italy, July 1997.
225. H. Itoh, I. Maekawa, R. Yanamoto, and H. Iwahara, Rev. High Press. Sci. Technol., 1998,7,
986-988.
226. C. Ellison-Hayashi, M. Zandi, F. J. Csillag, and Shih-Yee Kuo, US Patent 5, 135, 892, 19912.
227. C. Ellison-Hayashi, M. Zandi, D. K. Shetty, P. Kuo, R. Yeckley, and F. Csillag, US Patent
5330,937, 1994.
228. J. M. Leger, J. Haines, M. Schmidt, J. P. Petitet, A. S. Pereira, and J. A. H. da Jomada, Nature,
1996,383,401.
229. Federation Europkene des Fabricants de Produits Abrasifs (FEPA), Standard for Superabra-
sives Grain Sizes, 2nd Edn. September, 1997.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
2.1 Introduction
The ability of Chemical Vapor Deposition (CVD) technology to synthesize diamond
as large plates of high and controlled purity [l-51 is enabling a host of new
applications in science and technology that were hitherto hampered by the practical
difficulties inherent in the use of natural diamond. The basic principles of CVD for
the synthesis of polycrystalline diamond have been discussed in the chapter by
Klages et al. [6]. The aim of this chapter is to discuss a selected range of applications
to illustrate how CVD diamond is emerging as a key (and in some cases unique)
engineering material.
The impact that CVD diamond is starting to have on some areas of modern
technology is in some way similar to the impact that the invention of the high
pressure and high temperature (HPHT) techniques for the synthesis of diamond
had on the abrasive industry in the early sixties [7]. At that time the industrial
implementation of HPHT technology gave a renewed impetus to an existing but
relatively small diamond abrasive industry. HPHT synthesised diamonds have
been the foundation of what is now a relatively mature industry in which the
extreme hardness and wear resistance of diamond are exploited for numerous
abrasive and cutting tool applications [S].
The potential importance of diamond in technology follows from the wide range
of technically desirable thermomechanical, optical, and electronic properties
exhibited by diamond [9-131. In addition to its extreme hardness and exceptional
wear resistance, diamond is the stiffest known material, has a broad transmission
spectral range (from the ultraviolet to the far infrared and extending to microwave
frequencies), has the highest room-temperature thermal conductivity, one of the
lowest thermal expansion coefficients and is radiation hard and chemically inert
to all acid and base reagents.
What makes diamond so attractive, and in many cases unique, is not only the
extreme value of some of the above properties such as hardness, thermal
conductivity or the Young modulus, but the combination of two or more of
these properties in a high performance product, as will be discussed later in this
chapter.
The enormous interest that CVD diamond has generated since the mid eighties
has been based on the perception that this new diamond synthesis technology
could overcome all the limitations previously encountered with the availability of
large size natural diamond specimens [ 14,151.As was the case in the abrasives indus-
try, the use of natural diamond in other fields of technology has been relatively small
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
514 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
and restricted to niche applications in optics [ 161, the thermal management of small
electronic devices (such as laser diodes or impatt oscillator diodes) [17,18], and
others [19,20]. For these types of applications only the relatively rare high-purity
Type IIa form of natural diamond could be used. The most abundant form of
natural diamond, Type Ia, contains relatively large concentrations of nitrogen
which degrade properties such as optical transmission [21],thermal conductivity
[22],and electronic transport [23]. Another major difficulty in the use of natural
diamond is the stringent limitation to the size of specimens that are practically avail-
able. Most natural Type IIa windows sold commercially are below 5 mm in diameter
and typical heat sinks are smaller than 1 x 1 mm [16, 181. The cost of natural single
crystal windows of sizes in excess of 19 mm, if available, would be outside the budget
range of most practical applications [16]. The largest and the most spectacular nat-
ural diamond product ever made was a window, 18.2mm in diameter and 2.8 mm
thick, used for the main pressure cell of the pioneer Venus probe launched on
August 1978 [24]. Purity and size are also serious limitations in the use of HPHT
synthetic diamond in these types of applications because of the high intrinsic
single substitutional nitrogen concentration and the practical maximum size of typi-
cal HPHT diamond crystals [7].
It is within this perspective that efforts in recent years have been directed to the
development of a CVD diamond technology that allows the routine manufacture
of large plates of diamond of high and controlled purity. Some of this effort has
met with considerable success [l-5,251. In the following sections of this chapter
we will briefly review the properties of various types of CVD diamond materials
that have been developed to address specific applications. We will illustrate through
a few selected examples how the current ability to manufacture CVD diamond
(including large area plates and three-dimensional shapes) as a reliable and robust
engineering material is enabling the use of diamond in an increasing range of tech-
nically demanding applications.
Figure 1. Optical grade CVD diamond window, 100mm diameter, 0.7mm thick
different values of fracture strength, intrinsic bulk absorption, and total forward
scatter [26,27].
The majority of the properties described in this section refer to the type of CVD
diamond that is currently available as a commercial product sold under the trade
name of DIAFILM [28]. With sufficient control of synthesis conditions, the extra
degree of freedom conferred by the polycrystalline structure has been used to
formulate different 'grades' of CVD diamond in which properties are optimized
for specific applications. Optical grades, for instance, exhibit the highest optical
transparency and can be made routinely in discs up to 100 mm in diameter. In con-
trast mechanical grades are dark in color but have a higher mechanical strength and
can be made in plates up to 160mm in diameter. Various thermal grades have been
developed for thermal management applications which offer a compromise between
good thermal conductivity and cost. Most of the applications currently pursued
with CVD diamond are relatively new and in certain cases the quality of the
corresponding grade is subject to continuous development in which factors such
as cost and performance are gradually optimized.
Figure 1 shows an example of a l00mm diameter, 0.7mm thick, optical grade
CVD diamond window which has been polished on both sides to an optical
finish. The as-grown grain morphologies of optical (2 mm thick) and mechanical
(0.5 mm thick) grade samples are shown in Fig. 2. Figure 3 shows the Raman spec-
trum of an optical grade CVD diamond sample measured by a Renishaw instrument
with an excitation wavelength of 633nm. The main Raman line is centered at
1332cm-' and the full-width-half-maximum (FWHM) line width is approximately
3.0cmP', which compares well with typical values of 2.8 cm-' to 3.1 cm-' observed
in high-quality single-crystal Type IIa natural diamond specimens measured with
the same instrument.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
576 2 Applications of' Diamond Synthesized by Chemical Vapor Deposition
Figure 2. (a) SEM of the as-grown morphology of an optical grade DIAFILM sample 2mm thick.
(b) SEM of the as-grown morphology of a mechanical grade DIAFILM layer 0.5 mm thick.
10,000
.-yE
n
8,000
4Q 6,000
a
Y
.w 4,000
C
3
0
0 2,000
n
(200 1,300 1,400 1,500 1,600 1,700
Raman shift from 633 nm (cm-1)
Figure 3. Raman spectrum of an optical grade CVD diamond specimen measured with a Renishaw
instrument with an excitation wavelength of 633 nm.
Beyond the intrinsic, multi-phonon absorption bands (between 4000 cm-' and
1500 cm-') [30], the transmission of the CVD diamond specimen reaches the maxi-
mum reflectivity limited value (Fig. 5) of 71.4% (using a refractive index value of
2.375 [32,3]) and is identical, within experimental accuracy, to that of high-quality
single crystal Type IIa samples measured with the same instrument.
80
Single crystal Type Ha, 0.5 mm thick
I
60
.-0
C
.-
v)
u)
40
E
s
20
n
u
500 1,000 1,500 2,000 2,500
Wavelength (nm)
Figure 4. Transmission spectrum for an optical grade CVD diamond window 1.0mm thick in the
UV-visible-near IR spectral range compared to that of a high quality natural Type IIa window
0.5 mm thick.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
518 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
---------- 1
.-
v) Optical grade DlAFlLM
v) 1.0 mm thick sample
71.4% transmission
L - 1 ~
0
3,000 2,000 1,000 500
Wavenumber cm-I
Figure 5. Transmission spectrum for an optical grade CVD diamond window 1.Omm thick in the
infrared spectral range.
Transmission measurements have been recently reported [33] extending the range
of the spectrum of Fig. 5 to much longer IR wavelengths (to 20 cm-' or 500 pm)
showing that there are no measurable absorption features within that range.
The intrinsic absorption at 10.6 pm wavelength of optical grade samples measured
by calorimetry is in the range 0.10-.029 cm-' [27,34]. It is thought that some of these
values may be affected by surface absorption. This aspect is currently being investi-
gated. As will be discussed below, values of intrinsic absorption below 0.1 cm-' are
acceptable for most practical applications, due to the very high thermal conductivity
of diamond, coupled with the low change in refractive index with temperature.
The low angle scattering of optical grade CVD diamond at ~ 1 pm 0 wavelength
has been assessed by measurements of modulation transfer function (MTF) using
samples of optical grade CVD diamond with accurately polished surfaces to evalu-
ate one aspect of their performance as imaging windows [4,35]. The MTF is the best
parameter available for defining image quality. A modulation transfer function
curve effectively shows the contrast in the image as the spatial frequency of that
image changes (so that the curve will always drop to zero at a high enough spatial
frequency). The MTF of the components was determined from the width of the
image of a slit source. Any component which is placed in the light path which has
a poor optical performance (whether because of form shape errors or scatter) will
increase the broadening of the image of the slit source (the line spread function,
LSF).
An infrared interferometer was used to determine the LSF due to the form shape
error and this can be subtracted from the total broadening as measured in the infra-
red MTF instrument. What remains is the broadening due to the scatter alone.
Figure 6 shows a comparison of the LSF for the system with and without a
sample showing that the CVD diamond optical window introduces only a very
small broadening. The computer converts these data into an MTF trace such as
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Properties of' Chemical Vapor Deposited Diamond 579
2,500
2,000
e
$-= 1,500
b
'i1,000
32
- FWHM = 115 pm
500
n
-200 -300 -200 -100 0 100 200 300 400
Position (microns)
Figure 6. Comparison of the Line Spread Function measured with and without a sample in the
system.
that in Fig. 7 (MTF vs RSF, reduced spatial frequency) which also shows the MTF
degradation caused by form shape errors and the M T F of the system with no
window in place (the diffraction limited curve). The effect of scatter is therefore
reflected in the difference between the lower and middle curves. The conclusion is
that optical grade CVD diamond shows minimal scatter and degradation in MTF
Figure 7. Modulation Transfer Function (MTF) measured with a CVD diamond window (lower
curve), the M T F degradation caused by form shape errors (middle curve) and the MTF of the
system with no window in place showing the diffraction limited case (upper curve).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
580 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
in the far infrared and is therefore a good candidate for windows required in ima-
ging systems. Such an application will be discussed in 2.3.1.
The fracture stress of CVD diamond has been evaluated using the three-point bend
geometry [15]. The details of these measurements and its relative merits have been
discussed elsewhere [4,35].
The fracture stress is plotted against thickness in Figs 8 and 9 for the growth
(coarse grain) and nucleation (fine grain) surfaces respectively, for the optical and
mechanical grades of material. All of the samples in this study were in the as-
grown state. These strength data suggest several conclusions.
First, the measured strength depends strongly on thickness. There is an increase in
grain size on the growth side with thickness and it is speculated that the size of the
critical flaw is limited by the grain size close to the tested face, as discussed
previously [4,25,35]. The variation of the stress level through the thickness is
another factor which may influence the thickness dependence of the strength. A
more precise analysis has been performed in which the experimental data seem to
be in good agreement with the model, as reported elsewhere [35].
Second, the reproducibility of the strength values is high. If the data from samples
of thickness between 0.4 and 1.4 mm are assessed, they give the Weibull moduli [ 151
shown in Table 1 (a higher value indicates a more reproducible strength, c.f.
sapphire = 2.1, zinc sulphide = 5.4, silicon carbide = lo). These can be corrected
to allow for the thickness effect [25,35]. The corrected values for the growth surface
are also given in the table.
The observations suggest that CVD diamond has bulk (rather than surface)
200
0 0.5 1 15 2 2.5 3
Thickness (mm)
Figure 8. Fracture stress as a function of thickness measured with the growth face in tension.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.2 Properties of Chemical Vapor Deposited Diamond 58 1
1400
T 1200 -
2i
* loo0 -
I
v)
P 600
8oo
I I I I I
400
strength controlling flaws [26] which are related to the size of the crystal grains.
Table 1. The Weibull moduli calculated from the data in Figs 8 and 9
1160
-2 1140
1120
9
-g
1
1100
1080
-& 1060
c
ga 1040
1020
1000
0 100 200 300 400 500 600 700 800 900
Temperature [“C]
Figure 10. Young modulus of an Optical Grade CVD diamond specimen in the 20°C to 800°C tem-
perature range [37].
I I
A
Y
E.
4000
P
3 3000
u
S
U
s 2000
0
I I
150 200 250 300 350 400 450
Temperature (K)
Optical Thermal 1 Thermal 2 Thermal 3 Type 118 natural
0 4 * -
Figure 11. Temperature dependence of the thermal conductivity for different grades of DIAFILM
CVD diamond: optical grade and three thermal grades.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Optical Applications 583
DIAFILM specimens. These results are compared with literature values of natural
Type IIa single crystal diamond [39]. The difference between the thermal conductiv-
ity of optical and thermal grade specimens becomes more appreciable at lower tem-
peratures. The optical grade specimen seems to follow the Type IIa trend reaching a
value of 4000 W mK-’ at 200 K, probably the highest reported so far for polycrys-
talline CVD diamond at that temperature. For the ‘thermal-3’ grade, which shows
the lowest conductivity and is of dark color, the temperature dependence is not very
pronounced, in agreement with previous publications [40,41]. The ‘thermal-1’ grade
has a thermal conductivity better than 1800W mK-’ at room temperature which is
adequate for most demanding applications, as discussed in 2.5.2. The ability to
define different grades of thermal material is an example of the cost-performance
optimisation that is possible with CVD diamond.
The extreme hardness and abrasion resistance of diamond makes this material an
ideal choice for applications in which the optical components are exposed to aggres-
sive environments. The use of diamond for optical elements for infrared seeking mis-
siles and other military uses constitute probably the best examples of this type of
application as discussed in more detail in this section. There are, however, a host
of other applications in industry outside the military field in which the attributes
of diamond optics are of great advantage. These include windows used in the mon-
itoring of chemical reactions, or in the analysis of fluids which contain abrasive com-
ponents.
Originally conceived during the 1940s to allow armed forces to operate effectively
under the cover of darkness, infrared thermal imaging fulfilled a need that could not
be met with radar, acoustic, or visible sensors [44]. Thermal imaging relies on sen-
sing the heat emitted by a body, by virtue of its temperature alone; thus thermal ima-
ging is purely passive in its operation. The wavelength of the emitted light is
temperature dependent and the IR detector is optimized for maximum sensitivity
over a limited wavelength (and hence temperature) range.
Infrared detector materials are delicate and require protection from the environ-
ment by a transparent window. In many military applications this environment can
be extremely harsh, such as the environment encountered by a high-speed, heat-
seeking missile or a tank window exposed to a sandstorm. The damage caused by
a high velocity rain drop impact onto a zinc sulphide window [45] is clearly illu-
strated in Fig. 12. The erosion/impact damage seriously affects the component
strength and also the ability to form an acceptable image behind the window.
Owing to the absence of strong absorption by water or carbon dioxide molecules
in the atmosphere, two wavebands are basically transparent and are generally used
by IR imaging equipment. The mid waveband is between 3 and 5 pm wavelength and
the long waveband is between 8 and 14 pm wavelength. For a body at 300 K there is
considerably more emission per unit area (termed excitance) in the 8-14 pm wave-
band than the 3-5 pm waveband (the excitance is about thirty times more). For a
body at 600K, the levels of excitance in each waveband are similar but at 2000K
the 3-5 pm excitance is a factor of about nine times higher [15]. This is one of the
main reasons why the 3-5 pm operating waveband has been favored for heat-
seeking missiles designed to follow the hot air trail left by a jet engine. The other
main reason why the mid waveband seekers are currently preferred is that there
are several durable window or dome materials that are transparent in that spectral
range, such as sapphire or magnesium oxide [ 151 and that are currently available on
a commercial basis. Figure 13 shows typical heat-seeking missiles fitted with 3-5 pm
wavelength transparent domes.
In contrast, the longer wavelength infrared (LWIR) band is desirable for imaging
cooler bodies (such as the aircraft itself rather than just the hot engines, which can be
‘masked’) and although LWIR detector materials and missile seekers have been
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
585
Figure 12. Example of the damage caused by a high velocity water drop onto a zinc sulphide
Figure 13. Examples of air-to-air heat-seeking missiles fitted with 3 to 5 pm wavelength transparent
2.3 Optical Applications
domes.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
586 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
Figure 14. Example of a CVD diamond hemispherical dome, 70 mm in diameter and over 1.O mm
thick that has been semi-processed on both surfaces.
available for some time, they have not been exploited due to the lack of erosion resis-
tant windows and domes that transmit in the 8-14pm waveband [14]. Materials
such as germanium and zinc sulphide, for instance, are transparent in the longer
wavelength IR band but even with the best available protective coatings they are
not able to withstand the rain erosion encountered during air carriage, let alone a
missile flight scenario.
Diamond has been recognized to be an ideally suited material for this application
[15,14,25] because of its good IR transmission and its extreme mechanical proper-
ties. It was only with the introduction of CVD diamond technology and its potential
for synthesizing diamond films over large areas that the IR seeker community
started to consider the practical possibility of using diamond for this application.
One approach has been the use of CVD diamond as a protective coating for
other softer IR materials such as ZnS or Ge [46]. However the successful develop-
ment of thick-film or bulk CVD diamond [2f, 131 motivated active development
projects on the synthesis of free-standing CVD diamond domes. This work has
met with considerable success [25,47] resulting, for instance, in the synthesis of
CVD diamond hemispherical dome geometries 70 mm diameter, 2 1 mm thick
that have been semi-processed on both surfaces as shown in Fig. 14. For imaging
applications the form shape needs to be very precise and the radii of the two surfaces
have been consistently machined to better than fl pm, which is technically difficult
to achieve in such a hard material.
Figure 15 shows the thermal image of a commercial aircraft through a CVD
diamond dome fitted (such as that shown in Fig. 14) to an 8-14 pm thermal imaging
camera. There is a small loss of image resolution due to imperfections in both the
surface finish (causing scattering) and the form shape of the dome being tested; how-
ever, the engines are clearly visible as being hotter than the bulk of the plane, and the
plane shape is clearly identified.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Optical Applications 587
Figure 15. Thermal image of a commercial aircraft through a CVD diamond dome such as that
shown in Fig. 14. The engines are clearly visible as being hotter than the main body and the
shape of the aircraft is clearly identified.
Table 2 Property comparison between high quality CVD diamond and alternative window
materials that transmit in the long wave (8-14 pm) thermal imaging region
-
-
at 473 k
Thermal shock (mild) FOM 1000 2.6 2.1 6.1
( x l o ' w m - ') (3)
Refractive index at 10.6 pm 2,376 2.19 2.19 4
dn/dT (K-') (at 300 K)
*
9.6 x 41 x 10 36 x 400 x
Dielectric const. D (35 GHz) 5.68 0.15 8.35 8.39 -17
YOincrease in D at 773 K (Yo) 4.3 -8 -8 -
(1) *** Band cdge moves towards shorter wavelength,introducing significant absorption.
Mass loss due to oxidation.
***
Free-carrier absorption makes material opaque.
(2) Calculated sand erosion mass loss values at 80 m/s for ZnS and Ge (to allow a comparison
with CVD diamond) are based upon up-scaling theories (using erosion rate velocity scaling
factors) from work undertaken at the Cavendish Laboratory [45].
(3) FOM = (Material strength x (1 - Poisson's Ratio) x thermal conductivity/thermal expan-
sion coeff. x Young modulus) [IS].
90
\
CVD diamond, growth surface
70
,
40
30
I’ I I
sapphire
I I
Figure 16. Reduction in transmission as a function of sand erosion time for a CVD diamond
window compared to a sapphire window.
explains why CVD diamond is emerging as one of the preferred material options for
applications requiring multi-spectral transmission in harsh environments.
A window can absorb heat from the laser beam by intrinsic bulk absorption or at the
surfaces (through intrinsic surface absorption mechanisms or at the anti-reflection
(AR) coatings). The absorbed heat induces radial and axial temperature gradients
which in turn result in two distinct forms of optical distortion:
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
590 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
(i) The thermo-refractive effect is the result of changes in refractive index with
temperature and is described by the thermo-refractive coefficient dnjdT,
where n is the refractive index of the window material.
(ii) Thermo-elastic effects which are a caused by bulging (or other form-shape dis-
tortions) or by changes in the optical constants by thermal stress gradients
through the photo-elastic effect.
Materials which have relatively small thermal gradients and do not change geometry
with temperature will show the lowest values of thermal distortion. Critical param-
eters of relevance in these effects are thermal conductivity, thermal expansion
coefficient, thermo-refractive constant, bulk absorption coefficient, absorption at
the AR coatings and the photo-elastic constants. The total wave-front distortion
can be described by the variation in phase (or optical-path-length) of a transmitted
beam with radial position. For the thermo-refractive effect, for instance, the phase
change will be (dn/dT) x t x A T (where t is the thickness of the window and AT is
the temperature change) and is thus smaller for a thinner window. The minimum
acceptable thickness of a window will be determined largely by the ability to support
vacuum or pressure without breaking or distorting [50]. Critical parameters in these
cases are the fracture strength and the Young modulus respectively.
The current standard material for optical elements in C 0 2 lasers is ZnSe [27]
because of its very low intrinsic absorption at 10.6 pm [14, 15,341. Table 3 compares
the critical parameters of CVD diamond and ZnSe showing that CVD diamond, for
the reasons discussed above, has the potential to handle much greater beam powers.
This has been long recognized [51] and the thermal effects in diamond laser windows
have been theoretically modelled in previous work [27,51-531. Some of the earlier
results [5 1-53] were derived before reliable data were available on the properties
of polycrystalline diamond and were therefore very speculative. The following is
an up-date of the predicted thermal effects in optical grade CVD diamond and
ZnSe windows where some of the earlier calculations arc revisited.
The analysis presented here assumes that bulging and thermo-elastic effects in
either diamond and ZnSe can be neglected [27,52-541. The assumed configuration
for each case is that of a circular window, 25mm in diameter, which is edge
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Opticul Applications 59 1
Figure 17. Schematic of the axially symmetrical temperature gradients generated in an edge-cooled
window fitted with AR coatings when traversed by a laser beam.
Using Eq (l), the ratio of the CVD diamond-to-ZnSe thickness (for the same
diameter) is:
/
tCVDD tZnSe = { (kCVDafCVDDSZnSe) / (kZnSefZnSeSCVDD) 1112 (3)
Using the parameters listed in Table 3, assuming that the k parameter is similar in
both materials and using the same safety factor for ZnSe and CVD diamond results
in tCVDD/tZnSe z 1/3. Equations (1) and (2) predict a minimum t / D ratio of 0.017 for
CVD diamond (using a safety factor of 4 and a fracture strength value of 400 MPa)
or a thickness above 0.4mm for a 25 mm diameter window. In practice, it is found
that a better flatness and overall shape stability can be achieved if the thickness is
close to 1.Omm. For these calculations it will be assumed that the CVD diamond
window is 1.O mm thick and the ZnSe 6.0 mm, which is the typical thickness used
for ZnSe windows in the industry [27]. The higher thickness used for ZnSe may
be partly justified by the need to use a higher safety factor due to the lower value
of its Weibull modulus [35].
The absolute bulk absorption of optical grade CVD diamond windows has been
found to be in the range 0.1--0.03cm-' [27].A value of 0.05 cm-' has been assumed
for this calculation.
The analysis has been done for a continuous wave, 5 kW incident beam that is
assumed to have a Gaussian profile with a 14mm diameter at 14% (lie2) of maxi-
mum intensity. For this level of power the temperature increases are relatively low
and the temperature dependence of parameters such as dn/dT [54] have been
ignored, but the temperature dependence of the thermal conductivity for optical
grade CVD diamond (Fig. ll), although not very pronounced, has been taken
into account.
The analysis solves the fundamental heat propagation equations using a two-
dimensional computational model in which values of the radial and axial variations
of the temperature and the refractive index are calculated. The calculated tem-
perature gradients for the two cases are shown in Fig. 18. Figure 19 shows the
radial dependence of the average (across the thickness) temperature. The tempera-
ture in CVD diamond shows a maximum variation (center to edge) of 3.4"C
compared to 23.4"C for the ZnSe case. The temperature variations over the effective
diameter of the beam are 1.7"C and 12.4"C for the CVD diamond and ZnSe
windows respectively.
The phase or optical path length variation A@ = (dn x t ) is a function of radial
position in the ZnSe and CVD diamond windows and is shown in Fig. 20. The
variation for CVD diamond over the effective beam width is over a factor of 250
smaller. Three factors combine to yield this large difference: the smaller temperature
rise in CVD diamond due to the higher thermal conductivity ( A T a factor of more
than 7 smaller in CVD diamond), the smaller value of dn/dt for diamond (a factor of
5.8), and the thickness of the window, which for CVD diamond is only 1/6th of that
in ZnSe. For an example similar to that shown in Fig. 18 the thermal lensing effect in
the ZnSe window was found to be equivalent to a lens of 3-5 m focal length, whereas
for the CVD diamond window the effect was negligible [27,34].
As a general rule it can be assumed that the lensing effect can be neglected if the
difference in A@is smaller than hjl0 [51],or 1 pm for a C 0 2 laser beam. As shown in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.3 Optical Applications 593
Figure 18. Calculated temperature gradients in a CVD diamond and a ZnSe windows, edge cooled
and with AR coatings.
Fig. 20, A@is approximately 8 pm for ZnSe but under 0.04 pm for CVD diamond.
Recent results have been reported using electronic speckle pattern interferometry
(ESPI) [56] using a C 0 2 laser of power up to 1 kW on a CVD diamond and a
ZnSe windows of 25mm aperture and AR coated on both sides. The ZnSe
window starts to deform at moderately low power levels (50W). In the CVD
diamond window discernible distortions start at powers over 700 W. These
0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3
Radial Position (mm)
Figure 19. Radial dependence of the average (across the thickness) temperature for the CVD
diamond and the ZnSe windows derived from the calculations shown in Fig. 18.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
594 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
10
8 --...._
-...--.
.-
a
--......
h
fn *%.
C
$ 6 *.' ..%.
.-
E
Y
*%.
CI
x 4 -*...
rn
C
2 -*...+.-
5.
-...
-*..
....*-.
I .%
0
Figure 21. lnterferometric surface profile of a 25mm CVD diamond window showing a flatness
better than one fringe (633 nm illumination).
with ZnSe). For a 160 kW beam, this implies a minimum beam diameter of 8 mm,
considerably smaller than actual beam sizes in lasers of this power level. What has
not yet been investigated is the potential decrease in LIDT that could be introduced
by surface and bulk features typical of polycrystalline CVD diamond [2,35]. For
pulsed lasers, considerably higher LIDT values have been reported [58]. Clearly,
more experimental work is required before the true power handling capability of
CVD diamond is truly understood.
Beam distortion can be generated when the form shape of the surfaces of a window
deviate from perfect flatness. The flatness of optical elements as measured by a inter-
ferometer is expressed as the number of fringes (twice the wavelength of the light
beam probe) shown in the interference pattern of the optical element placed against
a perfect flat surface [59]. Depending on the application, an acceptable number of
fringes is specified. Applications in laser optics often require a flatness of < 4 fringes
and sometimes < 1 fringe using visible illumination (typically 633 nm wavelength).
Figure 21 shows the interferometer pattern of the surface of a CVD diamond
Window which has been polished to a flatness better than 1 fringe (633 nm illumina-
tion). This is typical of what can be achieved presently for CVD diamond showing
that current technology is sufficient to address applications such as exit windows,
where a flatness of 1 to 3 fringes is considered adequate. Better results have been
achieved using techniques currently under development as shown in Fig. 22 where
a flatness of 0.5 fringe has been achieved in a window 48mm in diameter. This
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
596 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
Figure 22. lnterferometric surfxe profile of a CVD diamond window showing that by the use of an
improved process, a flatness of 0.5 fringe can be obtained over a 48 mm diameter.
suggests that CVD diamond will be able to address a wider range of applications as
an optical element in laser systems.
The ability to process CVD diamond to controlled form shapes has also allowed
the production of samples with wedge angles up to 1 degree, as required to eliminate
interference fringes for broad band IR transmission, as is the case in synchrotron
beam lines [4,33].
Figure 23. The Rofin Sinar ‘Diffusion cooled’ DC-series COz slab laser with output power up to
3500 W [60].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.4 Windowsf o r High Power Gyrotron Tubes 597
,Bending mirror
Figure 24. Schematic diagram showing the location of the CVD diamond window in the ‘Diffusion
cooled’ Rofin Sinar DC-series COz slab laser [60].
2.3.2.3 Scattering
Scattering in the infrared has been discussed in 2.2.2 and by Union et al. [34] con-
cluding that, at ~ 1 pm,
0 a good quality optical grade CVD diamond window does
not show a level of scattering that is likely to affect its infrared imaging properties
and is capable of transmitting a diffraction limited image. For systems operating at
wavelengths in the UV-visible spectral range, scattering effects may be more pro-
nounced, although recent measurements indicate that a good quality CVD diamond
window could exhibit values of scattering at 633 nm similar to that of a single crystal
Type IIa specimen [54].
CVD diamond windows are now being used in commercial high power COz
lasers. An example is the Rofin Sinar Diffusion-cooled C 0 2 slab laser with an
output power of 3500 W [48] shown in Fig. 23. In this laser, the CVD diamond
window is used to seal the slab cavity, as shown schematically in Fig. 24.
Figure 25. Example of a 1.3 MW, 140 GHz Gyrotron tube built at the Forschungzentrum Karlsruhe
WI.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Additionally, the barrier windows in the Plasma Vacuum Torus should be able to
withstand high doses of radiation without degradation.
The critical parameters for high-power windows are the dielectric characteristics
of the windon, material;;he dielectric loss-factor tan 6 and the permittivity E:. (or the
refractive index 11 = E:.' ' - ) because they affect power absorption and reflection. The
power absorption coefficient 'x is related to tan 6 by [43]
'x = 2x11 tan ijj h
where h is the wavelength of the microwave radiation in free space.
To minimize power reflection at a given frequency the thickness t of the window
needs to be a11 integer number of half dielectric wavelengths [64]
t = Nhj21l
where N is an integer. This requirsment can be relaxed if a Brewster angle window is
used. but the penalty is that a much larger diameter window is necessary [67].
The requirement of CW operation puts extremely high demands on the material
properties of the dielectric vacuum barrier window that serves as both the primary
tritium containment boundary and as the output window of the Gyrotron. The
window options available to Gyrotron manufacturers have been discussed in
detail by Th~iiiiiiiet ril. [64]. From this analysis it is appreciated that all other options
apart from diamond require sophisticated and, in some cases, impractical cooling
system involb ing, for instance. cryogenically edge-cooled windows, liquid-surface-
cooled double-disc windows or distributed windows i n which segments of dielectric
are metal bonded to microchannel cooled metal ribs. Another shortcoming of some
of the materials considered. such as sapphire or gold doped silicon, is the
pronounced temperature dependence of the loss factor and refractive index,
which may lead to thermal runaway and lor detuning of the window resulting in
high levels of reflected power.
This background explains the keen interest in diamond as an alternative material.
Theoretical thcrmal modeling results such as that illustrated in Fig. 26 [66] show that
a simple edge cooled diamond window approxiinately 100 mm in diameter and with
a value of loss tangent below 1 O-'. should be able to withstand powers in excess of
1 M W with a temperature rise at the centre of the window below 240 C, which is far
below the temperature (600-C)at which diamond starts to degrade by oxidation,
although thermal stress may be a limiting factor at these temperatures. This
relatively low temperature excursion is a direct consequence of the high thermal
conductivity of diamond. Table 4 compares properties of diamond relevant to
this application with those of other competitive materials.
510
520
sw
480
460
440
420
4lm
380
360
340
320
3w
0.00
Figure 26. Results from a theoretical thermal model showing the predicted temperature rise at the
center of an edge cooled diamond window for a 1 MW incident microwave beam. Results are shown
for different values of dielectric loss and for two different cooling diameters assuming an exposed
aperture of 100mm diameter [66].
Figure 27. Spatial distribution of the dielectric loss in a 100mm diameter CVD diamond window
1.6mm thick measured with a spatial resolution of 6mm. The isolated points of higher loss at the
edge may be caused by scattering losses. Over most of the area the measured dielectric loss is equal
or below tan 6 = which is close to the sensitivity of the measuring technique. This may account
for some of the measured isolated spurious higher values of loss.
specimens (DB 1 to 4) can also be achieved in full size windows. The spatial unifor-
mity of the dielectric loss in this window is shown in Fig. 27 which shows a map of
tan 6 measured with a beam diameter of 6 mm [69]. It can be seen that over most of
the active aperture of the window (a central area of approximately 80mm diameter)
tan 6 is below lop5 and close to the sensitivity of the measuring technique, which
may account for spurious isolated higher values of loss. The temperature depen-
dence of this window over the 70 to 370K temperature range indicates that the
loss is largely independent of temperature over this range and that the average
tan 6 exhibits a value of 0.6 (* 0.2) [70], which is the lowest loss tangent so
far reported for a CVD diamond window.
Sample DB 6 is probably the largest Gyrotron window ever made (named 'Star
of FZK') and exhibits a loss of 2 x over the same temperature range as
DB 5 [70].
The window DB 8, which has a relatively high loss tangent of over 13 x lop5,was
the first CVD diamond window to be mounted in a high power Gyrotron tube
(shown in Fig. 28) at the Japan Atomic Energy Research Institute (JAERI) [71-
731. This window was tested in 1998 using the 170 GHz JAERI/Toshiba Gyrotron
at output powers of 0.5 MW for 8 s with a water cooled edge showing a temperature
increase a t the center of not more than 130°C,consistent with theoretical modelling
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
f o r High Power G y o t r o n Tubes
2.4 Windo~vs 603
Figure 28. The first high-power Gyrotron (JAERI/Toshiba, 170 GHz) fitted with a CVD diamond
window. Measurements using beam powers up to 1 MW for 10 seconds have already been
performed.
[73]. The second window for the JAERI Gyrotron (DB7) has a lower loss tangent of
approximately 2 x lo-’. This window was tested at the end of 1999 using also the
170GHz Gyrotron for powers up to 1 MW for 10s. The temperature increase at
the center was less than 3 0 T , also in very good agreement with theoretical model-
ling [73]. Because diamond reaches almost steady state for times in excess of 3 s the
above experiments show convincingly that CVD diamond windows are able to
transmit millimeter wave powers of at least 1 MW continuous wave. At the
moment of going to press, in excess of 20 full size CVD diamond windows have
been delivered to members of the international nuclear fusion community. These
exhibit routinely values of loss tangent of 2-3 x lo-’.
It is not clear yet if the low values of dielectric loss of samples DB 1 to DB 7
represent a fundamental limit of a loss mechanism in diamond or if they are still
affected by impurities or defects in the polycrystalline CVD diamond structure.
From measurements in previous samples [74], there is clear evidence that, in
samples of relatively low quality (tan 6 greater than 70 x darker regions of
the sample have higher values of loss. Even in the highest quality polycrystalline
CVD diamond, such as that of the sample shown in Fig. 27, it is possible to observe
localized features related to the grain structure of the layer. It can be speculated that
these features may be a contributing factor to the dielectric loss of polycrystalline
CVD diamond. However, from a practical point of view, this is probably not of
great concern in the foreseeable future since it has been estimated that diamond
windows with values of tan 6 below 2 x lo-’ (a factor of two above that of the
window shown in Fig. 27) should have a power handling capability of at least
2.5 MW [64].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
604 2 Applications of' Diamond Synthesized by Chemical Vapor Deposition
10
5
7
<
0
s 2
*
E
$ 1
C
m
I- 0.5
v)
v)
0
2
A 0.2
Q,
ii 0.1
0.05
1 100 150 200 250 300 350 1 0
TtKl
Figure 29. Temperature dependence of tan 6 for a CVD diamond sample measured at 145GHz
compared with sapphire [75].
Figure 31. Examples of CVD diamond windows mounted onto vacuum flanges. The two at either
end show 100 mm diameter windows mounted to conventional CONFLAT vacuum flanges. The one
in the middle illustrates a double flange configuration designed specifically for Gyrotron tube assem-
blies. All these vacuum assemblies have been tested to be vacuum tight to better than 10-9 mbar I s C '
Figure 32. Window disc with slots to reduce hoop stresses and to facilitate bonding alignment
Figure 33. Package configuration analysed in the work consisting of the laser diode array mounted
on a CVD diamond spreader which is mounted on a copper heat-sink block.
and separated by a spacing of 200pm. The length of the heat-spreader has been
assumed to be constant at 11mm. This is a three-dimensional calculation which
simultaneously takes into account heat dissipation parallel and perpendicular
to the direction of the LDA bar, in contrast with previous two-dimensional
calculations reported in the literature [79].
Figures 34 and 35 show that the thermal resistance decreases as a function of heat-
spreader thickness and depth. Heat dissipation through the rear (or depth) of the heat
spreader is important, because in an array of diodes there is thermal cross-talk
between junctions in which the temperature of one junction is affected by that of
adjacent junctions. Figure 34 shows that heat-spreader depths between 2 and 3 mm
are required for a thickness of 300pm. Also, because of thermal cross-talk, the
0.2'
0
I
100
I
200
1
300
I
400
I
500
I
600 700
Thickness of CVDD layer (microns)
Figure 34. Thermal resistance decrease as a function of the thickness of the CVD diamond heat
spreader for the configuration shown in Fig. 33.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.5 Thermal Management of Laser Diode Arrays 609
8c 0.32 ~
.-4
3 0.28
-m
'
p?
0.24
c
I-
0.2
0
' 0.5 1 1.5 2 2.5 3 3.5 4
Length of CVDD (mm)
Figure 35. Thermal resistance decrease as a function of the depth of the CVD diamond heat sprea-
der for the configuration shown in Fig. 33.
influence of heat spreader thickness is not as pronounced as would have been expected
in a single diode junction and very little gain is achieved for a thickness above 300 pm,
as shown in Fig. 35.
Figure 36 shows the decrease in thermal resistance as a function of the thermal
conductivity of the spreader, showing that, for a total dissipated power of 90 W
(or 3.6 W per junction), a reduction close to 5°C is obtained by increasing the
thermal conductivity from 1000 W m-' K-' to values close to 2000 W m-' K-I.
This reduction in temperature is significant in view of the increase in life of a
factor of two expected for a decrease on lO"C, as mentioned above. As discussed
in 2.2.5, values of thermal conductivity of 1000W m-' K-' are typical of the
lower thermal grade whereas the higher grades exhibit a thermal conductivity
close to 2000 W m-' K-l (Fig. 33). The result of Fig. 36 shows, therefore, that for
this application it is advantageous to use a higher thermal conductivity CVD dia-
mond grade.
18'
500 1000 1500 2( 10
Themal Conductivity (WlmK)
Figure 36. Decrease of the thermal resistance as a function of the thermal conductivity of the CVD
diamond heat spreader.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
610 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
These results indicate that an ideal CVD diamond heat-spreader would reduce the
thermal resistance of a 25 element array by approximately 36% from 0.36 K W-' to
0.23 K W-I. Experimental results [79] show that, in a similar array, the measured
reduction in thermal resistance is only 22%. This may be a consequence of the
extra thermal resistance introduced by the metallization. If a metallization thermal
resistance value of 0.12 K W-' is added, the results from the above calculations
agree well with experimental results. Using a novel microchannel heat-spreader
made of CVD diamond reductions in thermal resistance up to 75% have been
predicted [81].
Figure 37. Example (courtesy of Jenoptik [83]) of a laser diode array bar which has been pre-cracked
to reduce effects of thermal stress.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2.6 Cutting Tools. Dressers and Wear Parts 61 1
stress can seriously affect the reliability and life of these devices [79,80,82]. A novel
approach has been proposed called stress induced dicing (SID) where the tensile
stress in the laser diode bar after soldering and during cooling down is relaxed by
cracking the bar at defined lines between emitters [83]. Figure 37 shows an example
(courtesy of Jenoptik) of such a laser diode bar mounted on a pre-metallized CVD
diamond spreader using a Au:Sn (80:20) solder alloy.
Figure 38. Schematic diagram showing the relative merits of various cutting tool materials for the
machining of different materials. It can be seen that CVD diamond occupies a specific niche in the
machining of materials requiring exceptional wear resistance.
2.6.2.1 High Speed Finish Machining of High Pressure Laminate Wood Flooring
During the 1990s, the wood working industry has seen a phenomenal growth of the
use of PCD cutting tools in sawing, milling, and profiling of high and medium
density fibre board, chipboard, particle board - all known to be abrasive materials.
In the last five years, high pressure laminate (HPL) - a generic wood product, with
its decorative aluminum oxide coated surface for high wear resistance, has emerged
as a flooring material. Figure 39 shows the makeup of a HPL wood flooring
material. In this application, CVD thick film diamond has recently been tested
alongside pcD and single crystal diamond tools for the finish machining of the
highly abrasive coated layer of HPL wood flooring material, with very encouraging
results [84] as described below.
For CVD diamond materials, the tool edges were prepared by mechanical grinding
using a standard cutter grinder and diamond wheels. The cutting edge of synthetic
single crystal diamond was prepared by scaife polishing, and the pcD tools' edges
were prepared by the standard Electric Discharge Grinding (EDG) method.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 14 2 Applications of’Diamond Synthesized by Chemical Vupor Deposition
The criterion by which tool life is measured is chipping of the board edge. When
chips of an ‘unacceptable’ size are produced, then the tool is changed. ‘Unaccepta-
ble’ chip size is determined visually by the machine operator.
2.6.2.1.2 Test Results Figure 40 shows, for each test, the average linear length
machined by the diamond tools, before the tools were taken out on the basis of visi-
ble edge chipping of the floor panels.
The edge wear of CVD tools appears to have been more influenced by grain pull
out and erosion at the grain boundaries than wear.
All the materials evaluated in these tests are capable of machining HPL flooring,
however, it is quite clear that single crystal diamond gives the longest tool life and
CVD diamond gives a marginal advantage compared with PCD. Both synthetic
single crystal diamond and CVD diamond materials, while providing high tool
life, need to be evaluated in the greater cost benefit equation for specific workpiece
materials, the requirements of individual HPL flooring products and the manufac-
turing environment pertaining.
0.35
PCD 025
3 0.3
E PCD 002
0.25
m
g 0.2 CVD thick film
5 0.15
Y
2.6.2.2.2 Test Results Figure 41 shows the flank wear on the respective tool relative
to the machining time. In the case of the CVD thin coated tool, initially the rate of
tool wear is low but as soon as the ‘thin’ coating is worn through the flank wear
increases rapidly. The lowest rate of tool wear is observed on the CVD thick film
(0.5 mm) insert. Although testing is limited, the CVD thick film flank wear indicates
that the edge has ‘bedded in’ and steady state wear has been established. This sug-
gests that the ultimate tool life could be quite considerable.
The machining of MMCs, particularly high volume fraction, > 35%, generates
high temperature and involve having to efficiently shear hard silicon carbide parti-
cles in the cut zone. The high thermal conductivity/resistance to thermal degrada-
tion and abrasion resistance of CVD are particularly suitable for machining high
volume fraction MMCs.
Table 6 summarizes the relative merits of the three major diamond cutting tool
materials - pcD, CVD diamond, and synthetic single crystal as current technology
~
Table 6. The relative merits of different diamond cutting tool materials: tool performance.
Toughness JJJ JJ J
Abrasion resistance JJ JJ JJJ
Thermo-chemical resistance J JJJ JJJ
’
Dresser Wear (mm)
0. 5 Dnrarw paramatem
Meel AI-0xde WA46KV (300mm dla x 50 rnm d e )
Speed 2300rpm
In-Feed 0 020 mm I wheel dla
Feed 4 rnrnlrec
Caolant Floodmolant
0.05 -
0 ,
0 200 400 600 800
Volume of Wheel Removed (cm3)
CVD Diamond
-m- MONZESS
Figure 42. CVD diamond and Monodress single point dresser wear relative to volume of grinding
wheel removed.
The test parameters were identical in each case, with a contact stress of 15 MPa
and sliding speed of 150 m min-' . The tests were performed for 48 h where possible
(to steady state or until sliding speed diminished through increased friction).
The tests all began at nominal room temperature of 20°C and, where required, the
lubricant used was a typical heavy white mineral oil; excess lubricant was removed by
a repeated solvent wash prior to weighing the pin to determine the specific wear rate.
The temperature rise associated with both the dry and lubricated wear tests did
not exceed 90°C. Such a temperature rise can be associated with the low coefficients
of friction (and carrier mechanisms during the lubricated tests). Such increases are
not sufficient to noticeably alter the properties of the CVD diamond or to break-
down the mineral oil lubricant.
The coefficient of friction determinations have been made at specific times into the
sliding wear tests. These have an associated uncertainty of approximately flo%,
predominantly from the (unsteady) change in friction coefficient with time as the
sample surface is modified by wear and vibration/noise associated with the friction
monitoring transducer.
It is worthwhile noting that there are several sources of error in the determination
of the specific wear rate which give approximately an order of magnitude uncer-
tainty in the quoted values, in particular, the mass loss through wear of the pins
is minuscule - a consequence of the excellent abrasion resistance of CVD diamond
~ and such a low loss is easily affected by any residual lubricant.
As can be seen from the data presented in Table 7 the surface finish of the CVD
diamond has a considerable effect on both the wear rate and coefficient of friction.
However, whereas the hardened steel could not be run without a lubricating layer,
the CVD diamond does not require lubrication or might be self-lubricating.
It is clear that there is a great potential for very low wear rate and low coefficient
of friction CVD diamond materials prepared with the required surface finish for
wear part applications, where component lifetime is determined by sliding wear in
low lubricant or lubricant-free applications.
References
1. R. S. Sussmann, DIAFILM, A New Material for Optics and Electronics, Indust. Diamond Rev.,
1993, 53,555, pp. 63-72.
2. R. S. Sussmann, J. R. Brandon, G .A. Scarsbrook, C . G. Sweeney, T. J.-Valentine, A. J. White-
head, and C . J. H. Wort, Diamond Re/. Muter., 1994, 3, 303-312.
3. C. J. H. Wort, C. G. Sweeney, M. A. Cooper, G. A. Scarsbrook, and R. S. Sussmann, Diamond
Re/. Muter., 1994, 3, 1158-1 167.
4. R. S. Sussmann, C. S. J. Pickles, J. R. Brandon, C . J. H. Wort, S. E. Coe, A. Wasenczuk, C. N.
Dodge, A. C. Beale, A. J. Krehan, P. Dore, A. Nueara, and P. Calvani, Nuovo Cimento, Section
D , 1998, 12 (April), 503-526.
5. R. S. Sussmann, J. R. Brandon, S. E. Coe, C. S. J. Pickles, C . G. Sweeney, A. Wasenczuk, C. J.
H. Wort, and C . N. Dodge, CVD Diamond, A. New Engineering Material for Thermal Dielec-
tric and Optical Applications, Proceedings of the Industrial Diamond Association of America
meeting on Ultrahurd Materials, Windsor, Ontario 28-30 May 1998, and Indust. Diamond
Rev., 1998, 58(578), 69-77.
6. C. P. Klages, Chemical Vapor Deposition of Diamond Films, in Handbook qf Ceramic Hard
Materials, R. Riedel (Ed.). Wiley-VCH, Weinheim, 2000, pp. 390-419.
7. G. J. Davies, Section 2 of Diamond Materials and their Applications, in Hundbook of Ceramic
Hard Materials, R. Riedel (Ed.), Wiley-VCH, Weinheim, 2000, pp. 485-504.
8. M. W. Bailey, Section 1 of Diamond Materials and their Applications, in Handbook ofceramic
Hard Materials, R. Riedel (Ed.), Wiley-VCH, Weinheim, 2000, pp. 479-485.
9. R. Berman Chapter I , J. E. Field Chapter 9, D. Tabor Chapter 10, A. T. Collins Chapter 3, C . D.
Clark, E. W. Mitchell and D. J. Parson Chapter 2, and C. A. Brooks Chapter 12, in The
Properlies qf Diamond, J. E. Field (Ed.), Academic Press, London, New York, San Francisco,
1979.
10. C. D. Clark, A. T. Collins, and G. S. Woods Chapter 2, R. Berman Chapter 7. J. E. Field
Chapter 12. C . A. Brookes Chapter 13. D. Tabor and J. E. Field Chapter 14, J. Wilks and
E. M. Wilks Chapter 15, and J. E. Field Chapter 18, in The Properties of Natural andSynthetic
Diamond, J. E. Field (Ed.), Academic Press London, San Diego, New York, Boston, Sydney,
Tokyo and Toronto, 1992.
11. J. E. Field in Proceedings of the N A T O Advanced Study Institute on Diamond and Diamond-like
Films and Coatings 22 July 1990, R. E. Clausing, L. L. Hardon, J. C. Angus, and P. Koidl (Eds),
Plenum Press New York, London, 1991, pp. 17.
12. Chapters I , 3, 5 , 7 and 8. in Handbook uf’lndustrial Diamond~sand Diamond Films M. A. Prelas,
G. Popovici. and L. K. Bigelow (Eds), Marcel Dekker Inc. New York, Basel, Hong Kong,
1998.
13. Chapters I , 2, and 5. in Properties andGrovvth ofDiamond, G. Davies (Ed.), EMIS Data Review
Services, INSPEC IEE, London, 1994.
14. J. Save, in Infrared Optical Materials und Their Antirefection Coatings, Adam Hilger, Bristol,
Boston, 1985, p. 119.
15. D. C. Harris in Itfrared Windows and Dome Materials, Tutorial Texts in Optical Engineering,
Vol. TTlO SPIE Optical Engineering Press, Bellingham, WA (USA), 1992, Chapter 5.
-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
620 2 Applications of Diamond Synthesized by Chemical Vapor Deposition
16. M Seal and W. J . P. van Enckevort, Proceedings of SPZE- Diamond Optics, 1988,969, 144152.
17. M Seal in Industrial Diamond Review, P. A. Daniel (Ed.), 1971, 31 (Nov), 464469.
18. J. Doting and J. Molenadr, Proceedings ofthe Fourth Annual ZEEE Semiconductor Thermal and
Temperature Measurement Symposium, Sun Diego California, February. IEEE CH2530-
418810000-0113, 1988, pp. 113-117.
19. R. J. Caveney in Chapter 20 of Ref. [9].
20. M. Seal in Chapter 16 of Ref. [lo].
21. G. S. Woods in Chapter 3 of Ref. [13].
22. J. W. Vandersande in Chapter 3. of Ref. [13].
23. A. T. Collins in Chapter 9. of Ref. [ 131; B. A. Fox, L. Harstell, D. M. Malta, H. A. Wynands, G.
Tessmer, and D. L. Preifus in Proceedings of the Materials Research Society Symposium, 1996,
416, 319-329.
24. M. Seal, Zndust. Diamond Rev., 1979, 39(461), 115-118.
25. J. A. Savage, C. J. H. Wort, C. S. J. Pickles, R. S. Sussmann, C. G. Sweeney, M. R. McClymont,
J. R. Brandon, C. N. Dodge and A. C. Beale. SPZE Proceedings of the Windowand Dome Tech-
nologies and Materials Symposium, Orlando, Florida, April, 1997, 3060, 1 4 4 159.
26. D. C. Harris, Development of Chemical Vapour Deposited Diamondjor Infrared Optical Applica-
tions, Status Report and Summary of Properties, Report No N AWCWPNS TP 8210, Naval Air
Warfare Center Weapons Division, China Lake, CA (USA), 1994.
27. M. Massari, P. Union, G. A. Scarsbrook, R. S. Sussmann, and P. Muys, Proceedings ofthe SPZE
Laser-Znduced Damage in Optical Mateilals, Boulder, Colorado, November, 1995, SPZE, 2714,
177-184.
28. DIAFILM is a. trade mark of De Beers Industrial Diamond Division.
29. R F. Potter in Handbook of Optical Constants of Solids, E. D. Palik (Ed.), Academic Press Inc.
Orlando. San Diego, New York, London, Toronto, Montreal, Sydney, Tokyo, 1985, Chapter 2.
30. M. E. Thomas and W. J. Tropf, SPZE, 1994,2286, 114&151.
31. V. Vorlicek, J. Rosa, M. Vanacek, M. Nesiadek, and L. M. Stals, Proceedings of the 7th Euro-
pean Conference on Diamond, Diamond-like and Related Materials, Tours France, September
1996, Diamond Rel. Muter., 1997, 6 , 704707.
32. D. F. Edwards and H. R. Philipp, in Handbook of Optical Constants of Solids, E. D. Palik (Ed.),
Academic Press Inc. Orlando, San Diego, New York, London, Toronto, Montreal, Sydney,
Tokyo, 1985, subpart 3, pp. 668.
33. P. Dore, A. Nucara, D. Cannavo, G. De Marzi, P. Calvani, A. Marcelli, R. S. Sussmann, A. J.
Whitehead, C. N. Dodge, A. J. Krehan, and H. J. Peters, J. Appl. Optics, 1998,37 (July), no. 21.
34. P. Union, P. Muys, D. Vyncke, B. Depuydt, and P. Boone, SPIE, 1995,2870, 521-527.
35. C. S. J. Pickles, J. R. Brandon, S. E. Coe, and R. S. Sussmann, Proceedings of Symposium ZV,
Diamond Films, 9th Cimtec - World Forum on New Materials, Florence, Italy, June 1998,
P Vincenzini (Ed.), Faenza, Italy, 1999, pp. 435454.
36. T. J. Valentine, A. J. Whitehead, R. S. Sussmann, C. J. H. Wort, and G. A. Scarsbrook, Dia-
mond Rel. Muter., 1994, 3, 1168-1 172.
37. F. Sziics, University of Berlin and California Institute of Technology, Pasadena, California.
Private communication, 1998, see also: F. Sziics, M. Werner, R. S. Sussmann, C. S. J. Pickles,
and H. J. Fecht, J. App. Phys., 1999,86 (December) no. 11, 6010-6017.
38. F. Sziics, C. H. Moelle, S. Klose, H. J. Fecht, S. Fassbender, N. Meyendorf, and M. Werner, in
Proceedings of the International Conference on Micro Materials, April I997 (MICRO-MA ’97)
B. Michel and T. Winkler (Eds), 1997, pp. 569-572.
39. J. W. V. Vandersande in Chapter 1 of Ref. [13].
40. J. E. Graebner, M. E. Reiss, L. Seibles, T. M. Hartnett, R. P. Miller, and C. J. Robinson, Phys.
Rev. B, 1994, 50(6), 3702.
41. E. Wormer, C. Wild, W. Muller-Sebert, R. Cocher, and P. Koidl, Diamond Rel. Materi., 1996,5,
688-692.
42. See for instance P. A. Rizzi, Chapter 2.6 in Microwave Engineering, Prentice-Hall International,
Englewood Cliffs, NJ (USA), 1988, p. 41.
43. A. C. Lynch and R. N. Clarke, IEE Proceedings A , 1992, 139, 221.
44. A. P. O’Leary in Electro-Optics, Special Report on Thermal Imaging Systems, 2nd edn, Jane’s
Information Group, Antony Rowe Ltd, Chippenham, Wiltshire, UK, 1994.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 62 1
45. E. J. Coad, The Response of C V D Diamond and other Brittle Maierials to Multiple Liquid
Impacts, PhD Thesis, University of Cambridge, 1996.
46. M. D. Hudson, C. J. Brierley, A. J. Miller, and A. E. Wilson in the Proceedings of the Windou.
and Dome Technologies and Materials Synzposium, Orlando, Florida, April 1997, SPIE, 3060,
196202.
47. C. J. H. Wort, J. R. Brandon, B. S. C. Dorn, J. A. Savage, R. S. Sussmann, and A . J. Whitehead
in Applications o f Diamond Films and Related Materials: Third International Conference, A.
Feldman, Y. Tzeng, W. Yarbrough, M. Yoshikawa and M. Murokawo (Eds), NIST Special
Publications, 1995, 885, 569-572.
48. Rofin Sinar Technologies, Annual Report, 1997, pp 6. 7.
49. C. A . Klein, Optical Engineering, 1990, 29(4), 343-350.
50. M . Sparks and M. Cottis, J. App. Phys., 1973, 44, 787-794.
51. S. Singer, Diamond optics, A. Feldman and S. Holly (Eds), SPIE, 1988, 969, 168-177.
52. C. A. Klein in Laser Induced Damage in Optical Materials, 1991, SPIE, 1624, 475492.
53. C. A. Klein, Dianiond and Relaied Materials, 1993, 2, 102&1032.
54. C. Wild, Chapter 10.3 in Lou-Pressure Synthetic Diamond, B. Dischier and C. Wild (Eds),
Springer Veriag, Heidelberg, 1998.
55. S. P. Timoshenko and J. N. Goodier, Chapter 12 in Theory of Elasticity, 3rd edn, 1970,
McGraw-Hill, NY, pp. 385.
56. B. Depuydt, P. Boone, P. Union, and P. Muys in Proceedings y f t h e 4th International Worshop on
Laser Beam and Optics Characterisation, Munich, June 1997, A. Giesen and M. Morin (Ed.),
1997, 657-663.
57. N. Ellis and D. E. Greening Diamond and Related Materials, 1993, 2, 572-581.
58. R. S. Sussmann, C. A. Scarsbrook, C. J. H. Wort, and R. M. Wood. Diamond Rel. Muter., 1994,
3, 1 I73 1177.
-
59. H.-K. Karow, Fabrication Methods fbr Precision Optics, John Wiley and Sons, 1993, p. 620.
60. Tim Holt, Rofin Sinar 1998, Private communication.
61. R. Heidinger. Digest 19th International Confirence on Injiared and Millimeter Waves,Lausanne
( C H I , SPIR, 1991, 1576,441.
62. L. Redcliffe, Digest 16th International Conference on Injiared and Millinieter Waves.Sindon ( J ) ,
JSAP Cat No AP 941228, p. 277.
63. M. Thumm, State of the art of high power Gyro-devices and,free-electron Masers, Wissenscliaf~
tliche Berichte F Z K A 5564, Forschungszentrum Karlsruhe GmbH, Karlsruhe 1995, (ISSN
0947-8620).
64. M. Thumm, J. of Infrared and Millimeter Wave, 199X, 19, No. I , pp. 3-14.
65. M. Thumm, Fusion and Engineering Design, 1995, 30, 139-170.
66. M . Thumm, FZK Institute, 1997, Private communication.
67. 0. Braz, G. Dammertz M. Kujntze, and M. Thumm, International Journal of Infrared and
Milliineter Waves, 1997, 18(8), 14651477.
68. R. Heidinger, A . Ibarra, and J. Molia, J. Nuclear Materials, 1998, 258-263B, part B, pp. 1882-
1826.
69. R. Sporl, R. Schwab, R. Heidinger, and V. Parshin, C V D Diamondfor High Power Gyrotrons:
Characterisutian of Dielectic Properties, 1TG-Pachbericht, Berlin, April 1998.
70. R. Heidinger, R. Sporl, M. Thumm, J. R. Brandon, R. S. Sussmann, and C. N. Dodge,
Progranime Commitire, 23rd International Conference on Infrared and Millimeter Waves,
Colchester, Essex. UK. ‘CVD Diamond Windows for High Power Gyrotrons’, September
1998, pp. 223-225.
71. 0. Braz, A. Kasugai, K. Sakamoto, K. Takahashi, M. Tsunekoa, T. Imai, and M. Thumm, Int.
J . Infrared Millimeter Waves, 1997, 18(8), 1495-1 503.
72. A. Kasugai, K. Sakamoto, K. Takahashi. M. Tsuneoka, T. Kariya, T. Imai, 0. Braz, M.
Thumm, J. R. Brandon, R. S. Sussmann, A. Beal. and D. C. Ballington, Rev. Sci. Instrum,
1998, 68(5).
73. K. Sakamoto, Japan Atomic Energy Research Institute, Private communication.
74. R. Heidinger. R. Schwab, R. Sporl, and M. Thumm, 22nd International Conference on Infrared
and MiNimeter Waves, Wintergreen. Virginia, July 1997, P. Freund (Ed.).
75. R. Heidinger. FZK Institute, Private communication.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
622 2 Applications of Diamond Synthesized bqi Chemical Vapor Deposition
76. S. M. Hu, Critical stress in silicon brittle fracture, and effect of ion implantation and other
surface treatments, J. Appl. Phys., May 1992, 53(5), 35763580.
77. R. C. Eden, in Handbook of Industrial Diamonds and Diamond Films, M. A. Prelas, G. Popovici,
and L. K. Bigelow (Eds), Marcel Dekker Inc, New York, Basel, Hong Kong, 1998, Chapter 30,
1073-1102.
78. J. Bonhaus, D. Lorenzen, E. Kaulfersch, A. Denisenko, A. Zaitsev, S. Kwong, K. Man,
H. Reichi, and W. R. Fahrner, Proceedings of the International Conference on Micro Materials,
Micro Mat '97, Berlin, April 1997, B. Michel and T. Winkler (Eds), 1997, 505-508.
79. D Lorenzen, S. Heineman, J. Bonhaus, Proceedings of the International Conjerence on Micro
Materials, Micro M a t '97, Berlin, April 1997 (Ed.: B. Michel and T. Winkler), 1997, 513-516.
80. D. Lorenzen, Jenoptik Laserdiode, 1998, Private communication.
81. K. E. Goodson, K. Kurabayashi. and R. F. W. Pease, IEEE Transactions on Components
Packaging and Manujacturing Technology, 1997, Part B. - Advanced Packaging 20-1, 104109.
82. S. Weiss, E. Zakel, and H. Reichi, IEEE Transactions on Components Packaging and Manu-
facturing Technology, 1996, Part A. Vol. 19-1, 46-53.
83. Patent assigned to Jenoptik Laserdiode, DE 196 44 941 CI (D. Lorenzen, 1998. Private
communication).
84. P. K. Sen, M. W. Cook, and R. Achilles, To be published: Proceedings of the Industrial Diamond
Association of America Meeting on Ultrahard Materials, Windsor, Ontario 28-30 May, 1998.
85. S. J. Bull and A. Matthews, Diamondjor Wear and Erosion Applications, Diamond and Related
Materials, 1992, 1, 1049-1064.
86. B. Lux and R. Haubner, Phil. Trans. R . SOC.Lond. A, 1993, 342, 297-311.
87. T. Le Huu, H. Zaidi, and D. Paulmier, Wear, 1977, 203-204,442446.
88. M. N. Gardos in Protective Coatings and Thin Films,Y. Pauleau and P. B. Barna (Eds), Kluwer
Academic Publishers, The Netheriands 1997, 185-196.
89. D. Cooper, F. A. Davies, and R. J. K. Wood, J . Phys. D: Appl. Phys., 1992, 25, A195-A204.
90. A. Edemir and G. R. Fenske, Proceedings of the 51st Annual meeting of the S T L E , Cincinnati -
Ohio, May 1996, 1-8.
91. R. A . Hay and J. M. Galimbarti, Handbook of Industrial Diamonds and Diamond Films, M. A.
Prelas, G. Popovici, and L. K. Bigelow (Eds), Marcel Dekker Inc, New York, Basel, Hong
Kong, 1998, Chapter 33, 1135-1 147.
92. M. J. Neale (Ed.), Tribology Handbook, Butterworths, London, 1973 (2nd Edition 1994).
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
3.1 Introduction
The term ‘diamond-like carbon’ was coined in the early 70s in order to denote amor-
phous carbon films, prepared by physical vapor deposition techniques, which were
expected to approach genuine diamond, as far as density and mechanical properties
are concerned. Today the expression (frequently abbreviated to ‘DLC’) is used for a
fairly wide range of amorphous films containing carbon atoms in different states of
hybridization and possibly also further elements such as hydrogen, silicon, or metals
(‘Si-DLC’, ‘MeeDLC’). There are presently no generally accepted criteria for the
use of the term DLC or for a clear distinction between ‘diamond-like’ on the one
hand and ‘graphite-like’ or ‘polymer-like’ on the other. This review will therefore
pragmatically emphasize those kinds of amorphous carbon-based films which
have found the greatest interest from a scientific or an applicational point of
view. Aside from hydrogen-free amorphous carbon (a-C) and especially highly
tetrahedrally coordinated amorphous carbon (ta-C), deserving the denotation
‘diamond-like’ most of all, hydrogenated films (a-C :H) and metal-containing
hydrogenated films (MeC: H) will be treated in more detail, owing to the technical
importance for mechanical-tribological applications which these materials have
gained. Section 3.2 will present the deposition methods which have been used or
are being used for the preparation of DLC films. Microstructures and properties
will be treated in sections 3.3 and 3.4 and the final section will deal with applica-
tional aspects of DLC coatings.
Substrate
Process pressure: 1 to 10 Pa
-
Current density: 0.6 mA cm-’
Figure 1. Schematic drawing of the radio frequency (RF) plasma technique for deposition of a-C: H
films with some typical process parameters.
practice this condition is usually fulfilled. The self-bias drops across the plasma
sheath in front of the negatively biases substrate electrode down to the plasma
potential Up. The maximum ion energy Eimax of the ions arriving at the substrate
surface is e( Up - u b ) , where e is the unit electron charge. In asymmetric systems
the mean plasma potential is a few volts positive. Therefore u b is a sufficiently
good measure for the maximum ion energy. However, the ions can lose energy
if their mean free path in the plasma sheath is smaller than the thickness of
the sheath. For the pressure range typical for RF de osition of a-C:H ( z 10Pa)
P
Bubenzer et al. [13] derived the relation Ei M U b / P‘ ’ . Catherine and Couderc [6]
confirmed this relation for the R F process. The self-bias u b varies with the discharge
parameters according to Ub (W/P)’’’ where W is the R F power and P the
N
pressure of the process gas [6]. For the film growth rate the relation R z U b P was
found experimentally [ 131. The deposition rate also depends on the type of precursor
C,H, used. The highest rates can be realized with benzene (C6H6)and cyclohexane
(C6HI2),the lowest with methane. One possible explanation is that the rate
decreases with increasing ionization potential of the source gas [14].
The negatively self-biased substrate surface is bombarded with ions which are
generated in the discharge (C,H; and noble gas ions if noble gases are added).
However, not only the C,H; ions introduce carbon into the film. Comparing the
calculated mass deposition rates of major ions (CH; for CH4 gas and C 6 H l as
well as C6Ht for C6H6 precursor) with the experimentally determined deposition
rates Catherine and Couderc [6] found clear discrepancies in the correlations
between rate and R F power and between rate and pressure. In order to overcome
these, the adsorption of activated species and their incorporation into the growing
film must be taken into account additionally to explain the experimental data.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.2 Preparation Methods for Diamond-like Carbon Films 627
stress. The densification can cause a change of the hybridization from sp2 to sp3.
This process occurs in an energy range which obviously depends on the preparation
method used [19]. Considering the experimental data, high sp3 fractions in the a-C
films ( > 70%) were found in the energy range between 20 and 600eV. For too low
energies the ions can not penetrate, will be stuck at the surface and form sp2-carbon.
For too high energies the ion dissipates the energy in a ‘thermal spike’ and relaxation
of the densified regions occurs. At present several methods to prepare ta-C films are
known, but mostly on a laboratory level. Therefore the deposition technique is still
not suitable for broad technological applications as they are already realized in
different fields for a-C: H or MeC: H coatings.
DOS t valence
bands
conduction
bands
EF Energy
Figure 2. Schematic drawing of electron density (DOS) of states of amorphous carbon phases.
EF = Fermi energy (after [14]).
The most important methods which have been in use to reveal the structure and
bonding of a-C and a-C:H films have been summarized in review articles [39, 401.
In order to understand these methods and the electrical and optical properties of
DLC, a closer look to its electronic structure is necessary. Figure 2 is a simplified
schematic diagram of the electron density-of-states (DOS) of amorphous carbon, as
it is frequently shown in the literature (e.g. in [ 141). A pair of o and o*bonding and
antibonding states is due to the bonds forming the skeleton of the a-C or a-C:H
network. Owing to the weaker n: bonds, the n:/n:* couple is located more closely to
the Fermi energy E F .(Strictly speaking, the nomenclature of o and n: molecular orbi-
tals would require the presence of a mirror plane within a molecule [38]. Nevertheless,
this simplified concept is of certain value for a discussion of bonding in DLC.)
The hydrogen content in a-C:H films can be determined by a number of methods,
including classical chemical combustion analysis, nuclear reaction analysis (NRA),
elastic recoil detection (ERD), nuclear magnetic resonance (NMR) spectroscopy, or
an analysis of C-H band intensities in infrared vibrational absorption spectroscopy
(IR). The most reliable information about the relative amounts of sp2- and sp3-
P-
bonded carbon (s bonded C atoms usually do not play an important role) comes
from solid state 1 C-NMR spectroscopy [41] and electron energy loss spectroscopy
(EELS) [4244] (see also the references to other original papers in [39] and [40]). The
results of Weissmantel rt al. already pointed to the existence of a special carbon
structure in films deposited either by dual beam sputtering or by condensation of
species from a low pressure ionization system operated with benzene. Fink et al.
used R F plasma deposition of a-C: H from benzene and measured the EELS spectra
for samples grown at different substrate bias voltages before and after annealing up
+
to 1000°C. The o n plasmon energy positions and intensities are significantly
dependent on the sp3 and sp2 fractions. The concentration of carbon in sp3 config-
uration was found to be roughly two third in as-grown films and increased after
annealing. The hydrogen content declined strongly from 60 to 25 at-%, when the
bias voltage was raised from 200 to 12OOV, accompanied by a significant density
increase (1.49 to 1.75g ~ m - ~A) strong
. dependence of the sp2 carbon and the hydro-
gen content, in a-C:H films grown by R F or DC plasma and ion beam deposition,
was also noted in the NMR measurements performed by Kaplan et al. [41].
A compilation of the sp3 percentages and H contents of typical a-C and a-C:H
films was given by Robertson [45]. Table 2, containing also density, hardness, and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
632 3 Diamond-like Carbon Films
Table 2. Structural characteristics and properties of various kinds of carbon, data taken from an
earlier comuilation 1451.
band gap energy values, compares these special forms of carbon and hydrogenated
carbon to a few 'conventional' materials. It is important to note, that a 100% sp3
content, accompanied with a band gap energy in excess of 5 eV, is found in diamond
as well as in polyethylene, representing the extremes of the hardness and density
scales. In DLC, a high amount of tetrahedral bonding is present in very soft, poly-
mer-like a-C :H films with high H content and low density as well as in amorphous
a-C films, grown by mass selected ion beam deposition (MSIB), which have a low or
zero H content and densities as well as hardness values approaching those of dia-
mond. The latter, ta-C (tetrahedral amorphous carbon) films, which can also be
grown by filtered vacuum arc deposition [46], have in recent years gained consider-
able scientific interest. Several models have been proposed to explain the formation
of these very dense amor hous carbon phases [39], [47]. A prerequisite to obtain
i:
maximum density and sp content is a suitable average kinetic energy of carbon
ions in the 20-600eV region. If the energies deposited per carbon atom during
film growth are much smaller, e.g. during evaporation of graphite ( E M 0.1 eV) or
plasma deposition of an a-C:H film from a hydrocarbon at small substrate bias
voltages, soft films of low density and hardness will be deposited, consisting
predominantly of sp2-bonded carbon (in a-C) and strongly hydrogenated sp3-
bonded carbon (in a-C : H).
Among the methods mentioned above, IR spectroscopy is by far the most
inexpensive and - if the analysis of absorption band intensities is done with great
care - a powerful tool for structural investigations of hydogenated amorphous
carbon films. The interrelation of deposition parameters and structural data of
Table 3 which were reported by Dischler [48], are partly based on IR spectroscopic
data (namely the ratios of sp"-hybridized carbon atoms and ratios of CH,-bonded
segments) and can be considered as typical for plasma deposition of a-C :H. Growth
of a dense, hard phase with high refractive index is favored by high ion energies of
carbon-containg species deposited on the substrate, which are achieved under con-
ditions of high bias voltages and relatively low pressures. In case of the softer a-C: H
film, a large amount of hydrogen is present in methyl groups which do not contri-
bute to the rigidity of the film. For the harder sample, methyl groups are virtually
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.3 Microstructure and Bonding of Diamond-like Carbon 633
absent and all sp3-bonded carbon atoms make at least two bonds to neighboring
carbon atoms, thus contributing to the cross-linking of the network.
Quantitative investigations of the connectivity in a-C: H random networks were
introduced by Robertson [49] and Angus [50], based on the pioneering work on
random covalent networks of Phillips, Thorpe and Dohler et al. [51-531. In consid-
erations of this kind, the number of translational degrees of freedom of atoms (= 3
for one atom in three-dimensional space), forming the random network, are
balanced out against the number of constraints N,,, on their movement imposed
by bonds to neighboring partners. It can be shown that in three-dimensional net-
works whose bonding is dominated by directed valence bonds to nearest neighbors
the number of constraints for one atom is
Ncon(r)= r2/2 for Y 5 2 and Ncon(r)= 5 r / 2 - 3 for r 2 2
(1)
r is the coordination number. As long as there are less constraints than degrees of
freedom in a random network, atoms can still move in order to minimize strain
energy, the network is 'floppy'. As soon as the degrees of freedom are exhausted
by an increased number of bonds, the network will be fully constrained, additional
cross-linking will then increase the strain energy rapidly. For a monatomic network
this happens at an average coordination number r,, = 2.4. For a completely con-
strained network consisting of sp2- and sp3-bonded carbon atoms as well as
hydrogen atoms with molar ratios x3,x4 and x I , Angus [50] deduced an equation
for the sp3/sp2ratio (x4/x3)as a function of the hydrogen content xl :
xq/x3 = (8x1 - 3)/(8 - 13x1) (2)
In order to be composed exclusively from sp3 carbon atoms (corresponding to a
zero denominator in Eq. (2)) and still be not overconstrained, a network should,
according to Eq. (2), contain a hydrogen atom fraction x1 of 8/13 = 0.615 [54].
Modification have to be introduced, however, in order to take account of the clus-
tering of sp2-bonded carbon [55]. In early papers on DLC structure and properties,
extended clustering of the sp2-bonded carbon atoms to graphite-like aromatic
planes was invoked in order to explain, among other properties, the small band
gaps of a-C and a-C:H. For a band gap of, say, 1.2eV, typical for hard a-C:H,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
634 3 Diamond-like Carbon Films
W-C:H
0,Ol 0,'
q
Figure 3. Small angle X-ray scattering (SAXS) curves for WC: H films with differing metal contents.
Continuous lines = simulation, see text. (From [57] by courtesy of K. I. Schiffmann).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
636 3 Diamond-like Carbon Films
volume fraction, among a majority of smaller ones. The continuous lines in Figure 3
are simulated curves, based on a model of hard homogeneous spheres in a
homogeneous matrix. In order to represent the particle sizes, a combination of
two log-normal distributions was used, one main distribution representing the
majority of particles with a relatively small width, plus an additional broad back-
ground distribution representing few, much larger particles. Particle radii and dis-
tances ranging from 0.6 to 0.8nm and 2 to 3nm, resp. could be deduced from
these investigations for W-C:H films with W contents between 7 and 25 at-%.
Metal and carbide particle distances and radii in MeC:H films have also been
measured by applying scanning tunnel microscopy (STM) imaging. In order to
take into account tip convolution effects, leading to an appearant particle enlarge-
ment and to particle hiding, statistical methods were proposed by Schiffmann et d.,
using an off-line tip radius determination by analysing appearant particle radii and
a Monte Carlo model to compute fractions of hidden particles. A very good
agreement was generally observed for radii and distances as measured by STM,
TEM, SAXS and (for the radii only) XRD [57, 601.
The anisotropies in two-dimensional SAXS measurements as observed by Fryda
[59] using sample tilting with respect to the incoming synchrotron radiation beam
are, according to TEM micrographs of film cross sections [57], not due to an
elongated shape of the segregated particles, but possibly to a chain-like ordering
along the growth direction. The particles themselves are usually nearly spherical
and do not exhibit a pronounced shape anisotropy along the growth direction.
