You are on page 1of 211



Edited by and

J. H. Westbrook Brookline Technologies, Ballston Spa R. L Fleischer Union College, Schenectady, New

Chichester • New York • Weinheim • Brisbane • Singapore • Toronto

Copyright © 2000 by John Wiley & Sons Ltd, Baffins Lane, Chichester, West Sussex PO 19 IUD, England National 01243 779777 International ( + 44) 1243 779777 e-mail (for orders and customer service enquiries): Visit our Home Page on or All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London WlP 9HE, without the permission in writing of the Publisher Other Wiley Editorial Offices John Wiley & Sons, Inc., 605 Third Avenue, New York, NY 10158-0012, USA WILEY-VCH Verlag GmbH, Pappelallee 3, D-69469 Weinheim, Germany Jacaranda Wiley Ltd, 33 Park Road, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons (Canada) Ltd, 22 Worcester Road, Rexdale, Ontario M9W ILl, Canada Library of Congress Cataloging-in-Publication Data Basic mechanical properties and lattic defects of intermetallic compounds / edited by J. H. Westbrook and R. L. Fleischer. p. cm. 'This volume is one of four now being published, each of which consists of reprints of chapters from the 1995 comprehensive two-volume set—Intermetallic compounds, principles and practice . . . selected sets of chapters are collected, each set being on a single theme . . . reprint volume 2'—Pref. Includes bibliographical references and index. ISBN 0-471-61175-1 (pbk. : alk. paper}—ISBN 0-471-60814-9 (set) 1. Intermetallic compounds—Mechanical properties. 2. Crystals—Defects. 3. Crystal lattices. 4. Alloys. 5. Physical metallurgy. I. Westbrook, J. H. (Jack Hall), 1924II. Fleischer, R. L. (Robert Louis), 1930- III. Title: Intermetallic compounds. TA483.B37 2000 620.1'692—dc21 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN ISBN 0 471 61175 1 0 471 60814 9 (set)


Typeset by Dobbie Typesetting Ltd, Tavistock, Devon Printed and bound in Great Britain by Antony Rowe, Chippenham, Wiltshire This book is printed on acid-free paper responsibly manufactured from sustainable forestry, in which at least two trees are planted for each one used for paper production.

To the memory of

John Herbert Hollomon 1919-1985
Wise, vigorous, effective advocate of the relevance and value of scientific research in industry. His strong belief in the synergetic interaction of Principles and Practice in the field of metallurgy impelled him to assemble an innovative, diverse staff at General Electric, and to inspire independent exploration that benefited both science and engineering.

Clyde L. Briant Division of Engineering Brown University Box D, 182 Hope Street Providence, RI 02912 9104, USA

Gerhard Sauthoff Max-Planck-Institut fur Eisenforschung GmbH, D-40074 Dusseldorf, Germany

Joel Douin Laboratoire d'Etude des Microstructures, Unite Mixte CNRS/ONERA, UMR 104, 29 Avenue de Ia division Leclerc, 92322 Chatillon Cedex, France

Yong-Qian Sun Department of Materials Science and Engineering, University of Illinois, Urbann, IL 61801, USA

Lawrence M. Howe 2 Frontenac Crescent, P.O. Box 697, Deep River, Ontario, Canada KOJ 1P0

Takayuki Takasugi Institute for Materials Research, Tohoku University, Katahifa 2-1-1, Aoba-ku, Senda 980, Japan

Han R. P. Inoue (now Kanryu Inoue) Department of Materials Science and Engineering, University of Washington, Seattle, WA 98195, USA

Patric Veyssiere Laboratoire d'Etude des Microstructures, Unite Mixte CNRS/ONERA, UMR 104, 29 Avenue de Ia division Leclerc, 92322 Chatillon Cedex, France

Morihiko Nakamura National Research Institute for Metals, 1-2-1 Sengen, Tsukuba-shi, lbaraki 305-0047, Japan

C. Marvin Wayman Department of Materials Science and Engineering, University of Illinois, Urbana, IL 61801, USA

Charles de Novion Laboratoire des Solides Irradies, ICEA-CEREM, URA CNRS No. 1380), Ecole Polytechnique, 01128 Palaiseau Cedex, France

structure and properties? The Practice volume considers commercial production and engineering applications of intermetallic compounds. From the 1967 book we remind the reader that 'The first published paper dealing with intermetallic compounds appeared in 1839. for others. First. The number of intermetallic entries in the Permuterm Subject Index shows a doubling period of less than nine years. . Each author was asked to set forth the principles of his or her subject in terms that are meaningful to scientists and engineers who are not specialists in the author's field. The burgeoning literature on intermetallics and the lack of a comprehensive single source of up-to-date descriptions of where we are. but the authors have aimed to reveal what the obstacles are so that the reader may make his or her own assessment (and possibly provide a solution!). and practices. Even the metal-metal aspect of the definition is often relaxed by including some metal-metalloid compounds. tellurides. is directed at the science of intermetallics—-how do we understand their formation. and semiconductors. This work was planned to provide state-of-the-art assessments of theory. Here we include as intermetallic compounds all metal-metal compounds. such a compound is a structure in which the two or more metal constituents are in relatively fixed abundance ratios and are usually ordered on two or more sublattices. world-wide use is in the 1000s of tons. Apart from scholarly interest. for some. their value as vital components of electronic circuits is in the billions of dollars. This work is divided into two volumes in order that each be of manageable size. binary and multicomponent. the first review paper by Neville in 1900. The scope of the present work is as broad as that of its 1967 predecessor. as a result of low ordering energy or the intervention of some external agency. for example III-V semiconducting compounds. limited sub-topical subject areas have become common. what we need to know. Some of the hoped-for uses are fated never to be realized. .Preface to the 1995 Edition Intermetallic compounds were last comprehensively reviewed in 1967 in a volume that was edited by one of us (JHW). devoted exclusively to this subject. properties. one by Desch in England and one by Giua and Giua in Italy*. However. to actual commercial application—depending on the specific subject. and then to progress to include knowledge that workers in their own areas would wish to have. radiation. experiment. each with its own distinct population of atoms. and practice that will form a solid base for workers who wish to know more than their own particular area. the increased volume of activity in intermetallics and the increased significance of their applications have necessitated an expansion from the 27 chapters of the earlier work to the 75 chapters of the present treatise. having reached roughly 1800 entries per year in 1993. etc. . At that time the field was described as of special interest because it was undergoing 'exponential proliferation'. for example extreme cooling rates. and they provide useful examples of principles. and more than sixty years elapsed before . both ordered and disordered. new results were then appearing so rapidly that fifteen years later two books were printed. such as silicides. However. The first. That trend continues to the present. More recently. Concluding sections of most chapters give the authors' critical assessment of the state of their subject and of where they believe further effort is merited. Often substantial or complete disorder may obtain. conference volumes that deal exclusively with intermetallics but typically only within specific. such as Ni3Al. intermetallics have now become of substantial commercial significance. necessitating a partial disorder. The reader who browses carefully will recognize that the immediacy of the practice described ranges from hoped-for use. on the theme Principles. We believe this inclusion is appropriate since the phenomenology of many such compounds is nearly identical to metal-metal ones. and what we can do with intermetallics created the incentive for the present pair of volumes. to beginnings of use. Deviations from precise stoichiometry are frequently permitted on one or both sides of the nominal ideal atomic ratios. what are intermetallic compounds? Generally. although the quantities employed are not in tonnage numbers.

We planned this work as an aid to both scientists and engineers. Thus to be definitive. Thus.We conferred carefully with many people in order to identify authorities for each subject. New York Upon these considerations. both the structure type and the Pearson symbol need to be given. or wish. to expand their knowledge of related science or practical technology. Schenectady. C s have 2-to-l atomic abundance ratios. having recruited contributors for the project. WESTBROOK. It is hoped that passage of the next 25 years will reveal some influence of this treatise on the further development of this field. finally Mary Carey at Rensselaer Polytechnic Institute. the purpose of which is to instruct rather than to amuse. H.g. first by Phillis Liu. Vanessa Lutman and Cliff Morgan. Structure type compounds are the specific ones used to designate a particular structure. then Constance Remscheid at General Electric. Nomenclature for crystal structure types is often complex. We expect that many chapters are appropriate source matter for special topic or seminar courses at the advanced undergraduate and various graduate school levels. and we thank them for their hard work and high achievement. or the Pearson and the Strukturbericht symbol. Generally we have asked authors to include both of two commonly used types of symbols as they introduce structures. —Isaac Ware. J. Villars and Calvert list two cP4s. Equally it is a resource to workers who are already active in the field and need. we have been induced to undertake the present extensive work. Thus B2 compounds are also referred to as CsCl compounds. and Pearson symbol. L. in order that the reader may be able to locate any subject that is addressed^ Although the purpose of such combined indices is not to induce the owner of a single volume to purchase the other. The Pearson symbols tell the crystal symmetry and the number of atoms per unit cell. We sought experts wherever they were to be found. It can serve as a base for those who wish to know about intermetallics as an area in which to begin research. and our international set of nearly 100 authors turned out to be almost equally divided between the United States and 14 other countries. B2. but the principal regard will be shown to what is necessary and useful. The two-part table following this preface lists many of the common types—by Strukturbericht symbol. We have benefited from outstanding secretarial help during the three years of this project. Although there is only one cP2 structure. B's are equiatomic ordered structures. but it is not necessarily unique. three cF12s and twenty-two hP9s. and some of the authors have introduced their own. Strukturbericht symbols are only partly significant and systematic: A's are not compound structures but consist of a single lattice of atoms (except for A15!). Laves. in which nothing will be omitted that is elegant or great. Heusler. New York R. The Pearson designation is informative. CsCl has a primitive (P) cubic (c) structure with 2 atoms per cell and hence the Pearson symbol (cP2). etc. fluorite. Irene Cooper. it possibly may help to reduce the barrier to such action. FLEISCHER. Philip Hastings. Manuscripts have in fact come from all inhabited continents. We appreciate the hospitality of the General Electric Research and Development Center during the inception and middle period of preparing these volumes. Many structures are better known to metallurgists and mineralogists by names other than the formula of the structure type chosen by crystallographers. Assembling the final product has been eased for us by the continuing efforts and cheerful good counsel at John Wiley of Jonathan Agbenyega. e. DO's 3-to-l. Such names have been added in selected cases. As an assist to readers we have provided in the following pages a consolidated acronym list and some crystallographic tables. we then strove to assist them in achieving clarity and thoroughness from outline to draft to final manuscript. The index in each volume includes the subjects in both volumes of this work. The contributors cooperated superbly. prototype name (termed a structure type). Balhton Spa. 1756 .

contains three from Property Fundamentals. Electrical. reprint volume 2. a margin mark (*) alerts the reader to refer to the added pages at the end of the chapter. and finally reprint volume 4. In this format readers who are interested in a particular aspect of intermetallic compounds can have a less weighty volume specific to their subject. the tables of contents of which appear at the end of this volume. Crystal Structures of Intermetallic Compounds. contains the ten chapters from the original work under the heading Crystal Structures. and Optical Properties and Applications of Intermetallic Compounds. Although each chapter is reprinted nearly intact (only typographic and factual errors corrected). In the present volumes selected sets of chapters are collected. The subjects in most cases are taken from one or more chapter groupings of the original Volume 1 or 2: Hence reprint volume 1. Some have chosen to do so. either in the preferred case because they felt none was needed or because the four-month window of opportunity they were given to satisfy our and the publisher's desire for promptness did not fit their work schedule. and two from Kinetics and Phase Transformations. Corrections to the original chapters that were so lengthy that they would upset the original pagination are to be found in the addenda at the end of each relevant chapter. some have not.Preface to the Reprint Volumes from Intermetallic Compounds: Principles and Practice This volume is one of four now being published. the author or authors were given the option of adding a brief addendum in order to add whatever new perspective has arisen over the intervening few years. and that includes a modest updating of the subject matter. four chapters from Defect Structures. contains two chapters from the section on Property Fundamentals. Structural Applications of Intermetallic Compounds contains the thirteen chapters that were under that same topic. original 1900page set. JHW RLF . seven from Electromagnetic Applications and one from Miscellaneous. Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. each set being on a single theme. Magnetic. each of which consists of reprints of chapters from the 1995 comprehensive two-volume set Intermetallic Compounds: Principles and Practice. Where an addendum is particularly relevant to a portion of the original chapter being reproduced. reprint volume 3. a volume that can be produced more economically than the full. Cross-references to other chapters relate to the original 1995 two-volume work.

atomic number atom probe atomic property antiphase boundary antiphase domain avalanche photodetector atomic property expression augmented plane wave antireflection angle-resolved inverse photoemission spectroscopy angle-resolved photoemission spectroscopy atomic-sphere approximation augmented spherical wave bond charge body-centered cubic Bardeen-Cooper-Schrieffer (theory of superconductivity) body-centered tetragonal BH BIS BM BSCCO BSE BT BW BZ CAM CANDU CAP CAT CBLM CC CCD CCGSE CCIC CCMAI c. CCT CD CD CEBAF CEF CERN CFT CMC CN buried heterostructure bremsstrahlung isochromat spektroskopie Bowles-Mackenzie (theory of martensitic transformation) bismuth-strontium-calcium-copper oxide back-scattered electrons Bhatia-Thornton (partial structure factor for liquid alloys) Bragg-Williams (theory of ordering) Brillouin zone c-axis modulated Canadian deuterium-uranium (power reactor) consolidated under atmospheric pressure computer-assisted tomography cluster Bethe lattice method cluster center charge-coupled device concentric-circle grating surfaceemitting (laser) cabled conductor in conduit crystal chemical model of atomic interactions cubic close-packed continuous cooling transformation compact disc climb dislocation continuous electron-beam accelerator facility crystalline electric field Centre Europeenne Recherche Nucleaire concentration-functional theory ceramic-matrix composite coordination number .c. BCS b.c.c. two-dimensional three-dimensional six-dimensional angular correlation of annihilation radiation angular correlation of positron annihilation radiation atomic environment Auger electron spectroscopy atomic environment type argon induction melting atom location by channeling enhanced microanalysis atomic layer epitaxy air mass Advanced Materials Technology.t.Acronyms 2D 3D 6D ACAR ACPAR AE AES AET AIM ALCHEMI ALE AM AMT AN AP AP APB APD APD APE APW AR ARIPES ARPES ASA ASW BC b.p.c. Inc.