The average distance d of the particles, at a given volume fraction directly related
to their average size, correlates quite closely with the melting point T, of the
corresponding bulk metal. The plot against T i 1 in Figure 4 is based on the evalua-
tion of the position of the maximum or shoulder in the Guinier region of SAXS
scattering curves for a number of MeC:H films [59] plus the data for Au from
[57]. Qualitatively, an increasing mobility of the atoms of lower melting metals
80
70
60
.a 50
\
TJ 40
30
20
10
2 3 4 5 6 7 8
T~
; I I O - K-?
~
Figure 4. Plot of the average center of mass distances d (in A) of segregated metal and carbide par-
ticles, respectively, in MeC: H films against the reciprocal melting points l/T, of the corresponding
bulk metals, according to data taken from [59] and, for Au, [57].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
3.4 Physical Properties of DLC Filnis 637
during film growth and segregation of the metal or metal carbide phases can be
made responsible for their increasing sizes and distances.
and the question of its dopability is presently under discussion. Data published for
undoped ta-C exhibit a large spread in room temperature conductivity, from to
10-12 0-1 cm-l
, even for quite similar sp3 fractions [39]. Amaratunga et al. were the
first to grow ta-C (filtered arc deposition) with additions of N, P, and B [68]. The
conclusions drawn in this paper and in subsequent publications from this group
have recently been critically reviewed by Hofsiiss [69] on the basis of their own
experimental results. According to his investigations, the conduction in undoped
ta-C is mainly due to hopping in band-tails states and (at higher electrical fields)
Poole-Frenkel excitation of electrons from traps into the conduction band.
Doping with N or B leads to an increase of conductivity by 3 4 orders of magnitude
at room temperature, presumably due to a generation of localized states related to
three-fold coordinated B or N sites, contributing to hopping as well as Poole-
Frenkel conduction. This view is supported by completely symmetrical I- V curves
measured for layered diode-like structures containing N- and B-doped material.
Electrical properties of MeC :H films, deposited by reactive sputtering from
metal or metal carbide targets in hydrocarbon atmosphere, have been described
elsewhere [32]. In case of Ta as a typical refractory, carbide-forming metal, the
electrical behavior can be divided into three categories, depending on the metal
content:
- a 'dielectric' region with thermally activated conductivity according to Eq. (3) at
less than about 13 at-% Ta,
- a transition region with metallic behavior (i.e. residual conductivity at vanishing
temperature) but still a positive or zero slope of log(o)-T curves and
o > lo2R-' cm-' at medium Ta contents (14 and 22 at-% Ta) and
- a metallic region with with negatively sloped curves log(o)-T and
o > lo3R-' cm-' for 44 and 94 at-% Ta.
In contrast to Au- and Cu-C:H, for example, showing a percolation-type sudden
increase of room temperature conductivity at a medium metal content of several
10 vol-%, the o ( R T ) values in Ta-C:H films tend to increase continuously by
about six orders of magnitude right from the lowest concentrations in nearly
metal free films up to less than 10 at-% Ta. It is only at low temperatures, that
an insulator/metal transition is observable as a four orders of magnitude jump in
o(3 K) between 11.5 and 14.3 at-%. A detailed discussion of the electrical properties
of MeC :H films is given elsewhere [70].
Similar to the electrical conductivities optical properties like absorption
coefficients and refractive indices are critically dependent on the preparation
conditions used to grow a-C or a-C :H. Except for absorption coefficients below
about lop4 cm-', the photon energy E dependence of the absorption coefficient M
can generally be described by the Tauc relation (4) [71]:
for absorptions in and near to the visible region. Typical values for a-C and a-C: H are
shown in Table 2. While an upper limit of 1.5eV is still quoted in this survey for ta-C,
prepared by mass selected ion beam deposition, considerably higher values approach-
ing or even exceeding 3 eV seem now feasible according to the review [39]. Thus, the
very hard and wear resistant hydrogen-free ta-C films might gain importance for opti-
cal applications in the visible region.
Refractive indices for ta-C films are typically close to the value of diamond
(n = 2.4) [39]. Incorporation of hydrogen lowers n to values between about 1.8
and 2.3 in a-C:H. In plasma deposition of these films, hydrocarbon pressure P
and substrate bias voltage UB are the most important growth parameters allowing
a tuning of n. Dense films with high refractive indices are favored by low P and high
UB,To a first approximation, holding for n up to 2, constant refractive indices can be
achieved for a constant ratio U B / P[62].
interesting MeC: H films with low metal contents up to about 20 at-%. However, the
essential results basically hold also true for DLC coatings in general. According to
these considerations, the exceptional mechanical-tribological properties of DLC can
be attributed to
~ hardness values (~10-100GPa) similar or even exceeding those of ceramic
inorganic materials (10-20 GPa) and by far larger than for metals (< 8 GPa),
- elastic moduli (~100-1000GPa) similar or exceeding metals and ceramics
(< 600 GPa), combined with
- elasticities as measured by H / E ratios (0.1-0.2) which are throughout larger than
for metals (<0.03) and ceramics (<0.06) and approach the lower limit of a range
which is more typical for organic polymers (0.2-0.5), and
- for hydrogenated film surfaces, surface energies (30-50 x N/m) in a region
typical for polymers and much less than for metals (300-2000 x N/m) and
ceramics (200-1300 x lop3N/m) (see [33] for references to the sources used).
Friction coefficients, being not a physical property in the strict sense, but depending
on the contact partners and many other external parameters, will be reported in the
next section. The combination of low friction coefficients, typically 5 0.2, with
mechanical and surface properties which are partly polymer-like (elasticity, surface
energy) and partly ceramic- or metal-like (hardness, Young modulus) qualitatively
explains the outstanding position of DLC films as low-friction, highly wear-resistant
coatings on which most of their present applications are based [77].
. =.
b
0
0
0 o without
0 ' I I
0 2 4 6 8 1 0 1 2 1 4
6[ P ~ I
Figure 5. Critical load L, against a-C: H film thickness df for deposition with Al- and Ti- interlayers
or without interlayers, respectively, on polished steel substrates.
adhesion the critical load, L,, or the crack pattern caused by a Rockwell indentor is
frequently used. The critical load is a measure determined by scratch tests [84],
utilizing a diamond tip moving over the film with continuously increasing normal
load. L, is the load where an intensive acustic emission occurs, the friction coefficient
between tip and sample drastically increases or a failure of the film can be observed
by a microscope. All three criterions are in use. Therefore a comparison of data from
different laboratories is difficult. Figure 5 shows the effect of aluminum and titanium
interlayers on the critical load as a function of the a-C:H film thickness df.L, was
determined with the second and third criteria. The values determined with both
methods were nearly identical. These films were prepared by R F plasma deposition
from acetylene after magnetron sputtering of the metal interlayers (thickness about
0.5 pm) onto polished steel substrates in the same coating machine. In particular in
the range df > 3 pm a pronounced improvement of the adhesion was found. The
hardness of these coatings was about 25GPa. Even films thicker than lOpm
having excellent adhesion can be achieved by this method. It should be noted that
the effect of interlayers is still more pronounced if substrates with rougher surfaces,
like ground substrates are to be coated.
However, the reliable and reproducible preparation of well-adhering a-C :H films
is commonly not a simple task and requires a good control of several process param-
eters. Traces of oxygen in the films, to give just one example, can cause a drastic
deterioration of the film adhesion.
On the other hand, there are very reliable industrial processes for deposition of
MeC: H coatings. MeC: H films have lower stresses than a-C: H (< 1 GPa) [34, 851
Therefore their tendency to delaminate is considerably reduced. Furthermore,
intermediate layers can be prepared very easily using non-reactive sputtering from
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
642 3 Diamond-like Carbon Films
-
ap 120
I
140[
E 100
a
L
80-
Q
'=m
>
- 60-
PI
K
40 -
"
Hard steel Ti N MeC D LC
metal (100Cr6)
Figure 6. Abrasive wear of DLC coatings compared to other coating and substrate materials. As
abrasive medium S i c with a particle dimension of 12 Fm was used.
the metal target (see Section 3.2.3.), before the MeC: H deposition process is started.
1ooc
0
w
0 looor w
.-: 1oc
s-
?!E
4-
gF
L
1c
V
:g
.->'
aJ
v)
1
r
a
-0
0.1 AI E
Y
r" Y2
Figure 7. Adhesive wear measured in a ball on disk arrangement: (a) MeC: H and TiN coated discs
and a disc of ball bearing steel 100Cr6 and (b) wear of the counterpart (100Cr6 ball). Measurements
were performed under ambient conditions.
o'2
01
0.01
I 0.1 1
I
10
Normal load "1
Figure 8. Friction coefficients against normal load for MeC:H, TiN, TiAlN and steel measured
against ball bearing steel under ambient conditions (up to about 50% relative humidity).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
644 3 Diamond-like Carbon Films
drastically declined from p > 0.2 to p < 0.05 with decreasing relative humidity
(< 1%) [89,90]. For silicon-containing films very low p values (<0.1) were measured
even up to relative humidities of 75% [88, 911. A possible reason for the different
friction behavior could be a material transfer from the film to the steel ball (decreas-
ing p) or in the opposite direction (increasing p) [90]. In the case of silicon-
containing films reactions of Si with the atmosphere were assumed. The generated
silicon ceramics are known to have extremely low friction coefficients in water
[911. However, there are still several open questions concerning the friction
mechanisms of DLC films. It should be noted here that any published values of
friction coefficients must be considered with caution because measuring conditions
like humidity, loads, material of counterparts or sliding distance, which can have a
strong influence on the result, can be extremely different [78].
films prepared by ion beam decomposition, R F plasma and pulsed arc deposition,
for coating implants. In spite of different film properties all three materials turned
out as very suitable.
Ion beam deposited DLC films are used to increase the abrasion and chemical
resistance of sunglass lenses, of float-glass windows in supermarket laser scanners
and of infrared windows [95].Very thin amorphous carbon films are used to protect
magnetic disks and heads. Recently it was found that especially nitrogen containing
carbon films (CN,) have an excellent wear resistance (about four-fold value of a-C)
[96]. One reason for this difference is presumably the different roughness of the films.
Using an atomic force microscope (AFM) roughness values r.m.s. of 0.25 nm and
0.63 nm, respectively, were measured for CN, and for a-C.
For the application of coatings, also cost and benefit aspects must be considered. It
shall be noted here that Matthews and Eskildsen [78]discussed this topic in detail for
four different DLC production methods taking into account prices for deposition
equipments to be obtained in the UK.
References
1. L. Holland and S. M. Ojha, Thin Solid Films 1976,38,L17-LI9.
2. D.S. Whitmell and R. Williamson, Thin Solid Films 1976,35,255-261.
3. A. Grill and V. V. Patel, Diamond Films Technol. 1992,4,219-233.
4. L.Holland, J . Vac. Sci. Technol. 1977,14, 5-15.
5. K. Bewilogua and D. Wagner, Vacuum 1991,42,473476.
6. Y.Catherine and P. Couderc, Thin Solid Films 1986, 144, 265-280.
7. C.Weissmantel, K. Bewilogua, K. Breuer, D. Dietrich, U. Ebersbach, H.-J. Erler, B. Rau, and
G. Reisse, Thin Solid Films 1982,96, 3144.
8. C. Weissmantel, in Thin Films from Free Atoms and Particles, J. Klabunde (Ed.), Academic,
New York, 1985,pp. 153-201.
9. T. Namba and Y. Mori, J . Vac. Sci. Technol. 1983,Al, 23-27
10. M. W. Kroesen, D. C. Schram, and M. J. F. van de Sande, Plasma Chem. Plasma Proc. 1990,
10,49-69.
11. L.Martinu, A. Raveh, A. Domingue, L.Bertrand, J. E. Klemberg-Sapieha, S. C. Gujrathi, and
M. R. Wertheimer, Thin Solid Films 1992,208,4247.
12. B. Chapman, Glow Discharge Processes, Wiley, New York, 1980,Chapter 5.
13. A. Bubenzer, B. Dischler, G. Brandt, and P. Koidl, J . Appl. Phys. 1993,54,45904595.
14. J. Robertson, Surf. Coat. Technol. 1992,50, 185-203.
15. W. Moller, Appl. Phys. A 1993,56, 527-546.
16. H. Kersten and G. M. W. Kroesen, J . Vac. Sci. Technol. 1990,A8, 3842.
17. D. Roth, B. Rau, S.Roth, J. Mai, and K.-H. Dittrich, Surt Coat. Technol. 1995,7475,637-641.
18. S.Aisenberg and R. W. Chabot, J . Appl. Phys. 1971,42,2953-2958.
19. Y. Lifshitz, Diamond Re/. Muter. 1996,5, 388400.
20. J. J. Cuomo, D. L. Pappas, J. Bruley, J. P. Doyle, and K. L. Saenger, J . Appl. Phys. 1991,70,
1706171 1.
21. H. Hofsass, H. Binder, T. Klump, and E. Recknagel, Diamond Rel. Muter. 1993,3, 137-142.
22. N.Savvides, Thin Solid Films 1988, 163, 13-32.
23. N . Savvides and B. Window, J . Vac. Sci. Technol. 1986,A4, 504-508.
24. A. Anders, S. Anders, and I. G. Brown, Plasma Sources Sci. Technol. 1995,4, 1-12.
25. H.-J. Scheibe, B. Schultrich, and D. Drescher, Surf. Coat. Technol. 1995,74-75, 813-818.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
646 3 Diamond-like Carbon Films
26. H.-J. Scheibe, D. Drescher, B. Schultrich, M . Falz, G. Leonhardt, and R. Wilberg, Surf: Coat.
Technol. 1996, 85,209-214.
27. A. A. Voevodin and M. S. Donley, Su$ Coat. Technol. 1996,82, 199-213.
28. F. Rossi and B. Andre, Jpn. J. Appl. Phys. 1992, 31, 872 -879.
29. J. Robertson, Diamond Rel. Muter. 1993, 2, 984-989.
30. H. Dimigen and H. Hubsch, Philips Tech.Rev. 1983/84,41, 186-197.
31. H. Dimigen and H. Hiibsch, R. Memming, Appl. Phys. Lett. 1987,50, 10561058.
32. C.-P. Klages and R. Memming, Muter. Sci. Forum, 1989, 52-53, 609-644.
33. H. Dimigen and C.-P.Klages, Surf. Coat. Technol. 1991,49, 543-547.
34. E. Bergmann and J. Vogel, J . Vac. Sci. Technol. 1987, A5, 70-74.
35. K.Bewilogua and H. Dimigen, Surf. Coat. Technol. 1993,61, 144-150.
36. D. Hofmann, H. Schuessler, K. Bewilogua, H. Hubsch and J. Lemke, SwF Coat. Technol. 1995,
73, 137-141.
37. J. T. Harnack and C. Benndorf, Diamond Rel. Muter. 1992, 1, 301-306.
38. F. L. Pilar, Elementary Quantum Chemistry, MacGraw-Hill, New York, 1968.
39. Y. Lifshitz, in The Physics of Diamond, A. Paoletti and A. Tucciarone (Eds), 10s Press, Amster-
dam, 1997, p. 209.
40. J. Robertson, in Diamond and Diamond-like Films and Coatings, R. E. Clawing et al. (Eds),
Plenum, New York, 1991, p. 331.
41. S. Kaplan, F. Jansen, and M. Machonkin, Appl. Phys. Lett. 1985, 47, 750-753.
42. C. Weissmantel, K. Bewilogua, D. Dietrich, H.-J. Erler, H.-J. Hinneberg, S. Klose, W. Nowick,
and G. Reisse, Thin Solid Films 1980, 72, 19-31.
43. S. Miihling, K. Bewilogua, and K. Breuer, Thin Solid Films 1990, 187, 65-75.
44. J. Fink, Th. Muller-Heinzerling, J. Pfluger, B. Scheerer, B. Dischler, P. Koidl, A. Bubenzer, and
R. E. Sah, Phys. Rev. B 1984, 30,47134718.
45. J. Robertson, Diamond Rel. Mater. 1992, 1, 397406.
46. D. R. McKenzie, D. Muller, B. A. Pailthorpe, Z. H. Wang, E. Kravtchinskaia. D. Segal, P. B.
Lukins, P. D. Swift, P. J. Martin, G. Amaratunga, P. H. Gaskell, and A. Saeed, Diamond Rel.
Muter. 1991, 1, 51-59.
47. H. Hofsass, H. Feldermann, R. Merck, M. Sebastian, and C. Ronning, Appl. Phys. A , 1988,
A66, 153-181.
48. B. Dischler, in Proceedings of E-MRS Symposium on Amorphous Hydrogenated Carbon, Sympo-
sium C, Strasbourg, France, Les Editions de Physique, Paris, 1987, XVII, pp. 189-201.
49. J. Robertson, Adv. Phys. 1986,35, 317-374.
50. J. C. Angus, in Proceedings of E-MRS Symposium on Amorphous Hydrogenated Carbon,
Symposium C , Strasbourg, France, Les Editions de Physique, Paris, 1987, XVII, pp. 179-188.
51. J. C. Phillips, Phys. Rev. Lett. 1979, 42, 153-181.
52. M. F. Thorpe, J. Non-Cryst. Sol. 1983, 57, 355-370.
53. G. H. Dohler, R. Dandoloff, and H. Bilz, J. Non-Cryst. Sol. 1980, 42, 87 96.
54. J. C. Angus, F. Jansen, J. Vac. Sci. Technol. 1988, A6, 1778-1782.
55. J. Robertson, Diamond Rel. Mater. 1997, 6, 212-218.
56. C.-P. Klages, A. Dietz, T. Hoing, R. Thyen, A. Weber, P. Willich, Surf. Coat. Technol. 1996,80,
121-128.
57. K. I. Schiffmann, Thesis, University of Hamburg, published as Fraunhofer-IST-Berichte aus
Forschung und Entwicklung, Fraunhofer IRB Verlag, Stuttgart, 1997, Nr. 1.
58. P. Willich and R. Bethke, in Secondarji Ion Mass Spectroscopy ( S I M S X i , A. Benninghoven,
B. Hagenhoff and H. W. Werner, Wiley, New York, 1996.
59. M. Fryda, Fortschritt-Berichte VDI, Reihe 5 , VDI Verlag, Dusseldorf, 1993, 303, p. 20.
60. K. I. Schiffmann, M. Fryda, G. Goerigk, R. Lauer, and P. Hinze, Ultramicroscopy 1996, 66,
183-192.
61. K. I. Schiffmann, M. Fryda, and G. Goerigk, Mikrochinz. Acta 1997, 125, 107-1 13.
62. J. C. Angus, P. Koidl, and S. Domitz, in Plasma Deposited Thin Films, J. Mort and F. Jansen
(Eds), CRC, Boca Raton, FL, 1986, p. 89.
63. Th. Frauenheim, U. Stephan, K. Bewilogua, F. Jungnickel, P. Blaudeck, and E. Fromm, Thin
Solid Films 1989, 182, 63-78.
64. H. Overhof, P. Thomas, Electronic Transport in Hydrogenated Amorphous Semiconductors,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 647
Springer Tracts in Modern Physics, Vol. 144, G. Hohler (Ed.), Springer Verlag, Berlin, 1989, p.
10.
65. R. Zallen, The Physics ofAmorphous Solids, Wiley, New York, 1983, p. 252.
66. B. Meyerson and F. W. Smith, Solid State Commun. 1982, 41, 23-27.
67. J. U. Thiele, B. Rubarth, P. Hammer, A. Helmbold, B. Kessler, K. Rohwer, and D. Meissner,
Diamond Re/. Muter. 1994, 3, 1103-1 106.
68. G. A. J. Amaratunga, D. E. Segal, and D. R. McKenzie, Appl. Phys. Lett. 1991, 59, 69-71.
69. H. Hofsass, in Proc. 1st Int. Specialist Meeting on Amorphous Carbon ( S M A C '97).
Amorphous Carbon: State of the Art, S.R.P. Silva et al., (Eds.), Singapore: World Scientific,
1998, p. 296&310.
70. H. Koberle, PhD Thesis, University of Hamburg, 1989.
71. J. Tauc, R. Grigorovici, and A. Vancu, Phys. Stat. Sol. 1966, 15, 627-637.
72. P. Wierenga, A. J. J. Franken, J . Appl. Phys. 1984, 55, 4244-4248.
73. J. L. Loubet, J. M. Georges, 0. Marchesini, and G. Meille, J . Tribol. 1984, 106, 4 3 4 8 .
74. X. Jiang, K. Reichelt, and B. Stritzker, J . Appl. Phys. 1989, 66, 5805-5808.
75. X. Jiang, K. Reichelt, and B. Stritzker, J . Appl. Phys. 1990, 68, 1018-1022.
76. M. Fryda, K. Taube, and C.-P. Klages, Vacuum 1990,41, 1291-1293
77. E. Rabinowicz, Friction and Wear of Materials, Elsevier, Amsterdam, 1987.
78. A. Matthews and S. S. Eskildsen, Diamond Re/. Muter. 1994, 3, 902-91 1.
79. M. Tamor, in Proc. Third Int. Conf on Applications of Diamond Films and Related Materials, A.
Feldman et a/. (Eds), NIST Special Publications 885, 1995, p.691.
80. R. Butter, M. Allen, L. Chandra, A. H. Lettington. and N. Rushton, Diamond Re/. Muter. 1995,
4, 857-861.
81. M. Grischke, K. Bewilogua, K. Trojan, and H. Dimigen, Surf: Coat. Technol. 1995,7475,739-
745.
82. H. Ronkainen, J. Vihersalo, S. Varjus, R. Zilliacus, U. Ehrnsten, and P. Nenonen, Surf. Coat.
Trchnol. 1997, 90, 190-196.
83. D. P. Monaghan, D. G. Teer, P. A. Logan, I. Efeoglu, and R. D. Arnell, Surf. Coat. Technol.
1993, 60, 525-530.
84. P. J. Burnett and D. S. Rickerby, Thin Solid Films 1988, 157, 233-254.
85. M. Grischke, K. Bewilogua, and H. Dimigen, Mater. Manuf. Proc. 1993, 8, 407417.
86. H.-J. Scheibe and D. Klaffke, Surf. Coat. Technol. 1993, 57, 111-1 15.
87. S. Miyake, Surf: Coat. Technol. 1992, 5455,563-569.
88. K. Oguri and T. Arai, Thin Solid Films 1992, 208, 158-160.
89. K. Enke, H. Dimigen, and H. Hiibsch, Appl. Phys. Lett. 1980,36,291-292.
90. R. Memming, H. J. Tolle, and P. E. Wierenga, Thin Solid Films 1986, 143, 3 1 4 1 .
91. K. Oguri and T. Arai, J . Mater. Res. 1992, 7, 1313-1316.
92. K. Taube, M. Grischke, and K. Bewilogua, Surf. Coat. Technol. 1994, 68/69, 662-668.
93. R. S. Butter and A. H. Lettington, Proc. Third Int. Con6 on Applications of Diamond Films and
Related Muterials, A. Feldman et aI. (Eds), NIST Special Publications 885, 1995, p.683.
94. S. Mitura, P. Niedzielski, D. Jachowicz, M. Langer, J. Marciniak, A.Stanishevsky, E. Tochitsky,
P. Louda, P. Couvrat, M. Denis, and P. Lourdin, Diamond Re/. Muter. 1996, 5, 1185-1 188.
95. F. M. Kimock and B. J. Knapp, Surf. Coat. Technol. 1993,56, 273-279.
96. D. Li, E. Cutiongco, Y.-W. Chung, M.-S. Wong, and W. D. Sproul, Diamond Films Technol.
1995,5, 261-272.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
Table 1. Commercially available tool materials for machining metals. HVl (testing load 1 kg) and
HVlO (10 kg load) hardness values are given for grinding and cutting materials to avoid the misuse
of microhardness data
generally less compared with alumina because of its poor ability to machine ferrous
workpieces).
On the other hand, there has been some stagnation in the field of tools with
geometrically defined cutting edges after the development of advanced A1203/
ZrO, and A1203/TiC composites at the beginning and of tough Si3N4 cutting
tools and Sic-whisker reinforced grades at the end of the eighties (high-strength
tetragonal zirconia polycrystals cannot be used for machining metals). At least
until 1997 it was the position of most manufacturers that they can satisfy the current
requests of their customers with the available ceramics. Consequently, many labora-
tories are busy performing extended screening tests with commercial grades. A
somewhat special position has to be attributed to efforts intended to improve the
chemical resistance of Si3N4 by surface coatings. A very recent tendency is the
increasing interest in a new class of materials with submicrometer microstructures,
a trend which is discussed in Section 4.4.
The strong motivation for these developments in grinding materials is the real
possibility of new inventions able to surpass the present state of the art by 5&100%
(in terms of rate of material removal or maximum time of use of a tool), the chance
to find materials that redouble the productivity of the industrial grinding operation
and reduce substantially the time-out for the change of tools.
Significant recent changes in the market shares of the different kinds of cutting
ceramics shown in Table 1 should be noted. There has been an increase in the
use of ceramic cutting tools, attributed to increasing use of silicon nitride tools.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
650 4 Ceramics Based on Alumina: Increasing the Hardness,for Tool Applications
Two different reasons exist. First, the requirement for high reliability in automated
manufacturing lines (without running a tool to the possible maximum time of use)
favors the application of tool materials with a high fracture toughness (like Si3N4)
tolerating, in the mean time, their lower hardness. Second, a shift in the ratio of
machined materials from steel towards globular and vermicular cast iron also
requires tougher tools like silicon nitride, especially when high feed rates are neces-
sary. A drawback of nitride tools is their rather high chemical reactivity with iron:
against steel and cast iron (like GG25), at feed rates of about 0.3 mm rev-' the wear
of Si3N4is usually higher than for A1203-basedceramics. Hence, even for machining
cast iron, nitride tools can be applied with advantage only when it is impossible to
use the less tough AI2O3/ZrO2or A1203/TiCcomposites due to the required high
feed rates. Attempts have been made to overcome this problem by TIN and/or
oxide coatings, but real success was obtained only in some very special cases.
There is another trend, which has not yet altered market shares, because the new qua-
lities are not commercially available: this is the substitution of certain grinding opera-
tions by highly precise cutting approaches, especially in machining very hard materials.
It is expected that turning may cut the machining time to 1/3 compared with grinding,
and is sometimes associated with significant savings of consumed energy, and even the
compromise of first turning with subsequent final grinding reduces the time to 2/3 com-
pared with an all-grinding process. Additional advantages are the less expensive recy-
cling of turning chips compared with the waste disposal of the grinding silt and
investment costs of the equipment which can be reduced by a half for precise turning
[ 1,2]. On the other hand, investment in new machine tools is imperative for precise cut-
ting of hard materials (the mechanical stability of the most machine tools used today in
the automotive industry is probably not sufficient for the new demands). More wide-
spread use of cutting rather than grinding is restricted by the limited lifetime of today's
cutting ceramics and the increase in workpiece roughness after rather short cutting
paths. Investigations by large-scale manufacturers indicate a lack of real alternatives
to cBN for turning hard workpieces with high feed rates, and drilling and milling is
the field of hard metal tools, but for precise finishing AI2O3/TiCcomposites exhibit
an advantage in the cost-performance ratio [2]. It is expected that such findings will sti-
mulate the development of significantlymore fine grained and harder cutting tool cera-
mics based on corundum. For example, it was in 1997 that a German company
launched a new submicron A1203/Ti(C,N)composite. Indeed, such A1203/Ti(C,N)
tools show the fewest problems when compared with other commercial ceramics
(which, however, all have not been developed to meet the demands of precise machining
of hard surfaces), but an analysis of the technical demands under thls special condition
points to better opportunities with other ceramics (cp. Section 4.4.4).
understanding of the required basic mechanical parameters, and of the real chances
new technologies offer for their improvement. Also, it must be emphasized that the
properties of different new tool materials can be compared only if the effect of their
different microstructures is not obscured by processing-related imperfections: the
investigation of new materials requires samples that are different in their micro-
structures but with similarly small frequency and size of defects. Much care has
to be taken to perform this recent research (the results of which are described
below) because for extremely fine-grained microstructures this demand is not
easy to meet.
Figure 1. Relicts of hard powder granules (from spray drying) in pressureless sintered alumina cera-
mics (solid state sintering).
Typical flaws like the small pores in Fig. 2 are too small to be seen with the optics
of most hardness testers. Hence, an evasive positioning of indents is impossible,
and the testing gives a hardness which will be considerably lower than the inherent
hardness of the perfect submicrometer microstructure. In fact, this reduced hardness
represents a ‘true’ technical property of the material (because the flaws will affect
other hardness-related properties as well, e.g. the tribological behavior), and the
real drawback comes from the limitation of assessable properties and not from
the influence of measuring procedures.
flaw populations are characteristic of the various technical approaches and cause
processing-related property-limits which often cannot be overcome by materials
development [31.
As to new raw materials, the most significant step has been the development of
chemically derived substitutes for materials of fundamental industrial importance
like alumina or zirconia (most grades of which were originally manufactured
from mineral sources). Since the late eighties, new corundum powders have been
commercially available with fine particle sizes (rz 0.2 pm), which offers chances
for developing extremely fine-grained microstructures with a high density. The
small size and the increased curvature of the particle surface are expected to
increase the driving forces of solid state diffusion on sintering, to reduce the
required temperature, and to reduce grain growth. However, these potential
improvements are often offset by the increased tendency of the fine-grained
powders to agglomerate on sintering a body which contains agglomerates, densifi-
cation proceeds first inside these agglomerates (where the particles are in closer
contact), and, as a consequence, with the shrinkage of the agglomerates their
boundaries will move away from each other [4]. Then, to eliminate the sintering-
related new (and rather coarse!) interagglomerate porosity, much higher tem-
peratures are required than expected for the fine-grained powder used. And, of
course, the impact of the high temperatures on these fine particles is an undesired,
strong grain growth (first starting inside the agglomerates). Therefore, for many
years the published progress achieved with using these new raw materials was
much less than hoped.
Success came from increasing understanding of the relationship between sinter-
ing temperature, obtained density, and grain size as a function of spatial homoge-
neity and of interfaces, similarly to the microstructural dependence of the strength
and the hardness of ceramics [3,4]. The general idea concerning the mechanism of
sintering has been outlined above. The expertise starts from the simultaneous or
sequential application of several dispersal approaches to bring the extremely
fine-grained powders into a state that gives the right basis for the adjusted use
of different shaping procedures. Successful experiments were reported with direct
coagulation casting (DCC) [5] and centrifugal compaction [6] as well as with
cold isostatic pressing of freeze dried granules, pressure filtration, and gelcasting
[3]. Cold isostatic pressing easily gives dense sintered products with an average
crystal size of 0.5 pm. Gelcasting means shaping by casting a (preferentially
aqueous) slurry which contains a small percentage of a polymerizable additive.
After polymerization, the soft but geometrically stable body is transferred to
thermal processing. A bending strength of 700-900 MPa was reported for
pressureless sintered single phase alumina bodies (which is about twice the conven-
tional value) [3].
It is interesting to note that it is really the improved homogeneity of compaction
of the powder particles in the ‘green’ bodies that is the key to smaller grain sizes
with an increased hardness, wear resistance, and strength: in spite of a similar
green density (similar starting porosity at the beginning of sintering) of all
bodies produced by different shaping, the sequence dry (uniaxial) pressing +
cold isostatic pressing + pressure filtration + gelcasting gives a dramatic decrease
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.2 Technological Essentials f b r Producing Hard and Strong Tool Ceramics 655
of the sintering temperature by about 250°C (always using the same raw powder
and avoiding any doping additives). Therefore, the improved homogeneity of com-
pacted particles in the given sequence is the one possible reason for the improved
sintering behavior.
For most machining operations where really hard tools are required (e.g. grinding
or cutting of hardened steel), either more complex strength properties than a
maximum bending strength of compact bodies are required, or the macroscopic
strength is generally less important owing to the applied small feed rates (as in
precise cutting). Therefore, advanced principles for obtaining a high macrostrength
for sintered shaped bodies manufactured from these tool ceramics [ 3 ] will not be
discussed here in detail, whereas special emphasis shall be given to prospects for
improved hardness and wear resistance.
v
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
656 4 Ceramics Based on Alumina: Increasing the Hardness f o r Tool Applications
2*1o'6-
1'1 0-l6
-
8'1 0 --
1~
6*1Oei7-
4*10-17-
2*10-j7
0.3 0.6 1.0 3.0 0.3 0.6 1.0 3.0 0.3 0.6 1.0 3.0
D @ml D WnI D @mI
Figure 5. Crystallite size and environmental effects on fretting wear of Al2O3-Al2o3couples.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
657
Figure 6. Increasing grain size promotes grain boundary fracture and grain pull-out (fretting wear
in dry air; cf. Fig. 5 ) . Average grain sizes: (a) 3 pm; (b) 1.6pm; (c) 0.6pm (bottom) [I I].
4.2 Technological Essentials for Producing Hard and Strong Tool Ceramics
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
658 4 Ceramics Based on Alumina: Increasing the Hardness,for Tool Applications
When comparing data from different sources, the condition of the surface inves-
tigated has to be noted: machining introduces high dislocation densities [8], and the
hardness of ground ceramic surfaces is usually higher than measured on polished
surfaces, an effect that is the more significant the more fine-grained the microstruc-
ture is (Fig. 3).
10 15 20 25 30
Vickers hardness HV1 [load = 1 kg] (GPa)
Figure 7. Influence of the hardness of the grinding grit on the rate of removed material on grinding
the cross-section of a steel tube (flexible disc, different alumina grits; the results have been normal-
ized for a unified grit toughness of KIc = 3.6MPam”’ to demonstrate the hardness effect without
the additional influence of Klc) [12].
procedures had to be found to obtain a ‘green’ density of the intermediate (not yet
sintered) product that is high enough to enable sintering until a high density with
less than 1-2% of residual porosity is reached. Presently approaches that associate
minimum crystal sizes with a maximum density of the sintered grits are the primary
target of research.
Efforts to increase the hardness by using composite materials with harder addi-
tives are much less important for grits than in the field of compact ceramics. The
reason is the chemical nature of available hard materials (carbides like Sic, Tic,
but also nitrides or, in a few experiments, diamond). The inability of Sic to
grind ferrous materials is well known, and a similar observation had to be made
for carbide reinforced composite alumina grits, in spite of their improved hardness.
Further, a carbide additive in an oxide matrix does not always result in an
increased hardness because the low diffusion of the carbides (dominated by
covalent bonds) often gives increased porosity or (after sintering at a higher
temperature) a coarser microstructure. On the other hand, highly pure (and
therefore chemically very stable) sintered corundum grades have been developed
recently, which after pressureless sintering exhibit a hardness similar to or
even higher than the hardest hot-pressed composites with high carbide concen-
trations (cp. Fig. 5). All these features limit the motivation for new composites
and increase the trend towards advanced, thermodynamically stable sintered
grits on a pure oxide basis.
temperature), the final grinding of functional surfaces in ball bearings can hardly
be substituted by other approaches today. The interest in grinding materials that
exhibit a high grinding power at low process temperatures has been further pro-
moted by the need to improve the environmental compatibility of manufacturing,
a trend that is manifested by efforts to decrease or to avoid the consumption of
coolants (reduced coolant concepts, dry grinding). There are manifold ways to
develop such grinding materials. One approach is microstructural engineering:
the shape of the crystals in the microstructure governs the fracture behavior of
the grinding grain, the shape of the cutting edges and tips, and, finally, the chip
formation. One of numerous ideas was the composition of grinding tools with
hollow sphere-shaped bodies that, during grinding form always new (sharp)
grain edges.
250
200-
1r
- sintered AI,O, (soVgel)
K,
(MPadrn)
HVI
Wa)
2.83.9 15-20
4.2 16.3
1 .O-2.5 17.5
I I I I I I I I I I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
time (min)
Figure 8. Comparison of the material removal rate of sol/gel derived sintered and of fused grits on
the basis of AI2O3(flexible fiber disc, 6000min-’, contact pressure 30N; workpiece: steel tube [ST W
22 DIN 15431 with diameter 195 mm, wall thickness 1 mm).
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
K (MPdrn) K, (MPaJrn)
Figure 9. Microstructural dependence of the fracture toughness in grinding grains on the basis of
alumina prepared by sol/gel approaches (left) and in fused grits (right).
grinding performance of the grits under different conditions [16]. Figure 9 shows
different toughness distributions obtained in sol/gel by producing microstructures
with different porosity, crystallite size, and shape (as exemplified by Fig. 10).
Obviously, the polycrystalline nature of sintered materials is a strong tool for modi-
fications of the mechanical properties which are impossible in fused corundum grades.
From Fig. 9 it can be understood that the sol/gel derived grits owe their increased
grinding efficiency (Fig. 8) first of all to their improved global stability (fracture
toughness, strength). On the other hand, it is very instructive to understand the
grinding performance of the A1203-Zr03 eutectic: with a very high toughness
(Fig. 9) and with a hardness which is smaller than observed for zirconia-free
fused grits but in no way inferior to the sol/gel grades (Fig. 8), it is impossible to
attribute the only intermediate grinding efficiency of this material (Fig. 8) to deficits
in its toughness or hardness. On the contrary, the A1203-Zr02 eutectic is a very clear
example for the power-limiting effect of too high a strength of grinding grains,
decreasing the frequency of microfracture processes which are needed to re-generate
sharp cutting edges.
Of course, the concept of an adjustable strength that enables an optimum fre-
quency of microfracture events does not mean tolerance of hardness deteriorating
defects. From Fig. 2 it can be shown that within the framework of sol/gel
approaches the frequency and size of such defects can be minimized but hardly
avoided. In spite of the high number of patent applications devoted to improved
sol/gel-derived grinding materials, it has been impossible to overcome the funda-
mental limit associated with this technology: the discrepancy between the unique
chances to use nanoscale raw materials and the spatial heterogeneity introduced
by the indispensable but always coarser seeds. With the obviously negative impact
of flaws like those in Fig. 2 on the hardness and wear resistance even if the porosity is
no more than 1-2%, it has to be expected that sol/gel derived products will not be
able to exceed some limiting level of properties.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
664 4 Ceramics Based on Alumina: Increasing the Hardness for Tool Applications
Figure 10. Microstructure of sol/gel derived sintered alumina grits: (a) platelet shaped microstruc-
ture (‘Cubitron’, manufactured by MMM, USA), (b) bimodal grain size distribution after seeding
with transitional alumina nuclei, (c) anisotropic grain growth (Ce doping) [ 161.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.3 Tool Materials with Undefined Cutting Edge: Sintered Grinding Materials 665
HV1lnhei
250- sintered alumina Wa)
(by powder processing) 21-25
200 -
1
0) sintered alumina
-
v
z
.-
$150-
E
-0
a,
>
E 100-
??
50-
01 I I I I I
0 1 2 3 4 5 6
time (min)
Figure 11. Comparison of the grinding performance of known (fused and sol/gel derived) grinding
materials on the basis of A1203with a new single phase corundum grit produced by a powder
processing approach (same test conditions as given for Fig. 8) [16].
Most of the machining of lzavdened steel and hard cast iron, on the other hand, is done
by grinding, a process that today is necessary for the final finish of functional surfaces,
like in ball bearings. Nevertheless, there are several limitations of grinding which
motivate the search for other methods: compared with cutting operations (whenever
applicable) grinding is a time-consuming process (up to a factor of three), the equip-
ment is more expensive, and disposal or recycling of the grinding silt is difficult from
both ecological and economic points of view. It has been shown that even without a
complete substitution the working time can be cut to 2/3 by combining a first turning
step with final grinding [2]. If turning operations shall be used for the final finish as
well, important tool life criteria are the surface roughness of the machined workpiece
and the stability of the geometrical position of the cutting tip (which governs the accu-
racy of the produced piece). This paragraph deals with the technical demands on cut-
ting tools which might be able to give fresh impetus to such efforts without increasing
the costs (as it is the case with tools manufactured of diamond or cubic boron nitride).
Therefore, the discussion will be focused on ceramics on the basis of alumina as
outlined already in Section 4.1.
Table 2. Demands on basic properties of cutting tool ceramics for machining hardened steel:
x = important, .xx = very important
On precise machining
hand, however, special measures may be required to balance the increasing high-
temperature creep usually associated with small grain sizes. Reinforcement with
whiskers (Sic whiskers are most common [18]) is one way to reduce the creep
rate [19], an approach that also increases the thermal shock resistance [20]. Another
way to reduce the creep rate is grain boundary engineering. Besides the prevention
of impurity associated amorphous grain boundary films it was observed that a small
concentration of yttria (0.1%) can reduce the creep rate of sintered alumina [21].
As it is also known from other fields of working metals, the different forms of the
chips from machining cast iron and steel cause a much more intense interaction with
the cutting plate for the latter material. Therefore, the demands concerning the
chemical and oxidation stability are more stringent on working hardened steel
than on machining hard cast iron (Table 2).
If the wear resistance is increased by decreasing the grain size of the ceramic
microstructures, an additional advantage comes from a feature illustrated by
Fig. 6. The reduced wear rate is associated with a decreasing amount of grain
pull-out at smaller grain sizes, and the same effect can improve the quality of the
ground ceramic surfaces. With a constant grinding procedure in finishing alumina
ceramics (diamond wheel, grit size 4&50pm), the roughness of a ground 0.6pm
alumina ceramic is about one half the value measured on a ground 3.6 pm micro-
structure [7]. Obviously, the improved quality of the ceramic cutting edge increases
the chances for tool life at a small roughness of the machined hardened steel.
To prevent spontaneous microstructural changes in the workpiece and to obtain a
high efficiency of the machining process, all approaches try to realize conditions with
low cutting forces and a low heat input into the finished surface. Whereas this is gen-
erally true for both grinding and cutting operations, it is also accepted that the use of
cutting ceramics is promoted if local temperatures are high enough for some softening
of the (usually brittle) surfaces of hardened steel or hard cast iron [2]. To this end,
ceramics with a low thermal conductivity are preferred. Table 3 reveals little differ-
ences with the exception of the most expensive tool materials (cBN, diamond)
which are on the less favorable side (in fact, the high thermal conductivity of the
latter gives an improved thermal shock behavior which, however, is considered as
less important when small feed rates are applied on precise machining hardened
steel). Much more significant differences exist for the hardness of the different
ceramics (Table 1) and for their mechanical resistance at high temperatures. As to
its oxidation stability and its chemically almost inert behavior in contact with iron
or steel, pure sintered corundum is second to no other ceramic.
A1203+ x 15% Z r 0 2 = 28
A1203 x 30
A1203+ x 30% Ti(C, N) x 35
Si3N4 30-35
Cubic boron nitride (cBN) 40-120
Diamond 500-600
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 New Trend for Cutting Hard Workpieces 669
Corn S
E
-
Y
c
0
S
E 10-
.-
Corn
5
c
E
z
5-
0
Figure 12. Turning hard workpieces with different ceramic composite tools. Corn = advanced
commercial AI20,/Ti(C,N) grade, concentration of Ti(C,N) about 25-30 vol-%. S = laboratory
grade: average sizes of A1203 and TIC subregions 0.8-1 .O pm, TiC concentration 35 vol-%.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
670 4 Ceramics Based on Alumina: Increasing the Hardness Jor Tool Applications
an average grain size of 0.5 pm [22]. On turning hardened steel (X63CrMoV5) with a
discontinuous cutting regime, the comparison with an unspecified commercial tool
showed lifetime (or cutting path) that was three times greater. Another recent devel-
opment uses oxycarbide and oxycarbonitride compositions (Ti(C,O), Ti(C,O,N))
with different stoichiometry instead of TIC and obtains crystallite sizes of about
0.6 pm (after sintering at 1700°C and hot-isostatic post densification at 1435°C);
the hardness is HVlO = 22.5-23 GPa [23]. Figure 13 shows the microstructure of
these materials which is much more fine grained than an advanced (also claimed
submicrometer) commercial material with the same carbide concentration (on the
market since 1997). As another example, Fig. 13 includes a new material with an
increased concentration of 35 vol-% of the covalent phase which can be manufac-
tured with the same small grain size in spite of the high carbide loading.
The cutting performance of these new materials was tested in a direct comparison
with similarly hard but single phase (absolutely oxidation resistant) pure alumina
ceramics. The results are discussed in Section 4.4.4.
Figure 13. Latest tool composites based on A1203:(a) commercial ‘submicronstructured’ Al2O3/
671
Ti(C,N) composite and new laboratory grades [23] with (b) Ti(C,O;N), (c) and with an increased
4.4 N ~ MTrend,for
J Cutting Hard Workpieces
"""1
f = 0.2 mm/rev
a = 1.5 mm
0.0 0 1 2 3 4 5
Table 4. Properties of new tool ceramics compared with advanced commercial grades
Pure alumina
AC4 1 A1203 0.56 22.08 f 0.90
by high loads at the cutting tip. A new commercial grade of an A1203/Ti(C,N) tool
ceramic was claimed ‘submicronstructured’ by the producer and introduced into the
market 1997 especially for machining hard materials; it was compared here with an
advanced standard grade SH 1 (CeramTec AG, Plochingen, Germany) and used as a
reference to evaluate the cutting power of new laboratory grades (Table 4). Both the
submicrometer and the SHl tools are composites based on A1203with about 30%
of Ti(C,N).
On the turning hardened steel at high cutting velocities > 250 m min-’ there is a sig-
nificant input of heat not only into the formed chips but also into the cut surface of the
shaft. Depending on the state of wear of the tool, cutting forces and process tempera-
tures may increase to an extent that causes softening of the hardened steel, and the
hardness was observed to drop from HRC = 57-60 to values of less than 50. There-
fore, the decrease of the hardness has to be recorded as an additional parameter in
such investigations, and an intermediate machining operation with a low velocity
’
‘u = 180 m min- was required before every new run to restore the original hardness.
4.4.4.2.1 Intermediate Demands: Machining Hard Cast Iron Figures 15-17 show
the results for continuous cutting chilled globular cast iron G-X300CrMo 153
(Rockwell hardness HRC = 43-45). The flank wear of the tools is given in Fig. 15.
The high hardness of the oxycarbide reinforced submicrometer composite reduces
its wear by an amount similar to the progress of the commercial submicrometer
tool compared with the standard material SHI. Much more surprising is the quali-
tative leap in the lower wear of the pure alumina insert which cannot be explained by
its hardness: the hardness is only just intermediate between the (lower) hardness of
the commercial composites and the (higher) value of AT60A. Also, the grain size of
the sintered corundum material is not very different from the microstructural data
given in Table 4 for the A1203/Ti(C,0)composite AT60A.
Therefore, Fig. 15 suggests that reducing the grain size of TIC, Ti(C,N), or
Ti(C,O)-reinforced composites may indeed increase the hardness and reduce the
wear in turning hard cast iron, but the effect is limited, probably because of chemical
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
674 4 Ceramics Based on Alumina: Increasing the Hardness f o r Tool Applications
0.3.
0.25.
E
E
! 0.2-
m
>
5
50.15-
Y
c
m
-
c
5 0.1.
3
0.05.
i
p
if
0.0. I I I I I I
0 10 20 30 40 50 60 70
Cutting time (min)
Figure 15. Flank wear width of submicrometer ceramic tools on turning hard globular cast iron
(Rockwell hardness HRC = 43459, see Table 4 for tool characterization.
interactions between the iron and the carbide phases of the tools. To eliminate this
interaction by using a pure alumina tool can greatly improve the wear resistance as
far as the process temperature at the cutting tip is not too high to initiate additional
wear due the low high-temperature creep resistance of the submicrometer alumina
microstructure.
The cutting edge displacement (CED) characterizes the obtained precision in
turning hard materials. Figure 16 (CED) shows that complex wear processes may
give a different ranking of the tool materials compared with Fig. 15 (flank wear).
Again, all of the new laboratory submicrometer grades exhibit much less wear
than the advanced commercial tools. However, with the criterion of the cutting
edge displacement it is the ceramic with the highest hardness which shows the
best performance (composite AT60A, Table 4).
Another important criterion for the quality of tools in precision turning of hard
materials is the obtained roughness of the cut metal surface. The graphical
presentation in Fig. 17 gives the average roughness R,. A fixed ratio between the
statistically averaged maximum depth of roughness RZD and R, was observed in
all the test (Fig. 17). The arbitrary scatter of data was rather large, and no significant
difference was observed between the surfaces cut by the SHl or by the commercial
submicrometer tools. However, all other inserts show the same ranking as observed
for the flank wear in Fig. 15. Improved results are provided by the oxycarbide
reinforced composites (AT60), but a more significant progress results from the
use of pure submicrometer alumina inserts.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 New Trend f o r Cutting Hurd Workpieces 615
70
cast iron HRC = 42 - 44
v = 250 mlmin =
mm
f = 0.1 mmlrev
a = 0.2 mm K = 45"
E
c
c
E
-
$40-
Q
0)
.-
U
g30-
73
g 20-
0,
.
0
= i AT6OA
10- (Al,OJri(C,O))
0 I
I)-.. ..
I
....4
I I I I
Figure 17. Surface quality of cut surface. Average grain sizes of the tool microstructures (in com-
posites: sizes of single phase subregions) are D(SH1) = 1.5pm, D(AT60) = 0.67pm, D(AC41) =
0.56 p m (see Table 4).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 Ceramics Based on Alumina: Increasing the Hardness for Tool Applications
60.0
hardened steel criterion:
HRC = 57 - 60 flank wear width
40.0
'. VB = 0.15 mm
.E20.0
c
h
e
3
0
c
0
E
-
.-
10.0
8.0
6.0
4.0 ,
150 200 250 300 350
v (m/min)
Figure 18. Life time of advanced ceramic inserts on machining hardened steel (see Table 4 for
materials characterization). The arrows indicate frequently observed deviations.
(al 70 (bl 70
hardened steel HRC = 57 - 60 hardened steel HRC = 57 - 60
60-
I
E
150-
c
E
8 40-
n
$
g30-
P
5 0I
20-
0
10-
0 I
*
I
.
... .-.
I I I I
0 5 10 15 20 25 30
Cutiing time (rnin)
Figure 19. Cutting edge displacement on turning hardened steel with advanced ceramic tools: (a) at
v = 220mmin-' (above), (b) at 300mmin-' (below); see Table 4 for materials characterization.
between 200 and 250 m min-' . Flank wear is small at v = 220 m min-l , but already
at 'u = 235 m min-l a greater crater-wear is observed. At v = 235 m min-I, more
than 50% of the tested commercial inserts exhibited a sudden, strong wear increase
during the first 10-20min associated with local or global fracture at the cutting edge,
and it becomes difficult to record a continuously increasing flank wear up to
VB > 0.2mm.
Whereas the cutting edge displacement was immeasurably small during the first
5 min, it was overtaken by fracture soon later and could not be determined optically
(even at v = 220 m min-' the cutting edge displacement of commercial tools was
often shadowed by local features of microfracture). Hence, it was impossible to
measure the cutting edge displacement of the reference tools at v = 235 m min-].
Therefore, tests started with v = 220mmin-', but some of the new laboratory
grades exhibited extremely few wear at this velocity, and additional experiments
had to be performed at higher rates. It should be noted, however, that already at
v = 235 m min-' the reference curves for SH1 and the commercial submicrometer
material characterize but a minority of inserts standing this condition (about 50%
failed at shorter times by cutting edge fracture). Therefore, these data (Fig. 18) over-
estimate the real behavior of the commercial grades at this velocity.
Whereas even at the lower velocity of u = 220mmin-' almost 50% of both
commercial grades failed by fracture during the first 5-1 5 min of cutting hardened
steel, no one such event happened with the laboratory grades (where crater-wear
was also observed but to a lesser extent compared with the commercial ceramics).
This behavior correlates with another surprising observation made on preparing
the hardened shaft for the tests. This preparation is difficult with shafts of hard
cast iron or hardened steel because it associates the high hardness of the counterpart
with discontinuous cutting conditions at a changing frequency and power of
impacts. Cubic boron nitride inserts (dreborid$/, Lach company, Hanau/Germany)
failed on machining the hard (globular) cast iron, and hard metal inserts (WC/
6% Co, Vickers hardness M 16 GPa, Kc M 9 MPam'/2) were not able to cut the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
678 4 Ceramics Based on Alumina: Increasing the Hardness for Tool Applications
rough outer shell of the hardened steel shaft at velocities between 50 and
100mmin-'. The same observation had to be made with the commercial ceramic
composites (SHI and the submicrometer grade fractured after t < 2min). On the
contrary, both the pure alumina insert AC41 and the composites AT60A and
AT62 were successfully used with w = 120 m min-', feed rate f = 0.3 mm rev-',
and a depth of cut a = 1.5mm. These new, submicrometer cutting ceramics did
not only not fracture macroscopically during one hour of severe cutting (with one
tip) but retained microscopically nearly perfect cutting edges providing a high sur-
face quality of the hardened shaft in the first cut ( R , < 1 pm for t 5 15min [32]).
Surprisingly, under these severe conditions the general performance of the pure
alumina ceramics was even more prospective than the behavior of the new sub-
micrometer composites [32].
For SH 1, the manufacturer's data sheet gives a strength of 600 MPa (no available
data for the submicrometer grade). The bending strength of the Ti(C,O) reinforced
composite AT60A is about 800MPa [23],and a lower strength of about 650-
700MPa was observed for the pure submicrometer alumina ceramic [3]. The
fracture toughness of all of the laboratory grades is 3.3-3.8 MPa m1/2, whereas
surprisingly high values of 5.5 MPa m'I2 (SHI) and 6.6 MPa rn'" (submicron com-
mercial composite) are given by manufacturer's information for these references.
None of these data explain the much smaller fracture risk of the submicrometer
laboratory grades on machining hardened steel. Obviously, complex wear-induced
flaw-generation processes will have to be considered to understand the improved
global stability of the new laboratory grade microstructures.
No differences were observed in the roughness of the hardened steel surface
when different ceramics were used. At 'u = 235 rnmin-' ( f = 0.1 mm rev-',
a = 0.2mm), R, was 5 1 pm for the first 10-15min of cut for all of the tested
ceramics, increasing to about 1-1.5 pm after 30 min (K = 45": see comment at
the end of Section 4.4.4.2.1!). On the other hand, the tools showed very different
flank wear and cutting edge displacements. Figure 18 gives the influence of high
cutting velocities on the life of advanced cutting ceramics on machining hardened
steel (flank wear criterion VB = 0.15mmin-'). As already mentioned, 'u = 200-
220 m min-' is the upper limit of reliable use for the latest generation of commer-
cial A1203/Ti(C,N)composites. At this velocity, best results are obtained with the
new submicrometer composites with Ti(C,O) which exhibit the highest hardness of
all of the tested ceramics (see Table 4). At higher cutting velocities, however, the
relationship compared with the pure submicrometer alumina changes, and pure
A1203 again gives the greater advantage. A similar smaller effect of the cutting
velocity on A1203 inserts compared with composites was noticed on cutting hard
cast iron (Fig. 15, Fig. 17).
Contrary to the new ceramics, most commercial submicrometer inserts fractured
within the first 5-10min when used at w = 300mmin-', and the one continuous
wear curve measured up to 15min and displayed in Fig. 19b was obtained at a
softer part of the steel shaft (HRC = 52-57).
At all cutting velocities the new laboratory grades are clearly superior to advanced
commercial inserts. Contrary to a sometimes different behavior of flank wear and
cutting edge displacement on turning hard cast iron, both parameters were well
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4.4 New Trend,for Cutting Hard Workpieces 679
the surface of the machined steel. For example, on machining the hardened steel
with the submicrometer composite ceramic AT6OA the hardness of the cut surface
decreased from originally HRC = 59-60 to values of 57-59 (25min) and 55-56
(30 min) [32]. This influence of high-velocity turning on the hardness of workpieces
is important for different aspects in the development of new machining technologies.
On the one hand, the softening of the hardened steel promotes unexpectedly high
materials removal rates. In this way, turning of hardened steel with cutting ceramics
that are much less expensive than cubic boron nitride (cBN) seems possible now. If,
on the other hand, for the final step of precision machining a constant high hardness
of the cut surface is required, critical limits exist for the cutting velocity (about
2 5 0 m m K ' in the investigations described above) or for the tolerable tool wear
(here at ZI = 300 m min-' : VB 5 0.15 mm). These technical demands are, of
course, affected by the choice of the ceramic tool depending on its thermal and
wear properties.
4.5 Summary
Recent developments show somewhat different trends for sintered ceramic grinding
materials and for cutting tools.
(i) In the field of grinding grits, extended innovative work is directed towards a sub-
stantial further increase of the grinding efficiency (high materials removal rates)
and increased tool life within narrow tolerances for the surface quality of the
ground workpiece.
(ii) For cutting tools, innovation is focused on two aspects: the development of very
tough and highly reliable grades for high feed rates and cutting depths (e.g. for
globular and vermicular cast iron), and of extremely hard and chemically resis-
tant microstructures for the precise machining of hardened steel and hard cast
iron.
Apart from temperature dependent transformation toughening (Zr02 additives),
high values of the fracture toughness are obtained first of all by anisotropic crystal
shapes (whisker reinforcement, rod shaped crystals with high aspect ratio in
Si3N4, platelets); generally new mechanism are not likely to be found in the
near future.
Both the demands for advanced grinding tools and for the precise machining of
hard cast iron and hardened steel with cutting ceramics are expected to promote the
further development of powder technologies for the manufacture of highly perfect
submicrometer microstructures for applications in fields where diamond or cBN
tools are too expensive.
For grinding materials, corundum (a-A1203) ceramics are dominant at present.
Here, contrary to other fields of ceramic technologies, precursor (sol/gel) approaches
represent an advanced industrial state of the art. Extended developments are yet in
progress, but, on the other hand, it is just the wide industrial use of these procedures
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Rejkrences 68 1
which gives insight to some general limitations. New powder routes revealed such a
surprisingly large increase of the observed grinding efficiency of the grits compared
with sol/gel derived materials, that another leap in the quality has to be expected
similar to the progress enabled ten years ago by the introduction of sol/gel
approaches.
At present, submicrometer cutting ceramics for machining hard workpieces are
developed in two classes of materials. In the industry, first tests in Europe are
known for carbide reinforced composites based on A1203,whereas more fundamen-
tal consideration (Sections 4.4.1 and 4.4.3) and first laboratory results from Asia
and Europe (Section 4.4.4) indicate advantages of new grades of pure corundum
tools for cutting both mild steel and for machining hard cast iron/hardened steel.
Besides of precision machining, first surprising results suggest that pure A1203
cutting tools may be applied for the machining of hardened steel even under
conditions where at present only cubic boron nitride can be used (discontinuous
cut). On turning hardened steel with pure alumina, high materials removal rates
are possible if at high cutting velocities some softening of hardened steel workpieces
can be tolerated in a preliminary machining step.
References
1. A. Raab and W. Zielasko, Schneidstoffeinsatz in der Groaserienfertigung, Lecture at the meet-
ing of the VDI-Division Schneidstoffanwendung, 2 I . 1 I . 1996, Siegmaringen.
2. G. Wellein and J. Fabry, Hartdrehen ist mehr als der Einsatz von CBN, Werkzeug-Technik
1996, 45, 7-10.
3. A . Krell and P. Blank, The influence of shaping method on the grain size dependence of strength
in dense submicrometre alumina, J . Europ. Ceram. SOC.1996, 16, 1189-1200.
4. F. F. Lange, Sinterability of agglomerated powders, J . Am. Ceram. Soc. 1984, 67. 83-89.
5. T. J. Graule, F. H. Baader, and L. J. Gauckler, Shaping of ceramic green compacts direct
from suspensions by enzyme catalyzed reactions, cfi Ber. Dtsch. Keram. Ges. 1994, 71, 317-
323.
6. S. Tashima, Y. Yamane, H. Kuroki, and N. Narutaki, Cutting performance of high purity
alumina tools formed by a high speed centrifugal compaction process, J . Jpn. Soc. Powd.
Powd. Metall. 1995, 42, 1464-1468.
7. A. Krell, Fracture origin and strength in advanced pressureless sintered alumina, J . Am. Ceram.
SOC.1998, 81, 1900-1906.
8. B. J. Hockey, Observations by transmission electron microscopy on the subsurface damage
produced in aluminum oxide by mechanical polishing and grinding, Proc. BY. Ceran?. SOC.
1972, 20, 95-1 15.
9. A. Krell and P. Blank, Grain size dependence of hardness in dense submicrometer alumina,
J . Am. Ceram. Soc. 1995,78, 1118-1 120.
10. S. D. Skrovanek and R. C. Bradt, Microhardness of a fine-grain-size A1203,J . Am. Ceram. Soc.
1979,62, 215-216.
11. A. Krell, Improved hardness and hierarchic influences on wear in submicron sintered alumina,
Matrr. Sci. Eng. 1996, A209, 156-163.
12. A. Krell, Fortschritte in der spanenden Metallbearbeitung durch Keramiken mit Submikrometer-
Gefiige, in Hartstofle, Hartstoflychichten, Werkzeuge, Vrrschle&chutz, Proc. of the Hagen
Symposium for Powder Metallurgy, 13-14 Nov. 1997, R. Ruthhardt (Ed.), Werkstoff-
Informationsgesellschaft, Frankfurt, 1997, pp. 57-76.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
682 4 Ceramics Based on Alumina: Increasing the Hardness for Tool Applications
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
5.1 Introduction
S i c is considered by many as the most important carbide, because of its extreme
properties and the variety of present and potential commercial applications.
According to Kieffer [l] silicon carbide belongs to the group of nonmetallic hard
materials, that is materials whose great hardness and high-melting point result
from a high fraction of covalent bonding. Superhard compounds are obviously
formed by combination of the four low atomic number elements: boron, carbon,
silicon, and nitrogen to form a quarternary system (see Fig. 1); carbon as diamond,
boron-nitrogen as cubic boron nitride, boron-carbon as boron carbide, and silicon-
carbon as silicon carbide belong to the hardest materials hitherto known.
5.1.1 History
Berzelius [2] first reported the formation of silicon carbide in 1810 and 1821. It was
later rediscovered during various electro chemical experiments especially by
Despretz [3], Schiitzenberger [4], and Moissan [5]. However it was Acheson [6],
who first realized the technical importance of the carbide as a hard material and,
by 1891, had managed to prepare silicon carbide on a large scale. Originally,
Acheson had completely different things in mind: he wanted to produce artificial
diamonds by recrystallizing graphite in an aluminum silicate melt; at a time when
high-pressure and high-temperature chemistry was unknown. His efforts proved
fruitless, but he was rewarded by the discovery of a far more versatile material,
silicon carbide.
Believing that this was a compound of carbon and corundum, he named the new
substance ‘carborundum’ (Fig. 2). Today, silicon carbide has become by far the
most widely used nonoxide ceramic material, with an annual world production of
about 700 000 tonnes. Owing to its great hardness, heat resistance, and oxidation
resistance it has become firmly established as an abrasive as well as a raw material
for making refractories such as firebricks, setter tiles and heating elements. Another
major use of silicon carbide is as a siliconizing and carburizing agent in iron and steel
metallurgy.
Self-bonded silicon carbide materials in molded form and of high S i c content in
order to make full use of the excellent material properties, have been available only
from the late 1960s onwards. They are being increasingly used as refractory
materials as well as for structural components in mechanical engineering, and
have proved highly successful for use under extreme conditions involving abrasion,
corrosion, and high temperatures.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
684 5 Silicon Carbide Based Hard Materials
( raphite)
(%iamond)
rarer [7]. Its existence in interstellar space has been shown by spectral measurements,
and it has also been found in samples of moon rock.
4000
3826
3232
-
3000
L+C
-
Y
v
-5 1.5 >99.97
2
3
;2000
L
Q) -0.5
? 1414
1413 *1 Si + S i c
-
49.4 e 0.6 /
1000
Si + Sic
0 20 40 80 )0
Carbon (at. %)
Figure 3. Phase diagram of the silicon-carbon system at 1 bar total pressure after Kleykamp and
Schumacher [ 121.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
686 5 Silicon Carbide Based Hard Materials
Modification Polytype
Since silicon carbide predominantly exists in its beta form at temperatures below
2O0O0C,this is referred to as the low-temperature modification. Cubic PSiC is meta-
stable and, in accordance with Ostwald' s rule, is formed initially in Sic production
from silicon dioxide and carbon. PSiC (polytype 3C) can also be prepared at about
1450°C from simple mixtures of silicon and carbon or by hydrogen reduction of
organosilanes at temperatures below 2000°C. Above 2000"C, only the hexagonal
and rhombohedra1 types are stable if there are no stabilizing influences which
raise the transformation temperature [ 14,151. Thus, in a nitrogen atmosphere at
high pressure, PSiC is the stable form, so that above 2000°C aSiC transforms to
PSiC. On the other hand, a hexagonal polytype, the wurtzite or 2H modification
of Sic, has been obtained at temperatures as low as 1300-1600°C by the decom-
position methyltrichlorosilane CH3SiCI3 [ 161. This 2H polytype, which does not
occur in commercial Sic, and the most common 3C, 4H, 6H and 15R structures
are called short-period polytypes. Long-period polytypes are much less common
and consist of blocks of short-period polytypes which are broken by regularly
occurring stacking faults.
The basic element of the silicon carbide structure is the tetrahedron [17] due to sp'
hybridization of the atomic orbitals. This tetrahedron consists of a silicon or a
carbon atom at the spatial center, surrounded by four atoms of the other kind.
The Sic- bond is 88% covalent. The tetrahedra are arranged in such a way that
units of three silicon and three carbon atoms form angled hexagons which are
arranged in parallel layers as shown in Fig. 4.
The layer sequences can repeat themselves in the cycles ABC, ABC . . . (zinc
blende, type 3C) or AB, AB . . . (wurtzite, type 2H), according to cubic or hexagonal
close packing. In addition, numerous others stack sequences are formed in the case
of silicon carbide, resulting in many similar polytypes.
About 180 different polytypes are currently known. They all have the same
density, 3.21 g ~ r n - ~
The
. formation and stability of the various polytypes are not
yet clearly understood. Written polytype nomenclature [ 181 indicates the number
of layers in the repeating layer pack by a numeral, while the crystal system is denoted
by the letters C, H, or R.
The amounts of the most frequently encountered polytypes 3C, 4H, 15R, and 6H
can be quantitatively determined by X-ray diffraction techniques [ 191, which can
indicate the temperature of formation of the silicon carbide since the stability of a
given polytype depends on, among other things, temperature. According to Inomata
et al. [20] 2H-Sic seems to be stable below 14OO0C, 3C between 1400 and about
1600"C, 4H between 1600 and 2100"C, 6H above 2100°C and 15R above 2200°C.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.2 Structure and Phase Relations o j S i C 687
u - S i c l6H)
Figure 4. Crystal structure of the SIC polytypes 3C and 6H.
There is no doubt, however, that impurities also play a part in the formation of the
various polytypes, together with the surrounding gas atmosphere and growth-
kinetic processes. Preferential stabilization of the 4H polytype is achieved with
additions of aluminum up to concentrations of ~ 0 . weight-%, 4 which is thought
to substitute for silicon in the silicon carbide lattice [21-241. According to Lundquist
[21] pure CcSiC mainly crystallizes as 6H, with small amounts of aluminum 15R also
occurs, and above 0.1 weight-% aluminum, 4H predominates (see Fig. 5). Also it is
well known today that nitrogen can be incorporated into the S i c lattice, whereby it
stabilizes not only the cubic PSiC form but also the wurtzite 2H polytype [14,25].
Whereas the nitrogen content obtained in 3C-Sic by gas/solid diffusion is low,
e.g. only < 1 atom YO,extremely high nitrogen contents of up to = 14 atom YOwere
A l content. w t % -
Figure 5. Proportions of polytypes in technical-grade S i c as a function of Al content
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
688 5 Silicon Carbide Bused Hard Muterials
30 m
I- 4
Resistance core
Reaction mixture
Current conductors
/
~
I
Gas collection duct Porous bed
outwards, with decreasing temperature, the crystal growth rate decreases and the
S i c roll becomes more and more finely crystalline. The outer zone is composed of
very fine crystals of SSiC. While in the case of the classic Acheson furnaces, with
their laterally positioned electrodes and side walls, the carbon monoxide burns
away on the surface, polluting the atmosphere inside and outside the furnace
building, the ESK (Elektroschmelzwerk Kempten) process [30] developed in the
1970s (Fig. 6), is ecologically safe. It operates with floor electrodes, which enter
the reaction mass from below a porous floor, and the carbon monoxide is collected
by a gastight plastic canopy and transported through gas collecting pipelines to a
power station where it is converted into electrical energy after burning. This
energy is recycled and amounts to 20% of the energy necessary for the production
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
690 5 Silicon Carbide Based Hard Materials
of silicon carbide. The silicon carbide cylinders inside the linear or U-shaped resis-
tance furnaces are up to 60 m long and weigh about 300 tonnes after a run of 10 days
(see Fig. 7). A comparative review of some Acheson plants has been published [3I].
The lumps of silicon carbide obtained on dismantling the furnace are taken to a
sorting area and broken down with hydraulic equipment and compressed air
hammers. The finely crystalline outer zone (the so-called amorphous zone) is
further size reduced (e.g., in a jaw crusher) and is usually returned to the furnace
process by adding it to the coke/sand mixture. Alternatively, this 9&95% Sic material
is a valuable raw material for the production of grain refiners for the ferrous metal
industry.
The coarsely crystalline compact pieces from the inner zone of the Sic roll are
sorted according to criteria such as color, structure, etc. This top-grade material
forms the largest fraction of the Sic yield from a furnace charge.
Further processing of the silicon carbide into the various macro- and microgrits
depends upon such factors as the required particle size distribution [32] and shape,
and involves breaking, grinding, sieving, air classification, and elutriation.
For applications that require a particularly high grade of silicon carbide, such as
industrial ceramics, the refractory industry, and ceramic-bound abrasives, the
granular product must be washed with alkali or acid to remove adhering traces of
elemental silicon, metals, metal compounds, graphite, dust, and silica.
Commercial Sic grades vary in composition from 90% to 99.5% Sic. Light green
S i c is 99.8% pure, and, as the purity decreases to 99%, the color changes to dark
green, and, at 98.5% to black.
Before aSiC is converted into molded parts by sintering, it must be intensively
milled into the submicrometer range, that is to a mean particle size of less than
1 pm (see Fig. 8b), in attritors [33,34] or ring-gap mills. Adherent impurities
which interfere with the sintering process, such as SiOz or wear from milling
equipment, are removed by wet chemical purification [35].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production of' S i c 69 1
-
equation
Si02 + 3C < 1.900'C PSiC + 2CO.
According to the reaction mechanism proposed by Weimer [36] the reaction takes
place in several steps:
SiO,(S) + C(S) + CO(,)
+ + C02(g)
+
pressure (solid/solid reaction). The high rate of SiO(,) formation suggests that the
further two gas/solid reactions make the final carbon reduction of the SiO to Sic
possible once the C/Si02 contact points are consumed. Carbon crystallite diameter
was found to have a substantial influence on the rate of reaction and the size of
synthesized Sic.
This process has been used in the early 1960s to fabricate pure ultrafine PSiC
powders especially suited for the semiconductor and pigment industry [37,38].
During the first step of the General Electric Process [37] a silica gel is formed by
hydrolysis of silicon tetrachloride in an aqueous solution of sugar (sucrose).
Upon dehydratization at 300°C the sugar pyrolizes and an intimate amorphous
carbon/amorphous silica mixture is obtained which is subsequently fired at
1800°C in an inert atmosphere to form PSiC.
Later the process was modified for the fabrication of PSiC sintering powder
[39, 401, using lower reaction temperatures and as starting materials the com-
binations tetraethylsilicate/sugar and fumed amorphous silica/sugar, respectively.
Particle size control corresponding to powder surface areas in the 10-20 m2 g-'
range (see Fig. 8a) was necessary, because ultrafine powders have high oxygen
contents (surface silica), and often low green densities result due to their high surface
area.
Since 1981 a continuous version of the Acheson process has been commercially
operated by the Superior Graphite Company [41]. A mixture of silica sand and
petroleum coke is fed continuously to the furnace and heated electrothermally to
a temperature of 1900°C. The run-off-furnace product is a free flowing grain
generally smaller than 3 mm, comprising agglomerates of PSiC crystallites and
free carbon. The grain is friable and is readily processed to make sized microgrits
and sintering powder by dry milling, wet grinding, classification, and wet chemical
purification (removal of 5 weight-% free carbon).
5.3.2.1.2 Synthesis from the Elements Prochazka [39] synthesized submicron PSiC
powder with a surface area of 7 m2 g-' using direct reaction of high-purity silicon
powder with carbon-black at temperatures between 1500 and 1650"C, but saw
little densification with boron and carbon additions, suggesting that free silicon is
detrimental to densification.
The in situ formation of silicon carbide according to the reaction
Si(l,,) + C(S) = SIC(,)
is mainly used in the manufacture of reaction sintered Sic components. After
shaping a mixture of fine-grained Sic powder and carbon and exposure at high
temperatures to liquid or vapor-phase silicon, the carbon is reacted to secondary
silicon carbide, which bonds together the primary Sic grains. Excess silicon
(about 10vol-Yo) is usually left to fill the pores, thus yielding a nonporous body
(Si-Sic).
5.3.2.1.3 Deposition from the Vapor Phase Submicron PSiC can be continuously
produced by decomposing gaseous or volatile compounds of silicon and carbon
in inert or reducing atmospheres, at temperatures above 1400°C [42]. Particle size
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production ojsic 693
SiC14(,) + CH4(g)--%
H
+
Sic(,) 4HC1(,).
If boron trichloride is used for the boron dopant together with a small excess of the
carbon feedstock, sinterable Sic powders are obtained.
The laser synthesis allows laboratory scale production of high-purity S i c powder
with controlled particle sizes. Using a C 0 2 laser and a mixture of silane (SiH4) and
methane (CH4) or acetylene (C2H2) PSiC particles in the size range 5-200 nm have
been obtained. An advantage of using SiH4, rather than SiC14, as a reactant is that it
has a strong absorption band near the wavelength of the laser (10.6 pm).
Amorphous S i c powders with metal impurity levels below 60 p.p.m. total have
been synthesized by thermal decompostion of a methylsilane gas stream at tempera-
tures above 600°C according to the equation
CH3SiH3 Sic + 3H2
in an induction heated graphite reactor [46].
5.3.2.2.1 The Vapor-Liquid-Solid Process In the VLS process silicon and carbon-
rich vapors (usually CH4, SiO, or SiC14) react at 1400°C to from S i c on a liquid
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
,,.,.
694 5 Silicon Carbide Based Hard Materials
<
Vapor Feed Temperature = 1400°C Vapor Feed
sio
c+sI CSolid
Sic Clystals
Liquid Catalyst
Supersaturated with Si+C
Solid Catalyst
(Steel 30vrn)
\1
6%
/////// / / / / / / / / / / / /
Substrate (Carbon)
I 77
Figure 9. VLS process for S i c whisker growth (1400°C).
alloy:
+
Si(d) C(d)
1400°C
___t Sic.
(d) dissolved in liquid iron alloy
Many different VS-processes are known according to the starting raw materials and
catalysts species involved. In all cases, a catalyst is used for whisker formation to be
favored versus particulate S i c formation. Commercially available VP-whisker
grades are typically less than 1 pm in diameter (submicron) and up to 30 pm in
length and show strength/Young modulus values of 1-5 GPa/400-500 GPa.
During whisker handling and processing, rigorous protection has to be provided
since smaller whiskers (< 3 pm diameter) can become lodged in the lungs and
represent a health hazard. An inexpensive VS-process utilized calcined rice hulls,
which contain both silica and carbon, as a precursor to S i c whiskers.
The characteristics of commercial S ic whiskers as claimed by producers are given
in Table 2.