CO CP CPA CRSS CS CSF CSL CSRO CT CTE CVD CVM CW CW CW CWM cubo-octahedron coordination polyhedron coherent-potential approximation critical resolved shear stress chemisorption complex stacking fault coincidence-site lattice chemical short-range order chisel toughness coefficient of thermal expansion chemical vapor deposition cluster variation method cold worked concentration wave continuous wave Connolly-Williams method (theory of phase transformations) donor-acceptor diffusion bonding ductile-brittle transition temperature direct chill (casting) direct current direct configurational averaging density functional distributed feedback density-functional theory double heterojunction double hexagonal close-packed de Haas-van Alphen (effect) directional levitation zone melting density of states displacement per atom demonstration poloidal coil dense random packing directional solidification displacement shift complete electron/atom (ratio) embedded-atom method electron beam physical vapor deposition effective cluster interaction embedded-cluster method electro-optic directional coupler electrodischarge machining energy-dispersive X-ray (spectroscopy) electron energy-loss spectroscopy electromotive force effective pair interaction extrinsic stacking fault ESR ETP EXAFS f.h. f.p. FENIX FET FIM FLAPW FLASTO FLMTO FOM FP FT FZ GB GB GFT GGA GITT GPM GRPA GS GT HB HBT HCF h.t.c.c. HEMT HIP HPT HR HREM HRTEM HSCT HTS HVEM HVTEM electroslag refined electrolytic tough pitch (copper) extended X-ray absorption fine structure face-centered cubic face-centered tetragonal Fusion Engineering International Experimental Magnet Facility field effect transistor field ion microscopy full-potential linearized augmented plane wave full-potential linearized augmented Slater-type orbital full-potential linearized muffin-tin orbital figure of merit Fabry-Perot (laser) phase transformation floating zone gain x bandwidth (product) grain boundary glass-forming tendency generalized gradient approximation galvanostatic intermittent titration technique generalized perturbation method generalized random-phase approximation ground state Goody-Thomas (electronegativity) horizontal Bridgman heterojunction bipolar transistor high-cycle fatigue hexagonal close-packed high-electron-mobility transistor hot isostatic pressing heterojunction phototransmitter high resolution high-resolution electron microscopy high-resolution transmission electron microscopy high-speed civil transport high-temperature superconductor high-voltage electron microscopy high-voltage transmission electron microscopy irregular atomic environment irregular atomic environment type D-A DB DBTT DC DC DCA DF DFB DFT DH d.p. dHvA DLZR DOS DPA DPC DRP DS DSC e/a EAM EBPVD ECI ECM EDC EDM EDX EELS EMF EPI ESF IAE IAET .c.c.c.

rr + . o .• -n . Mirror Fusion Test Facility metal-insulator-semiconductor field effect transistor McDonald jelly roll (superconducting cable construction) multi-layer reflector metal-matrix composite Mendeleev number magneto-optical metal-organic chemical vapor deposition metal-oxide-semiconductor metal-oxide-semiconductor field effect transistor metal-organic vapor phase epitaxy multiple quantum well magnetic resonance imaging maximum resolved shear stress orthodontic NiTi alloy muffin tin martensitic transformation diagram million volt-amperes National AeroSpace Plane Next European Torus (fusion device) normal hydrogen electrode National Maglev Initiative nuclear magnetic resonance nearest neighbor nearest-neighbor histogram KKR KSV KTP KW LA LAPW LASTO LCAO LCF LCT LCW LD LDA LEC LED LEED LEISS LHC LKKR LME LMTO LNT . junction field-effect transistor Korringa-Kohn-Rostoker (bondcalculation method) Khantha-Cserti-Vitek (deformation model) potassium titanyl phosphate Kear-Wilsdorf (dislocation locking mechanism) longitudinal acoustic (wave) linearized augmented plane wave linearized augmented Slater-type orbital linear combination of atomic orbitals low-cycle fatigue large coil task Lock-Crisp-West (radiation analysis) laser diode local-density approximation liquid-encapsulated Czochralski light-emitting diode Low-energy electron diffraction low-energy ion scattering spectroscopy Large Hadron Collider Layered KKR (structure calculation) liquid metal embrittlement linearized muffin-tin orbital liquid nitrogen temperature LO LPCVD LPE LPPS LPS LRO LSDA LSI ^SR MA MAPW MB MBE MBT MC MCS MD MEE MESFET WT-™-< MFTF MISFET MJR MLR MMC MN MO MOCVD MOS MOSFET MOVPE MQW MRI MRSS MRT MT MTD MVA NASP NET NHE NMI NMR NN NNH longitudinal optical (wave) low-pressure chemical vapor deposition liquid-phase epitaxy low-pressure plasma spraying long-period superstructure long-range order local spin-density approximation large-scale integration muon spin relaxation mechanical alloying modified augmented plane wave Martinov-Basanov (electronegativity) molecular beam epitaxy metal-base transistor Monte Carlo Monte Carlo simulation molecular dynamics migration-enhanced epitaxy metal Schottky field-effect transistor X . . . . .IC IC IDOS IEM IGC IHPTET ILS IMC IMC IMC IPM IPS IQC IR ISF IT ITER IV TT7™ JFET integrated circuit investment cast integrated density of states interstitial-electron model Intermagnetics General integrated high-performance turbine engine technology invariant line strain intermetallic compound intermetallic matrix composite inverse Monte Carlo (method) independent-particle method (approximation) invariant plan strain icosahedral quasicrystal infrared intrinsic stacking fault (positive) inner tetrahedron International Thermonuclear Experimental Reactor intermediate valence . ^ %-. ..T .


next nearest neighbor negative OR (logic operator) notch/strength ratio oxidation-affected zone oxygen dissolution reaction oxide dispersion-strengthened optoelectronic integrated circuit octahedron Oak Ridge National Laboratory (negative) outer tetrahedron orthorhomic Ti-matrix composites positron annihilation spectrosxopy periodic bond chain permeable-base transistor phase-change material pressure-composition-temperature phase diagram pair distribution function phonon density of states planar flow casting Pearson's Handbook phase computation primary knock-on atom photoluminescence powder metallurgy phenomenological martensite transformation concept periodic number type of photothyristor partial pair distribution function path-probability method Paidar-Pope-Vitek (Ll 2 hardening model) Pearson symbol phase transformation phenomenological theory of martensite crystallography physical vapor deposition lead zirconate titanate (ceramic) quasicrystal(line) quantum confined Stark effect quantum formation diagram quantum number quantum structural diagram quantum well Rutherford back scattering ribbon comminution replacement-collision sequence radial distribution function


RKKY rms




rate-dermining step rare earth (metal) radiofrequency reversible hydrogen electrode ngu *" lon m o ^ e l t v ... v Ruderman-Kattel-Kasuya-Yoshida (electron interactions) root mean square residual resistivity rabo rapidly solidified ra ld P sohdincaUon processing resolved shear stress room temperature resonance ultrasound spectroscopy selected-area diffraction selected-area electron diffraction stress-assisted grain-boundary oxidation separate absorption and multiplication avalanche photodetector simple cubic semiconductor standard colomel electrode separate confinement heterostructures specific damping capacity spin-density wave scanning electron microscopy superlattice extrinsic stacking fault stacking fault spin glass self-propagating high-temperature synthesis semi-insulating very large-scale integration self-interstitial atom self-interaction correlation stress-induced martensite secondary-ion mass spectrometry superconductor-insulatorsuperconductor superlattice intrinsic stacking fault static inductance transistor semimetal second-moment approximation shape-memory alloy shape-memory effect superplastic forming superconducting quantum interference device short-range order solid-state amorphizing reaction structural stability diagram



superlattice stacking fault Atlas of Crystal Structure Types scanning transmission electron microscopy scanning tunneling microscopy Sodani-Vitole change of Paidar et al. model transverse acoustic (wave) tight binding topologically close-packed thoria dispersion temperature dependence of flow stress thermoelectric transverse electric (field) thermoelectric cooler thermoelectric generator transmission electron microscopy triethylphosphene Teatum-Gschneidner-Waber (atomic radius) thermally induced porosity Takeuchi-Kuramoto (dislocation locking mechanism) transition metal transverse magnetic (field) titanium-molybdenum-aluminum (alloy) transverse optical (wave) two-photon absorption topological short-range ordering truncated tetrahedron tubular tin source time-temperature-transformation ultra-high frequency ultra-high vacuum


ultra large-scale integration ultrasonic wave ultimate tensile strength ultraviolet vacuum arc refined vertical-cavity surface-emitting laser valence-electron concentration vertical gradient freezing very high frequency vacuum induction melting vapor-liquid-solid very large-scale integration vapor phase epitaxy vacuum plasma spraying vacuum ultraviolet weak beam waveguide photodetector Wechsler-Lieberman-Read (theory of martensitic transformation) Wigner-Seitz (cell) Winterbon-Sigmund-Sanders (model of irradiation damage) weight parts per million exchange-correlation exothermic dispersion (synthesis process) X-ray inspection module X-ray photoelectron spectroscopy X-ray diffraction extreme ultraviolet yttrium aluminum garnet zero insertion force

Crystal Structure Nomenclature*
Arranged Alphabetically by Pearson-Symbol Designation Pearson symbol cFA cFS cFYl cF16 cFIA cF32 cFS2 cFS6 cF6S cFSO cFU2 cFU6 cI2 cl 16 c/28 c/32 cIAO cI52 cISA cISS c/76 cISO cI96 c/162 cP\ cP2 cPA cP5 Strukturbericht designation Al AA BX 53 Cl Space group FmIm FdIm FmIm F43m FmIm F43m Fm \m Fm 3 m FdIm Film FdIm Fd3m FmIc Fm 3m FdIm Fd\m Fm 3m Fd3m FdIm Fm3c Fm 3m Fm3m Im3m «,3 IA3d ImJ Im 3m IA3d IA3m Im 3m Im 3m /43m IA3d Ia3 IaI Im3 Pm 3m Pm13m Pm\m Pm 3m Pm 3m Pm 3 m PA3m Pearson symbol cP6 cPl cPS cP12 cPTO cP36 cP39 cPS2 hPl hP2 Strukturbericht designation C3 D2} AlS BUO H24 C2 FO1 A13 D2t DSC DS3
A l A3 Bh C32 C6 L'3

Prototype Cu C (diamond) NaCl (rock salt) ZnS (sphalerite) CaF2 (fluorite) MgAgAs AlCu2Mn (Heusler) BiF3 (AlFe3) NaTl AuBe5 SiO2 (0 cristobalite) Cu2Mg (Laves) CuPt3 UB12 Al2MgO4 (spinel) Co3S4 Co9S1 Sb2O3 (senarmontite) Fe3W3C (ij carbide) NaZn13 Cr23C6 Mn23Th6, Cu16Mg6Si7 (G-phase) W CoU Th3P4 CoAs3 (skutterudite) Ge7Ir3 Pu2C3 Cu5Zn1 (7 brass) Fe3Zn10 (7 brass) Sb2Tl7 otMn (x-phase) Cu15Si4 Mn2O3 AlLi3N2 Mg32(Al1Zn)4, aPo CsCl AuCu3 ReO3 AlFe3C (perovskite) CaTiO3 (perovskite) Fe4N

Prototype Ag2O CaB6 Cr3Si(^W) FeSi Cu3VS4 (sulvanite) FeS2 (pyrite) NiSbS (uUmanite) /3Mn BaHg11 Mg2Zn11 Cu9Al4 (7 brass) HgSn6-10 Mg WC AlB2 CdI2 Fe2N LiZn2 7Se crLa BN C (graphite) NiAs ZnS (wurtzite) La2O3 Ni2Al3 CaCu5 CoSn Cu2Te HgS MoS2 Ni2In Na3As Ni3Sn TiAs CrSi2 Fe2P fAgZn SiO2 (high quartz) Pt2Sn3 CuS MgZn2 (Laves) SiO2 (0 tridymite)

Space group Pn3m Pm3m Pm3n P2,3 P43m Pa3

L2, DO3 £32 ClSb C9 CIS D2f Hl1 Dl2 DS9 D54 E% D2, DS4 DS. A2 B. Dl3 DO2 DSj DS2 DSx L22 A 12 DS6 DS3 E9d DS9 A> B2 Ll2 DO9 LV2 Elx LV


AS A3' Bk A9 BS1 BA DS2 DS13 D2d B3S


7*4^2 Pm 3m Pm3 PA3m PS/mmm P63/mmc Plm2 PS/mmm P3ml P63/mmc P63/mmc P3t2l P63/mmc P63/mmc P63/mmc P63/mmc

B9 Cl BS2 DOn DO19 B1 CAO C22 Bb CS DSb BlS C14 ClO

hPS hP9


P3ml P3ml P6/mmm PS/mmm P6/mmm />3,21 P63/mmc P63mmc P&t/mmc P^/mmc P^/mmc PS2H P62m P3 P6222 P^/mmc P63/mmc P6j/mmc P^/mmc

continued •Adapted (with additions and corrections) from ASM Handbook, Vol. 3, 10th ed, ASM International, Materials Park, OH.

Arranged Alphabetically by Pearson-Symbol Designation (continued)
Pearson symbol HP IA hP 16 hP\% HP20 hP2A HP2S hR\ hR2 HR 3 hR4 hR5 hR6 hRl hR\0 hR\2 hR\3 hR\5 hR20 hR26 hR 32 mC6 mC% mC\2 mC\A mC\6 mPYl mP20 mP22 mP2A mP32 mP6A oCA oCS Strukturbericht designation DS, DS, DO11 E9t "'. DlO2 DSk DO2x C36 DSn AlO
* i

Prototype W2B5 Mn5Si3 Ni3Ti Al4C4Si AIgFeMg3Si6 Mg2Ni Fe3Th7 Th7S12 Cu3P MgNi 2 (Laves) Co2Al5 aHg 0Po cxAs aSm NaCrS2 Bi2Te3 Ni 3 S 2 CaSi2 NiS (millerite) Al4C3 Mo2B5 CrAl2O3 (corundum) BaPb 3 Fe 7 W 6 (/i-phase) B4C HoAl 3 Cr5Al8 CuPt AuTe2 (calaverite) CuO (tenoritc) ThC 2 6Ni3Sn4 FeKS2 AgAuTe 4 (sylvanite) ZrO2 As2S3 Co 2 Al 9 FeAsS AsS (realgar) 0Se aSe aU CaSi ceGa CrB I2 P (black) ZrSi2 BRe3 PdSn4 PdSn2 Al6Mn TiSi2 Mn4B CuMg2 GeS2 aS SiS2 Ta3B4 Al4U Ga2Mg5 AuCd FeS2 (marcasite) CaCl2 aNp TjNiSi

Space group P6i/mmc P62/mcm P63/mmc P6jmc PlIm P6222 P62mc P63/m P63cm P63/mmc P6i/mmc R 3m RZm

Pearson symbol oP%

Prototype 0Cu3Ti FeB GeS SnS MnP TiB Co2Si, NiSiTi (E-phase) Co2Si HgCl2 Al3Ni AsMn3 BaS3 CdSb CuS2Sb (wolfsbergite) Fc3C (cementite) Cr3C2 Sb2S3 Sb2O3 (valentinite) AuTe2 (krennerite) CuFe2S3 (cubanite) TiO2 (brookite) Cr7C3 aPa In 0Sn CaC2 MoSi2 ThH2 Al3Ti Al4Ba MoNi4 Al2Cu ThSi2 Al2CdS4 Al3Zr CuFeS2 (chalcopyrite) Cu2FeSnS4 (stannite) Ir3Si MoB SiU3 TlSe Fe8N Mn12Th MnU6 Cr5B3 Ni3P W3Si3 6CuTi 0Np AuCu CuTi3 7CuTi PbO Pb3Sr PtS Cu2Sb PbFCl TiO2 (rutile) Pb4Pt Si2U3 PdS B4Th 0U aCrFe Al7Cu2Fe Zn3P2 yB

Strukturbericht designation DO. B21 Bid B29 B31 &3 C37 C28 DO20 D0d DO11

Space group Pmmn Pnma Pnma Pmcn Pnma Pnma Pnma Pbnm Pmnb Pnma Pmmn PA2xm Pbca Pnma Pnma Pnma Pnma Pccn Pma2 Pnma Pbca Pnma lA/mmm lA/mmm lA/amd lA/mmm lA/mmm lA/mmm lA/mmm lA/mmm lA/m lA/mcm lAx/amd IA lA/mmm IAId /42m lA/mcm lAx/amd lA/mcm lA/mcm lA/mmm lA/mmm lA/mcm lA/mcm IA lA/mcm PA/mmm P42.2 PA/mmm PA/mmm PA/nmm PA/nmm PA/mmm PA2/mmc PA/nmm PA/nmm PA2/mnm PA/nbm PA/mbm PA1Zm PA/mbm PA2/mnm PA2/mnm PA/mnc PA2/nmc PA2/nnm

oPYl 0PI6

Al C19 C33 D% C12 B13 Dlx Z>8, Z>5, D8 5

R3m R 3m R 3m R32 R 3m R 3m R 3m R 3m R3c K$m JGm R3m R3m R3m R 3m C2/m C2/c C2/c C2/m C2/c P2/c P2x/c P2x/c P2x/c P2x/c P2x/c P2x/c P2x/c Cmcm Cmmc Cmca Cmcm Cmca Cmca Cmcm Cmcm Aba2 Aba2 Cmcm Fddd Fddd Fddd Fdd2 Fddd lbam lmmm lmma lbam Pmma Pnnm Pnnm Pnma Pbnm oP20 oP2A oPAO tI2 tIA //6 //8 //10 //12 //14 //16



E9e C21 DlO 1 A. A6 AS

Lix C34 526

CIl. L"lb DO21 Dl3 D\a C16 Z)O 23

% DS1 Di6 EQ1 B1

h Si

B37 D2'b D2C DS1 DO, DSm L 2. Ad

//18 //26 //28 //32 tP2 tPA


oC\2 oC\6 oC20 oC2A oC2% oF2A oFAO oFAS oF12 oF\2% ol 12 ol 14 o/20 o/28 oPA oP6 oPS

B33 AlA All CA9

Ii: Si
C5A % CAA A 16 CA2 Dlb Dlb D%, B19 C18 C35

ilo X.6. SIl BlO

tP6 /PlO /P16 /P20 /P30 /P40 /P50

B17 C38 £0, CA D\d DS. S34 D\, DSb DS, A.


( = B33) Pearson symbol til tP30 oPS tPA hPl /P 50 cPl HRl mP64 mP31 cFA cTl hPl hPA cFS tIA til HRl hP3 hPA hRl oCS cISS cPIO oCS cPS oFHS oCS oCA c/16 hP9 oCS oPS oP 16 oCS //16 hPl hPS hPA mP32 oPS cFS cPl cFS hPA hPA hP6 hP6 tPA tPA HR 6 oPS (PA hPll oPA Space group IA/mmm PA2/mnm Pnma PAl(L P6/mmm PA2/nnm Pm 3m R3m Pl1Zc PlxZc Fm3m Im3m P63Zmmc Pe^Zmmc FdIm IAxZamd IA/mmm R 3m P3. Ac A.Arranged Alphabetically by Strukturbericht Designation Strukturbericht designation A.3 PJ Cmmc Pbnm Pbca Cmcm IAxZamd Plml P63Zmmc Pe^Zmmc PlxZc Pnma Fm 3m Pm3m FA3m P63mc P6-immc P6^mmc P3.21 P6zZmmc R3m Cmca /43m P4. NiSiTi (E-phase) HgCl2 AlB2 Bi2Te.21 PAZnmm PAZnmm R3m Pnma PA2Zmmc P^Zmmc Pmma Strukturbericht designation £20 B16 BIl B19 B31 B 31 B33( = Bf) B3A B3S B31 Pearson symbol cP8 mCS oPS oPS oPS cFl6 oCS /P16 HP6 //16 hPIS oFAS till oCIA mCll hP6 hP3 cFll cFll CPU cP6 fP6 hP3 HP6 hP9 cFIA hPll Space group P2.32 Cmca Pm 3/i Fddd Cmca Cmcm /2. Af A' At Ak A1 Al Al Al AV AA AS A6 Al AS A9 AlO All All A 13 AH A 15 A 16 All A 20 B. cb ce ce § ck Cl Ch Cl C3 CA C6 Cl CS C9 ClO cna cnb CIl ClA CIS tie tie hRe tin SS) — AuBe5 Al2Cu FeS2 (marcasite) aSm TiO2 (brookite) Fe2P Co2Si.3 Cl/c Pnma Pmcn Pnma FdIm Cmcm PA2Zm PeZmmm IAZmcm Pe2Il Fddd IAxZamd Abal ClZc PeZmmm Pe^Zmmc Fm3m FA3m Pa_3 Pn 3m PA2Zmnm P3ml Pe^Zmmc Pe2Ii Fd3m Pe^Zmmc IAZmmm IAZmmm R 3m Pe^/mmc Fd3m FA3m IAZmcm Pnnm R 3m Pbca P61m Pnma Pmnb PeZmmm R 3m ClZm Pnnm Pe^Zmmc Pbnm PAZnmm Pe2Il Ibam PIiZc Fddl Pma2 Cmcm Fddd Pmmn IAZmcm continued Prototype aPa /3U aNp 0Np HgSn6-10 7B aPo /3Po aSe /8Se Cu W Mg aLa C (diamond) /3Sn In crAs 7 Se C (graphite) aHg crGa aMn (x-phase) 0Mn I2 Cr3Si 03-W) aS P (black) aU CoU ^AgZn CaSi I7NiSi CdSb CrB MoB WC TiAs BN AsS (realgar) TiB NaCl (rock salt) CsCl ZnS (sphalerite) ZnS (wurtzite) NiAs Ni2In HgS (cinnabar) PbO TCuTi NiS (millerite) GeS PtS (cooperite) CuS (rovelite) AuCd Prototype FeSi CuO (tenorite) FeB SnS MnP NaTl CrB PdS CoSn TlSe Mg2Ni CuMg2 ThSi2 PdSn2 ThC2 Cu2Te LiZn2 CaF2 (fluorite) MgAgAs FeS2 (pyrite) Ag2O TiO2 (rutile) CdI2 MoS 2 SiO2 (high quartz) SiO2 {b cristobalite) SiO2 (/3 tridymite) CaC2 MoSi2 CaSi2 c. A. BZ1 B9 BlO BU £13 516 BIl BlS £19 k mce oPe hPlA oPll tpe hP9 Ol Il mPll oFl'l oPIA oCll oFIA oPS //16 . Bb Bc Bd Be B. AuTe2 (calaverite) CaCl2 MgNi2 (Laves) Co2Si Cu2Sb CrSi2 SiS2 ZrO2 GeS2 AuTe2 (krennerite) ZrSi2 TiSi2 /3Cu3Ti SiU3 hPll cFIA cFIA cisb C16 CIS C19 C21 C22 C23 CIS C31 C33 C3A C3S C36 C31 C3S CAO CAl C43 C44 C46 C49 CSA D0a D0c oPe hR3 oPlA HP9 oPll oPll hP3 hRS B1 Bk B1 Bn Bl Bl B3 BA BB.