PSiC whisker have been added to a wide variety of ceramics (A1203,Si3N4,AlN,
MoSi2, mullite, cordierite, and glass ceramics) in an effort to achieve increased
toughness. In most cases, hot pressing has been required to achieve near-theoretical
density of the matrix. A real success story in the advanced ceramic market was the
introduction of a 25-30 weight-% S i c whisker reinforced hot-pressed alumina for
the high-speed cutting of high nickel content alloys. Thus, by incorporation of
S i c whiskers into the alumina matrix a composite with increased strength (by
over %YO),fracture toughness (by loo%), increased fatigue and thermal shock
resistance was obtained [48]. In this composite, whisker pull out and crack bridging
have been identified as the major toughening mechanisms. Crack bridging consists
in bridging the crack behind the crack tip by unfractured whiskers or fibers. The
effect depends on the strength of the fiber and on its aspect ratio. Inhibiting crack
Phase
Length (pm)
P
20-50
P
5-200
a
3-20
P P
30-200
P
Diameter (pm) 0.3-1.0 0.05-1.5 0.15-0.25 0.3-1.0 1-3 0.5-1.7
Aspect ratio 20-200 20-50 20-50 5-20
Density (gcm-') 3.2 3.18 3.21 3.21 3.21
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
696 5 Silicon Carbide Based Hard Materials
opening reduces the stress intensity at the tip of the crack and stops or slows crack
propagation.
The fiber or whisker pull out has the potential for the greatest toughening effect
and consists in a debonding of fiber from the adjacent matrix consuming large
energy. This mechanism depends strongly on the aspect ratio of the second phase
and is favored by weak interface bonds.
Sic whisker also can be combined with metals (MMCs, metal matrix composites)
to increase the high temperature strength of the material as well as provide a
comparable substitution for heavier traditional materials, such as steel. MMCs
can be fabricated by infiltrating a Sic whisker preform with aluminum or by
addition of S i c whiskers to molten aluminum.
(-Si(CH3) - CH2-),
I
-
1300°C
PSiC + CH4 + H2.
H
Dimethyldichlorsilane is dechlorinated with metallic sodium in the presence of
xylene to polydimethylsilane (R2Si),, which is converted to polycarbosilane (PCS)
by polycondensation with introduction of silicon-carbon bonds into the chain.
The pyrolysis of the PCS resin to PSiC is accomplished by a slow heating to
1300°C and is accompanied by an appreciable loss of mass (=60%) and consider-
able shrinkage. This melt spinning-curing-pyrolysis route has been adopted by
Nippon Carbon to produce a fiber called Nicalon [55] which has a composition
almost identical to Si3C40thus, it is not pure Sic, but a Si-C-0 composite, in
which nanometer-sized PSiC, Si02 and carbon are uniformly dispersed. These
impurities affect the thermal stability of the fiber and its strength, which has been
reported to degrade above 1200°C. However, these fibers have successfully been
used to reinforce aluminum, refractories and Li-Al-Silicate glass ceramics. The
coating of fibers, for instance by carbon, allows a decrease in the interface bonding
with the matrix, which increases the strength and the toughness by favoring pull-out
mechanism.
UBE Chemicals synthesized amorphous Si-Ti-C-0 fibers from the PCS-titanium
alkoxide compound polymer. These so called Tyranno fibers show excellent proper-
ties and can be spun thinner than the Nicalon fibers (see Table 4). However heating
above 1000°C results in a crystallization of the fiber. In the mid 1990s the German
company Bayer AG synthesized an amorphous Si-B-N-C fiber, by pyrolysis of a
polyborosilazane polymer [56]. This SiBN3Cfiber (see Fig. 11) has a tensile strength
of 3 GPa and maintains its amorphous character up to 1800°C.The advantage of the
production route from liquid to solid to produce S i c has also attracted attention for
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production of sic 699
Figure 11. SEM of SiBN3C fibers made by Bayer AG, Germany (courtesy H.P. Baldus, [56]).
5.3.2.4.3 Sintered Powder Derived aSiC Fibers (SP-fibers) Recently, the Carbo-
rundum Company developed a single phase polycrystalline aSiC fiber having a
diameter of approximately 20-1 50 pm and density of over 96% of theoretical. It
is made by first producing green filaments by melt extrusion or suspension spinning
of plasticized mixtures of sinterable aSiC powder with organic additives, such as
polyethylene or polyvinylbutyral plus novolac, respectively. The green filaments
are subsequently debindered during free fall from the extruder in presintering
furnaces and finally undergo pressureless sintered at 2 100°C in argon atmosphere.
These polycrystalline aSiC fibers are stable up to 1600°C in air and 2250°C in inert
gas, far superior than most of the commerciallly available PSiC based fibers [49].
According to Prochazka [%],an increase in the strength of aSiC fibers is obtain-
able by using refined aSiC powders, having a surface area greater than
20m2gp', a median S i c particle size <0.25pm, and no particles larger than
1.5 pm.
Figure 12. Microstructure of various S i c ceramics; (a) KSiC, nitride bonded S i c (AnnaSicon 25);
(b) RSiC, recrystallized Sic (AnnaNox CK); (c) SiSiC silicon infiltrated B C (AnnaNox CD), (d)
SSiC, BjC doped (Ekasic F), (e) LPSSiC, equiaxed SIC grains (Ekasic T), (f) LPSSiC, platelet
S i c grains (ESK), (g) HPSiC,Al-doped (ESK), (h) HIPSIC, undoped (ESK).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5 Silicon Carbide Based Hard Materials
achieved by choosing the right S i c and carbon contents in the initial mix, and ensur-
ing correct porosity in the preform and by proper temperature control.
5.3.3.4.1 Solid-state Sintered S i c Once it was thought that silicon carbide was
unsinterable. It was only in the early 1970s that it became possible to sinter silicon
carbide without applied pressure and thus to achieve densities more than 95% of the
theoretical density. This was largely due to the pioneering work of Prochazka [75] at
the General Electric Corporation Research Center in Schenectady. The starting
materials were submicrometer PSiC [76] or later aSiC [77] with the simultaneous
addition of up to 2% carbon and boron. A minimum of 0.3 weight-% boron was
required, although larger additions did not improve the density, which Prochazka
attributed to the need to exceed the solubility limit of boron in S i c (0.2 weight-%
according to Shaffer [78]). Carbon is added in excess to ensure complete deoxida-
tion, that is removal of the Si02 film from the S i c grains, which in turn leads to
the presence of carbon inclusions. Additions of excess carbon are observed to inhibit
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
104 5 Silicon Carbide Bused Hard Materials
grain growth, and produce a more equiaxed grain structure [79],probably as a result
of the energy required from the grain boundaries to overcome the carbon inclusions.
Instead of boron and its compounds [80-831 one can also use aluminum and alumi-
num compounds [84-901, or beryllium and its compounds [91], the net result being
the same. Sintering of AljC doped [92] or A1203-doped[90] Sic powders is believed
to occur in two stages: transient liquid phase sintering; then solid state sintering
during the final stage.
This opened up an inexpensive method of producing dense and complex parts
consisting of pure silicon carbide (SSiC). The powders (see Fig. 8) can be molded
into a green body by any of the methods used in ceramic molding [93,94], depending
on the shape required and the number of pieces involved (see Table 5). Sintering is
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production of Sic 705
Table 6. Microstructure and Sic-polytype phase composition in sintered silicon carbide (B/C, AljC and B/Al/C doped SSiC materials).
s?
Powder T Sintered body
b
Sample SiC- Addition to Sintering Chemical analysis Polytype analysis Micro- ir!
3
no. modification SIC (weight-%) temperature (weight-%) (weight-%) structureb 0
(colour) ma A1 B 6H 15R 4H 3c n
Q
1 P 1B,2C 2050 - 1.1 13.3 6.4 1.2 79.1 FG 5
2 P 2.5 AIN, 2 C 2050 1.o - 3.9 8.9 74.6 12.6 FG f
tn
3 a-black 2 c 2050 0.44 - 5.6 1.5 92.9 - FG
-
e2L
4 a-black 1 B, 2C 2050 0.65 0.86 - 6.5 93.5 FG
- - Q
5 a-black 2 AIN, 2 C 2050 1.34 - 2.9 97.6 FG QQ
- -
6 a-green-I 1.2 B, 2C 2050 - 1.12 100 - FG
7 a-green-I 0.2 B, 2 AlN, 2 C 2050 1.03 0.22 73.2 .- -
%
26.8 FG f?
8 cc-green-I 0.2 B, 1 Al, 2 C 2050 0.38 0.21 92.6 - 7.4 - FG T
- - 6'
9 a-green-I 0.2 B, 2C 2050 - 0.25 100 - FG 6
-
10 a-green-I1 1B,2C 2050 0.92 83.5 1.3 15.1 - FG
11 cc-green-I1 1 B,2C 2200 - 0.93 40.9 14.3 44.8 - AGG
12 a-green-I1 1 AIN, 2 C 2050 0.49 - 85.9 2.1 12.0 - FG
13 cL-green-TI 1 AIN, 2 C 2200 0.45 - 83.2 1.2 15.6 - FG
14 a-green-I1 2.5 AlN, 2 C 2050 1.20 - 63.7 5.3 31.0 - FG
-
15 a-green-I1 2.5 AIN, 2 C 2200 1.29 48.8 7.0 44.1 - FG
16 a-green-I1 1 Al, 2 C 2050 0.76 - 76.1 2.4 21.4 - FG
-
17 a-green-I1 1 Al, 2C 2200 0.74 68.5 3.2 28.3 - FG
(a) 1 bar argon, graphite-tube-resistance furnace, 15 min hold
(b) FG . . . FINE GRAINED AGG . . . ABNORMAL grain growth
Characterization of Starting Powders (Specific Surface Area: 15-20 m2 g-I):
P carbotherm. red. silica, 100 w/o 3 C, 20 p.p.m. Al-content.
-
a-black Milled Acheson SIC, 89.9 w/o 4 H - 2.5w/o 15 R 7.6w/o 6 H, 0.71 w/o Al -
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production of Sic 707
(35 kbar, 2300°C) to 100% of the theoretical density [115]. In axial hot-pressing with
graphite tools it is, however, necessary to use small amounts of sintering aids: 0.5-
3% boron, aluminum, aluminum oxide, beryllium oxide, yttrium oxide, or tungsten
carbide are added to fine uSiC powder before molding [26, 116-1231. The molding
temperature is 1900-2000°C, depending on the particle size of the powder and the
amount of sintering additives used. The pressure that can be used is limited by
the strength of the graphite, the maximum being 50 MPa. The conventional
method requires large amounts of energy and mold material and can normally be
used only for parts with simple, uncomplicated shapes. Owing to this limitation, pre-
cision parts can only be produced by machining with diamond tools [124] these, of
course, are expensive.
A more convenient method than conventional hot-pressing, also still more
expensive, is hot-isostatic pressing (HIP). During the development of this tech-
nique [ 125-1271 it proved possible to be prepare high-purity S ic products with
silicon carbide contents of more than 99.5% by hot isostatic pressing of S i c
powder or SIC preforms in vacuum-sealed casings to a final density exactly equiva-
lent to the theoretical density and with an uniform, fine-grained microstructure (see
Fig. 12h). Owing to the higher isostatic pressure of 2 kbar (argon gas), compared
with normal hot-pressing, no sintering aids need to be added; that is, the thin film
of silica on the silicon carbide particles (which represents residual oxygen impurity
of the S i c powder) is sufficient to achieve complete consolidation.
The postdensification of SSiC or LPSSiC by hot isostatic pressing is less compli-
cated because no gastight encapsulation is necessary. Not only is it possible to
achieve more than 99% of the theoretical density, but the variation of density
and strength can also be reduced. Pressureless sintering and hot isostatic pressing
can now be performed in the same cycle (see Fig. 13). This new process, called
sinter-HIP, offers economic benefits [ 128-1 301.
The importance of hot-pressed silicon carbide has decreased considerably since the
introduction of pressureless sintering, although it is currently still the most suitable
method of obtaining the best mechanical properties for pure, monolithic SIC.
5.3.3.6 Chemical-Physical-Vapor-DepositedSic
CVD-SIC is obtained by the chemical reaction of volatile silicon- and carbon-
containing compounds in the presence of hydrogen and in the temperature range
1000-1800°C, e.g.
CH3SiC13 2 SIC + 3HC1.
Besides the formation of S i c powders, S ic fibers, and S i c thin or thick films, nowa-
days monolithic bodies in sections up to 25 mm thick can be prepared by this method.
CVD-SIC is of high purity (99.999%), cubic p-modification, very fined grained,
and shows the lowest of oxidation rates. Bulk cubic phase CVD-SIC is produced
for applications such as electronic packaging, components for wafer handling,
kiln furniture for diode manufacture, laser optics for high temperatures, and
substrates for computer storage media [131, 1321. It can be polished to variations
of less than 0.3 nm. Coatings of CVD-SIC were initially developed in the 1960s
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
708 5 SiEicon Carbide Based Hard Materials
I
I
Shaping 1
Encapsul.
- HIP Containerless-HIP
for nuclear fuel particles employed to reduce the diffusive release of metallic fission
products from the fuel kernel.
Today Sic thin films (< 1 pm thickness) can also be produced by physical vapor
deposition (PVD), for example by sputtering, which method allows lower substrate
temperatures, but works more slowly. Electrically conductive B/N-doped sintered
clSiC with up to 9weight-% free carbon has been developed as target material,
[133]. Novel applications for PVDSiC include films for computer storage media,
protective coating for lenses, and microwaveable packaging for food.
The gases diffuse through a porous graphite retainer and along carefully programmed
thermal and pressure gradients. The primary gaseous species of silicon reacts with the
graphite walls of the growth cell to from additional Si2C and Sic2 which recombine
on the growing crystal according to a ‘double condensation’ reaction
2SiC2 + 2Si2C = = 6H Sic.
Hence the silicon plays the role of a carbon-transporting agent. Water-clear boules
of the pure 6H polytype having diameters > 50 mm and lengths > 60 mm have been
grown by this technique. It is believed that by the year 2001 the size of the boules will
have increased to 15cm in diameter.
Based on the growth of epitaxial thin films of single crystal polytypes on boule-
grown substrates, S i c is now becoming the material of choice for high-power,
high temperature (> 5OO0C), and high-frequency devices [ 1341.
As a spin-off, currently colorless S i c gemstones, 1/2 to 1 carat in size, cut from
6H-Sic wafers are entering the jewellery market at about 10-1 5% of the price of dia-
monds.
5.3.3.8 S i c Nanoceramics
Sintering nanosized powders with the aim of flaw avoidence from nanosized sintered
bodies (grain size 5 100 nm) is recognized as a promising way of improving mechan-
ical properties and reliability of S i c ceramics. However, nanosized S i c powders are
not easy to process [42,45,136], and several difficulties must be overcome: the
powder flows badly, exhibits low oxidation resistance [40], has low filling and
compaction density [40] and is currently too expensive for large scale use.
5.3.3.8.1 Fabrication by Solid State Sintering In 1991 Vassen et al. [ 1371showed that
polycrystalline PSiC bodies with a density of at least 95% of the theoretical density
and a fine grain size of 150nm can be prepared by encapsulated HIPing of BjC-
doped laser-synthesized powders with particle sizes below 20 nm. Necessary HIP-
temperatures were 1500°C, which is 250°C below the temperature needed to densify
conventional submicron powders. This fabrication process was later optimized
[138] by preheating the shaped S ic bodies in vacuum during an annealing step
before encapsulation. Owing to this additional annealing step, residual oxygen
contents were minimized and a mean grain size of only 60nm could be obtained in
the HIPed Sic shapes. It was found that the reduced final grain size has a strong
effect on mechanical properties; as a consequence of grain size reduction from 1 pm
to 150nm the Vickers hardness increased from 2000 HVlO to 2500 HV10, whereas
fracture toughness decreased from 4 to 3 MPamI”, respectively [139]. However, by
adoption of a bimodal grain-size distribution, with introduction of larger sized S ic
grains into a nano-sized matrix, an incrase in fracture toughness to 6 MPa m1I2was
achieved [140,141]. For a fine grained HIPSiC (300nm) even under a stress of
100 MPa at 1600°C very moderate creep rates of 1 x s-’ were measured [142].
Figure 14. Potential toughening mechanisms for S i c based ceramics: 1 crack deflection, 2 micro-
crack formation, 3 crack bridging with metallic ligaments, e.g. TiB2, 4 crack bridging and pullout
by platelet or fiber - reinforcement.
be considered (see Fig. 15). In the following sections today’s state of the art in some
more or less important Sic-nonoxide composites is reviewed.
5.3.3.9.1 Sic-Tic Very promising composites have been developed in the SIC-TIC
system with S i c as the matrix phase [146-1481. Dispersed TIC particles significantly
Orientation
- worse- -better-
Figure 15. Microstructural design for optimized crack deflection in composites
(Courtesy R. Telle, RWTH Aachen/Germany [ 1451.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
712 5 Silicon Carbide Based Hard Materials
Figure 16. SEM micrograph showing crack deflection and crack bridging in Sic-Tic and SiC-TiBz
composites: light areas = TIC or TiB2, dark matrix = SIC
(Courtesy D.Ly Ngoc, MPI Stuttgart/Germany [149].
improve both the strength and the toughness. Although an addition of T i c does not
reduce the densification temperature significantly below 2 1OO'C, the coarsening of
S i c is completely retarded which raises the strength to 700-800 MPa [146-1491.
The increase in KIc to 6.5-7.5 MPam'I2 is attributed to the misfit of the thermal
expansion coefficients of T i c and Sic, introducing considerable radial tensile stres-
ses at the phase boundaries and hoop compressive stresses in the matrix. These stres-
ses enable crack deflection (see Fig. 16), crack branching, and microcracking above
a critical particle size of = 3 pm. The optimum volume content of T i c ranges
between 20 and 30 vol-%.
the particular mechanisms to the total fracture toughness, stating that the stress
induced microcracking is operational in a process zone of approximately 150 pm width.
Typical conditions for densification by axial hot-pressing are 2000-21 OOOC, at a
pressure of 20-60MPa for 30-60 min which results in 96-99.8% density. The
particle sizes of the matrix and dispersed phases range between 1-5 and 4--8pm,
respectively. An optimum volume fraction of reinforcing particles of 25-30 vol-Yo
has been reported, yielding a flexural strength of 710 MPa and a fracture toughness
of 5.0-5.7 MPam'/2 [149]. Composites with a lower TiB2 content of 15 vol-Yo exhi-
bit a mean strength of 485 MPa combined with a KIc of 4.5 MPa [147].
The strength of SIC based materials with 50 vol-YOZrB2, HfB2,NbB2 or TaB2 par-
ticles also ranges between 400 and 500 MPa [ 1501. Similar strength values (480 MPa)
combined with a exceptionally higher fracture toughness of 7-9 MPa m1/2have been
reported for large scale lots of pressureless sintered 16 vol-YO TiB2 composites
[ 1513. Since the sintering was carried out with temperatures exceeding 2000°C yielding
98-99% of the theoretical density and an average particle size of 2.0 pm, it is obvious
that the reinforcing phase also acts as a grain growth inhibitor for Sic. The high
temperature strength of SiC-TiB2 and SiC-ZrB2 composites was found to remain
nearly constant at 480MPa up to 1200"C, and is hence superior to that of many
sialons [150,151].
Tani and Wada [ 1541fabricated optimized SiC-TiB2 composites by reactive sintering
starting from an intimate mixture of Sic, Ti02, B4C, and C powders. The mixture was
either hot pressed or pressureless sintered and post-HIPed at temperatures of > 1900°C.
Titanium diboride was formed in situ according to the reaction
Ti0 2 + 0.5B4C+ I .5C t TiB2 + 2C
during an intermediate heating step at 140CL15OO"Cin vacuum or argon atmosphere.
Overstoichiometric amounts of B4C and C (1-2 weight-% each) can be adjusted to aid
sintering. The primary advantages claimed for this reaction sintering process are the
use of water in powder processing due to the disuse of highly reactive, preformed TiB2
powders, and the very small size of reinforcing TiB2 particles formed in situ due to the
use of ultrafine Ti02, B4C, and C starting powders.
Effective reactive pressureless sintering of SiC-TiB2 composites was recently
reported by Blanc, Thevenot, and Treheux [155]. In addition they studied the
tribological behavior using a pin on flat configuration (flat: Sic, pin: SiC-TiB2).
In dry conditions the composites showed less wear resistance than monolithic
Sic, however, with water as lubricant the opposite was the case.
In the very recent study of Kuo and Kriven [156]indentation-strength tests were used
to determine the retained strength, flaw tolerance, and toughness-curve characteristics
of two kinds of SiC-TiB2 composites. BSiC-TiB2 composites which were hot-pressed
with an A1203sintering aid, were compared with the well-studied aSiC-TiB2 compo-
sites, which were pressureless sintered with boron and carbon additives. TiB2 (I 5 vol-
Yo)in the B- and C-doped aSiC only increased the retained strength without a
significant improvement in the toughening. On the other hand, TiB2 (3ovol-%)
along with the effect of AI2O3 sintering aid for the BSiC-TiB2 composite greatly
improved properties with a higher retained strength in long crack regions, better
flaw-tolerance behavior and a sharply rising toughness vs. crack size curve. The
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7 14 5 Silicon Carbide Based Hard Materials
t
76.8
19.2
3.7
58.6
37.0
42
39.0
560
4,7
20.0
74,5
5,5
'/. c c
a./
=/e
SIC
6
'
Figure 17. Flexural strength of pressureless sintered (= S, unhatched bars) and post-HIPed
(=HIPS, hatched bars) SiC-B4C particulate composites [ 1571.
different toughening behaviors for rx and pSiC-TiB2 were related to the weak nature of
the Sic-Sic and SiC-TiB2 interfaces as well as the fraction and size of TiB2.
5.3.3.9.3 Sic-B4C In the mid 1980's pressureless sintering and posthipping were
developed by Schwetz et al. [157] to produce 100% dense Sic-B4C composite
materials having SiC:B4C weight ratios within the range of from 90: 10 to 10:90
and a free carbon content of 4-5 weight-%.
These composites combine the good thermal shock resistance and oxidation resis-
tance of silicon carbide, with the hardness, wear resistance, and low specific gravity
of boron carbide. In this way a maximum strength of 550 MPa (four point bend) was
achieved for a composite of 59 weight-% Sic-37 weight-% B 4 C 4weight-% C (see
Fig. 17). The composite can be used in oxidizing atmospheres up to 1200 "C. Its micro-
structure is characterized by equiaxial B4C and graphite grains of < 1 pm diameter,
which were embedded in a matrix of S ic grains with an average grain size of 1.5 pm
(see Fig. 18). However, no improvement in fracture toughness was achieved, since the
fracture mode was almost 100% transgranular. Similar results on sintered Sic-B4C
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.3 Production of’ S i c 7 15
composites were obtained by Thevenot [158] and later by Tomlinson et al. [l 591, who
observed a 20% increase in strength when 25 vol-% B4C was added to Sic.
Excellent tribological properties for SiC-B4C-C composite materials were
recently encountered by Kevorkiijan et al. [ 1601.
In this study SiC-B4C-C seal rings for magnetic pumps were prepared by
pressureless sintering and characterized by a pin-on-disc method (medium: water,
pressure: 16-25 MPa, speed: 35-75 m sC’). The introduction of a lower level of
B4C (5-20 weight-%) particles into the S i c matrix resulted in an almost linear
decrease of wear rate. Moreover, further addition of B4C ( 2 0 4 0 weight-%) led to
an almost parabolic wear rate response. For example, with 40weight-% B4C a
decrease of the relative wear volume of seal rings by more than 55% was achieved.
These results recommend use of wear resistant SiC-B4C composites for heavily
loaded mechanical face seals in the pairing hard/hard against themselves. They
may be likewise suitable for the production of shaft protection sleeves and compo-
nents for sliding bearings whose wear resistance is to be improved.
5.3.3.9.4 SiC-AlN A series of solid solutions between S ic and A1N over the whole
composition range was concurrently discovered at the Universities of Utah and
Newcastle upon Tyne [ 1611, and has since received considerable attention [I 62-1 651.
A 2H wurtzite-type structure is formed by the carbothermal reduction of fine Si02
and A1203with a carbon source under nitrogen atmosphere at 1600°C:
3 s io 2 + 0.5AI2O3+ 4.5C + N2 1600°C ___i. 3SiC.AlN + 1.5CO + 0.5N2.
Kinetically favored is the carbothermal reduction of ct’ SiAlONs or a’ SiAlON -
CaSi9A130N15 1OC +
+
precursor mixtures (3Si3N4 3A1N CaO):
- +
1800 c
3(3SiC.A1N) + Ca + CO + 6N2.
Because the diffusion coefficients in covalent solids are extremely small, solid
solution was thought unlikely to be obtained by heating and annealing of the pow-
dered solid components.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
716 5 Silicon Carbide Based Hard Materials
Figure 19. Tentative SIC-A1N phase diagram after Zangvil and Ruh [166].
Zangvil and Ruh [166] however, obtained SIC-AIN solid solutions by hot
pressing powder mixtures
(1 -x)SiC+xAlN+ (SiC),-x(AIN)x
According to the phase diagram proposed by Zangvil and Ruh (see Fig. 19) at
temperatures above 2O0O0C, a 6H-4H-2H series of solid solutions appears with
increasing amounts of AlN. AlN strongly stabilizes specific polytypes (4H and
2H) at certain composition ranges, enabling the engineering of single-phase Sic
materials with discrete physical properties. At temperatures below about 19OO0C,
a miscibility gap was first proposed by Rafaniello et al. [163], a suggestion supported
by several later studies. Xu et al. [165] obtained strong SIC-AlN materials with
flexure strength up to 1 GPa. Several mechanisms of grain refinement resulting
from Sic-polytype transformations into a wurtzite (2H) solid solution, were
reported. Kuo et al. [167] discovered the formation of modulated structures
within the miscibility gap and Lee and Wei [168] reported that pressureless sintering
with 2weight-% Y 2 0 3 as sintering aid at 2050°C produced a duplex structure
composed of large (Sic), (AIN), grains and small Sic grains. Further solid solution
treatment (> 2225°C) followed by annealing within the miscibility gap (1860°C)
resulted in spinoidal decomposition, giving various duplex/modulated structures
with improved fractured toughness of the alloys.
In addition to their possible use as high temperature structural ceramics, materials
in the system SIC-AIN have potential as wide band-gap semiconductors and for
opto-electronic applications [ 1661.
A14C3.The resulting ceramic was fine grained (average grain size less than 5 pm) and
consisted of S i c (starting polytypes) and A120C as the two major phases. The
properties of the hot pressed ceramics varied with the amount of A120C, but at
an optimum composition of about 5-10 wei ht Y ' A120C, the strength (660 MPa)
- O.
and fracture toughness (KIc = 3.1 MPa m I$) obtained were comparable or super-
ior to the corresponding properties of commercial grades of sintered Sic. Huang
et al. [170] found encapsulation to be necessary for effective sintering with additions
of A1203 and A14C3, the densification occurring above M 1860°C. They attributed
densification to a transient liquid phase in the system A1203-A1404C in the route
to forming A120C with an eutectic temperature of ~ 1 8 4 0 ° C Four . point bend
strength, hardness and fracture toughness for the SiCAlON materials have been
reported [170]. The strength decreased with A4OC content in hot-pressed samples,
from ~ 6 0 MPa0 at 10 weight-% A120C to around 250 MPa at M 50% A120C. Most
significantly, the fracture toughness of some SiCAlON com ositions appeared to be
higher than that of S i c ( ~ 4 . compared
2 with 3.0MPam I f : ), using an indentation
technique. Lihrmann and Tirlocq [ 17I] proposed fabrication of sintered or hot-
pressed S i c based composites containing S i c as well as 5-30weight-YO solid
solutions composed of A120C and AIN. Starting from SiC-A1N-Al4C3-Al2O3
powder mixtures, densification was greatly enhanced by occurrence of a transient
liquid phase originating in the A1203-A14C3 system at temperatures above
1800°C. The composites with 5-10% (Al2OC-A1N) solid solution exhibit a mean
grain size < 2 pm and a mean strength of 620-670 MPa combined with a K,, of
5.1-6.8 MPam'/2. Strength and fracture toughness both retained their values up
to 14OOcC, before weakening, thus demonstrating the highly refractory nature of
the A120C-A1N second phase.
Table 7. Mechanical and physical data for continuous fiber reinforced S i c matrix composites
(MAN-Technology, Germany [ 1731.
At high temperatures (> 600°C) the composites degrade in strength and toughness
due to oxidation of C fibers and/or these interface layers and prevention of fiber
pullout (brittle fracture mode). Studies are in progress to increase oxidation resistance
by use of a CVD-SIC overlayer which seals the surface of the porous composites.
infiltration ceases [174]. Finally the body is heated to temperatures between 1000
and 1500°C for crystallization of the S i c matrix (LPI-C/SiC).
between 0.1 R cm and lo’* Ocm, depending on the concentration of the dopant [178].
Whereas compact, homogeneous SIC obeys Ohm’s Law, aggregates of S i c grains
show nonlinear current-voltage behavior. At low applied voltages they behave as
insulators, but when the applied voltage is increased above a certain value the
current increases exponentially. Thus, the points of contact between the grains
cause the electrical resistance to be voltage dependent [178,179].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
720 5 Silicon Carbide Based Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
122 5 Silicon Carbide Based Hard Materials
I 900 1
SIC-CERAMICS
800 -
600 -
solid st sintered
500 ---------------
..-- _/--
reaction banded
400 - (with free Si)
1 ,
\ ,
I
recrystallized
__
---------1---------'
100
ceramic-bonded
- -
-
- L
0
~ ~
1
Load P
Distance d
Load P
0 0.1 0.2
Distance from Edge d, mm
0.3 0.4 0.5
Figure 22. Edge toughness of LPS-Sic (EKasic T) compared to sintered Al,03 and Si3N4:
b
(a) flaking load against distance from edge, (b) schematic of experimental setup (Courtesy
Edge Flake
L.S.Sig1, ESK-Kempten/Germany)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Properties of Silicon Carbide 723
sintered S i c as seal face material was determined [201,202]. Since then, the mechan-
ical engineering industry has focused attention on this material and tribological
studies have been simulated to understand better those physical and chemical prop-
erties of SSiC that will affect its behavior when in contact with itself, other ceramics,
or metals. Model investigations involving Sic have shown the coefficients of fric-
tions against various materials, even itself, to be a function of contact stress [203]
and have documented the anisotropic wear behavior of monocrystalline S i c crystals
[204]. Lashway et al. [205] found that a controlled amount of porosity improved the
ability of sinteredSiC to retain a hydrodynamic film with lower friction. Seal tests
also indicated lower power dissipation at varying pressure-velocity values for com-
binations with sinteredSiC seal face materials.
Consistently positive practical experience gathered by Knoch et al. on sintered-
Sic seal rings [206,207] has helped the material to gain popularity rapidly for use
in situations involving wear problems.
Excellent results have been achieved in developing sliding bearings for hermeti-
cally sealed pumps [208]. All of them are absolutely leak-proof, whether in operation
or shut down and they are therefore of great value environmentally. They all have
the same design requirement that the sliding bearing of the pump shaft must be
flushed and lubricated by the pumped medium. Traditional materials simply
could not cope with the harsh conditions of the chemical industry and were quickly
destroyed by corrosion and abrasion. Sintered silicon carbide solved the problems.
Liquids containing abrasive particles do not restrict the use of sintered Sic.
Many sliding-wear problems occurring in the field are attributable to interruption
of the ideal (that is properly lubricated) running conditions, in which case the sliding
faces of the bearing or seal in question make contact with each other, giving rise to
solid-state or dry friction marked by a pronounced increase in the coefficients of fric-
tion. Local frictional heat leads to peak thermal stresses that may be of such inten-
sity as to cause a breakout of microstructural constituents. Then, when lubrication
(and cooling) is restored, the material is in danger of cracking or fracturing due to
thermal shock. Sintered silicon carbide, however, is better able to cope with such
situations than other ceramic materials, because it is stronger and has a lower ther-
mal expansion coefficient, and a higher thermal conductivity. Consequently, brief
periods of dry running can be survived.
The nonlubricated wear behavior of sintered silicon carbide under unidirectional
sliding at room temperature was studied by Derby et al. [209] as a function of load
and sliding time. At low loads polishing and ploughing mechanisms were observed
and the microstructure revealed an etched appearance. With increased loads and slid-
ing times microcracking at the grain boundaries occurred leading to subsequent grain
pull-out. In the study of Cranmer [210] it was shown that for sintered silicon carbide
sliding against itself also surface plastic deformation, ploughing and cracking were
operative as wear mechanisms. Miyoshi et al. [211] carefully studied friction and
wear of the pair SiC/iron in vacuum environment at temperatures up to 1500°C.
By using X-ray photoelectron spectroscopy it was found that the surface chemistry
of sintered silicon carbide (graphite and Si02) as well as the alteration of this chem-
istry by the temperature are very influential in dramatically changing the friction co-
efficient. Breznak et al. [212] quantified the role of initial surface roughness on the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Properties of Silicon Carbide 725
friction behavior of sintered silicon carbide rubbing against itself in cyclic oscillatory
motion. It was shown that the coefficient of friction decreased from an initial value of
0.40 to 0.25 as a result of surface polishing. Further work by Breval et al. [213]
indicated that adding graphite to Sic reduces the wear of SiCjSiC couples somewhat,
but improving their initial surface finish has the opposite effect: it leads to a greater
loss of material during running in. Wear debris of SiCjSiC couples exhibited a bi-
modal particle size distribution: some particles were micrometer size and the others
ranged in nanosize from 5 to 50nm. Using pin on disk, as well as abrasion wheel
test, Wu et af. [214] observed that the amount of wear increases with increasing
grain size for sintered silicon carbide.
Smythe and Richerson [215] conducted experiments in which the dynamic sliding
contact behavior at temperatures greater than 1000°Cwas studied. It was found that
surface film formation at higher temperatures governs the friction behavior. It is
believed that Si02 or Si02 modified by impurities from the SIC or the environment
with flow properties contributed to the lowering of the friction coefficient [216].
Tomizawa et al. [217] performed pin-on-disk experiments on friction of SSiC against
itself in water at room temperature. The authors found a friction coefficient of 0.26 and
noted that wear of sintered silicon carbide occurs by a combination of tribochemical
dissolution and the formation of pits by fracture of Sic grains. The amount of material
removal varied from one Sic grain to the other, due to a strong dependence of tribo-
chemical wear on crystallographic orientation of Sic grains.
Knoch and Kracker [2 181 observed that this anisotropic tribological behavior is
most conspicuous with EKasic D sintered silicon carbide, which has a bimodal
grain structure: about 3ovol-% of larger platelets of hexagonal a grains (about
l00pm long) and about 70vol-Y0 of smaller a grains (about 10pm long) as shown
in Fig. 23a. This material shows superior performance particularly if paired against
a softer carbon material. Figure 23b shows the surface topography developed in a
finely lapped surface under in-service conditions, as water lubricated EKasicj
carbon seal. This relief structure, with the depressions in a textured surface, obviously
acts as reservoirs for lubricant, thereby improving the emergency running properties
of the components. Figure 23c shows a sketch of the process. Boch et al. [219] studied
the dry friction behavior of SiCjSiC couples (fixed ball rotating disk) at temperatures
from 20 to 1000°C and observed a rapid decrease in the coefficient of friction from
0.45 to 0.16 when temperatures increases over 400°C. From 400 to 700°C, the
debris agglomerates in the form of SiOz rolls, which arrange themselves perpendicu-
larly to the sliding direction. These rolls act as minute roller-bearings decreasing the
coefficient of friction and preventing the formation of cracks in the wear track.
Habig and Woydt [220] studied unlubricated friction and wear of selfmated SSiC
sliding couples (EKasic D) at temperatures between 22 and 1000°C using a station-
ary pin and a rotating disk assembly. As shown in Fig. 24, at room temperature the
coefficient of friction decreases with increasing sliding velocity fromf = 0.8 to 0.6.
At temperatures >400°C a scatter of friction coefficients between 0.2 and 0.7 is
observed. By using small-spot-ESCA (electron spectroscopy for chemical analysis)
it was shown that the low friction coefficients of SiCjSiC were due to formation of
thin oxide layers formed by tribo-oxidation, while the higher friction coefficients
accompanied by higher wear coefficients were related to thicker oxide layers.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
126 5 Silicon Carbide Based Hard Materials
Figure 23. (a) SEM micrograph of a polished and plasma-etched section of sintered Sic, ‘EKasic D’,
note: bimodal grain size distribution. (b) SEM micrograph of EKasic D sliding face after about
4000 h of successful service as a mechanical seal face (SiC/carhon pair, lubricated with water,
maximum pressure 50 bar). (c) Schematic of the operational condition of a relief structured sliding
surface in the lubricated state and after breakdown of the lubricant film.
These results are in excellent agreement with those of Yamamoto et al. 12211, who
observed a decrease in the coefficient of friction when SiCjSiC couples had been
heated for 1 h at 1000°C with formation of a thin oxidation layer. However, at a
higher oxidation level the coefficient of friction increased rapidly. A low coefficient
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Properties of Silicon Carbide 727
of friction for the SiCjSiC couple was also measured by Martin et al. [222] who
studied friction under an oxygen partial pressure of 50 mPa, which should admit
only a very thin oxidation layer.
Sasaki [223] studied the influence of humidity on friction and wear behavior of
SiCjSiC couples. It was shown that humidity decreased the coefficient of friction
from 0.5 (dry air) to 0.2 (wet air), simultaneously a decrease in the coefficient of
wear was observed from (dry air) to mm3 N-l m-I.
Denape and Lamon [224] considered the important contribution of wear particles to
the wear mechanisms. The main findings were: polishing at low loads (5 N) is due to
fine individual wear particles smaller than 1 pm circulating in the sliding interface,
and abrasion and grain pull-out (at high loads of 20N) are associated with large
accumulations of particles adherent to the sliding phases. Wear tests performed
under water showed again an anisotropic wear of individual S i c crystallites. It
was concluded that the circulation of wear debris in the sliding interface is control-
ling both the wear rate and the friction response.
Woydt et al. [225] evaluated the tribological behavior of SIC-TIC and SiC-TiB2
ceramic-matrix composites using a special high temperature tribometer. In an
normal ambient environment these composites are able to form ‘lubricious oxide’
reaction layers on the hard substrate; the low friction coefficient of the couple
Sic--TiC/SiC-Tic at room temperature with values of f = 0.2. and 0.3 was
explained by the formation of self-lubricating layers of TiOz and SiC,O,,. The
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
728 5 Silicon Carbide Based Hard Materials
friction coefficient of self-mated SiC-TiB2 sliding couples was in the range 0.56-0.72
and compared to sintered aSiC materials no improvement was observed in the wear
resistance of the SiC-TiB2 composite [226].
In a study on the friction and wear behavior of lubricated ceramic journal bear-
ings it was shown by Maurin-Perrier et d.[227] that both SiCjSiC and SiC/Si3N4
couples provide a significantly better behavior in terms of film stability compared
with the classical materials used for water- or hydrocarbon-lubricated bearings. A
self-improvement of surface roughness during the running-in period by tribochem-
ical reaction significantly increases the range of stability of the film; consequently it
is not necessary to produce a highly expensive surface finish to increase the perfor-
mance of the SiCjSiC couple. The lifetime of the components is mainly determined
by the behavior during the start and the stop phases where boundary lubrication
occurs. Sliding wear on sintered silicon carbide leads to a very smooth surface
between residual porosity. Further it was found [228] that in dry run situations,
sintered Sic of bimodal grain structure (EKasic D) outperforms other Sic
materials. The bimodal EKasic D shows the lowest development of heat, and there-
fore it has the lowest coefficient of friction. In a mayor US study on 'Tribological
Fundamentals of Solid Lubricated Ceramics' [229] the bimodal grain size/shape
distribution was judged best in terms of wear resistance, when compared with
other sintered aSiC materials.
Loffelbein et al. [230] tested self-mated sliding couples of SSi3N4,HIP-RBSi3N4,
SSiC, SiSiC, MgO-Zr02 and A1203 in different aqueous solutions (H20, NaOH,
KOH, NH40H, HN03, H2S04, H3P04, CH3COOH, HCl, HC104) under condi-
tions of boundary or mixed lubrication. The best frictional behavior was observed
with couples of SSiC, with steady state friction coefficients of 0.05. The lowest wear
coefficients were measured for couples of the two types Si3N4,SSiC and A1203with
values of approximately 10-7mm3N-'m-'. If low friction and low wear are
required, couples of SSiC seem to be the best choice in most aqueous solutions. A
recent study by Anderson and Blomberg [231], based on tests with sintered silicon
carbide sliding unlubricated on itself in point (pin-on-disc), line (journal bearing)
and plane (mechanical face seal) contacts. Tribo-oxidation and surface fracture
were identified as the dominating deterioration mechanisms. The oxidation
products formed silicon dioxide and, within narrow operational regimes, silicon
monoxide. The highest wear rates occured in the pin-on-disc configuration, while
the lowest rates were obtained in the journal bearing tests.
Kitaoka et al. [232] investigated the effects of temperature and sliding speed on the
tribological behavior of sintered silicon carbide (0.1 weight-% B/l .O weight-% C) by
sliding on the same material in deoxygenated water from room temperature to
300°C under high vapor pressures (120°C/2 bar, 3OO0C/85 bar). The wear mechan-
ism appears to consist of hydrothermal oxidation of Sic according to the equations:
Sic + 2 H 2 0 + Si02 + CH4,
S i c + 4H20 4 Si02 + C 0 2 + 4H2
S i c + 2H20 Si02 + 0.5C2H6+ 0.5H2,
4
and dissolution of reaction products such as silica. Fine mirrorlike worn surfaces
were observed without wear debris under all sliding conditions.
For description of the tribotechnically used machine-elements and more informa-
tion about tribology, the book ‘Tribology Handbuch Reibung und Verschleip’
~
Figure 25. SEM micrographs of sliding faces (left) and polished sections (right) of sinteredSiC
grades: EKasic D (top), TRIBO 2000 (middle), and TRIBO 2000-1 (bottom). The graphite in the
TRIBO 2000-1 microstructure was revealed as white particles after coating the polished section
with a carbon film.
shows stable hydrodynamics down to slower sliding velocities, before the lubricant
film breaks down and the bearing comes into solid state contact. Figure 28 is an
example.
In this case, perhaps unexpectedly, the TRIBO 2000-1 material shows both a
higher hydrodynamic friction and a deviation with increasing velocity when com-
pared with TRIBO 2000 and EKasic D.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Properties qj" Silicon Cavhidp 73 1
250
-5 200
a
-
150
a
.->
5 100
50
00
0
Flexural Strength Weibull Fracture Youngs Thermal Thermal Corrosion coeffof ~ ~ ~ c t ~
4-Point Modulus Toughness Modulus Conductlvity Expansion Resistance khaml Seal
p.25 bar v=9 4 in's
Figure 26. Comparison of properties of the pore-containing material (TRIBO 2000) and the pores-
plus-graphite-particles-containing material (TRIBO 2000- 1) with EKasic D
Summarizing all Stribeck test data (load, media, and up to 10000 repeats) the
TRIBO 2000 material shows the best performance in terms of stabilizing the hydro-
dynamic film down to the lowest velocities. This is important for sliding bearings
and has been confirmed many times. The poorer performance of the TRIBO
2000-1 material in this typical sliding bearing test is not clearly understood, but it
Lubricant
stationary body
Surface roughness R
0,OI
Viscosity . Rubbing speed v
__f
load
Figure 27. Schematic of a Stribeck curve and the four modes of lubrication
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
132 5 Silicon Carbide Based Hard Materials
5
a
7
=c 4
.-
c
0
.o 3
IL
o ~ ~ ' ' ' i ' ' " i ' ' " ~ ' " ' ~ ' ' " l
0 5 10 15 20 25
Sliding Velocity [m/s]
Figure 28. Coefficient of friction as a function of sliding velocity for self-mated sliding pairs of sin-
tered S i c materials.
may have something to do with the poor wetting behavior of graphite in water under
normal pressure.
In mechanical seal applications, under differential pressures of less than 15 bar the
same ranking of materials is observed in terms of friction and wear. Under extreme
situations, however, like very high differential pressure (producing continuous
boundary lubrication) or pump cavitational run (causing breakdown of the hydro-
dynamic film under frequent violent loads), the TRIBO 2000-1 material showed the
best performance (see Fig. 29).
To diminish the effects of hydrodynamic grain-boundary corrosion in applica-
tions with hot water, such as chipping out of fine grains in hot spots, tribochemical
reaction with water
S i c + 2H20 = Si02 + CH4
1.2 I I
.a
.-0 - +
m n
a"
Y
.-
0
In
0
Figure 29. Relative wear volume on seal rings of sintered Sic material (self-mated couples).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.4 Properties of Silicon Carbide 733
Figure 30. Microstructure of EkasicW: coarse-grained SSiC with platelet structure for mechanical
seals and bearings (hard/hard combinations) in hot water applications.
and formation of damaging SiOz layers on the sliding surfaces, another two new
modified SSiC materials with a predominantly coarse-grained, bimodal platelet
structure have been developed [235]. Catastrophic failure of coponents is avoided
since the large S i c platelets near the surface are anchored to a depth at which
there is no grain-boundary corrosion.
EKasic W silicon carbide is a dense material with a predominantly coarse-grained
bimodal platelet structure. The effectiveness of the coarser microstructure (see
Fig. 30) in improving the corrosion resistance has been clearly demonstrated in prac-
tical tests on a mechanical seal test rig. Even after 500 h of testing (deionized water,
60°C, 6 bar, hard/hard couples) no SO2 layer was formed on the functional surface.
Figure 31. Microstructure of EKasic HW: coarse-grained SSiC with platelet structure and graphite
particles (black phase 4 0 4 0 pm in diameter) for mechanical seals and bearings (hard/hard combi-
nations) in hot water applications.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
734 5 Silicon Carbide Based Hard Materials
In order to characterize S i c powders and sintered ceramics the total oxygen and
nitrogen content as well as the contents of metallic impurities are analyzed. Total
oxygen and nitrogen contents are usually determined by an inert gas fusion
method (Leco T C 436) using powdered samples, whereas metallic impurities (Na,
K , Ca, Mg, V, Fe, Ti, Al, Cr and Ni) and boron content are determined in acidic
solutions by inductive plasma emission (ICP) spectroscopy [240-2421.
Besides chemical analysis, physical properties such as particle size, particle size
distribution, and bulk density are also important.
The particle size distribution of coarse abrasive materials (5 mm-50pm) is
determined by sieving. For particle sizes less than 50 pm, sedimentation and laser
diffraction methods are used.
Both the particle size distributions and the methods for their determination are
standardized. The standard for bonded abrasive applications [243] differ from
those for coated abrasives [244]. The bulk density depends both on the particle
size distribution and on the particle shape. It is measured by weighing a known
volume of S i c grains [245]. The packed density is often determined instead of the
bulk density [246].
According to the analysis of the microstructure [247] and the physical properties
measurements, the reader is referred to the literature [93,248-25 11. The major
sequential steps in conducting analysis and properties evaluation for SIC ceramics
are shown in Fig. 32.
Raw Materials
to Ceramic Materials
material, for example, in linings and skid rails for furnaces and hot cyclones, and as
a kiln furniture, especially in saggers [61,265-2691. The good electrical conductivity
of the material at high temperatures, coupled with its outstanding oxidation resis-
tance, led to its early use in the electric heating industry [270-2721, which markets
its products mainly in the form of rods and tubes that operate up to 1500°C.
Recrystallized S i c igniters are used in home gas appliances, replacing pilot lights.
High-purity S i c shapes are used in the electronic industry as furnace components
for processing of silicon wafers. The thermoelectric properties of S i c suggest the
use of sintered S i c rods as high-temperature thermoelements [273] and as Seebeck
elements [274] for high-temperature thermoelectric energy conversion. Voltage-
dependent resistors (varistors) consist of ceramic- or polymer-bonded S i c and are
used in overvoltage protection equipment.
Silicon carbide is an outstanding material for the construction of electronic equip-
ment. Blue light-emitting diodes having an improved 470 nm peak wavelength are
being produced and marketed as the first commercial S i c semiconductor device.
The continual development of the deposition of S i c thin films and of large
diameter single crystal S i c wafers, the associated technologies of doping, etching
and electrical contacts have culminated in a host of new solid state devices including
field effect transistors capable of operation up to 650°C [275].
High-density, high-strength SiSiC, SSiC, HP-, and HIPSiC materials, which have
been on the market only since the 1970s, have opened up a new field of application
[276], namely, in mechanical and high-temperature engineering. SiSiC and,
especially SSiC, are displacing the chemically less resistant tungsten carbide (hard
metal) and the thermal-shock-sensitive aluminum oxide in modern mechanical
seals where they are used in the form of slide rings (see Fig. 33). The excellent
wear resistance of sintered silicon carbide, its excellent chemical resistance, and
outstanding tribological characteristics ensure that mechanical seals made of this
material last longer, resulting in much reduced maintenance and production costs
for pump-dependent processes in the chemical industry [277].
SiSiC seal rings (in contrast with those made of pure SSiC) can only be used in acid
media because of the accompanying silicon phase, which is attacked by alkalis [207].
For similar requirements involving radial loads, sliding bearings (see Fig. 34) are
manufactured from SSiC. The erosion and chemical resistance of S i c enable the
designer to position the bearings in the medium to be transported; that is, to elim-
inate lubrication and sealing problems. Other components include shaft protection
sleeves for waste gas exhaust fans, and precision spheres for dosing and regulating
valves.
Due to the tailored properties of liquid phase sintered silicon carbide (LPSSiC) it
is used as dewatering elements in the paper machinery and as rings for highly
stressed gas seals. It is a price competitive alternative to silicon nitride materials
and outperforms alumina and tungsten carbide materials. In addition, LPSSiC is
proposed as neutral matrix in ceramic matrix composites containing plutonium to
burn the world’s stockpiles of military plutonium in thermal or fast reactors [278].
Hot-pressed SIC is the preferred material to replace oxide ceramics for rods
fixtures, and punches in high-temperature strength testing equipment. In view of
the low level of plasma contamination, the low induced radioactivity, and excellent
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
5.7 Uses of Silicon Carbide 739
Figure 35. Monolithic SSiC gas turbine rotor (courtesy A.Lipp, ESK-Kempten/Germany).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
740 5 Silicon Carbide Based Hard Materials
reentry vehicles [173]. Due to the high thermal shock capability of this class of
materials and to the high fracture toughness of some of these composites some
civil applications are gaining importance: highly loaded brake discs, for high
speed trains, and highly loaded journal bearings; in both applications the con-
ventional materials (metals for brakes, monolithic ceramics for bearings) are not
applicable because of the thermal loads and the brittle failure mode respectively.
Acknowledgments
Some parts of the text appeared in the earlier review by K.A.S. (Silicon Carbide in:
Encyclopedia of Advanced Materials pp. 2455-2461,1994) and Elsevier Science Ltd.,
Kidlington OX5 16B, UK is kindly thanked for permission to use them.
References
I. R. Kieffer and F. Benesovsky, Hartstoffe, Springer Verlag,New York, 1963.
2. J. J. Berzelius, Am. Phys. Chem. 1824, 1, 169-230.
3. C. M. Despretz, Compt. Rend. 1849, 29, 709-724.
4. P. Schutzenberger and A. Colson, Compt. Rend. Akad. Sci. 1881,92, 1508.
5. H. Moissan, Der elektrische Ofen, Verlag M. Krayer, Berlin, 1900.
6. E. G. Acheson US Pat. 492 767 (28 Feb. 1893), DE 76629 (1894), DE 85195 (1896).
7. Gmelin Handbook of Inorganic Chemistry, 8th edition, Silicon Supplement Vo.B 3, System No.
15, H. Katscher et al. (eds.), Springer Verlag, Berlin, 1986.
8. H. Moissan, Compt. Rend. Akad. Sci. 1905, 140, 405.
9. A. J. Regis,Land . B. Sand, Bull. Geol. Soc. Am. 1958, 69, 1633.
10. A. P. Bobrievich, V. A. Kalyuzhnyi, and G. I. Smirnov, Proc. Akad. Sci. USSR: Geol. Sci.
1957, 115, 757.
11. J. Bauer, J. Fiala, and R. Hrichovi, Am. Miner. 1963, 48, 620-34.
12. H. Kleykamp and G . Schumacher, Ber. Bunsenges. Phys. Chem. 1993,97, 799-805.
13. A. Dietzel, H. Jagodzinski, and H. Scholze, Ber. Dtsch. Keram. Ges. 1960, 37, 524-537.
14. R. Kieffer, E. Gugel, P. Ettmayer, and A. Schmidt, Ber. Dtsch. Keram. Ges. 1966,43,621-623.
15. N. W. Jeeps and T. F. Page, J. Am. Ceram. Soc. 1981, 64, C177-C178.
16. R. F. Adamsky and K. M. Merz, Z. Kristallogr. 1959, 111, 350-361.
17. H. Ott, Z. Kristallogr. 1925, 61, 515-532; 62, 210-218; 63, 1-19.
18. L. S. Ramsdell, Am. Miner. 1947, 32, 64-82.
19. J. Ruska, L. J. Gauckler, J. Lorenz, and H. U. Rexer, J . Muter. Sci. 1979, 14, 2013-2017.
20. Y. Inomata, Z. Inoue, M. Mitomo, and H. Suzuki, Yogyo-Kyokai-Shi, 1968, 76, 313-319.
21. D. Lundquist, Acta. Chem. Scand. 1948, 2, 177-191.
22. P. T. B. Shaffer, Mater. Res. Bull. 1969, 4, S13-24.
23. M. Mitomo, Y. Inomata, and M. Kumanomido, Yogyo-Kyokai-Shi1970,78, 365-369.
24. M. Mitomo, Y. Inomata, and H. Tanaka, Muter. Res. Bull. 1971, 6, 759-764.
25. M. M. Patience, Silicon Carbide Alloys, PhD. Thesis, University of Newcastle upon Tyne,
1983.
26. D. Foster, Densification of Silicon Carbide with Mixed Oxide Additives, University of
Newcastle upon Tyne, 1996.
27. N. W. Jeeps and T. F. Page, J . Microscopy 1979, 116, 159-171; 1980, 119, 177.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 741
66. K. M. Taylor, Improved K TSiC for high temperature parts, Muter. Meth. 1956, Vol. 44, No.
4, 92-95; US Patent 3 189 472, 1965.
67. R. Kieffer, E. Gugel, and A. Schmidt, Verfahren zur Herstellung eines dichten Formkorpers
auf Basis Sic, DE-OS 1 671 092, 1967.
68. P. Kennedy and J. V. Shennan, Engineering applications of REFEL silicon carbide, in Silicon
Carbide 1973, R. C . Marhall et a/. (Eds), Univ. of South Carolina Press, Columbia, 1974,
pp. 359-366.
69. G. Q. Weaver, H. R. Baumgartner, and M. L. Torti, Special Ceramics 6, P. Popper (Ed.), Brit.
Ceram. Res. Association, UK, pp. 261-281.
70. W. B. Hillig, R. L. Mehan, C. R. Morelock, V. J. De Carlo, and W. Laskow, Am. Ceram. Soc.
Bull. 1975, 54,1054-1056.
71. T. Hase et al., J . Nucl. Muter. 1976, 59, 4248.
72. Anon. A super ceramic called REFEL, Interceram. 1982, No. 1, 50-51.
73. G. Willmann and W. Heider, Werkstofftechnik1983, 14,158.
74. K. Taylor, Dense silicon carbide, US Patent 3 205 043, 1962.
75. S. Prochazka, The sintering process for Sic, a review, Techn. Report 81-CRD-314, General
Electric, Schenectady NY, 1981, pp. 16.
76. S. Prochazka, Sintering of Sic, in Ceramics for High Performance Applications, J. J. Burke et al.
(eds.), Metals and Ceramic Info. Center, Columbus, OH, 1974, pp. 239-252.
77. J. A. Coppola and G. H. McMurtry, Substitution of ceramics for ductile materials in
design, Nut. Symp. on Ceramics in the Service of Man, Carnegie Institution, Washington
DC, 1976.
78. P. T. B. Shaffer, The Si c phase in the system SiC-B4C-C, Muter. Res. Bull. 1969,4, 213-220.
79. R. Hamminger, Carbon inclusions in sintered Sic, J . Am. Ceram. SOC.1989, 72, 1741-1744.
80. Y. Murata and R. H. Smoak, Densification of Sic by the addition of BN, BP and B4C, and
correlation to their solid solubilities, in Proc. Int. Symp. Densijcation and Sintering, Oct. 1978,
Hakone/Japan, S. Somiya and S. Saito (Eds), Gakujutsu Bunken, Tukyu-Kai, Tokyo, 1979,
pp. 382-399.
81. W. Boecker and H. Hausner, Powder Metall. In?. 1978, 10, 87-89.
82. D. R. Stutz, S. Prochazka, and J. Lorentz, 1. Am. Ceram. SOC.1985, 68,479482.
83. T. Fetahagic and D. Kolar, Ceram. Acta 1990, 2, 31-37.
84. K. A. Schwetz, A. Lipp, The effect of boron and aluminum sintering additives on the proper-
ties of dense sintered aSiC, Science of Ceramics 10, Berchtesgaden Sept. 1979, H. Hausner
(Ed.), Verlag Deutsche Keramiske Ges 1980, pp. 149-158.
85. W. Grellner, K. A. Schwetz, and A. Lipp, Fracture phenomena of sintcred alphasic, in Proc.
7th Symp. on Special Ceramics, Bedford College, London, Dec. 1980, D. Taylor and P. Popper
(Eds), Br. Ceram. Res. Assoc., Stoke on Trent 1981, pp. 27-36.
86. W. Boecker, H. Landfermann, and H. Hausner, Powder Metall. Int. 1979, 11, 83-85.
87. H. Hausner, Pressureless sintering of non-oxide ceramics, in 4th CIMTEC, Energy and Cera-
mics, St. Vincent, Italy 1979, P. Vincenzini (Ed.), Elsevier Scientific Publ. Comp., Amsterdam,
1980, pp. 582-595.
88. H. Tanaka, Y. Inomata, K. Hara, and H. Hasegawa, J . Muter. Sci. Lett. 1985, 4,315-317.
89. K. Suzuki, Pressureless sintering of SIC with addition of AI2O3,Rep. Rm. Lab. Asahi Glass Co.
1986,36, 25-36.
90. K. Suzuki, Pressureless sintering of S i c with addition of AI2O3,in SIC Ceramics - 2, S. Somiya
and Y. Inomata (Eds), Elsevier, London, 1991, pp. 163-182.
91. R. H. Smoak, Pressureless sintering beryllium containing SIC powder composition, DE-OS 27
51 851, 1977.
92. A. Mohr, Untersuchungen zur Minimierung der Additivgehalte fur die drucklose Sinterung
von aSiC, Diplomarbeit Institut fur Keramik im Maschinenbau, Universitat Karlsruhe, 1989.
93. D. W. Richerson, Modern Ceramic Engineering, Dekker, New York, 1992.
94. F. Thummler and R. Oberacker, Introduction to Powder Metallurgy, The Institute of Materials
Series, University Press, Cambridge, 1993.
95. C. A. Johnson and S. Prochazka, Microstructures of sintered Sic, in Ceramic Microstructures
1976, R. M. Fullrath and J. P. Pdsk (Eds), Westview Press, Boulder, CO, 1977, pp. 366-378.
96. K. A. Schwetz, F. Isemann, and A. Lipp, Injection molded sintered turbine components of Al-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 143
doped alpha-Sic; in Proc. 1st Internat. Symp. Ceramic Comp. For Engine, S. Somiya et a/.
(Eds), KZK Scientific, Tokyo, 1984, pp. 583-594.
97. Showa Denko, DE 3 927 300, 1988.
98. Ceramiques et Composites, European Patent 486 336, 1991.
99. Carborundum Company, US Patent 5 589 428, 1996; US Patent 5 635 430, 1997.
100. ESK-GmbH, European Patent 685 437, 1995.
101. Carborundum Company, European Patent 145 496, 1984.
102. J. Greim, L. Sigl, and H. Thaler, Sintered SIC for high-performance bearings and seals, Maga-
zine of Wacker Chernie-GmhH Werk+ Wirken: Int. Edn, 1997, 1, 23-25.
103. M. Omori and H. Takei, J. Am. Ceram. Soc. 1982, 65, C-92; and US Patent 5 439 853,
1995. .
104. E. Kostic, Powder Metall. Int. 1988, 20, 28-29.
105. R. A. Cutler and T. B. Jackson, Liquid phase sintered Sic, in Ceramic Materials and Components
for Engines, Proc. 3rd. Int. Symp., V. J. Tennery (Ed.), Am. Ceram. Soc., 1988, pp. 309-318; and
US Patent 4 829 027, 1988.
106. W. Boecker and R. Hamminger, Interceram. 1991, 40, 520-525.
107. F. K. Van Dijen and E. Mayer, J. Europ. Ceram. Soc. 1996, 16, 413420.
108. C. Wolf, H. Hubner, and J. Adler, Mechanical behavior of pressureless sintered SIC at high
temperature, in 3rd Euro-ceramics, Vol.3, P. Duran and J. F. Fernandez (Eds), Faenze Editrice
Iberica, Castellon, Spain, 1993, pp. 465470.
109. K. Y. Chia and S. K. Lau, Ceram. Eng. Sci. Proc. 1991, 12, 1845-1861.
110. K. A. Schwetz, P. Matje, M. Mohr, and K. P. Martin, Development of a pressureless sintering
process for ceramics in the system SiC--AlN (0),Final Report on German BMFT Project No.
03-zC-153, 1986 (unpublished).
111. I. Wiedmann, M. Nader, M. J. Hoffmann, and F. Aldinger, in Symposium 7, Material-
wissenschaftliche Grundlagen, Werkstqfimhe '96, F. Aldinger and H. Mughrabi (Eds),
DGM Informdtionsgesellschaft mbH, Oberursel, 1997, pp. 51 5-520.
1 12. L. S. Sigl and H. J. Kleebe, J. Am. Ceram. Soc. 1993. 76, 773-776.
113. M. P. Padture, J. Am. Ceram. Soc. 1994, 77, 519-523; 2518--2522.
114. S. K. Lee and C. H. Kim, J. Am. Ceram. Soc. 1994, 77, 1655-1658.
115. J. S. Nadeau. Ceram. Bull. 1973, 52, 170-174.
116. R. A. Alliegro, L. B. Coffin, and J. R . Tinklepaugh, J. Am. Ceram. Soc. 1956.39, 386389.
117. S. Prochazka and R. J. Charles, Am. Ceram. Soc. Bull. 1973, 52, 885-891.
118. J. Kriegesmann, Ber. Dtsch. Keram. Ges. 1978, 55, 391-397.
119. F. F. Lange, J. Mafer. Sci. 1975, 10, 314320.
120. J. M. Bind and J. V. Biggers, J. Appl. Phys, 1976, 47,5171-5174.
121. D. Broussaud, Independence of composition of hot-pressed Sic, in Ceramic Microstructures
'76, R. M. Fullrath and J. A. Pask, (Eds), Westview Press, Boulder, CO, 1976, pp. 679-688.
122. T. Iseki, K. Arakawa, H. Matsuzaki. and H. Suzuki, Yogyo-Kyokai-Shi1983, 91, 349.
123. K. Nakamura and 0. Asai, Kagaku Kogyo 1982, 33, 977.
124. H. Kessel and E. Gugel, Industrie Diamanten Rundschuu 1978, 12, 180-185.
125. J. Kriegesmann, K. Hunold, A. Lipp, K. Reinmuth, and K. A. Schwetz, European Patent
71241. 1981.
126. T. J. Whalen, R. M. Williams, and B. N. Juterbock, HIP of Sic and Si3N4structural ceramics,
in IOth Plansee Seminar Proceedings, Reutte, Tyrol, 1985, p. 783.
127. K. Hunold, Powder Metall. Int. 1984, 16, 236-238; 1985, 17, 91-93.
128. K. Hunold, Sinter/HIP of Sic, in Proc. Adv. Mat. Technology Ceramic Workshop No. 4 ,
Advances in Materials, Processing and Manufacturing Science, Nagoya Japan, March 3 4 ,
Japan Fine Ceramics Center, 1988, pp. 49-62.
129. R. Oberacker, A. Kiihne, and F. Thummler, Powder Metall. Int. 1987, 19, 43-50.
130. F. Fetahagic, R. Oberacker, and F. Thummler, Process development for sinter-HIPing of Sic,
in Ber. KFA-Jiihlich 1989, Juel. Conf: -77, Emerging Mat. Adv. Process., 9th German-Yugosla-
vian Meeting on Materials Science Development, April 1 6 1 9 , 1989, pp. 313-325.
131. T. Hirai and M. Sasaki, S i c prepared by CVD, in S i c Ceramics-I, S. Somiya and Y. Inomata
(Eds), Elsevier, London, 1991, pp. 77-98.
132. Anon, Material Innovations: CVD scaled up for commercial production of bulk Sic, Am.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
744 5 Silicon Carbide Based Hard Materials
improved strength, in Silicon Carbide '87, J. W. Cawley and C. Semler (Eds), Am. Ceram. Soc.,
Columbus, OH, 1989, pp. 227-240.
170. J. L. Huang, A. C. Herford, R. A. Cutler, and A. V. Virkar, J . Muter. Sci. 1986,21, 1448-1456.
171. J . M. Lihrmann and J. Tirlocq, Process for producing dense products based on S i c and
composite products thus obtained, World Patent, PCT, WO 97j06119, 1997.
172. D. P. Stinton, R. A. Lowdon, and R. H. Krabill, Mechanical property characterization of
fiber-reinforced Si c matrix composites, ORNL/TM- 11524, April 1990.
173. H. Wurtinger and A. Miihlratzer, Cost effective manufacturing methods for structural ceramic
matrix composite (CMC) components, A S M E Paper 96-GT-296, 1996.
174. E. Fitzer and R. Gadow, Fiber reinforced Sic, Am. Ceram. Soc. Bull. 1986, 65, 326335.
175. G. S. Corman, M. K. Brun, P. J. Meschter, K. L. Luthra, and R. Eldrid, Toughened silcomp
ceramic composites for gas turbine applications, Proc. 39th Internat. S A M P E Symposium,
April 1 1-1 3 1994, pp. 2300-23 13.
176. Ullmanns Encyclopedia of Technical Chemistry (E. Bartholome et al., eds.), 4th edn., 21,
Verlag Chemie, Weinheim, 1982, 43 1438.
177. P. Wecht, Feuerfest-Siliciumcarbid, Applied Mineralogy, Vol. 11, Springer Verlag, New York,
1977.
178. K. Ziickler, in Halbleiterprobleme 111, W. Schottky (Ed.), 1956, pp. 207-229.
179. S. H. Hagen, Ber. Dtsch. Keram. Ges. 1970, 47, 630-634.
180. E. Gugel, P. Schuster, and G. Senftleben, Stahl Eisen 1972,92, 144-149.
181. Anon, JANAF Thermochemical Tables, 2nd edn., NSRDS-NBS37, Washington DC, June
1971.
182. K. Niihara, Am. Ceram. Soc. Bull. 1984,63, 1160-1164.
183. T. F. Page, Proc. Brit. Ceram. Soc. 1978, 26, 193-208.
184. W. Kollenberg, B. Mossner, and K. Schwetz, VDI-Berichte 1990, Nr. 804, 347-358.
185. M. E. Washburn and W. S. Coblenz, Ceram. Bull. 1988,67, 356-363.
186. F. Thummler, Sintering and high temperature properties of Si3N4 and Sic, in Sintering
Processes, G. C. Kuczynski (Ed.), Plenum, New York, 1980, pp. 247-277.
187. J. Kriegesmann, A. Lipp, K. Reinmuth, and K. A. Schwetz, Strength and fracture toughness
of SIC, in Ceramics for High Performance Appl. 111, E. M. Lenoe et al. (Eds), Plenum, New
York, 1984, pp. 737-75 1.
188. R. Hamminger, R. G. Grathwohl, and F. Thummler, Microchemistry and high temperature
properties of sintered Sic, Proc. 2nd Int. ConJ Science of Hard Materials, Rhodos Sept.
1984, Inst. of Physics Conf. Series No. 75, Hilger, Bristol, 1986, pp. 279-292.
189. G. Grathwohl, Th. Reets, and F. Thiimmler, Creep of hot-pressed and sintered S i c with
different sintering additives, in Sci. Ceram. 1981, 11, 425431.
190. J. L. Chermant, R. Moussa, and F. Osterstock, Rev. Int. Hautes Temp. Refract. 1981,18,5-55.
191. G . Wiebke, Ber. Dtsch. Keram. Ges. 1960, 37, 219-226.
192. J. Schlichting, Ber. Dtscli. Keram. Ges. 1979, 56, 196199 and 256261.
193. J. Schlichting and K. Schwetz, High Temp. High Press. 1982, 14, 219-223.
194. B. Frisch, W. R. Thiele, R. Drumm, and B. Miinnich, Ber. Dtsch. Keram. Ges. 1988,65,277-284.
195. K. G. Nickel, Z. Fu, and P. Quirmbach, Trans A S M ;J. Eng. Gas Turb. Power 1993,115,7682.
196. K. Konopicky, I. Patzak, and H. Dohr, Glus Email Keramo. Tech. 1972, 23, 81-87.
197. D. W. McKee and D. Chatterji, J . Am. Ceram. Soc. 1976, 59,441444.
198. N. W. Jepps and T. F. Page, J . Microscopy 1981, 127, 227-237.
199. P. T. B. Shaffer, Ceram. Age 1966,82, 4244.
200. T. Iseki, T. Kameda, and T. Maruyama, J . Muter. Sci. 1984, 19, 1692-1698.
201. T. Labus, Lubr. Eng. 1981,37, 387-394.
202. J. H. Eisner, Chemie-Anlagen + Verfahren, 1982, p. 46, p. 51, p. 54.
203. D. W. Richerson, Muter. Sci. Res. 1981, 14, 661-676.
204. K. Miyoshi, Proc. Int. Con$ Wear of Mat., ASME, San Francisco 1981.
205. R. W. Lashway, S. G. Seshadri, and M. Srinivasan, Various forms of Sic and their effects on
seal performance, 38th Ann. Mtg. of the ASLE, Houston, Texas, 1983.
206. H. Knoch, J. Kracker, and A. Schelken, Chemie-Anlagen + Verfahren, 1983, 28-30.
207. H. Knoch, J. Kracker, and A. Schelken, Cliemie-Anlagen + Veufahren, 1985, 101-104.
208. H. Knoch, J. Kracker, and A. Schelken, World Ceramm 1985, Vol. 2, 96-98.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
746 5 Silicon Carbide Based Hard Materials
Braunschweig/Wiesbaden, 1992.
234. H. Knoch, L. Sigl, and W. D. Long, Product development with pressureless sintered Sic, 1
Oct. 1990, 37th Sagamore Army Mat. Res. Conf. AMTL-Watertown MA.
235. J. Greim, H. Thaler, and L. Sigl, Gesintertes Siliciumcarbidfur hochbeanspruchte Komponenten
in der Lager- und Dichtungstechnik, in VDZ- Bericht 1331, Dusseldorf, 1997, pp. 153-159.
236. DIN 51 075 part 1-5, 1982, Beuth Verlag GmbH, Berlin/Germany.
237. I S 0 9286, 1997, Beuth Verlag GmbH, Berlin/Germany.
238. K. A. Schwetz and J. Hassler, J . Less Common Metals, 1986, 117, 7-15, and DIN 51079-3,
1997.
239. DIN 51 076, part 1, 1991, Beuth Verlag GmbH, Berlin/Germany.
240. F. F. van Dijen, Interceram 1993, 42, 92-94.
241. J. A. C. Broekdert, R. Brandt, F. LeiP, C. Pilger, D. Pollmann, P. Tschopel, and G. Tolg, J . Anal.
Atom. Spectr. 1994,9, 1063-1070.
242. G. Zaray, F. Leis, T. Kantor, J. Hassler, and G. Tolg, Analysis of silicon carbide powder by
ETV-ICP-AES, Fres. J . Anal. Chem. 1993,346, 1042-1046.
243. DIN-IS0 8486, part 1-2, 1997, Beuth Verlag GmbH, Berlin/Germany.
244. DIN 69 176, part 1-3, 1985.
245. FEPA 44-D-1986, Fachverband Elektrokorund-und Sic-Hersteller, Fr.a.M.
246. I S 0 787, part 11, 1991, Beuth Verlag GmbH, Berlin/Germany.
247. V. Carle, U. Schafer, U. Taffner, F. Predel, R. Telle, and G. Petzow, Ceramography of high
performance ceramics, Part I1 silicon carbide, Pruct. Met. 1991, 28, 420-434.
248. W. Kollenberg, Prufverfahren keramischer Hochleistungswerkstoffe und deren Grundlagen,
KfA-Bericht Nr. 470, Zentralbibiothek der Kernforschungsanlage Julich GmbH, 1988.
249. L. L. Hench and R. W. Gould, Characterization of Ceramics, Dekker, New York, 1971.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 747
250. 0. Van der Biest, Analysis of High Temperature Materials, Applied Science Publishers,
London and New York, 1983.
25 1. H. J. Hunger, Werkstqffinalytische Verfahren, Deutscher Verlag fur Grundstoffindustrie,
Leipzig/Stuttgart, 1995.
252. FEPA-ESK, S i c material safety data sheet, according to 9ljl55 EEC and ISO-Standard
11014, Edition 1, issued 22-05-97, 5 pages.
253. D. Henschler (Ed.), Gesundheitsschadliche Arbeitsstoffe, Toxikologisch-arbeitsmedizinische
Begriindung von MAK-Werten, VCH, Wcinheim, 1987.
254. J. Bruch, Institut fur Hygiene und Arbeitsmedizin, Universitatsklinikum Essen, 1993.
255. I. D. Birchall et al., J . Muter. Sci. Lett. 1988, 7, 350-352.
256. ASTM, Standard Practice for Handling S i c whiskers, E 1437-91, Subcommittee E 34.70,
American Society for Testing and Materials, Philadelphia, USA, I99 1.
257. L. Coes, Jr., Abrasives, in Applied Mineralogy, Vol. I , Springer Verlag, Wien, 1971.
258. K. Martin, Neue Erkenntnisse iiber den Werkstoffsabtragsvorgang beim Llppen, Fachber-
ichte Oberflachentechnik 1972, 10, 197-202.
259. G. Spur and I. Sabotka, ZwF 1987,82, 275-380.
260. G. Spur and D. Simpfendorfer, ZwF 1988, 83, 207-212.
261. H. B. Britsch, Ber. Dtsch. Keram. Ges. 1976, 53, 143-149.
262. T. Benecke, S. Venkateswardn, W. D. Schubert, and B. Lux, GiePerei 1993,80, 256-662.
263. S. V. Nair, J. K. Tien, and R. C. Bates, Int. Met. Ref. 1985, 30, 275-290.
264. A. Lipp, Feinwerktechnik 1970,74, 150-154.
265. E. Gugel, Ber. Dtsch. Keram. Ges. 1966, 43, 354359.
266. A. F. Fickel, Sprechsaal 1980, 113, 517-531; 1980, 113, 737-747.
267. R. Rasch and H. Maatz, Maschinenschaden 1978, 51, 145-147.
268. Z. Stavric and M. Hue, Crystar, Keram. Z. 1975, 27, 125-128.
269. G . Bierbauer, Keram. Z . 1972, 24, 142-145.
270. 0. Rubisch and R. Schmitt, Ber. Dtsch. Keram. Ges. 1966, 43, 173-179.
271. E. Buchner and 0. Rubisch, Corrosion behavior of S i c heating elements, in Silicon Carbide -
1973, R. C. Marshall et al. (Eds), Univ. South Carolina Press, Columbia, SC, 1974, pp. 428434.
272. Y. Nakamura and S. Yajima, Am. Ceram. Soc. Bull. 1982, 61, 572-573.
273. W. Huether, Thermoelement zur Temperaturmessung und Verfahren zur Herstellung dessel-
ben, European Patent 72 430, 1982.
274. K. Kuomoto, M. Shimohigashi, S. Takeda, and H. Yanagida, J. Muter. Sci. Lett. 1987, 6,
1453-1 455.
275. R. Davis, Recent advances regarding the definition of the atomic environment, film growth
and microelectronic device development in Sic, in The Physics and Chemistry of’ Carbides,
Nitrides and Borides, R. Freer (Ed.), Kluwer, Dordrecht, 1990, pp. 589-623.