Dl. D\. Dl1 Dl. Dld Dl. Di1 Di. D8 6 DS. Di. AlFe3 ReO3 Fe3C (cementite) BaS3 Na 3 As Ni3Sn Al3Ni Cu3P Al3Ti Al3Zr Ni3Ti MoNi 4 AI4U PdSn4 Pb4Pt B4Th Mn4B B4C Al4Ba Mn12Th MnU 6 CaCu5 BaHg11 UB 12 Fe8N Al6Mn CaB 6 NaZn13 Si2U3 Pt2Sn3 Pu2C3 Ni 3 S 2 As2S3 OtAl2O3 (corundum) La2O3 Mn2O3 Sb2O3 (senarmontite) Sb2S3 Zn 3 P 2 Cr3C2 Sb2O3 (valentinite) Ni2Al3 6Ni3Sn4 Ta3B4 Al 4 C 3 Co3S4 Th 3 P 4 Mn23Th6. Di./c ImJ Fm 3m PmSm Pnma PA2xm P63/mmc P6i/mmc Pnma P6 3 cm IA/mmm IA/mmm P63/mmc IA/m Imma Aba2 PA/nbm PA/mbm Fddd R 3m IA/mmm IA/mmm lA/mcm P6/mmm Pm3m Fm 3m IA/mmm Cmcm Pm3 m Fm 3c PA/mbm P63/mmc IA3d R 32 Pl1Zc RIc P3ml Ia3 Fd3m Pnma PA2/nmc Pnma Pccn P3ml C2/m Immm R 3m Fd3m IA3d Fm3m PA2/mnm PmI as //16 hP 16 //1O o/20 oC20 /PlO /P20 oF40 hR 15 //1O //26 //28 hP6 cP36 cFS2 //18 oC28 cPl cFUl /PlO hP\Q cIAO hR5 mP20 hR IO hP5 clSO cFSO oP20 tPAO oP20 oP2Q hP5 mClA o/14 HRl cF56 c/28 CF116 /P 30 cP39 D\. Co 2 Al 9 Mg 32 (AUZn) 49 Ge 7 Ir 3 Ga 2 Mg 5 W2B5 Mo 2 B 5 Th 7 S 12 Cr5B3 W 5 Si 3 Fc 3 Zn 10 ) DS. Dl. Di. Cu5Zn8J 7 brass DS2 Cu 9 Al 4 ] DS1 Cr 23 C 6 DS4 Fe7W6 (. Di. Di. Di. Z)O 24 (continued) Stmkturbericht designation D%d DSe DSr D*'g DSh DSi DSk DS1 Prototype Ir3Si AsMn 3 Ni3P CoAs 3 (skutterudite) BiF3. DO1. Di. OS 10 DS 11 D5 1 3 Dl. DO1 DO. DO11 DO11 DO1. Dl1 Dl1 Di.i-phase) DSi Cu13Si4 DS6 Mn5Si3 DS9 Co 9 S 8 DS9 Cr5Al8 DS10 Co 2 Al 5 DSn Cr7C3 DlO1 Fe3Th7 DlO2 PbFCl EO1 FeAsS EO1 MgCuAl 2 AgAuTe 4 (sylvanite) E\ CuFeS 2 (chalcopyrite) Elx CaTiO3 (perovskite) E2X Al 2 CdS 4 E3 Al7Cu2Fe E% Al 8 FeMg 3 Si 6 E% Mn3Al9Si E9C AlLi 3 N 2 E% CuFe 2 S 3 (cubanite) E9e Fe3W3C (TJ carbide) E% Al4C4Si E9A NiSbS (ullmanite) FO1 FeKS2 NaCrS2 CuS2Sb (wolfsbergite) Al 2 MgO 4 (spinel) Hl1 Cu 3 VS 4 (sulvanite) H24 Cu2FeSnS4 "26 Fe 4 N LY CuPt3 LK AuCu LU AuCuII Ll0(M) CuPt Ll1 AuCu 3 Ll 2 AlFe3C (perovskite) Ll2' 5CuTi ThH 2 ^ . Di. Di. DS1.3 C2/c R 3m Pnma Fd3m PA3m /42m Pm3m FmIc PA/mmm Imma R3m PmZm Pm3m PA/mmm IA/mmm Fm3m Irnlm P6Jmmc PA/mmm . AlCu 2 Mn (Heusler) Llx Sb2Tl7 Ll1 Fe 2 N Ly CuTi3 ImJ Im3m Ibam P63/mmc R 3m P6 3 /m IA/mcm IA/mcm Im3m /43m PA3_m Fm 3m R 3m IA3d P63/mcm Fm3m R3m P63/mmc Pnma P63mc PA/nmm P2x/c Cmcm P2/c IAlA Pm 3m IA PA/mnc Pl2m P^/mmc Ia3 Pnma Fd 3m P63mc P2.Arranged Alphabetically by Strukturbericht Designation Strukturbericht designation DVC D0d DO. Dl. Dl1. Dh DX. D2C Dl. DO. 00» DO21 Z)O22 DO1. D7» Dl1 Dl1 Dl. Cu16Mg6Si7 (G-phase) oCrFe Mg 2 Zn n Pearson symbol //16 o P 16 //32 c/32 cF\6 cPA oP\6 0PI6 hPS hPS oP\6 hP24 Space group IA/mcm Pmmn IA Prototype Pearson symbol mP22 cl162 c/40 o/28 hPlA hRl hP20 //32 //32 c/52 c/52 cP52 CF116 HR13 c/76 hP\6 cF6S hR 26 hP2S oPAQ hP20 /P6 mP2A oC 16 mP 12 //16 cP5 //14 /P40 HP 18 HP26 c/96 oP24 cF112 AP18 cP12 mC16 hR A oP 16 cF56 cPS //16 cP5 cFXl tPA o/40 hR7>2 cPA cP5 XFl //6 cFlb c/54 hP3 tPA Space group P2.

...5 1.6 2....................................................................................................................... 2 B2 (cP2) Compounds ...................................................... 25 25 26 34 36 37 39 39 vii .......................................................................................... L1 (cP4) Compounds .............................................. References .1 1...................................8 1.................................................................................... Summary and Conclusion ................................................................ Characteristics of Elastic Moduli ................6 1..... Intergranular and Cleavage Fracture .........................................................2 2........................ Anisotropy of Elastic Properties for Single Crystals ............................................................................................................................................................................................................................................................ Acknowledgement .................... Preface to the 1995 Edition ..............7 Introduction ..................3 2............................................................................................... References ........................................................................................7 1..................................................................................4 2.........1 2.....9 Introduction ...................... Other Intermetallic Compounds .................................................... 1.....3 1................................................ Acronyms .......................................................................... 2.............................................................. Elastic Constants of Single Crystals ...............Contents Contributors ............ Elastic Properties of Dislocations ................................................................................................................................ This page has been reformatted by Knovel to provide easier navigation................................................ Elastic Moduli of Polycrystalline Intermetallics .2 1.................... 1..............................................................................................................................................4 1................................................. Elastic Properties .. Crystal Structure Nomenclature .................................................... ix xi xiii xv xxi 1 1 1 2 8 10 14 20 20 20 22 Addendum .......5 2............ Models for Brittle Fracture .......................... Conclusions ................................................ 2... Preface to Reprint Volumes ................................................................

............................................ 4.................1 5.................................. APBs in Dissociated Dislocations ............................................................................................ Acknowledgements ............4 5.............................. Point Defects .................................................. 131 6................................... References .......................................................................................... Conclusion .......................... Dislocations ....................................... 5.....................2 4.........................................................................5 3................4 3.............................. Dislocations Debris ............................................................................................1 6........................1 4..........................................................7 Introduction ...........................4 4..............7 5.................................................................................5 5................................................ References .............................. Stress and Temperature Dependence of Strength and Ductility (Deformation Maps) .....viii Contents 41 41 43 48 49 58 58 58 64 3.............................................................................................. Core-Related Mechanical Properties ....8 Introduction ............3 4..................2 5....................................................................... Plasticity-Controlling Factors ............. Sessile Configurations ..5 4.......................................... 67 67 69 81 83 86 86 87 89 Addendum ............................ Addendum ........................................................................6 3.................................................................................................................. 4....................................... This page has been reformatted by Knovel to provide easier navigation............... 3.................................................7 Introduction ...........................3 3.............................1 3.............................................................. Grown-in APBs and APDs ..... 131 132 ......................... Deformation Mechanisms ......................6 4... Structure of Antiphase Boundaries and Domains ...................................... Acknowledgements ................................................................................. References ........... Experimental Techniques for Studying Point Defects ................................................... Concluding Remarks ............................................................................................................................................................................................... 91 91 95 113 115 118 122 124 125 6......3 5................................................... 5...........2 3...........6 5.................................................... Plastic Deformation ................................................................................................... Perspectives of Materials Developments ............................................................... Slip Systems in Intermetallics and Dislocation Cores .......2 Introduction .......... The Structure and Energy of APBs ........................................................................................ The Fine Structure of Superdislocations in Relation to Crystal Structure ......................................................................

...............................5 7....................................................... Thermodynamics of Point Defects ....................................................3 6...................... Structure of Grain Boundaries .................6 6.......................................................2 8......................................................................................... Acknowledgements ...........................................................................1 8....... Specific Systems ............. Direct Observations of Structure ..... ix 136 138 140 142 145 146 150 152 152 153 153 6....................................................... Point Defects in Intermetallic Compounds ..............................2 7............. Migration Properties of Point Defects .................................5 6........... 181 183 189 198 212 212 215 Addendum .......................................................................................................................... 6.................. Acknowledgements ................................ Summary ..................................................................................................................................................6 Introduction .......................................................................... Interpretation of Various Property Measurements by Grain-Boundary Structure and Chemistry ......................................................... Modeling of Grain-Boundary Structures ...................7 7............................................................ Direct Observations of Chemistry . 6.. Production and Structure of Interstitials ....... 7............................ References ......................................................................................................................................8 Introduction ............... ... 181 8........................................................1 7.......................................................................7 6................................................. This page has been reformatted by Knovel to provide easier navigation................8 6............................. Electronic Bonding Aspects ...................................... Irradiation Damage ........................................................................................................................9 Experimental Studies of the Formation of Vacancies and Antisite Defects ...................................................................................................... References ............................................... Recovery of Point Defects after Irradiation or Quench ..............3 7.......................................................10 Influence of Point Defects on Physical Properties ..............................................Contents 6............................................ 157 7.........4 6...................... Irradiation-Enhanced Ordering and Irradiation-Induced Disordering of Ordered Alloys ..................................5 8........ Irradiation-Induced Amorphization .........6 7..............................11 Conclusion ...................................................................................................................4 7.......... 157 158 163 167 170 175 176 178 8...3 8..................4 8.............. Theory of Point-Defect Formation Properties ........12 References ..................................

......................................... Massive Transformations ................................................... 239 240 242 Index ............ Stress...............................................................6 9.....................................................................................................................................................................................................................7 9......................................................and Strain-Induced Martensitic Transformations ........5 9.................................................. 221 9.............. Martensitic Transformations in Ceramics ................2 9..................x Contents 9................................................... Kinetics of Martensitic Transformations ..............................9 Scope ...10 Twinning . ...... 221 221 225 225 225 226 229 229 237 9......... 9............................................. 249 This page has been reformatted by Knovel to provide easier navigation.... Bainite Transformations .............................1 9..................................3 9........... Addendum ..........................................4 9............................................................8 9............... Crystallographic Transformations ................................................ Shape-Memory Effects ............................................................................................. Nucleation of Martensite ............................................................................................. Crystallography ............................11 Applications of Phase Transformations to Structural and Functional Materials ...................................................... 238 9.....12 References ..................................................................................................................................................

The fracture strength ac of brittle materials is represented by Griffith's relation (Griffith. 1920): oc = 2(yE/cy> (2) where 7 is the surface energy. Tsukuba-shi. the effects of various factors like temperature on the elastic moduli are discussed. b is the magnitude of the Burgers vector. followed by a summary of the elastic moduli of polycrystalline intermetallics. The velocity v of the ultrasonic wave in a material is given by the following equation: v = (M/p)XA (3) Here. Westbrook and R L Fleischer ©1995. Nabarro. the properties of dislocations like Peierls stress. The relation between stress a^ and strain Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. 1-2-1 Sengen. Ibaraki 305-0047 1. H. They give information about the stiffness of the material and are used for understanding of mechanical properties. The Peierls stress rP is given by the following equation (Peierls.v) ] exp [ . and d is the distance between slip planes.. In this chapter. 1940.v) ] (1) Finally.Chapter 1 Elastic Properties Morihiko Nakamura National Research Institute for Metals. v is Poisson's ratio. 2000 John Wiley & Sons Ltd . which results from the adiabatic behavior of the specimen during the propagation of an ultrasonic wave pulse in the latter. For example. are explained by elastic theory. etc. Elastic Constants of Single Crystals Elastic modulus values are classified into two groups: one is the static modulus.2-KdIb(X . elastic constants of single crystals are first summarized. and c is the size of an internal crack. E is Young's modulus. 1973). The former is called the isothermal modulus' and is obtained from the linear relationship between load and displacement of a specimen. and the anisotropy of elastic properties of single crystals is discussed using Young's modulus. G is the shear modulus. 1947): rp = [2G/(1 . The latter is called the 'adiabatic modulus' and is determined from the resonance frequency or the velocity of an ultrasonic wave (USW) in a specimen. p is the density and M is the appropriate elastic modulus. 2. The difference between them is caused by thermal expansion. Edited by J. interaction between dislocations. For example. Then. since we have a textbook that reviews experimental procedures (Schreiber et al. a relatively large specimen is needed for the static measurement of a small strain. and the other is the dynamic modulus. the elastic modulus is usually determined from the velocity of an ultrasonic wave in a single crystal of a material. etc. for which it is difficult to prepare a large specimen. self-energy. Poisson's ratio. and are often used for the estimation of an interatomic potential that is applied in a computer simulation. Some difficulties cannot be avoided in the determination of the isothermal modulus. Here. Thus. Introduction Elastic constants are fundamental physical constants that are measures of the interatomic forces in materials. the elastic properties of dislocations are discussed using elastic constants of single crystals.. The experimental details for the determination of elastic constants are not described.