276. M. Srinivasan, The silicon carbide family of structural ceramics, in Structural Ceramics Trea-
-
tise on Materials Science and Technology, Vol. 29, J. B. Wachtman (Ed.), Academic, San
Diego, CA, 1989, pp. 100-159.
277. D. Zeus, How the use of advanced ceramics as tribomaterial has effected the evolution of
mechanical seals, cfi Ber. Dtsch. Keram. Ges. 1991, 68, 3-5.
278. V. D. Krstic, M. D. Vlajic, and R. A. Verrall, SIC ceramics for nuclear applications, in
Advanced Ceramic Materials, H. Mostaghaci (Ed.), Key Eng. Muter. 1996, 122-124, 387-396.
279. F. Porz, G. Grathwohl, and R. Hamminger, J . Nucl. Muter. 1984, 124, 195-214.
280. G. Heuschmann and G. Willmann, Interceram. 1986, 1, 24-29.
281. D. P. Butt, R. E. Tressler, and K. E. Spear, Discontinuous phase formation and selective
attack of S i c materials exposed to low oxygen partial pressure environments, in Corrosion
of Advanced Ceramics, K. G. Nickel (Ed.), Kluwer, Dordrecht, 1994, pp. 153-164.
282. R. A. Penty and J. W. Bjerklie, S i c for high-temperature heat exchangers, in Ceram. Eng. Sci.
Proc. 1982, 3, 120-127.
283. J. Heinrich, J. Huber, S. Foster, and P. Quell, Advanced ceramics as heat exchangers in
domestic and industrial appliances, in High Tech Ceramics, P. Vincenzini (Ed.), Elsevier,
Amsterdam, 1987, pp. 2427-2440.
284. M. C. Kerr, Advanced ceramic heat exchangers utilizing Hexology-SA, single phase S i c tubes,
in High Tech Ceramics, P. Vincenzini (Ed.), Elsevier, Amsterdam, 1987, pp. 2441-2449.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
748 5 Silicon Carbide Based Hard Materials
285. W. Heider, Verwendung von SiSiC im Apparatebau, Proc. Techkeram. '87, WiesbadenlFRG,
S . Schnabel and J. Kriegesmann (Eds), Demat Exposition Managing, 1987, pp. 18.01-18.19.
286. W. Hof, Hexoloy S i c SA ein neuer Werkstoff fur Rohrbiindel-Waremetauscher, Chemi.
-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
6.1 Introduction
Silicon nitride ceramic inaterials have been intensively studied for many years
because of their great potential for use in structural applications at room and
high temperatures. This is due to their excellent mechanical properties in combina-
tion with good corrosion and thermal shock resistance.
Over the past decade a continuous increase in the number of application
fields has been observed. Besides parts for the automotive industry (cam rollers,
valve plates for coninion rail systems). which have not been produced in the
volume that was predicted 10 years ago, applications such as cutting tools. ball
bearings, wear parts. applications at high temperatures, and application in the
electronic industry are becoming increasingly important [1-4]. A new cooking
system based on a silicon nitride plate is penetrating into the market at the
moment [51.
This wide range of applications leads to an increase in the variety of materials
with different microstructures (see Fig. I ), making precise control of the microstruc-
ture and propertics of the materials necessary (Table 1).
The materials differ in the production technology used to make them, their phase
content and properties. The different production technologies used are summarised
in Table 2. A inore detailed overview is given in [ l . 61.
The reaction-bonded materials (RBSN) produced by the nitridation of compacts
of silicon powder (see Table 2) are not dense. whereas the other kinds of materials in
the table are dense.
The range of applications for RBSN and sintered reaction-bonded silicon nitride
has shrunk over the last few years. This is connected with the decreasing cost of
silicon nitride powders. reducing the production costs of SSN and the lower strength
and lifetime of RBSN in comparison to SSN in inany applications. For this reason,
these materials will not be explained in detail in this work; reference is made to
special reviews [ 6 , 7 ] . The following paragraphs concentrate on dense silicon nitride
materials made from silicon nitride powders.
The most common sintering technology used for silicon nitride materials is gas-
pressure sintering. This technology lcads to improved reliability and strength in
comparison to norinal gas-pressure sintering and is accompanied by only a moder-
ate iiicrease in cost. The HIP processes are only used for special applications (e.g.
balls for ball bearings) due to the high cost of these technologies. Hot pressing is
used for evaluation of materials or for some special applications requiring simple
geometries and only ;I very limited number of parts.
Besides the different production technologies, the materials differ in their compo-
sition and microstructure (see Fig. I ) . The wide variation of the microstructure and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Property Microstructure
Material type Reaction Sintered RBSN Sintered silicon Gas pressure Hot-pressed Sinter HIP HIP-ed silicon
bonded silicon (SRBSN) nitride sintered silicon silicon nitride silicon nitride nitride
nitride (RBSN) (SSN) nitride (HPSN) (Sinter-HIPSN) HIP-SN
(GPSN)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.2 Crystal Structure and Properties qf the SijN, Modifications 753
Table 3. Crystal structures of the modification of the Si,N, modification and the resulting materials
names.
Property Modifications
Space group P63:m No. 176 [9, 101 P31c No. 159 [9] Fd3, or Fd3 m
Lattice parameter
a, nm 0.7586 0.766 0.7738
c, nm 0.2902 0.5615
Hardness (1 00) plane: 2 100 (1 10) plane: 2250 20-40 GPa
(001) plane: 1326 (001) plane: 2200
HV 0.025 [12] HV 0.300 [ 1 11
Coefficient of 3.39 3.64
thermal expansion 1131 [I31
0-1000”C, 10-6/K
materials based on nearly all SSN-, cc’-SiAION ?
the modification GPSN-materials, cc’/P’-SiAION
P'-Si AlON
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
754 6 Silicon Nitride Based Hard Materials
SI (OCt)
SI (tetr).
(4
c7 b
a
formed with the addition of BeO/BeN [13]; however, materials made from these
solid solutions have no practical use due to the toxicity of Be.
Analysis of the a-Si3N4crystal structure shows that there is an empty position
+
with a coordination number of eight (7 1). This position can be partially occupied
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.3 DensiJicntion 755
2iR207)
A
/
/
\‘ ‘, 4/3(AIN A1,03)
\
/ ‘
i _- >2(A1203) ,,
/ ’ -
3(Si02)e
\
3/2(Sl*N20) ‘\
‘4
i
I I
S-SlAlON Si3N4 I
SALON-
Ct-SiAION Polytypolds
Figure 3. Stability range of a’- and 0-SiAIONs in the system Y -Si-AI-O-N (see also Section 6 4.2
[ I , 14,151)
by ions with an atomic radius of about 0.1 nm. Additionally, Si and N must be
replaced by A1 and 0 to obtain electroneutrality. The resulting phase is the so-
called a-SiAION with the formula: R,Si12-n-mAlm+nN16-nOn. Occupation of the
+
(7 1)-coordinated position leads to stabilisation of the metastable a-Si3N4
phase. The lowest x value for trivalent cations is 0.33, for Ca 0.3 [14-161. This
means that the a’-SiAlON solid solution does not include the composition of pure
a- Si3N4.
The different kinds of the materials based on the different modifications are given
in Table 3 .
The hardness values of the crystals are given in Table 3. These data show that the
p modifications have much lower hardnesses than the a modifications. N o data for
the chemical stability of the high-pressure modification have been published up to
now. First measurements of the microhardness show that this structure must have
a high hardness which is similar to that of diamond and c-BN.
The microhardness was determined for the different crystal planes of B-Si3N4.The
data differ widely in different orientations.
6.3 Densification
The high energy of the covalent chemical bonds in the Si3N4 crystals, which is
the basis for the excellent properties, presents a disadvantage in the fabrication of
the materials. The self-diffusuion coefficient is very low in comparison to other
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
756 6 Silicon Nitride Based Hard Materials
ceramic materials. The mass transport necessary for sintering can only be achieved
at temperatures where the decomposition of silicon nitride occurs to a great extent
even at high pressure (i.e. >2O0O0C). Dense materials without sintering additives can
only be produced by encapsulation-HIP processes. The green bodies which are thus
formed are sealed in a dense glass container and densified at 1800-1900°C at 2000
bar [l]. The resulting material consists of Si3N4,with some Si02 at the grain bound-
ary. These materials have good oxidation and creep resistance (see Section 6.5.2) but
low fracture toughness and strength (see Section 6.5.1).
Usually the densification of Si3N4 is achieved through the acceleration of mass
transport by the formation of a liquid phase during sintering. In the liquid-phase-
sintering process, silicon nitride can be sintered in a temperature range of 1550-
1900°C. The liquid phase used for sintering is an oxynitride liquid formed by the
reaction of the sintering additives with the Si02 existing on the surface of the
Si3N4powder particles. The most commonly used additives are A1203 in combina-
tion with Y203, La203, lanthanoids or MgO. For high-temperature applications,
pure rare-earth additives are used. AlN additives are necessary for the production
of a’-or P’-SiAlON.
Under sintering conditions stable oxides or nitrides, which form a liquid phase
with Si3N4 and Si02 during sintering, can be used as sintering additives. For
example, alkalis and alkaline earths beside MgO can be used as sintering additives.
The problem in using these elements is their high vapour pressure under sintering
conditions.
Besides Y203,Sc203and La203,the oxides of the d-elements, Zr02 or HfO,, can
be used as sintering aids. Under sintering conditions, Ti02 reduces to TIN; hence
Ti02 can only accelerate the sintering during the initial stage of sintering. The d-
elements in the 51hto gth group form silicides under sintering conditions, forming
separate inclusions.
The amount of liquid phase formed and the resulting grain-boundary phases can
be predicted by the phase diagrams (see Fig. 3). Detailed information about the
phase diagrams are given in [l, 3,13,17].
Densification is improved when the liquid is formed at low temperatures and has
a low viscosity. For this reason, the densification behaviour worsens in the order
MgO/A1203< MgO/R203;MgO; R203/A1203 << R 2 0 3(R = Y, Sc, La and lantha-
noids).
The mechanisms of liquid-phase sintering of ceramic materials are explained in
detail in [19]. Figure 4 shows typical shrinkage curves of Si3N4 materials with
different amounts of liquid phase of the same composition. Densification starts at
1200-1300°C. At these temperatures, the chemical reaction involving the additives
starts and the formation of the liquid is initiated. The first maximum in the
densification rate is connected with the rearrangement of the Si3N4 particles. At
1500-18OO0C, densification by the solution/diffusion/precipitation mechanism
takes places. The phase transition of a-Si3N4 to P-Si3N4 occurs simultaneously.
In [18] it was shown that the a/P-Si3N4 ratio has practically no influence on the
densification when the grain sizes of the a and P phases are the same. This is an
indication that the solubilities of the a- and P-Si3N4 modifications in the liquid
do not differ significantly [20].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.3 DensiJication 151
dLILo, 9
21.0
18.0
15.0
12.0
9.0
6.0
3.0
0.0
1200 1300 1400 1500 1600 1700 6 12 18
Temperature, "C Time, min
dLIdt, %bin
Figure 4. Typical densification curves for Si3N4 materials (1-3.5 vol.% additives, 2-6 vol.%
additives; 3-8.5 vol.% additives; AI2O3/YzO3weight ratio 1 : 2).
The sintering behaviour is improved with increasing additive amount (see Fig. 4).
However, for additive contents higher than 15-20 vol.%, formation of gas bubbles
hinders densification. The bubbling is connected with the low solubility rate of Si3N4
and the increased formation of gaseous SiO [21,22 1.
After sintering, the additives are situated at the grain boundary (3-20 vol.%). The
majority of the grain-boundary phase is concentrated in the triple junctions. The
thickness of the grain boundary films between two grains is in the range of 0.8-
1.5 nm [41].
The grain-boundary phase significantly influences the material properties. There-
fore, one usually attempts to reduce the amount of additives. However, a lower addi-
tive content also leads to a lower sinterability, a lower reliability and a broader
distribution of the properties. Thus for materials for room-temperature applications
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
758 6 Silicon Nitride Based Hard Materials
/\
LT
'f ...........................
...............
1 P
I 'dp
dt
- t
(e. g. wear parts), the additive content usually lies between 6 and 15 vol.%, representing
a compromise between reliability and properties. For high-temperature applications,
special refractory grain boundary compositions are used (see Section 6.5.2).
The most common method used for sintering silicon nitride materials is gas-
pressure sintering. The principle of gas-pressure sintering is the same as that of
the sinter-HIP process (see Fig. 5). The material is sintered to closed porosity
( > 95% theoretical density) at a low gas pressure. Then an outer gas pressure of
up to 10 MPa (GPS: gas-pressure sintering) or 100-200 MPa (sinter-HIP) is applied.
The difference between the outer pressure and the pressure in the closed pores is an
additional driving force for densification, leading to a better healing of defects. The
increase in cost for the GPS process in comparison to sintering under 1 atm (SSN) is
relatively low. In many cases these costs can be offset by the better reliability and
properties of the materials.
the starting powder must be in the range >0.5-1%. This is the case for commercially
available a-Si3N4 powders. In [39], it was found that at normal heating rates
the nitrogen content in the oxynitride liquid during heating is lower than at low
heating rates and low additive contents. This means that for normal heating rates,
the liquid phase is undersaturated with respect to (j-Si3N4and no nucleation can
take place.
These experimental data suggest that (j-Si3N4nucleation has no influence on the
microstructure of silicon nitride materials under common sintering conditions and
using common powders.
The consequence of this is that the different microstructures are a result of the
different amounts and sizes of pre-existing p nuclei as well as the growing conditions.
The amount of growing (j-Si3N4nuclei depends on the (j content in the starting
powder, the crystallite size and the sintering conditions. In Fig. 9 the dependence
of the equilibrium concentration in the oxynitride liquid during sintering on the
:I
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.4 Microstructurul Development 76 1
0
, ;J:,
vl
53 /
c /+
10
__.--
0 7
1200 1300 1400 1500 1600 1700
Temperature, "C
100 j
3
E
C
'1
20 i
1200 1300 1400 1500 1600 1700
Temperature, "C
Figure 7. Change in crystallite size and phase content as a function of sintering temperature,
amount of sintering additives and starting powder 1. diimid powder 5% Y2O,/Al2O3; 2. gas
phase synthesised powder 5% Y2O3/Al2O3;3. gas phase synthesised powder 11.6% Y2O3/AI2O3.
crystallite size of a- and P-Si3N4is shown. Nuclei with a crystallite size lower than a
critical value have a higher equilibrium concentration than the Si3N4concentration
in the liquid. These crystals dissolve in the liquid and cannot growth. During the a/P
transformation under isothermal conditions, the critical size of P-Si3N4 is deter-
mined by the crystallite size of a-Si3N4(r(crtl), see Fig. 9). During heating, espe-
cially for high heating rates, the oxynitride liquid is undersaturated (see Fig. 9,
lines 2 and 3).
As a result of the undersaturation, the critical crystallite size becomes infinitely
large and only the largest nuclei survive because they require more time to dissolve
than smaller nuclei. The resulting microstructure of materials heated at high heating
rates in the temperature range of 1200-1600°C is coarser than for materials heated
at low heating rates [20,29,39].
After the alp transformation, the critical crystallite size is determined by the
crystallite size distribution. The critical crystallite size increases with increasing
sintering time and temperature, i.e. only the largest crystallites which were present
in the starting powder determine the microstructure.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
162 6 Silicon Nitride Based Hard Materials
The needle-like grain shape seems to be kinetically determined and therefore not
the equilibrium crystal shape [30,38]. The estimated equilibrium aspect ratio (ratio
of grain length to thickness) is 1.3, not as high as 20, as was found to be the case for
the silicon nitride materials. The microstructural formation must therefore be
explained on the basis of growth mechanisms.
The explanations for the microstructure given in the ’80s were based on the
assumption that needle-like grain growth can take place only during the ct/p trans-
formation. Later it could be shown that anisotropic grain growth could also take
place after the a l p conversion [31,32,42] (see Fig. 10).
Figure 10. Microstructure of a material hot pressed at 1800°C (starting powder: fine P-Si3N4)and
the same material after additionally heat treatment at 1900°C. It is observed that anisotropic grain
growth takes place after a/P conversion.
Anisotropic grain growth occurs due to the different growth mechanisms in the
different directions of the B-Si3N4 crystals. The (001) basal planes of the needle-
like (j-Si3N4 crystals are atomically rough; therefore, their growth is diffusion-
controlled [4244]. Transport occurs by two mechanisms: diffusion through the oxy-
nitride liquid and surface diffusion in the (100) planes [44]. The (100) prismatic
planes are atomically flat and therefore the growth is interface-controlled. The
growth rate of the (001) plane has been assigned values of 30 [38] and greater
than 27 [35]; thus, it is at least ten [42] times higher than the growth rate of the
(100) plane. The first value seems to be the more exact one because it was determined
by growth observation on crystals in oxynitride glasses. The different growth
mechanisms are supported by TEM observations [43,44], theoretical calculations
[38] and by the observed crystal habits [4547]. The relationship between growth
rate and oversaturation is shown schematically in Fig. 11. For a high degree of
oversaturation, the (1 00) planes become atomically rougher, causing the growth
rates to approach the rate of the other planes, the growth to be isotropic and the
grain morphology to be equiaxed. At the lower saturations found in common sinter-
ing conditions, extended anisotropic grain growth occurs. The boundary between
the two regimes shifts to lower oversaturations with increasing temperature.
For a given crystal, the oversaturation depends on the size and is different for the
different crystal planes. Using the differences in surface energy calculated by the
broken bond method [38], the oversaturation as a function of grain thickness is
given in Fig. 12. The equilibrium concentration above the basal planes depends
only on the thickness of the grains, whereas the equilibrium concentration above
the (100) planes depends on both the thickness and the length of the grains.
Using these data and the dependencies of the growth rate on the oversaturation,
it can be shown that the largest crystals with the highest aspect ratios grow
anisotropically, whereas the small crystals with a low aspect ratio dissolve. Thin
crystals with a high aspect ratio grow stably whereas crystals of the same thickness
but a low aspect ratio dissolve. This behaviour was also determined experimentally
[30,48].
At first these results and the growth mechanisms discussed seem to be in contra-
diction to the results of the works from the 1980s showing a reduction in the aspect
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
164 6 Silicon Nitride Based Hard Materials
region of anisotropic
grain growth - * .-.-
grain growth //
//
/
.-.-
region of isotropic,/'
,
/
/
A
0
c ,
e
$e I !
i
1
i
/interface controlled
Oversaturation
o(1OOyD; o(100YD + o(oo1)n
Figure 11. Growth rate of the basal plane (diffusion-controlled) and the prismatic planes (interface-
controlled) as a function of oversaturation in the oxinitride liquid. The oversaturation after the alp-
transformation depends on the surface energy (o), the grain thickness (2D) and the length of the
needle like grains (2L).
1.1:
(C(100) A-I
1.1
Y
0
1.05
Grain thickness, pm
Figure 12. Calculated dependence of the oversaturation of the basal (001) plane and the prismatic
(100) planes on the thickness of the grains, according to [30] using data [38,63] (A = aspect ratio).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.4 Microstructural Development 765
ratio with increasing p-Si3N4content and at longer soaking times. These differences
can be explained by taking the crystallite size of the P-Si3N4in the starting powder
into consideration. The powders used for these investigations were produced by
direct nitridation of silicon. The higher p content of these powders was a result of
the faster nitridation and larger crystallite size. The larger crystallite size has two
consequences: 1. the oversaturation, which is inversely proportional to the grain
size, decreases and consequently the grain growth rate also decreases; and 2. the
thicker crystallites must undergo more growth to become needle-like. Both pro-
cesses lead to a reduced aspect ratio. Additionally, the powders had a wide grain
size distribution. With increasing sintering time, the critical crystallite size (see
Fig. 9) increases and only the largest crystals survive. However, the largest
crystallites had lower aspect ratios. The aspect ratio is the result of two competing
processes: anisotropic grain growth (preferred growth in the direction of the needle),
and the dissolution of small elongated grains. When the second process is the
dominating process, the overall aspect ratio decreases.
An additional effect influencing the different microstructures which are formed is
connected with the different grain sizes of the materials. The oversaturation depends
on the mean grain size. With increasing grain size, the curve characterising the
dependence of the oversaturation on grain size is flatter (see Fig. 9); for a given
mean grain size, the oversaturation cannot compensate for the difference in surface
energy between the (001) and (100) planes. As a result, anisotropic grain growth
cannot take place and the grain shape approaches the more equiaxed equilibrium
grain shape. Based on the investigations of microstructural formation, it can be
concluded that these processes are only significant for mean grain thicknesses of
0.7-1.5 pm or larger. The materials investigated in the 1970s and early 1980s were
characterised by mean grain thicknesses of 0.7-2 pm, whereas today’s materials
are based on finer powders having normally finer grain sizes.
Recent investigations [49-521 show that needle-like microstructures can be
achieved from p-Si3N4 powders when the crystallite size of the starting powder is
small enough or the sintering temperature and soaking time are large enough.
Materials with high fracture toughnesses of up to 10-12 MPam1/2[48] are observed.
Low-cost powders with higher P-Si3N4 contents, produced by special nitridation
processes which allow the exothermic nitridation to be controlled [53], have
relatively small crystallite sizes. Materials with strengths higher than 800 MPa and
fracture toughnesses similar to materials based on a-Si3N4powders can be produced
from these powders [49].
At high sintering temperatures, abnormal grain growth can take place [35,54,55].
The reason for this is the presence of large p-Si3N4 nuclei (seeds) which can exist
in the starting powder, formed during densification [35,56,62] or added as seeds
[57-591. The formation of large seeds during sintering can be intensified by low
heating rates or dwell times at temperatures higher than 1500°C but lower than
the temperature at which rapid densification occurs [56]. The amount of additional
seeds needed to produce a reinforced material ranges from 1-5 or 10 vol.%. The
seeds are usually between 10 and 100 times larger than the nuclei in the starting
powder [56-58, 60, 611. The addition of larger seeds is intended to increase the
fracture toughness without drastic decreasing the strength (see Section 6.5.1).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
166 6 Silicon Nitride Based Hard Materidv
mean
crystallite size
in the
starting
powder (r,,)
1 10 50 100
4/3 (AINxA1203)
4/3 (AINxA1203)
area of elongated
grain growth
R203 x 9A1N
7
b)
Figure 14. Stability area of the a’-SiAlON solid solution R,Si,,_,~_,A1,+,N16_nO~ a) in the systems
with R = Nd, Sm, Dy, Y and Yb at 1800-1900°C (after [14]) and b) in the area where elongated grain
growth takes place in the a’-SiAlON/P’-SiAlON-plane(after [14,15,77,79]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.4 Microstructural Development 769
[73,74]. When the a’-SiAlON formation is completed, no more liquid exists for
densification. This is the reason why pure a’-SiAlON materials with low degrees
of substitution n and m are difficult to densify to full density. Better densification
can be realised at higher n and m values (Fig. 14) [75-771 or in composite dip’-
SiALON materials [78].
The resultant materials had microstructures consisting of equiaxed grains, lead-
ing to reduced fracture toughness and strength. This was a reason for the limited
use of these materials [I]. In the past few years materials with elongated grains
have been produced by different research groups. The materials with elongated
a-SiAION-grains have relatively high n and m values (greater than 1.2) (Fig. 14)
or mixed cations (rare earth’s and Sr or Ca [14,75]). In early studies it was sug-
gested that elongated grain growth occurs only when p-Si3N4 powder is used [76].
However, later investigations have shown that the use of a-Si3N4 powder also
leads to elongated grain growth [75,14, 15,791. The reason for the anisotropic
grain growth of a’-SiAlON is not completely clear because it was not investigated
in the same manner as the p-Si3N4 grain growth. The a’-SiAlON materials with
low n and m values and equiaxed grains possess a low grain size. This is the
consequence of the fast disappearance of the liquid. In contrary, the microstruc-
tures of the a’-SiAlON materials with elongated grains show a coarser micro-
structure; i.e. more intensive grain growth takes place. This must be connected
with the amount of liquid existing during sintering, which is higher when
higher substitution levels of the d-SiAIONs are used. Recently it could be
shown that materials with elongated a’-SiAlON and low n and m values can
be produced by gas pressure sintering using an appropriate liquid phase (see
Fig. 15) [80]. This suggests, that the needle like grain shape is governed by the
grain growth kinetics.
The stability of a’-SiAlON at elevated temperatures (1 100-1 500°C) depends on
the nature of the cations R entering the a’-SiAION structure. For the trivalent
rare-earth cations (i.e. La-Yb; Y), the stability increases with decreasing atomic
radius [14,81]; i.e. the temperature below which a phase transformation a’ -+ p’
can thermodynamically take place decreases. The stability also increases with
increasing substitution level. Thermodynamic stabilisation of the a’-SiAlON
with respect to the p’-SiAION can be achieved by the use of mixed cations [14].
Besides the nature of the cation which is incorporated in the a’-SiAlON, the rate
of a’-$-transformation depends on the existence of $-SiAlON nuclei and the
amount and viscosity of the amorphous grain-boundary phase. The transformation
between a’-SiAlON and D’-SiAlON is reconstructive and must occur via a liquid in
which one phase can dissolve and the product precipitate. Therefore the transforma-
tion can be retarded by the absence of a glassy phase or by a very high viscosity
[14,81].
Better understanding of the microstructural formation of the a’-SiAlON
materials in the future will be necessary for the improvement of the mechanical
properties, especially fracture toughness and strength, and the wider application
of this group of materials, having higher hardnesses than the B-Si3N4 and p-
SiAlON materials (see Fig. 16).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 15. Polished and etched micrograph of an a’-SiAION material with n,m < 1; a) plasma-
chemically etched material (dark area: P-Si3N4;grey area: d-SiAION; bright area: grain-boundary
6 Silicon Nitride Based Hard Materials
110 HV10
- 1000
cl
iE
3
Ei
-900 $
,
''
,
' 5
-3
E
-800
t20 to
c
3 .
2
5
-700
--2"
I
m
;a
.-
II
- 600
- 500
T - 400
B-Si3N4 - Conventional High toughness
materials a '-SiAlON a '-SiAION
materials materials
Figure 16. Comparison of the properties of the conventional a'-SiAION materials with P-Si3N4
materials and with high toughness a'-SiAION materials with elongated grains [80].
[871
[551 seeds oriented
" .0
5 10 1
to have a grain size lower than that of the defect, because larger grains can act as
strength-determining defects. This means that high-strength materials must be
fine-grained. On the other hand, the fracture toughness increases with the square
root of the grain size [36,69,87,88]. These two adverse dependencies produce the
relationship between strength and fracture toughness shown in Fig. 17. For low
fracture toughness values, flaws other than the largest grains are the strength-
limiting defects. In this region of KIC, further improvement of the strength is
possible. For high KIc values, the strength is determined by large elongated
grains. Thus the structure itself determines the strength, not processing-dependent
defects. Materials with large, needle-like grains in the matrix are usually called
insitu-reinforced materials. Attempts to improve the strength of such materials by
seeding a matrix with large, needle-like grains with a narrow size distribution are
underway [57,59,60]. Results for tape-cast silicon nitride materials with large,
oriented P-Si3N4needles represent a possible approach [60] (Fig. 17) to increase
the strenght and fracture toughness. However, such materials exhibit anisotropic
properties, e.g. low strength in the direction of orientation of the needles.
Additionally, the insitu-reinforced silicon nitride materials exhibit an increased
Weibull modulus [86]. This has two reasons: firstly, the concentration of the
large, strength-limiting grains is so high that the defect size has a narrow size distri-
bution; and secondly, these materials exhibit pronounced R-curve behaviour, which
makes the material tolerant of larger cracks. The high thermal shock resistance of
these materials also seems to be connected with this fact [30].
The subcritical crack growth rate in silicon nitride materials is relatively low in
comparison to other ceramic materials. The growth exponent is in the range of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties of Si3N, Materials 773
N = 30-300 for static loads and n = 20-50 for cyclic loading [90,91,97-1001. Up to
now, no clear correlation between the microstructure and the growth exponent
could be made. For cyclic loading it can be shown that materials with the lower
grain size have the higher growth exponent, i.e. have a lower degradation of strength
during loading and a higher lifetime at a given strength level [91,88].
The fracture toughness depends strongly on the microstructure. Two main factors
influence the fracture toughness: 1. the grain shape and size, and 2. the composition
of the grain-boundary phase.
The high fracture toughness of silicon nitride materials in comparison to other
ceramic materials is connected with toughening mechanisms which are similar to
those in whisker-reinforced composite materials: grain bridging and pull-out,
crack deflection and grain branching around large, elongated grains [88,89,94].
Due to these mechanisms, the fracture toughness increases with increasing
volume fraction and square root of the mean grain thickness of the elongated
grains (i.e. grains with aspect ratio >4) [37,70,88,89].
Recent investigations show that fine-grained materials have a higher toughness
than coarse-grained materials in the small crack region but lower toughness in the
large crack region [82,84,90]. This can be important for the use of such materials
under conditions in which high local stresses exist (e.g. for ball bearings). The
dominant toughening mechanism depends on the grain thickness of the elongated
grains. Elastic bridging and pull-out were observed for thin, needle-like grains
(with thickness < 1 pm). Crack deflection was mainly observed for thick, elongated
grains (with thickness > 1 pm), whereas frictional bridging was detected indepen-
dently of the grain size [89].
The toughening mechanisms described above can only operate when the domi-
nant fracture mode is intergranular. The ratio of transgranular to intergranular frac-
ture depends on the relative strengths of the grain boundary and the grains. For a
material with a high toughness, the grain boundary must be weak in comparison to
the grains (see Fig. 18).
The strength of the grain boundary is connected with two different mechanisms:
local residual stresses [92,101] and special chemical interactions between the grain-
boundary phase and the silicon nitride grains [93,94].
The amorphous or partially crystallised grain-boundary phases generally have
different thermal expansion coefficients than silicon nitride. When the thermal
expansion coefficient of the grain boundary is higher than that of the silicon nitride,
the grain-boundary phase is under tensile stress and the fraction of intergranular
fracture increases. As a consequence the fracture toughness increases (see Fig. 19).
Similarly, a material with a grain-boundary phase under compression (e.g. HIPSN
without sintering additives) shows a low fracture toughness [ 1021 due to a high per-
centage of transgranular fracture (see Fig. 18a).
The residual stresses can be influenced by the sintering cycle. This can be brought
on by a change in the composition of the grain boundary (e.g. by evaporation of
S O 2 ) during sintering, by crystallisation of the grain boundary or by partial relaxa-
tion of the stresses (e.g. through slow cooling). These changes are usually out-
weighed by the grain size and shape or the starting composition of the material,
which have a greater influence on the stress state and KIc.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
114 6 Silicon Nitride Based Hard Materials
Figure 18a and b. Crack path in silicon nitride materials a) HIPed material (only S O 2 sintering
additive, mainly transgranular fracture; chemical etched), b) material with Y20,/A120,additives
(mainly intergranular fracture, plasma etched).
Figure 19 shows the additional influence of the grain size (see curves 1-3). The
composition of the grains also influences the fracture toughness by determining
the special chemical interactions that occur between the grains and the grain
boundaries. The formation of P-SiAlON layers on the Si3N4 grains was found to
result in an increase in the amount of transgranular fracture and a decrease in the
fracture toughness [93, 1001. This can be a reason for the lower fracture toughness
of the P-SiAlONs with a high degree of substitution, in addition to the lower
aspect ratio (see Fig. 18d).
The recently developed a-SiAlON materials have a microstructure consisting of
relatively coarse, elongated grains (see Fig. 15) and show a high percentage of inter-
granular fracture; thus, these materials exhibit a high fracture toughness (see Fig. 16).
Summarising the fracture toughness data, it can be concluded that the fracture
toughness of the silicon nitride materials depends strongly on the microstructure.
For high fracture toughness values, elongated grains and a grain-boundary phase
which promotes intergranular fracture are necessary.
The hardness of different silicon nitride materials is given in Fig. 20. Materials
with a high a-Si3N4 content, i.e. materials with a high amount of metastable a-
Si3N4which did not transform during sintering (normally produced by hot pressing
or HIP’ing [l, 59]), show a hardness which is as high as that of the d-SiAlON
materials.
Analysis of the hardness values of different P-Si3N4 materials reveals that the
hardness increases with decreasing amount of grain-boundary phase and decreasing
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties q f S i 3 N 4 Mureriuls 775
Figure 18c and d. Fracture surfaces of silicon nitride materials c) fracture surface of material with
Y2O7/AI2O1additives (mainly intergranular fracture), d) fracture surface of a P’-SiAION material
with a high degree of substitution z (z = 4) (mainly transgranular fracture).
grain size; i.e. the dependence of hardness on the microstructure is opposite to that
of the fracture toughness.
Usually the hardness values of b-Si3N4 materials (HVlO) are in the range of
12 GPa (coarse grain size, high additive content, residual porosity) to 16 GPa (fine
grain size, low additive content). The hardness of the mixed a/b-SiAlONs change
lineary with the phase content [ 1311.
The hardness at elevated temperatures depends additionally on the softening of
the glassy grain-boundary phase. Materials with MgO/Al2O3as sintering additives
show a faster degradation of the hardness than materials with more refractory grain-
boundary phases (Fig. 20).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
716 6 Silicon Nitride Based Hard Materials
Glass
Si02
i
ain boundary
Silicon nitride
---___
----_--__ /
SiO2-grain
boundary phase
Si3N4
2J
I
0 2 4 6 8
Thermal expansion coefficient, 10-6/K
Figure 19. Residual stresses of grain-boundary phase and the silicon nitride grains as a function of the
thermal expansion coefficient of the grain-boundary phase (calculated after [92]); used constants for
Si3N4:E = 320, v = 0.27, s( = 3.39 x K; and for the glass: E = 140, v = 0.29, volume fraction
of glass = 0.15; a) and dependence of the fracture toughness on the thermal expansion coefficients
(data 1-3 after [92]) 1. mean thickness of the grains 0.55; 2. 1-1.3; 3. 1.5-1.7 (data 4 after [64]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties of Si,Nd Materials 777
25
20
15
10
5
500 loo0 1500
Temperature, "C
Figure 20. Hardness of different Si3N4materials; 1 . additive free material with remaining a-Si3N4
[59], 2. hot pressed with 10% Ce02 and 6.6% AIN contain 20% a-SiAION [84]; 3. sintered p-silicon
nitride material with 3% Y 2 0 3and 2 % Al2O3;4. d-SiAlON.
Figure 21. SEM images of polished cross sections of the Si3N4material 1 (Table 1) after 2500 h of
oxidation treatment at 1500°C.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties of' Si3N4 Muteriuls 779
process and the ability of this layer to prevent oxygen from diffusing into the
material. In this case the formation of a surface layer of pure silica causes the rate
of oxygen diffusion into the material to be very low [103, 1041.
A more refractory grain boundary can only be obtained by removing the silica
from the grain boundaries and triple junctions. Each addition of only a small
amount of impurities or sintering additives will weaken the material by producing
a change in the chemistry of the grain-boundary phase [105-1071. The result is
degradation of the high-temperature behaviour of the materials. Besides the
costly fabrication, the main disadvantage of these material is their inadequate
mechanical properties (especially toughness) at room temperature. The microstruc-
ture, which is mainly globular, and the strong bonding of the intergranular phase do
not allow toughening mechanisms like crack deflection or bridging to be active.
With the addition of sintering aids these disadvantageous properties can be
improved; however, the improved densification, the microstructure with elongated
grains, and the improved toughness are the consequence of a changed grain-boundary
phase. Depending on the kind of sintering additive used, a silicate phase will be
formed during sintering and will remain as an amorphous or partially crystallized
phase in the grain boundaries and triple junctions between the Si3N4grains. Besides
the chemical composition and the amount, special emphasis is placed on the crystal-
linity of the intergranular silicate phase. With a high amount of crystallized silicates in
the grain boundary, the amorphous phase is minimized; this is found to be especially
beneficial to the creep and slow crack growth behaviour of the Si3N4 materials.
Glass-forming and stabilizing sintering additives which, with the silica of the Si3N4
powder, form a silicate phase with a low softening point and viscosity such as MgO,
AlzOi or AlN are principally not useful in the fabrication of high-temperature silicon
nitride materials. More refractory silicates that form an intergranular phase with a
high crystalline content were obtained by using yttria or the rare-earth oxides as sin-
tering additives. Examples for these materials are given in Table 4. When A13+ions are
present, the softening point and viscosity of the grain boundary are lowered [ l , 31,
with the consequence of degradation of the mechanical properties at elevated tem-
peratures. Figure 22 summarizes the creep curves of the alumina-containing silicon
nitride, material 3 (Table 4), in comparison with the additive-free material and the
material with yttria as a sintering aid. The HIP'ed material exhibits the highest
creep resistance. Superior creep behaviour is also found for Si3N4with Y203 as a sin-
tering additive. This is the consequence of the high degree of crystallized grain-bound-
ary phase having the apatite structure. The creep resistance of the Si3N4materials is
found to be lowered by an amount which depends on the amount of A13+ions present.
With the addition of a small amount, the crystalline grain-boundary phase does not
change significantly; however, a degradation of the creep resistance is observed,
caused by the concentration of A13+ ions in the residual amorphous grain-boundary
phase. In the material with 5% Y 2 0 3and 3% A1203,the grain-boundary phase is
found to be mainly amorphous, with a lowered softening point and viscosity. The
result is a very poor high-temperature resistance: the material failed after only a
few minutes during creep testing at 1400°C and 200 MPa.
A similar tendency is found by comparison of the oxidation behaviour of the
materials. With the addition of the sintering aid the composition of the protective
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
780 6 Silicon Nitride Based Hard Materials
0,Ol
0,008
0,006
0,004
1/
0,002
0
0 20 40 60 80 100
Figure 22. Comparison of the creep behaviour of the materials (1 in Table 4) HIP’ed Si3N4without
additives, (3 in Table 4) Si3N4with 8% Y203/o.6%A1203and (5 in Table 4) Si3N4with 8% Y203as
sintering additives. The tests were conducted at 1400°C in air at 200 MPa.
oxidation layer formed during oxidation changes, with the consequence of a higher
oxygen diffusion into the materials. The weight gain of the Si3N4with Y2O3 is about
three times higher than that of the HIP’ed Si3N4without sintering aids, and that of
the material with the small amount A1203about one order of magnitude higher than
that of the HIP’ed Si3N4without sintering aids. Figure 23 shows the microstructure
+
of the materials 3 (8% Y2O3 0.6% A1203) and 5 (8% Y203)after oxidation at
1500°C. Oxidation processes are observed in the bulk of both materials and lead
finally to alteration of the microstructure. However, due to the addition of A1203,
the oxidation of material 3 (in Table 4) occurs faster. Already after 1000 h oxidation
the microstructure is found to be strongly degraded. The same damaging processes
occur with the Y203-containingSi3N4material; however, this material exhibits, due
to its more refractory grain-boundary phase, a higher oxidation resistance (see
Figure 23. Microstructural damage in the surface region of the Si,N4 materials with Y2O3/Al2O3
(3 in Table 4) after 1000 h (A) and Y 2 0 3(5 in Table 4) after 2500 h (B) oxidation at 1500°C.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties of Si3N4 Materia1.r 78 I
microstructure in Fig. 23 after 2500 h). As a result of the oxidation processes, degra-
dation of the mechanical properties is observed; this is found to be dependent on the
extent of oxidation as represented by the residual bending strength after oxidation
(see Table 4).
Interesting materials with respect to high-temperature behaviour are mixed a’/p’
SiAlON [3,78]. Due to the ability of the a’-SiAlON grains to incorporate cations
from the sintering aids into their lattice, it is possible to modify the grain boundary
of the silicon nitride material. In a material with a high amount of a’-SiAlON, a
skeleton from strong grain boundaries between the a’-SiAlON is formed due to
the incorporation of the ions of the sintering aids into these grains, with the
consequence of an improved creep resistance. The oxidation resistance cannot be
improved by the increase in the d-SiAlON content. The high amount of A13’
ions in the material is the reason for the formation of a weak surface layer of alumo-
silicates during oxidation, allowing a high rate of diffusion of oxygen into the
material 1781.
One goal in recent studies in the field of high-temperature Si3N4materials was the
improvement of the behaviour at elevated temperatures by the formation of compo-
sites. Typical examples which have been reported in the literature are Si3N4-Sic
micro- or nanocomposites or Si3N4-MoSi2 composite materials. In the case of
Si3N4-SiC nanocomposites, a structural synergism between the matrix Si3N4and
the nano-Sic was supposed to be responsible for the high level of the properties
of these composites at elevated temperatures [ 1091. Recent studies, however,
showed that the improvement of the mechanical properties at high temperatures
was the consequence of a chemical modification of the composition of the grain-
boundary phase and a significant improvement of the oxidation resistance
[104, 108,110,ll I]. Similar behaviour was found for Si3N4-MoSi2 composite
materials [112]. After long-term oxidation tests at 15OO0C, a considerably less
damaged microstructure was observed. Figure 24 shows the comparison of the
material 5 (Si3N4 with 8 % Y203) with a Si3N4-MoSi2 composite (material 7).
Apart from the 10% MoSi2 in the composite, both materials had the same composi-
tion. The microstructural development of the materials was found to be different. A
significantly lower amount of damage was observed in the Si3N4-MoSi2 composite,
although the oxidation processes also occurred in the bulk of the composite.
Figure 24. Comparison of the microstructure of (A) the Si3N4material ( 5 in Table 4) and (B) the
Si3N4-MoSi2 composite (7 in Table 4) after 5000 h oxidation at 1500°C in air.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
182 6 Silicon Nitride Based Hard Materials
1-1
bending stress / MPa
900
i
700
500
300
I00
time I h
Figure 25. Comparison of time to failure behaviour of silicon based nonoxiode materials at 1400°C
in bending (curve 1: material 7; curve 2: material 6; curve 3: material 5; curve 4: SBC).
This behaviour was the consequence of different processes in the surface region
of the materials, resulting in a changed oxidation mechanism in the composites.
The important feature of the Si3N4--MoSi2and Si;N,-SiC composite was the
increased formation of Si20N2 instead of SO2, producing an additional layer
between the oxidation surface and the bulk of the silicon nitride composite materials.
The significantly reduced defect size gives an idea of the potential of these
composite materials in terms of stability and time-to-failure behaviour at elevated
temperatures (Fig. 25). Principally, the amount, composition and condition of the
grain-boundary phase are the key factors which must be considered for the suc-
cessful development of Si3N4materials for applications at elevated temperatures.
Materials without sintering additives exhibit the highest stability at elevated
temperatures due to the high refractory grain-boundary phase of pure SO2. The
addition of sintering aids ultimately leads to a degradation of the properties at
elevated temperatures. However, by a purposeful design of the grain-boundary
phase and the formation of high temperature stable composites which are stable
at high temperatures, it is possible to obtain silicon nitride materials which can be
used up to temperatures of 1500°C.
Figure 26. Typical cracks on the track of ball bearings (dry running under 1.75GPa Hertzian
pressure. 18 turns).
Silicon nitride is the preferred ceramic material for use in ceramic and hybrid ball
bearings due to its high strength, fracture toughness and high resistance to sub-
critical crack growth. The advantages of ceramic ball bearings in comparison to
metal ball bearings are: reduced wear, higher hardness, better temperature and
corrosion resistance, lower specific weight and higher electrical resistance. This
combination of properties makes lubrication of such bearings either almost or
completely unnecessary [2, 113, 1 15, 1171. The field of application of such bearings
is fast growing. The bearings show a good performance and a high reliability. To
improve the behaviour a development of silicon nitride materials for ball bearing
was started in the last few years [ I 13-1 16,1181. The main failure mechanism of
ceramic ball bearings are connected with local and temporal changing high tensile
and compressive stresses during operation. These local and temporally changing
stresses result in the formation of cracks perpendicular to the rolling direction of
the balls in the ball bearing (see Fig. 26) [I 141.
The lifetime and load capacity of a ball bearing can be increased by optimisation
of the geometry of the track [ 1 13-1 15,1181, by reduction of the material damage due
to cyclic Hertzian contact, and by reduction of the friction coefficient [I 14,1161.
The influence of the microstructure on the Hertzian contact damage was investi-
gated in detail [119, 1201. Coarse-grained silicon nitride materials are subjected to an
accelerated fatigue under cyclic Hertzian pressure, unlike fine-grained materials.
Therefore materials for ball bearings have to be fine-grained. Apart from that, the
materials must exhibit high strength (i.e. >800 MPa) and a high Weibull modulus
because every defect on the track can be a starting point for failure.
The influence of the friction coefficient on the load capacity is connected with the
fact that, besides rolling, sliding always exists in the ball bearing [114, 1151. This
friction of the ball on the track is the reason for an increase in the local tensile
stress in the track during operating. Therefore reduced friction will increase the
load capacity and the lifetime. The load capacity of ball bearings operating in wet
conditions is more than 2 times higher than in dry conditions. This is due to the
lubricating effect of the water [I 151. The development of Si3N4 composites with
small amounts of submicro-TiN or BN inclusions leads to an increase in the load
capacity of up to 60% under dry run (see Fig. 27) [115]. Further improvements
can be expected through surface modification by ion implantation [I 171, the use
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
184 6 Silicon Nitride Based Hard Materials
Friction Figures p
0.02
Idry run in air
water lubrication
0,015
0,Ol
0.005
0
Standard
Standard Si3NUBN Si3N3/TN Silicon nitride
Silicon nitride (grease lubriicated )
Figure 27. Friction figure (maximal value) of angulare contact bearings 7006 E under dry running in
air and in water at 500N preload [115].
Figure 28. Surface of silicon nitride materials after oscillating sliding tests.
high strength, hardness and fracture toughness. These properties are necessary but
not sufficient for a good cutting behaviour: there are several Si3N4materials which
fulfil these conditions, but which have a bad cutting behaviour due to enhanced
chemical interaction with the workpiece.
A wide range of substances in different ratios are used as sintering additives for
cutting tools: CaO, MgO, A1203,Y 2 0 3 ,Zr02 and Hf02 [2,171,125-1311. The use
of sintering additives that reduce the high-temperature strength is surprising, but is
less so when one considers that the softening of the glassy grain-boundary phase
starts at temperatures above 800”C, whereas the temperature at the cutting edge
in the machining of cast iron is lower than 800°C. An increased Si02 content in
the grain-boundary phase reduces the lifetime of the cutting tools [133]. This
seems to be connected with the acceleration of the chemical interaction.
A decrease in the amount of sintering additives improves the cutting behaviour
[125]. The problem of optimising the cutting tool material is that on the one hand
it is necessary to minimise the amount of the grain-boundary phase and on the
other hand it is necessary to attain a reliable densification to achieve full density,
high strength and high reliability. One possibility to improve the densification of
the materials with a low additive content is to use MgO-containing additives
[71,129]. Another way to improve the sintering behaviour without increasing the
amount of the residual grain-boundary phase is to form p’- or P’/a’-SiAlONs
[130]. The substitution grade z in the P-SiALON must be small because high z
values cause the fracture toughness and strength to be reduced. In [ 1321 an optimum
for Si6-,Al,0,N8_, is found with z = 1-1.5, but most of the commercial cutting
tools have an lower substitution degree. p’/a’-SiAlON composite materials are
used especially for cutting superalloys. They have a slightly reduced strength and
fracture toughness [128, 1311.
A low grain size is favoured for cutting tools because the hardness increases with
decreasing grain size. Nevertheless excellent cutting performance is found for self-
reinforced Si3N4materials with large, elongated grains [7 11.
The examples given above show that it is necessary to find a compromise between
the different material properties. The compromise strongly depends on technologi-
cal possibilities (e. g. improvement of the sintering behaviour of the powder) and on
changes in the metal machining. The chemical stability of the cutting tools can be
increased through the use of coatings, e. g. A1203/TiN[129].
Table 5. Classification of corrosion conditions and main features of the corrosion behaviour
Figure 29. a) SEM micrograph of an corroded in H2S04 100 h 90°C material with Y203/A1203
additives and b) SEM micrograph of an material with optimized grain boundary corroded in
H2S04 500 h 90°C material (plasmaetched).
Recent investigations show that, besides the amount of the additives, the main
parameter governing the corrosion resistance of the silicon nitride materials is the
Si02 content in the grain-boundary phase (see Figures 29-31). A material with a
high S O z content in the grain-boundary phase is more stable in acids than one
with a low Si02 content. This correlates with the stability of the grain-boundary
phase. The acid corrosion of glasses shows a similar behaviour [147,155]. The net-
work modifier is leached from the glass network by the acids and the glass network
hydrated. Therefore a network with a low amount of network modifiers has a
higher stability [ 1471. Additional to the factors influencing the corrosion of glass,
the corrosion of silicon nitride is affected by the fact that corrosion takes place
only in small channels between the Si3N4 grains. During corrosion the Si3N4
grains also produce hydrated Si02, which can reduce the corrosion rate and
change the corrosion mechanism. This seems to be one of the reasons for the
low corrosion rate of Si3N4 materials with a low amount of sintering additives
[141].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6.5 Properties of Si,N4 Muterials 789
P
I OY 203/A1203- 1
I
m
. 5Mg0
10 MgO/Y203/A1203
*- *-
<
N
1
5
;;4
E
-
v1
0
- -- 3,2 Vol.%
Time, h
b)
105 1
751 \
3,2 Val.%,
65
0 100 200 300 400
Time, h
C)
Figure 30. Weight loss of different Si3N4-materialsin H2S04 at 60°C (additive composition is given
at the curves) a) and weight loss b) and residual strength c) of different materials with different
amount of Y203/A1203sintering additives in HCI at 60°C.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
190 6 Silicon Nitride Bused Hard Muterials
0,li ,
25 30 35 40 45 50 55 60 65 70
' ' " ' ' ' 8
Si02-content, Mol%
Figure 31. Weight loss of silicon nitride materials with different Si02 content in the grain boundary
(200 h corrosion in H2S04 at 60°C).
7%Y203/A1203
Time. h
7%Y 203/Al203
I O%Y203/A1203
Time , h
Figure 32. Weight loss and residual strength of silicon nitride materials in 1 N NaOH at 60°C.
in the grain boundary are less stable under hydrothermal conditions than materials
with a low Si02 content in the grain-boundary phase.
Under hydrothermal corrosion at 270"C, a significant dissolution of the Si3N4
grains takes place. In materials with Y203/A1203as sintering additives, the dissolu-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
192 6 Silicon Nitride Bused Hard Mutevials
Si3N4
MgOiA1203
r4
‘E 1-
HIPSN
E
Sl
y:
without additives
-0
0,014
10 100 1000
Time, h
Figure 33. Weight loss of silicon nitride materials under hydrothermal conditions at 210°C.
tion rate of the grains is higher than that of the grain-boundary phase (see Fig. 34).
The weight loss due to hydrothermal corrosion depends linearly on time. Only
materials with high corrosion rates (MgO-containing and Si02-rich materials)
show an increase of the corrosion rate with time.
Investigations of the corrosion behaviour of silicon nitride materials in acids,
bases and under hydrothermal conditions show that a main parameter governing
the corrosion resistance is the composition and amount of the grain-boundary
phase. Materials with a high Si02 content in the grain-boundary phase are stable
in acids and less stable in bases and under hydrothermal conditions. Materials
with a low Si02 content in the grain-boundary phase are stable in bases and
under hydrothermal condition but less stable in acids.
Up to now no materials which are simultaneously stable in acids, bases and under
hydrothermal conditions exist.
A more detailed understanding of the corrosion mechanisms and the influence of
the medium on the lifetime of the materials is necessary before silicon nitride
materials can be used in a broad range of applications under corrosive conditions.
The results of the investigations indicate that tailoring of the microstructure and
the grain boundary composition is necessary for the application of silicon nitride
ceramics in corrosive environments.
The development of the silicon nitride materials in the future will be divided in
two main directions; the increased penetration in the market of the state of the
art silicon nitride materials by improving the technology and reliability and decreas-
ing the production costs and the development of new materials with improved wear,
corrosion, high temperature properties.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
794 6 Silicon Nitride Based Hard Materials
In the last decade a continuously increase in the number of application fields was
observed. The technical feasibility and the reliability of the materials and compo-
nents were shown in many applications, such as valves and other parts for the
automotive industry, bearings, household applications. For the wider penetration
into the market there is a engineering work necessary to optimise the production,
stabilise the reliability and reduce the production cost. First steps in this direction
were made by using low cost powders and optimising processing, sintering and
finishing.
The majority of applications can be realised with strength levels lower 1000-
1200 MPa.
Higher strength level need a special more expensive finishing procedure increasing
the costs of parts. Additionally a strength level of 1200 MPa corresponds to a defect
size of 10-15 pm (at the existing KTClevels) which is in the range of defects created
during operating of the materials.
Whereas the direction of further material development mentioned above is
connected with optimising the technology, the second direction is connected with
adapting of the microstructure and properties of silicon nitride materials on differ-
ent applications. As like as exists different kinds of steel, which are well designed for
the given applications, as well in the future will exist different classes of silicon
nitride materials aimed for applications under different conditions e.g. wear, corro-
sion, operation at high temperature and so on (see Table 1).
These differentiation processes stay only at the beginning and need as a bases a
controlled formation of the microstructures and the knowledge of the relation
between the microstructure and properties.
The main factors and processes influencing the microstructure of Si3N4materials
are known and the microstructure can be tailored in a wide range. In the future the
use of different p-silicon nitride powders as seeds for a precise tailoring of the micro-
structure will be used more intensively. New aspects can be expected by the
development of fine grained materials. The correlation between microstructure,
strength and toughness are well understood, whereas the understanding of the
correlation between microstructure on the one side and wear, corrosion properties
and long term behaviour on the other side need further intensive interdisciplinary
investigations.
An intensive development of high fracture toughness a’-SiAlON-materials can be
expected. The better understanding of the microstructure formation of these
materials and a further improvement of strength and toughness by a reduced
grain size are main topics of the investigations.
Frequently, there is a need to make a compromise between the different properties
due to the opposite dependencies on the microstructure. How such compromise
looks like depends on the technological possibilities and on the powder properties,
which developments offer new possibilities.
An improvement of the materials behaviour can be expected by the development
of homogeneous composite materials based on silicon nitride, as it was shown for
Si3N4/MoSi2or Si3N4/SiCcomposites for the high temperature long term applica-
tions. Especially in this area the use of metallorganic precursors can lead to an
improvement of the materials. The reduction of sintering temperature and in
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 795
some modifications the shrinkage (concept of active filler) can allow the production
of composites of thermodynamic unstable components. The development in this
direction is only at the beginning.
Acknowledgements
The authors are grateful to the colleges of the IKTS which take part in the technical
completion of the paper especially A. Bales and D. Hermannutz. They are indebted
to the coworkers in the IKTS especially to Dr. I. Schulz, Dr. T. Reich, A. Bales and
G. Michael involved in research projects in the field of silicon nitride materials which
results were included in this paper. The work is based on different projects supported
by the BMBF, AiF and the DFG in the field of silicon nitride materials.
References
1. S. Hampshire, “Nitride Ceramics in Structure and Properties of Ceramics”, ed. R. W. Cahn,
P. Haasen and E. J. Kramer, Materials Science and Teclznology, 11, 121-168.
2. R. N. Katz, “Application of Silicon Nitride based Ceramics in the U.S.”, Mat. Res. Soc. Symp.
Proc., 287, 1993. 197.
3. K. H. Jack, “SiALON Ceramics: Retrospect and Prospect”, Mat. Res. Soc. Symp. Proc., 287,
1993, 15-18.
4. J. G. Heinrich and H. Kruner, “Silicon Nitride Materials for Engine Applications”, M. J.
Hoffmann and G. Petzow (eds.), Tailoring of’ Mechanical Properties of Si,N4 Ceramics,
Kluwer Academic Publishers, 1994, 1941.
5. “Der Kochherd der Zukunft arbeitet mit Si3N4-Platten”,cfi/Ber. DKG, 76, 1999, D 18-20.
6. J. Heinrich, “Siliciumnitridkeramik”, editor: J. Kriegsniann, DKG Technisclze Keramisclze
Werkstofe, 1999, 4.3.1 .O.
7 . H. M. Jennings, “Review on Reactions between Silicon and Nitrogen, Part 1 Mechanisms”,
Journal of Material Science, 18, 1983, 951 -967.
8. A. Zerr, G. Miehe, G. Serghiou a. o., “Synthesis of Cubic Silicon Nitride”, Nature, 400, 1999,
340-344.
9. W. Ching, Y. Xu, J. Gale and M. Riihle, “Ab-Inito Total Energy Calculation of cc- and p-sili-
con nitride and the Derivation of Effective Pair Potentials with Application to Lattice
Dynamics”, J . Am. Ceram. Soc, 81, 1998, 3189-3196.
10. I. Kohatsu and J. W. McCauley, “Re-Examination of the Crystal structure of a-Si3N4”,Mat.
Res. Bull, 9, 1914, 9 17-920.
11. H. Suematsu, J. J. Petrovic and T. E. Mitchell, “Plastic deformation of silicon nitride single
crystals” Materials Science und Engineering, A 209, 1996, 97-102.
12. J. Ihsza, T. Eschner and K. Rundgren, “Hardness anisotropy in bimodal grained gas pressure
sintered Si3N4”,Journal of Materials Science Letters, 16, 1997, 16641667.
13. Andrievskii, Spivak I. I. , “Nitrid kremnija i materiali na evo osnove”, Moskau, Metallurgijr,
1984, 19-20.
14. H. Mandal, D. P. Thompson and K. H. Jack, “cc H f3 Phase Transformations in Silicon
Nitride and Sialons”, preprint.
15. L.-0. Nordberg, “cc-Sialon Ceramics and Y-cc-Sialon Composites; Composition, Microstruc-
tures and Properties”, Doctorial Dissertation, 1997, 1-124.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
196 6 Silicon Nitride Based Hard Materials
41. H.-J. Kleebe, “Structure and Chemistry of Interfaces in Si,N4 Ceramics Studied by Transmis-
sion Electron Microscopy”, Journal of’the Ceramics Society of Japan, 105, 1997, 453475.
42. T.-Y. Tien, “Silicon Nitride Ceramics-Alloy Design”, Mut. Res. Soc. Symp. Proc., 287, 1993.
43. T. Y. Tien and C. J. Hwang, “Microstructural Development in Silicon Nitride Ceramics”,
Mater. Sci. Forum, 47, 1989, 84109.
44. L. L. Wang and T. Y. Tien, “Morphology of Silicon Nitride Grown from a Liquid Phase”,
J. Am. Ceram. Soc, 81, 1998, 10, 2677-2686.
45. D. Y. Yoon, Korean Advanced Institute of Science and Engineering, Report 16.6. 1999 in the
IKTS.
46. Y.-J. Park. N.-M. Hwang and D. Y. Yoon, “Abnormal Growth of Faceted (WC) Grains in a
(Co) Liquid Matrix”, Metal/. Mater. Trans., 27A, 1996, 1-1 1.
47. D.-D. Lee, S.-J. L. Kang and D. N. Yoon, “Mechanism of Grain Growth and a-p’ Transfor-
mation During Liquid-Phase Sintering of ff-Sialon”, J. Am. Ceram. Soc., 71, 1988, 803-806.
48. N. Hirosaki, Y. Akimune and M. Mitomo, “Microstructural Design by Selective Grain
Growth of a-Si3N4”, Mat. Res. Soc. Symp. Proc., 287, 1993, 405410.
49. M. Herrmann, I. Schulz and J. Hintermayer, “Materials From Low Cost Silicon Nitride
Powders”, Proc. 4th EcerS Conf ., Riccione, Ed. by C. Galassi, Gruppo Editoriale Faenza
Editrice, 1995, 21 1-216.
50. N. Hirosaki, M. Ando, Y. Akimune and M. Mitomo, “Gas-Pressure Sintering of p-Silicon
nitride Containing Y z 0 3 and Nd z0 3” ,J. of Ceranz. Soc. ofJapan, 100, 1992, 817-820.
51. M. Mitomo, M. Tsutsumi, H. Tanaka, S. Uenosono and F. Saito, “Grain Growth During
Gas-Pressure Sintering of p-Silicon Nitride”, J . Am. Ceratn. Soc., 73, 1990, 244145.
52. M. Mitomo, H. Hirotsuru, H. Suematsu and T. Nishimura, “Fine-Grained Silcon Nitride
Ceramics Prepared from P-Powder”, J . Am. Ceram. Soc., 78, 1995, 21 1-14.
53. J. Hintermayer, G. Ernst, W. Gmohling, G.Schrol1 and W. Kobler, “Verfahren zur Herstel-
lung von Siliziumnitrid” EP 0377132B1, 1994.
54. E. Hidseyuki and M. Mitomo, “Control and Characterization of Abnormally Grown Grains
in Silicon Nitride Ceramics”, Journal of the European Ceramic Society, 17, 1997, 797-804.
55. M. Mitomo, “In-situ Microstructural Control in Engineering Ceramics”, Key Engineering
Materials, 161-163, 1999, 53-58.
56. P. Branvall and K. Rundgren, “Grain growth mechanisms in in situ reinforced Si3N4”,Brit.
Ceranz. Proc., 60, 1999, 7-9.
57. K. Hirao, T. Nagaoka, M. E. Brito and S. Kanzaki, “Microstructure Control of Silicon
Nitride by Seeding with Rodlike 8-Silicon Nitride Particles”, J . Am. Ceram. Soc., 77, 1994,
1857-62.
58. K. Hirao, T. Nagaoka, M. E. Brito and S. Kanzaki, “Mechanical Properties of Silicon
Nitrides with Tailored Microstructure By Seeding”, Journal of the Ceramic Society of
Japan, 104, 1996, 55-59.
59. R. A. Andrievski, “Some high-temperature properties of silicon nitride”, High Temperature -
66. N. Hirosaki, Y. Okamoto, M. Ando, F. Munakata and Y. Akimune, “Effect of Grain Growth
on the Thermal Conductivity of Silicon Nitride”, Journal of the Ceramic Society ofJapan, 104,
1996, 50-54.
67. D. A. Bonnell, M. Riihle and T.-Y. Tien, “Redistribution of Aluminum Ions During Proces-
sing of Sialon Ceramics”, J. Am. Ceram. Soc., 69, 1986, 623-27.
68. M. Kitayama, K. Hirano, M. Toriyama and S. Kanzaki, “Anisotropic Ostwald Ripening in 0-
Si3N4With Different Lantanide Additives”, G . L. Messing, F. F. Lange, S.-I. Hirano: Ceramic
Processing Science, 5 17--24.
69. M. J. Hoffmann, “Analysis of Microstructural Development and Mechanical Properties of
Si3N4Ceramics”, M . J. Hoffmann, G . Petzow, Tailoring of Mechanical Properties of Si3N4
Ceramics, Kluwer Academic Publishers, 1994, 59-72.
70. H. J. Kleebe, G. Pezzotti and G. Ziegler, “Microstructure and Fructure Toughness of Si,N4
Ceramics”, J. Am. Ceram. Soc., 82, 1999, 1857-1860.
71. A. R. Prunier and A. J. Pyzik, “Self reinforced Silicon Nitride for Cutting Tool Application”,
Key Eng. Mat., 89-91, 1994, 129-134.
72. A. J. Pyzik and D. R. Beaman, “Microstructure and Properties of Self-Reinforced Silicon
Nitride”, J. Am. Ceram. Soc., 76, 1993, 273144.
73. M. Menon and I.-W. Chen, “Reaction Densification of a’- Sialon: I, Wetting Behavior and
Acid-Base Reactions”, Journal of the American Ceramic Society, 78, 1995, 545-552.
74. M. Menon and I.-W. Chen, “Reaction Densification of d-Sialon: 11, Densification Behavior”,
Journal of the American Ceramic Society, 78, 1995, 553-559.
75. H. Mandel and M. J. Hoffmann, “Novel developments in a-SiALON ceramics”, Brit. Ceram.
Proc., 60,1999, 11-13.
76. Schutzrecht USA WO 98/23554 (1998-06-04).
77. I-W. Chen and A. Rosenflanz, “A tough SiAlON ceramic based on cc-Si3N, with a whisker-
like microstructure”, Nature, 389, 1997, 70 1-704.
78. H. Klemm, M. Herrmann, T. Reich, C. Schubert, L. Frdssek, G. Wotting, E. Gugel and
G. Nietfeld, “High-Temperature Properties of mixed a’/$-Sialon Materials”, Journal of the
American Ceramic Society, 81, 1998, 1141-1 148.
79. T . 4 . Sheu, “Microstructure and Mechanical Properties of In Situ 0-Si3N4/d-SialonCompo-
site”, Journal of the American Ceramic Society, 77, 1994, 2345-2353.
80. M. Herrmann, annual report 1998, IKTS, 1998.
81. N. Camascu, D. P. Thompson and H. Mandal, “Effect of Starting Composition, Type of Rare
Earth Sintering Additive and Amount of Liquid Phase on a’/p-Sialon Transformation”,
Journal of the European Ceramic Society, 17, 1997, 599-613.
82. K. Urashima, Y. Ikai and S. Iwase, “Features of a Superior Strength Si3N4 Ceramic and its
Applications”, 6th International Symposium on Ceramic Materials and Components for
Engines, 1997, 167-112.
83. G. Riedel and H. Kruner, “Si3N4-material with high strength (1400MPa)”, Third Euro
Ceramics, Madrid, 1993, 453-458.
84. R. F. Silva and J. M. Vieira, “Hot hardness of Si3N4-basedmaterials”, Journal of Material
Science, 30, 1995, 5531-5536.
85. M. Yoshimura, T. Nishioka, A. Yamakawa and M. Miyake, “Grain Size Controlled High-
Strength Silicon Nitride Ceramics”, J. Ceram. SOC.Japan, 1995, 407.
86. T. Nishioka, K. Matsunuma, T. Yamamoto, A. Yamakawa and M. Miyake, “Development of
High Strength Si3N4Material for Automobile Parts”, Sumitomo Electric Technical Review, 36,
1998, 77.
87. N. Hirosaki, Y. Akimune and M. Mitomo, “Effect of Grain Growth of 0-Silicon Nitride
on Strength, Weibull Modulus and Fracture Toughness”, J. Am. Ceram. Soc., 76, 1993,
1892-94.
88. P. F. Becher, H. T. Lin, S. L. Hwang, M. J. Hoffmann and I. W. Chen, “The Influence of
Microstructure an the Mechanical Behaviour of Silicon Nitride Ceramics”, Mat. Res. Soc.
Symp. Proc., 281, 1993, 147-157.
89. P. Sajgalik, J. Dusza and M. J. Hoffmann, “Relationship between Microstructure, Toughen-
ing Mechanism, and Fracture Toughness of Reinforeced Silicon Nitride Ceramics”, J. Am.
Ceram. Soc., 78, 1995, 2619-24.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 799
90. C. W. Li, C. J. Gasdaska, J. Goldacker and S. C. Lui, “Damage Resistance of in Situ Rein-
forced Silicon Nitride”, Mat. Res. SOC. Symp. Proc., 287, 1993, S. 473-480.
91. B. Speicher, G. A. Schneider, W. Drepler, G. Lindemann, H. Boder and V. Knoblauch,
“Reliability of Ceramic Valve plates for Common-Rail Injection Pumps”, preprint Materialica
1999, 1999.
92. I. Peterson and T. Tien, “Effect of the Grain Boundary Thermal Expansion Coefficient on the
Fracture Toughness in Silicon Nitride”, J . Am. Crram. Soc., 78, [9], 1995, 2345-52.
93. E. Y. Sun, P. F. Becher and K. P. Plucknett, “Microstructural Design of Silicon Nitride with
Improved Fracture Toughness: 11, Effects of Yttria and Alumina Additives”, J . Am. Ceram.
SOC.,81, 1998, 2831-2840.
94. A. G. Evans, “Perspective on the Development of High-Toughness Ceramics”, J . Am. Ceram.
SOC.,73, 1990, 187-206.
95. H. Kessler. H. Kleebe, R. Cannon and W. Pompe, “Influence of Internal stresses on Crystal-
lization of Intergranular Phases in Ceramics”, Acta metall muter., 40,1992, 2233-2245.
96. I. Tanaka, G. Pezotti, Y. Miyamoto and T. Okamoto, “Fracture Toughness of Si3N4and its
Si3N4 Whisker Composite Without Sintering Aids”, J . of Mat. Sci., 26, 1991, 208-210.
97. J. Zhihao and Q. Guanjun, “Fatigue Behavior of some Advanced Ceramic Materials”, 6th
International Symposium on Ceramic Materials and Componentsfor Engines, 1997, 760-765.
98. T. Ogasawara and Y. Mabuchi, “Effect of Humidity and Cyclic Loading Condition on
Fatigue behavior of Silicon Nitride”, J . Ceram. Soc. Jap., Int. Edition, 101, 1993, 1122-1 127.
99. D. Jacobs and I. Chen, “Mechanical and Environmental Factors in the Cyclic and Static
Fatigue of Silicon Nitride”, J . Am. Ceram. Soc., 77, 1994, 1153-61.
100. P. F. Becher, E. Y. Sun, K. P. Plucknett, K. B. Alexander, C. H. Hsueh, H. T. Lin, S. B.
Waters and C. G. Westmoreland, “Microstructural Design of Silicon Nitride with Improved
Fracture Toughness: I, Effects of Grain Shape and Size”, J . Am. Ceram. SOC., 81, 1998,
2821-30.
101. R. Schober, M. Herrmann, W. Krehler, H.-J. Richter, W. Hermel and G. Naumann, “Verdich-
tungsverhalten und Eigenschaften von nachgesintertem reaktionsgebundenem Siliziumnitrid
(SRBSN)”, editor: J. Kriegsmann, DKG Technische Keramische Werkstofe, 3. Erg.-Lfg. 1999.
102. I. Tanaka, G. Pezzotti, T. Okamoto, Y. Miyamoto and M. Koizumi, “Hot Isostatic Press
Sintering and Properties of Silicon Nitride at Elevated Temperatures”, J . Am. Ceram. SOC.,
74, 1992, 752-59.
103. H. Klemm, M. Herrmann and Chr. Schubert, “High Temperature Oxidation and Corrosion of
Silicon-Based Nonoxide Ceramics”, A S ME Turbo Expo ‘98,Stockholm, Sweden 1998, ASME
Paper 98-GT-480.
104. H. Klemm, M. Herrmann and Chr. Schubert, “High Temperature Oxidation of Silicon Nitride
Based Ceramic Materials”, Proc. 6th Internat. Symp. on Ceram. Muter. & Comp.for Engines,
Arita, Japan, 1997, 57681.
105. I. Tanaka, G. Pezzotti, K. Matsushita, Y. Miyamoto and T. Okamoto, “Impurity Enhanced
Cavity Formation in Si3N4at Elevated Temperatures”, J . Am. Ceram. SOC.,74, 1992,752-59.
106. H. Klemm and G. Pezzotti, “Fracture Toughness and Time-Dependent Strength Behavior of
Low- Doped Silicon Nitrides for Applications at 1400C”, J . Am. Ceram. Soc., 77, 1994,553-61.
107. G. Pezzotti, K. Ota and H.-J. Kleebe, “Viscous Slip along Grain Boundaries in Chlorine-
Doped Silicon Nitride,” J. Am. Ceram. SOC.,80, 1997, 234148.
108. M. Herrmann, Chr. Schubert, A. Rendtel and H. Hiibner, “Silicon Nitride/Silicon Carbide
Nanocomposite Materials: I, Fabrication and Mechanical Properties at Room Temperature”,
J . Am. Ceram. Soc., 81, 1998, 1095-108.
109. K. Niihara, “New Design Concept of Structural Ceramics - Ceramic Nanocomposites”, The
Centennial Memorial Issue of the Ceramic Society of Japan, 99, 1991, 974.
110. M. Herrmann, H. Klemm, Chr. Schubert and W. Hermel, “Long-Term Behavior of Sic/
Si3N4-Nanocomposites at 1400-1500”C”, Key Engineering Muter., 132-136, 1977-80, 1997.
11 1. G. Pezzotti and M. Sakai, “Effect of a Silicon Carbide ‘Nano-Dispersion’ on the Mechanical
Properties of Silicon Nitride”, J . Am. Ceram. SOC.,77, 1994, 303941.
112. H. Klemm. K. Tangermdnn, Chr. Schubert and W. Hermel, “Influence of Molybdenum
Silicide Additions on High-Temperature Oxidation Resistance of Silicon Nitride Materials”,
J . Am. Ceram. Soc., 79, 1996, 2429-35.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
800 6 Silicon Nitride Based Hard Materials
113. R. Sternagel and M. Popp, “Innovationen fur Gleitlager, Walzlager, Dichtungen und Fuhrun-
gen”, VDI-Ber., 1331, 1997, 131-138.
114. M. Rombach, R. Schofer and W. Pfeiffer, “Tragfahigkeit und Lebensdauer fur Wdzlager aus
Hochleistungskeramik”, Report of a BMBF- project 03 M2107B, 1996.
115. M. Popp, R. Sternagel and G.Wotting, “Hybride- and Ceramic Rolling Bearings with modi-
fied Surface and low Friction Rolling Contact”, Euromat 1999, 1999.
116. Ch. Schubert and M. Herrmann, “Materials development of silicon nitride for ceramic ball
bearings”, Report of the BMBF- project Nr. 03 M2107B4, 1996.
117. M. Herrmann, I. Schulz, Chr. Schubert and W. Hermel, “Siliziumnitridwerkstoffe mit
niedrigem Reibkoeffizienten und hohem Verschleipwiderstand”, Werkstqfwoche 98, Vol. V ,
Symposium 6 , 1998, 241-247.
118. M. Popp and R. Sternnagel, “Neue Entwicklungen bei Walzlagern aus Hochleistungskera-
mik’, Werkstoffwoche 98, Vol. V , Symposium 6 , 1998, 247-253.
119. B. R. Lawn, S. K. Lee and K. S. Lee, “Fracture and Deformation Damage Accumulation in
Tough Ceramics”, Key Engineering Materials, 161-163, 1999, 3-8.
120. S. K. Lee and B. R. Lawn, “Role of Microstructure in Hertzian Contact Damage in Silicon
Ntride: 11, Strength Degradation”, Journal qf’ the American Ceramic Society, 81, 1998, Nr.
4, 997-1003.
121. S. W. Lee, “Tribological Characterization in Ceramic Based Composite”, Key Engineering
Materials, 161-163, 1999, 593-598.
122. K. H. Habig, “Tribologisches Verhalten von Ingenieur-Keramik”, Ingenieur- Werkstoffe, I,
1989, 78-83.
123. A. Skopp, “Tribologisches Verhalten von Siliziumnitridwerkstoffen bei Festkorpergleitrei-
bung zwischen 22°C und lOOO”C”, Dissertation 1993.
124. S. K. Lee and B. R. Lawn: Role of Microstructure in Hertzian Contact Damage in Silicon
Ntride: 11, Strength Degradation. In: Journal of ihe American Ceramic Society, 81, 1998,
997-1003.
125. H. Tanaka, “A Recent Tendency of Si3N4Cutting Tools”, Advanced Materials: Ceramics,
Powders, corrosion and Advanced Processing, 14A, 1994, 541-545.
126. H. K. Tonshoff and B. Denkena, “Wear of Ceramic Tools in Milling”, Journal of the Society of
Tribologists and Lubrication Engineers, 47, 1991, 772-778.
127. H. K. Tonshoff, Chr. Blawit, M. Rodewald and H.-G. Wobker, “Entwicklung und
VerschleiPverhalten von Si3N4-Schneidkeramik”, Mat.-wiss. u. Werkstofftech., 26, 1995,
255-262.
128. G. Brandt, “Development of ceramic cutting tools”, Maieriaux & Techniques, 85, 1997, 3-12.
129. R. D. Nixon and P. K. Mehrotra, “Optimization of Sintering Aids in Silicon Nitrides for
Cutting Tool Application”, Ceramic Engineering & Science Proceedings, 18, 1997, 457464.
130. H. K. Tonshoff and C. Blawit, “Wear mechanism of silicon nitride in dependence of workpiece
material”, BMFT- Project, Final Report No. 03T0028A, 1993, 1-96.
131. H. Miao, L. Qi and G. Cui: Silicon Nitride Ceramic Cutting - Tools and their Applications”,
Key Engineering Materials, 114, 1995, S. 135-172.
132. G. Brandt, “Wear- and thermal shock-resistant SIALON cutting tool material”, EP 97-850038
9703 1 1.
133. US- Patent 432 325.
134. Y. R. Liu, J. J. Liu, B. L. Zhu, Z. R. Zhu, L. Vincent and P. Kapsa, “Wear Maps of Si3N4
ceramic cutting tools”, J . Mater. Eng. and Perform., 6, 1997, 671-675.
135. Gmelin Handbook, Si-Suppl, b5d2., 11&117; 1995, 184-210.
136. K. Komeya, T. Meguro, S. Atago, C.-H. Lin, Y. Abe and M . Komatsu, “Corrosion Resis-
tance of Silicon Nitride Ceramics”, Key Engineering Materials, 161-163, 1999, 235-238.
137. M. Herrmann, A. Krell, J. Adler, G. Wotting, T. Hollstein, W. Pfeifer and M. Rombach,
“Innovationen fur Gleitlager, Walzlager, Dichtungen und Fuhrungen”, VDZ-Ber., 1331,
1997,251-258.
138. T. Hollstein, T. Graas, K. Bundschuh and M. Schutze, Keramische Zeitschrijt, 50, 1998, 416 -
421.
139. T. Sato, Y. Tokunaga, T. Endo, M. Shimada, K. Komeya, et al., “Corrosion of Silicon Nitride
Ceramics in Aqueous Hydrogen Cloride Solutions”, J. Mat. Science, 23, 1988, 3440-3446.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
RL.firences 80 1
140. S. Iio, A. Okada, A. Akira, A. Tetsuo and Y. Masahiro, “Corrosion of silicon nitride ceramics
in solutions. Part 3. Corrosion behavior in hot sulfuricacid and microstructure of the corroded
layer”, Journal oftlie Ceran?ic Society of Japan, h t . Edition, 100, 1992, July, 954-956.
141. M. Herrmann, Chr. Schubert and G . Michael, “Korrosionsstabile keramische Werkstoffe fur
Anwendungen in Walzlagern und im Anlagenbau”. Fortschrittsberichte der DKG, 14, 1999,
13~151.
142. M. Herrmann and G. Michael, “Corrosion Behaviour of Engineering Ceramics in Acids and
Basic Solutions”, Britidl Ceramic Proceedings, 60, extended Abstracts, Vol I, 1999, 455457.
143. A. Okada and M. Yoshimura, “Mechanical Degradation of Silicon Nitride Ceramics in Cor-
rosive Solutions of Boiling Sulphuric Acid”, K q Engineering Materials, 113, 1996, 227-236.
144. Y. G. Gogozi and V. A. Lavrenko, “Corrosion of High Performance Ceramics”, Springer
Verlag, Berlin, 1992, 7 6 7 8 .
145. M. Shimada and T. Sato, “Corrosion of Silicon Nitride Ceramics in H F and HCI Solutions”
Ceram. Trans., 10, 1989, 355-365.
146. Q. Fang, P. S . Sidky and M. G . Hocking, “The Effect of Corrosion and Erosion on Ceramic
Materials”, Corrosion Science, 39, 1997, 51 1-527.
147. A. Paul, “Chemical Durability of Glasses: a Thermodynamic Aproach”. J . Mat. Sci., 12, 1977,
2246-2268.
148. K. Kanbara, N. Uchida, K. Uematsu, T. Kurita, K. Yoshimoto and Y. Suzuki, “Corrosion of
Silicon Nitride Ceramics by Nitric Acid”, Mat. Res. Soc. Symp. Proc., 287, 1993, 533-538.
149. T. Sato, Y. Tokunaga, T. Endo, M. Shimada, K Komeya, K. Nishida, M. Komatsu and
T. Kameda, “Corrosion of Silicon Ceramics in Aqueous H F Solutions”, J . q f M a t . Sci., 23,
1988, 3440-3446.
150. T. Sato, S. Sato, K. Tamura and A. Okuwaki, “Corrrosion Behaviour of SiliconNitride
Ceramics in Caustic Alkaline Solutions at High Temperatures”, Br. Ceram. Trans. J., 91,
1992, 117-120.
151. T. Sato, T. Murakami, E. Shimada, K. Komeya, T. Komeda and M. Komatsu, “Corrosion of
Silicon nitride ceramics under hydrothermal conditions” J . Mat. Science, 26, 1991, 1749-1754.
152. M. Yoshimura and S. Yamamoto, “Corrosion and alteration of ceramic materials in high
temperature-high pressure water”, Serumikkusu, 30, 1995, 995-998.
153. K. Oda, T. Yoshio. Y. Miyamoto and M. Koizumi, “Hydrothermal corrosion of pure, hot
isostatically pressed silicon nitride”, J . A m . Cerum. Soc., 76, 1993, 1365-1368.
154. K. Oda and T. Yoshio, “Properties of YzO3-Al2O3-SiO2Glasses as a Model System of Grain
Boundary Phase of SijN4 Ceramics (Part 2)”, J . Ceraniic Soc. qfJupan, 99, 1991, 11 10-1 112.
155. S. Iio, A. Okada, T. Asano and M. Yoshimura, ”Corrosion Behavior of Silicon Nitride
Ceramics in Aqueous Solutions (Part 3)”, J . o j t h e Ceramic Soc. ofJc~pan,100, 1992,965-967.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
7.1 Introduction
Materials based on boron compounds have been explored for many decades because
of their exceptional properties in respect to chemical bonding, crystal structure, and
phonon and electron conduction. Especially in the field of energy conversion, elec-
tron emission, and neutron absorption, borides occupy many niches of application
for which no other material can be employed. Until approximately 1980, the main
interest in borides always came, however, from basic research aimed at the
understanding of their electronic structure, being either responsible for the unique
transport properties or the peculiarities in chemical bonding. It is, therefore, no
wonder that the most information about borides was at that time created from
the viewpoint of physicists and chemists.
Although even some of the interesting thermal and mechanical properties of
borides, e.g., the generally high melting point and high hardness, have already
been exploited and led to application as wear-resistant parts or grinding grits a
long time ago, the interest in borides increased dramatically together with the
fundamental understanding of technical ceramics. The exploitation of materials
with exceptional mechanical, chemical, electrical and thermal properties yielded
boron compounds as potential candidates for “high-tech” applications besides the
well-developed oxide ceramics, silicon nitride, silicon carbide, and hard metals or
cermets.
For a successful economic application, these advanced ceramics should be
available without any limitations; prepared by relatively simple methods, they
should exhibit a low specific gravity, high reliability, long lifetime and of course
they should be available at low costs. Unfortunately, most boron-based materials
do not satisfy all of these requirements and, hence, disadvantages and disappoint-
ments have to be tolerated in development and application. In particular, densifica-
tion of powder-derived parts is extremely difficult due to their high amount of
covalent bonding resulting in low diffusion coefficients and, therefore, costly.
Furthermore, oxide impurities on surfaces and grain boundaries create undesired
and unexpected effects on the properties whereas the affinity of boron to oxygen
limits its use at higher temperature in an oxidizing atmosphere. Additionally,
phase diagrams and crystal structures of the particular compounds are the subject
of permanent revision since improvement of the purity of raw materials, more
sophisticated synthesis procedures and characterization methods, such as precur-
sor-based processing, advance chemical analysis, atomic-resolution transmission
electron microscopy, and high-temperature X-ray diffraction, yielded more precise
but also more confusing information about the true characteristics of the boride
structures and phases.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.2 Chemical Bonding and Crystal Chemistry of Borides 803
Besides the relatively high costs a lack in understanding is the main reason why
boride-based materials are not yet in as widespread use as the other ceramics
materials, although some properties such as extreme hardness make them superior
to oxides and nitrides. Thus, some “eras” of boride exploration can be distin-
guished, one supported by the needs of the nuclear energy and weapon industry
around 1960, another between 1970 and 1980 for thermo-electric power generation
and coming up again at the present, the next driven by the need for exceptional wear
properties between 1980 and 1990, and finally between 1985 and 1992 triggered by
aerospace and military research. The interest in borides rising from a political and
strategic point of view has contributed to the fact that information about borides is
still difficult to obtain and even more difficult to interpret since most of the data have
been restricted for a long time and their sources and quality cannot be ascertained in
all cases. Thus, sometimes contradictory data are available that cannot be easily
discerned as being right or wrong.
Besides the “high-tech” applications of novel materials, it should not be forgotten
that there was and still is a tremendous market for borides in the metallurgy of steel
and iron, e.g., for antioxidizing additives in refractory linings or as alloying ingredi-
ents for the metals.
Thus, this contribution is aimed at the state of the art in boride ceramics with their
problems in densification, microstructural peculiarities and exceptional mechanical
properties. Starting with the unique interaction of metallic, covalent and ionic types
of bonding and the crystal structures of technically important compounds, phase
diagrams will be presented as far as they are of technical interest. The major part
consists of the description of the synthesis and properties of ceramics and cermets,
reflecting the development of suitable sintering procedures and the consequent
improvement of the thermal and mechanical properties.
electrons. These electrons are thus more or less delocalized. The resulting sp2 hybri-
dization leads to the plane B3X3 hexagon as the main structural element in BN,
B203,H3B03 and related compounds, and to the B3 triangle as a fraction of the
typical five-fold symmetric icosahedron of elemental boron, the group of boranes
and their derivatives. Depending on the saturation of the electron deficiency, soft
and non-conducting, salt-like compounds or semimetallic to metallic materials of
exceptionally high melting point and hardness and excellent electrical conductivity
exist. As pointed out in the following section, the latter boron compounds may
contain ionic, metallic and covalent fractions of bonding forming very stable
compounds due to the well-balanced electron transfer between metal and boron
sublattice.
1-
i
Single Multiple
x
0-0
8
0-0
sp2-Type B,-TYPe
have been the subjects of investigation for many years [1,2]. In general, compounds
with a boron-to-metal ratio of less than 1.O are built up of isolated boron atoms or
pairs with a low B-B interaction (e.g., Ni3B, Ru7B3,Fe2B, Cr5B3),in zigzag chains
with additional, isolated B (e.g., o-Ni4B3).At a ratio of 1.0 to 1.3, infinite chains are
formed which may be parallel to one or even two crystallographic axes (e.g., m-
Ni4B3, FeB, CrB, MOB), whereas in M3B4borides double chains are predominant
(e.g., Cr3B4). With increasing boron content, two-dimensional nets are stable,
yielding preferential stoichiometries between M2B3 and MB4. The most important
structure type group thereof is the A1B2 type, which is covered in more detail later
on. Three-dimensional frameworks exist in so-called higher borides with typical
stoichiometries of MB4, MB,, MBI2,and MB,,. Channels with rectangular cross-
sections were found in, e.g., CrB4 and MnB4, which is unique for the three-center
bond of boron [3]. A rigid boron skeleton consisting of B, octahedra is a character-
istic of the CaB6 structure group (important member: LaB,), whereas the UB12
structure contains BI2cubo-octahedra. Other borides of MB6 and MB12 stoichio-
metry or a higher boron-to-metal ratio, especially the main group element borides,
can be derived from the trigonal rhombohedra1 a-boron or (3-boron structure with
the B12 icosahedron as the most important structural unit. The SiB6 structure
consists of a special type of boron arrangement built up by 18 boron icosahedra
(BIZ units), 4 icosihexaedra (BI5 units) and 8 single atoms where some of the
boron positions are occupied by Si [4]. A structural curiosity are the yttrium borides
with B/Y > 25. The member being richest in boron, YB66crystallizes in cubic sym-
metry and contains 1584 boron atoms and approximately 24 yttrium atoms (Fig. 2).
The boron is arranged in 27 (B12)13 units made up by 13 interconnected icosahedra
and in 8 Bso clusters being occupied by approximately 42 B atoms. The yttrium is
arranged in Y-Y pairs of most probably Y3+ state, providing all the electrons neces-
sary to stabilize this framework [5, 61. Even this compound is of high technical
importance for monochromators of synchrotron radiation.
Some general systematics on chemical bonding and crystal chemistry have been
published by Matkovich and Economy [7], and Aronsson et al. [8, 91, who also
refer to the structural similarities in silicides and phosphides.
: Y-Y pair
co I
0 @.
Metal Boron
A
H C
K’ A, B: metal
a1 -c A H, K’: boron
Figure 5. High-resolution TEM micrograph of WlB, stacks, view of [110] zone axis, distance of
stacks: 13.9 A (courtesy, B. Freitag [368]).
diborides the boron 2p0 and 2p7t orbitals are the main constituents of the states at
the Fermi edge, while for the transition metal diborides it is the localized metal 3d
orbitals which are the predominant component of the valence and conduction
bands. Since the boron sublattice is electron deficient all diborides exhibit an elec-
tron transfer from the metal atom to the boron, which gives rise to a strong ionic
contribution to the bonding. In the transition metal diborides, the charge transfer
decreases from 2.28 electrons in ScB2 to 1.09 electrons in MnB2 [16]; lower values
have been presented by Samsonov and Kovenskaya [17, 181. The additional elec-
trons occupy the 2p7t orbitals of the boron where the electrons are involved in
both the B-B bonding as well as the metal-boron interactions. Cluster calculations
of main group element diborides show that the metal-metal bonds are weak, the
metal-boron interaction is significant and the boron-boron interactions are very
strong. In the transition metal diborides the metal-metal bonds within the layers
are considerably stronger than in main-group diborides and reaches a maximum
for VB2. This internal bonding within the layers is clearly of a metallic type and is
thus responsible for the metallic transport properties. The metal-metal interlayer
bonding, as well as the metal-boron interactions also increase from ScB2 to
MnB2, whereas the contribution of the boron-boron bonding decreases in this
order. Due to the existence of vacancies in the boron layer and the possible occupa-
tion of interstitial sites by additional boron atoms, the boron sheets may also exhibit
some metallic or semimetallic conductivity. The considered metallic fraction does
not, however, account very much for the transport properties. In contrast, the inter-
action between metal and boron layers contains a more efficient metallic portion,
which explains the electric conductivity along the c-axis [19]. In ideal boron
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.2 Chemical Bonding and Crystal Chemistrv of Borides 809
layers, the donor capability of the metal governs the extent of electron localization
in the sp states of the boron atoms. Thus the covalent character of the B-B bonds
decreases from group IV to group VI metal diborides [20].
As it has been established that the boron network is rather rigid and governs the
lattice expansion in the a direction whereas the lattice dilatation perpendicular to the
metallic layers strongly depends on the metal species, it seems likely that the metal
atoms are distorted in some cases [21, 221. The c / a ratio is thus a function of the
Y ~ ~ ratio
~ and~ depends
~ / furthermore
Y ~ ~on the
~ valency
~ ~ electron concentration
~91.
Since the layers of the A1B2 type structure consist of very distinct kinds of
material, namely metals and boron, exhibiting entirely different binding character-
istics, more attention was recently paid to the (0001) terminating surface layers of
single crystals. Hayami et al. [23] and Souda et af. [24] have studied HfB2, TaB2,
and WB2 single crystals by impact-collision ion scattering spectroscopy at
5 x lo-' Pa inert gas pressure, which gives information on the ultimate surface
structure by atomic shadowing effects of scattered He+ and Lif ion beams at low
incident angles. As a result, the HfB2 basal plane is shown terminated by metal
atoms whereas B is entirely absent even in some layers down further. In contrast,
TaB2 and WB2 basal planes exhibit boron surface layers emerging from boron
diffusion to the surface at above approximately 500°C, and compensating for the
boron defects due to ion bombardment. This effect is attributed to the compara-
tively stable graphite-like boron layers and the higher affinity of Hf to oxygen
compared to Ta and W.
- --
Atom Symbol Site
!
trigonal
OXIS k)
B,C: @ 6h, B(1) !
B,C,Si : 0
B,C?,A1:
stoichiometry [3042]. At that time, the majority of the authors agreed that the
linear chain is of the C-B-C type and does not change with the C content, which
fits the stoichiometry of the most stable compound in the system, B13C2.Accord-
ingly, the linear chain thus contains a closed shell of ten valence electrons, which
is achieved by a charge transfer to the B12icosahedra to which 38 valence electrons
are formally assigned. The charge difference between B and C within the chain
results in even stronger electrostatic binding forces [35], but the energetic differences
estimated by density-of-state calculations are very small [43]. Since there is a
deficiency of one electron in the icosahedron, additional C as an electron donor
preferentially replaces B in one of the icosahedral sites [42, 44, 451. The B2 site
was established as the most favorable position for such a replacement [41, 43, 46-
491. Thus the total structure can be written as (BI1C)-(CBC)+.
To re-investigate the carbon-rich limit of solid solubility, boron carbide being
in eutectic equilibrium with graphite was prepared by arc fusion or diffusion couples
and subsequently analyzed with a microprobe [50] yielding a composition of B4.3C
(i.e., a maximum carbon concentration of 18.87 at.-%). Optical absorption spectros-
copy of this material indicated that 8 1.4% of the linear chains had a C-B-C structure
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.2 Chernicul Bonding and Crystal Chemistry of Boridrs 8 11
Table 1
and 18.6% consisted of a C-B-B array; the chains were statistically distributed [51].
Concluding from Fourier transform infrared (FTIR) spectroscopic data, Kuhlmann
rt al. [51] argue that in contrast to the generally accepted continuous substitution of
B I 2 for B l l C icosahedra and C-B-B for C-B-C chains with decreasing carbon con-
tent, the structure consists of statistical mixtures of these units. With decreasing C
content, a growing portion of unit cells without any central linear chain is formed.
Earlier, the boron-rich side of the homogeneity range was established at 8.7 mo1.-%
C (i.e., B : C = 10.4) by Bouchacourt and Thevenot (1981) [49] who assigned the
structural composition B12_\-C\.(C-B.yCI - ,-C) with 15.33 atoms per unit cell to
the resulting formula B10.4C.The density of boron carbide at the boron-rich end
is with 2.465 g cmP3lower than that at the carbon-rich corner with 2.51 g cmP3.
Together with the assessment of the binary B-C phase diagram the homogeneity
range of boron carbide and the other compounds was modeled by means of the
Compound Energy Formalism using the following sublattice models (Table 1)
where vacancies are denoted as V:
In comparison to Lim and Lukas [52] who still consider the existence of a linear
C-C-C chain, Kasper [53] favors the model taking vacancies in the chain into
account. His calculation reveal a prediction of the occupation state of the icosahe-
dral and chain position (Fig. 7) being in accordance with experimental results on
IR absorption bands and heat capacity.
As already mentioned, the linear chain can accommodate other main group
elements such as Al, Si, P, As, and 0 without a change in the structure type.
Solid solutions formed with (B4C), however, are only known for A1 and Si, which
can partially occupy one of the positions within the C-B-C chain, e.g., A1 is
placed in the centrosymmetric B3 site which causes a slight kinking of this linear
array, whereas Si replaces up to one third of the carbon sites at the ends of the
chain. The total solid solubility of both species is comparatively small (max. 2.5
at.-%) [28, 541, although a complete solid solution series between BI2(C,B, Si),
and the silicon boride “SiB4” was considered [55]. Due to the large size of the Si
atom, however, the formation of a two-atom chain is favored, as in the cases of
P, As, and 0. The isotopic binary borides of these elements thus fit into the general
stoichiometry B12(XVX). This arrangement favors a charge transfer to the BI2
icosahedron. The ideal structure of BI2O2 may be written as (B12)2-(00)2+,
which exactly yields a filled-band configuration. In reality, for all these compounds
a considerable disordering has to be taken into account, which results in the compo-
sitions B12P1,8, B12A~2.0,and B1201.82 (corresponding to B6.60)[25]. In the case of
“SIB4”, the real stoichiometry is (B10,4Sil.6)(SiSi) = SiB2.89,which is attributed to
a Si-Si chain and two sites in the icosahedron being partially occupied by Si
[7, 21, 561. Sublattice structure modeling carried our by Kasper [53] to establish
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
812 7 Boride-Based Hard Materials
XC
Figure 7. Calculated occupation of sites in the icosahedra and the chain of B4C solid solution [53].
the ternary B-C-Si system considered that the Si-Si chain may also at least partially
be the structural unit for the Si-containing boron carbide solid solution [Eq. (l)]:
B4+8C(Si) = (B, C)12(C2B,CB2, B2V, Si2) (1)
The solid solubility limit was then simulated by a statistical combination of all the
kinds of chain occupancies known.
then selected ternary systems which are of technical interest. In the subsequent
sections particular phase systems will be treated in respect to sintering of B4C and
TiB2 or in the context of microstructural design and mechanical strengthening.
2600
2400
- 2200
2
- 2000
a,
L
3
3 1800
L
a,
1600
a,
I-
1400
1200
1000
B 10 20 30 40.0C
(a) -C (atom-%)-
2800
I;)
t 2400
/
-
! ! l 2000
a
E
F 1600 Boron +
boron
carbide
1200
8.6 1 18.81 I I
B 10 20 30 40
(b) Carbon, atom% ---+
Figure 8. The B-C phase diagram (a) calculated, (b) according to Schwetz and Karduck [50].
carbon is given at 2375 f 5°C and 29-31 at.-% C, which is in a good agreement with
thermodynamic calculations [53, 731 stating 29.1 at.-% as the eutectic composition.
The phase diagram is presented in Fig. 8 comparing the calculated version (Fig. 8a)
with the one (Fig. 8b) suggested by Schwetz and Karduck [50]. The reaction of
boron carbide with elemental boron was modeled as a eutectic at 0.2 at.-% C and
2073"C, i.e., 2 K below the melting point of boron.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 8 15
Calculated Experimental
-+-
[78] Lukas [52] [53] [82] Abbaschian [80] [83]
-
SIB, SIB, + liquid 1898 1850 1850 1989 1850 1900
SIB, (Si)sol,dH liquid 1385 1384 - 1385 1385
SiB, SiBb + liquid 1377 - - 1377 -
SiB, + (Si) H liquid 1340 - 1340 -
++ +
SIB3 (Si),,l,, SiB6 - 1270 1198 1270 I358
2200
1200-
- SiB, SiB, + Si
-
1000 , I ' I ' I ' I ' I ' I I I '
Si
X 60Bro [3BlMal
A BOVia X &Via
0 .2 .4 .6 .0 1.0
B Molar Fraction Si
Figure 9. (a) The binary B-Si system according to experimental results with powders of technical
purity. Homogeneity fields of the silicon borides according to Ettmayer et al. [76], Lugscheider
et al. [77], Armas et al. 1981 [82]. (b) The binary B-Si system according to recent calculations by
Lim and Lukas [52].
(AlB2 structure). While TiB and Ti3B4 decompose peritectically at 2180 and
22OO0C, respectively, TiB2 melts congruently at 3225 f 25°C [85]. TiB has a
narrow homogeneity range of about 49-50 at.-% B [86] and reacts eutectically
with Ti solid solution at 1540°C and 7 at.-% B. The existence of the Ti3B4
phase was proven by Fenish [86] and its peritectic reaction with TiB2 and liquid
was placed at 2020°C. Rudy and Windisch [85], however, omitted this phase
from their binary diagram, probably because it could not be observed in melt-
derived samples close to the decomposition point of TiB. In 1981, Ti3B4 was re-
investigated by Neronov et al. [87] in reaction layers between Ti and B; later, in
1986, it was confirmed by Spear et al. [88] by arc melting and annealing
studies. TiB2 reacts eutectically with elemental boron at 2080 f 20°C and
approximately 98 at.-% B. Batzner [89] re-calculated the system with slightly dif-
ferent liquidus lines and omitting the little solid solubilities of TiB and TiB2,
respectively, reported by Murray et al. [84]. The congruent melting point of
TiB2 was set to 3216"C, the eutectic between TiB2 and B to 2059°C instead of
*
2080 20°C after Murray et al.
m a ss % Zr
03050 70 80 90 100
---
0 10 20 30 LO 50 6 0 7 0 80 90 100
B at.% Zr Zr
In the other transition metal boron phase systems of groups 111, IV, and V, the
MB2 phase is also the dominating compound with respect to the high melting
point. Exceptions are the Y-B, lanthanide metal-B and actinide metal-B systems,
which possess very stable MB4- and MB6-type compounds. In the Y-B system
additional higher borides of the stoichiometry YBI2, YBZ5,and YB6, exist. In
group V the number of known phases with a B/M ratio < 2 increases, and the
tendency for the MB phase to be more stable than MB2 is obvious on advancing
from V to Ta. Destabilization of the A1B2 structure to the benefit of the MB struc-
ture is also evident in group VI where the MB2 phase forms an individual structure
type that can be derived from the A1B2 structure by the introduction of variations in
the stacking sequences. In the VIIth and VIIIth group metal-boron systems the
melting points of the borides decrease becoming significantly lower than the melting
points of the elements. A summary of the known boron-containing binary systems is
given by Spear [64].
Mol-% A1
-.
W
m+ @--
; SiB, + Si
w
B I
z
w
W SiB3 + Si
Figure 13. Two possible view of the B-Si system (after [53]).
+
SiB3 H SiB6 Sisolid.Fig. 14 shows a calculated isothermal section of the system
at 2000K according to Lim and Lukas [52].
The ternary B-C-Ti system was intensively studied by Rudy et al. [99] and has been
considered for the fabrication of ceramic cutting tools [loo-1031. The most recent
critical assessment was published by Duschanek et al. [104, 1051 who, in contrast
to the previous authors, take the re-established Ti3B4phase into account. No tern-
ary phases have been discovered. The ternary solid solubility of the particular binary
compounds except TiC0.81B0.17 is generally less than 1 at.-%. TiB2 coexists with
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
822 7 Boride-Based Hard Muteriuls
S Ji
C
(B4C) Molar Fraction C
Figure 14. Isothermal section of the B-C-Si system at 2000 K (calculated by Lim and Lukas [52]).
Calculated Experimental
1380 - 1380
si + SiB6 H SiB, + sic 1345 - 1345
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 823
I I I I
3500 -\
3000 -
2500 -.
s
TiB2+TiC ,.x
2000 -
1500 -
Ti, B, +TIB,+TiC,.,
1000 I I I I
A o TiB,
0.1 0.2
xc
0.3 0.4
TIC
and B by reaction sintering. Composites of T i c , --x and TiB2 have been investigated
for coherent grain boundaries [103]. Furthermore, Tic, --y is stable together
with TiB up to approximately 2160°C (calc.: 2113°C) and then decomposes to
Ti3B4 and liquid. The previously established ternary peritectic at 21 60°C
+ +
(TiB2 T i c l p.y + TiB liquid, [99]) has also to be replaced by three transition
equilibria due to the existence of Ti3B4. Table 4 presents the ternary reactions
accordingly [ 1051.
-
temperature ["C] temperature ["C]
quasibinary eutectic
quasibinary eutectic liquid H TiB2 + B4C 2310 2381
eutectic liquid ++ TiB, + TIC, -, + C 2400 2394
eutectic liquid TiB2 + B4C + C 2240 2246
transition liquid + TiB2 tf Ti3B4+ T i c l 1; 2180 2121
transition liquid + Ti3B4 tf TiB + TIC, -, 2160 21 13
eutectic TiB2+ B4C,,+ B tf liquid 2016 2058
eutectic TiB + T i c , ~~z+ PTi u liquid 1510 1535
transition PTi + T i c , -,tf TiB + aTi 890 927
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
824 7 Boride-Based Hard Materials
1.0 *
pTitTiBtTiC l.x
Figures 16 and 17 show isothermal sections of the Ti-B-C system at 1400°C and
21OO0C, respectively [104, 1051. The diagrams based upon the work of Rudy et af.
[99] are wrong below 2100°C as they neglect the Ti3B4phase. Technically interesting
isopleths are shown in Fig. 18 (TIC-B), Fig. 19 (TiB2-B4C),Fig. 20 (B4C-Ti), and
Fig. 21 (B4C-Tic). The equilibria of the various titanium borides with TIC1-,
have been studied in more detail by Brodkin and Barsoum [106].
The quaternary B-C-Si-Ti system has not yet been established completely.
Evidently TiB2 and Sic form a eutectic and can be cast after arc-melting
[107]. TiB2-BI2(B,C, Si)3-SiC composites are chemically stable and may form a
quasiternary eutectic. Composites of these three hard materials can be prepared
either by direct hot pressing or by reactive infiltration of porous B4C bodies with
an Si-TiSi2 eutectic melt [108, 1091. Of technical interest is the ternary boundary
system C-Si-Ti studied by Brukl [I lo], Borisova et af. [l 1 I], and Holleck [112],
calculated by Touanen et al. [ 1 131 and experimentally re-examined by Wakelkamp
et af. [114].
1.o
1 .o
C
B
Figure 17. Isothermal section of the B-C-Ti system at 2100°C [105].
0.8
0.8
I
Mole fraction C
0.6
0.6
I
X.B
L B
0.4
I
0.2
0.2
I
Tio.5Co.5
Ti
0
I
3500
A
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1 .o
Ti
-
0.8
I
0.6
Ti
I
7 Boride-Based Hard Materials
TiB, +B4C
0.2
A o
2500 -
2000 -
1500 -
.
826
+
Y
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 827
3500 i 7
L+Bd C
Figure 22. Isothermal section of the Fe-Ti-B system at 1050°C (calculated by Golczewski and
Aldinger [320]).
1500
2 1300
E
a,
E
a,
I- 1200
1100
0 20 40 60 80 100
100
7.3 Phase Sq’stems
Figure 25. Isothermal section of the Ti-Ni-B system at 800°C in at.-% (after [125]).
T: Ni,,Ti,B,
80
Ti
0
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
830 7 Boride-Based Hard Materials
I ' ' ' I ' 1 ' I I 1 ' 1 " ' 1 ' 1 I ' I I
20 40 60 80
TiB, a t % Ni Ni
Figure 26. Isopleth of the Ni-TiB, section (after [127]).
react forming various metal borides with a low melting point (approx. 90&1 100°C)
and a suitable wetting behavior.
In contrast to the very convenient Ti-Fe-B system, the Ti-Ni-B phase diagram
contains a congruently melting ternary phase, denoted z, with a composition of
Ni21Ti2B6and a Cr23C6structure (Fig. 25) [125-1271. At 800°C, the z-phase is in
equilibrium with Ni, Ni3B, Ni3Ti, and TiB2. A pseudo-binary eutectic with Ni
exists at 1077 5"C, whereas the relationships in the TiB2-rich corner are more
complicated. An isopleth (Fig. 26) across the line Ni-TiB2 reveals a solid-state equi-
librium below 980°C involving TiB2, z,and Ni3B [127]. A liquid phase forms above
that temperature because of the decomposition of z. Above llOO°C, Ni3B also
decomposes completely in the presence of TiB2, which increases the amount of
liquid phase and thus accelerates the densification.
The Boron-Transition Metal Systems of Group (IV-VI) and Group VIII Elements
Cemented borides with a metallic matrix have also been fabricated successfully from
the ternary transition metal borides of so-called z-, cp- and a-types since these com-
posites can easily be liquid phase sintered with metallic melts.
The z-phase with a general composition of Mi1Mi'B6 or (MI, Mr1)23B6 has been
observed in ternary systems with M' = group VIIIB elements such as Fe, Co, or Ni
and M" = group IVB-VIB elements such as Ti, Zr, Hf, Nb, Ta, Mo or W, and with
M' as the liquid phase [125-1281. It forms in coatings or particulate reinforced
metals as a consequence of the reaction between diborides such as TiB2, ZrB2,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 831
Figure 27. Isothermal section of the B-Co-Mo system at 1000°Cwith indicated field of liquid phase
formation [128].
HfB2, NbB2, and TaB2 and Ni-Co-Cr based alloys. The refractory and extra-
ordinarily hard MB2 phases are hence in equilibrium with a comparatively soft
and brittle ternary compounds.
The ternary systems of these kinds of metals with boron reveal a more complex
structure because of the presence of many other ternary phases denoted to as p-
and o-phases. The stoichiometries of these p- and o-phases are M'MI'B and
M;M"B2, respectively, where M' represents Cr, Mo, Ta, or W and M" holds for
Fe, Co, or Ni and solid solutions thereof. Other ternary phases have the composi-
tion and M;M1'B2\-, e.g., TaNiB2, Mo2FeB4,and Mo3CoB6.As an example, an iso-
thermal section of the B-Co-Mo system is shown in Fig. 27 in which both the t- and
the p-phases are linked with Co as the binder [128]. However, in systems with Fe
replacing Co, a p-phase does not exist. Hence o is in equilibrium with liquid
metal and is thus likely to form a cermet material with Fe like the z-phase
Mo2Fe13B5(Fig. 28). Phase compositions located in the pseudo-binary equilibria
with a metal can easily be pressureless liquid phase sintered at temperatures between
1500°C and 1700°C. Wear-resistant parts have been developed from Mo2FeB2-Fe
cermets with Ni or Cr additives [129-131, 3071. Figure 29 presents an isothermal
section of the Ni-Ta-B system at 950°C [126] with three ternary phases where
only z is in equilibrium with metallic Ni.
Figure 29. Isothermal section of the Ta-Ni-B system at 950°C (after [126]).
Ta
Figure 28. Isothermal section of the Mo-Fe-B system at 1000°C [128].
100
80
Mo
TapNi
60
YoTa
7 Boride-Based Hard Materials
TaNi
A
B
At
40
B
Mo,Fe,
TaNi,TaNi,
,
Ni
J
m-Ni4B3
o-Ni, B,
20
8
p"
"O
\
0
832
Ni
Fe
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 833
[ 132-1 361. Precise experimental data are, however, rare or are not readily available.
Makarenko [137] simply mentions that most of the transition metal diborides of
group IV-VI are fully soluble between 2000 and 3000°C and probably also at
room temperature. As the only limiting factor the difference in atomic radii was con-
sidered which should not exceed 15% like in case of V and Cr, and Hf and Zr [132].
This general statement should be taken with care since low-temperature miscibility
gaps have been observed for (Cr,Nb)B2 and (Cr,Ta)B2 [133, 1381 and other quasi-
binary systems were proved of eutectic character. Complete solid solubility with
an miscibility gap at lower temperatures was also proven for the TiB2-CrB2
system, limited low-temperature boundary solubility and large homogeneity
ranges at high temperatures have been observed for the TiB2-WB2, and CrB2-
WB2 systems which will be discussed in more detail.
Studies in the TiB2-CrB2system are difficult due to the comparatively high vapor
pressure of Cr and the little densification of powder blends at temperatures below
2000°C. Hot-pressing at 2000°C for 3 w 2 0 min yields a homogeneous (Ti,Cr)B2
solid solution [136]. Koval’chenko et al. [139] have also studied the sintering
behavior of pre-reacted powders and the physical properties of a (Ti,Cr)B2 solid solu-
tion containing 20 mo1.-YOCrB2. Klimenko and Shunkowski [ 1401 have investigated
the activated sintering of a “titanium-chromium mixed boride” at 2000 and 2200°C
being obviously a completely single phase. Unfortunately, no X-ray nor chemical ana-
lyses have been reported for the particular annealing steps. In spite of the problems
upon obtaining equilibrium conditions in the overall material chemical and X-ray
analyses by Telle et al. [I361 yielded evidence for the existence of an immiscibility
gap below 2000°C in the TiB2-rich corner. In the temperature-concentration range
between 1500°C and 2000°C and correspondingly 0 to 55mol.-% CrB2 no solid
solutions possessing an intermediate CrB2 concentration were observed whereas
homogeneous (Cr,Ti)B2 particles with CrB2 contents exceeding 55 mo1.-% at
1500°C could be detected in contact with (Ti,Cr)B2 phase of a CrB2 content less
than 2mol.-Yo. Samples with overall concentrations of 35, 75, and 90mol.-% CrB2
have been treated by dilatometry and hot-pressing up to 2200°C to prove the
formation of a liquid phase but no melting could be monitored at temperatures
lower than the melting point of CrB2. This indicates that the quasi-binary system
does not contain a liquidus temperature minimum nor is of eutectic character.
From the equilibrium limits a room temperature solubility of TiB2 in CrB2 to 15-
20mol.-%0 is estimated. The melting points of TiB2 and CrB2 in Fig. 30 have been
adopted from Rudy and Windisch [85] and Liao and Spear [142]. These results
confirm the existence of a continuous solid solubility above 2100°C as stated by
Post et al. [132]. Zdaniewski [I351 also reported on properties of a continuous series
of (Cr,Ti)B2 solid solutions, however, without giving details on the temperature range.
Post et al. [132] have investigated the CrB2-WB2 system at 2 1 0 0 f 100°C but
note that no exact limits of the homogeneity range could be determined. Telegus
and Kuz’ma [ 1431 studied 15 compositions in the quasi-binary CrB2-WB2 section
represented as a part of an isothermal section of the B-Cr-W system at 1500°C.
They report a maximum solid solubility of 6 at.-% Cr in W2B5 corresponding to
3mol.-% CrB2 in WB2 whereas the published section shows 16.7mol.-% CrB2
being dissolved in the W2B5 structure. This work allows an extrapolation of
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
834 7 Boride-Based Hard Materials
3 500
..
0 10 20 30 40 50 60 70 80 90 100
Ti62 Mole - O h CrBz CrBz
Figure 30. The quasi-binary TiB2-CrB2system.
the homogeneity range of (W, Cr)B2 to about 89 mo1.-Yo CrB2 or less at 1500°C.
Moreover, in contrast to Rudy [104, 1441 the binary homogeneity range of the
“W2B5” phase is more extended. In the CrB2 corner Telegus and Kuz’ma report
less than 2at.-% W in solution which could not be confirmed in the materials
investigated by Telle et al. [ 1361who synthesized the system up to 2070°C. Unfortu-
nately, the liquidus equilibria were not established, so the accurate composition of
the eutectic is not determined yet. At temperatures between 1750°C and 2000°C
samples with overall compositions between 40 and 86 m0l.-YOWB2 consist of two
phases, namely (Cr,W)B2 solid solutions with WB2 concentrations up to 38-
40mol.-% and (W, Cr)B2 phases containing always less than 9-14mol.-% CrB2.
These results allow the formulation of the boundaries of the homogeneity fields
2500 -
Liquid 2365
-_____ -- -
-
*.-;:
,,
I
; ; 2200 -----==--------__
- I
,
’
0
-W
L
2000 - 2030250
3
z
c
W
a
E,
1-
1500 - .. (Cr,W)B, + (W,Cr)B,
1000- I , I I I . I I ,
0 10 20 30 40 50 60 70 80 90 100
CrB2 Mole -YOWB, WB2
Figure 31. The quasi-binary CrB2-WB2system.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.3 Phase Systems 835
3500
3225
- 3000
-
0
0
f 2500
,2365
Y
I
2000
+W
1500 / (Ti.W)B, + (W,Ti)B,
.L AIB,-type W2B,-type *
1000
0 1 0 20 30 10 50 6 0 70 80 90 1 0 0
TIB, Mol% WB, WB,
of (Cr, W)B2 and (W, Cr)B2 phases with AlB2 and W2B5 structure, respectively. As
shown in Fig. 31, the CrB2-WB2 system is of a simple eutectic type
(T, = 2030 f 50°C) with a solid solubility of CrB2 in WB2 of 1Omol.-% and WB2
in CrB2 of 37mol.-% at eutectic temperature. The tungsten content in CrB2
decreases strongly with decreasing temperature whereas the chromium content in
the W2B5 structure is almost temperature independent.
The quasibinary TiB2-WB2 system as a part of the Ti-W-B system has been
treated by Telle et al. [136] up to 2250°C. The borders of the (Ti, W)B2 homogeneity
range have been intensively studied between 1500°C and 17OO"C, at 2000°C and
around the quasi-binary eutectic temperature. Thus a huge solid solubility up to
approximately 63mol.-YOWB2 could be established. In the W-rich corner of the
phase system, the TiB2 concentration in (W, Ti)B2 is with 1-2 mo1.-% almost con-
stant up to 1700°C. At higher temperatures the solid solubility of the W2B5-type
structure for TiB2 increases slightly to approximately 3 mo1.-% at the eutectic tem-
perature. A further decrease in TiB2 along the supersolidus equilibrium could be
established up to 2250°C as expected according the phase rules. Taking inaccuracies
in the temperature measurement of the dilatometric experiments into account, TiB2
*
and WB2 react eutectically at T, = 2230 40°C and 90 f 3 mo1.-% TiB2 (Fig. 32).
The solid solubility of WB2 in TiB2 at that temperature is approximately 63 mo1.-%
whereas the solid solution of (W,Ti)2B5 type contains only 3mol.-% TiB2 at the
eutectic equilibrium. The homogeneity range of the (Ti, W)B2 solid solution narrows
significantly with decreasing temperature and is 46-49 mo1.-Yo at 2000°C and 8-
10mol.-% WB2 at 1500°C.
Post et al. [I 321 have treated this system before and found a homogeneity range of
the A1B2 structure type to at least 50mol.-Yo WB2. The preparation of a pure single
phase WB2 of AlB2 type structure has failed. Pastor [134] also states that the WB2
concentration in (Ti, W)B2 may exceed 50mol.-%. Yasinskaya and Groisberg [145]
have examined the interaction of TiB2 with metallic W up to 2700°C. Although no
data are given in the text a maximum solubility of 5 at.-% metallic W can be taken
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
836 7 Boride-Based Hard Materials
from the given pseudo-binary section TiB2-W. The corresponding points are shown
as a quadrangles in Figs. 32 and 33. The isothermal section of the B-Ti-W system
presented by Kuz’ma et af. [146] reveals a maximum of 6-8 mo1.-% WB2 dissolved
in (Ti, W)B2 and 34mol.-% TiB2 in (W, Ti)B2 at 1400°C (Fig. 33). These data are
more or less consistent with the phase diagram presented by Telle et al. [136] and are
thus shown as triangles in Fig. 32. Kosterova and Ordan’yan [147] have retreated
the ternary B-Ti-W at 1400°C in order to specify more accurately the phase equili-
bria in the metal-rich region but have basically found the same situation in the TiB2-
WB2 section as Kuz’ma et al. According to their isothermal section the TiB2 content
in (W,Ti)B2 is considered in the range of 10mol.-%, which is certainly overesti-
mated. Contrary to Kuz’ma et al. [146], Kosterova and Ordan’yan [147] take the
existence of a WB4 phase into account being in equilibrium with a W-rich TiB2
and WB2. The presentation of a metastable B-Ti-W phase diagram at room
temperature by Ariel et al. [148] notes a maximum W content in (Ti, W)B2 of 2
at.-%, which was obviously measured on samples of the TiB2-W section and thus
corresponds to the data of Kosterova and Ordanyan [147]. Unfortunately, the
more recent publication on the formation of A1B2 structure type solid solutions
[135] does not give any data on temperaturexoncentration relations in this
system. A partial re-treatment of the Ti-W-B system by Ahn et af. [149] for the
development of TiB2-W cermets was based on the isothermal section of Kuz’ma
et al. [146] but proved the existence of a distinct WTiB2 phase in the WB-TiB
monoboride section. Hot isostatic pressing of TiB2 with 1-30vol.-% W at 1900°C
resulted in TiB2-WTiB2 composites whereas a higher amount of W yielded
WTiB2-W cermets with little residual TiB2. The WTiB2 possesses P-WB structure
and contains 25.7 at.-% W, 24.8 at.-% Ti, and 49.3 at.-% B which is close to the
stoichiometry 1 : 1 :2. In spite of this, the isothermal section given by Ahn et al.
[149] show an extension of the homogeneity range to WB.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 837
Current EXAFS- and HR-TEM studies in the TiB2-WB2by Pohl et al. [150,151],
Mitra and Telle [ 152, 1531, and Schmalzried and Telle [ 1541 indicate, however, that
annealing of (Ti, W)B2 solid solutions in the TiB2-“W2B5”two phase region at
lower temperatures results not only in the precipitation of ‘‘W2B5’’ particles but,
besides segregation of metallic glasses and elemental boron, also of metastable p-
WB-type phases and spinodal ex-solutions of W-rich A1B2-type phases which
have formerly been erroneously interpreted as “W2B5” [136, 1551. This is attributed
to the much faster diffusion of boron compared to Ti and W [I501 and shows that
equilibrium conditions are difficult to obtain.
The quasi-ternary CrB2-TiB2-WB2 system is of special interest in respect to
anisotropy effects of the thermophysical properties of the diboride solid solutions
and has been studied between 1900°C and 2100°C by Telle et al. [I361 up to now. It
is obvious that homogeneous ternary (Ti, Cr, W)B2 solid solutions of A1B2 type struc-
ture are formed in a wide range of composition. The miscibility gap emerging from the
TiB2-CrB2edge extends to towards the TiB2-WB2edge but disappears at < 80 mo1.-%
CrB2, in the WB2 rich corner a W-rich ternary liquid was observed. It is concluded
that the quasi-ternary system exhibits a ternary eutectic at 1800 f 50°C. Close to the
TiB2-WB2edge, a relatively small amount of chromium stabilizes a (Ti, Cr, W)B2 solid
solution with a high tungsten content exceeding by far that of the quasi-binary com-
pound. Also the extend of TiB2 solid solubility in “W2B5”is increased by the presence
of CrB2. But also in this system, the previously mentioned metastable phases have been
observed which make the interpretation of the entire equilibria difficult.
(i) from the elements by arc melting at 25OO0C, or self-propagating synthesis above
1100°C:
4B + C + B4C (8)
(ii) by chemical vapor deposition reducing boron trichloride in the presence of
carbon in a hydrogen atmosphere:
4 BC13 + 6 H2 + C -+ B4C + 12 HCI (9)
(iii) by pyrolysis of boron trihalides with methane or carbon tetraiodide as carbon
carriers, in high-frequency furnaces:
Density I%]
Figure 34. De-Hoff diagram showing the loss of specific surface area during densification of various
ceramics (after [ 1711).
the specific surface area (driving force for densification) is consumed to a much
greater extent for pore and particle coarsening (Ostwald ripening) than for grain
boundary movement and pore removal. Figure 34 shows a so-called DeHoff
diagram correlating the specific surface area and the fractional density on which
the path of an ideally densifying material is illustrated by the diagonal line. The
lines plotted for A1203, Sic and B4C powders make it obvious that in the first
step of sintering, the surface energy is dissipated very fast due to coarsening [171].
Since a doubling in particle size corresponds to a decrease in the densification
rate by a factor of ten, it is no wonder that densification comes to an end before
pore closure is achieved. As pores are favorably removed by grain boundary move-
ment, it is essential to generate a pore size distribution below a critical size above
which pores are stable or even tend to grow (i.e., the driving force for pore shrinkage
is 5 0) and to induce grain growth at moderate rates so that vacancies may be
suitably removed from the surface of the pores.
Another reason for the poor sinterability is the extraordinarily high vapor
pressure of boron oxides and suboxides. Since boron carbide powders are generally
coated by a B203 layer [172] which quickly reacts to form boric acid, H3B03,in
humid atmosphere, vapor phase reactions are active at higher temperatures, in
particular above 15OO0C, providing a fast transport of boron compounds. Redox
reactions such as
B2°3 vap. + B4Csolid * Bovap. + covap.
may be shifted to the benefit of the left or right side of Eq. (12) depending on the local
chemical potential of the particle surface being defined by the local curvature. The
Gibbs-Thomson equation correlates not only the particular chemical potential p
with the local surface radii but also the local vapor pressure. As a result, B 2 0 3 may
pick up boron from B4C particle areas with convex surface curvatures and move it
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Curhide Ceramics 84 1
as BO to areas of concave curvature where B4C is deposited again due to the lower
vapor pressure compared to the starting point. This reaction releases B203 again,
which starts the process cycle another time. Depending on the oxygen vapor pressure
also other boron suboxides such as B202and B 2 0 may be involved. This mechanism,
denoted as evaporation/reprecipitation in the science of sintering, is known to contri-
bute significantly to grain growth without any shrinkage (e.g., [173]).
In conclusion, the sintering of boron carbide requires (i) oxygen-removing
additives such as graphite, carbon black or organic deoxidation agents such as
formaldehyde resin, (ii) very fine powders of high surface area and therefore high
driving force, (iii) high temperatures to enable grain boundary and volume diffusion,
and, if everything fails, (iv) high pressures.
1
5
Q
0.98 i
s
.-
c.
rn 0.96
s
C
free carbon, wt%
a, 0.94
.->
3
-
2 0.92
S i c and 0-10 mass-% A1 additives to > 85% density. In similarity to the decom-
position of an A14C3 addition, metallic A1 is dispersed very homogeneously by
evaporation and condensation in the still porous ceramic body [180,202]. Starting
from submicron powders, Schwetz et al. [203] prepared composite materials consist-
ing preferentially of 9-10 mass-% Sic and 1-3 mass-% C with 97-99.7% density at
2000-2100°C. Residual porosity was removed completely by a post-HIP (hot iso-
static pressing) treatment at 1950-2050°C. Both C and S i c may also be introduced
in the form of organometallic precursors, e.g., by infiltration of a porous B4C body
with polycarbosilane and phenolic resin, or by coating of boron carbide powders
followed by pressing and subsequent pyrolysis. [107,204]. Bougoin and Thevenot
Figure 36. Microstructure of boron carbide pressurelessly sintered with carbon black (after [ 1971).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 845
0 L- ' 1 1 8 1 8 I
2050 2100 2150 2200
Sintering Temperature, 'C
Figure 37. Grain growth In boron carbide pressurelessly sintered with carbon (after [197]).
’05 1
In lo4 + 35 vol%
m U 25.1 vOI%
I
n
E ----.f- 13.7 VOl%
103
v
u 4.5 vol%
A
.-
c
v) lo2
0
0
.-In
> 10’
loo
0 2 4 6 8 10 12 14
PH
Figure 38. Viscosity as a function of pH and volume fraction of B4C in aqueous suspension (after
[2081).
-sr
a
105
a
E
4
lo4 _C_ As Recvd
6 Months Old
.-I*n
L
lo3 _f_ freshly Dried
0 __o_ freshly Washed
.-
0
In
7 lo2
10'
4 5 6 7 8 9 1 0 1 1
PH
Figure 39. Influence of surface conditions to the viscosity for a 35vol.-% B4C suspension (after
[2081).
450°C. Vacuum sintering between 2125 and 2225°C for 2 h yielded >96% of
theoretical density which was improved to > 99.5% by subsequent hot isostatic
pressing at 2050°C at 200 MPa.
--t- As Recvd
+ Washed
- U 0.2M B(OH),
-
>
-
E
.-
m -20 -
-
c
C
aJ
-40-
a
0
c
m
aJ -60 -
N
-I
L
- s o ! . I ' I ' 1 . ' ' 1 ' 1 . 1
0 2 4 6 8 10 12 14
PH
Figure 40. Zeta potential of B4C (0.01 mass-% suspension) as a function of pH [208]
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
848 7 Boride-Based Hard Materials
70 -
99-100
60 - 97-98
h
A 95-96
CI A
"i*
O \
90-94
-
L 50- o
2 o 80-89
Q, 40- A 70-79
L
v 60-69
3
30-
2! % th. 0.
0- 20-
10 -
for wear parts or shieldings which are not subjected to high stresses because these
materials exhibit a low strength ( C T ~< 300MPa) and a low fracture toughness
(Klc< 3MPa m1I2). Hence, for high densification at reasonable temperatures a
hot-pressing treatment that causes particle rearrangement and plastic flow is
required. Grain boundary sliding, strain-induced twinning, creep and, at a later
stage, bulk diffusion combined with recrystallization were identified as the mechan-
isms of mass transport [211-2131. Densification maps and diffusion diagrams of B-
rich boron carbide and C have been established by Beauvy and Angers [214], and
Bouchacourt et al. [215]. Suitable preconditions are (i) the use of submicron pow-
ders, (ii) temperatures in the range 2100-22OO0C, (iii) pressures of 2 5 4 0 MPa, (iv)
15-20 min hold, and (v) a vacuum or an argon atmosphere. To resist the high pres-
sures at these temperatures and to provide carbon as a sintering aid, the use of boron
nitride-coated, graphite crucibles is favored. In Fig. 41, literature data on the
obtained fractional densities of undoped B4C are related to the particular hot-press-
ing conditions. It is obvious that both high temperatures and high pressures are
required to achieve a density of >95%. Only formation of sinter necks is obtained
at 20 MPa pressure and 2000°C (Fig. 42). However, a strong coarsening has to be
taken into account at higher temperatures and average particle sizes of > 100 pm
in commercial ceramics are not rare.
Similar to pressureless sintering, additives may be used for hot pressing of boron
carbide to reduce temperatures required for grain boundary and bulk diffusion and
to retard grain growth. Figure 43 shows a typical microstructure of a hot-pressed B-
doped material with strain-induced polysynthetic twinning. Suitable dopants are B
[216,217], almost doubling the strength, C [192], Mg, Al, Si, Ti, V, Cr, Fe, Ni, and
Cu [94,183-185,2161. As demonstrated by Telle and Petzow [218], combined B-Si or
B-Si-Ti additions lubricate the grain boundary sliding and prevent coarsening by
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 849
Figure 42. Formation of sinter necks in undoped boron carbide after hot-pressing at 20 MPa pres-
sure and 2000°C.
forming a thin Sic or TiB2 grain boundary phase (Fig. 44), which pins the grain
boundary movement and controls the surface diffusion. Compounds used as addi-
tives are various glasses, alumina, sodium silicate with Mg(N03)2,and Fe203,which
may reduce the hot-pressing temperature down to 1750°C [219]. MgF2, AlF3 [181]
and ethyl silicate [220] are other additives which are active at particle surfaces and
grain boundaries: Hot pressing with 1-5 mass- % of the above-mentioned additives
usually requires a temperature of 1750-1900°C to obtain >95% density. In most
cases, grain size refinement and distributed second phases result in improved
mechanical properties such as strength and fracture toughness.
Figure 43. Hot-pressed B-doped boron carbide with strain-induced polysynthetic twinning.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
850 7 Boride-Based Hard Materials
Figure 45. Fracture surface of pure B4C hot isostatically pressed at 1700°C
growth to 350MPa and less. The fracture toughness as studied by Schwetz et al.
[2 101 by means of the bridge method (indentation and subsequent quasi-three-
point loading) was found to be grain size dependent and ranges between 2.5 and
3 MPa m1/2 at particles sizes < 2 p m and between 3.7 and 4 MP a m1l2 at 10m
mean grain size, which can be attributed to grain size dependent residual stresses
in the microstructure causing crack deflection.
500 2
400:
rn I
1
350 1-
3
9
300
0 1 2 3 4 5 6 7
"1 o
5 Wi-%C
0
0 2 4 6 8 10 12
Mean Grain Size , prn
Figure 48. Fracture toughness (Bridge method) vs. mean grain size and free carbon content (after
[2 101).
Formula -
B4.3C B10.4C
Density g/cm3 2.52 2.465
Composition B, C at.% 81.1, 18.9 91.3, 8.7
Crystal structure Space R3m R3m
group
Lattice constants a, c nm 0.5607, 1.2095 0.5651, 1.2196
Cell volume nm3 0.32894 0.33938
Melting point "C 2380 (eutectic with C) 2250 (liquid + C-rich)
Congruent melting point "C 2450
Thermal expansion 10-'/K 2.6 + 4.5 . T (25-8OO0C)
coefficient CL 4.5 (600°C)
4.6 (25HOO"C)
3.016 + 4.3. T - 9.18. lo-' T~
Thermal conductivity X W1m.K 35 (25°C) 4.0 (20°C) BqC
28 (200°C) 4.7 (200°C)
23 (400°C) 5-7 (400°C)
19 (600°C) 5-6.9 (600°C)
16 (SOO'C) 5-6.7 (80O0C)
Thermal diffusivity CL cm2 s-' 1 .O x lo-' (400 K)
3.0 x lo-' (600 K) 1 x lo-' (600K)
2.1 x (800K) 0.8 x (800K)
Electric resistivity p Rcm 0.1-100
Thermal coeff. of electric
conductivity K-' 3.2 10-~
Seebeck coefficient of
thermoelectric power PVIK 100 (400 K) 220 (400 K)
140 (750 K) 233 (750 K)
196 (1000K) 256 (1000K)
225 (1300K)
Self diffusion coefficient:
boron Cm-* 5.18 x lo-'' (lO0OT)
neutron absorption cross
section barn 400-750 at 0.025 eV
Fracture energy ys mJ 5.45 f0.44
Hardness HKo,, kp/mm2 2950
Strength (4-pt. bending) MPa 450-550
Fracture Toughness MPam'I2 2.5-3 for < 2 pm grain size
(bridge method) 3 . 5 4 for > 2 pm grain size
Young's modulus E GPa 440460 (25°C)
430 (250C)
420 (500°C)
415 (750'C)
400 (lO0OoC)
Porosity dependence of E GPa +
Ep = E, [(l - P)/(1 2.999P)], E, = 460 GPa
Poisson ratio - 0.15
Sound velocity ms-l 14.000
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 855
- 167 -
60
50
40
30
20
I B12(C,6 ,Si)3- single crystal
10 B12(C,6,SiJ3 + Si B,
B,C, +Si
0 i '
4
' " " "
7 2
lo0 i01 i02
Load (N)
Figure 49. Hardness vs. test load (after [71,83]).
46 C/g Ar
_____----
-t
untreated
I
r-.
iIu1
---*---
left untreated at 20°C in air was monitored by Heuberger [230] and reported by Telle
[231,232]. Similar experiments were carried out by Matje and Schwetz [172]. Con-
trolled exposure of Ar-sputtered and implanted powders resulted in a strong,
immediate weight gain which was considerably higher for the treated powder
compared to the untreated (Fig. 50). After approximately 20 h, the untreated and
the Si-sputtered material was stable whereas the implanted powder exhibited a
continuous weight gain with a final oxygen content of almost 4 mass-% compared
to the contamination of 1.6 mass-% in the as-received powder. The studies by Matje
and Schwetz [172] proved a similar parabolic weight-gain in atmosphere of 92%
humidity, whereas the increase in weight was linearly dependent on time in air
of 52% humidity (Fig. 51). The tremendous oxygen pick-up by Si-sputtered and
3t
El
x
B
Figure 51. Weight gain of B4C powders in dry and humid air [172].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 857
Figure 52. Microstructure of a B4C-SiC composite derived from polysilane-coated B4C powders.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
858 7 Boride-Based Hard Materials
are, for the time being, only useful for the fabrication of corrosion and oxidation
resistant parts with, however, comparatively small fracture toughness.
The oxidation behavior of single-phase B4C and B4C-Sic composites in dry and
in humid air was studied by Telle [2311 in more detail depending on the microstruc-
ture. B4C-SIC composites were prepared for this purpose (i) by the conventional
powder route yielding isolated Sic particles in B4C matrix and (ii) by the precursor
route [204] giving Sic layers between the B4C particles with the same volume
fraction. Figure 53 illustrates the weight change for the heating period and the
5 - 0.5 -
.-m
- 1.0 I ~ I I , ~
-E 4 - BJ+SiC lprec.)
Figure 53. Oxidation behavior of B4C and B,C-SiC composites in humid air.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Cerumics 859
U
L
L
2 -
o
.- r -
so&- .
The B bonds decrease only slightly in strength confirming that a boro-silica glass
of changing composition is present. X-ray analysis of oxidized single-phase B4C and
B4C-SiCpowder composites reveals also the existence of H3B03and HB02 after cool-
ing if treated at 700°C and IOOO'C, respectively, whereas no derivatives of boric acid
are found at room temperature after heating to 1200°C except of crystals formed
newly during storage. In contrast, B~C-SiC,,,,,rs,, did not exhibit any indication
of the presence of boric acid.
It may be concluded that in the initial stage the oxidation of all materials is
governed by the build up of a boron oxide layer, which in the case of single-phase
B4C and B4C-SiCpowderimmediately undergoes hydrolysis in humid air yielding
boric acid, which is removed by evaporation and hence is responsible for the subse-
quent weight loss according to the reactions. Only in dry systems a stable liquid
B203 layer may be formed before volatile suboxides are generated at locally low
oxygen partial pressures and temperatures exceeding 1000°C:
Figure 55. SEM micrograph of a reaction hot-pressed Tic-B powder blend. Light: TiB2, dark: B4C.
- 100 I 1
,\"
v
x 8o
.-
4-
In
C
-,-
C
60
H
I
Y
8 LO
a
Q)
.-> 20
4-
+
0
a,
L O
0 200 LOO 600 800 1000 1200 ILOO 1600 1800 2000
Temperature ("C )
Figure 56. X-ray analysis of sintering reactions during hot pressing of B4C-Si-B-TiC-WC-Co
blends. Holding time at the particular temperature step 1 min.
-
900-1 IOO'C
2WC+ 1 3 B W2B5+2B4C
~ (20)
TIC +6B 900-1 100°C
TiB2 + B4C
8 WC + 7 B4C
1000-1 100°C
4W2B5 + 2 B4C + I3 C (22)
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
864 7 Boride-Bused Hard Materials
Figure 57. Microstructure of hot-pressed material. Dark matrix: boron carbide, gray: TiB2-solid
solution, white: W2BS.
1000-1100°C
2TiC+B4C- 2TiB2+3C (23)
> 600°C
4B+C-B4C (24)
These reactions are strongly exothermic and belong to the type of materials synthesis
denoted as “self-propagating high-temperature combustion synthesis”. The standard
Gibbs Free Energy of the reaction (21) is with AGO = -200 kJ/mol B4Cinsensitive to
the MC compound involved and fairly independent of temperature. Furthermore,
W2B5 tends to decompose above 1600°C in the presence of TiB2 forming a solid solu-
nTiB2 -
tion (Ti,W)B2 with increasing solubility of W by increasing temperature:
+ mW2B2> 1600°C (Ti,W,)B2 with n = 1 - m and mmax= 0.63 at 2230°C
(25)
The microstructure of a hot-pressed sample is characterized by a homogeneous
distribution of B4C, W2B5, and (Ti,W)B2 (Fig. 57). The average grain size is 0.8-
1.O pm. The final phase composition consists of 72 v01.-YO B4C, 20 v01.-YO
(Ti,W)B2, and 8vol.-% W2B5. The porosity is less than 3vol.-%. Boron carbide
and titanium diboride particles are of more or less equiaxed morphology whereas
it is a characteristic of the W2B5 phase to grow in an elongated shape. Hot pressing
at temperatures above 1900°C results in an exaggerated grain growth of boron
carbide and the transition metal borides. The average grain size may reach more
than 10pm if fired at 2000°C. Boron carbide grows most probably by transient
liquid film sintering, consuming smaller particles, whereas the transition metal
borides grow by the dissolution of W2B5 in the liquid film and precipitation as
(Ti,W)B2 solid solution. The liquid phase consists possibly of a Si-Co-Ti-W-B
alloy which is consumed by the formation of B12(B,C,Si)3and (Ti,W,Co,Si)B2
solid solution in the areas close to the grain boundaries.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 865
Hardness HV,,
4-pt. bending strength
*
32.1 3.2
830 rt 113
GPa
M Pa
Fracture toughness
ISB 4.2 f 0.4 MPa ml’*
ICL,lON 5.0 f0.8
Young’s modulus 425 zt 10 GPa
Shear modulus +
185 5 GPa
Poisson’s ratio 0.15 zt 0.01 ~
The hardness (HV,,) reaches 32 GPa as a maximum. The maximum bend strength
of 830 MPa is obtained by hot pressing at 1820°C. Compared to single-phase boron
carbide, the strength is increased more than two times. In comparison to other TiB2-
reinforced boron carbide ceramics, the increment in strength is still 150%. Higher
hot-pressing temperatures result, however, in a decrease to 600 MPa due to coarsen-
ing. Fracture toughness was also improved from 2.5 to 4.2-5.0 MPam’/2 by crack
deflection. The properties are summarized in Table 6. Figure 58 shows the micro-
structure of the composite in Fig. 57 failed by dynamic fracture toughness measure-
ment applying the multiple impact test (courtesy of B. Ilschner and R. Zohner).
The microstructure of hot isostatically pressed samples does not differ so much
from that of hot-pressed material. The average grain size is slightly smaller than
that of hot-pressed material, i.e., in the range of 0.8pm if treated at 1600°C and
1.2pm if HIPed at 1900°C. Porosities of <0.55% at 1600°C have been achieved.
The bending strength is further enhanced to 1129 i 85 MPa. The fracture toughness
exhibits a remarkable maximum of 5.2 MPam'I2 at 1700°C HIP temperature. The
entire data set is listed in Table 7.
A further improvement of both hot-pressing cycles and additives, in particular
concerning the transition metal diborides resulted in a microstructural optimization
with in-situ grown W2B5 particles. After annealing, the average of the diameter of
the W2B5 particles was about 4.5-7 pm (Fig. 59).
According to the proceeding growth of W2B5platelets an increase of toughness
with increasing sintering temperature and time was observed. As predicted by the
theory of particle reinforcement of brittle matrices introduced by Faber and
Evans [247] the increasing aspect ratio makes crack deflection a more efficient tough-
ening mechanism (Fig. 60).
In Figs. 61 and 62 the data of high-temperature strength and K,, measurements
are shown. At 1000°C the strength is only half of the room temperature strength.
The toughness at this temperature is still more than 60% of the room temperature
toughness and stays constant for higher temperatures. Since crack deflection
responsible for the relatively high fracture toughness is mainly caused by internal
stresses emerging from the misfit of thermal expansion between matrix phase and
Figure 59. SEM micrograph of annealed B4C-TiB2-W2B5sample. Note the elongated W2B5
particles.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Cerumics 867
6-
E0
a 5-
2E
U
U
U
Y
3
1.0 1!5 2.’0 2
20/7mm,51rin air
500
-a 400
300 _ _ _ _ _ _ _ _ _ _ _ - - _\ _ -4_ _ _ _
+-\*
Ll
f. unreinforced matrix
b” 200 \I\
100
% O { , . . , , , , , , , ,
Figure 63. Microstructure of B4C-TiC sintered with 3 mass-% C. White: TiB2, black: free C (after
~071).
0 1 1 - 1 1
2050 2100 2150 2200 2250
Sintering Temperature, "C
Figure 64. Dependence of the grain size of B4C on free carbon content and sintering temperature
(after [207]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
870 7 Boride-Based Hard Materials
Table 8. Mean strength and toughness of sintered and post-HIPed B,C-TiB2 composites as a func-
tion of sintering temDerature; n.d. = not determined.
a
I
B,C-TIC
L 0 L
0 200 400 600 800
Flexural Strength, MPa
Figure 65. Fracture toughness and strength of B,C-TiC materials (after [207]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Ceramics 871
4 t I*--
1 /
/
F
i
1
where the flexural strength is plotted versus toughness: This plot reveals a general
trend for the strength of boron carbide materials to drop with increasing toughness.
Figure 66 shows a plot of the asymptotic fracture toughness of B4C materials
versus grain size. The fracture resistance data of B4C-TiC materials closely follow
a scatter band, which is also characteristic for B4C doped with amorphous
carbon. KI, starts at a toughness of 2 MPaJm, reflecting the inherent fracture resis-
tance of B4C, and increases up to a shallow maximum of = 4 MPaJm at a grain size
of about 10-15 pm, until KI, gradually drops again. From a microstructural point of
view, this behavior is accompanied by a continuous switch from inter- to trans-
granular fracture. This trend is outlined in Fig. 67, which summarizes the
dependence of flexural strength on grain size for B4C materials doped both with
TIC and amorphous carbon. As in many ceramic materials, the strength data
could be fitted to a grain size dependence, where strength scales with the inverse
square root of grain size, d-'I2.
Since toughening mechanisms by changing the crack path from transgranular to
intergranular fracture depend on both internal stresses and interfacial strength it is
worth while studying the grain boundary structure in B4C-TiB2 materials in more
detail as well as to explore the possibility for microcrack formation. In contrast
to Telle and Petzow [218] who have used transient liquid film hot pressing providing
a relatively strong B4C-MB2 interface, Sigl and Schwetz [254] and Sigl and Kleebe
[255] have exploited the opportunity for crack deflection and microcracking by
weakening the B4C-TiB2 phase boundary by the incorporation of free carbon. It
was suggested that microcracking accounts most for the toughness increment.
Sintering and post-HIP of B4C blends 20 and 40vol.-% TiB2 and 0, 1.2, and
4.8 mass-% free carbon derived from phenolic resins yielded full density samples
which have been examined by SEM and TEM. A SEM micrograph of a 20vol.-
Yn TiB2 composite is shown in Fig. 68. Despite the fact that 1.2 mass-% free
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
>
872 7 Boride-Based Hard Materials
800 7117
(d B4C-Tic
n 0
z 600
Strength/Grain-Size Dependence
of B,C sintered with amorphous C
L
Q, 400
5
L
3 200
X
a,
LL
0 I J - I
0 5 10 15 20
Grain Size, pm
Figure 67. Strength-grain size relation (after [207]).
carbon should be present according to the chemical analysis, only two phases,
i.e., TiB2 (white) and boron carbide (gray), are visible. Particulate carbon
precipitates are obviously absent in B4C-TiB2composites with < 2 mass-% free
carbon. Discrete graphite particles appear in 4.8 mass-%-free carbon composites
Figure 68. SEM micrograph of post-HIPed 1.2% free carbon containing B4C-TiB2 composite.
Arrows indicate circumferential microcracks after cooling from sintering temperature (after [255]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.4 Boron Carbide Crrumics 873
Figure 70. TEM micrograph of C interlayer between B4C and TiBz particles (after [255]). The
inserted diagram presents EELS-analyses of the corresponding points 1-3 of the micrograph.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7 Boride-Based Hard Materials
81 BT 80120
P
7
0
0 2 4 6 8
Relative Carbon Content, Cfre$B,C, wt%
Figure 71. Fracture toughness-free carbon relation (after [255]).
densities of pure TiB2 do not exceed 90% even if heated between 1900 and 2500°C
(e.g., [257,258]). Thus it is nearly impossible to achieve completely dense bodies by
pressureless sintering, as no shrinkage between 1900°C and 2100°C occurs, and
further densification by volume diffusion and plastic flow, is accompanied by
exaggerated grain growth. This behavior was attributed by Coble and Hobbs
[259] and Kislyi and Zaverukha [260] to the competing mechanisms of evapora-
tion-recondensation and volume diffusion exhibiting the same rate of mass trans-
port, whereas gas-transport reactions are favored due to the lower activation
energy. The powder size-dependent sintering behavior of TiB2 was studied by
Kislyi et al. [26 I]. Starting with high-purity submicron-size powder, synthesized
from TiC14 and BC13 in hydrogen in a plasma-arc heater, Baumgartner and Steiger
[262] achieved densities of 98.4-99.4% at 2000-2100°C combined with a com-
paratively fine microstructure (average grain size 1-18pm) due to TIC and T i 0
inclusions. Further heating or a prolonged holding time generate exaggerated
grain growth to 80pm, whereas the density does not improve. This is attributed
to entrapped porosity, which can only be removed by volume diffusion. In contra-
diction to reports of the other authors, no significant weight loss was observed even
after several hours hold at 2250"C, which could be related to active evaporation
reactions. Since the small grain size of 1 pm could be retained up to 2O0O0C, the
porosity was preferentially removed by grain boundary diffusion. Thus, contrary
to carbothermically produced TiB2 powder, high densities have been obtained
below the critical temperature promoting rapid grain growth. This implies that
both the initial particle size and the presence of impurities significantly influence
the densification kinetics. Baik and Becher [263] have studied the effect of oxygen
contamination of submicron TiB2 powders and concluded that in the case of
hot-pressing between 1400 and 1700"C, oxygen promotes grain coarsening by
enhanced evaporation-recondensation of B203. Upon pressureless sintering
between 1700 and 2050°C oxygen remains primarily as titanium oxides and sub-
oxides, which increase the surface diffusivity and thus the pore and particle
coarsening rather than the densification. A maximum total amount of oxygen of
less than 0.5 mass-% in the powder or reducing additives such as carbon is recom-
mended. Sintering is usually carried out under vacuum in a resistance furnace with
a graphite, tantalum or tungsten resistor or in a high-frequency furnace with a
graphite susceptor. Ar or H2 gas atmospheres can also be used. If carbon crucibles
are used, boron nitride diffusion barriers have to be inserted to prevent eutectic
melting of the borides and carbon in the temperature range of 20O&250O0C.
The considerable losses of volatile boron or boride species may be reduced by
powder-bed sintering [ 1341.
The physical properties of diborides have been reviewed in many papers, e.g., by
Clougherty and Pober [269], Samsonov et al. [270], and Castaign and Costa [271].
The reported value depend, however, on purity, final porosity, grain size, and
other factors that are usually not well-documented. Therefore, a comparison of
the data is difficult. Especially the mechanical properties differ very much as the ani-
sotropy of thermal expansion eventually results in microcracking during cooling
from sintering temperature, or at least to high stress concentrations at grain bound-
aries. Hardness and fracture toughness, for instance, vary therefore with grain size
and testing conditions. Table 9 presents a data set for TiB2 and ZrB,. Only data
from polycrystalline materials with densities > 95% and preferentially small grain
size, if available, have been used. Thermal expansion of some diborides have been
measured by high-temperature X-ray analysis by Lonnberg [272] and by Fendler
et al. [250] and Telle [232].
The electric resistivity in the solid solution system TiB2-ZrB2 was studied by
Rahman et al. [273]. Billehaug and 0 y e [274] present a study of several transition
metal diborides for cathode materials in Hall-H&-ould cells and come to the conclu-
sion that TiB2 should be an excellent candidate because of its stability against the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
880 7 Boride-Based Hard Materials
aluminum-kryolithe melt and high thermal and electric conductivity. The sensitivity
to thermal shock, the infiltration of liquid phase along grain boundaries and, finally,
the high price are, however, the limiting factors. In combination with other, cheaper
or properties-adjusting materials, there are excellent chances for this kind of appli-
cation. Ssrlie and 0ye [275] favor today a dense TiB2 coating on graphite to reduce
costs and to gain full advantage of the corrosion resistance and energy saving of
TiB2. In connection with BN and AlN the excellent electric conductivity of TiB2
and the perfect wetting by A1 has been used for the manufacturing of resistance-
heated evaporator boats for the metal deposition on plastic sheets. Data about
the corrosion behavior have been presented by Bannister and Swain [276].
A
-
0 TIC 1,5um K =3.LMPa\liii
o T i c 8,Oym ICO
T 70 -
- TiB 4,Oum
3
aJ
60 -
& 50 -
3 to
!-
-
30 -
.I-
0
a -
W
20
2
2
c-(
10 -
0 c
0 10 20 30 LO 50
TiB and Tic-Content in Sic [vol.-%]
strength of SiC/TiB2 and SiC/ZrB2 composites was found to remain nearly con-
stant at 480MPa up to 1200”C, and is hence superior to that of many sialons
[283,284]. SiC-TiB2 composites have been prepared a in situ synthesis by Ohya
et al. [253] by adding TIC and B powder to Sic. If the dopants are well-balanced,
these incompatible compounds react between 1000 and 1600°C to form isolated
and homogeneously distributed TiB2particles. The reaction is, however, accompa-
nied by a strong expansion, which can be overcome by isothermal soaking at
1500°C and final pressureless sintering at 2200°C for 30 min. The fracture tough-
ness of the >98% dense samples is around 3.7MPa m112 at lOvol.-% TiB2.
A systematic study on the rule of mixtures for the mechanical and electrical prop-
erties of TiB,-TiC-SiC composites was carried out by de Mestral and Thevenot
[290]. They modeled “iso-property” curves in the quasi-ternary phase diagram for
mechanical parameters by fitting 20 independent coefficients of a third-order poly-
nomial developed by Phan-Tan-Luu et al. [291] to the results of experimental test
points. Calculated iso-bend strength curves as well as tests on hot-pressed materials
indicate a maximum of 1 100 MPa close to the TiB2-Tic binary edge of the system
(Fig. 73). The best fracture toughness value of 6.4MPa m112was obtained on the
binary SIC-TIC edge (measured and calculated), on .the binary TiB,-Sic edge and
in the ternary region close to the TiB2 phase (67mol.-% TiB,, and 16.5mol.-%
S i c and Tic, respectively) (Fig. 74). The calculated rule of mixtures could also be
confirmed in the case of hardness measurements.