. C13 = C23. or an all-electron self-consistent for full-potential linear muffin-tin-orbitals (F-LMTO) method (Alouani et al. Kayser and Stassis. etc.c 1 2 ) / 2 . specific heat at constant pressure C p . MoSi2. For example. respectively. Pt 3 Al. In contrast. 1991). the first-principles full-potential linearized augmented-plane-wave (F-LAPW) total-energy method (Fu and Yoo. The isothermal moduli can be calculated from the adiabatic ones using the thermal expansion coefficient a. Anisotropy of Elastic Properties for Single Crystals When a tensile or compressive test is carried out. although C12 is close to C66 ( = C44) for intermetallic semiconductors like GaAs. and c. 1987. C33^O. The number of independent components ctj decreases with increase of symmetry from 21 for a triclinic crystal to two for an isotropic material. C 12 ^O. Recently. The elastic constants are closely related to the potential energies between atoms. The number of independent ctj is three for a cubic structure like L l 2 (cP4). the applied stress is usually controlled. It is five for a hexagonal structure like DO19 (hP8). and C^ = O for / > 3 o r y > 3 and /Vy. It is two for an isotropic material: C11 = C22 = C33.y O'. c-ti is the Voigt notation for the components of the elastic stiffness tensor. Cn = C13 = C23. it is more convenient to use the following equation instead of equation (4) for the elastic stress-strain relationship: .. Table 3 shows the comparison between the adiabatic and isothermal elastic stiffness constants for cubic crystals. C l l b (tI6). 1989). 1981. C44 = C55 = C66 = (C1. Ni3Al and Ni3Fe have similar constants to Ni. C33 *0. and density Q (Garber and Granato. The calculated elastic stiffness constants are shown in Table 2. In general. C12^O. 1959). Rausch and Kayser. C44 = C55. B2 (cP2). Table 1 shows that generally the Cauchy relationships are not true for intermetallic compounds. It should be noted that the calculated values are in good agreement with the experimental ones for Ni3Al and MoSi 2 (see Table 1). 1986). C 1 2 = C 1 3 = C 2 3 . Mdetermined from the velocity of the transverse wave in the [ 1 0 0] direction on (0 0 1) and (0 1 0) planes is C55 and C66. Cauchy's relationships hold true for the elastic stiffness constants (Love. by measuring the velocity of the longitudinal wave in some crystal orientations and that of the transverse wave on certain crystal planes in some other orientations. and ZrAl3 exhibit higher stiffness constants than their constituent metals.3) for Cartesian coordinates is represented by the generalized Hooke's law. etc. DO23 (tI16). When the interatomic force is a central force. 3. Table 1 shows the elastic stiffness constants ctj of some intermetallic compounds at room temperature. the elastic constants of Ni 3 Al.e. C40 (hP9). a linear augmented-plane-wave (LAPW) method (Mehl et al. the jc3-axis is selected for the most symmetrical axis of a crystal structure: and the elastic constants of Cu 3 Au are similar to those of Au and Cu. and the strain is measured. C 4 4 = C 5 5 = C 6 6 .. Thus.y = 1. C44 = ^55vC66 = ( c n . Thus. The methods of first-principles calculation of elastic constants are summarized in Chapter 9 by Mehl et al.: CxI = C22 = C 3 3 .. C13 = C23. 1977). etc. when the elastic constants were calculated using a central-force approximation for the forces between atoms. intermetallics do not always have higher stiffness constants than those of the constituent elements. depending on crystal symmetry.. and c. C66^O. an elastic-stiffnessconstant matrix [c£ is obtained. The former are about 21Vb higher than the latter. a n d ctj = 0 f o r / > 3 or j> 3 and /Vy. CoAl.2. is C11. WSi2. averaged. MoSi 2 .c 1 2 ) / 2 . etc.y = 0 / > 3 ory > 3 and /Vy*. 1990).y = 0 for / > 3 o r y > 3 and /Vy.: Cn=C22. TiAl3. which is determined from the velocity of the longitudinal wave in the [ 1 0 0 ] direction. The appropriate elastic modulus M. SbY. 1944): C 23 = C 44 C3! = C 5 5 C12 = C66 (5) C (4) 14 = C 56 C25 = C 4 6 C45 = C36 Here. It is six for a tetragonal structure like DO22 (tI8). they disagreed with experimental values (Girifalco and Weizer. Thus. in this volume. Chen et al. etc.: Cn = C22. Although alloys often exhibit some maxima of Young's modulus at compositions corresponding to intermetallic compounds. have been calculated using an embedded-atom method (EAM) (Foiles and Daw. 1975.

22 Cn 0.2935 1.4366 1.491 Cn 0.2123 Cu n Cu 0.3845 0. CoSi.086 Cn 3.2240 0.872 1. cF24 Shannette and Smith (1969) C15.852 Cn 2.". cF24 Mock and Guntherodt (1984) C15.175 2. cF8 Smithells (1976) B3." Mg2Si Ni3Al Ni3Ai Ni3Al Ni3Fe Ni.317 Cu 0.248 C44 Cu Cn 0.7047 C44 Cu n 0.3311 2.165 0.428 3. cP2 Yoo (1987b) B2.25 Cu u 1.445 1.9443 2. 1993) C40. (1991) DO3.83 u Cu 0.903 0. cF12 Nakamura (1994) C l .82 Cu u 1.6253 C 0. cF24 Cheng (1967) A15.6253 0.23 1. (1972) A15.03 1.6163 0.340 0.489 Cu Cu 0.35 Cn 0.Table 1. (1965) B3.121 Cu Cu c c 0.29 n 0. cP4 Yasuda and Koiwa (1991) Ll.506 1. Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn c c n Cn n n c n c Cn Cn Cn c n Cn c C n n Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn n c n Cn c 0.192 0.920 1.175 Cn 2.375 Cn 0.826 Cn 0. (1990) C l .401 Cn Cu 0.266 Cn 1.437 Cn u u 1.042 0.179 Cu Cu 1. cP2 Yoo (1987b) B2.228 2.46 2.87 1.578 3. hP9 Nakamura (1991b. cP4 Yoo (1987b) Ll. cF12 Whitten et al.567 n C44 0.907 1.705 CM 2.1271 0.3659 0.24 u u 1.722 1.428 n Cu Cu c 0.823 Cu Cu C 1. cP4 Siegel (1940) L T2.84 0. CeAi. (1992) LU. Villars and Calvert (1985).210 2.838 1.7084 1.430 n 1. cP2 Zirinsky (1956) B2.28 1.5947 Cu Cn Cu c 0.8366 Cu Cn C44 1. cF8 Smithells (1976) B3.9702 1.269 1.870 0.681 Cn 0.370 Cu Cu 1. cF24 Penney et al.357 C Cu < 0.980 2.230 Cu Cu c 1.606 2. cF12 Guenin et al. (1967) B2.829 Cn 0.07 u Cu c 1.2330 0.088 4. cP4 Miller et at.68 u u 1.204 4.089 0.885 0.170 4.4303 cu n Cu 0.56 n 0.822 Cu Cu 0.398 C66" Cu 1. cP2 Schiltz e/ al.046 0. hP8 Nakamura (1991a) C54. (1981) B3.404 c 0.1805 Cu C 66 0. (1981) B2.J/Z n U.091 Cn 0. (1990) C40.116 2.6051 1. cF8 Miller et al. cP4 Yasuda et al. (1971) B2. MgCu2 Nb3Sn Nb3Sn Nb3Sn V3Ge V3Si HgTe GaAs GaAs GaSb InSb InP GaP PbS ZnS CdS "C66 = 2.240 Cu Cu c c 1.710 0.93 C44 0. HfCo. tI6 Nakamura et al.220 0.48 Cu u c 0.367 c 0. cP4 Yoo (1987b) Ll 2 . cP4 Yasuda et al.577 44 0..1735 0.433 Cn u Cu 0.412 1.150 C^ Cu C 66 0.966 2.810 Cn 0.5622 Cu Cn Cu c 0.7047 cu Cn Cu c 0.6631 Cu Cu c c U.042 Cu 2.766 Cn 0. cF16 Leamy et al. cF8 Hickernell and Gayton (1966) B3.15 C12 CM CM CM CM CM C M c C n M CM C M CM CM C M M CM CM C CM CM CM CM CM CM CM CM CM CM CM CM CM C M CM C M CM C C M M CM CM C C C M M M CM CM C C M CM CM CM M M CM C 0.217 Cu 0. cP2 Smithells (1976) B2.4755 Cu Cu C 1.455 Cn 0. cF8 Weil and Groves (1971) B3. cP2 Zirinsky (1956) B2.491 u 0. GdAf.576 0.4013 cu Cn Cu 1.113 0. cP8 Chang and Barsch (1980) AiS.300 c 0.5635 cu 0. cP8 Rehwald et al.653 0.485 Cu 0. cF8 Smithells (1976) B3.406 Cn C44 u 1.175 1.8600 Cn CM CM CM C M Cn CM CM C C Cn CM CM CM CM C M M CM CM CM CM CM C M CM C M M CM C CM C M M CM C Cn CM CM CM C C C C C M CM CM C M M M CM Cn C M M M M CM C M CM C C CM CM M 0.9030 cu Cn C44 0.570 f3 3 DO22.3960 Cu Cn Cu 1. cP2 Wasilewski (1966) B2. (1982) C15.23b 2.412 Cn Cu Cu c 1.5625 0.".5554 1.758 1. cP2 Cheng (1967) B2.5361 1. cF24 Shannette and Smith (1969) C15.721 Cn 0. cF24 Butler et al.3404 0.29 1. (1990) C l l b .4600 Cn 0. CeAl" ZrCo-.496 2.53 1.022 0. HoCo.223 Cn 3.202 0.279 0.5452 Cu Cu 0.91 2.124 0. hP12 Sumer and Smith (1962) C15. 1993) DO19.612 0. hP4 Nadgomyi (1988) 'Pearson (1958). cP8 Rosen and Klimker (1969) A15.581 0.538 n C44 Cu 0. (1971) B2.263 Cn 0.418 0. cF24 Schiltz and Smith (1974) C15.450 Cu Cu c c 1.4731 Cu Cu c 1.132 Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn Cn c n NiAl AgMg AgMg AgMg CuZn CuZn AuCd AuZn AuZn YZn CoTi Fe3Al CaMg 2 CaAl 2 YAK LaAl.770 2.Fe Ni3Ga Ni3Ge Ni3Mn Cu 3 Au Cu 3 Au Al3Li Cn C2I C33 C 44 C C66 Cn Cn C 23 Structure* References 2.306 Cn 1. (1982) C15.672 1.083 Cn 5.432 Cn 1.10 cu 0. cP8 Testardi et al.115 0.145 Cn 5. cP2 Yasuda et al. cF8 Nadgornyi (1988) B4. cP2 Chang et al. cF24 Schiltz and Smith (1974) C15.47 1.390 3.177 2.50 1. tI8 Nakamura and Kimura (1991) DO23.042 1. cP2 Schiltz et al. oF24 Nakamura (1994) C l . cP4 Yasuda et al.030 Cu Cu 1.938 0.383 Cn 0.297 0.. (1967) C14.3655 Cu Cn Cu 0. (1965) Ll.838 Cn 0.907 (Cn-C12)Zl. cP8 Keller and Hank (1967) \ 1 5 .936 1.3200 0. cP4 Yoo (1987b) LU9 cP4 Ono and Stern (1969) Ll. til6 Nakamura and Kimura (1991) C l l b . cF24 Schiitz and Smith (1974) Cl5.460 Cn 1. cF8 Weil and Groves (1971) B3.523 Cn 4. (1992) Ll 2 . (1992) Ll.453 C 1.8096 Cu Cn Cu c 0.68 Cu Cu 1.428 Cu u 1.538 Cn 1. tI6 Nakamura et al. hP9 Nakamura (1991b.159 0.48 Cn 1. .21 2.. Elastic stiffness constants (102 GPa) of intermetallic compounds at room temperature Intermetallics TiAl 3 ZrAl 3 MoSi2 WSi2 VSi2 CrSi2 Ti3Sn TiSi2 CoSi.706 0.846 0.538 2. cF24 Schiltz and Smith (1974) C15. cP4 Dickson and Wachtman (1969) Ll 2 .508 1.460 n Cu Cu 0.3679 0. cF8 Nadgomyi (1988) B3.022 1.40 Cn 0.653 Cn 1.48 Cn 1.464 Cu Cu c 1.630 2.23 Cn 1.6978 Cu Cn Cu 1.47 2.412 1.461 Cn Cu 0.599 0.302 Cn u Cu c 0.