Sic-based composites with W2B5have been discussed by Telle [232]. The fabrica-
tion method was similar by starting with mechanically alloyed WC-B-containing
S i c powders. Due to the strong tendency of W2B5 to grow anisotropically in a
platelet shape crack deflection and crack bridging was efficiently applied, Several
kinds of diboride additives have been studied by Tanaka and Iyi [292] to reinforce
Sic. Pressureless sintering of P-SiC composites with 15-17 vo1.-YO NbB,, TaB,,
TiB2, and ZrBz at > 2000°C resulted in > 99% density. Surprisingly, the additives
strongly influenced the p-cl transformation of Sic during sintering. TaB2 addition
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.5 Transition Metal Boride Ceramics 883
Ti 8
Si C Ti C
Figure 73. Iso-bend strength curves in SiC-TiC-TiB2composites [290].
clearly stabilized the 3C polytype up to 2200°C whereas ZrB2 addition extended the
stability of the 15R structure. NbB2 doping results in a very smooth transition from
3C to 6H between 2000 and 2200°C. The 4H polytype was only found in TaBz
containing materials at a larger content. Although the grain size and shape of the
S i c matrix is heavily influenced the increment in toughness reaches only 20%.
The highest vaIue was obtained for ISvol.-% TaB2 with KI, = 4.751MPam”~.
Ti B ,
Si C Ti C
layered structure, the formation of solid solutions has been extensively investigated
and used for hardening effects. As an example, the quasi-binary system CrB2-TiB2
exhibits a continuous mutual solid solubility approximately between 2000°C and
2100 f 50°C (Fig. 30) [132, 1361, but there is evidence of a solubility gap below
2000°C where the solubility of TiB2 in CrB, is about 40mol.-% at 1500°C and
the solubility of CrB2 in TiB2 is less than 1 mol.-% below approximately 1800°C.
The presence of CrB2 aids the densification of TiB2 due to its higher diffusion
coefficient. Above 21OO0C, CrB2 containing materials partially melts, which is due
to an almost horizontal solidus line between approximately 40 mol.-% CrB, and
pure CrB,. This fact enables liquid phase sintering of TiB2 but with the risk of
exaggerated grain growth and evaporation of chromium and chromium borides,
since the vapor pressure of Cr is four orders of magnitude higher than that of Ti.
Pre-reacted and hot-pressed materials of that system exhibit a flexural strength of
350-500MPa [139,140].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.5 Transition Metal Boride Ceramics 885
Figure 75. SEM micrograph of large W2B5 precipitates white and thin TiWB, layers in (Ti,W)B,
solid solution matrix.
Figure 76. Aged in-situ reacted (Ti,W)B2-WB2composite with grown WB, particles.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
886 7 Boride-Based Hard Materials
Figure 77. Interlocking grain boundaries bridged by WB2 particles. Note the thin TiWB, stacks in
the host crystals.
1200 I
600 -
200 -
, r l , , , l l l l , I I I ~
Figure 78. Strength-grain size relation of various TiB2-MBZ-Mcomposites (data from [293]).
1500 and 1700°C for 60 to 120min yielding 98-99% relative density. Densification
starts above 980°C due to the formation of a liquid phase in the Ti-B-Co/Ni system
(Fig. 25). At this early stage, a rigid skeleton of TiB2 and WC develops. Due to dis-
solution and reprecipitation, a (Ti,W)B2 solid solution grows on the residual TiB2
particles. Subsequently, crystals of w-phase (see Section 7.8.3) form with compositions
of WCoB or W2NiB2,respectively. Upon cooling, the residual liquid phase crystal-
lizes as C-and Ti-enriched Co3B or Ni3B solid solution. A typical microstructure is
shown in Fig. 79. Sintering at 1700°C for 2 h yields an average particle size of
0.8 m resulting in a flexural strength oh= 600-680 MPa. The KIc of 6.5-7.5 MPa
mIP is mainly attributed to crack deflection because of the weak Co- or Ni-boride
intergranular phases.
Figure 80. SEM micrograph of a (Ti,Zr)B2-(Zr,Ti)02material. Note the core-rim structure of the
dark TiBz particles.
-0
900
Q
E
u
800
5 700
C
2”v, 600
- 500
2
3
400
LL
300
200 I I I I 1
0 10 20 30 40 50
Z r 0 2 [voI.-% ]
Figure 81. Strength of TiB2-Zr02composites (data from [218,300]).
I
remaining tetragonal fraction ranges between 0 and SO%, which could almost be
entirely transformed to monoclinic during machining. Between 20 and 30 mass-%
ZrOz, there is, however, about 15 mass-% untransformable tetragonal Zr02. The
presence of these untransformable tetragonal ZrOz particles without an yttria addi-
12.0
1f I+
11.0 1
10.0
-
E
c
9. 0
8. 0
t
a
3 7. 0
::
Y 6.0
5. 0
4. 0
3. 0
0 10 20 30 40 a
Vol% ZrO,
Figure 82. Facture toughness of TiB,-ZrO, composites (dots: data from [300],squares: data from
P181).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.7 Boride-Zirconia Composites 89 1
1.400 -
a
0
z 1200 -
f
m
-E 1000-
C
VI
m
5
C
800-
m
600 -
0 20 LO 60 80 100
Tetragonal ZrO, content (mass%)
8.5
7.5
B
6.5
4.5
3.5
0 10 20 30 40 50 60
Mass -% ZrO,
time is obvious [239,305]. Composites of comparably fine grain size and homoge-
neous phase distribution, i.e., isolated ZrO2 particles of z 1-2 p size, undergo a con-
tinuous phase redistribution and coarsening if hot-pressed up to 1900°C. Zirconia is
then arranged in clusters with favored Zr02-Zr02 interfaces and has grown to
> 5 pm size. As in the case of unstabilized Zr02, Ti is found in the surface-near
areas of the Z r 0 2 particles and, vice versa, Zr is also present in the outer rims of
the TiB2 particles. Thus, a grain boundary diffusion of both Zr and Ti, and presum-
ably also B and 0 is most likely. Plasma etching makes an intergranular phase
clearly visible. This kind of coalescence resulting in Z r 0 2 clusters incorporated by
a contiguous TiB2 matrix with an optimized number of TiB2-TiB2grain boundaries
is considered the reason for geometrical and mechanical destabilization of the tetra-
gonal phase. The spontaneous t-m transformation is assisted by the large ZrOz grain
size and the radial tensile stresses at the contact to TiB,. Since these interfaces are
the weakest anyway, microcracking occurs here preferentially, and the residual tet-
ragonal Zr02 particles in the cluster transform by autocatalysis.
At higher temperatures, coarsening continues up to 2100°C where the material
decomposes by internal oxidation [239]. Both boron and oxygen are then very
mobile and exhibit a high vapor pressure so that the material is blown up drastically.
Volume changes of more than 200% can be observed. The ceramic then consists of
Ti02, titanium borates, and oxygen-deficient Zr02-Ti02 solid solutions whereas
B 2 0 3 vapor creates round entrapped pores causing the strong bloating. Rupture
was not observed, most probably because the grain boundary phase is already
liquid or at least viscous at this temperature.
The high enthalpy of formation of both Z r 0 2 and transition metal borides can be
used to enhance densification by a chemical driving force starting from, e.g., Ti02
and ZrB, [303,304]:
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
a94 7 Boride-Based Hard Materials
m
Q
E
---
c
F
65
600
._
U
a
l
c
m
400
I L
200 ' I
10 20
I
30
I
40
I I
50
Zr 0 , content (wt%)
Figure 87. Mechanical properties of Ti(C,N)-50 mass-% TiBz-ZrOz composites (after [3lo]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 895
5Ovol.-% ZrOz at 1600°C yields almost dense composites with a flexural strength of
2300MPa and a hardness of HV 1800 (HRA = 93.4). A reaction between zirconia
and tungsten carbide has not been observed but the appearance of a liquid phase
is reported. A similar behavior of the mechanical properties has been observed
for composites of Ti(C,N) containing Y-doped tetragonal zirconia composites
where a significant increase in the strength and hardness with Y-Zr02 content is,
however, accompanied by a modest increment in the toughness. In Ti(C,N)-50
mass-% TiB2-Zr02a decrease in hardness and strength combined with a parabolic
increase in toughness was measured with increasing Z r 0 2 content [310]. The best
performance was achieved by 12% ceria-stabilized ZrOz (Fig. 87). A characteristic
of all the Zr02-reinforced borides is the comparatively high scatter of the mechan-
ical properties measured.
of liquid A1 into a porous body of B4C at temperatures between 700°C and 1200°C.
Since A1 melts at 600°C and exhibits a significant vapor pressure at only slightly
higher temperatures, the equilibrium between 1000 and 1880°C at which liquid A1
is stable with an Al-saturated B12(B,C,A1)3-solidsolution [73] cannot readily be uti-
lized for liquid-phase sintering with small volume fractions of liquid. As shown by
Halverson et al. [206], it is more effective to infiltrate compacted, or presintered,
porous B4C bodies with liquid Al. The resulting material is a metal-reinforced B4C
cermet rather than a liquid-phase sintered B4C ceramic. The wetting behavior is
strongly influenced by oxidation layers formed on the surface of the B4C particles
[312], but can be improved by superheating the melt. Between the melting point of
A1 and approximately 1000°C, wetting angles of 100-150" are observed, which
decrease to reasonable values with prolonged soaking for thousands of hours
[206]. Hence, in that temperature range, only hot-pressing or hot isostatic pressing
result in high-density cermets. Above 1000-1200°C, a suitable wetting behavior is
obtained within minutes of annealing. Due to capillary forces and phase reactions
both densification and adhesion of the metal-ceramic interface are excellent.
During infiltration, reactions of A1 with B4C occur. Below 1200"C, A14BC, A1B2,
A1B12, and AIBI2C2are formed within tens of hours whereas above 1200°C the
generation of A14C3,A1B12,and A1B24C4is more favored [206]. If the composite is
prepared by fast heating, infiltration, and rapid cooling, most of the aluminum
matrix is retained unreacted. The matrix can then be hardened by a subsequent
heat treatment at 800°C for 20 hours due to the precipitation of aluminum carbides
and borides. Since the mechanical properties are determined by the A1 matrix, a K,,
of 5-16 MPa m1I2and a flexural strength of 200-680 MPa can be obtained, depend-
ing on the quality and volume fraction of the metallic binder. The Vickers microhard-
ness of 15.7 GPa for a 3 1 v01.-YOA1 composite is improved by annealing to 19.4 GPa.
A re-investigation of the Al-B-C system with special emphasis on B4C-A1cermets
was published by Pyzik and Beaman [90]. The composites consisted of both isolated
and interconnected B4C particles and A1 matrix and only a small amount of ternary
aluminum carbo-borides by fast firing and a post-sintering heat treatment between
600 and 1050°C. Cermets with a high B4C contiguity had the highest strength of
550 MPa if treated at < 600°C. Then the strengths drops to 400420 MPa and
remains constant at higher heat treatment temperatures while structures with iso-
lated B4C particles show a maximum strength of 420 MPa after a 600°C treatment
with a similar decrease at higher temperatures. As expected, the latter composite
with the continuous Al-matrix phase gives the higher fracture toughness of
11.7 MPa m1/2 after a 600°C annealing for 20 hours compared to 8.5 MPa m1I2
for the contiguous B4C microstructures, being best without any heat treatment
(Fig. 88).
Reactions of B4C with metals have also been employed for the bonding of boron
carbide-based ceramics to metallic substrates. Nishiyama and Umekawa [3131 have
studied the intermediate phases in B4C-Fe couples with A1 and Al/Mo interlayers.
Depending on the treatment temperature and thickness of the foils, B12C2A1and
intermetallic phases are formed that allow reactive brazing with a tensile strength
of 55-60 MPa. In a similar way, porous boron carbide of several B:C ratios was infil-
trated by Si-TiSi2alloys with a 1330°C melting point [94,109,233]. After wetting and
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 897
infiltration, the liquid phase reacts with boron carbide to form TiBz and SIC. There-
fore, these composites were discussed in Section 7.4.4.2.
1500 I
I I
I I
I I
-
1400
L + Ti&
h
0
v -
9
3
+ -
2 1300
1268 'C
a,
Q -
E L + TiB, +a
1200 1183'C -
L + TiB, + y 1167 "C
-
Federov and Kuzma [ 1201 and Shurin and Panarin [ 1211, however, showed evidence
for the existence of a pseudo-binary Fe-TiB2 equilibrium, which is characterized by
a eutectic point at a temperature of 1340°C and 6.3mol.-% TiB2 (Fig. 23). After
Smid and Kny [122] the two phase equilibrium is limited at 800°C by the tie-lines
TiB2-Fe2B and TiB2-Fe (5 at.-% Ti). Ottavi et al. [123,124] finally proved that
not pure Fe is in equilibrium with TiB2 but a Fe solid solution with Ti. The bound-
+
ary tie-line interconnects TiB2 with Fe 0.5 at.-% Ti. Accordingly, a small excess of
boron by result in the formation of FeB2. The first liquid phase in the pseudo-binary
section appears at 1170°C coming from the Fe-FeB2 eutectic, the ternary eutectic is
set to 1240°C (Fig. 24). The contradictions in these experiments have been attributed
by Sigl and Schwetz [116,119,196] to the instability of the system TiB2-Fe in the
presence of C or B4C impurities, which originate from the TiB2 synthesis by the
carbothermic reduction of Ti02-B203or Ti02-B4C mixtures. These impurities
react with Fe and TiB2 and form Fe2B and TIC. Thermodynamic calculations by
Golczewski and Aldinger [320], however, take the y-a transition of iron more
precisely into account and indicate that TiB2 is compatible with Ti-bearing ferrite
only below 900°C (Figs. 89 and 90). The pseudo-binary section TiB2-pure Fe there-
fore contains FeB2 below 1167"C, and between 1 167 and 1268°C solid y- or a-Fe
instead plus liquid. At above 1268"C, the two phase equilibrium TiB2-liquid is
entered (Fig. 89) whereas in the TiB2-Fe (2 at.-% Ti) section FeB2 and y-Fe are
absent (Fig. 90). Although there are still some uncertainties on the phase diagram,
which the synthesis of pure two-phase cermets is based upon, the authors agree that
the presence of oxygen and carbon impurities is detrimental to the wetting behavior
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 899
1600 I
1500
L + Ti&
1400
1300
'-
L + TiB, +a
1100 '
0
I I
5
I I I I I I1
I
10 15
2o
Concentration TiB,, (Mol%)
Figure 90. Isopleth of TiB2-(Fe + 2 at.-% Ti) section (after [320]). Note the difference to Figure 89
concerning the stability of CL 4 y Fe and the appearance of the liquid phase.
and responsible for the presence of the embrittling but hard Fe2B phase, which also
controls the sintering behavior and thus the properties.
After the development of cemented borides with a Fe matrix by Funke et al. [316]
and Funke and Yudkovskii [32 11, this cermet system was recently intensively studied
again by Yuriditsky [115], Sigl and Schwetz [116, 1961, and Jiingling et al. [322,323],
and, in connection with an enhanced European research activity, independently by
Ottavi et al. [123, 1241, Ghetta et al. [117], Pastor et al. [324,325], and Sanchez et al.
[326].
+
Figure 91. SEM micrograph of an 8Ovol.-% TiB, 2Ovol.-% Fe cermet sintered at 1600°C. Note
both the incomplete and complete wetting of TiB, by liquid Fe.
by Ghetta et al. [328] have demonstrated that the spreading of liquid Fe in fact wor-
sens with increasing oxygen content of TiB2 substrates (Fig. 92) whereas powders
with < 0.26 mass-% oxygen show a wetting angle of < 5" after 10 min soaking time.
Generally, transition-metal diborides should be wetted well by liquid metals due
to their missing gap between the valency band and the conduction band whereas
oxides exhibit bad wetting behavior due to their partially ionic and partially
covalent character in bonding [329-33 11. Although an improvement of wetting by
liquid Fe may be obtained by addition of Cr or Ni [332], breaking up the Ti02-
layer is thus considered the key to well-processible TiB2 powders.
80 7 7 - r -1
70
0
a- 60
50
40
30
20
0 500 1000 1500 2000
Time, sec.
Figure 92. Oxygen contamination affecting the wetting behavior of Fe liquid on sintered TiB, at
1300°C (redrawn after [117]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 90 1
Figure 93. SEM micrograph of a TiB2-Fe cermet with a nominal composition of 80voI.-%
TiB, + 20vol.-% Fe. Note the existence of Fe,B in the binder phase (after [119]).
It also appears that the impurities in commercial TiB2 promote the formation of
iron boride, Fe2B. A TiB2-Fe alloy with a typical two-phase microstructure of the
binder, i.e. with both Fe and Fe2B making up the matrix between the TiB2 grains,
is shown in Fig. 93. Fe2B is a hard but fairly brittle compound [333] and its forma-
tion consumes a substantial fraction of the ductile Fe phase such that Fe2B is
believed to embrittle the composite seriously. Avoiding that phase has consequently
been a major goal for developing tough TiBz cermets.
Commercial TiB2 powders are typically produced by the carbothermic reduction
of Ti02 with B2O3 or B4C. Both reactions
Ti02 + B 2 0 3+ 5 C + TiB2,+5 CO (41)
and
2Ti02 + B4C+ 3 C + 2TiB2 + 4CO (42)
are being industrially utilized. A typical analysis of the as-received TiB2 is listed in
Table 10.
Since this powder originates from the B4C-processingroute, i.e. Reaction (41), it
is not surprising that about 90% of the total carbon is present as boron carbide and
only 10% is free carbon. Assigning the oxygen content to particular compounds is
not unambiguous, but it is estimated that approximately 1/2 of the oxygen is present
as Ti02 and B203,respectively. Nitrogen is completely bonded as TIN, as BN is
usually not found. It is therefore concluded that B4C, B2O3, free carbon, and
Ti02, generally make up the critical impurities in the TiB2 powder. More carbon is
introduced to the powder blend when organic binders are added and it should be
noted that the oxygen content increases significantly upon milling [118].
The key questions for a successful liquid phase sintering of TiB2 with iron are,
therefore, (i) how to break up the Ti02-films on the TiB2 grains and (ii) how to
avoid the formation of Fe2B.
+
Sigl and Jiingling [119] have shown that powder compacts of TiB2 20 v0l.-% Fe
heated to 1000°C contained Fe2B and Ti203.Notably, their concentration did not
increase significantly at temperatures up to 1600"C, suggesting that the major part
of the reaction had been completed at that time already. Since Fe2B and Ti203
already appear at low temperature, solid-state reactions are considered responsible
for the formation of this compound. Of particular interest are the reactions of Fe
with the boron-containing impurities, i.e. with B4C and B203.According to Table
11 the following reactions are candidates to develop Fe2B:
B4C+ 8 Fe --+4Fe2B + C (43)
2 B203+ 7 Fe 2Fe2B + 3 FeO
---f
(44)
The change in Gibbs Free Energy, AG, has been calculated for Eqs. (43) and (44)
with the data from Janaf [334]. The results are plotted versus temperature in
Fig. 94. Obviously a reaction between Fe and B203is unlikely in the whole tempera-
ture range of interest, which has been confirmed by corresponding model reactions.
It is, however, well known that Fe and B4C react heavily even at low temperatures
forming Fe2B and free carbon [335]. Due to the favorable properties of Fe2B, this
fact is commercially utilized for depositing thin layers on steel surfaces [336].
As mentioned above, borides can be synthesized by the carbothermic reduction of
B203 with carbon and a metal or a metal oxide [337], e.g. according to Eq. (40).
Since free carbon is present in the as-milled powder mixture and also develops during
Reaction (43), the following process must also be considered a potential source of Fe2B:
+
B203 4Fe +3C t 2Fe2B + 3 CO. (45)
Although the AG, which is plotted as a dashed line in Fig. 94 appears to favor
Reaction (45) above 800"C, it is still considered unlikely. According to Sigl and
Jiingling [119], the following steps rather than the gross Reaction (45) are proposed
to operate instead
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 903
a,
,f -400
-800 L -
600 800 1000 1200 1400 1600
Temperature, "C
Figure 94. The change in Gibb's free energy for reactions yielding Fe2B at 0.1 mbar total pressure.
3COz+3C+6CO
Given the CO/CO2 equilibrium according to Eq. (48), the Gibbs Free Energy for
Reactions (45) and (47) can be calculated and is plotted in Fig. 94 at a total pressure
of 0. I mbar. Though Fe and B would favorably react, CO is unable to reduce B 2 0 3
below 1000°C and therefore the carbothermal formation of Fe2B is operating only
beyond 1000°C. This hypothesis has been confirmed experimentally: While a B203-
Fe-C powder blend with a composition according to Eq. (45) does not change its
phase composition after annealing the mixture at 1000°C for 1 h, the X-ray diffrac-
tion of identical samples heated up to 1600°C shows ample evidence for Fe2B. This
suggests that processes involving liquid iron can play an important role during the
formation of Fe2B as elaborated below.
In the regime of liquid-phase sintering TiB2 is dissolved in the Fe melt according
to
TiB2 {Ti} + 2 {B} (49)
where the brackets { } denote the dissolved state of an element in liquid Fe.
Obviously impurity compounds such as T i0 2 , Ti203, C, B203, TIN, and Fe2B
will also dissolve in the liquid. During cooling some of these compounds reprecipi-
tate from the melt in their original composition, e.g. TIN or T i2 0 3and hence do not
influence the equilibrium concentration of Ti and B in the liquid. Species such as
carbon, Ti02, and B 2 0 3which introduce a lot of {C}, {0}, and {B} but only little
{Ti} into the melt do, however, precipitate as compounds that are richer in Ti than
the previously dissolved species, e.g. as
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
904 7 Boride-Based Hard Materials
Fe Fe,Ti FeTi Ti
-Ti [at.-%] +
Figure 95. Isothermal section of the Fe-Ti-B system at 1000°C. The point indicates the cermet com-
position and the arrow the change in composition if the titanium content in the liquid phase is
decreasing.
Figure 96. Optical micrograph of a TiB2-Fe cermet containing Ti203emerging from powder con-
taminations.
2 T i 0 2 + C 4Ti203 CO + (53)
need to be studied. The Gibbs free energy for Reaction (52) is plotted in Fig. 97 as a
function of temperature at a total pressure of 0.1 mbar. Though carbon should be
able to reduce T i 0 2 at temperatures above 850°C according to the thermodynamic
calculations, a reaction between a thin oxide film and solid carbon is considered
unlikely. Instead, carbon monoxide rather than free carbon is believed to deoxidize
T i 0 2 according to
2Ti02 + CO 4 +
Ti203 C02 - (54)
since the driving forces for Reactions (52) and (53) are identical. A further reduction
7
Y
)r
F
a
5
a
9
LL
B
0
(3
Temperature, "C
Figure 97. The change in Gibb's free energy for the reduction of titanium oxides by carbon monox-
ide at 0.1 mbar total pressure.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
906 7 Boride-Based Hard Materials
of Ti203according to
+
Ti203 CO + 2TiO + C02 (55)
is also supported by thermodynamic arguments (Fig. 97), but CO seems unable to
reduce T i 0 to Ti at a pressure of 0.1 mbar.
Since experiments have shown that TiB2 powders with more than 2 weight-%
oxygen can be sintered to full density [116, 196,3221 and that Ti203 particles preci-
pitate during sintering, it is considered that the free carbon in the as-received TiB2
powder and the carbon being set free from the reaction of Fe with B4C generates
CO, which eventually breaks up the oxide layers on the TiB2 grains. The result
that Ti203appears instead of TiO, as may be inferred from Fig. 97, is attributed
to the small quantity of carbon normally available. Principally, a carbon quantity
sufficient to form either Ti203 or T i 0 should promote the wetting and consequently
the sintering behavior of TiB2 powders. This conclusion is supported by Fig. 96,
which suggests that Ti203 particles hardly disturb the wetting by liquid Fe.
Excess carbon can, however, lead to the formation of Fe2B via the liquid-phase
processes described above. Since small quantities as in the previous case will be
used up during the reduction of Ti02, the carbothermic formation of Fe2B from
Fe and B203according to Eq. (45) cannot occur. Hence it is concluded that Fe2B
originates mainly from the reaction of Fe with B4C and that the gross reaction
B4C + 2 Ti02 + 8 Fe + 4Fe2B + Ti203 + CO 7' (56)
operates at temperatures below 1000°C.Though a good part of Fe2B is due to solid-
phase reactions, it is anticipated that B203 will also give rise to both Fe2B and
Ti203. However, the role of B203is as yet not fully understood.
In conclusion, deoxidizing the thin TiOz films appears to be an absolute pre-
requisite for the sintering of TiB2 with liquid iron. Though carbon monoxide
cannot fully reduce Ti02 at the sintering conditions (0.1 mbar), the formation of
particulate lower oxides such as Ti203 seems to be sufficient for supplying a good
wetting behavior.
1800 4
9 1400 0
2i
3
5 1000 -4
8
a, A-
Q d
600 -8
200 -12
0 100 200 300
Time, min
Figure 98. Densification behavior of model alloys with 30 v01.-% binder with and without Ti addi-
tion (after [340]).
3 1800
.-r 2 1400 $I
-.
s
E
n!
3
2
1000
a,
Q
-1 200
0 50 100 150 200 250 300 350
Time, min
Figure 99. Densification rate of model alloys with 3Ovol.-% binder with and without Ti addition
(after [340]).
the temperature range of 1500-1700°C and reach nearly 99% T.D. The addition of 7
mass-% Ti results in a strong decrease in the density of as-sintered materials, but at
> 1650°Ca density of >95% T.D. and closed porosity are obtained. Further densi-
fication can be achieved by subsequent hot isostatic pressing. Materials with 5, 10,
and 20 v0l.-YOFe-Cr-Ni-binder with 7 mass-% Ti addition nearly approach their
theoretical densities by a post-HIP treatment at 1460°C with a pressure of 100
bar Ar (Table 13).
The spontaneous shrinkage for materials with Fe2B at 1140°Cis explained by the
formation of a liquid phase and a subsequent rearrangement of particles. This con-
clusion is compatible with the eutectic temperature of 1177°Cin the system Fe-Fe2B
--
"1
100 I --I , , , r
Grade F + carbanyllc Fe
d
95-
c
c
90 - i
+20 VOl%
Sintering Temperature, OC
Figure 100. Densification behavior of TiB2 + 15 or 20 v01.-% Fe cermets (after [340]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 909
(57)
- B4C + 2Ti --i 2TiB2 + C (58)
c + c02 + 2 c o c + co*+ 2co (59)
2 TiOz + CO + +
Ti203 C 0 2 2 Ti02 + CO + Ti203+ COz (60)
[341]. In this way Fe2B contributes a lot to the excellent densification behavior of
these materials but it also consumes a major part of the ductile Fe-binder phase.
In materials without Fe2Bthe beginning of shrinkage is delayed. A first maximum
in densification rate appears at 125OoC,about 90 K below the eutectic temperature
of Fe-TiB2. Again, this densification peak is attributed to the formation of a
liquid phase, which coincides with the eutectic Fe-Fe2Ti at 1289°C rather than
with the Fe-TiB2 eutectic [342]. This can be explained by the solution of some Ti
in the binder during heating. As proved by X-ray analysis, the addition of 7 mass-
% Ti fully prevents the formation of Fe2B. Smaller amounts of Ti yield
corresponding quantities of Fe2B. Table 12 compares the particular reactions of
materials with and without Ti additions.
After initial particle rearrangement the densification rate decreases in both
materials. It is accelerating again when a sufficient amount of liquid phase is
formed due to the eutectic Fe-TiB2 at about 1320°C. During cooling the liquid
phase finally solidifies at 1140°Cfor samples with and at 1250°Cfor samples without
Fe2B as proved by exothermic reactions producing the small expansion peak in
Fig. 99. The last expansion peak at 800°C is due to the a 4 y transformation of
Fe. As expected, this peak is much higher for the composite that does not contain
Fe2B.
In conclusion, the formation of FeB2 can be avoided by the control of powder
impurities, sintering atmosphere, and compensation of excess boron and carbon
by titanium addition. The wetting behavior depends strongly on oxide surface
layers and is excellent if surface contamination of TiB2 is entirely removed. The sin-
tering temperature and Fe addition must ensure that a suitable amount of liquid
phase is available at the lowest possible vapor pressure. Since the reducibility of
T i 0 2 and the vapor pressure of liquid Fe determine the atmospheric pressure, the
interval (0.01-0.1) mbar < p < 1 mbar is considered best suited. The optimum
amount of liquid phase is generated at 1450 < T < 1650°C if grain growth is
taken into account as another limiting factor. Accordingly, a feasibility diagram
of suitable sintering conditions can be constructed (Fig. 101) showing the optimum
environmental parameters.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 91 1
10'
1 ' 1 ' 1 ' 1 2
reduction of Ti0,
1oo
-ki grain
growth
E
Y 10' -
lo3
1300 1400 1500 1600 1700 1800
Temperature [ "C]
Figure 101. Optimum parameters for the pressureless sintering of TiBz-Fe cermets limited by grain
growth, evaporation, suboxide formation, and suitable liquid phase content.
post-HIP treatment. As expected, samples with a low binder content are harder than
those with high binder fraction, provided the binder composition remains
unchanged. The observation that materials with Fe-Cr-Ni-Fe2B binders are
harder than materials with Fe-Fe2B binders is attributed to the smaller grain size
of the former materials (Fig. 103). Simultaneously, the fracture toughness increases.
Thus avoiding the formation of Fe2B strongly increases the fracture toughness but
decreases the hardness only moderately, such that the combination of hardness and
fracture toughness is clearly improved [343] compared to WC-Co-based hard
metals.
The bending strength does not vary significantly with increasing binder content
(Fig. 104) but the presence of FeB2 results in considerably lower values. This
not only attributed to the brittleness of this phase but is also due to the residual
2500 1
s> 2000
r
ui
u)
a,
c
E 1500
m
I
conventionalcemented carbi
-
10001' " " " " I ' " ' I ' " ' 1
0 5 10 15 20
Fracture Toughness hC,M P a h
Figure 103. Hardness and fracture toughness of TiB2-Fe cermets in comparison to conventional
hard metals (after [340]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 913
a 1200
a
.- 1000
a
?t 800
600
a,
L 1
400 1 I
3 Fe-Cr-Nt-Fe,B
0 5 10 15 20 25
Binder Content, vol%
Figure 104. Flexural strength dependence on binder volume of TiB2-Fe cermets with and without
FezB (after [343]).
2 % porosity. A post-HIP treatment may overcome this problem and also yields
a lower scattering of data. Additions of metals such as Mo, Cr, Ni, and Co to
the Fe matrix may be used to fabricate composites with improved mechanical and
corrosion properties. Figure 105 shows significant variations in the bending strength
with increasing amounts of Mo in the binder phases of different volume fractions
[115].
Figure 106 illustrates the hot-hardness of composites with various binder content,
again in comparison to WC-Co and WC-Fe-Co-Ni hard metals [343]. It is evident
that TiB2-based materials with a small amount of binder are superior in the entire
temperature range measured.
Figure 105. Variation of strength with Mo content in the binder phase of TiB,-Fe cermets (after
[115]). Binder volume: A, 12.5vol.-%; M, 15vol.-%; 0 , 17.5vol.-%.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
914 7 Boride-Based Hard Materials
25001 I I I I I I I I I I I I I I I I
1 2000
uj-
1500
a,
v)
C
E
ca
T
1000
Figure 107. Optical micrograph of a TiB2-Ni composite, pressurelessly sintered at 1600C. Note the
incomplete wetting of TiBz by liquid Ni and the formation of Ni borides in the matrix phase.
The liquid phase intensifies the mass transport but causes an accelerated grain
growth. The microstructures of composites prepared by liquid-phase sintering are
similar to those of hard metals. The TiB, particles form a rigid skeleton of faceted crys-
tals whereas the binder, e.g., Ni3B, Ni2B, Ni3B4,or comparable compounds of Fe, Cr,
or Co, is the matrix phase. The TiB2 grain size usually exceeds 20pm (Fig. 107).
Depending upon the wetting behavior, which is influenced by the surface oxidation
of the hard material phase, round pores may accumulate at particle/matrix interfaces
or close to triple junctions that have not been completely infiltrated by the liquid phase.
Moreover, the evaporation of Fe-, Co-, or Ni-borides may cause entrapped gas pores.
Hence, hot-pressing is still required for a homogeneous distribution of the liquid phase,
particle rearrangement, and complete removal of the residual porosity. In contrast to
hard metals, the matrix phase is very brittle, e.g., the K,, of Ni3B equals 1.4-1.9 MPa
m"2 [347], and hence does not improve the mechanical properties.
Nishiyama and Umekawa [348] report on sintering and wear application of a TiB,
cermet with Ni7Zr2 binder. Pressureless sintering between 1450 and 2100°C in
hydrogen results in an incomplete melting of the alloy added with wetting angles
of around 90". The phases detected are Ni4B3, Ni7Zr2,and solid solutions thereof
with Ti. Since TiB2 forms a rigid skeleton the hardness HV, reaches 30-33 GPa at
binder contents of < 8 mass-%, dropping almost linearly to 17 GPa with 30 mass-%
Ni7Zr2 addition. Due to the toughness of the matrix phase, the 3-point bending
strength improves from 500-600 MPa at < 5 mass-% binder to 850-900 MPa with
> 15 mass-% of additives.
7.8.2.3 Transition-Metal Diboride Cermets with Co, Ni, Cr, Mo, and W
The research for high-strength/high-toughnesscomposites in the TiB2-transition
metal systems resulted in many attempts to avoid brittle matrix phases and to
introduce hardening particles while making use of the opportunity of liquid-phase
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
916 7 Boride-Based Hard Materials
-
:
x
I
1 0 0 0 - :;9 0 .
0
0
TiB,+lrn%CoB
TtBz+5rn%TaB,+lrn%CoB
TiB +, 5 rn% W B
, st 1m % C o B
2 200-1
m
0 5 10 15 20
A v e r a g e groin size (prn)
Figure 108. Strength-grain size relation in TiB,-COB-based composites (drawn according to data in
[2931).
sintering. Pastor [I341 reports on the manifold of diborides that have been sintered
with Fe, Co, Ni, Cr, Mo, W, and even Re additives, yielding closed porosity or
almost total density. Unfortunately no phase compositions were presented in
most cases. In order to avoid brittle ternary phases, many liquid-phase sintering sys-
tems started with combinations of binder metals, sometimes even with their low
melting boride phases, which will be reviewed in the following paragraph.
Hot-pressed composite materials developed from the more complex systems of
the type TiB2-M1B2-M" with M' being Hf, V, Nb, Ta, Mo, or Mn and MI' being
sintering additives such as Co and Ni, exhibit bending strengths between 850 and
1000 MPa, which are due to the grain growth inhibiting influence of the 1-5 mass-
% of M"B2 particulates (Fig. 108) [293-2961. During liquid-phase sintering in a
Co- or Ni-boride melt, both TiB2 and M' are partially dissolved and reprecipitated
as a solid solution. The effect of grain growth retardation as well as of strength and
hardness increments is attributed to stresses at the TiB2/(Ti,M')B2 phase boundaries
generated by the mismatch of the lattice parameters between the unreacted TiB2
acting as a nucleus and the epitaxially precipitated (Ti,M')B2 solid solution. In
the case of a TiB2-5 mass-% W2B5/TaB2material with 1 mass-% COB binder the
lattice strain ranges between 9 x and 14 x depending on the hot-pressing
temperature [293]. An improved sintering behavior of this material was obtained by
the addition of 1.7% TIC, which reduced the porosity from 0.3-0.7 to 0.1-0.2 vo1.-YO
after hot-pressing at 1500°Cand by intensive ball milling resulting in an average par-
ticle size of 1 pm, but increased oxygen contamination. Watanabe and Shoubu [297]
reported the formation of a (Ti,Ta)(C,O) solid solution, which is considered to initi-
ate the improved densification resulting in a flexural strength of l000MPa. The
strength-grain size correlation in Fig. 108 can be interpreted in the way that the frac-
ture is flaw-controlled at average grain sizes < 7-8 pm and microcrack-controlled at
> 8 pm.
In a similar multiphase system, transition metal carbides were used as additives for
pressureless sintering of TiB2, yielding composites of binary and ternary borides
[218,296]. Attrition milled powder mixtures of TiBz with 3-10 mass-% Co or Ni
and 20-35 mass-% WC have been sintered in a vacuum at temperatures between
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 917
1500 and 1700°C for 60 to 120min yielding 98-99% density. Densification starts
above 980°C due to the formation of a liquid phase in the Ti-B-Co/Ni system
(Fig. 26). At this early stage, a rigid skeleton of TiB2 and WC develops. Due to dis-
solution and reprecipitation, a (Ti,W)B2 solid solution grows on the residual TiB2
particles. Subsequently, crystals of o-phase form with compositions of WCoB or
W2NiB2.Upon cooling, the residual liquid phase crystallizes as C-and Ti-enriched
Co3B or Ni3B solid solution. A typical microstructure is shown in Fig. 109. Sintering
at 1700°C for 2 h yields an average particle size of 0.8 pm, resulting in a flexural
strength ob= 600-680 MPa. The K,, of 6.5-7.5 MPa m1/2is mainly attributed to
crack deflection because of the weak Co- or Ni-boride intergranular phases.
As already shown, an interesting characteristic of the mixed diborides is the so-
called core-rim structure, which is well known from complex cermets based on
(Ti,Ta . . . )(C,N) compounds. During sintering or hot-pressing of the particular
carbonitride powder blends solid solutions are formed at the phase contacts but
the annealing time is insufficiently long to obtain equilibrium conditions. Conse-
quently, interdiffusion is stopped, an unreacted core of, e.g., Ti(C,N) remains
whereas an outer layer, the rim, consists of, e.g. (Ti,Ta)(C,N) solid solution with
a very distinct boundary to the residual host crystal. Exactly the same effect is
observed in borides of all kinds of composition if sintered for a not appropriately
long time (Fig. 1 lo). In case of liquid-phase sintered borides, this layer is formed
by epitaxial precipitation of the thermodynamically more stable solid solution
onto a nucleus of undissolved starting material. Recently, Telle et al. [349] observed
dislocations at the interface of TiB2-(Ti,W)B2solid solutions in hot-pressed TiB2-
WB2 composites (Fig. 11 l), indicating that stresses are generated at the epitaxial
interface as considered by Watanabe and Kouno [293].
Another approach to fabricate metal-matrix-based boride and carbide compo-
sites according to the Lanxide process starts with reactive blends of B4C and Ti
or Zr metal. Upon conversion to TiB2 or ZrB2, respectively, a large release of
heat is observed, which can easily lead to partial melting of the composites. Depend-
ing upon the starting composition, residual metallic Ti or Zr, or B4C may be found
after reaction. Interesting microstructures can also be obtained if TIC or ZrC are
added as fillers [350].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
918 7 Boride-Bused Hurd Material5
Figure 110. Core-rim structure of (Ti,W)B2 solid solution on a TiB, host crystal.
Similar metal-matrix composites with B4C, Sic, and TiB2 as fillers have been
fabricated by the so-called Lanxide- or Dimox-process (direct metal oxidation)
where an Al- or Ti-based liquid mixed with ceramic particles -preferably of whisker
or platelet shape - is slowly converted in air, oxygen, or nitrogen to alumina or tita-
nium nitride, respectively [351-3531. This self-propagating reaction yields columnar
crystals of the oxide or nitride phase, with B4C, Sic, or TiB2 inclusions and residual
metal-filled channels, which contribute significantly to the strength and tou hness.
SiC-A1203compositeshave an excellent fracture toughness of 8-1 5 MPa rn1Ig and a
flexural strength of 500-800 MPa. They have also been demonstrated to be highly
resistant against erosive wear [354].
Mo
Ni
.B
Figure 112. Crystal structure of w-type ternary borides MiM"B2 with M' = Mo and M" = Ni or
Cr. Depending on the Cr content, the orthorhombic structure (a) or the tetragonal structure (b) are
stable (after Ozaki et al., 1994 [358]).
(1 B o r o n s o u r c e powder
0 0 0 Metal powder
T e r n a r y complex b o r i d e
Liquid
i n i t i a l stage
mixed powders
solid state
diffusional
l i q u i d phase
s i n t e r i ng
final stntered
compact (3 Metal m a t r i x
b o r o n i z ing
reaction
Figure 113. Schematic illustration of the "reaction boronizing sintering process" (Takagi, 1993)
[359].
from the M' boride across the contact points. This solid-state diffusion is enhanced
by the concentration gradient in boron and both M' and M" metals as well as by
the chemical driving force for the formation of the ternary M;M"B2 phase. Both
diffusion and heat release result in the growth of sinter necks between the starting
particles and the precipitation of very small ternary boride particles being homo-
geneously arranged around the M" grains. The source powder for boron is used up
completely by this reaction, and, upon further heating, the liquid phase is created
first on sites where the eutectic composition exists. Since the ternary borides have
been precipitated in situ, no oxide layer is present, preventing the wetting. During
spreading of the melt particle rearrangement by capillary forces is active and
causes shrinkage and a more homogeneous distribution of all phases. Due to the
higher temperature and the enhanced diffusion via the liquid phase, the residual
metal powder dissolves in the melt, filling the residual pores and allowing grain
growth of the stable ternary borides. Finally, after cooling, the cermet consists
of the hard boride phase dispersed in a metallic matrix phase or with a binder
phase at the triple points, depending on the volume fraction of binder. The micro-
structure resembles that of ordinary hard metals based on WC-Co with some pecu-
liarities, however, which will be discussed later.
To enable the solid-state reaction at the beginning and to control wetting at higher
temperatures, oxide impurities on the starting metal powders have to be removed.
This can be performed by reducing agents such as carbon or by evaporation of
boron oxides and suboxides during vacuum sintering. A hydrogen atmosphere
can also be used for sintering. Pressureless sintering is carried out slightly above
the quasi-binary eutectic temperature, usually between 1200 and 1350°C. The sinter-
ing time is comparatively short, e.g. 20-30 min to prevent grain growth by dissolu-
tion and reprecipitation.
A detailed study of the sintering phenomena of a Mo2FeB2-(Fe-alloy)cermet is
presented by Sivaraman et al. [361] who focused on the commercially available
KH-C5O material by Toyo Kohan Co., Ltd. consisting of 42.71 mass-% Mo,
39.28% Fe, 10.20% Cr, 4.96% B (as FeB), and 2.85% Ni. Sintering was followed
by DTA, TG, dilatometry, and X-ray diffraction; the microstructure was observed
after quenching from particular heating steps. As a result, the sintering reactions
could easily be monitored by both shrinkage effects and exo- and endothermic
effects, The first solid-state reaction is the formation of Fe2B at 486°C followed
by the precipitation of tetragonal Mo2FeB2, the reduction of metal oxides, the
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
922 7 Boride-Based Hard Materials
formation of the liquid phase, and the incorporation of Cr into the Mo2FeB2.Table
14 summarizes the reaction steps and the resulting intermediate microstructures.
The sintering as monitored by dilatometry is illustrated in Fig. 114 [361]. Evi-
dently, the solid-state reaction to form Fe2B as well as the initial state sintering
by neck growth and particle center-approach yield a small contribution to shrink-
age. The most dramatic effect is, however, the generation of a liquid phase at
above 1150°C which yields a shrinkage rate of 6%/min, i.e., after 5-10min the
final density of 99.6% of theoretical density is achieved. This fast sintering can be
attributed to the sudden rearrangement of the boride particles of less than 1 pm
in size. At temperatures exceeding 1200-1220°C only coarsening is observed by dis-
solution and reprecipitation mechanisms but the particle size remains around 1 pm.
The change in particle size in Mo2NiB2cermets with Cr-Ni binder was investigated
in more detail by Matsuo et al. [362]. They started from MOBas a boron source to
compare the influence of Cr addition on structural changes. Table 15gives the
composition of the boride phase aimed at whereas the binder content was kept
constant at 25 mass-% (Ni f 10% Cr).
Sintering at 1260-1300°C for 20min revealed that MOBwas consumed totally by
the formation of Mo2NiB2, which exhibited an orthorhombic crystal structure,
while the addition of Cr yielded Moz(Ni,Cr)B2with a tetragonal structure. The micro-
structures consist of blocky faceted crystals of somewhat elongated habit in all cases.
It is, however, evident that Mo2NiB2tends to coarsen strongly (mean particle size
(M 8-10 pm) compared to Moz(Ni,Cr)B2 (mean particle size (M 3-5 pm), Fig. 1 15.
Another peculiarity is that the first phase forms euhedral crystals with sharp edges
and corners whereas the latter phase shows rounded edges and corners under exactly
the same conditions. This observation is attributed to the higher surface energy ani-
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Boridees 923
Boride composition B Mo Ni Cr
obtained clearly indicate that some of the elements have no effect on strength and
hardness at all, namely, Fe and Co, whereas Ti, Mn, Zr, Nb, and W yielded even
lower strength than the cermet without additives. Cr and V, however, caused
more than 40% increase in strength with a slightly higher hardness. The
addition of V gives rise to a steep increase in both strength and hardness if only
2.5 mass-% are added (Figs. 116-1 18).
It was proved that all boride solid solutions except the Cr- and V-containing
one crystallize in the orthorhombic structure type. As mentioned before, the
lower anisotropy of the surface energy of the tetragonal phase is the reason for
the formation of more spherical particles. The inhibition of grain growth was
accordingly found in Cr- and V-containing cermets but also in Nb- and W-bearing
materials (Fig. 119). In the latter cases, however, the microstructure is very inhomo-
geneous, which may be the reason for the comparatively low strength. In case of Co
and Mn addition the particles have grown considerably, which again explains the
low strength since no other brittle phase was detected. If Ti and Zr are added,
third phases have been observed but could not be identified.
A variation of the Cr content similar to the studies of the influence of V on the
properties was carried out by Komai et al. [356].Also in this case, an improvement
of both strength and hardness was achieved as a function of the additive amount.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 925
Figure 116. Strength of Mo2NiB2-based cermets with various additives (after [365]).
7 p
92
2.5
2 2.0
-
2
tn
.
ln
0 88 g
E
2+ 1.5 I
86
c
- a4
1.0
Figure 118. Transverse rupture strength (TRS) and hardness of V-containing Mo2NiB2 (after [365]).
Figure 119. Back-scattered SEM micrographs of a Ni-6 mass-%, B-58.6 mass-%, Mo-10 mass-%
metal cermet. Note the change in grain size and shape due to the metal additive (after [365]).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 927
50-2. 92
90
88
2
-86 \
In
-a4 g
[z
I
82
-
80
0.501 I 78
0 5 10 15 20 25
Cr content (mass%)
Figure 120. Transverse rupture strength and hardness as a function of C r addition to Mo2NiB,
(after [356]).
and 1993. The tests were aimed at (i) screening the cutting properties of TiB2-Fe and
(ii) providing useful information for further development of these materials.
The properties of TiB2 cermets suggest that workpiece materials that demand
tools of the lower P, M, and K qualities would be most adequate for cutting
tests. Thus a medium carbon steel Ck 45, a globular cast iron GGG 50, and a
hypo-eutectic cast G-AlSil2CuMgNi alloy were selected by Sigl et al. [366] for
continuous turning experiments having a Brine11 hardness of <210, <220, and
< 115, respectively.
Two series of TiB2 cutting inserts with ferritic iron binders and austenitic steel
binders (nominal composition 18 weight-% Cr, 10 weight-% Ni, bal. Fe) were
tested. Due to a Ti deficiency, the ferritic TiB2 grades with a nominal binder fraction
of 15 v01.-YO contained approximately 1Ovol.-% cc-Fe and 5vol.-% Fe2B. Ti was
added to prevent the formation of Fe2B in the steel cutting grades, such that the
second series of TiB2 cermets could be prepared with 10vol.-YOof fully austenitic
binder. Uncoated ferrite-binder cermets were used to machine the A1 and cast
iron alloys, whereas the Ck 45 steel was treated with both uncoated and TiN-
coated (3 pm CVD-TIN layer) inserts with austenite binder. Both composites
represent “generation-I” TiB2 cermets, in the sense that they feature a much coarser
grain size (on the order of 5 pm) than the composites available today. The tests were
carried out in comparison to conventional cutting tool materials. Thus the cast iron
grade was machined with a commercial alumina-zirconia material (10 weight-%
Zr02), while the carbon steel and the hypo-eutectic A1-Si alloy were treated with
a commercial Ti(C,N)-Ni cermet. Important mechanical properties of the cutting
tool materials are compared in Table 16. The actually tested combinations of work-
piece and cutting tool materials are compiled in Table 17.
Continuous turning tests were performed. The SNGN 120408 tool geometry and
parameters summarized in Table 18. Cutting was interrupted after a cutting length
of 500m for the GGG 50 and after l000m for the Ck 45 and G-AlSil2 alloys,
TiB2-a(Fe)-Fe2B X X
TiB2-y( Fe-Cr-Ni) X X
A1203-Zr02 X
Ti(C,N)-Ni X X
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.8 Cemented Borides 929
respectively. The flank wear, wB, the depth of cut notch depth, 6,the surface rough-
ness of the workpiece material, the cutting forces, and the power consumption of the
machine were monitored at these intervals. The test was performed until the failure
criterion or a total cutting length of 8000 m was reached.
In all cutting materials the flank wear and the cutting forces increased con-
tinuously up to the ultimate lifetime. With very few exceptions of cutting edge
fracture, all inserts faded due to flank wear. Failure due to excess depth-of-cut
notch depth was not observed. Cratering at the rake face did occur, but was in no
instance the reason for insert breakdown. The surface roughness of the workpiece
materials was rated acceptable in all cases. Thus a flank wear, wB of 0.4mm was
defined as lifetime criterion. The cutting results using this criterion are summarized
in Figs. 121-124 for the various combinations of tool and workpiece materials.
Ranking the cutting tool materials, TiB2 with a ferritic binder is superior to the
Ti(C,N) cermet and the alumina-zirconia ceramic for light cutting operations such
as finishing of GGG 50 and G-AlSil2. However, increased thermal loads favor cut-
ting materials with enhanced high-temperature endurance, i.e., the Ti(C,N) cermets
and the alumina-zirconia ceramics. Figures 121- 124 illustrate the details of these
general observations: TiB2 cermets outperform commercial Ti(C,N) tool materials
100
.-c 10
E
0.1
I000
100
10
1
10 100 1000
Cutting Speed v, m/min
in A1 cutting over the whole range of cutting speeds up to 700 m/min, and they work
well at medium cutting speeds up to 200 m/min during finishing operations of gray
cast iron. Tool life decreases dramatically, however, upon increased cutting loads,
i.e., for light roughing of GGG 50 and for steel finishing operations (Figs. 123,
124). In particular, the Taylor slope of TiB2-Fe in cast iron cutting is much steeper
than for alumina-zirconia tools (Figs. 122, 123). A similar, yet less pronounced differ-
ence in Taylor slopes is observed between Ti(C,N) and TiB2 steel cutting grades
(Fig. 124). The TIN coating significantly improves the life of TiB2 tools, but it
cannot fill the gap of the exceptional quality of Ti(C,N) cermets.
100
10
10 100 1000
10 100 1000
Cutting Speed v,, m/min
Figure 124. Life-time diagram for finishing Ck 45
An optimization of the microstructure towards fine grain size of TiB2 and binders
with sufficient high-temperature strength will further improve the cutting capacity of
TiB2 cermets specifically for steel machining operations. The major disadvantage
that has prevented the production of TiB2-Fe composites on a commercial scale is
their delicate processing behavior. Thus the cost-performance relation is still the cri-
tical point for the current market conditions.
Figure 125. Injection molding machine parts composed of MolFeB2-based alloys: (a) cylinder; (b)
auger; (c) check valve; (d) seal ring; (e) auger head.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.9 Future Prospects and Fields of Application 933
compete with all kinds of hard metals and cermets with respect to performance
and costs.
8000
6000
-
N
E
E
\
-.-
m
Y
2 1000
Ln
In
a,
C
?
I
2000
0
0 LOO 800 1200
Temperature 1°C)
Figure 126. High-temperature hardness of ceramic materials in comparison to diamond and cubic
boron nitride.
machining aluminum alloys, whereas ternary boride cermets have been successfully
developed for injection molding nozzles, bearings, wire drawing cones, and other
wear resistant parts that are not exposed to high temperatures. Composites of
TiB2/TiC, TiB2/TiN, TiB2/ZrB2/TaN, as well as B4C/TiB2, B4C/SiC, and B4C/
SiC/TiB2 have been exploited as cutting tools for brass, bronze, and A1 alloys,
drilling tools for rocks, and concrete due to their comparatively high toughness.
An obvious conclusion from the various efforts in strengthening and toughening
of boride-based materials is that the maximum strength obtained by crack-
propagation influencing mechanisms operational in composites is about 1-
1.2 GPa and the fracture toughness usually obtained in optimized microstructures
is about 6-7 MPa m1I2.It is also clear that simple two-phase composites are gener-
ally superior to materials of more complex composition. Metal-matrix composites
may exhibit a much higher toughness than ceramics but are limited in high-
temperature use due to their sensitivity towards creep and corrosion. With the
possible exception of the very promising TiB2-Fe-based and ternary boride-based
cermets, borides are thus currently of more interest for strengthening metals similar
to oxide-dispersed superalloys. A certain challenge is still the transformation tough-
ening of borides by Y-stabilized Zr02, increasing the toughness by a factor of two
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
7.9 Future Prospects and Fields of Application 935
demand for future key technologies, there is a chance that these materials may
increase in importance.
References
1. Kiessling, R. Acta Chem. Scand. 1950, 4, 209.
2. Lundstrom, T. Arkiv Kemi 1969, 31(19), 227.
3. Anderson, S. and Lundstrom, T. Acta Chem. Scand. 1968, 12, 3103.
4. Vlasse, M., Slack, G. H., Garbauskas, M., Kasper, J. S., and Viala, J. C. J . Solid State Chem.
1986, 63, 31.
5. Richards, S. M. and Kasper, J. S. Acta Crystallogr. 1969, B25, 237.
6. Higashi, I., Kobayashi, K., Tanaka, T., and Ishizawa, Y. J. Solid Stute Chem. 1997, 133, 16.
7. Matkovich, V. I. and Economy, J. in Boron and Refractory Borides, Matkovich, V. I. (Ed.),
Springer, Berlin, 1977, pp. 77-94.
8. Aronsson, B., Lundstrom, T., and Rundqvist, S. Borides, Silicides and Phosphides: A Critical
Review of Their Preparation, Properties and Crystal Chemistry, John Wiley & Sons Inc., New
York, 1965.
9. Aronsson, B., Lundstrom, T., and Engstrom, I. in Anisotropy in Single Crystal Refractory
Compounds, Plenum Press, New York, 1968, Vol. 1, pp. 3-22.
10. Lundstrom, T. Arkiv Kemi 1969,31(19), 227.
11. Higashi, I. and Takahashi, Y. J . Less-Common Met. 1986, 123, 277.
12. Okada, S., Atoda, T., Higashi, I., and Takahashi, Y., J . Mater. Sci. 1987, 22, 2993.
13. Lundstrom, T. Arkiv Kemi 1968, 30(11), 115.
14. Romans, P. A. and Krug, M. P. Acta Crystallogr. 1966, 20, 313.
15. Woods, H. P., Wawner, F. E., and Fox, B. G. Science 1966, 151, 75.
16. Armstrong, D. R. in Proc. 9th Int. Symp. on Boron, Borides, and Related Compounds, Werheit,
H. (Ed.), University Press, Duisburg, 1987, pp. 125-131.
17. Samsonov, G. V. and Kovenskaya, B. A. in Boron and Refractory Borides, Matkovich, V. I.
(Ed.), Springer, Berlin, 1977, pp. 5-18.
18. Samsonov, G. V. and Kovenskaya, B. A. in Boron and Refractory Borides, Matkovich, V. I.
(Ed.), Springer, Berlin, 1977, pp. 19-30.
19. Aravamudham, R. Z. Metallkd. 1967, 58, 179.
20. Samsonov, G. V., Goryachev, Yu. M., and Kovenskaya, B. A. Izv. Vuz Fiz. 1972, 6 , 37.
21. Hoard, J. L. and Hughes, R. E. in The Chemistry of Boron and Its Compounds, Muetterties,
E. L. (Ed.), Wiley, New York, 1967, pp. 25-154.
22. Spear, K. E. J. Less-Common Met. 1976, 47, 195.
23. Hayami, W., Souda, R., Aizawa, T., and Tanaka, T. Surf. Sci. 1998,415, 433.
24. Souda, R., Asari, E., Kawanowa, H., Otani, S., and Tanaka, T. Surf. Sci. 1998,414(1-2), 77.
25. La Placa, S. and Post, B., Planseeberichtefiir Pulvermetallurgie 1961, 9, 109.
26. Lipp, A. Ber. Deutsche. Keram. Ges. 1966, 43(1), 60.
27. Lipp, A. and Roder, M. Z . Anorg. Allg. Chem. 1966, 343, 1.
28. Lipp, A. and Roder, M. Z . Anorg. Allg. Chem. 1966, 344, 225.
29. Matkovich, V. I. and Economy, J. in Boron and Refractory Borides, Matkovich, V. I. (Ed.),
Springer, Berlin, 1977, pp. 98-106.
30. Clark, H. K. and Hoard, J. L. J. Am. Ceram. Soc. 1943,65, 2115.
31. Allen, R. D. J. Am. Chem. Soc. 1953,753582.
32. Neidhard, H., Mattes, R., and Becher, H. J. Acta Crystallogr. 1970, 26B, 315.
33. Ploog, K. J . Less-Common Met. 1974, 25, 115.
34. Will, G. and Kossobutski, K. H. J . Less-Common Met. 1976, 47, 43.
35. Kirfel, A., Gupta, A,, and Will, G. Acta Crystallogr. 1979, B35, 1052.
36. Werheit, H. and de Groot, K. Phys. Status Solidi 1980, 97, 229.
37. Conard, J., Bouchacourt, M., and Thevenot, F. J . Less-Common Met. 1986, 117, 51.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 937
38. Tallant, D. R., Aselage, T. L., and Campbell, A. N. Phys. Rev. B 1989, 40, 5649.
39. Aselage, T. L., Emin, D., and Wood, C. in Trans. 6th Symp. Space Nuclear Power, Albuquer-
que, NM, January 8-12, 1989, pp. 430433.
40. Aselage, T. L., Tallant, D. R., Gieske, J. H., Van Deusen, S. B., and Tissot, R. G. in The
Physics and Chemistry of Carbides, Nitrides, and Borides, Freer, R. (Ed.), NATO AS1 Ser.
E, Vol. 185, Kluwer, Dordrecht, The Netherlands, 1990, pp. 97-1 12.
41. Aselage, T. L. and Emin, D. in Boron-Rich Solids, Proc. 10th Int. Symp. on Boron, Borides, and
Related Compounds, Emin, D., Aselage, T. L., Switendick, A. C., Morosin, B., and Beckel, C. L.
(Eds), AIP Conf. Proc. 231, American Institute of Physics, New York, 1991, pp. 177-185.
42. Morosin, B., Aselage, T. L., and Emin, D. in Boron-Rich Solids, Proc. 10th Int. Symp. on
Boron, Borides, and Related Compounds, Emin, D., Aselage, T. L., Switendick, A. C., Mor-
osin, B., and Beckel, C. L (Eds), AIP Conf. Proc. 231, American Institute of Physics, New
York, 1991, pp. 193-196.
43. Kleinman, L. in Boron-Rich Solids, Proc. 10th Int. Symp. on Boron, Borides, and Related
Compounds, Emin, D., Aselage, T. L., Switendick, A. C., Morosin, B., and Beckel, C. L.
(Eds), AIP Conf. Proc. 231, American Institute of Physics, New York, 1991, pp. 13-20.
44. Armstrong, D. R., Bolland, J. and Perkins, P. G. Acta Crystallogr. 1983, 39, 324.
45. Emin, D. in Boron-Rich Solids, Emin, D., Aselage, T, Beckel, C. L., Howard, I. A,, and Wood,
C. (Eds), AIP Conf. Proc, American Institute of Physics, New York, 1986, pp. 189-205.
46. Silver, A. H., Bray, P. J. J . Chem. Phys. 1959, 31, 247.
47. Hynes, T. V., Alexander, M. N. J . Chem. Phys. 1974,54, 5296.
48. Becher, H. J., Thevenot, F. Z. Anorg. Allg. Chem. 1974, 410, 274.
49. (a) Bouchacourt, M. and Thevenot, F. J . Less-Common Met. 1981,82,219. (b) Bouchacourt,
M. and Thevenot, F. J . Less-Common Met. 1981, 82, 227.
50. Schwetz, K . A. and Karduck, P., J . Less-Common Metals 1991, 175, I .
51. Kuhlmann, U., Werheit, H., and Schwetz, K. A. J . Alloj's Compd. 1992, 189, 249.
52. Lim, S.-K. and Lukas, H.-L., in Hochleistungskeramiken Herstellung, AU!fbau,Eigenschaften,
-
Petzow, G., Tobolski, J., and Telle, R. (Eds), VCH, Weinheim, and Deutsche
Forschungsgemeinschaft, 1996, pp. 605 -616.
53. Kasper, B. Dr. rer. nut. Thesis, University of Stuttgart, 1996.
54. Telle, R. in The Physics and Chemistry of Carbides. Nitrides, and Borides, Proc. Freer, R. (Ed.),
NATO AS1 Ser. E, Vol. 185, Kluwer, Dordrecht, The Netherlands, 1990, pp. 249-268.
55. Meerson, G. A., Kiparisov, S. S., and Gurevich, M. A. Sov. Powder Metall. Met. Ceram. 1966,
5, 223.
56. Magnusson, B. and Brosset, C. Acta Chem. Scand. 1962, 16,449.
57. Hansen, M. Constitution of Binary Alloys, McGraw-Hill, New York, 1958.
58. Elliott, R. P. Constitution qf Binary Alloys, 1st suppl., McGraw-Hill, New York, 1965.
59. Shunk, F. A. Constitution of Binary Alloys, 2nd suppl., Mc-Graw-Hill, New York, 1969.
60. Mofatt, W. G. The Handbook of Binary Phase Diagrams, General Electric Co., Schenectady,
NY, 1976.
61. Mofatt, W. G. The Index to the Binary Phase Collections, General Electric Co., Schenectady,
NY, 1979.
62. Binary Alloy Phase Diagrams, Massalski, T. B. (Ed.), ASM Int., Materials Park, OH, 1990,
Vol. 1: Alloys, Vol. 2: Phase Diagrams.
63. Ternary Alloys, Petzow, G., Effenberg, G. (Eds), VCH, Weinheim, 1988ff, Vols. 1 ff.
64. Spear, K. E. in Application of Phase Diagrams in Metallurgy and Ceramics, National Bureau of
Standards, Gaithersburgh, 1977, Special Publication SP-496, pp. 744-762.
65. Lundstrom, T. in Boron undRefractory Borides, Matkovich, V. I. (Ed.), Springer, Berlin, 1977,
pp. 351-376.
66. Spear, K. E. in Phase Diagrams, Materials Science and Technology, Vol. IV, Alper, A. M.
(Ed.), Academic Press, New York, 1976, Ch. 1 I , pp. 91-159.
67. Guillermet, A. F. and Grimvall, G. Phys. Rev. B 1989, 40, 1521.
68. Grimvall, G. and Guillermet, A. F. Boron-Rich Solids, Proc. 10th Int. Symp. on Boron,
Borides, and Related Compounds, Emin, D., Aselage, T. L., Switendick, A. C., Morosin,
B., and Beckel, C. L. (Eds), AIP Conf. Proc. 231, American Institute of Physics, New York,
1991, pp. 423-430.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
938 7 Boride-Based Hard Materials
69. Samsonov, G. V. and Schuravlev, N. N. Fiz. Met. Metullovad. Akad. Nauk, S S S R Ural Filial 3
1956, 109.
70. Schuravlev, N. N. and Makarenko, G. N. Ed. Acad. Sci. USSR,OTN, Metallurgia i Toplivo I ,
1961, 133.
71. Bouchacourt, M. and Thevenot, F. J. Less-Common Met. 1979,67, 327.
72. Beauvy, M. in Conf. Abstract, 8th Int. Symp. on Boron, Borides, Carbides and Related Com-
pounds, Tsagareishvili, G. V. (Ed.), Acad. Sci. Georg. S.S.R., Tbilisi, 1984, p. 25.
73. Lukas, H. L. in Constitution of Ternary Alloys 3, Petzow, G. and Effenberg, G. (Eds), VCH,
Weinheim, 1990, pp. 140-146.
74. Ekbom, L. B. Sci. Ceram. 1977, 9, 183.
75. Lange, D. and Holleck, H., Proc. 11th Int. Plansee-Conf., Planseewerke, Reutte, 1985, HM 50,
741.
76. Ettmayer, P., Horn, H. C., and Schwetz, K. A. Microchim. Acta, 1970, Suppl. IV, 87.
77. Lugscheider, E., Reimann, H., and Quaddakers, W.-J., Ber. Deutsche Keram. Ges. 1979,
65(10), 301.
78. Dorner, P. Dr. rer. nut. Thesis, University of Stuttgart, 1982.
79. Emin, D. in The Physics and Chemistry of Carbides, Nitrides and Borides, Freer, R. (Ed.),
Kluwer, Dordrecht, The Netherlands, 1990, pp. 691-704.
80. Olesinski, R. W. and Abbaschian, G. J. Bull. Alloy Phase Diagrams 1984, 5, 478.
81. Mitra, I., Biittner, D., Telle, R., Babushkin, O., and Lindback, T. in Proc. 8th CIMTEC,
Ceramics Charting the Future, Advances in Science and Technology, Vol. 3B, Vincenzini, P.
(Ed.), Techna Srl., Milano, 1995, pp. 787-792.
82. Armas, B., Male, G., and Salanoubat, D. J . Less-Common Metals 1981, 82, 245.
83. Telle, R. Dr. rer. nut. Thesis, University of Stuttgart, 1985.
84. Murray, J. L., Liao, P. K., and Spear, K. E. Bull. Alloy Phase Diagrams 1986, 7, 550.
85. Rudy, E. and Windisch, S. in Ternary Phase Equilibria in Transition Metal-Boron-Carbon-
Silicon Systems, Part I, Vol. VII, Technical Report No. AFML-TR-65-2, Wright Patterson
Air Force Base, OH, 1966.
86. Fenish, R. G . Trans. A I M E 1966, 236, 804.
87. Neronov, V. A,, Korchagin, M. A,, and Aleksandrov, V. V. J . Less-Common Met. 1981, 82,
125.
88. Spear, K. E., McDowell, P., and McMahon, F. J . Am. Ceram. Soc. 1986,69, C4.
89. Batzner, C. Dr.rer.nat. Thesis, University of Stuttgart, 1994.
90. Pyzik, A. J. and Beaman, D. R. J . Am. Ceram. Soc. 1995, 78, 305.
91. Kieffer, R., Gugel, E., Leimer, G., and Ettmayer, P. Ber. Deutsche. Keram. Ges. 1972,49(2), 44.
92. Mogilevsky, P., Gutmanas, E. Y., Gotman, I., and Telle, R. J. Eur. Ceram. Soc. 1995, 15, 527.
93. Lukas, H. L. Final Report, Deutsche Forschungsgemeinschaft Project, Contract No. Lu 283/
2-2, 1992.
94. Telle, R. and Petzow, G. in High Tech Ceramics, Vinzencini, P. (Ed.), Material Science Mono-
graphs, Elsevier, Amsterdam, 1987, pp. 961-973.
95. Werheit, H., Kuhlmann, U., Laux, M., and Telle, R. J . Alloys and Compounds 1994, 209, 181.
96. Secrist, D. R. J. Am. Ceram. Soc. 1964, 47, 127.
97. Shaffer, P. T. B. Muter. Res. Soc. Bull. 1969, 4, 213.
98. Nishiyama, K., Mitra, I., Momozawa, N., Watanabe, T., Abe, M., and Telle, R. J . Jpn. Res.
Inst. Muter. Technol. 1997, 15, 292.
99. Rudy, E. and Windisch, S. Ternary Phase Equilibria in Transition Metal-Boron-Carbon-Silicon
Systems, Part 11. Related Tenary System Vol. VIII, TI-B-C System, Technical Report No.
AFML-TR-65-2, Wright Patterson Air Force Base, OH 1966.
100. Nowotny, H., Benesovsky, F., and Brukl, C. Monatsh. Chem. 1961,92,403.
101. Ordan’yan, S. S. and Unrod, V. V. Sov. Powder Metall. Met. Ceram. 1975, 14, 729.
102. Holleck, H. Planseeber. Pulvermetall. 1982, 2, 849.
103. Holleck, H., Leiste, H., and Schneider, W. in High Tech Ceramics, Proc. 6th CIMTEC,
Vincenzini, P. (Ed.), Elsevier, Amsterdam, 1987, pp. 2609-2622.
104. Rudy, E. Ternary Phase Equilibria in Transition Metal-Boron-Carbon-Silicon Systems, Vol. V,
Compendium of Diagram Data, US Atomic Energy Comm. Publ. AFML-TR-65-2, Wright
Patterson Air Force Base, OH, 1969.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 939
105. Duschanek, H. and Rogl, P., Lukas H. L. J . Phase Equilibria 1995, 16, 46.
106. Brodkin, D. and Barsoum, M.W. J . Am. Ceram. Soc. 1996,79, 785.
107. Telle, R., Brook, R. J., and Petzow, G. J . Hard Muter. 1991, 2, 79.
108. Taffner, U., Telle, R., and SchHfer, U. Z. Prukt. Metallogr. 1988, 21, 17.
109. Klingler, H., Franz, E.-D., and Telle, R. Spreclisaal 1995, 128, 11.
1 10. Brukl, C. E. Ternary Phase Equilibria in Transition Metal-Boron-Carbon-Silicon Systems Part
11; Ternary Systems, Wright, Patterson Air Force Base, OH, 1965, Vol. VII, Tech. Rep. No.
AFML-TR-65-2.
11 1. Borisova, A. L., Borisov, Yu. S., Polyanin, B. A., Shvedova, L. K., Kalinovskii, V. R., and
Gorbatov, I. N. Sov. Powder Metall. Met. Ceram. 1986, 25, 769.
112. Holleck, H. in Proc. 12th Planseeseminar ofPoivder Metallurgy 3, Bildstein, H., Ortner, H. M.
(Eds), Planseewerke, Reutte, 1989. pp. C3, 1-12.
113. Touanen, M., Teyssandier, F., and Ducarroir, M. in Proc. 7th Eur. Con6 on Chemical Vupour
Deposition, Ducarroir, M., Bernard, C., and Vandenbulcke, L. (Eds), J . Phys. 1989, 50, 105.
114. Wakelkamp, W. J. J., van Loo, F. J. J., and Metselaar, R. J . Eur. Ceram. Soc. 1991, 8, 135.
115. Yuriditsky, B. Y. Refractory Muter. Hard Muter. 1990, 3, 32.
116. Sigl, L. S. and Schwetz, K. A. in Boron-Rich Solids, Proc. 10th Int. Symp. Boron, Borides, and
Related Compounds, Emin, D., Aselage, T L., Switendick, A. C, Morosin, B., and Beekel, C. L.
(Eds), AIP Conf. Proc. 231, American Institute of Physics, New York, 1991, pp. 468472.
117. Ghetta, V., Gayraud, N., and Eustathopoulos, N. Solid State Phenomena 1992, 25/26, 105.
118. Jiingling, T., Oberacker, R., Thiimmler, F., Sigl, L. S., and Schwetz, K. A. Powder Metall. Int.
1991, 23(5), 296.
119. Sigl, L. S. and Jiingling, T. J . Hard Muter. 1992, 3, 39.
120. Federov. T. F. and Kuzma, Y. B. Izvest. Akad. Nuuk S S S R - Neorg. Muter. 1967, 3, 1489.
121. Shurin, A. K. and Panarin, V. E. Izvest. Akad. Nauk. S S S R Metally 1974. 4, 235.
-
122. Smid, I. and Kny, E. I n f . J . Refractory Met. Hard Muter. 1988, 8, 135.
123. Ottavi, L., Chaix, J. M., Allibert, C., and Pastor, H. Solid State Phenomena 1992,25/26, 543.
124. Ottavi, L., Saint-Jours, C., Valignant, N., and Allibert, C. Z. Metullkd. 1992, 83, 80.
125. Schobel, J. D. and Stadelmaier, H. H. M~~tallwissenschuft Tech. 1965, 19, 715.
126. Lugscheider, E., Reimann, H., and Pdnkert, R. Z. Metullkd 1980, 71, 654.
127. Lugscheider, E., Reimann, H., and Pankert, E. Metall. 1982, 36, 247.
128. Haschke, H., Nowotny, H., and Benesovsky, F. Monatsh. Chem. 1966, 97, 1459.
129. Takagi, K., Ohira, S., and Ide, T. in Modern Developments in Powder Metallurgy, Vol. 16:
Ferrous and Non-ferrous Materials, Aqua, E. N., and Whitman, C.I. (Eds), MPI & APMI
Publ., Princeton, NJ, pp. 153-166.
130. Takagi, K. et al. Int. J . Powder Met. 1987, 23, 157.
131. Takagi, K., Ohira, S., Ide, T., Watanabe, T., and Kondo, Y. Met. Powder Rep. 1987,42,483.
132. Post, B., Glaser, F. W, and Moskowitz. D. Acta Mefall. 1954, 2, 20.
133. Kuz’ma, Y. B., Telegus, V. S., and Kovdlyk, D. A. Poroshkovuya Mefallurgiya 1969,5(77), 79;
Sov. Powder Metall. Met. Ceram. 1969, 4, 403.
134. Pastor, H. in Boron and Refractorj, Borides, Matkovich, V.I. (Ed.), Springer, Berlin, 1977,
pp. 457493.
135. Zdaniewski, W. A. J . Am. Ceram. Soc. 1987, 70, 793.
136. Telle, R., Fendler, E., and Petzow, G. J . Hard Muter. 1992, 3, 21 1.
137. Makarenko, G. N. in Boron and Rcrf).actoryBorides, Matkovich, V. I. (Ed.), Springer, Berlin,
1977, pp. 310-330.
138. Kuz’ma. Y . B., Telegus, V. S., and Marko, M. A. Poroshkovaya Metallurgiya 1972,4(66), 215
139. Koval’chenko, M. S., Ochkas, L. F., and Vinokurov, V. B. J . Less Common M e t . 1979,67,297.
140. Kliinenko, W. N. and Shunkowski, G. L. Planseeher. Pulverrnetall. 1981, 2, 405.
141. Okada, S., Satoh, M., and Agata. T. Nippon Kagaku Kuishi 1985, 685.
142. Liao, P. K. and Spear, K. E. Binar?, Alloy Phase Diagrams Vol. I , ASM Internat., Metals Park,
OH, 2nd Ed. 1990,471474.
143. Telegus, V. S. and Kuz’ma, Y. B. J . Sov. Powder Metallurgy and M e t . Ceram., 1968, 2, 133.
144. Rudy, E. and Windisch, S. Ternary Phase Equilibria in Transition Metal-Boron-Carbon-
Silicon Systems, Part I, Related Binary Systems, Technical Report No. AFML-TR-65-2,
Wright Patterson Air Force Base. OH, 1966.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
940 7 Boride-Based Hard Materials
145. Yasinskaya, G. A. and Groisberg, M. S. Sov. Powder Metall. Met. Ceram. 1963, 2, 457.
146. Kuz’ma, Yu. B., Svarichevskaya, S. I., and Telegus, V. S. Soviet Powder Metall. Met. Ceram.
1971, 10,478.
147. Kosterova, N. V. and Ordanyan, S. S. Zvest. Akad. Nauk U S S R Neorg. Muter. 1977, 13,
-
1411.
148. Ariel, E, Barta, J. and Niedzwiedz, S. J. Less-Common Met. 1970, 29, 199.
149. Ahn, D.-G., Kawasaki, A,, and Watanabe, R. Muter. Trans. JIM 1996, 37, 1078.
150. Pohl, A., Kitzler, P., Telle, R., and Aldinger, F. Z . Metallkunde 1994, 85, 658.
151. Pohl, A., Telle, R., and Petzow, G. 2. Metallkunde 1995, 86, 148.
152. Mitra, I. and Telle, R. in Werkstoflwoche 1996: Symposium 7: Materialwissenschaftliche
Grundlagen, Aldinger, F. and Mughrabi, H. (Eds), DGB Informationsgesellschaft, Oberursel,
1996, pp. 521-525.
153. Mitra, I. and Telle, R. J. Solid State Chem. 1997, 133, 25.
154. Schmalzried, C. and Telle, R. in Proc. Werkstoffwoche 98:\Band 7: KerumiklSimulation Kera-
mik, Heinrich, J., Ziegler, G., Hermel, W., and Riedel H. (Eds), Wiley-VCH, Weinheim, 1999,
pp. 377-382.
155. Telle, R. in Materials Science and Technology, Cahn, R. W., Haasen, P., and Kramer, E. J.
(Eds), Verlag Chemie, Weinheim, 1993, Vol. 11, Ch. 4, pp. 175-266.
156. Joly, A. C. R. Acad. Sci. 1883, 97, 456.
157. Moisson, H., C. R. Acad. Sci. 1894, 118, 556.
158. Ridgeway, R. R. Trans. Electrochem. Soc. 1934,66,117; Norton Co. US Patent I8972I4, 1933.
159. Lipp, A. Tech. Rundschau 1965, 57(14), 5; 57(28), 19; 57(33), 5.
160. Lipp, A. Tech. Rundschau 1966, 58(7), 3.
161. Dufek, G., Wruss, W., Vendl, A,, and Kieffer, R. Planseeber. Pulvermetall. 1976, 24, 280.
162. Gray, E. G. Eur. Patent Appl. 1 IS2 428, 1980.
163. Schwetz, K. A. and Lipp, A. in Ullmanns Encyclopedia of Industrial Chemistry, VCH, Wein-
heim, 1985, A4, pp. 295-307.
164. Thevenot, F. J . Eur. Ceram. Soc. 1990,6, 205.
165. Thevenot, F. in The Physics and Chemistry of Carbides, Nitrides, and Borides, Freer, R. (Ed.),
NATO AS1 Series E, Vol. 185, Kluwer, Dordrecht, The Netherlands, 1990, pp. 87-96.
166. McKinnon, I. M. and Reuben, B. G. J . Electrochem. Soc. 1975, 122, 806.
167. Knudsen, A. K. Adv. Ceram. 1987, 21, 237.
168. Walker, B. E., Rice, R. W., Becher, P. F., Bender, B. A,, and Coblenz, W. S. Ceram. Bull. 1983,
62, 916.
169. DeHoff, R. T., Rummel, R. A., LaBuff, H. P., and Rhines, F. N. in Modern Development in
Powder Metallurgy, Hausner, H. H. (Ed.), Plenum, New York, 1966, pp. 31&330.
170. Greskovich, C. and Rosolowski, J. H. J. Am. Ceram. Soc. 1976, 59, 336.
171. Prochazka, S. WhyIs It DifJult to Sinter Covalent Substances? General Electric Corp. Res. and
Dev. Center Technical Information Series, Report No. 89CRD025, Schenectady, NY, 1989.
172. Matje, P. and Schwetz, K. A. in Proc. 2nd Int. Conj: on Ceramic Powder Processing Science,
Hausner, H., Messing, G. L., and Hirano, S. (Eds), Deutsche Keramische Gesellschaft, Koln,
1989, pp. 377-384.
173. Shaw, N. J. Powder Met. Int. 1989, 21, 16; 31.
174. Grabchuk, B. L. and Kislyi, P. S. Sov. Powder Metall. Met. Ceram. 1976, 15, 675.
175. Adlassnig, K. Plunseeber. Pulvermetall. 1958, 6, 92.
176. Grabchuk, B. L. and Kislyi, P. S. Poroshkovaya Metallurgiya 1974, 8, 11.
177. Katz, J. D., Blake, R. D., and Petrovich, J. Met. Powder Rep. 1988, 43, 835.
178. Dole, S. L. and Prochazka, S. in Ceramic Engineering and Science Proc. 6(7/8), Smothers,
W. J. (Ed.), American Ceramic Society, Westerville, OH, 1985, pp. 1151-1160.
179. Kislyi, P. S. and Grabchuk, B. L. in Proc. 4th Eur. Symp. Powder Metallurgy, SOC.Franqaise
Metall., Grenoble, 1975, p. 10-2-1.
180. Kriegesmann, J. Ger. Patent DE 37 I I 871 C2, 1989.
181. Lange, R. G., Munir, Z. A., and Holt, J. B. Muter. Sci. Res. 1980, 13, 311.
182. Kanno, Y., Kawase, K., and Nakano, K. J. Ceram. Soc. Jpn. 1987, 95, 1137.
183. Glasson, D. R. and Jones, J. A. J. Appl. Chem. 1969, 19(5), 125.
184. Janes, S. and Nixdorf, J. Ber. Deutsche. Keram. Ges. 1966, 43, 136.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 941
185. Stibbs, D., Brown, C. G., and Thompson, R. US Patent 3 749 571, 1973.
186. Zakhariev, Z. and Radev, D. J . Muter. Sci. Lett. 1988, 7, 695.
187. Prochazka, S. U S . Patent 4005235, 1977.
188. Kriegesmann, J. in Teclznische Keramische Werkstoffe, 4th Suppl., Kriegesmann, J. (Ed.),
Deutscher Wirtschaftsdienst, Koln, 1991.
189. Schwetz, K. A. and Vogt, G. Ger. Patent 2 751 998, 1977; U S Patent 4 195 066, 1980.
190. Henney, J. W. and Jones, J. W. S. Br. Patent Application 20141934, 1978.
191. Suzuki, H. and Hase, T. in Proc. Conf. Factors in Densification of Oxide and Nonoxide
Ceramics, Somiya, S. and Saito, S. (Eds), Japan, 1979, p. 345.
192. Schwetz, K. A. and Grellner, W. J . Less-Common Met. 1981, 82, 37.
193. Dole, S. L., Prochazka, S., and Doremus, R. H. J . Am. Ceram. Soc. 1989, 72, 958.
194. Bougoin. M., Thevenot, F., Dubois, F., and Fantozzi, G. J . Less-Common Met. 1985, 114,
257.
195. Matje, P., internal report, Elektroschmelzwerk Kempten GmbH, Germany, 1990.
196. Sigl, L. S. and Schwetz, K. A. Powder Metall. Int. 1991, 23, 221.
197. Sigl, L. S. and Schwetz, K. A. Euro-Ceramics II 1991, 1, 517.
198. Grabchuk, B. L. and Kislyi, P. S. Poroshkovaya Metallurgiya 1975, 7(151), 27. Engl. Transl.:
Sov. Powder Metall. Met. Ceram. 1975, 14, 538.
199. Oh, J. H., Orr, K. K., Lee, C. K., Kim, D. K., and Lee, J. K. J . Korean Ceram. Soc. 1985,22,
60.
200. Weaver, G. Q. U S Patent 4320204, 1982.
201. Weaver, G. Q. UK Patent GB 2093481 A , 1982.
202. Borchert, W. and Kerler, A. R. Metall 1975, 29, 993.
203. Schwetz, K. A,, Reinmuth, K., and Lipp, A. Sprechsaal 1983, 116, 1063.
204. Lorcher, R., Strecker, K., Riedel, R., Telle, R., and Petzow, G. in Solid State Phenomena 8&9,
Proc. Int. Conf: Sintering of Multiphase Metal and Ceramic Systems, Upadhyaya, G . S. (Ed.),
Sci-Tech Publications, Vaduz, Gower Publ. Co., Brookfield, VT, 1990, pp. 479492.
205. Bougoin, M. and Thevenot, F. J . Muter. Sci. 1987, 22, 109.
206. Halverson, D. C., Pyzik, A. J., Aksay, I. A,, and Snowden, W. E. J . Am. Ceram. Soc. 1989,72,
775.
207. Sigl, L. J . Europ. Ceram. Soc. 1998, 18(11), 1521.
208. Williams, P. D. and Hawn, D. C. J . Am. Ceram. Soc. 1991, 74, 1614.
209. Pyzik, A. J., Aksay, A,, and Sarikaya, M. Muter. Sci. Res. 1987, 21, 45.
210. Schwetz, K. A., Sigl, L. S., and Pfau, L. J . Solid State Chem. 1997, 133, 68.
21 1. Kuzenkova. P. S., Kislyi, P. S., Grabchuk, B. L., and Bodnaruk, N. I. J . Less-Common Met.
1979, 67, 217.
212. Ostapenko, I. T., Slezov, V. V., Tarasov, R. V, Kartsev, N. F., and Podtykan, V. P. Porosh-
kovaya Metallurgiya 1979,197,38. Engl. Transl.: Sov. Powder Met. Met. Ceram. 1979,19,312.
213. Brodhag, C., Bouchacourt, M., and Thevenot, F. in Ceramic Powders, Material Science
Monographs 16, Vincenzini, P. (Ed.), Elsevier, Amsterdam, 1983, pp. 881-890.
214. Beauvy, M. and Angers, R. Sci. Ceram. 1980, 10, 279.
215. Bouchacourt, M., Brodhag, C., and Thevenot, F. Sci. Ceram. 1981, 11, 231.
216. Ekbom, L. B. and Amundin, C. 0. Sci.Ceram. 1980, 10, 237.
217. Champagne, B. and Angers, R. J . Am. Ceram. Soc. 1979, 62, 149.
218. Telle, R. and Petzow, G. Muter. Sci. Eng. 1988, A105/106, 97.
219. Vasilos, T. and Dutta, S. K. Ceram. Bull. 1974, 53, 453.
220. Furukawa, M. and Kitahira, T. Nippon 7ungsten Rev. 1979, 12, 55.
221. Larker, H. T., Hermansson, L.. and Adlerborn, J. Ind. Ceram. 1988, 8, 17.
222. Schwetz, K. A., Grellner, W., and Lipp, A. in Science of’Hard Materials, Proc. 2nd Int. Conf.
Science of Hard Materials. Almond. E. A,. Brookes. C. A,. and Warren. R. (Eds’l. Inst. Phvs.
\
227. Wood, C., Emin, D., and Gray, P. E. Phys. Rev. B 1985, 31, 6811.
228. Aselage, T. L. and Emin D. Proc. 10th Int. Symp. Boron, Borides, and Related Compounds,
Albuquerque, NM, Boron-Rich Solids, AIP Con$ Proc. 1990, 231, 177.
229. Heuberger, M., Telle, R., and Petzow, G. Powder Metallurgy 1992, 35, 125.
230. Heuberger, M. thesis, University of Stuttgart, 1995.
231. Telle, R. in Boron-Rich Solids, Proc. 10th Int. Symp. Boron, Borides, and Related Com-
pounds, Emin, D., Aselage, T. L., Switendick, A. C, Morosin, B., and Beekel, C. L. (Eds),
AIP Conf. Proc. 231, American Institute of Physics, New York, 1991, pp. 553-560.
232. Telle, R. Ceramics for High-Tech Applications, Materials Science, European Concerted Action
COST 503 Powder Metallurgy-Powder Based Materials, Vol. V, Valente, T. (Ed.), The Eur-
opean Communities, Brussels, 1997.
233. Telle, R. and Petzow, G. in Proc. 9th Int. Symp. on Boron, Borides, and Related Compounds,
Werheit, H. (Ed.). University Press, Duisburg, 1987, pp. 234-245.
234. Efimenko, L. N., Lifshits, E. V., Ostapenko, I. T., Snezhko, I. A,, and Shevyakova, E. P. Sov.
Powder Met. and Met. Ceram. 1987, 26(3), 318.
235. Litz, L. M. and Mercuri, R. A. J. Electrochemical Soc. 1963, 110(8), 921.
236. Motzfeld, K. Acta Chem. Scand. 1964, 18, 1596.
237. Hofmann, H. and Petzow, G. J. Less-Common Met. 1986,117, 121.
238. Telle, R. and Petzow, G. in Horizons of Powder Metallurgy 11, Proc. 1986 Int. Powder Metall.
Conf. Exhibit., Kaysser, W. A. and Huppmann, W. J. (Eds), Verlag Schmid, Freiburg, 1986,
pp. 1155-1158.
239. Telle, R., Meyer, S., Petzow, G., and Franz, E. D. Muter Sci. Eng. 1988, A105/106, 125.
240. Kim, D. K. and Kim, C. H. Adv. Ceram. Muter. 1988,3, 52.
241. Nishiyama, K. and Umekawa, S. Trans. Jpn. Soc. Ceram. Met. 1985, 11(2), 53.
242. Rice, R. W., Richardson, G. Y., Kunetz, J. M., Schroeter, T., and McDonough, J. in Proc.
10th Annu. Con$ Composites and Advanced Ceramic Materials, Messier, D. R. (Ed.), American
Ceramic Society, Westerville, OH, 1986, p. 737.
243. Richardson, G. Y., Rice, R. W., McDonough, W. J., Kunetz, J. M., and Schroeter, T. in Proc.
10th Annu. Con$ Composites and Advanced Ceramic Materials, Messier, D. R. (Ed.), American
Ceramic Society, Westerville, OH, 1986, p. 760.
244. McCauley, J. W., Corbin, N. D., Resetar, T., and Wong, P. in Proc. 10th Ann. Con$ Com-
posites and Advanced Ceramic Materials, Messier, D. R. (Ed.), American Ceramic Society,
Westerville, OH, 1986, pp. 538-554.
245. Holt, B., Kingman, D. D., and Bianchi, G. M. Muter. Sci. Eng. 1985, 71, 321.
246. McCauley, J. W. Am. Ceram. Soc. Bull. 1988, 67, 1903.
247. Faber, K. T. and Evans, A. G. Acta Metallurgica 1983, 31(4), 565.
248. Selsing, J. J . Am. Ceram. SOC.1961, 44,419.
249. Fendler, E., Babushkin, O., Lindback, T., Telle, R., and Petzow, G. in Proc. IntSymp.
Ceramic Materials and Components,for Engines, Gothenberg, Sweden, June 10-12, 1991.
250. Fendler, E., Babushkin, O., Lindback, T., Telle, R., and Petzow, G. J . HardMater. 1993,4, 137.
251. Lonnberg, B. J. Less-Common Met. 1987, 141, 145.
252. Rey, J. and Male, G. in Proc. 9th Int. Symp. on Boron, Borides, and Related Compounds,
Werheit, H. (Ed.), University Press, Duisburg, 1987, pp. 419420.
253. Ohya, Y. and Hoffmann, M. J., Petzow, G. J. Am. Ceram. Soc. 1992, 75, 2479.
254. Sigl, L. S. and Schwetz, K. A. Jap. J. Appl. Physics 1994, 10, 224.
255. Sigl, L. S. and Kleebe, H.-J. J . Am. Ceram. Soc. 1995, 78(9), 2374.
256. Thevenot, F. and Bouchacourt, M. L’lndustrie Ceramique 1979, 732, 655.
257. Binder, F. Radex-Rundschau 1975, 4, 531.
258. Fitzer, E. Arch. Eisenhiittenwesen 1973, 44(9), 703.
259. Coble, R. L. and Hobbs, H. A. in Investigation of Boride Compounds for Very High Tem-
perature Applications, NTIS Report AD 428006, Kaufman, L. and Clougherty, E. V. (Eds),
Clearinghouse for Federal Scientific and Technical Information, Springfield, VA, 1973,
pp. 82-120.
260. Kislyi, P. S. and Zaverukha, 0.V. Poroshkovaya. Metallurgiya 1970,7, 32. Engl. Transl.: Sov.
Powder Metall. Met. Ceram. 1970, 7, 549.
261. Kislyi, P. S., Kuzenkova, M. A,, and Zaveruha, 0. V. Phys. Sintering 1972, 4, 107.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 943
262. Baumgartner, H. R. and Steiger, R. A. J . Am. Ceram. Soc. 1984, 67, 207.
263. Baik, S. and Becher, P. J . Am. Ceram. Soc. 1987, 70, 527.
264. Cech, R., Olivierus, P., and Seijbal, J. Powder Metall. 1965, 8(15), 142.
265. Rasskazov, N. I. in Proc. 3rd Znt. Powder Metallurgy Conf. 2, Karlovy Vary, Czechoslovakia,
1970, pp. 228-238.
266. Krylov, Yu. I., Bronnikov, V. A., Krysina, V. G., and Prislavko, V. V. Sov. Powder Metall.
Met. Ceram. 1976, 15, 1000.
267. Munir, Z. A. Am. Ceram. Soc. Bull. 1988, 67, 342.
268. Ouabdesselam, M. and Munir, Z. A. J . Muter. Sci. 1987, 22, 1799.
269. Clougherty, E. V. and Pober, R. L. Nucl. Met. 1964, 10, 422.
270. Samsonov, G. V., Goryachev, Y. M., and Kovenskaya, B. A. Izv. Vuz Fiz. 1972, 6, 37.
271. Castaign, J. and Costa, P. in Boron and Refractory Borides, Matkovich, V. I. (Ed.), Springer,
Berlin, Heidelberg, New York, 1977, pp. 390412.
272. Lonnberg, B. J . Less-Common Metals 1988, 141,145.
273. Rahman, M. and Wang, C. C., Chen, W., Akbar, S. A. J . Am. Ceram. Soc. 1995, 78, 1380.
274. Billehaug, K. and 0ye, H. A. Aluminium 1980, 56, 642.
275. Serrlie, M. and 0ye, H. A. Cathodes in Alurniniuni Electrolysis, 2nd ed., Aluminium-Verlag,
Dusseldorf, 1994.
276. Bannister, M. K. and Swain, M. V. Ceram. Znt. 1989, 15, 375.
277. Murata, Y., Julien, H. P., and Whitney, E. D. Ceram. Bull. 1967, 45, 643.
278. Venables, J. D. Philos. Mag. 1967, 16, 873.
279. Williams, W. S. Trans AZME 1966, 236, 211.
280. Ramberg, J. R. and Williams, W. S. J . Muter. Sci. 1987, 22, 1815.
281. Kang, E. S. and Kim, C. H . J . Muter. Sci. 1990, 25, 580.
282. Ly Ngoc, D. Dr. rer. nut Thesi.s, University of Stuttgart, 1989.
283. Jimbou, R., Takahashi, K., and Matsushita, Y. Adv. Ceram. Muter. 1986, 1, 341.
284. McMurtry, C. H., Bocker, W. D. G., Seshadri, S. G., Zanghi, J. S., and Garnier, J. E. Am.
Cerani. SOC. Bull. 1986, 66, 325.
285. Takahashi, K., Jimbou, R. Commun. Am. Ceram. Soc. 1987, C-369.
286. Janney, M . A. Am. Ceram. Soc. Bull. 1986, 64(5), 357.
287. Janney. M. A. Am. Ceuam. Soc. Bull. 1987, 66(2), 322.
288. Cai, H., Gu, W-H., and Faber, K. T. in Proc. Am. Soc. Composites, 5th Techn. Conf. on Com-
posite Materials, 1990, pp. 892-901.
289. Faber, K. T., Gu, W.-H., Cai, H., Winholtz, R. A,, and Magleg, D. J. in Toughening Mechan-
isms in Quasi-Brittle Materials, Shah, S. P. (Ed.), Kluwer, Dordrecht, The Netherlands, 1991,
pp. 3-17.
290. de Mestral, F. and Thevenot, F. in The Physics and Chemistry of Carbides. Nitrides, and Borides,
NATO AS1 Series E, Vol. 185, Freer. R. (Ed.), Kluwer, Dordrecht, The Netherlands, 1990,
pp. 457-482.
291. Phan-Tan-Luu, R., Mathieu, D., and Feneuille, D. Mbthodologie de la Recherche Expirimen-
tale, Fascicules de cours, L.P.R.A.I., Universite d’Aix-Marseille, 1989.
292. Tanaka, H. and Iyi, N. J . Am. Ceram. Soc. 1995,78, 1223.
293. Watanabe, T. and Kouno, S. Ceram. Bull. 1982, 61, 970.
294. Watanabe, T. J . Am. Ceram. Soc. 1977, 60, 176.
295. Watanabe, T. Am. Ceram. Soc. Bull. 1980, 59, 465.
296. Petzow, G. and Telle, R. in Advanced Ceramics, Somiya, S. (Ed.), Terra Scientific Publ. Co,
Tokyo, 1987, pp. 131-144.
297. Watanabe, T. and Shobu, K. Yogo-Kyokai-Shi1988, 96(7), 778.
298. Shobu, K. and Watanabe, T. Yogo-Kyokai-Shi1987, 95(1), 991.
299. Telle, R., Meyer, S., Petzow, G., and Franz, E. D. Marer. Sci. Eng. 1988, A105/106, 125.
300. Watanabe. T. and Shobu, K. J . Am. Ceram. Soc. 1985,68, C34.
301. Shobu, K., Watanabe, T., Drennan, J., Hannink, R. H. J., and Swain, M. V. in Proc. 4th Int.
Conf: Sci. Techn. Zirconia Advanced Ceramics, Zirconia 86, American Ceramic Society,
Westerville. OH, 1987, pp. 1091-1099.
302. Swain, M. J . HardMater. 1991, 2, 139.
303. Franz, E.-D., Tumback, M., and Telle, R. Sprechsual 1992, 125, 55.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
944 7 Boride-Based Hard Materials
304. Tumback, M., Franz, E.-D., and Telle, R. Sprechsaal 1992, 125,415.
305. Miiller, R. thesis, University of Stuttgart, 1995.
306. McHale, A. E. and Scott, R. S. J. Am. Ceram. SOC.1986,69, 827.
307. Takagi, K., Komai, M., Ando, T in Sintering ’87, Vol. 2, Proc. 4th Inst. Symp. Science and
Technology of Sintering, Somiya, S., Simada, M. (Eds.), Elsevier, Amsterdam, 1988, pp.
1296-1 301.
308. McMeeking, R. M. J. Am. Ceram. SOC.1986, 69, C-301.
309. Ogata, T. Toray Zndustries Internal Report, private communication, 1989.
310. Watanabe, T., Shobu, K., Tani, E., and Nakanishi, T. in Proc. 11th Znt. Symp. Boron. Borides
and Related Compounds, JJAP Series 10, Uno, R. and Higashi, I. (Eds), Publication Office, Jap.
J. Applied Physics, Tokyo, 1994, pp. 208-21 1.
3 11. Hamjian, H. J. and Lidman, W. G. J. Am. Ceram. SOC.1952, 35, 44.
312. Halverson, D. C., Pyzik, A. J., and Aksay, I. A. Ceram. Eng. Sci. Proc. 1985, 6, 736.
3 13. Nishiyama, K. and Umekawa, S. in Achievements in Composites in Japan and the United States,
Kobayashi, A. (Ed.), Proc. Vth Japan-US Conference on Composite Materials, Tokyo, 1990,
pp. 371-378.
314. Hayes, F. H., Lukas, H. L., Effenberg, G., and Petzow, G. Z . Metallkunde 1989,80, 361.
315. Zupanic, F., Spaic, S., and Krizman, A. Mater.Sci. Technol. 1998, 14, 1203.
316. Funke, V. F., Yudkovskii, S. I., Samsonov, and G. V. Russ. J. Appl. Chem. 1961,34,973.
317. Tangermann, I. Neue Hutte 1963,8, 973.
318. Samsonov, G . V. Metallov. i. Obrab. 1958, 12, 35.
319. Kieffer, R. and Benesovsky, F., Hartstoffe, Springer, Vienna, 1963.
320. Golczewski, J. and Aldinger, F., poster presentation at Annu. Mtg. Germany Society ,fbr
Materials (DGM), Gottingen, 1994.
321. Funke, V. F. and Yudkovskii, S. I. Zhur. Fiz. Khim. 1963, 37, 1557; translation into English:
Russ. J. Phys. Chem. 1963,37, 835.
322. Jiingling, T., Oberacker, R., Thiimmler, F., Sigl, L. S., and Schwetz, K. A. J. Hard Muter.
1991, 2(34), 183.
323. Jiingling, T., Oberacker, R., Thiimmler, F., and Sigl, L. S., in Proc. Znt. Conf.Advances in Hard
Materials Production, 1992, pp. 15/1-15/14.
324. Pastor, H., Allibert, C. H., Ottavi, M. L., Albajar, M., and Castro, F. French Patent Applica-
tion No. 91 08030, 1991.
325. Pastor, H. et al., in HMP’Y2 - Advances in Hard Materials Production, Proc. 4th Int. Conf.,
MPR Publishing Services, Bellstone, UK, 1992, pp. 23/1-23-11.
326. Sinchez, J. M., Barandika, M. G., Gil-Sevillano, J., and Castro, F. Scripta Metall. Muter.
1992, 26, 957.
327. Brynestad, J., Bamberger, C. E., Land, J. F., Fandinch, C. B. J. Am. Ceram. SOC.1983,66, C215.
328. Ghetta, V., Gayraud, V., and Eustathopoulos, N. presented at Sintering ‘ Y I , Vancouver,
August, 1991.
329. Li, J. G. J. Am. Ceram. SOC.1992,75, 3118.
330. Li, J. G. Rare Met. 1992, 11, 177.
331. Savov, L., Heller, H.-P., and Janke, D. Metall. 1997, 51, 475.
332. Panasyuk, D. and Umansky, J. J. Less-Common Met. 1986, 117,336.
333. Schwetz, K. A., Reinmuth, K., and Lipp, A. Radex Rundschau 1981, 3, 568.
334. Janaf, Thermochemical Tables, 2nd ed., NBS, Washington, DC, 1971.
335. Kubaschewski, 0. and Alcock, C. B. Metallurgical Thermochemistry, Int. Series on Materials
Science and Technology, Vol. 24, Pergamon, Oxford, 1979.
336. Matuschka, A. V. Borieren, Carl Hanser Verlag, Munich, 1977.
337. Schwetz, K. A., Reinmuth, K., and Lipp, A. Radex Rundschau 1981, 3, 568.
338. Sigl, L. S., Jiingling, T. Proc. 4th Int. Conf. Sci. Hard Materials (ICSHM4), Funchal, Madeira,
NOV.10-15, 1991, p. 321.
339. Ottavi., L., Chaix, J. M., Allibert, C., and Pastor, H. presented at Sintering ‘91, Vancouver,
August, 1991.
340. Jiingling, T., Dr.-Zng. Thesis, University of Karlsruhe, Germany, 1992, IKM-Series 007.
341. Liao, P. K. and Spear, K. E. in Binary Alloy Phase Diagrams, Vol. 1, Massalski, T. B., et al.
(Eds.), American Society of Metals, Metals Park, OH, 1986, p. 355.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 945
342. Murray, J. L. in Binary Alloy Phase Diagrams, Vol. 1, Mas-salski, T. B., et a/. (Eds.), American
Society of Metals, Metals Park, OH, 1986, p. 1118.
343. Jiingling, T., Sigl, L. S., Oberacker, R., Thiimmler, F., and Schwetz, K. A. Int. J . Refractory
Met. Hard Muter. 1993194, 12, 71.
344. Angelini, P., Becher, P. F., Bentley, J., Brynestad, J., Ferber, M. K., Finch, C. B., and Sklad,
P. S. in Science of Hard Materials, Proc. 2nd Int. Cohf. on Science of Hard Materials, Almond,
E. A,, Brookes, C. A,, and Warren, R. (Eds), Inst. Phys. Conf. Ser. 75, Adam Hilger, Bristol,
1986, pp. 1019-1032.
345. Sklad, S. and Yust, C. S. in Proc. 1st Int. Con$ Science ofHardMaterials, Hagenmiiller, R. and
Thevenot, F. (Eds), Bordeaux, I98 1.
346. Takatsu, S. and Ishimatsu, E. in Proc. 10th Plansee-Seminar on Powder Metall, Planseewerke,
Reutte, 1981, Vol. 1, p. 535.
347. Finch, C. B., Cavin, 0. B., and Becher, P. F. J . Cryst. Growth 1984, 67, 556.
348. Nishiyama, K. and Umekawa, S. in Proc. Composites '86: Recent Advances in Japan and in the
United States, Proc. 3rd Japan-US Conf. on Composite Materials, Tokyo, Kawata, K., Ume-
kawa, S., Kobayahi, A. (Eds.), 1986, pp. 433440.
349. Telle, R., Schmalzried, C., and Schunck, B, unpublished report to the Volkswagen-Founda-
tion, 1999.
350. Johnson, W. B., Nagelberg, A. S., and Breval, E. J . A m . Ceram. Soc. 1991, 74, 1093.
351. Newkirk, M. S., Urquhart, A. W., and Zwicker, H. R. J . Muter. Res. 1986, 1, 81.
352. Antolin, S., Nagelberg, A. S. J . Am. Ceram. Soc. 1992, 75, 447.
353. Nagelberg, A. S . , Antolin, S., and Urquhart, A. W. J . Am. Ceram. Soc. 1992, 75(2), 455
354. Weinstein, J. in Proc. Int. Symp.Advances in ProcessinE andCharacterization of Ceramic Metal
Matrix Composites, CIMjICM; Vol. 17: Mostaghaci, H. (Ed.), Pergamon, Oxford, 1989,
p. 132.
355. Lugscheider, E. and Eschnauer, H. in Proc. 9th Int. Symp.on Boron, Borides, and Related Com-
pounds, Werheit, H. (Ed.), University Press, Duisburg, 1987, pp. 202-212.
356. Komai, M., Yamasaki, Y., Takagi, K., and Watanabe, T in Advances in Powder Metallurgy
and Particulate Materials 8, Capus, J. M. and German, R. M. (Eds), Metal Powder Industries
Federation, Princeton, NJ, 1992, pp. 81-88.
357. Rieger, W., Nowotny, H., and Benesowsky, F. Monatsli. Chem. 1966, 97, 378.
358. Ozaki, S., Yamasaki, Y., Komai, M . , and Takagi, K. in Proc. 11th Int. Symp. Boron, Borides
and Related Compounds, JJAP Ser. 10, Uno, R., Higashi, I. (Eds.), Publication Office, Jap.
J. Applied Physics, Tokyo, 1994, pp. 220-221.
359. Takagi, K. in Proc. 11th Int. Symp. Boron. Borides and Related Compounds, JJAP Series 10,
Uno, R. and Higashi, I. (Eds), Publication Office, Jap. J. Applied Physics, Tokyo, 1994, pp.
200-203.
360. Sivaraman, K., German, R. M., and Takagi, K. in Proc. Int. Con/ Powder Metallurgy PM'96,
Toronto, 1996, pp. 9/117-9/129.
361. Sivaraman, K., Griffo, A,, German, R. M., and Takagi, K. in Proc. Int. Confi Powder Metal-
lurgy PM'96, Toronto, 1996, pp. 11/67-11/79.
362. Matsuo, S., Ozaki, S., Yamasaki, Y., Komai, M., and Takagi, K. in Sintering Technology,
German, R. M., Messing, G. L., and Cornwall, R.G. (Eds), Marcel Dekker, New York,
1996, pp. 261-268.
363. Warren, R. J . Muter. Sci. 1968, 3, 471.
364. Schmalzried, C. and Telle, R. in Proc. Int. Conf: on Boron, Borides and Related Compounds
( I S B B '99). Dinard, France, October 1999, to be published in J . Solid State Chem.
365. Takagi, K., Yamasaki, Y., and Komai, M. J . Solid State Phys. 1997, 133, 243.
366. Sigl, L. S., Schwetz. K. A., and Dworak, U. Int. J . Refractory and Hard Materials, 1994, 12,
pp. 95-99.
367. Ramberg, J. R., Wolfe, C. F., and Williams, W. S. J . Am. Ceram. Soc. 1985, 68,C78.
368. Freitag, B. and Mader, W. J . Microsc. 1999, 194, 42.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
8.1 Introduction
WC-Co hardmetal is a sintered material consisting of brittle tungsten carbide (WC)
crystals bonded by a tough cobalt-based binder. Figure 1 shows the microstructure
of a typical commercial WC-Co hardmetal.
WC-Co is used mostly on account of its hardness which is high for a metallic
material, although other outstanding properties, such an extremely high Young
modulus, a high thermal conductivity, and a low coefficient of thermal expansion,
contribute to its success in a wide range of technical applications.
This chapter reviews the information available on the dependence of the hardness
of WC-Co on microstructural parameters, composition and external conditions,
such as temperature. It also reviews information available on the effect of micro-
structure and composition on the relationships between hardness and toughness
and hardness and abrasive wear resistance.
The hardness of industrial hardmetal grades is characterized preferentially by the
Rockwell A scale [l] in the USA [2] and by the Vickers scale [3] in Europe. Most
research work, however, has been carried out, also in the USA, using the Vickers
scale. Most of the results reported below were obtained using standard pyramidal
Vickers indenters [ 3 ] .
(0001) 2100
(1 100) 1080
(1 101) 1060
34
32
-3
5 28
In
m
0)
f 26
m
r
24
22
I
201
1
0
Figure 2. Plot of the microhardness of a (0001) plane of a WC single crystal against the indenting load.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
948 8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
15!
10 15 20 25 30 35 40
l/d112(mm-’12)
Figure 3. Plot of the hardness of hot-pressed polycrystalline WC against d - ‘ I 2 , where d is the mean
size of the WC grains. Adapted from reference [9].
plot of the microhardness of the (0001) planes of WC crystals versus the indenting
load. It shows that the microhardness of WC decreases with increasing load, which
means that when measuring the microhardness of WC on sections of WC-Co samples
one obtains different results at different loads.
The hardness of WC has also been measured on polycrystalline samples [8].
Figure 3 reports results from several investigators [9] which satisfy the following
Hall-Petch relationship:
Figure 4. Plot of the hardness of the cobalt binder in WC-Co alloys against where X is the
mean width of the binder layers, or binder mean free path. Adapted from reference [9].
to room temperature only part of the cobalt transforms from the high temperature
(above 417°C) f.c.c. structure to the low temperature h.c.p. structure, and the
transformed regions are present as hexagonal lamellae in a predominantly cubic
material [lo].
The hardness of pure cobalt has been reported as being between 140 and 240 HV,
depending on the method of sample preparation [l I]. However, a solid solution of C
and W in cobalt can be up to 100% harder than pure cobalt [12], depending on the
composition. It has been difficult to measure directly the hardness of the binder in
WC-Co because the width of the binder regions in the hardmetal is typically of the
order of 0.1 pm.
However, Fig. 4 summarizes measurements by a number of investigators of the
in situ hardness of the cobalt binder [9]. The hardness is plotted versus X-”2,
X being the thickness of the cobalt layer where the hardness was measured, usually
called ‘the cobalt mean free path’. The results in Fig. 4 satisfy the following Hall-
Petch type relationship
Hco = Hoco + KocoX-”2 (2)
where Hco = the in situ hardness of the cobalt layers, X = thickness of the layers,
Hoco = mean hardness of the binder and Koco = Hall-Petch coefficient.
From Fig. 4 it has been calculated that Hot,, = 149kgfmmP2 and Koco =
16 kgfmmp’/2. Koco is related to the ease of slip transfer across Co-WC interfaces.
The contribution of the Co-Co grain boundaries to the hardness of the material can
be neglected because in WC-Co the size of the cobalt grains is of the order of lo3 pm
while the size of the WC grains is of the order of 1 pm. Thus the number of Co-Co
boundaries is negligible compared to the number of Co-WC interfaces.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
950 8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
..................................
.............................................................................................................................
......................................................................................
1000 - ...................................
800- .............................................
600- .....................................................................................................
400 I
0 5 10 15 20 25 30 35 40 45 50
I I I I I I I I
2200 I I I I I I I I I i
Figure 6. Plot of the hardness of WC-Co alloys against carbide grain size at the cobalt contents indi-
cated in the key. Reproduced by permission of L. Makhele.
where Hwcpc0= the hardness of the alloy, Hwc and Hco = the hardness of the
component phases as expressed by Eqs (1) and (2), and Vwc = volume fraction of
the WC phase.
However, Lee and Gurland found that this simple law does not agree with the
measured hardness values of WC-Co alloys. In order to reproduce the measured
values, they had to introduce the concept of ‘continuous carbide volume’ or
‘volume fraction of the carbide skeleton’, V,, which is defined as the volume fraction
occupied by that part of the WC grains which are involved in forming infinitely long
chains of connected particles while those parts which are surrounded by the cobalt
binder do not contribute to V,. Lee and Gurland derived that V, is related to the
volume fraction of the carbide phase, Vwc, via the contiguity, C:
v, = cvw, (4)
where the contiguity, C , is defined as the ratio of the grain boundary area and the
total surface area of the carbide phase.
By replacing Vwc in Eq. (3) with V,, Lee and Gurland obtained the following
equation:
Hwc-co = CVWCHWC + (1 - CVWC)HCO (5)
which is in excellent agreement with experimental results (see Fig. 7).
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
952 8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
1800
.
iii
E
E
m
& 1400
1600
/c-
I I
v)
5
(I)
1200
c
m
t
Y
0
1000
5
800
5
-0
0
2 600
400
4 0 600 800 1000 1200 1400 1600 1 DO
Calculated hardness (kg/mma)
Figure 7. Plot of the hardness of WC-Co alloys, calculated from Eq. (9,against measured hard-
ness. Adapted from reference [8].
0 2 4 6 a 10 12
mean free path (um)
Figure 8. Plot of the hardness of WC-Co alloys of four different grain sizes against the binder mean
free path U F = 0 7pm, F = 1 Opm, M = 3 Opm, C = 5 Opm Reproduced by permission of the
Int J of Refract Met & Hard Mater
the dissolved W and the dissolved C. The higher the total carbon, the lower is the
amount of W dissolved in the binder and the lower is the hardness of the binder
(Fig. 9) [19].
8.3.4 Porosity
The hardness of materials produced via powder metallurgy is known to decrease
with increasing porosity level. However, hardmetals are almost pore-free materials
and variations in their extremely low level of porosity does not affect their hardness
appreciably.
90.5
go - ......................................................
~ 89.5 - ......................................................
LII
v)
v)
89 ......................................................
E
88.5 _.....................................................
.
I
_
88 ......................................................
87 7
carbon -
87.5 _ ......................................................
---High Low carbon
0.1 1
Figure 9. Plot of the hardness of WC-Co alloys of different carbon contents against the binder mean
free path. Adapted from reference [19].
22
.;a 20
n
c
2
In 18
5
In
a,
16
0
.c
8 14
0
4 12
10
8
100 200 300 400 500 600 700 800 900
temperature (
'C)
Figure 10. Plots of the hardness of WC against temperature. Adapted from reference [20].
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8.3 Factors Affecting the Hurdness of' WC-Co Hardmetal 955
200
180
160
140
120
100
80
60
40
20
0 100 200 300 400 500 600 700 800
temperature (OC)
Figure 11. Band including possible values of the hardness of cobalt at various temperatures.
Adapted from reference [I 11.
where
and
with
24
22 -- d = 0.5 pm
m
h
a
c7
\
6 - r ~
15 wt% co
4
0 100 200 300 400 500 600 700 800 900 0 100 200 300 400 500 600 700 800
Temperature ("C) Temperature ("C)
Figure 12. Plot of the hardness of WC-Co alloys against temperature. The cobalt content of the
alloys and the grain sizes are indicated.
linear at any Vwc, in agreement with the results in Fig. 14. Figure 14 shows that the
above relationship is linear also at all temperatures, which implies that the
relationships between Hwc and d and Hco and X (see Eqs (1) and (2)) are linear
at all temperatures.
Figures 15 and 16 show that Ho in Eq. (7) decreases rapidly with increasing
temperature up to about 600"C, while K,, decreases substantially with temperature
only above 600°C. Since Ho is a combination of the intrinsic hardness of WC and Co
(see Eq. (8)) and Ky a combination of the Hall-Petch coefficients of Eqs (1) and (2)
(see Eq. (9)), Figs 15 and 16 suggest that the softening of WC-Co with increasing
temperature is due to the intrinsic softening of the component phases up to about
600°C but is controlled by the ease of slip transfer across grain boundaries and
interfaces above that temperature.
Milman and co-worker? results [21] have established that the hardness of finer
grained material decreases with increasing temperature at a lower rate than the
hardness of coarser grained material and thus the advantage of using finer grained
material increases with increasing temperature.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8.3 Factors A.fecting the Hardness of' WC-Co Hardmetal 951
24
22
20
18
0
c2 16
g 14
C
$ 12
r
10
8
6
4
0 100 200 300 400 500 600 700 800 900
temperature (C)
20
10 wt% c o
18
-I? 16
14
g 12
C
$
c
10
-
8
6
d = 0.5 urn +- d = 1.2 urn +++ d = 2.3 urn
4
I 100 200 300 400 500 600 700 800 C
temperature ( C)
18
15 wt% Co I
16
-I? 14
12
9.
g 10
C
$ 8
c
6
2 0
temperature ( C)
Figure 13. Plot of the hardness of WC-Co alloys against temperature. The grain size of the alloys
and the cobalt contents are indicated,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
-
25 .........
20 -
m
5 '5-
In)
u
acl
2 10-
m
I
5~
15 20 25 30 35 40 45 50
d (rnm)i'nr
ll.l^l"lllll ................................ " "
I
_I___
__
*20 ' C
20 -
0200 ' C
A300 ' C
A400 ' C
0500 ' C
0 600 ' C
m700 ' C
0800 'C
XJOO ' C
~
I
i
j
20 ~
15 wt% co
A300 "C
1 A400 "C
i 0 6 O O "C
15 20 25 30 35 40 45 50
d (mrn)4'n)
Figure 14. Plot of the hardness of WC-Co alloys against d-''2, where d is the carbide mean grain
size, at temperatures in the range 20-900°C. The regression coefficients are all higher than 0.95 and
mostly higher than 0.99.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8.3 Factors Affecting the Hardness of WC-Co Hardmetal 959
700
600
500
.
h
E 400
E
0
Y
Y
0 300
I
200
100
04 I
o loo 200 300 400 500 600 700 aoo 900
temperature ("C)
Figure 15. Plot of the parameter Ho of Eq. (8) against temperature for WC-Co alloys of different
cobalt content. The cobalt contents are indicated in the key. Reproduced by permission of the Znt. J .
of Refract. Met. & Hard Mater.
40 I
35
30
E
A
E
25
m
& 20
y"
15
10
5 I
100 200 300 400 500 600 700 aoo goo
temperature ("C)
L"
-
6 wt% CO
20
19
10 Wt% cc 18
3
a 17
16
w
c E 15
E
2 14
! 3 12
1
d = 0.5 pm d = 1.2pm
14.1 I I 11
-200 -150 -100 -50 0 50 -200 -150 -100 -50 0 50
Temperature ("C) Temperature ("C)
- 16
m
& 15
u)
$ 14
E
2 13
12
11
104 I I
-200 -150 -100 -50 0 50
Temperature ("C)
Figure 17. Plots of the hardness of WC-Co alloys of different cobalt contents against low tempera-
tures. The grain size of the alloys is indicated in the keys.
Milman and co-workers tested WC-Co alloys of grain size ranging from 0.5 to
2.3pm and cobalt content from 6 to 15 weight-% also at temperatures ranging
from -196°C to 20°C. The results are shown in Figs 17 and 18 [21].
124
-200 -150 -100 -50
I
0
I
50
temperature ( C)
24
22
- 20
cr
n
18
u)
16
;
I?
c
14
12
10
-200 -150 -100 -50 0 ! I
temperature ( C)
18
17
16
:
cr
15
3 14
c
-0
13
c
12
11
10
-200 -150 -100 -50 0 50
temperature ( C)
Figure 18. Plots of the hardness of WC-Co alloys of different grain size against temperature. The
grain sizes and the cobalt content are indicated.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
962 8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
1000 I I
8 10 12 14 16 18 20 22
Fracture Toughness (MPafi)
Figure 19. Plot of the hardness of WC-Co alloys of constant grain size against toughness. The grain
sizes of the alloys are 2.2 pm and 6 pm. Adapted from reference [22].
done to establish the relationships that exist between hardness and toughness and
hardness and abrasive wear resistance.
1000
........................
.-0IE
E
n l
U- ...................................
+F
............................
.................................. -UF "M *c
0.1
400 600 800 1000 1200 1400 1600 1800 2000
Hardness fKglrnrn2J
Figure 20. Plot of hardness against abrasive wear resistance for a wide range of WC-Co alloys.
U F = 0.7 pm; F = 1.0 pm; M = 3.0 pm; C = 5.0 pm. Reproduced by permission of Int. J . Refract.
M e t . & Hard Mater.
using alumina particles as abradors. The alloys tested ranged in cobalt content from
3 to 50 weight-% and in grain size from 0.6 to 5.1 pm [17]. Figure 20 shows that up to
a hardness of approximately 1000 HV there was a one-to-one correspondence
between hardness and abrasion resistance, which suggests that the main wear
mechanism was plastic deformation. Above 1000 HV the wear resistance of coarser
grained alloys was higher than that of finer grained alloys of equal hardness, up to
approximately 1600 HV. This suggests a wear process controlled by microfracture,
since in that range of hardness coarser alloys are tougher than finer alloys. There-
fore, the abrasive wear resistance of hardmetal increases with increasing hardness
but, in the hardness range which is of interest to most applications, there is not a
one-to-one correspondence between hardness and abrasive wear resistance since
resistance to wear is determined also by other properties, such as toughness.
8.5 Conclusions
The knowledge of the factors affecting the hardness of WC-Co alloys has increased
considerably in recent years, which has allowed a more precise selection of grades.
However, much work remains to be done to quantify the relationships that exist
between hardness and other properties which affect the performance of the material.
Acknowledgments
The author wishes to acknowledge the contribution of several colleagues to the research
work carried out over the years and reported here. Professor Yu. V. Milman,
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
964 8 The Hardness of Tungsten Carbide-Cobalt Hardmetal
References
1. ASTM Standard B 294-92.
2. J. J. Oakes, Teledyne Advanced Materials, Private Communication.
3. I S 0 Standard 3878.
4. S. Luyckx, F. R. N. Nabarro, Siu Wah Wai, and M. N. James, Acta Metall. Muter. 1992, 40,
1623.
5. T. Takahashi and E. J. Freise, Philos. Mug. 1965, 12, 1.
6. S. Luyckx and J. Katzourakis, Muter. Sci. Technol. 1991, 7,472.
7. S. Luyckx and L. C. Demanet, Unpublished Results.
8. H. C. Lee and J. Gurland, Muter. Sci. Eng. 1978, 33, 125.
9. L. S. Sigl and H. E. Exner, Muter. Sci. Eng. 1989, A108, 121.
10. J. Freytag, PhD Thesis, University of Stuttgart, 1977.
11. W. Batteridge, Cobalt and its Alloys, Chichester: Ellis Horwood, 1982.
12. B. Roebuck and E. A. Almond, Muter. Sci. Eng. 1984, 66, 179.
13. H. E. Exner and J. Gurland, Powder Metall. 1970, 13, 13.
14. D. G. F. O’Quigley, MSc Thesis, University of the Witwatersrand, Johannesburg, South Africa,
1995.
15. L. Makhele, University of the Witwatersrand, Unpublished results.
16. D. F. Carroll, 14th International Plansee Seminar, Plansee AG, Reutte, Austria, 1997, 2, 168.
17. D. G. F. O’Quigley, S. Luyckx, and M. N. James, Int. J. Refract. Met. Hard Muter. 1997,15,73.
18. H. E. Exner, Int. Muter. Rev. 1979, 4, 149.
19. T. Sadahiro, J. Jpn SOC.Powd. Powd. Met. 1979, 26, 33.
20. M. Lee, Metall. Trans. A , 1983, 14A, 1625.
21. Yu. V. Milman, S. Luyckx, and I. T. Northrop, Int. J . Refract. Met. Hard Muter. in press.
22. D. G. F. O’Quigley, S. Luyckx, and M. N. James, Muter. Sci. Eng. 1996, A209, 228.
23. S. Luyckx, V. Richter, D. G. F. O’Quigley, and L. Makhele, Proc. Int. Con$ Deformation and
Fracture in Structural Materials, Institute of Materials Research, Slovak Academy of Science,
Kosice, Slovakia, 1996, 2, 109.
Handbook of Ceramic Hard Materials
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
9.1 Introduction
Hard materials are used as thin hard coatings of some microns thickness for
wear protection of tools and machine parts because of their high abrasive wear
resistance. For the selection of the coating material the physical, mechanical, and
technological properites of these coatings, required by the application, are decisive.
The following data collection presents fundamental and available material proper-
ties for approximately 130 hard materials as a result of a literature search on
carbides, nitrides, borides, silicides, and oxides.
* This contribution is the revised and translated version of the publication Datensammlung zu
Hartstoffeigenschaften, published in Materialwissenschaft und Werkstofftechnik 1997, 28, 57-76,
0VCH Verlagsgesellschaft mbH, D-69451 Weinheim
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
966 9 Data Collection of Properties of Hard Materials
Good correspondence is found for crystal structure, lattice parameters, density and
melting point for one material.
The data collection (Table 1) does not claim to present all existing investigations
on hard materials. The aim of the presented table is to give a representative overview
on a special material. The number of quotaions gives a first impression on the
number of published values (see e.g. TIN which is well known as layer for tribolo-
gical applications). Looking on the five material groups, carbides and nitrides are
dominant.
The database for the data collection comprises actual textbooks, theses, and
publications in technical journals and from conferences in recent years. All in all
approximately 3000 references were searched through. One result is that in new
publications single phase materials are seldom studied. The quotation of older
literature in these publications leads to the repetition of single values by different
authors.
Acknowledgement
Many thanks to Mrs. R. Kurth for listing all the data on PC and thanks to Mr. U.
Petzel as well as Mr. T. Eid for their help in increasing the data base.
Table 1. Data collection of properties of hard materials
Legend
materials: carbides, nitrides, borides, silicides, oxides (mo) monocrystalline (x) density as determines by X-rays, determined by
m. 0, y. S, 6, TI, T2: phases (PO) polycrystalline lattice parameters and atomic mass
HV: Vickers hardness, value after HV gives load in kp according to DIN (i) incongruent melting point (formation of liquified material and a second solid phase) (d) decomposition
HK: Knoop hardness, value after H K gives load in k p according to DIN (per) peritectic melting point (several solid phases and liquified material in equilibrium) (p) “heat proofness”
HU Universal hardness. value after HU gives load in N according to DIN (I) “thermal stability”
Crystal stmcbre:
ortho: orthorhombic cub: cubic fcc: face-centered cubic mono: monoclinic The crystal systems of single materials may he
hex: hexagonal cub-B1: cubic, NaCI-type bcc body-centered cubic tri: triclinic explained by different crystal stuctures
hcp: hexagonal close-packed rhom: rhombohedral/trigonal let’ tetragonal e.g. hexlrhom for AIzO,)
Carbides
No. Symbol Crystal Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
Structure parameters point expansion, 3 conductivity, X resistivity modulus hardness resistance
inm) (c: cm-’) i’C) K-’) (W m-’ K - ’ ) ncm) (kl mol-’) (LO’ N mm-’) (10 Nmm?) (IOO’C)
I B4C rhom [XI 0 5631/1.2144 [52] 2.52 [I] 2450 [1.48,57.7716 [I] 27.63 [I] lo6 [I41 72 (51 4.5 [I1 3700 [I. 331 11-14 [ZS]
7 rhom [31] 0.5804,1.2079 [55] 2.5 171 2470 [5] 4.5 [8] 29.3 [XI 5 x 10’ [311 40 [43] 4.48 [lo] 3000 I71 7-8(P) [35]
3 rhom [43] 0.5599/1.20?4 [SS] 2 52 1141 2350 [7] 7.32,’s 33 [30] 10.42 1351 4 x 10’ [38] 71 6 1461 4.4 I141 490&5000 HV IS] ’ L431
4 rhom [SZ] 0 56003/1.2086 [I121 2.52 [251 2350 [XI 4.5 [31] 29 [38] 101-10’ [48] 57.7 [48] 2.96 [XI 2940HK0.05 [I41 x [701
5 hex [I121 2.52 [?I] 2450 [I41 45 [35] 28 [43] 4 x 10’ 1571 71.6 [73] 4.41 [31] 3700 HV 0.05 1251
6 2.52 [35] 2450 [25] [381 35 [48] lo6 [681 45 [38] 300M000HV [3I]
7 2.52 [38] 2430 [30] [431 28 [57] 5 x 10’ I771 4.6 [43] 2800 HK 0 I [33l
8 2.52 1391 2450 [31] 4.5 [48] 27 2 [I061 5 x 10’ [Y71 4.5 [48] 3700-4700 HV [35]
9 2.52 1431 2450 [38] [571 4.48 [S7] 4950 HV 1391
10 2.51 [48] 2447 [39] ~701 42 [70] 2800 HK [44l
11 2.52 [52] 2447 [43] 4.5 [77l 441 [77] 3700 HV [13.44. 57,681
12 2.52 [57] 2500 [55] 4.5 [97] 4.41 [97 35w-4500 HV 0 I [48]
13 2.51 [70] 2450 [88] 4.5 [I061 2.9 [I061 2940 HK 0. I I581
14 2.52 [771 2350 [I061 4.4-4.7 [lox]
15 2.5 Wl 4000-5500
30W3500 HV 0 I [58]
1701
16 2.52 [97l 300&7000 [781
17 2.5 U061 30w-4000 HV [77]
18 2.45 11081 3500 HV 0.2 [881
19 3200 HV [I081
20 Cr,C: ortho [6] I 146/0.552/0.2821 161 6.68 [6l 1900 (per) [Sl 10 3 [XI 18 8 (81 75 [7] 88.8 PI 4 191 1500-2000 I71 12 [71
21 ortho [8] I 147/0.5545/0.2830 [lo] 6.68 [7] 1850 [7] 117 [9] 19 [9] 75 [9] 88 [6] 3.86 [lo] 1300 HV [8,871 11-14 [25]
22 ortho [lo] 1.147/0.5545:0.2830 [Ill 6.68 191 I890 [XI 10.3 [lo] 75 [40] 23 [Ill 3.7 I401 2100 [91 11-12 [53]
23 ortho [Ill 1.147:0.553j0.282 [25] 6 68 [lo] 1810 [9. 10,401 10.3 [40] 75 [75] 109.7 [46] 3.73 [57] 1300 16, 10.40.761
24 ortho [25] I 147/0.5545/0.2839 1281 6.68 [XI 1810 176.77.871 10.3 [57] 75 [77] 85.4 [I021 3.284 1741 100&1400 [I 11
25 ortho 1281 I. 148310.5531/0.2827 [52] 6.68 [391 I895 [I31 10.3 [76] 75 [87 94.2 [73] 37 [76] 2280 HV [I31
26 ortho [46] I. 1488/0.552710.2829 [74] 6 66 [53] 1895 [ZS] 11.7 [77] 75 [I061 4 1771 1300 HV 0.05 ~251
27 ortho [52] 1.14883,O 5527310 28286 [I121 67 [57, 1071 1890 [39,46.90] 10.3 [87] 3.7 [87] 1350 HV 139.901
28 ortho [57] 6.68 1741 1895(d) I531 10.3 [I061 3.73 [I061 2280 HV 1531
29 ortho [74] 6.68 1771 18Y5(d) I571 2280 [571
30 ortho [76] 6 68 [YO. 1061 1850 [I061 1350 1591
31 ortho [I121 2150HV [771
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
32 Cr7Ci hex [I] 14/ll.45 [I1 69 [I1 1780Id) [I] 106 [I] 190? [I] 3.5 [53] 2200 [I1 7 [53]
33 hex [6] 1.398 0.4523 181 6.85 [61 1780 [46] 178 161 3.726 [74] I600 [Ill
34 hex [lo] I401 0.4525 [lo] 69 1521 1600-1790 [I021 228 2 1461 1900-2200 1531
35 her [Ill I401~04?32 I62 iloi] 1200-1600 H V [SBj
36 hex [25] I 398,O 453 181 3 [731 2200-2400 H V 0 05 [61]
77 hex [46.76] I4006 0.4532 1910HV003 [641
38 rhom [52] 0.7015,1.?153 0.4532 1200-1600 H V 1681
39 ortho [74] 070149'1 2153 045320 I500 H V 10.025 [72]
40 ortho Ill!] 1200 H K 0 01 [1271
41 Cr2,C6 cub [I] I 06 1520 [I] 101 [I] 68 7 [I/ 1650 111
42 fcc [6] I0638 68 6 [8] 1000 [ I 11
43 fcc [lo] I0655 396 5 [73] IROOHV (581
44 fcc [Ill I 06&1 066 l8OOHV [681
45 fcc (251 I065
46 fcc 1521 1.0659
47 fcc 174. 7x1 I066
48 cub (1121 1.06599 [Hi]
49 HfC fcc [8] 0.464 [lo] 122 [7] 3830 151 673 [XI 63 [8] 37 [7] 209 7 [II] 3 55 [9] 2200 H V 2
50 fcc IIO] 0 464 [Ill 1267 (91 3890 [7,8.9.13] 66 [lo] 6 [PI 39 [9] 205 [45] 3 174.21 [lo] 2600
51 cub-Bl [I I ] 0.464 [??I I 2 6 7 ( x ) [lo] 3928 [lo] 678 1301 29.31 [32] 37 [32] 219 [4h] 4.71 [32] 2700 H V
52 cub-BI 1251 0.464 [321 12 2 I251 3890 j25] 66 [32] I3 [38] 40 [38] 2303 [73] 4 [38] 2800
53 cub-Bl [30] 0 4646 [52] I?7 1251 3927 [30] 6.4 1381 37 [57] 3 5 4 6 0 8 [74] 2700
54 cub-BI [52] 0.446 [I091 12 3 [38] 3950 [32] 6.9 [45] 37 [lO6] 4 64 [76] 2276 HK 0.1
55 fcc 1571 0 463765 [I121 -12 [39] 3890 1381 6 (621 3 52 [I021 2200
56 fcc 1741 12 7 [52] 3890 1391 6.6 [76] 1932-2900 H V 0.05
57 cub-Bl 1761 12.2 1741 3387 1451 67 1981 1000 H V
58 fcc [98] I2 1901 3830 [46] 659 IlOh] 2913 H V
59 cub-BI [I091 I22 [I061 3890 1571 3200
60 cub 1112) 122 [I081 3890 i62j 180&2500 H V
61 12.7 [I071 3928 [76] 2600
62 3890 [YO] 180@2500
63 3900 198. 1061 ?913HV
64 2750 H V
65 MoC-., hex [2] 0 29'0 284 2700 [I] 49 [6] 11.7 [5]
66 hex 161 0 2901/02785 2677 IS] 126 [I!]
61 hex [in] o 2896 o 2x09 2692 [6] 10 [73]
68 hex 1521 0.2896\0 2809 2600 [lo]
69 hex [I121 0,2932 I 0 9 7
70 MoC-0 fcc [Ill 0427
71 cub-BI 1251 0.428
72 Mo2C-n ortho [ l o ] 0 7244jO 600410 5199 2670 (71 49:./x2 [lo] 228 [lo] 1600 171 5-8 [2?]
73 ortho [ I l l 04733l06034'05206 2400 [I31 8.5'4 5 5 7 [28] 1500 [101
74 ortho (28. 1121 04732.0 6037 05204 2410 [39] 1479-1800 [I11
75 1950 H V [I31
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table I. Continued
No. Symbol Crystal Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
Structure parameters point expansion, a conductivity, X resistivity modulus hardness resistance
(nm) (gcm-7 (“c) K-’) (W m-‘ K - ’ ) (lo-‘ Qcm) (kJ mol-’) (10’ N mm-2) (10 N mm-’) (IOWC)
76 MolC-$ hcp [6] 0.2994/0.4722 [6] 89 [6] 2687 (61 5.6/7 5 1281 97.5 [6] 46 [Ill 533 (571 15OOHVU1410.05 1251
77 hex [10,112] 0.3005/0 4755 [lo] 9 18 1251 2690 1251 7.8 1571 97 (251 46 (461 5.4 1771 1499 HV p91
78 hex [Ill 0.3004/0 4722 I251 9.2 1571 2430(per) [46] 7.8-9 3 1771 133 is71 46.1 (731 5.4 1971 iaoo HK (441
79 hcp 1141 0 300210.4724 1521 9.18 1771 2400Id) 1571 7.8-9 3 1971 57 I771 533 IlO6l ISOOHV
80 hcp 1251 0.414 8.9 1901 2430 1731 7.8 [I061 57 [97] 1950
81 hex 128,461 0 3012 0.47352 9 18 1971 2517 1771 71 [I061 I500 HV
82 hen (521 8.9 [I061 2410 1901 1500
83 hcp 1571 8.Y (107] 2517 1971 1499 HV
84 fcc 1251 8.9 [I081 2697 (1061 I660 HV
85 NbC fcc [I] 045 7.78 [I] 3490 [I] 6.65 111 14.24 [I] 3.C74 [6] 139 8 IS] 3.4 [I] 2400
86 cub-BI 161 0.4171 7 51-7.82 161 3600 [S] 6.2 161 11.25 [6] 34 [?I 140.7 [6] 2.763.45 161 195k2700
87 fcc 1x1 0447 7.82 171 3477-3900 161 6.84 181 14.2 181 74 1251 140.7 [Ill 3.38 [lo] 2000-2400
88 cub-BI [lo] 0.4471 7.79(x) [lo] 3500 I71 66 110) 14 24 [lo] 51 1321 138.2 1461 3 45 1321 2400 HV
89 cub-81 (111 0.4461 7 82 I251 3480 181 52 1301 14 24 1251 35 1571 140.7 1731 3 38 1571 2055-2400
90 cub-Bl 1251 0.447 7.56 1391 3600 [lo] 6.6 I321 14 24 I321 19 166) 3 193-5 10 1741 2400 HV 0.05
91 cub-BI 1301 0.4449 7.8 1571 3500 [I31 67 (571 I4 (1061 19 (771 58 [7?] 1800
92 cub-BI 1461 0.447 76 1741 3500 125) 7.2 1771 19 1971 5.8 1971 1961 HV
93 cub-BI 1521 044698 7.78 1771 3500 1301 7.2 197) 35 11061 3.38 [IOh] 1265
94 fcc 1571 7 56 1901 3600 1321 6.65 (1061 2000
95 ccc 114) 7.6 11061 3480 1391 2004 HV
96 cub [I121 7.9 [I071 3500 (571 2400 HV 0.025
97 7.9 [lox] 3480 j73,901 1800 HV
98 3613 (771 1961 HV
99 3613 197) 1800 HV
100 3480 1106l
101 Nb,C-a ortho [lo. 641 1.092/0497/0.308 [lo] 785(x) 1521 3090 (per) [73l 4.1UIl.OU7 1281 1863 1461
102 ortho [I121 1.092010.4974~0.309 11121 191.5 1731
103 Nh,C-P hex 161 0.540j04974 UOI 7 85 (XI 161 3100 (61 1924-2322 161
I04 hex 110. 1121 0.3116,’04958 11121
105 hex 1281
106 Nb?C-y hcp [6] 03128104974 I61 -3500 (per) 161 6.67.0,8 7 (61 167-195 161
I07 hcp [lo] 0.3127i0.4972 1101 6.816.9 1281
108 hcp 1281 0.311910.4Y59 1521
109 hcp 1521
110 SIC Ir.hcx [22] B.0.4360 [61 3.2 111 2200 [I] 5 68 [I] 15.49 111 10’ [9] a: 71 6 161 4.8 [I] 3500 [I. 331 1617 1251
Ill pfcc [22] n:0.3-7.3,1-1.5 [22] 3.17 16) 27W [7] 53 191 63-155 191 10‘ 1141 p73.3 161 3.94 I 161 1400HV 141 I3-14iP) 1351
II2 a:hex 1311 0:0.43 (221 3.2 171 2760fd) 191 4.8 (141 42 1141 10’-10’? 1221 80 1221 4.8 [9] 2169-2428 161 1: 16 1431
113 B:cub 1311 0 0 4358 1521 3.22 191 2300(d) 1141 a:4.5 [22] 60 (221 10’ 1311 1 70.0 (431 4.7 1141 2600 17. 451 16 1701
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
II 4 ?:hex [43] X:O 3n73!1 508 (1121 3.22 1141 2700 2. 440 [22] 20W3000
115 flub (431 3.2 [22] -2200 fl. 4.0 [22] 3500 HV
116 xhex [47] 32 [ZSj -2500 4.8 [31] 2580 H K 0 05
II 7 Fcub [47j 3.21 [31] 2760 48 [38] 3nowoon
118 ?.hex [52] 3.2 [35] 2300 :I 3 0 4 . 0 [43] 350M000
119 Pub [52] 3 21 [38] -2227 48 [45] 3500 HV 0.05
I20 ?.hex [76] 32 [39] 2827 (per) 48 [57] 2600 HV
I21 Fcub [76] 3.21 [43] 2760 4 [671 2600 HV
122 hex 198) 3 22 I451 7200 4 1701 2500 HK 0 I
123 xhex [I121 32 [57] 2986 (d) 48 [77] 3nw3500 HV
I24 3.15 [67] 2760 (dl 48 [97] 20W6000 HV 0 05
I25 3.21 [70] 21x0 48 [98] 2585 HK
I26 3.22 [77] 2760 3544 [I081 3000 HV 0.2
127 32 [XR] 2760 200&60no
I28 3 22 (971 2700 2580 HK 0.1
I2Y 3.22 [98] 2500 -4000 H V
I30 25-34 [lox] 3000
131 32 [I061 2500 H V 30
132 ~4on-3nonHV 0. I
133 2400-3500 H V
151 Ta2C-P hex [ I , 1121 0 31110.495 PI [I] 3400 (d) [I] 6 2,4 8 [28] 40 [6] 71.2 Ill I700
152 hex [lo] o.3103/0 49378 [6] I 5 05 (XI [6] 3500 (per) [73] 188.4 I61 1000- I720
153 hex [Ill 0.3102/0 494 104.3 1000 HK
154 hcp [I41 0.3106:0.4945 203.1
I55 hex (251 0.3091:n.493 203.1
I56 hex [52] 0.31037/0 49394
157 0.31042:O 4941
I58 Ta,C-Z hex [6] 0.3 1046’0.49444 u121
159 TIC fcc [I] n 433 [I] 4.93 [I] 3067 I41 7.42 [I] 2093 [I] 52 [41 I838 4 141 3200 HV [1.13.35.44] 12 [7]
I60 cub-BI [3] 0.429~~33 (61 4.93 141 3150 [?I 7.4 [31 17-235 [R] 68 171 1796 3.22 [6] 3200 HV [58.68.95] II 14 [25]
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[9Ll 0092 [LLI 8p9Z 1161 zoz
SZOO AH 0095 ILSl o w 1061 IOZ
AH motmooz I6El 0185 ILL1 ooz
B H 099Z lor1 triz [PLl 66 I
AH b60Z [SZl OW [LSI 861
SO0 AH 008Z (9011 ZZP L90l1 09 [Ell OE87 [6E1 [6011 SIVO 161
AH 0562 1161 0 [ELI 6'001 IL6l 65 1011 8P9Z [SZl [PLI 91VO 961
000E-OSPZ ILL1 Z'P I9Pl 6'001 [LLI 65 I9011 Z'L [XI OERZ [Oll [zsl Z81P'O 561
AH 0085 b L I ZEW
' YZ [Ill 9'101 65 [LLI r'L [LI OLLZ [Ll [SZI 91P'O P6 I
0065 [LSI ZZ'P I91 6101 09 [LSI Z'L L91 OSLZ 1011 991P'O E61
WH O9LZQPEZ loll PE'P [sl 1'501 951 [Sl ZP [or] SSY [Sl OLLZ [91 riwo 561
0562 "1 87 [I1 Z'Lll OS1-09 [91 61'P [a1 5'9 "1 om "1 zvo 3A 161
SO0 AH 00LE-0085 06 I
z o AH oozr 681
AH 8865 881
5500 AH 007s E6b 181
SO0 AH OOOS 6P 9R I
SOOAHSESI 80P 581
SO0 AH OOZZ oLir X P P81
DOZS-00PE 1561 090s E6'P EX1
SO0 AH OOZS-OWE [9011 PL L [861 L~OE E6'P 281
s o 0 AH ommoor [861 9'8-8 [La] 190s 6P 181
1.0AH wzr L90E 6'P 081
AH ooomrs osir E6 P 61 1
006r LPlE E6'P 811
WH OLPZ 19011 89 oLor Ebb LLl
AH 000f (Z61 55 wt f6'P 91 I
OS AH mzt [861 ZS 091s r6b SLl
1'0 B H OLPZ [L61 ZS osir Z6f PLI
008Z [LXI 5s LPit 6P ELI
AH 0085 ZS WIE t o ZLI
SO0 AH OOZE 89 wit r6P ILl
omz zs [9011 IZ 190s 6P OLI
SO 0 AH OOLfHM8Z 1191 55 1191 ft 190s P6'P 691
SO 0 WH 0085 [LSl 89 ILSl oz 190s E6b 891
E'O AH 008Z-0091 [OPI ZS [Ed 62 OblE E6P 191
00Sf-0085 [8E1 OL ILEl IZ oLor E6b 99 I
AH o o x r - 0 ~ 6 7 [zrl SP [srl lwpz om 16P 891
0067 ZS HOE (X) 16P 09 I
OOZE 5'89 L90E 16') $91
AH 0062 5s 0x1s 80 P Z91
5 AH 0085 [91 L6tZ ZS L6l zr-iz OLlf E6'P 191
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
[iYI OtLi YIP0
[ESI (PlOOLi YSLP Oi8PhhiO
[Sil OELi 9liLPO'hOXlS 0
ILl LSXi I LP 0;Xhi 0
191 0XRi*iLi SSLp'O~iOOtO
(51 ooxi i L P 01662 0
X ~ L P0'10oi'n
5 0 AH OROZ
AHOOli
AH O S f i
Si0 0 AH OOCZ [XOll EY 51
OOSi "011 8 51
AH 0Wi [YO11 SI Ll
SO 0 AH 000Z-00f I [YO11 96 Y [861 ZL'SI
000Z-00i1 [X61 iL [Lhl ZL'SI
AH 0011 [Lhl ii [Oh] L'SI
1lZl [Sh] iL [xi L SI
AH 000; [LR] S6 Y [LLl iL'SI
AH OOPZ lox1 ii'L [PLI EY S I
YH onRi LLLl iL [LYI L'SI
AH ORLl 1911 26 Y [LSI L'SI
AH OROZ [Pi1 LOZ Y-Lil Y [is1 L'SI
AH 00Oi-oOY1 [hY] i'L- [SPI LSl
"El OOPi [LYI 96 Y [hfl L'SI
[rrl IOYH0081 [LS! 96'9 iRil 1E l
[ril A H nooi 00x1 [SP] 56 Y lit1 LSI
a o AH o o ~ i IOPI 56 Y [SCI LLSl
so o YH O S E ~ [Xi1 S'L ktl i'jl mi o Z Y O ~ Zn
AHO81i [iil ir~ [Ell OOYi [Sil L'SI 16i n
SXOZ [ti1 LO L [OII YLLi [PI1 L 51 8t8i 0 Y06i 0
161 OOti [El Li L [61 YLLZ [Oll ixlRS1 iixz n ~ 0 6 io
181 AH OWZ OOOi [PI1 L 1x1 OtLi 161 iLSl lit1 LlRZ 0 Y06i 0
Its1 x-c [YL'SP'OPI 110Ii [nil 69 Y [Ll LYXZ lLl L'SI [Sil ltXi 0 OOhi 0
[jtl idly; l i t '(11 'il 00 IZ I61 iL 191 SXLZ [YI L $1 6iRi 0 YO62 0
lSi1 x-S It1 AH O O U [PI L [PI YLLZ bl i L 51 L t R i 0 Y06i 0
[Ll 8 ILS 'I1 nxoi 111 iL [I1 OOYi [I1 L 51 X i 0 6i 0
L6SP 0 Y06i 0
LLSP o i o h i n
LLSP 0 ZOhi 0
[ti1 SYIi [YI S'P LhLSP 0 lpO6i 0
it05 0 SPLS 0 LLSP 0
SSP 0 YO0 I 6P1 1
ios o PLS n LSP o
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal Lattice Denskty Melting Linear thermal Thermal Electrical Enthalpiy Young Mxro Oxidation
Structure parameters point expansion, a conductivity, X resistivity modulus hardness resistance
(nm) (gem-') ("c) (10.' K-I) (W m-' K-') ncm) (kJ mol-') (los N miK2) (10 N mm-') (100°C)
252 zrc fcc [I] 0.47 6.5 [I] 3535 Ill 6.73 [1,6] 20.93 [I] 63-156 (61 263.8 [I] 38 [I] 2560 [L,571 12 171
253 cub81 [6,76] 0.46894.476 6.51 (xj 161 3400 IS] 6.93 [XI 20.5 [XI 42 11 181 7 151 4 191 2600 HV [6, 81 11-14 [25]
254 fcc 181 0 4698 6.51 [?I 3532 [6] 7.0-74 191 I2 191 42 [9] 196.8 1111 4.07 [lo] 2600 [7, 10,32.45]
255 cub-BI [lo] 0.4698 6 59 (x) [lo] 3530 [?I 6.7 (101 20 52 (251 75 (2Sl 202.1 (461 3 88 (321 2500 191
256 cub-Bl (11,251 04685 6.9 I251 3530 [XI 8.3 I301 20 52 1321 43 1321 196.8 1731 3.9 (381 260C-2900 1131
257 cub-Bl 130.521 04698 66 [38] 3445 [9,45. 77, 97, 6.7 (32,481 19 (381 50 (381 184.6 (981 4 [45] 2560HV 113. 17.981
9x1
258 tet 1461 0.4694 6.73 [39] 3420 1101 57 1451 209 1981 42 1571 3. .32& [74] 2600 HV 0.05 (251
4 02
259 fcc [571 0.4693 6.73 [74,97] 3530 [39. 57,901 7 a 7 . 4 1771 20.5 (106] 42 1661 4 (771 2925HV 139,901
260 fcc 1741 0.47 6.63 (97,981 3443 I301 7-7.4 197) 42 (771 4 (971 2700 1591
261 fcc (981 0 467 6 63 [9,45,771 3420 [32.38.461 7-7.4 I981 42 [97.981 4 (981 lOo(t2800 174
262 mb-BI IlOS] 0.4693 6.73 11081 3530 [13.25, 1061 6.73 [I061 42 [I061 3 48 [I061
263 cub (1121 651
264 6.7
Nitrides
265 AIN hex [6] 0.311/0.498 3.05 >2200 [5, 106) 6 [38] 10 1381 10" 1381 288.9 [S] 3.15 1431 I200
266 hex 1431 0.31 IjO 494 3 09 1900 [6] 5.4 (431 Il&170[43] 10" 1571 290 (43) -3 1.571 1230HV
267 hex 1521 0.31 10/0.4975 3.25 2230(d) [7] 48 [471 30 [57] 10" [66] 3186 1461 3 [74] 25OOHV
268 hex (741 0.311\0.4979 3.05 2300(dj 1381 7 1571 165 [I081 10" [771 318 I651 3 437 1771 1230HK 1
269 hex (I121 0.311 14/0.49792 3.26 2397 1391 57 (771 10" (971 318 2 1731 35 138,971 1200
270 3.26 2227(d) [43] 5.7 [97] 0~~
1 0 ~ ~ - 1(ioq 1200 HV
27 1 3 26 2400 [47, 571 5.3 [lox]
272 3.26 2517 1731
273 3 26 2250 (d) (77,971
274 3.09
275 3.2
276 3.25
277 BN hex 111 0.251/0.669 2.25 3000 [I] 3.8 (381 284.7 [I] 3 x lo" (261 252.5 (51 0.9 [I] 4400HV
2?8 hex 130) 0 2504/0.6661 2.25 3000 [5] -2.8/41.5 [30] 25 138) 10" 138) 119.3 [6] 09 (381 4700
279 hex (391 0.2504/0.666 I 2 25 3000 [30] 1431 25 1431 10'' 1661 225 1431 3 1431 4700HV
280 hex (431 0 362 2.27 3000 (d) 1381 4.8 [XI- Is0-2Ol[81- 10" 191 252.5 1461 6.6 [9] 5000
2x1 hex 1461 0.36158 2.34 2997 (391 2 . 5 4 7 [9] 42 [9] 10'' (141 252.5 [73] 5.9 (141 4700 HK 0 05
282 hex 1521 3 48 2727 [43] 2.5-4.7 [35] 41.9 [35] lozo 1261 6.8 1361 210(t3600
283 hex [78] 3.48 2300(d) (71 [43] 200 [36] 10'8 [?7] 7.2 [36] 9OOC-9500 HV (mo)
284 cub [Y] 2.2 2730 [9] 3.5 1571 42 [43] 10"-10!~ [IOX] 8 1431 7000-8000 HV (PO)
285 cub (141 3.4b3.49 (mo)(351 2730 1141 14.0 (IOX] 15-33 [I081 6.8 (571 4050HK
286 cub 1351 3.3-3.4 (p0)[35] 2730 1251 20 [I061 6.6 (771 4700HK
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
287 cub [36] 3 48 1361 3027 6.6 1971 4000 HK 1571
288 cub [39] 34 1361 3000 4700 HK 0.1 1581
289 cub [43] 3 45 1391 2730 6006-8000 H K 0 I [70]
290 cub 1571 3 48 [43.88] 3000 200&4000 [761
291 cub [58, 761 345 [57] 2300 5000 H V [771
292 cub [77, 781 3 48 [70.77] 8000 H V 0.2 [881
293 cub [I121 2 34 [I061
294 22 [I071
295 18 [lox]
296 CrN f c c j l ] 0415 [I1 61 Ill I050 23 [lo] II 72 {lo] 640 I571 118-124 [61 4 [60] 1806-2100 171 7-7.5 [I251
291 cub-Bl 13.61 04149 [3] 5 39-7 75 161 1450 (d) 23 [32] 1 I 72 1321 640 1661 123.1 1461 3.236 1741 1100 [10.32.591
298 cub [I121 0414 161
. I
77 I500 23 1601 640 (771 I24 8 [lo?] 4 [77] 2006-2200HV005 [23]
299 cub-Bl [lo] 04149 [I01 6 I !x! I500 23 [971 640 [97] 123 I [73] 4 1971 2606-2900 HKOO5 [45]
300 cub-BI [ZS] 0.4148 [25] 6.1 I500 (d) 2.3 [XI] 118 I [Ill] 33 [XI] 1090 [571
30 I cub-Bl [46] 04149 1321 5Y loso 23 [I101 2 65-2.9 [I041 1x06-2200 H V 0 05 [hO]
302 cub.Bl 1521 04148 [52I 6 I !x) I050 23 [Ill] 4 11101 1965~2295HV 1641
303 fcc [57] 0.414 174. 1091 6I I500 7 S(XS0C- [ I l l ] 4 illlj 1890HV001 j75j
304 fcc [60] 0.4149 5.9 -1040C) 2.45 [I201 IIOOHV 1771
305 fcc 174. 751 0416 6 12 2300 HV [791
306 cub-Bl 176. 1091 0415 6 12 13284140 HVOOOOO5 [Xh]
307 fca [XI] 04148 59 70rMI00 H U 0 I [I031
308 cub [l04. 1101 0414 680-780 HU 0 I [I041
309 - 100&1700 H V 0 0 I5 [II51
310 1300 H V 0.05 [I181
31 I 810HV [1201
312 175P2650 H V 0 25 [I251
313 2600-2900HK 0 0 5 11251
314 Cr:N hex [25] 0 4760,’O 4438 59 - I500 94 308 [Ill] 3 138 1741 2250 H V 0 01 1751 8 6 3 1103 161
315 hcp [6. 52.751 0 4759 0 4438 59 1590 9 41 252 [Ill] 27 [Ill] 2500HV 1791 114 3 [46. 731
316 hex 13.74.761 0 48 I I 0 4484 58 lLCU2330 HK 1851 1277P [lo?]