20 2.78 c n c n Cn 2.15 1.52 2.5 2. When a uniaxial stress an is applied in the X1-direction.20 1.8 C33 C44 <*66 Cn C 13 Structure0 Ll 2 .30 Cn Cu 0.98 5.35 4. when the rotational transformation of coordinates is given for the compliance-constant matrix Is0]. cP4 Ll 2 .57 3. "Linear augmented-plane-wave method.36 1.SnZs1x for a single crystal with different orientations.3 1. d All-electron self-consistent full-potential linear muffin-tin-orbitals method.60 0.23 Cu 1.8 Cn Cn 1. 'Lattice parameters of a face-centered tetragonal lattice are used for a Ll 0 (tP4) structure. Similarly.89 2. Comparison between adiabatic and isothermal elastic constants (102 GPa) at room temperature Adiabatic elastic constants Intermetallics Ni3Al* Fe3Sic Cn Isothermal elastic constants Cn C44 C 44 Cn Cn Structure" Ll 2 . Table 3. S66. cF8 B2. (1991) Fu (1990) Fu (1990) Alouani et al. *Kayser and Stassis (1981).2 Cn Cn Cn Cn Cn Cn 0.45 1. (1990) Mehl et al.9 0.7 1.72 4. (1986) Fu and Yoo (1989) Fu and Yoo (1989) Guo et al. Young's modulus is given by the relation between stress and strain in the direction of the applied load.75 4. (1990) Mehl et al.53 'Pearson (1958). cF16 2. cP4 DO3.40 Cu Cu C 5. (1990) 1.66 Cu 1. cP4 Ll 2 .02 0. cP4 Ll 2 . Young's modulus E1 in the X1 -direction of a single crystal is represented by the reciprocal of S11.88 1.02 4. .5 "Pearson (1958). cP2 B2. cP2 Ll 0 .36 1.94 0.32 1.20 2. and Poisson's ratio are given respectively by S11.52 2.48 1. Villars and Calvert (1985). cP4 DO22.45 2. *Embedded-atom method. [S0] is the elastic-compliance-constant matrix. (1990) Mehl et al. the shear modulus G12 in the X1-direction on the *2-plane or in the x2-direction on the X1 -plane is given by the reciprocal of S66. Calculated elastic stiffness constants (102 GPa) of intermetallic compounds Intermetallics Ni3Al Ni3Al Ni3Al Pt3Al Al3Li Al3Sc TiAl3 MoSi2 SbY CoAl RuZr NbIr' NbIi* Method EAM* EAM* F-LAPW F-LAPWC F-LAPWC F-LAPWC F-LAPWC F-LMTO^ LAPW* LAPW' LAPW* LAPW* LAPW* 2.08 Cu 0.26 Cu 1. tP4 References Foiles and Daw (1987) Chen et al.07 1.25 1. when a uniaxial stress is applied to the specimen: £n=si\°n or <ju = (l/su)en (7) Therefore.43 0. cP4 Ll 2 . that of the shear modulus. (1990) Mehl et al. Poisson's ratio *>12 in the x2-direction is given by: "12= -€22/*"= ~sn^n (8) Here.01 1. tI6 Bl.28 1.45 2. tP4 Ll 0 .36 1. tI8 Cll b .43 0.75 2. and .75 4.46 2.255 C44 1.78 Cu Cu 1.37 1. 8 Lattice parameters of a body-centered tetragonal lattice are used.Table 2. the reciprocal of Young's modulus. cP4 Ll 2 .36 2.23 2.3 5. (1991) Mehl et al.25 1. c Rausch and Kayser (1977).32 Cu Cu 1.77 C44 u Cu 0.57 2.37 1.58 1. symmetry Thus. Tirst-principles full-potential linearized augmented-plane-wave total-energy method.9 2.

the cross-section changes to an elliptical shape with the major axis in the [ 0 0 1 ] direction and the minor axis in the [ 0 1 0 ] one. and lowest in the [ 1 0 0 ] and [ 0 1 0 ] directions normal to the [ 0 0 1 ] direction. the X1 -direction on the (T 1 0) plane changes from the [ 0 0 1 ] to the [ 1 1 0 ] direction. when the crystal is rotated around the axis perpendicular to the (1 1 0) plane. i. When the stress is applied in the direction perpendicular to the (0 0 1) or (T 1 0) plane. that of the reciprocal shear modulus S66. for example. It is. tI6 structure) single crystal (Nakamura. represent the directional cosines at an Orientation of applied stress (deg) Orientation of shear stress (deg) Orientotion (deg) Figure 1. When the X1-. (a) S11. X2-. x' 2-. [h &/] r a = direction of rotation axis. the X1 -direction changes from the [ 0 0 1 ] to the [ 0 1 0 ] direction. when the specimen is loaded in tension in the [ 1 0 0 ] direction. and jc3-axes are rotationally transformed to the x' r . In Figure l(a). For example. api. because of the anisotropy of this structure. In contrast. Poisson's ratio Se6(IO-3GPa1) S11(IO3GPa-1) . the cross-section of the specimen hardly changes from a round shape. and is lowest in the [ 1 1 0 ] direction. etc. and Ar'3-axes. remarkably dependent on the stress direction. (h k /) sp = shear plane. Poisson's ratio is low and is slightly dependent on the crystal orientation. dependent on the orientation for the stress direction perpendicular to the (100) plane. 1991b) angle between the x. Orientation dependence of the compliances Sn and S66 and Poisson's ratio for a WSi2 single crystal.e.and x^-axes. and Poisson's ratio v2l. Then. 1991b). Figure l(c) shows the change of Poisson's ratio v2i with rotation around the x2-axis. Figure 1 shows the orientation dependence of the reciprocal of Young's modulus Sn. the direction perpendicular to the (10 0) plane. (b) S66. It is found for WSi2 (Cll b . when the crystal is rotated around the applied stress axis. the components of the compliance tensor [s'pqrs] after transformation are given by: (9) Here. (c) Poisson's ratio. calculated from the elastic stiffness constants shown in Table 1 for a WSi2 ( C l l b . The shear modulus (IZs66) is independent of the stress direction on the (0 0 1) plane. when a round bar specimen prepared from a single crystal is loaded in tension elastically in the [ 0 0 1 ] or [1 TO] direction.The compliance-constant matrix [s^] is a simplified notation of the tensor {$#*/). i. normal to (h k I) plane (Reproduced by permission of the Japan Institute of Metals from Nakamura. the X1 -axis changes from the [ 1 0 0 ] to the [ 0 1 0 ] via the [ 1 1 0 ] direction. Young's modulus (IA 11 ) is lower than in the [ 0 0 1 ] direction.e. It is highest in the [ 0 0 1 ] direction. when the crystal is rotated by 90° around the [ 0 0 1 ] axis. (h k /) A s = direction of applied stress.. for example. tI6 structure) that in one of the close-packed atomic arrays. The shear modulus on the (T 1 0) plane is. the [3 3 1] direction in which W and Si atoms are arranged in a certain order. Thus. and exhibits a low value in the [ 0 0 1 ] direction and a high value in the [ 0 1 0 ] direction. however. In Figure l(b). This indicates that the shear modulus does not always exhibit a low value in the direction of atomic arrays with a small interatomic distance. the compliance-constant matrix [spq] is obtained for a single crystal with a desired orientation when the tensor {spqrs} is rewritten in the Voigt notation. along which the uniaxial stress is applied.. however. and slightly dependent on the stress direction on the (1 0 0) and (1 0 3) planes.

1985) in the [ 0 0 1 ] direction is larger than that (0. Villars and Calvert. 1985) is . (c) Poisson's ratio -S12Zsn. (b) S66.3848 nm) in the [ 1 0 0] direction. The value of Sn is higher in the [ 1 0 0 ] direction than in the [ 0 0 1 ] one. \/E and 1/G represent the reciprocals of Young's and the shear moduli. S11 depends remarkably upon the crystal orientation. su is markedly high. That is. Orientation dependence of the compliances S11 and S66 and Poisson's ratio for a TiAl3 single crystal: (a) su.2884 nm. respectively (Reproduced by permission of Chapman & Hall from Nakamura and Kimura.S11(IO-3GPa1) Orientation of applied stress (deg) S66(IO-3GPa1) Orientation of shear stress (deg) Poisson's ratio Orientation (deg) Figure 2.4296 nm. Young's modulus in the [0 2 1 ] direction is not so large. Villars and Calvert. The values for a polycrystalline material are estimated from the elastic constants for a single crystal. along which Ti and Al atoms are arranged alternately in close-packed atomic arrays. Young's modulus is very high in the [ 1 1 0] direction. The interatomic distance in the [ 0 2 1 ] direction (0. In the [0 2 1 ] direction. and it is considerably low. 1991) shows similar relationships to Figure 1. This behavior is not obvious from the fact that the interatomic distance (0. that is. When the stress direction is rotated around the [ 0 0 1 ] axis. and is close to that in the [ 0 0 1 ] direction. 1991) Figure 2 for a TiAl3 single crystal with a tetragonal DO22 (tI8) structure (Nakamura and Kimura. compared with that in the [ 1 1 0] direction. along which Ti and Al atoms are arranged in closepacked arrays. similar to those of the [ 1 1 0] direction.

Figure 3 shows the orientation dependence OfS11 for a Ni3Al single crystal with a cubic Ll 2 (cP4) structure for the [11 0] rotation axis (Nakamura. about 0. The fact that Poisson's ratio exhibits a negative value means that the uniaxial tensile stress perpendicular to the {1 1 0) planes induces a positive strain in the <1 1 0> direction. Although Ni3Fe contains a large amount of Fe. 1991a). S66 on the (1 1 0) plane. Young's modulus is not the highest in the direction along which Ni and Al atoms are arranged alternately in the closest-packed manner. i. When the stress direction is rotated around the c-axis.e. (Reproduced by permission of the Japan Institute of Metals from Nakamura. When a uniaxial stress is applied in a direction perpendicular to the [0 0 Ij or (1 1 1} planes. Orientation dependence of the compliance S11 for a Ni3Al single crystal. However. a large elastic distortion is produced in the {1 1 0} plane normal to the stress direction. respectively. and close to that in the [11 0] direction (0. that is. the shear modulus is not always lower on the closest-packed {111} planes than on the (1 1 0} planes. That is. Poisson's ratio is rather dependent on the crystal orientation: Poisson's ratio in the [001] direction. S66. as shown in Figure 2(c).8) in the <0 0 1) direction and extremely low (i. rather it is the highest in the <1 1 1 direction. the minimum one.1 ) . In this figure. the maximum value. i. S11 is independent of orientation. It also exhibits little orientation dependence on the (110 0) and (112 0) planes.e. Poisson's ratio as well as S66 are large and independent S 11 (IO 3 GPa. For the stress axis perpendicular to the {11 Oj planes. The radial and angular coordinates represent S66 and the crystal orientation.4298 nm). Figure 5 shows the orientation dependence of S66 for Ti3Sn (DO19. The orientation dependences of S11 and Poisson's ratio were also calculated. Poisson's ratio is independent of the orientation. > The orientation dependences of S66 and Poisson's ratio were also calculated. S66 is low in the <0 0 1> directions. The compliance S66 exhibits no orientation dependence. low in the <1 1 1> one. as shown in Figure 2. When a uniaxial stress is applied in the [100] direction. respectively. [1 1 0O](OOO 1) means that the shear stress is applied to the [110 0] direction on the (0 0 01) basal plane. Ni3Fe.2) in the <1 1 0> direction (see Figure 4).2721 nm). and it is generally higher than on the closest-packed (112) plane. 1991a). although it has a common trend for single crystals with a tetragonal structure. and relatively low in the <1 1 0> one. These facts indicate that Young's modulus is strongly influenced by the arrangement of Ti and Al atoms. That is.e. especially in the [010] direction. and Poisson's ratio for the stress direction perpendicular to the (1 1 0) plane exhibit remarkable orientation dependences. and high in the <1 1 0> directions. and has a smaller orientation dependence than Ni3Al. S11 is low in the [0001] direction. The elastic moduli of Ni3Ge are much less dependent on the orientation than those of Ni3Al. The value of sn is high in the <1 00> direction. it is found that the shear modulus is lower in the [ 1 1 0] direction on the (1 1 0) plane than in the directions of the closest-packed atomic arrays on the closest-packed planes for TiAl3 as well as for WSi2. It is high on the (11 0) plane. and it is very high (i. Here.S66 is Orientation of applied stress (deg) Figure 3. 1991a). it exhibits an orientation dependence of the elastic moduli extremely similar to that of Ni. especially in the [1 1 0] direction. a large strain is observed in the [010] direction.considerably smaller than that in the [001] direction (0. which contain the component of the c-axis on the (1121) and (lT 01) planes. 1991a) independent of the stress direction on the (0 0 1) and (1 1 1) planes. AuCu3) structures. is four times larger than that in the [110] direction. and compared with a Ni single crystal with an Al (cF4) structure (Nakamura. the shear modulus is constant and low on the (0 001) plane. and a small strain in the [001] direction normal to the [0 1 0] one. hP8. When the stress direction is perpendicular to the (T 1 0) plane. Also.0. it changes remarkably with the crystal orientation. The value of . the curve for Ni3Ge is coincident with that for Ni. S66 is low in the [1126] and [T 10 2] directions. The values Of^11 for rotation around the [11 0] axis. Ni3Sn structure) (Nakamura. and Ni3Ge single crystals with Ll 2 (cP4. about . on the {1 1 0) planes. although it causes a large negative strain in the <0 0 I) direction normal to the <1 1 0) one. That is.e. The compliance S66 is low on the (100) plane. The orientation dependence of the elastic moduli changes with the intermetallic compound. Figure 4 shows examples of the orientation dependences Of1S11. and Poisson's ratio for Ni3Al. Therefore. ra = rotation axis.