317 hex [I041 04759 0.4438 6 51 IIOOHUOI [1031
318 hen [110.112] 02750447 1570HV005 [1111
319 0 481 I3 0.44841
320 HI% Ec [8] 0 452 13 8 3310 6.9 II 3 [XI 26 [71 369 4 1121 33341 [291 170CL2000
321 cub-Bl [lo] 0 4526 13 8 ( x ) ilbl 3300 69 2177 I101 I5 (321 369 [I71 4 64 1401 2000 H V
322 cub-BI 1251 0.451 I I 97 [74] 2700 6.9 21 77 1321 33 1571 369 3 [461 4.64 1451 1600
123 fcc 1521 0 4513 13 84 1981 3387 69 56. 5 11061 300 1651 3.8 1741 1500-3500 H V
324 fcc [57] 04518 1197 [XI] 3300 6.6 369.3 [73] 4.64 1871 150&8590 H V
325 fcc [74] 0 452 138 [I061 1387 66 3.8 1811 1600HVOI
326 cub-BI 1761 0 4525 13 8 [IO7] 3928 66 2700
327 fcc [YR] 0 451 3928 6.9 1640
328 fcc [Sl] 0 452 3330 6.9 1700
329 cub [II?] 0 45257 3300 6.9 2000
330 3300 160!-35W
33 I 3928 2700 H V
332 3200 2030
333 3305 1300 H U 0 I
334
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crqstal structure Lattice Density Melting Linear thermal Thermal Electrical tnthalpiy Young Micro Oxidation
parameters point expansion, cz conductivity. X resistivity modulus hardness resistance
(nm) (gem-') K-’) (W m? K-’) Ikm) (kl mol-’) (10’ N rnm-’) ( I 0 N mm-’) ( I 00°C)
~ ~~ ~ ~~~
335 Mo?N-y fcc [6, 101 0.41394.4160 [lo] 9.46 [6] 700(d) 1571 17.88 [6] 70 3-81.6 161 630 (571
336 cub [I121 0.4128 1101 527 (731 69.5 (461
337 fcc I251 0.41554.4160 [25l 69 5 1731
338 fcc 1521 0.4163 1521
339 Icc [57,76] 04163 Ill21
340 Mo2N-B tet [6] 0.4184.42110.8-0.820 [6] 93 [6]
341 let 176, 112) 0.421010.8060
342 NhN-6 fcc [I, 81 0.44 83 [I] 2300(d) [I] 10.1 161 2 9 N . 6 1 [6] -200 1571 237.8 [I] 1378-1422
343 cub-Bl [6] 0.439 8 . 2 6 8 4 [6] 2300 18) 10 1 18) 3 74 [8] 237.8 161 1400 HV
344 cub41 [25] 0.4392 8.4 125) 2050 1251 1400
345 fcc [57] 0.4375 2630(d) 1571 1400
346 NbN-E hex (61 02951,’1.1271 8.4 [7] 2570 (71 10 I [lo] 3.77 [lo] 60 1321 21 1.9 151 4 814 [741 I4on
341 hex [lo] 0.2958/1.1272 7.3 (x) [lo] 2300 191 10 I [32) 3.77 1321 58 [66] 2366 [I21 4.8 1771 1400
348 hex 114.461 0.2958/1.1272 8 1741 2204 110. 32,971 10.1 1771 58 (771 234 1461 4.8 1971 1400 HV
349 hex (74,761 0.295211.125 8.43 1771 2630 1461 58 1971 236.6 (731
350 hex-Bl 1521 0 298610.5548 843 I971 2204(d) i771
351 NbN hex [I121 0.2980:1.1270 84 1106, 1081 2573 [I061 10.1 [lO6] 200 [I061
352 73 [I071
353 Nh,N hex [I] 021,0.496 8.3 [I] 2430 121 3.26 [6] 4 27-8.71 161 255.8 [I] 2120
354 hcp [6. 521 0.3052,’0.4964 8.08-8 62 [6] 2400 [73] 255.8 [6] 1620-1 820
355 hex [25] o 3058~04961 8.33 [52] 253.3 [12.73]
356 oriho (461 0.305610.4948 8 31 (x) [52] 248 7 1461
357 SIN, hex [6] ~0.78~0.56 3.44 [1.106] 1900 (11 2.4 [I] 20-24 [Y] 10” [Y] 750.5 [6] 21 1410 HV
358 hex (221 ?: 0.75810.5623 3.44 [6] 1900 [61 2.5 [9] 17-30 1221 10’8 1141 760 [22] 2.06 1700
359 a$ hex 143) p: 0.7610 30 3.44 (71 1900 I71 3 1141 I2 1381 in“-io18 [22] 630 (431 1’3.20 1700 H K 0.05
360 hex 1761 1:0 77541/0 56217 3.19 [Y] 1900 I91 2.7-3 5 1221 1&15 1431 10” 138) 745 2 [46] p. 2.90 1700 HK
361 z.P.hex 11121 p: 0.76044,O 29075 3 I8 1141 1900(d) 1141 28 1381 30 [57] lo6 [57] 2.2 a 28OCL4000
362 3.2 1221 1900(d) [221 2.7-3.3 1431 32 1671 lo1* 1661 1.63 0 8: 2000-3000
363 3.2 [38] 1900(d) (381 4.2 (471 10-13 [I081 >lo’’ (671 -3 2300 HV 0.05
364 344 139) I897 1391 24 1571 1018 1771 3.1 3340 HV
365 3.19 [43] 1897(d) 1431 3.3 [67l 1018 1971 3 2 0 W 5 0 0 HK
366 2.75-2.95 [47] 1900 (571 3 [TO] 10”-10~’ [lox] 2. I 1800 HV
361 I 8-3 I 147) 1900 [771 -1.5 (761 2. I cl.1700HKOI
368 3.44 [SZ] 1900 [W 2.5 [77] 1.7-3.0 25WiW0 HV
369 3.15(x) [52] 1900 [I061 2.5 [97] 3300
370 3.25 1571 33 (1081 1$00 HV 30
371 32 167) 2200-3200 HV 0. I
372 3 21 1701 5@3900
373 3.19 (71.971 I720 HV
374 3.2 1881 1400 HV 0.2
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
375 34 [lo?] 80&2000 HV
376 2.4-3.44 [lox]
377 T d N 0 52 0.29 [I1 [11 3000 111 36 8 58 [6l 128 161 225.7 5 756 1741 3240 5-8 [25]
378 0 5185,10.2908 [6] 13.613 8 [6] 3090 [5,25.88] 5 96 [S] 135 [25] 243 1 1060
379 0.5185/0 2908 [lo] 14 35 (x) [6] 3360 [l] 3.6 8 79 [lo] 128 [32] 25 I 1500-3000
380 0.495/0.305 (251 16.3 (7.741 2090 [8] 36 8 79 [32] I28 1571 252 5 1300 HV
381 0 518510 2908 [32] 14 34(x) (lO,52[ 3093 [lo] 13&250 [84] 252 5 1000
3x2 0 5 18510 2908 1521 14.1 [25,107] 3093 1321 1060
383 0.519210.2908 [74] 13.6 [52] 2950 (461 1150-3000
384 0 51Y18\0 29081 [Il>] 13.8 (881 W30(d) (571 3230 HV 0.2
385 163 [lO6] 3360 [I061
386 I6 3 [I081
387 Ta2N hex [21 0 3010 493 [2] 1 8 [I] 3000 121 10 05 [6] 263 161 270 9 3wo
388 hcp [6.10l 0 30476 0 49187 [6] I5 46 [6] 2700 [46] 19&250 [84] 270 9 I220
389 ortho 146) 0.3048/0.4919 [lo] 15.46 1521 2727 1131 270 5
390 hcp [14,52] 0.3048 0 4919 1521 15 86(x) [52] 273
391 hex [I121 0 3044510 49141 [In]
392 TIN fcc [I] 0 423 [I] 5 21 [I] 2950 [I] 9 35 29.31 Ill 25 [7l 336 6 25 [I] 2450 HV
393 cub-BI [3) 0.4238 [3] 5 21 161 2950 [?I 94 28 9 [XI 25 191 303 I 59 191 180~2800
394 cub-BI [6] 0.423 161 5 21 171 3220 [6] 9.35 30 19) 21 7 [ZS] 336 2 2.51136 [I?] 2400 HV
395 rcc 1x1 0 424 [lo] 54 [9] 2950 [7] 9.4 19.26 [lo] 25-30 [??I 338 I 2 56 (321 2500-2800
396 cub-BI [lo] 0 423 [25] 5.39 (x) [lo] 2930 181 9.35 19.26 1321 75-500 [27] 336 2 56 1371 2000
397 fcc 1231 0.424 [2Y] 5 21 1251 2950 [9) 9.35 29.2 [37] 25 1321 336 6 2.6 [38] 2000-2500 HV 0.05
398 cub-BI [25] 0 424 1321 5 21 [37] 2949 [lo] 9 35 38 [38] 30 [38l 337 59 1401 1-6000 HV 0 001
399 fcc [41] 0 4245 93 30 1571 25 [40l 333 6 1451 2000 HV 0 015
400 cub-BI [46] 0 423 94 19 [9l] 25 (571 338.1 25 [53] 220&2800 HV 0 05
40 I cub-BI [52] 0 424 54 (451 2930-2950 (291 93 70 [96] 17 161) 336.6 59 1601 1770 HK 0 I
402 Pc [57] 0 4241 4 73 1521 2949 132) 93 25 [66] 2 56 [61] 2300-2600
403 fcc [60] 0 424 5 43 1531 2950 [37] 9 35 25 [77] 256 1621 1170HK0.1
404 fcc [74] 0 42 5 21 [57] 2950 1381 9.35 25 [87] 2 51-6.40 [741 2450 HV 30
405 cub-Bl 1761 0 42 5.3 [60] 3205 (391 8.3 25 I911 59 [77] 1994 HV
406 fcc [loll 0 422 5.21 [74] 2900 [40] 8.3 25 [97] 2 56 [SO] 2100
401 fcc 19x1 0.424173 5.4 [77] 7.950 (451 9 35 25 (981 2 . 7 9 4 5 1 [82] 2200 HV
408 fcc ISl] 52 1881 2950 [46] 94 I7 [92] 59 [87] 3300 HK 0.05
409 cub-BI [I091 5 43 [90] 2950 I571 93 22 (1061 2.66640 [XY] 2300
410 cub [I121 5 21 [91] 2950 [60] 94 5.9 [91] 160&2000 HV
41 I 54 [93) 2950 [61] 94 5.9 [93] 2000-2700 HV
412 52 1961 2950 [62] 8 339.4 2.6 [95] 1900-2400
413 5.4 [97] 2949 1761 9 35 59 [Y7l 2450 HV0.I
414 5.4 [98] 2950 1771 83 59 [98] I050 HV 0.1
415 5 43 [92] 2950 [88] 9 35 25 [92] 200&2400 HV 0 05
416 5.21 [XI] 2900 187) 9 35 4 1811 170C-2800HV 0 05
417 5.21 [I061 3205 [90] 15OC-2700 HV 0 05
418 54 [I071 2930 [91] 1725-1825 HV 0.02
419 5.22 [lox] 2950 (931 1775-1935 HV
420 2950 [Y8] 1800 HV
421 2950 [92] 2160 HV
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
Structure parameters point expansion, c( conductivity, X resistivity modulus hardness resistance
(nm) (g cm-’) (“C) K-‘) (W m-’ K-’) ncm) (kJ mo1-l) (10’ N mm?) (10 N mm-’) (100°C)
436 VN fcc [1,8] 0.41 [I] 6.13 [7.74] 2050 (d) [!I 8.1 [XI 11.3 [6] 85-100 [6] 147.8 [S] 4.6 [77] 1520
437 cub-BI [6.25] 0.4137 [6] 5.62 [6] 2050 [5,6,7,8,25,73]8.1 [lo] 11.3 [XI [32] 217.3 [6] 4.6 [97] 140CL1600
438 cub-BI [46,52] 0.414 [3,10,32] 6.04 (x) 161
.. 2177 [10,32] 8.1 [32] 11.3 [lo] [57] 217.3 [46] 1500 HV
439 fcc [57,74] 0.4126 PI 6 [lo] 2350 (461 9.2 [77] 11.3 [32] [66] 217.3 [73] 1500
440 cubdl [3,10] 0.4169 [52] 6.04 [25,52] 2350 (d) [571 9.2 [97] [771 1520
441 cub-BI [76,109] 0.4139 [74] 6.11 177.971 2177 I77.971 8.1 [I061 85 [97l 1560 HV
442 cub [I121 0.413 [lo91 6.13 ilOSl . 2050 SO61 200 [I061
443 0.413916 [I121 6.13
444 6.1
452 ZrN fcc [I] 0.46 [I] 6.93 [61 3000 [Il 6 [I] 16.75 [I] 13.6 [6] 365.5 [I] 5.1 [9l 2000 111 1 2 [ 7 1
453 cub-BI [6] 0.458 [3] 7.09 [7] 2980 [5.7,8] 7.9 [8] 10.9 [8] 21 [7] 336.2 [5] 5.1 I451 1300-2000 [7] 11-14 [25]
454 fcc [8] 0.459 [6] 7.32 [9] 2980 [25.38] 7.2 [9] 28 [9] 21 [9] 365.5 [I21 3.93-1.60 [74] 1900 HV [XI
455 cub-BI [lo] 0.4577 [lo] 7.3 (x) [lo] 2980 [39,90] 7.24 [lo] 20.52 [lo] 13.6 [25] 368.4 [46] 5.1 [77] 1600 19.591
456 cub-BI [25] 0.456 [25] 6.93 [25] 2982 [6.9.10] 7.24 [32] 20.52 [32] 21.1 1321 366 [65] 5.1 [93] I500 [10.32]
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
457 cub-BI 1461 0.4575 7.3 [38] 2982 [32,45] 6.5 1381 19 [38] 20 1381 365.5 1731 5 I [97] I510 HK 0.1
458 cub-BI [52] 0.4577 7.09 1391 2982 [77.93,97] 7.24 1451 21 1571 1520 HV
459 fcc 1571 0.4562 7.32 1451 2930-2982 1291 7.2 [77] 21 1661 2300
4Ml fcc 1741 0.4578 7.09 i74;106] 2950 (463 7.2 1931 21 1771 1520
461 cub-BI [76] 0.463 7.32 [77] 2955 157) 7.2 1971 21 1971 1250HV 0.1
462 cub-BI [I091 0.457756 7.09 [90] 2980 [I061 7.24 11061 14 [I061 1500-2600
463 cub [I121 7.32 [93] I600HV
464 7.32 [97] 2500 HV 0.05
465 7.09 [in81 2130 HV0.05
466 7.4 [in71
Borides
467 AIB? hex [52] 0.3@86/0.3252 3.17(x) (521 1975 1461 67 [46]
468 hex [I 121 0.30054;0.325276
469 AIB,, tet 1521 O.lZ58jO.lO20 2.57 1521 2070 (per) [5] 10” 1661 201 [46] 4.3 1771 a: 2400 HK 1581
470 tet [I121 0.1030/0.1433 2.58 1771 2070(per) 1461 2 x 10’2 1771 2600 HV 1771
47I 1.01611.428
472 mono 1521 1.884
i.704~1.100~ 2150(d) [77]
473 0.8522~1.100/0.7393
474 CrB ortho [6] 0.296910.7858/0.~932 6.05 [6] I550 [6] 1251 -75.4 1461 1 6 1 8 [25]
475 ortho [25] 0.296910.785810.2932 6.05 [25,52] 1550 [25] 75.4 1731
476 ortho 1521 0.296910.785810.2932 6.11(x) 1521 I550 1731
477 ortho [I121 0.29663/0.78666,
0.29322
478 CrBl hex [I] 0.279I0.307 5.6 [1.106] 2200 [I] 11.1 [I] 56 [7] 94.6 [5] 2.15 [I] 2250 (1.571
479 hex [6] 0.296910.3066 5.6 [6,107] 2170 [5] 11.1 1571 21 [25] 129.8 [6] 2.15 1571 1700 161
480 hex [25] 0.296910.3066 5.6 [7] 18OC-2300 [6] 10.5 1771 56 1571 -94.2 1461 5.4 1771 2200 [71
48I hex I521 0.296910.3066 5.6 1251 2150 [7] 10.5 1971 I8 [66] 94.2 [73] 18on~~o.i [251
482 hex I571 0.29730/0.30709 5.6 1571 1850 [25.106] 7.1 [I061 18 1771 1800 WI
483 hex [I121 5.58 1771 2200 [571 I8 1971 2250 HV 177.971
484 5.58 191 1850-1900 1731 21 [I061 2100 [591
485 2188 177.971
486 Cr$, tet [I] 6.1 [I] 1900 [I] 125.6 [I]
487 tet (61 6.12 [6] 1960 1251
488 let 1521 6.14 1251
489 tet 11121
490 FeB ortho [52] 6.3 [52] 1650 [5,46,73] 71.2 5.46.73 14M1190 171
49I ortho [I091 1550 1611
492 ortho [I 121
493 Fe,B let I521 -7.0 1521 1389 [46,61.73] 71.2 46.73
494 let [I091
495 tet [I121
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
Structure parameters point expansion, u. conductivity, X resistivity modulus hardness resistance
(nm) (g cm-’) (“C) K-’) (W m-’ K-’I ncm) (kl mol-’) (lo5 N mm-*l (10 N mm-’) (100°C)
498 HIB? hex 125) 0.3141/0.3470 1251 11.01 [7] 3200 5.3 [571 430 1981 I0 1251 336 6 [5] 2800 171 11-17 1251
499 hex 1521 0.3141/0.3470 1521 11.2 (25,1071 3380 5.7 19x1 15.8 1571 336.2 [46] 2900 [571
500 hex [57] 0.314245/0.347602 [I121 10.5 [52,108] -3060 5.3 [I061 12 (1061 336.2 (731 2900 [59
%I hex [I121 11.2(x) 1521 3370 2850 PSI
502 11 1571 3250 400M6000 HVO.O1 [ill
503 10.96 [I061 2147
504 3240
505 LaB6 cub [IOY] 0.415 IW 4.76 [YO] 2530 28 (1061 Z770 HV Pol
506 cub [I121 0.41569 11121 4.7 [l06,lO7l 2200
507 2.61 I1081
508 LaB, cub [52] 0.4156 1521 4.73 [7] 2530 5.84 130) 17 171 1400-2500 I71
509 cub [55] 0.4145 1551 4.76 1391 2200 6.4 [77] 15 1771 2770 HV 1-19]
510 0.4156 1551 4.73 [77] 2770 2530 HV 1771
511 MOB-6 tet 1251 0.31 10/1.695 1251 8.3 [ZS] -2500 50 [25] 1570 HV 0.1 1251
512 let 1521 0.3110/1.695 ~521 8.77 (x) [52] 1930 50 (571 1570 P31
513 te1 [I121 0.3105/1.697 1521 8.3 [52] 2350 25w 1571
514 0.3108/1.697 [I121 8
515 MOB-p ortho [25,52] 0.316!0.861/0.308 [25] 8.5 [571 2500 [591
516 ortho [57] 0.316/0.861/0.30X (521
517 ortho [I121 0.316/0.845/0.308 [I 121
518 MOB> hex [I] 0.3/0.31 [I1 7.8 [I] 2100 45 [25] 96.3 [I] l38DHV D.l (27 11-14 125)
519 hex [25, 521 0.305/0.3113 [251 7.12 [7] 2300 30 [57l 3000 [571
520 hCP 1571 0.306/0.310 1521 8 [25] 2250
52 I hex [I 121 0.304/0.307 ill21
522 Mo,B tet [I] 0.55/0.47 Ill 9.2 [I] 2140 40 [25] 106.8 [I] 2500 111
523 tet (251 0.5543/0.4735 125.521 9.1 [25] 1850 1660HVO.l ~ 5 1
524 te1 152, 1091 0.5547/0.4739 [I121 9. I [521 166C 1331
525 let (1121 0.554/0.474 11091 9.31 (x) [521
526 Mo,Br hex [I] 0.3/2.1 [I1 7.2 [I] 2300 8.6 [77] 18 1771 209.3 [I] 6.7 (771 3200 “I
527 hex 1251 0.3011/2.093 1251 7.48 (x) [52] 2140 8.6 [97] 18 [97] 6.7 [97] 2350 HV 1771
528 hex [52] 0.3011/2.093 1521 7.45 (77,971 2140 5 IW 25 [ID61 2350 HV P71
529 hex [ I I?] 0.30 I I74/2.09369 [I 121 7.12 [I061 2100
530 7.8 “071
531 Mo,B2 tet [I] 0.60/0.31 Ill 9 [I] 2240 175.8 [I] 2300 Ill
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
532 NhB ortho [6.25] 0 3298 0.X724 0 3166 [h. 251 I67 5 [6] 2200 [61
513 ortho [52] 032Y2!0.8713.03165 [5?)
534 ortho (1121 0 32973 0 87229 /I121
031883
535 NbBz hex [I] 031’033 “I 68 (I] 3000 7 1’96 [6] 1675 [I] 32 1151 150.7 (I] -2.6 [6] 2600 [1.6.57.59]11-14 1251
536 her [a] 0.3099,0.3271 I61 72 [6] 3000 [77] 16 75-35 2 [6] I? 1571 175.8 IS] 63 [77] 130&2600 [71
537 hen 1251 0.3086 0 3306 [28] [7. 571 3036 [97] 16 75 [?5] I? 1771 251 2 [46] 63 [97] 1800-1900 [281
538 hex [82l 0 3096,0.3306 (521 66 (35,521 -2900 h? (1061 16 7 (1061 I? (971 251 2 [731 2600 H V 1771
539 hex [57] 0.311133~032743 [Il?] 7.21 in) [52] 3036 32 [I061
540 hex [I121 6.98 [77.97] 3036
54 I 697 1106, 1081 2900
542 SiBh ortho [I121 1.4470 I835009946 [I121 243 [I] 1950 83 10’ [hb] 3.3 [77] 1910 [11
541 247 16. 1081 1947 10’ [77] 33 [97] 245&?800HV [39]
544 243 [3Y] 1900 54 10’ [Y7] 2300 H K 0 I 15x1
545 243 (701 1900 2400-2800 H V 0 I [70]
546 2.43 1771 m m HV (77. 971
547 TaB, hex [I] 0.31 0 33 [1,87] 3150 51 [I] 21 35 [6l 68 1251 209 3 [5] 2 62 161 2200 [1.57] 11-14 [25]
548 hex [6] 0 3085 0 3249 I2 I (x) [h] 3100 7 88 [6] 10 89 [25] ?I 1571 193 8 [6] 68 1771 2615 HK 0 I [61
549 hex [25] 0 3078/0.3265 I2 58 171 3040 82 1571 109 [I061 14 1771 209 3 [46] 2500 [7. 591
550 hex [52] 0 309803 0.322660 11.7 [25] -3000 82 [771 68 [I061 209 3 [73] 210&2400 [281
551 hex 1571 11.7 [521 3100 54 [I061 2500 H V [471
552 hex [I121 1?.6(x) (521 3150 2000-2700 [761
553 12.58 [771 3037
554 II I5 [106. 1081 3000
555 I2 6 [I071 2lOOHV [771
564 Ta,B, ortho [6. 251 0 329 I 4010 313 (6.25. 521 13 5 [6. 521 3350 1571
565 ortho [52] 032914010313 [I121 I36O(x) 16.521
566 ortho [I121
567 TiBl hex [I] 0 310 32 [I] 4.5 [1.107] 2900 6 39 [I] 2596 [I] 9 [71 1507 [I] 3.7 [I] 3480 [I] 11-17 [25]
568 hex 16) 0 302810 3228 [6] 4.38 [h. 1061 2900 78 191 26 13 16) 7 191 324 I 151 56 [9] 300&3400 171 13 [43l
869 hex [25] 0.3026’0 3213 [25] 4 38 [71 2920 4,8419.83 [30] 26-39 [9[ 7 [31] 280 [43] 5.6 [31] 250&3500 191
570 hex [30] 0.3028’0 3228 [52] 45 [9. 1081 3230 7.8 [31] 26.13 [25] 10 [38] 324.1 [46] 37 [38] 3400HVO05 1251
571 hex 1311
.. 0.303034 0 322953 [74] 44 [25] 3225 7.4 1381 27 1381 7 1401 305 1481 4.8 1401 3000 H V 131.97.981
572 hex [43] 03 [YX] 45 [31] 3225
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal structure Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
parameters point expansion, a conductivity, X resistivity modulus hardness resistance
(nm) (g cm-3) ("C) (10.' K-I) (W m-' K-') (lo-' Qcm) (kJ mol-') (10' N mm-') ( I 0 N rnm-,) (100T)
573 hex 1521 0 303034,0.322953 4.5 3225 [77,97,98] 8 [431 24-56 [48] 9 1511 324.1 1731 4.8 I451 3400 1331
574 hex [57] 0.3018-0.3023/ ... 4.5 2980 [39,40.76] 7.8 1451 26 9 1981 I (66) 150.7 1981 5.6 1481 3300 HV (39,90,77]
575 hex (741 .../0.3194 322 4.52 2980 [87,90] 6.618.6 (481 26 I (1061 7 (771 3.74 1571 3370 140. 761
576 hex 1981 4.5 2980 1401 6.39 [571 7 [871 4.8 [I61 3000 I451
577 hex 11121 4.52 3197 [43] 8 1761 7 1971 5.6 171 2900-3700 HV 1471
578 4.38 3325 1451 7.8 1771 7 1981 4.8 1x71 25W3000 1481
579 4.5 2790 [I061 8 1871 15, 5 [I061 5.6 1971 3480 157
580 4.5 7.8 1971 5.6 [98] 3300 1591
581 4.5 7.8 1981 3370 HV (871
582 4.4 5.9 IW 34x0 HV 0.2 1881
583 164631 IS HV 0.05 (1231
584 IR43HV0.05 [I191
585 2 2 W 9 0 0 HVO.O1 [I271
589 VB, her [I] 0.310.31 48 2400 L5,1 5.3 203 9 (731 5.1 1771 2080 [I, 571 13 I71
590 hex 161 0.2998/0.3057 4.56 2450 161 5.3 5.1 1971 2077 (61 8-14 1251
89 I hex (251 0.2'99810.3057 5.1 2100 [25] 7.6 13@2100 171
592 hex 1521 0.2998jO.3057 5. I 2400 (571 76 2100 i5Yl
593 hex [57] 0.29976110.305620 4.61 2747 177,971 6.1 2150HV 1771
594 hex 11121 5 10 (x) 2120 HV 1971
595 5
596 5.05
597 4.92
598 WB let [I] 031L7 15.5 2860 [I. 251 3750 111
599 te1 1251 0.311511.693 15.3 2685-2920 161
.. 3750 1571
600 6: tet i 5 i 1121 0.3115/1.693 15.3 i52j 2860(d) 1571
601 p: ortho 1251 0.319/0.840/0.307 1251 16.0 (x) [52]
602 ortho 1521 0.319/0 840/0.307 152) 15 5 1571
603 onho (57) 0.311655/1.69101 [I121 15.7 [I071
604 10.77 I1081
605 W,B tet [I] 0.56047 Ill 16.5 [I] 2770 [I] 4.7 [I061 21-43 [I061 2350 [I] 8-14[25]
606 tet 1251 0 556410.4740 1251 16 125,521 2670-2780 161
607 tet 1521 0.5564/0.4740 1521 16.72 (x) [52l 2710 1251
608 tet (1091 O.S56!0.474 IW 10.77 (1061 2900 [I061
609 tet [ I 121 0 556810.474 [I 121
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
610 WB: 1513 [I) 2920 [7] 2600 171
611 W?B< hex 03’1.3Y 13 I [I] 2300 [I,571 78 !I 1571 77 1771 2700
612 hex U.?YX? I S87 II (521 29x0 PI 78 19 1771 7.7 [97] 2700
613 hcp 0 2982 1.387 13 1 IX) [52] 2200-2980 161 IY [9i] 2700HV
614 hex O:YX~?O~~S 13.03 [J7,9?) 2365 177. 97
618 ZrB, hen [I1 0 32)O35 61 [I, 38. 571 3000 (1.30. 1061 6.83 3.5 2200
619 ha IM 0.3170 o 3531 5 64 [6. 71 3040 [S. 39.901 6 I2 3.5 2600
620 hex 0.316910 3530 6.17 ps. 521 2992 “4 6.8 3.5 22Ml HV 0 05
62 I hex 0 1170~0.7533 6.17 1391 3250 171 59 54 2252 HV
622 hex 0 32 -6 (471 2990 [25. 38. 57.881 6 83 54 2200 HV
623 hex 03l6870~0353002 6 09 ( X I 1521 3200 (461 59 54 965 H V 0 I
624 hex 6 II 177. 971 3245 [77,97. 981 59 2300
625 hex 6.11 1981 5.9 2300-3000
626 6 PXl 55 2300 H V
627 6 17 I901 2200 H V 0 2
628 61 IW
629 6 0x5 (1081
630 5 64 lW
631 ZrB,? cub 0 14 3 65 I 2 I4 1251 502.4 [I] 2500
632 fcc 0 7408 3.7 2500
633 cub 0 7408 3.7
634 cuh o 7408 3.63 (1)
Silicides
635 CrSi cub I11 0.462 5 38 [6] 1550 [I] 53.2 1461 1000
636 cub 0 4619 1550 (61 54.8 [73] 95&1050
637 cub 0 462 1475 [63]
638 cuh 0.46074 4629 1457 (pa) 1731
639 cub 0 462
640 CrSi? hex “I 0 44?:0 655 491 [61 1630 161 100.5 [I] 1100
64 I hex 0.443 l/0.6364 4.4 1251 1451 146) 80 1461 xw110
h42 hex 0.442210.635I 4.91 [SZl 1570 I251 80.1 1731 I IS0 HV 0.1
443 hex o 4427’0 6375 4.4 (1071 1457 1731
644 hex 0 4420,O6349 5.5 IW
635 hex [II?] 0.44281 0.63691
646 Cr1Si cu h 0.455 6.52 (61 1710 [I] 105 5 [I] 90&980
447 cub 0.45W 6.52 1251 1770 151 105.5 IS] 1005
648 cuh 0.4555 6.45 1521 1750 [61 92.1 146)
649 cu h 0 455W.4564 1770 146. 63. 73) 105.5 (731
650 cub 0 4558
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
~~~~ ~
No Symbol Crystal Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
structure parameters point expansion, 01 conductivity, X resistivity modulus hardness resistance
(nm) (g cm+) (“C) K-‘) (W m-’ K-’) ncm) (kJ rno1-l) (10’ N mm-*J (10 N mm-’) (IOO’C)
651 Cr3Si2 5.6 [6] 151&1610 (61 1420 [7] 10Sc-I2W I61
652 4.7 171 I550 1000 I71
653 Cr,Si tet [I] 0.919/0465 111 5.6 111 1560 [I] 1280 “I
654 Cr& tet [6] 0 9170/0.4636 1251 5.9 161 1647 1461 211.4 1461
655 tet 1521 0.9 I70,’0.4636 1521 5.6 1521 I647 [73] 223.2 1731
656 5.9(xJ 1521
657 LaSi2 tet 1301 0.4281/1.375 11121 -1520 [30] 10.78/17.05 (301
658 tet 11121 > 1500 163)
659 MoSi2 tet [I] 0.32/0.786 [I] 6.3 [1,107] 2050 [I,7,22.30] 8.4 [I] 221.9 [I] 21 [71 108.9 [I] 3.84 [I] 1290 III 17171
660 tet 1221 0.32/0.65 122) 6.31 17, 1081 2030 [25,39. 73,901 8 1221 50 122) 20 1221 131.9 1461 4.4 [22) 1300 17,221 > 17 I251
66 I tet [25.28] 1).320/0.786 1251 6.2 1221 2020 1501 5.32115.55 1301 21.5 125) 131.5 [SO] 1290 H V 0 I [251 17 I1001
662 tet [30.52] 0.3205/0.7848 [28] 6.12 1251 2190 (per) 1631 131.8 1731 1200 H V (39. YO1
663 tet 1461 0.3203:0.7886 1521 6 [3Y, 901 900HV0.1 1581
664 tet [I091 0.320/0.786 IIW
665 tet [I121 0.32047/0.78449 11121
666 MolSi cub [I] 0.489 Ill 8.8 111 2150 [I] 100.5 [I] 1310
667 cub 1251 0489 1251 8.4 [25] 2050 [25] 116.4 [5] 1310HV0.1
668 cub 1461 0.489 1521 84 (521 2025 1731 116.4 (461
669 cub 152, 1121 0.489 [I121 8.97 1521 116.5 I731
670 Mo,Si, tet [I.1121 0.96410.49 III 7.8 Ill 2100 [I] 280.5 [I] 1170
67 1 cub 1461 0.96210.490 1521 2180 [SO] 310.2 146. SO]
672 tet 1521 0.96483/0.49135 11121 2190 1731 309.8 1731
673 NbSi, hex [I] 0.4810.66 [I1 5.5 Ill 1950 [I] 8.4 161 6.3 1251 50.2 [I] 700
674 hex 161 0.4785/0.6576 161 251 5.69(x) 161 1950 [7] 35.6-67.8 [6] 66h1320 H V
675 hex 1251 0.4797/0.6590 1281 5.7 171 1950 1251 138.2 (461 lOSOHV0.1
676 hex [28, 521 0.479506589 1521 5.29 1251 1930 (521 138.2 (731
677 hex [I121 0.47971/0.6592 [lli] 5.37 [losl
678 5.6 [I071
679 Nh& hex [I] 0.36j0.50 111 7.8 111 1950 [I] 550
680 hex 161 0.35910.446 161 8.01 161
681 Nh& 01: tet 161 3:0.6570/1.1884 [6] a: 7.09 (x) 161 2480 151 0: 7.3/4.6 161 452.2 151
682 p: tet [6] p l.Ml0/0.5070 [6] B: 7.?O(xJ [6] 2480 [63] 452.2 (731
683 3:tet 152,721 a: 0.658311.1884
684 p: tet 152,711 p: l.OOl0/0.5070
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
685 tet ]I121 I1 6569Rll.lX87
686 TaSi2 hex 161 0 4773'0.6552 92 [7] 2200 8 Y18.8 161 38 1507 1410 161 11 171
687 hex [25] 0 4773,0.6552 Y.O?(x) (61 2400 46 1193 I200 17) 11-14 1251
688 hex (28.521 0 4783106567 883 (251 2200 8.5 1192 1560 HV 0 I 1251
689 hex 146) 0.4781,O 6564 9.14 152. 1081 I560 1331
690 her (I 121 0 47835110.656980 8.4 (A) (52]
69 I 9.1 [107]
692 Ta,Si iei [I] 061,050 13 5 [I] 2450 I24 83.7 I500 111
693 let [a] 0615705039 13 54 (xi (6. 521 126 4
694 tet 1521 06157,05039 I? 4 125 6
695 tet [ll?] 06157~05039
696 T a , S let (61 TI o 9880:~5060 2500 1631 T2 5 5'8 0 (61 I08 335 4
697 let [6] TZ.0 6516 I 1873 2500 1731 334 9
698 let (461 TI~O9820,05010 334 9
699 tet 1521 T206513,I 1864
700 tei Ill21 0 6516, I 1873
701 TiB! ortho (251 0823604773'08523 (h.251 4 39 [90] 1520 18 1344 892 HV (35. 901 II [?I
702 ortho 1281 0 8263'0 4800'0 8553 12x1 4 17. 1081 1470 123 I34 3 700 171 8-1 I 125)
703 urtho (521 0 8253,O 4783'0.8540 1521 4 39 125. 391 1540 134.4 870 HV 0 I
704 ortho [IOO] 0 8267 0 4800,O 8551 11001 44 [I071 -1527 870 HV I
705 ortho [IIZ] 0 82687 0 85534 407 (1001 -Is50
0 47983
706 1480
707 TI,% her [I] 0 75 0 52 43 [I] ?I20 7.03 15 61 1301 577 8 986
708 hex (251 074650516? 4 32 [loo] 2120 9.2 [I001 579 986 HV 0 I
709 hex 130.521 0 7465 0 5162 2130 579 9 968HV I
710 hex (I001 0?44805141 -2150 579. I
711 hex [IIZ] 07444,05143 579 5
723 WSil let [I] 03210788 .. 9.5 Ill 2165 11. 391 6.5 [??I 45 1221 125 92 I 1090
724 tet (61 0.32 0 78 (221 9.4 [7: 1081 2170 [7] 10 92 9 1100
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal structure Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
parameters point expansion, u conductivity, h resistivity modulus hardness resistance
(nm) (g cm-9 ("C) K-') (W m-' K ' ) ncm) (kJ mol-I) (10' N mm-') (10 N m K 2 ) (100°C)
725 tet I221 0.3212/0.7880 1251 9.8 1221 203W2W I221 33.4 1251 92.8 [73] I200 [221
126 let [25,28] 0.321310.7829 1281 9.3 125, 107) 2150 [25.46] 1090HV0.1 1251
727 let 146,521 0.321 1/0.7868 1521 2160 1731 1074 HV 1391
128 tet [IIZ] 0.3211/0.7829 [IIZ] 2210 1901
729 W& tet [I] 0.561,10.496 Ill 14.56 [I] 2320 [I,731 125.6 [I] 770 Ill
730 tet 161 0.96410.497 (521 13.06 (100] 2350 (99,631 134.4 151
73 I tct [5?1 0.960110.4972 Ill21 135.2 146)
732 tet [I121 134.6 1131
733 ZrSi hex 146,521 0.7005/1.2772 1521 5.65 (x) [52] 2090 (461 141.8 1461
734 onho [SZ] 0.6982/0.3786/0.5302 [52] 2107(per) (731 154.9 1731
735 ortho [ I 121 0.6981!0.3785/0.5301 [I121
736 ZrSi, ortho [6] 0.372/1.41610.367 16,251 4.87 [6] 1700 [25] 9.7 [loo] 161 [25] 159.4 (IOO] 2.348 [IMI] 1030HV0.1 (251 8-11 [ZS]
737 ortho [25,28] 0.3698/1.4761/0.3665 (281 4.88 125, 1081 1520 [I001 151.1 1631 942 HV 1 IW
738 ortho [52, 1001 0.3721/1.468/0.3683 152, 1001 4.83 [lo01 1517(per) 1731 159.5 173)
739 ortho [I121 0.36958/1.4751/0.36654 [I121 4.9 11071
140 Zr2Si tct [I] 0.66/0.54 Ill 6 [I] 2110 11,461 309.8 [I] 1230 111
74 I tet [46, 521 0.661210.5294 1521 6.22 152) 2107(per) I731 339.1 [46]
742 bl [I121 0.66oS/0.5298 [I121 208.5 (731
743 Zr5Si, hen [46] 0.7886/0.5551 [52] 2150 1461 614.2 1461
744 hex (52, 1121 0.7885j0.5558 [I121 576.1 I731
Oxides
745 Al?O,-a hex 0.5127 - a = 55'16.7' 1521 3.99 2043 8 [31 30. I [XI 1580.I 151 4 [7] 2100HV0.05 PI > 17 1251
746 hex 0.512 1571 3.98 2045 8.6 [a] 27-36 (Y] 1590 (431 4 (91 2200 HV 141 18 1431
741 rhom 0.513 --u = 55.3' I1091 3.9 2030 7.2-8.6 I91 4.2-16.7 1161 1678.5 3.6 1161 ZIOOHV [8. 77,87,98] 20 1701
748 rhom 0.554410.9024 I1121 3.99 2047 7-9 [I61 25 I381 1690 3.5 I251 IXW?XK! I91
149 rhom 0.47588i/1.2992 ju~j 3.8-3.9 2015 9.5110 (301 27 1431 1676.4 4. I 1381 250W3000 HV 0.05
750 hex 3.98 2050 8.3 1381 2s [57] 4 1401 ?Ow HK 0.1 1331
751 hex 4 2050 9 1401 27 [h3] 4 [.I31 2800 133. 691
752 rhomihex 3.8 2050 8 1431 28 167) 4 [A51 1800 1341
753 hex 399 2041 7.7 [ q 30.1 19x1 X.? IS31 28WHV
754 hcx/rhom 3.96 (x) 2300 6.8 1531 35 (1061 2.5 (531 2100
755 3.98 2047 ?: 8.5 1561 27-36 [I211 5.2 ph] 2020HK
756 3.96 2040 7.2 .rj7i. 4 (h?] 18WHV
757 3.97 2050 8.4 1621 4 16-31 2100
758 3.95 2050 8 3.7 (h71 ZOSOHV
759 3.99 2054 8.5 37 [hY] 1200-1600HV
760 3.96 2015 9 4 1701 250&3000 HV
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
761 3.98 [77] 2050 4 I761 2000
762 3.9 188) 2300 4 1771 ZIOOHVOI
763 398 (981 2047 3.8 [82] 20W2300 HV
764 3.98 (92) 23M) 4 [871 2080 H K 0 I
765 4 “5] 2050 38 [95] 2000-2500 HV
766 3.99 [lO6] 2047 4 1981 1200-1600 HV 0 1
767 3.9 [I071 2046 2.5 [92] 2300
768 3.9 [I211 2045 294 [I061 20w2500
769 2045 400 (121) 855-975 HV 0 02
770 2000 HV 30
771 15OLL2200 HV 30
772 2000-3000 HV 0 I
173 23OO HV 0 2
774 2000 HV
775 2600-2800 HV
776 A120,-? cub (svtnellj 1510HK
[521
777 fcc 0.3958 1521 2200
778 cub 0.79 lll2l
779 cub-El o 774.8 iiosi
2450 9 1431 264 (431 lo?’ (661 569 3 [?5] 1230-1490 HV
780 Be0 hex 0 269910.4401 [52l 3 (251 IS] [S]
781 hex 3 03 L39.771 2530 [?5] 9 1771 10’’ (771 580 [43] 3.9 [43] IS00 HV
782 301 1431 2570 1391 608.8 (461 3.9 [77]
783 3.01 1521 2567 [43] 599.1 PI
784 3 00(xj 1521 2580 [45]
785 2550 1771
788 CrO, ortho 0.573/0.852/0.474 [6] 2.81 [6] 170-198 [6] 579 9
789 ortho 0 574310.8557/0.4789 1521 2.7 1521 185 [46] 589.9
790 ortho 0 57494,’O 8556’0.4796 [I121 2.82 (x) [52] 578.6
791 Cr,O, rhom 0.536-%=55’ [6l 5.21 [6] 2440 [7] 6.7 [53] 10” [6] 1130.4 1000 HV 1341
792 rhom 0.5361 -rr=55’ I521 5.2 [7] 2300 [39] 6.7 [60] 1.3 x lo9 [I061 1140.5 2915 HV 1531
793 hex 0.495876/1.35942 [I121 5.21 [39] 2400 [46] 5.6 [I061 1130.4 2300 HV PI
794 5.22 1521 2343 [53] 12W1700HV 1531
795 5.25 (x) [52] 2343 [a] 12W1700HV0.1 IS81
796 5.41 1531 2708 1106) 2000 1591
797 5.4 1601 1200-1700 HV0.05 [60]
798 5.21 [I061 I100 HV 0.02 [64]
802 Hf02 mono 0 512 1521 9.7 [7] -2900 15.471 10 [43l 3 [43l 5 x 10’’ [I061 1053.4 900 [591
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No Symbol Crystal structure Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
parameters point expansion, a conductivity, X resistivity modulus hardness resistance
(nm) (g cm-7 (“C) K-’) (W m-’ K - ’ ) ncm) (kl mol-’) (10’ N mm-’) (10 N nim-’) (100°C)
803 mono [46] 0.51 [I091 9 68 [43] 2790 [7] 6.5 1771 1010 [43] 780 H V 1771
804 cub [IW] 0 5281~0.51819~0.5115[I121 10.2 I771 2897 [43] 1113.7 1461
805 mono [I121 9.63 [I061 2810 [46] 11138 [731
806 96 [I071 2790 [I061
807 MkQ cub [43] 0.4208 1521 3.6 [7] 2850 [5] [43] 36 [43] 10” [66] 568 6 [5] 3.2 [43] 745 H V
808 cub-El [461 0.42 [IOY] 3.5 [25, 1071 2800 [7.25.39. 531 [771 36 [I061 10” [77] 570 I431 3.2 [771 520HV
809 cub-El [52] 04213 [I121 3 65 [391 2837 (431 II 2 [I061 10” [I061 601.6 [46] 2.4 [I061 -400HV
810 cub-Bl [IOU] 3.58 [43. 531 2825 [46. 731 601.6 [73] 520 H V 0. I
811 cub [I121 3.77 1771 2827 [77] 750 H V
812 3 58 (106. 1081 3073 [lob]
813 Nb!O, mono [I121 2893/03827;1.758 [I121 4.5 [7] 1460-1520 [61 1.76 [6] 1905.8 161 740
814 447 [I061 1510 171 1903 3 [73] 700
815 4.6 [I071 IS12 (731
816 1785 [I061
821 SiO, quartr- quartr 233 [6] 1703-1729 [6] 04 [I061 I38 [I061 10;’ [hh] 911 [46] 0.5-10 [82] 113&1260 161
822 trigonal [I091 0.4093,’O 5393 [6] 2.2 [7.105] 1713 [7] 05-075 [I081 12-1.4 [I081 10.’ [I061 911.5 [73] 1.114 [I051 1200 1591
823 0 421.’0.539 [I091 162.7 [47] 1722 [46] H IIIS] 0 654.75 [I081 IOW1?00H V IS?]
824 2 1-2.2 (471 1373-1473 [73]
825 CriStoballtK: cristobalite 2.2 11061 1713 [IOS]
826 t ~ t I61 0.70,0.69 [6] 2.6 [I071 1983 [I061
827 cub [IOU] 0.704 [IOY] 2.18 [I081
828 Ta,O,-u tet [6] 0.3808 0.3567 [52] 8.37-9.48 [6] 1880 (51 2 161 1953 I IS] 66&1030 1991
829 tKt [52] 75 [7] 1887 [h]
830 8.53 (x) 1521 1880 [71
83 I 82 [99]
832 TdlOr-p ortho 161 0 619 0 367 0 389 [6] 8 18-891 [6] 1785 [6] 2 161 10” [h] 2045 3 [h]
833 ortho (461 0 6192 4 4019 0 I898 (521 8.30(xJ [52] 1900 (461 2047 3 [46]
834 ortho [5?] ?047 3 [73]
835 Pd!Ol ortho [II?] 0 6198 4 029 0 3888 [I121 83 [I051 1880 11061 [ 1061
836 7.53 (1061 1880 (IOS]
837 8.7 [I071
838 Tho, fcc [6] 05859 [6] 10.05 (x) [6] 3250 [5] 10 1431 10 1431 1016 [6h] I I71 I [5] I 3X [h] 950 H V
839 cub 143) 05595 [52] 9.7 [7] 2997-3250 [a] 9.3 [77] 10 [I061 10’‘ 1771 11711 (431 14 I251
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
840 fcc [46] 0.521 Y7 1251 2990 2 x 10" [I061 1227 5 1461 24 [431
841 fcc [52] 0.5597 10 [43] 3050 1227 2 1731 24 [771
842 cub [ I091 10 [??I 3217 148 [I061
843 cub [II?] YY [I061 3370
844 10 [I071 3220
R45 3300
846 3323
847 Ti0 cub-Bl 16.461 0417 4 88 161 1750 [6. 7. 32. 1051 7 6 [I061 II [I061 [32] 520 [??I 1100
848 Cub [II?] 0418 49 [7. 1051 2020 I461 1800
R4Y cub-BI [52,76] 0424 493 [I061 1700 11061 I OIIO
850 cub-Bl [IOY] 04177
851 Ti0: tet [4h] 0 4543:O 2959 4 19 [6l 1900 8 [Y] I.2 x lo"' [I061 945 4 I461 2 05-2 80 [Y] 767 I000 HK
852 let 1521 0.449,'o 2959 4I [7] 1855 5 [lox] 945 4 I731 2 05 1771 7oo I Ino
853 cub-BI [25] 045933'0?9592 425 [Yl 1867 0 8-20 [X?] 1~IOOHV
854 tet IW 2.40-2.49 1411 IXbO 205 [98] 600 1000 HV
855 tel [II?] 4 16 [53] 1840 065 I 4 9 11241 600- 1000 HV I1 I
856 425 [77.98] 1867 I100 HV
857 42 [88] 1860 600 7on
8% 424 [I061 1860 1000 HV 1) 2
859 42 [105.107] 1913 II50 HV
860 3.02-3.92 [I241
861 TI~O, rhom 1461 0 5454 - x = 59 5' 4.6 [5?, 1051 2130 IO'Y-IO:' [lox] 1433 I [I] YXO HV
862 rhom [52] 0 5139,l 3659 3.58 1x1 [52] 1842 IS?I 9 [46.73]
863 rhom [I121 4.05 [I081 1760
866 WO? mnno 16.46, 521 0 5560,O 4884,O 5546 [52] I? I [6] 1500-1600 161 (1-1. 5 ) x [hi 587 [61
10"
867 tet (521 0 4870,0.2776 [52] 11.05 1521 1724 (per) I461 590 I I461
868 tet [IOY] 0 55610.555 11091 1082(x) [52] 1724(d) [?)I 590 I I731
869 mnnn 11121 0.55754~048995.' [I121
0 55608
870 wo, 111 [6] 0.382 0.748)O 728 [6] 7 16 [hl 1473 161 0.6 [I061 843 4 1461
87 I mono [46. 521 0.3835/0.7517/0 7285 [52] 7 16 [7. 1061 1473 17. 1061 843 5 1731
872 tet 1521 0 525010 39 I 5 [52] 1472 [46]
873 Ill 11121 07309107522:07678 [I121
874 Y?O; bcc 1461 1.0601 [521 5 [7. 1051 2465 151 8 [iXl 14 1381 1906 7 1461 I8 [38]
875 b€C I521 1.06 [I091 4.5 1381 2415 [7] 81 [lox] XI? [I081 IY067 1731
a76 cub [IOY] 1.04041 [I121 5.03 [I081 2450 [38]
877 cub [I 121 4 84 [I061 2704 [46]
878 2420 [73]
879 2410 [I051
880 2683 [I061
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 1. Continued
No. Symbol Crystal structure Lattice Density Melting Linear thermal Thermal Electrical Enthalpiy Young Micro Oxidation
parameters point expansion. c( conductivity. X resistivity modulus hardness resistance
(nm) (B cm-7 ("C) (lo-' K - ' ) (W in-' K - ' ) ncm) (kJ mol-') (10' N mm ') (10 N m m - * ) (100°C)
881 Zr02 cub 1391 0.511 1521 5.6 17, 1051 2150 [S] 7.5- 10.5 [9] 0.7-2.4 [9] 10l6 (91 1035 IS] 1.63 [h] 1200 PI > 11 [ZS]
882 cub 1521 0.521/0.5209,'0.5315 161 5.7 19, 881 2681 [7] I0 1381 2 1381 10" 1381 1040 [43] 1.8 [9] I200 HV 139.771 17 1431
883 mono (39,431 0.522Oj0.5211;0.53XI [52] 5.8 (251 2611 (9,43,53,77] 9-10.5 (431 2 (431 loi6 (661 1101.5 (461 1.7 1251 ISOOHV 1531
884 mono 146,521 0,508410.5165 1521 5.56 138,431 2690 125, 391 10 1611 2.Y 1671 10" [77] 1098.2 1731 2.4 [38l IOOOHV 1581
885 tef 1521 0.508 [I091 5.56 (mo)[39] 2680 138,461 10 1771 1.7 [I061 2.05 1431 1500 HK 0.1 1581
886 cub [I091 6.21 (cub) 1391 2700 I881 17.2 [I061 2.1 [67] 1600 I591
887 mono 16, 1121 0.51463iO.52135iO.5311 [IlZ] 6.1 153, 1081 2681 11061 1.9 [771 1300 HV 30 1671
888 tet [I121 0.364/0.521 [IIZ] 6.01 1671 2700 [IOS] 400-1000 1761
889 5.76 1771 I100 HV 0.2 1881
890 5.6 [I061 150&1650 11161
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 99 1
References
I . R. Kieffer and F. Benesovsky, Hartstoffe. Springer, Vienna, 1968.
2. H. Holleck, Binare und ternare Carbide und Nitride der Ubergansmetalle und ihre Phasenbezie-
hungen. Gebr. Borntrager Verlag, Berlin, Stuttgart, 1984.
3. U. Konig, Untersuchung von kathodenzerstaubten binaren und ternaren Hartstoffschichten
zur VerschleiBminderung von Hartmetallen. Tribologie Reibung, Verschle$, Schmierung.
-
Band 9, Springer, Berlin 1985, pp. 282, 287, 296, 306, 31 I , 316.
4. H. Kolleck, Die Bedeutung von Phasengleichgewichten bei der Entwicklung VerschleiBfester
Werkstoffe. Chem. 2. 1982, 106, 216.
5. M. Riihle, Zum technischen Stand der Dispersionshartung - Teil I . Metall. 1982,36, 128 1-1285.
6. Gmelin: Handbuch der anorganischen Chemie. Verlag Chemie: Weinheim 1960. Angaben in den
jermeiligen Bande der Elemente (z.B. Cr, Ti,
7. R. A. Haefer, OberJachen- und Diinnschichttechnologie Teil I: Beschichtung von Oberjfachen.
Springer, Berlin, 1987, pp. 282-294.
8. H. Frey and G. Kienel, Diinnschichttechnologie. VDI-Verlag, Diisseldorf 1987, p. 121.
9. B. Rother and J. Vetter, Plasma-Beschichtungsverfahren und Hartstoflschichten. Deutscher
Verlag fur Grundstoffindustrie, Leipzig, 1992, p. 171-172.
10. L. E. Toth, Transition Metal Carbides and Nitrides. Academic, New York, 1971, pp. 5-7, 80-
81,92,95-97 and 148-150.
11. E. K. Storms, The Refractory Carbides, Academic, New York, 1967.
12. H. L. Schick, Thermodynamics of Certain Refractory Compounds. Vols. 1 and 2, Academic,
New York, 1966.
13. E. Hornbogen, Werkstofe: Aufbau und Eigenschaften von Keramik, Metallen. Polymer- und
Verbundwerkstoflen. Springer, Berlin, 1987, p. 286.
14. H.-J. Bargel and G. Schulze, Werkstoffe.CDI-Verlag, Diisseldorf, 1988, pp. 293-294.
15. V. Demarne and E. Bergmann, Verwendung von Hartstoff- und Hartmetallschichten fur
Schneidwerkzeuge. In I . Int. PVD-Tagung an der TH Darmstudr am 15.116. Murz 1983, Vol.
20, ISBN 3-88607-027-1, THD Schriftenreihe Wissenschaft und Technik 20, 1983, pp. 269-277.
16. K. H. Kloos et a/. Herstellung und Eigenschaften gesputterter Aluminiumoxidschichten. In 2.
Int. PVD-Tagung an der TH Darmstadt am 11.112. Marz 1986, Vol. 30, THD Schriftenreihe
Wissenschaft und Technik, 1986, p. 100.
17. H. Jehn et a/. Beeinflussung der Schichteigenschaften mangetrongesputterter Nitridschichten
durch Substrattemperatur und Sputteratmosphare. In 2. Int. PVD-Tagung an der T H Darm-
stadt am 11,112. Murz 1986, Vol. 30, THD Schriftenreihe Wissenschaft und Technik, 1986,
p. 138.
18. J.-E. Sundgren et a/. Structure and Properties of HfN Coatings grown by Reactive Sputtering.
In 2. Int. PVD-Tagung an der T H Darmstadt am 11,112. Marz 1986, Vol. 30, THD Schriften-
reihe Wissenschaft und Technik, 1986, p. 185.
19. P. Johannsen, Beschichtete HSS- und Hartmetall-Werkzeuge in der GroBserienfertigung. In 2.
Int. PVD-Tagung an der TH Darmstadt am 11.112. Marz 1986, Vol. 30, THD Schriftenreihe
Wissenschaft und Technik, 1986, p. 192.
20. U. Kopacz and H. Jehn, Harte- und Haftfestigkeitsmessungen an Nitridschichten auf
Schnellarbeitsstahl. In 2. Int. PVD-Tagung an der T H Darmstadt am 11.112. Marz 1986,
Vol. 30, THD Schriftenreihe Wissenschaft und Technik, 1986, pp. 217-218.
21. H. R. Stock and P. Mayr, Hartstofheschichtung mit dem Plasma-CVD-Verfahren. In 2. Int.
PVD-Tagung an der T H Darmstadt am 11.112. Marz 1986, Vol. 30, THD Schriftenreihe
Wissenschaft und Technik, 1986, p. 250.
22. A. Petzold, Physikalische Chemie der Silicate. Deutscher Verlag fur Grundstoffindustrie,
Leipzig, 1991, p. 182.
23. D. Kammermeier, Charakterisierung von binaren und ternaren Hartstoffschichten anhand
von Simulations- und Zerspanungsuntersuchungen. Fortschritt-Ber. VDI Reihe 2 Nr. 271,
VDI-Verlag, Diisseldorf, 1992, p. 2.
24. R. F. Bunshah, Deposition Technologies for Films and Coatings. Noyes Publications, New
Jersey, 1982, p. 149.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
992 9 Data Collection of Properties of Hard Materials
25. R. Kieffer and P. Schwarzkopf, Hartstoffe und Hartmetalle. Springer, Vienna, 1953.
26. S. u. Z. Marinkovic et al. Boron Nitride Coatings. Kernforschungsanlage Julich GmbH, 1989.
27. S. u. Z. Marinkovic et al. Titanium Nitride Coatings. Kernforschungsanlage Jiilich GmbH, 1988.
28. B. Lonnberg, Synthesis, Structure and Properties of some technologically important Carbides,
Borides and Silicides. Acta Univ. Uppsala, 1988, pp. 17-21, 26 and 30.
29. T. Wolf, Herstellung und Charakterisierung von TiN, ZrN und HjN. Dissertation Universitat
Karlsruhe, 1982.
30. H. Bleckmann, Rontgenographisclze Bestirnmung der therrnischen Ausdehnung an verschiedenen
metallischen und nichtmetallischen Hartstoffen. Dissertation RWTH Aachen, 1971.
3I . G. Hilz, Zum EinfluJ innerer Grenzflachen auf Aufiau und Eigenschaften mehrphasiger Hart-
stoffschichten. Kernforschungszentrum Karlsruhe GmbH, KfK 5022, 1992.
32. G. D. Brundiers, Herstellung, Aujbau und Eigenschaften von Hafniumverbindungen irn System
Hf-C-N-0. Kernforschungszentrum Karlsruhe GmbH, KfK 2161, 1975, p. 7.
33. P. Schwarzkopf and R. Kieffer, Cemented Carbides. Macmillan, New York, 1960, pp. 138 and
214.
34. A. R. Lansdown and A. L. Price, Materials to Resist Wear. Pergamon Press, Oxford, 1986,
pp. 7 and 99.
35. R. Reinhold and S. Becker, Superharte Schneidwerkstoffe der Spanungstechnik. VEB Verlag
Technik, Berlin, 1982, p. 17.
36. W.-E. Borys, Vergleichsuntersuchung zum Einsatz hochharter polykristalliner Schneidstoffe
beim Frusen. Dissertation Universitat Hannover, 1984, p. 3.
37. W. Konig and H. U. Schemmel, Untersuchung moderner Schneid.ytoffe - Beanspruchungs-
gerechte Anwendung sowie VerschlezJlursachen. Forschungsbericht des Landes Nordrhein-
Westfalen, Nr. 2472, 1975, p. 54.
38. F. Loffler, Eigenschaften von keramischen Hartstoffschichten auf Silicium- und Aluminiumbasis.
Fortschritt-Ber. VDI Reihe 5 Nr. 180, VDI-Verlag, Dusseldorf, 1990, p. 23.
39. R.-J. Peters, Beschichten mit Hartstoffen. VDI Technologiezentrum Physikalische Technologien.
VDI-Verlag, Dusseldorf, 1992, pp. 2-3.
40. H . Freller and H. P. Lorenz, Kriterien fur die anwendungsbezogene Auswahl von Hart-
stoffschichten. VDI-Verlag, Dusseldorf, 1992, p. 22.
41. C. Ribeiro, Silbernes TiN durch Magnetronsputtern und seine Anwendungsbeispiele. VDI-
Verlag, Dusseldorf, 1992, p. 124.
42. E. Erturk and H.-J. Heuvel, Neue Schichtsysteme mit dem Arc-PVD-Verfahren. VDI-Verlag,
Dusseldorf, 1992, p. 254.
43. F. Thummler et al. Fortschritte mit neuen Werkstoffen: Keramik fur den Maschinenbau.
Keram. Z. 1988,40, 158-159.
44. K. Dettling, VerschleiDfeste Werkstoffe fur Extrusionswerkzeuge. Keram. Z . 1988, 40, 802.
45. R. Riesenberg et al. Mechanische Eigenschaften von Schichten. Wissenschaftliche Schriften-
reihe TU Chemnitz, Heft 9/1990, p. 53.
46. 0. Kubaschewski et al. Metallurgical Thermochemistry. Pergamon,. Oxford, 1979, pp. 268ff.
47. C. E. Morosanu, Thin films by chemical vapour deposition. Thin Films Sci. Technol. 1990,7,
pp. 225,229, 381, 391 and 424.
48. R. Telle, Boride eine neue Hartstoffgeneration? Chemie in unserer Zeit 1988, 22, 95.
-
49. P. Hedenqvist, Evaluation of vapour-deposited coatings for improved wear resistance. Acta
Univ. Uppsala, 1991, p. 13.
50. T. C. Chou and T. G. Nieh, Anisotropic grain growth of d-NiMo. Thin SolidFilms 1991,219,61.
51. G. Berg; E. Broszeit, C. Friedrich, W. Herr and K. H. Kloos, Grundlageneigenschaften und
VerschleiBverhalten von HfB,- und Hf(B,N)-Schichtsystemen. Mat.-wiss. u. Werkstoftech.
1994.25, 175-179.
52. W. B. Pearson, Lattice Spacings and Structures of Metals and Alloys. Pergamon, London,
1958. DD. 219ff.
53. H. Simon and M. Thoma, Angewandte Oberjlachentechnik f u r metallische Werkstoffe.Hanser,
Munchen, 1985, pp. 103, 108 and 131.
54. G. Berg, C. Friedrich, E. Broszeit and K.-H. Kloos, Comparison of fundamental properties of
r.f. sputtered Ti,N-and Hf,N coatings on steel substrates. Surf. Coat. Technol. 1995, 74-75,
135-1 42.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Rejerences 993
55. H. Bolmgren, Synthesis and structural studies of some refractory borides. Acta Univ. Uppsala
1992, p. 19.
56. N. P. Cheremisinoff, Handbook qf Ceramics and Composites. Dekker, New York, 1990, p. 24.
51. W. Schedler, Hartmetali fur den Praktiker. VDI-Verlag, Dusseldorf, 1988, pp. 3, 8-9, 12 and
207.
58. K.-H. Habig, VerschlegundHurte von Werkstofen. Hanser, Munchen, 1980, pp. 125, 164,211
and 263-268.
59. H.-D. Steffens and W. Brandl. Moderne Beschichtungsverfahren. Lehrstuhl fur Werkstofftech-
nologie der Universitat Dortmund, 1992, pp. 242 and 261.
60. J. H. Mittendorf, Einsatz von Werkzeugheschichtungen,furnusgenlahlte Umjormverfahren. Dis-
sertation RWTH Aachen D82, 1992, p. 123.
61. H. Benninghoff, Industriereife CVD-Verfahren. Metalloberfliiclie 1976, 30, 476.
62. E. Erben et a/. CVD ein modernes und vielseitiges Beschichtungsverfahren. Metall 1981, 35,
-
1256.
63. M. Ruhle, Zum technischen Stand der Dispersionshartung Teil 2. Metall 1985, 39, 520, 522
-
and 528.
64. K.-H. Habig, VerschleiB-Schutzschichten - Entwicklungstendenzen zur Optimierung von
Eigenschaften und Verfahren. Metall 1985, 39, 915.
65. H. Holleck, Neue Entwicklungen bei PVD-Hartstoffbeschichtungen. Metal1 1989, 43, 622.
66. S . Hofmann, Charakterisierung keramischer Werkstoffe mit der Auger-Elektronenspek-
troskopie: Moglichkeiten und Grenzen. Material~~,i~~senschaft und Werkstofftechnik 1990, 21,
99.
67. H. Kolaska and K . Dreyer, Hartmetalle, Cermets und Keramiken als verschleiBbestandige
Werkstoffe. Metall 1991, 45, 227 and 232.
68. W. Wahl, VerschleiBfeste metallische GuBwerkstoffe. VerschleiBfeste Werkstoffe. Vortruge der
VDI-Taguiig Stuttgart 1973, VDI-Berichte Nr. 194 (1973), p. 82.
69. G . Schumacher, Hartmetall fur den VerschleiBschutz. Vortruge der VDI-Tagung Stuttgnrt
1973, VDI-Berichte Nr. 194 (1973) p. 113.
70. E. Gugel, Nichtoxidkeramische Werkstoffe fur die VerschleiBtechnik. Vortruge der VDI-
Tagung Stuttgart 1973, VDI-Berichte Nr. 194 (1973), p. 139.
71. B. Finnern, Konstitution, Verfahrenstechnik und Anwendung verschleiBfester Diffusions-
schichten. Vortriige der VDI-Tagung Stuttgart 1973, VDI-Berichte Nr. 194 (1973), p. 219.
72. K.-H. Kloos, WmkrtqfJkzmdeIV, Vorlesungsumdruck TH Darmstadt, 1994, pp. 2.4.2.5f.
13. I. Barin and 0. Knacke, Thermocliemical Properties of Inorganic Substances. Springer, Berlin,
1971 and 1973.
74. H. Jehn, G. Reiners et a/. Charakterisierung dunner Schichten. DIN-Fachbericht 39. Beuth,
Berlin, 1993, pp. 134, 155 and 171.
75. K. Kashigawi et a/. Chromium nitride films by radio-frequency reactice ion-plating. Vuc. Sci.
Technol. 1986, A4, 210 and 214.
76. J.-E. Sundgren et ul. A review of the present state of art in hard coatings grown from the Vapor
phase. Vac. Sci. Technol. 1986, A4, 2260ff.
77. H. Holleck, Material selection for hard coatings. Vac. Sci. Techno/. 1986, A4, 2663.
78. W. Gissler et a/. Mixed phase nanocrystalline boron nitride films: preparation and character-
ization. Thin Solid Films 199I , 199, 1 15.
79. R. A. Andrievski et ul. Structure and microhardness of TIN compositional and alloyed films.
Thin Solid Films 1991, 205, 173-174.
80. N. Zhang and V. Wang, Dislocations and hardness of hard coatings. Thin Solid Films 1992,
214, 5.
81. C. Friedrich, G. Berg, E. Broszeit and K.-H. Kloos, X-ray diffractometry analysis of r.f.
sputtered hard coatings based on nitrides of Ti, Cr, Hf. Surf. Coat. Technol. 1995, 74-75,
279-285.
82. E. Vancoille et a/. Mechanical properties of heat treated and worn PVD TiN, (Ti, Al)N, (Ti,
Nb)N and Ti(C, N) coatings as measured by nanoindentation. Thin Solid Films 1993,224, 170
and 175.
83. S. D. Marcus and R. F. Foster, Characterization of low pressure chemically vapor-deposited
tungsten nitride films. Thin Solid Films 1993, 236, 331 and 332.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
994 9 Data Collection of Properties of Hard Materials
84. Sun, Xin et a/. Properties of reactively sputter-deposited Ta-N thin films. Thin Solid Films
1993,236, 350.
85. W. Heinke, A. Leyland, A. Matthews, G. Berg, C. Friedrich and E. Broszeit, Evaluation of
PVD nitride coatings, using impact, scratch and Rockwell C adhesion tests. Thin Solid
Films 1995, 270,431438.
86. A. Schroer, Synthese und chemische, mikrostrukturelle und mechanische Charakterisierung von
Hartstojschichten ayf Stiihlen. Dissertation Heidelberg, 1992.
87. J. Berger Beschichten mit Hartstoffen. VDI-Verlag, Diisseldorf, 1992.
88. E. Dorre, Nichtmetallische Hartstoffe, in Abrasion und Erosion, H. Uetz (Ed.), Hanser,
Miinchen, 1986,433437.
89. C. Biselli and L. Chollet, Young’s modulus of TIN and TIC coatings, in Mechanics of Coatings.
Tribology Series 17 D. Dowson, M. Godet and C. M. Tayloer (Eds), Proccedings of the 16th
Leeds-Lyon Symposium on Tribology, Lyon, Amsterdam, Elsevier, 1990.
90. P. Mayr and H. Vetters, Gefiige, Oberflachenzustand und technologische Eigenschaften von
Hartstoffschichten. H T M 1993, 48, 281-287.
91. R. Wild, Hartstoffschichten leisten Widerstand. Metalloberflache 1990, 44, 355-361.
92. R. SaB, and P. Thienel, Oberflachenbehandlungen von SpritzgieRwerkzeugen. Metallober-
,fliiche 1991, 45, 313-322.
93. R. Rochotzki and J. Vetter, Festkorperreibung von PVD-Hartstoffschichten: TIN,, ZrN,,
TIC,, TiC,/i-C. Neue Hiitte 1989, 34, 272-271.
94. R. S. Bonetti and M. Tobler, Amorphe diamantartige Kohlenstoffschichten. Metallobery%iche
1990,44, 209-211.
95. K. Keller and F. Koch, CVD-Beschichtung von FlieRwerkzeugen. VDZ-Zeitschrift.1989, 131,
42-50.
96. M. Fieber and W. Konig, Kontaktbedingungen beim Einsatz beschichteter Werkzeuge. Tribo-
logie + Schmierungstechnik 1995,42, 135-140.
97. G. Kienel (Ed.): Vakuumbeschichtung 4 und 5, Anwendungen Teil 1 und 2. VDI-Verlag,
Diisseldorf, 1993.
98. B. W. Matthes, Abscheidung, anwendungsbezogene Prufung und Optimierung von PVD-
Hartstofieschichtungen des Systems Ti-B-N/Al2O3aus der Hochfrequenz-Kathodenzerstau-
bung. Dissertation, TH Darmstadt, 1993.
99. P. J. Martin et al. Properties of thin films of tantalum oxide deposited by ion-assisted deposi-
tion. Thin Solid Films 1994, 238, 183-184.
100. R. Rosenkranz, Strukturen. physikalische und mechanische Eigenschaften der hochsch-
melzenden Ti5Si3-,TiSi2- und ZrSiz-Phasen. Fortschritt-Ber. VDI Reihe 5 Nr. 291, VDI-
Verlag, Diisseldorf, 1993.
101. E. D’Anna et a/. Synthesis of thin films of semiconductor and refractory metal nitrides by laser
irradiation of solid samples in ambient gas. Thin Solid Films 1992, 218, 224-225.
102. J. L. Davis et al. Thin metal films on polyether imide. Thin Solid Films 1992, 220, 221.
103. G. Berg, C. Friedrich, E. Broszeit and C. Berger, Development of chromium nitride coatings
substituting titanum nitride. Surf. Coat. Technol. 1996, 86-87, 184191.
104. C. Friedrich, G. Berg, E. Broszeit and C. Berger, Measurement of the hardness of hard
coatings using a force indentation function. Thin Solid Films 1996, 290-291, 216-220.
105. A. G. Balzers, Coating materials and sources selector guide. Company information. Balzers
AG, FL-9496 Balzers, Fiirstentum Liechtenstein.
106. A. G. Leybold, Berechnungsgrundlagen fur Diinnschicht-Technik. Company information 01 -
017.9.70.61.02.21/KT 5.
107. Cerac Inc.: Advanced special inorganic materials. Company information 1995. Cerac Inc., PO
Box 1178, Milwaukee, WI 53201-1178, USA.
108. Goodfellow: Firmenkatalog zu Metallen, Keramiken und anderen Materialien. Edition 19951
96. Deutsche Niederlassung: Goodfellow GmbH, Postfach 1343, D 61213 Bad Nauheim.
109. K. Schafer and C. Synowitz, Taschenbuchfur Chemiker und Physiker. Bd. 3,3. Aufl. Springer,
Berlin, 1970.
110. W. Herr, B. Matthes, E. Broszeit, M. Meyer and R. Suchentrunk, Influence of substrate
material and deposition parameters on the structure, residual stress and adhesion of sputtered
Cr,Ny hard coatings. SUTJCoat. Technol.,1993, 60, 428433.
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
References 995
Downloaded from https://onlinelibrary.wiley.com/doi/ by Technische Universität Chemnitz, Wiley Online Library on [14/06/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Edited by Ralf Riedel
Copyright Q WILEY-VCH Verlag GmbH, D-69469 Weinheim
(Federal Republic of Germany), 2000
Index