Figure 6 shows examples of the elastic fields around a screw dislocation line with a Burgers vector of i [ 3 3 1] on the (103) plane for MoSi 2 (Cll b . respectively..Orientation of applied stress (deg) and S11 in the [ijk] direction. appears. and the associated stiffness-constant and compliance-constant ratios for various intermetallics (Nakamura. the ratio Of^11 does not correspond to the interatomicdistance ratio. In WSi2 and MoSi2 the interatomic-distance ratio corresponds relatively well to the ratio of C11. and (c) Poisson's ratio for the stress direction normal to (1 10) plane (Reproduced by permission of the Japan Institute of Metals from Nakamura. the x3-axis perpendicular to the figure represents a dislocation line. ZrAl3. and Ni3Al single crystals. and CuZn the former does not correspond to the latter. The value of c[ijk] is related to the tensile or compressive stress. 1968. and Poisson's ratio for Ni3Fe. and the normal stress field all9 which does not exist in an isotropic material. The stress field O13 is slightly distorted from that of an isotropic material. In general. when the stress is applied in a direction perpendicular to the (1 100) or (1 12 0) plane. This indicates that the atomic arrangement affects the elastic properties of intermetallics as described before. The stress fields around an edge dislocation are also slightly distorted. Elastic Properties of Dislocations Poisson's ratio Orientation (deq) Figure 4. corresponding to one of the closestpacked atomic arrays. (b) S66 on (1 1 0) plane. In TiAl3 or ZrAl3. 1991a). c[ijk] ands[/7Ar] represent C11 Anisotropy of the elastic constants results in a change of elastic properties of dislocations in intermetallics. although the atomic distance in the [110] direction is approximately equal to that in the [2 0 1 ] or [4 01] direction. S66. It is found that Poisson's ratio is low and dependent on orientation. We have textbooks for calculation of the elastic properties of dislocations using anisotropic elasticity theory (Hirth and Lothe. The atomic arrays of intermetallic compounds may affect the orientation dependence of the elastic constants. is lower than that in the [2 0 1 ] or [401] direction corresponding to other closest-packed arrays. In this figure. The elastic energy per unit length of dislocation between two cylindrical surfaces of radius r0 and R is given by: (Kb1ZA-K)XOg(RZr0) (10) . 1991a). The contour-line maps for ol3 and an are shown in Figures 6(a) and (b). Table 4 shows the ratio of the interatomic distance in the [ijk] direction to that in the [001] direction. and the *2-axis is perpendicular to the (1 0 3) plane. NiAl. The elastic stress field around a dislocation affects the elastic energy of a dislocation and the interaction between parallel dislocations. in WSi2 and MoSi2. for which the ratio of C11 is close to the interatomicdistance ratio. S66(IO-3GPa1) S11(IO-3GPa*1) Orientation of shear stress (deg) 4. The results calculated from the elastic constants of single crystals are summarized. the elastic deformation in the direction of the applied strain may be determined mainly by the interatomic force in the strain direction. Ni3Ge. Here. and Ni: (a) S n for rotation around [ 1 1 0 ] axis. tI6 structure) (Nakamura. 1973). the ratio of C11 in the [110] direction. Steeds. when the elastic deformation causes no strain in directions except for the [ijk] direction. 1991a) of orientation when the stress is applied in a direction perpendicular to the (0 0 01) plane. Therefore.. but in TiAl3. Orientation dependence of 5.

the tangential constant F6 is very large and dependent on the orientation of the dislocation lines. Here. Figure 7 shows the elastic energy factor K of a dislocation with a Burgers vector of \ [3 3 1] or I [001] for MoSi2. 1968. T h e r a d i a l a n d angular coordinates represent S66 a n d orientation. a pair of dislocations lie on a (1 1 1) plane. (b) ( 1 1 2 0) a n d (1 1 0 0). K equals G and G/(I . but the tangential force constant Fd(iso) does not exist. is approximately equal to that for an isotropic one. which is measured from the direction of the Burgers vector.2615 nm (Villars and Calvert. the radial force constant /v(iso) between a pair of dislocations is finite. However. Steeds.v) for screw and edge dislocations. Here.3 GPa. Shear plane: (a) (0 0 0 1). and the orientation of the dislocation lines varies from the direction of the Burgers vector. 1987a). (c) (1 1 2 1). This may correspond to the fact that dislocations with a Burgers vector of -3 [00 1] could not be observed for deformed MoSi2 single crystals (Kimura et al.3 GPa 1 ) S 6 6 (IO 3 GPa-I) S 66 (IO. respectively. when both dislocation lines lie in the same direction. The dislocation line with a Burgers vector of \ [001] has a higher energy factor than that with a Burgers vector of \ [3 3 1]. For an anisotropic material. the elastic energy of the former dislocation is higher than that of the latter.1 ) Figure 5. Figure 8 shows the interaction force between parallel dislocation lines with Burgers vector of \ [1 1 0] for Ni3Al (Yoo. the orientation dependence of the radial force constant F1. For an isotropic material. and K is the energy factor. 1991a) where b is the magnitude of the Burgers vector. Orientation dependence of t h e compliance S66 for a Ti3Sn_ single crystal. a n d (d) ( 1 1 0 1) planes ( F r o m N a k a m u r a . K is calculated using elastic constants ctj (Hirth and Lothe.S 6 6 (IO 3 GPa-I) S 6 6 (IO. Since the magnitudes of the Burgers vectors are approximately equal (b = 0. 1990). For an anisotropic material. This large tangential force constant may affect dislocation motion. In an isotropic material. and cause . [T 10].. respectively. 1985) for the i [3 3 1] and i [ 0 0 1 ] Burgers vectors). although account was also taken of the antiphase boundary (APB) energy. 1973). the orientation of the dislocation lines varies from the direction of the Burgers vector.2616 and 0. A pair of dislocations represent a (111) APB-type dissociation.

9997 1. In the figure. (b) a.682 0.064 1.707 0.Table 4.584 0. where K1 is the stress intensity factor for mode I loading and r is the radial distance from the crack tip. especially for or and p.859 0. the crack tip. the elastic moduli of many polycrystalline intermetallics have been measured.093 0. on od.063 0...095 1. Elastic Moduli of Polycrystalline Intermetallics the anomalous temperature dependence of the yield strength in Ni3Al (Yoo.139 direction of crack extension is the [01 1 ] direction.247 1.999 0.866 0. This may affect the plastic deformation at the crack tip. 1991). especially for intermetallics that permit no deviation from stoichiometry ('line compounds').840 0.998 0. The contour lines are labeled in units of MPa (From Nakamura. Hence.707 0. In Ni3Al.2% flow stress with increasing temperature contrary to the case for ordinary alloys (Copley and Kear. 1987a).9993 0. The measurement of the ultrasonic wave (USW) velocity is mainly used for the determination of the elastic moduli of polycrystalline materials. Suzuki et al. T^9 and p in the 0 direction. distance ratio.510 0.927 0.766 0.895 0. 1991a) ..656 0. od.707 0. and hence the fracture toughness. Thornton etal.965 0. 1989). and p) is in units of K1Z(IiCr)1'2. Comparison between interatomic-distance ratio and elastic-constant ratio (Nakamura.988 0. 1970.861 0.224 0.718 0. cP2 structure) (Yoo and Fu. The AT3-axis perpendicular to the figure represents a dislocation line.672 0. i.572 s wk] s[OOl] 1. 5. there is an increase in the 0. Examples of elastic fields around a screw dislocation line with Burgers vector \ [3 3 1] for MoSi 2 : (a) a13. The x r axis in the In general. represents the relative magnitude of the stresses.974 0. and that of the shear stress (r^) is in units of 2K1Z(I-KrY2. The magnitude of the normal stress (<jr. 1967.866 1.973 0. and the x2-axis is perpendicular to the (10 0) crack plane. and the jc2-axis is normal to the (10 3) plane.718 0. 1968).353 0. c ioo n Intermetallics [ij k] [ij A:]/[001] cmj WSi2 WSi2 MoSi2 MoSi2 TiAl3 TiAl3 TiAl3 ZrAl 3 ZrAl 3 ZrAl3 Ni3Al Ni3Fe Ni3Ge NiAl CuZn [100] [331] [100] [3 3 1] [10 0] [2 0 1 ] [110] [10 0] [40 1] [110] [110] [110] [110] [111] [111] 1.224 0.. Figure 9 shows the stress distribution at a crack tip of (1 0 0) type under mode I loading for a NiAl single crystal (B2. 1991a) Crystal Interatomicorientation.e.856 0.234 1. The stress distribution calculated from the anisotropic elastic constants is different from that calculated from isotropic elastic constants.633 0.313 1.829 1. a single crystal is more difficult to prepare than polycrystalline material.082 1.028 0.273 1. the radial distance from the origin. while the resonance frequency for different vibrational modes is used for the determination of elastic moduli at elevated Figure 6. The stress distribution at the crack tip in a single crystal is also affected by anisotropy of the elastic constants (Sih and Liebowitz. although it is hard to prepare samples in single-phase condition.671 0.

Poisson's ratio of many intermetallics is lower than that of metals and ordinary alloys. 1963): • For Voigt's approximation Ky = Ucn + C22 + C33) + 1 (c12 + C13 + C23) • For Reuss's approximation l/KR = (sn+S22 + S33) + 2(Sn + Sl3+S23) (12) (11) Fn F9(GPa) Orientation (deg) Figure 8. this lattice structure per se does not necessarily result in a low Poisson's ratio of 0. Interaction force between parallel dislocation lines with Burgers vector \ [110] for Ni3Al. etc. the bulk modulus K of a polycrystalline material is estimated from the following equations (Anderson. For all crystal structures. Fleischer. because the coupling materials between a specimen and the ultrasonic-wave transducer cannot usually be used at temperatures above 500-600 0 C. 1991b.29. The method in which the average value of the maximum and minimum values is adopted is called HiIFs approximation. and Be-base (Fleischer and Zabala. ZrAl3. 1988) alloys. 1989) and Ti-base (Fleischer et aL. YAl2. tI26 structure) exhibits a Poisson's ratio of 0.. The elastic properties of intermetallics may be mainly characterized by Poisson's ratio.25-0. The radial component Fn tangential component Fe.K factor (GPa) Orientation (Deg) Figure 7. First.099. The orientation of the T dislocation lines varies from the direction of the Burgers vector VSi2. the maximum and minimum values of the moduli are obtained using Voigt's and Reuss's approximations. Since Be12Nb (also with a D2 b . and the estimated values of Young's modulus are known to be in good agreement with the measured ones (Anderson. 1977).1) for Be12Ti (D2b.44 (Smithells. ZrAl2. Measurements of elastic moduli have been carried out for many polycrystalline intermetallic compounds by Fleischer and Field (1990). Ir-Nb (Fleischer et al. Ru-Al (Fleischer et aL. 1993a). 1963).y 1991a). in which it is about 0. respectively. 1987a) temperatures (Schafrik. 1976). The orientation of the dislocation lines varies from the direction of the Burgers vector. [110] (Reproduced by permission of Pergamon Press from Yoo. Elastic energy factor A of a dislocation with Burgers vector j [3 3 1 ] or \ [0 0 1 ] for MoSi2. It is low for . Fleischer and co-workers have investigated systematically the effects of composition and alloying elements on the elastic moduli at room temperature for Ru-Ta (Fleischer et al. tI26 structure) and Be13Zr (D23. cF112 structure). 1991. where r is the distance between dislocations and b is the magnitude of the Burgers vector. and remarkably low (<0. The elastic moduli of a polycrystalline material can be approximately estimated from the elastic constants of the single crystal. Table 5 shows the elastic moduli of various intermetallic compounds. and radial component for an isotropic material Fr (iso) are normalized by (2wrb2).91990).

Stress field at a crack tip of (1 0 0) type under mode I loading (tension across the crack) for a NiAl single crystal: (a) radial stress. Young's modulus E and Poisson's ratio v are given by E=9KGZ(3K+G) (17) The shear modulus G is also estimated from the following equations: • For Voigt's approximation Gv = B(ClI+C 2 2 + C23-lV(q2 + Cl3 + C23)) + 1(^44+ C 55 + C66) ( M ) and p = {(3K-2G)Z(3K+G) (18) The elastic-wave velocity in a polycrystalline material is given by the following equations. (c) positive pressure. and (d) shear component. (Units: K1Z(ITTr)* for ar> ae. and 2K1Z(IiTr)* for r^. i. 1991) • For Hill's approximation KR = i(Kw + KR) (13) Then. the crack tip. ar. For the longitudinal wave. an oe. TF9. r^.lsotropic Anisotropic Model crack Figure 9. The average wave velocity vm is given by the following equation: . p0 is the density of the material. represents the relative magnitude of the stresses. and/?. Vx= [(K + 4GZI)ZP0]* (19) • For Reuss's approximation 15/GR = 4(sn +522 + 5-33) -4(s 1 2 + S13+ S23) (15) + 3(S44+ S55+ S66) • For Hill's approximation GH = {(GW + GR) (16) and for the transverse wave.e. /?. (b) tangential stress.) The radial distance from the origin. ae. and p in the 6 direction (Reproduced by permission of Pergamon Press from Yoo and Fu. vs = (GZp0)* (20) Here.

14 0.58 2.21 1.18 0. Intermetallics K (102 GPa) NbIr Ru11Ta9 Ru49Ta51 TiAl VSi2 Al8V5 AlRu CoAl CoAl CoAl FeAl RuSc RuTi RuTi ZrAl3 Be13Zr Be12Nb Be12Ti YAl 2 Cr 2 Nb TbFe 2 ErFe 2 YFe2 ZrAl 2 TiCr2 V3Si Cr3Si Ti3Sn Ti3Sn SbY 3. cF24 C15.31 0.224 1. (1990b) Fleischer et al.17 1.829 1. hP12 C14. (1989) Clark et al.05 2.73 1. Elastic moduli of polycrystalline compounds estimated from elastic constants of single crystals using Hill's approximation Intermetallics TiAl3 ZrAl 3 MoSi2 WSi2 VSi2 CrSi2 Ti3Sn TiSi2 CoSi2 NiAl Ni3Al Ni3Fe Bulk modulus.149 0.31 0. (1973) Clark et al.730 1. (1989) Fleischer et al. cP2 B2.975 1.55 1.71 0.773 0.78 2. hP8 B l . v 0. cF24 C15.214 0. 0D (K) 681 577 759 625 791 793 335 726 528 500 478 469 .59 1.851 1.819 1. (1991b) Fleischer and Zabala (1990b) Harmouche and Wolfenden (1986) Fleischer (1993b) Harmouche and Wolfenden (1986) Fleischer and Zabala (1990b) Fleischer and Zabala (1990b) Fleischer et al.94 1.778 1. 1963): where h is Planck's constant. tP4 C l l b . tI16 D2 3 . (1991b) Fleischer et al.86 1.720 0.655 0.167 0.315 0.353 0.352 0.158 0. cP2 B2.033 0.35 0.06 0.55 2.056 1.66 1.67 2.167 0.019 2.700 0.806 Young's modulus. (1991a) Fleischer et al.51 2. G (102 GPa) 0.98 2. cF24 C15.17 0. cF24 C14. K (102 GPa) 1. hP8 DO19.160 0.61 1.672 1. cF8 References Fleischer et al.099 0. E (102 GPa) 2. tP4 Ll 0 .100 2.04 Poisson's ratio.018 4.660 1.841 2.42 0.426 3.660 1.341 0. Table 6.28 0.00 0.101 1.80 1.Table 5. (1990) Fleischer et al. cP8 A15.900 1.82 2.22 1. G (102 GPa) 0.84 2.07 Young's modulus. tP4 Ll 0 . cP2 B2.613 1. (1988) Fleischer and Zabala (1990b) Fleischer and Zabala (1990b) Fleischer and Zabala (1990a) Fleischer et al.546 1. Elastic moduli of polycrystalline intermetallic compounds at room temperature Bulk modulus.31 1.157 2.55 2.305 0.205 0.29 0.245 'Pearson (1958). cF112 D2 b .827 0.296 Debye temperature. Villars and Calvert (1985).07 1.13 3.156 0. tP4 Ll 0 . E (102 GPa) 2.036 1. and q is the number of atoms in the molecule. cP2 B2.097 2. cF24 C15.20 2.01 2.11 2. cl52 B2.75 0.584 1.533 0. cP8 DO19.151 0.55 2.31 0.47 0.37 0. (1988) Fleischer and Zabala (1990b) 0. ^n=[-3(2/^+l/^r1/3 (21) 6D = (h/k) [i {q/<K){Npo/M) ] l/3vm (22) The Debye temperature 0D of a material can be calculated by the following equation (Anderson.419 1.189 0. cP2 DO23.479 1. Mis molecular weight.892 2.904 0.286 1.59 2.27 2.62 1.462 0.73 3. (1988) Fleischer and Zabala (1990b) Fleischer and Zabala (1989) Fleischer and Zabala (1989) Fleischer and Zabala (1989) Fleischer and Zabala (1990b) Fleischer et al.489 2. k is Boltzmann's constant. (1973) Clark et al.700 1.89 2.906 0.03 0.234 0.298 0.214 Shear modulus.069 2.675 0. cP2 B2.267 0.48 2.42 2.326 0.43 1.318 0.341 0.855 Poisson's Ratio. tI6 D8 2 . tI26 D2 b .930 0.97 1.82 1.22 1. tI26 C15. Nis Avogadro's number.496 0. cP2 B2.90 1.01 Shear modulus.312 0.954 1. (1991a) Schafrik (1977) Fleischer et al.355 0.938 0.911 2. V Structure" Ll 0 .184 0.78 2.679 3.27 0.397 4.185 0.141 0.405 0.93 1. hP12 A15.286 0. (1973) Fleischer and Zabala.68 2. cP2 B2.993 0.75 2.

^-(C11-C12)(IO2GPa) C44(IO2GPa) Al (at. intermetallics are not necessarily ideally ordered. Cu-Al. offstoichiometry. Cu-Sn (Cabarat etal. and Al (Smithells. 1986). which have an extended solubility range for Co and Al (Harmouche and Wolfenden.. The elastic moduli exhibit maximum or minimum values at the composition of Fe3Al with a DO3 (cF16) structure. Al-rich. 1967) alloys. ifci 1-^12) an<* cUi for Fe-Al alloys (Leamy et al. the Fe-group aluminides exhibit both a predominant defect type (vacancies. Figure 10 shows the composition dependence of the elastic moduli.%) Figure 10. Characteristics of Elastic Moduli 6.. even at the stoichiometric composition. indicative of a change of defect type.. Furthermore. DO3. 1986) Young's modulus (GPa) . the Young's modulus and the shear modulus of which are also higher than those of the constituent elements.b). 1976). (Reproduced by permission of Pergamon Press) 6. Note that the Young's modulus and the shear modulus for TiAl3 (DO22. 1967). The estimated elastic moduli of VSi2 and ZrAl3 are approximately equal to the experimental values (see Table 5). This may be because. In polycrystalline AlRu (B2. The maximum value of Young's modulus is found not at the stoichiometric composition but at a deviation from it. Zr. because while we have structural alloys with high specific strength like maraging steels. etc. The maximum value of Young's modulus was observed at the composition of the intermetallic compound or ordered phase like CuZn. 7000 series aluminum alloys. tI8 structure) and ZrAl3 (DO23. The high elastic moduli of intermetallic compounds may be useful for structural elements of machines. substitutions.1 Composition Dependence of Elastic Moduli It is well known that elastic moduli are dependent on alloy composition.9 1948a. tI6 structure). because Fleischer (1993) has found that the approximately stoichiometric composition of CoAl intermetallics exhibits the maximum Young's and shear moduli and that the density-composition curve has a break at stoichiometry. tI16 structure) are more than twice those for the constituent elements. MgPb. Ti. etc. etc. the Figure 11.Table 6 shows examples of the elastic moduli of polycrystalline intermetallic compounds estimated from the elastic constants of single crystals using HiIFs approximation. Composition dependence of Young's modulus for CoAl at various temperatures (Reproduced by permission of Elsevier Sequoia from Harmouche and Wolfenden. although those values are half those for MoSi2 and WSi2 (Cllb. Figure 11 shows the composition dependence of Young's modulus for CoAl alloys. cF16) Fe3Al is represented by the shaded band. Composition dependence of room-temperature elastic moduli for Fe-Al alloys (from Leamy et aL. and several investigations have been carried out over a wide range of compositions for Cu-Zn. the deviation of the maximum modulus from stoichiometry may be an artifact of inexact chemical analysis. CuAl. The composition range of single-phase (DO3. cP2 structure). On the other hand. almost all ordinary structural alloys exhibit similar specific Young's and shear moduli. 1967). Fe-group-rieh) and minor types (the reverse or so-called triple defects). and Mg-Sn (Guillet and Le Roux.

Effect of alloying elements on elastic moduli Bulk modulus.29 0.198 0.03 1. This behavior of the shear modulus.291 0.52 2.278 0. cP2 B2. 1967).. cP2 B2. G (102 GPa) 1.13 1. cP2 B2. tP4 Ll 0 .47 2.605 Poisson's ratio.04 1..10 0. and an increase in Poisson's ratio.30 2. is for a mixture of B2 and A3.312 0.16 1.225 0.14 1.43 1.80 2.93 2.5 Al 5 0 Co 4 5 Mn 5 Al50Co485Re15 Al 5 0 Co 4 5 Re 5 A1 48.31 0. substitute primarily for Al in Ni3Al (Guard and Westbrook.58 1.85 1.05 0. The elastic moduli of RuTi-. 1993a. The latter also induces an increase in Poisson's ratio.shear modulus at room temperature increases linearly with increase of the Ru content from 47 to 60 at. cP2 B2.2 Temperature Dependence of Elastic Moduli In general.55 2.07 1.887 0.t 1992).532 0. tP4 References Fleischer et al.29 2.68 1.42 2. Si. (1991a) "Pearson (1958).639 0. C44 is approximately independent of temperature. In a Laves-phase HfCo 2 single crystal (Cl5. cP2 Ll 0 .66 2.68 2. V.98 1.72 2. Figure 12 shows the temperature dependence of the stiffness constant C44 for an Fe3Si single crystal (Rausch and Kayser. 1977).230 0. 1969.97 1.53 2.38 1. cP2 B2.338 0. Table 7. cP2 B2.906 0. cause a decrease in the Young's and shear moduli for CoAl. Villars and Calvert (1985). The effect of the alloying element Ti on the elastic constants has been studied for Ni3Al (Yasuda et al.806 0.904 0.219 0. Table 7 shows examples of the effect of alloying elements on the elastic moduli for RuTi-. cP2 B2.727 1.385 0. single-crystal Fe3Al (Leamy et al. respectively.29 1. 1988. 1991).21 2. cP2 B2. and that of Al by Sc.10 0. tP4 Ll 0 .63 1.5C°50Ti1. Fleischer.5 Al45Co50Ti5 Ru11Ta9 Ru49Ta51 Co15Ru35Ta50 Co20Ru30Ta50 Fe 75 Ru 425 Ta 50 Fe15Ru35Ta50 0.47 2.91 2.19 2.289 0.62 1. A similar temperature dependence of the elastic moduli is observed for polycrystalline Zr3Al (Schulson. 1991.98 0. 1991 a.78 2. RuAl-. The latter also results in an increase of Poisson's ratio. 1967). The substitutions of Ru by Re or Co. and above 77 K it decreases almost linearly with increasing temperature. At temperatures below 77 K. compared with the values in Table 1.237 0.216 0. cP4 structure) are given by C11 = 238 GPa. 1959). These elastic constants are a little higher than those of Ni3Al.50 2. The substitution of Ru by Co or Fe also causes a decrease in the Young's and shear moduli.286 0. cP2 B2. 6. .68 2.78 2. tP4 structure) compounds have also been studied using polycrystalline alloys (Fleischers/ al. K (102 GPa) 2. Cheng. and RuTa-base intermetallic compounds.372 0. result in a decrease of the Young's and shear moduli for RuTi and RuAl.28 0. cP2 B2..14 2. Alloying elements like Ti. and C12= 141 GPa.75 1. the elastic moduli decrease gradually with increasing temperature for intermetallic compounds without order-disorder transition or phase transformation. C44 = 123 GPa.74 Shear modulus. and that of Al by Ti. cF24 structure).55 2.29 0.b). single-crystal LaAl2 (Schiltz and Smith. etc.214 0.09 1.341 0.50 2. and then decrease linearly with increasing temperature (Shannette and Smith. tP4 Ll 0 .341 0. (1988) 50 C °48. cP2 structure) and RuTa-base (Ll 0 . V Intermetallics RuTi Ti50Ru35Re15 Ti 50 Ru 20 Re 30 Ti 50 Ru 40 Co 10 AlRu Al43Ru52Sc5 Al40Ru50Sc10 Al25Ru50Sc25 AlCo Al Crystal structure" B2.374 Fleischer et al.62 1.76 2.30 Young's modulus. (1991b) Fleischer (1993b) Fleischer et al. E (102 GPa) 2. RuAl-.04 1. 1974). tP4 Ll 0 . and others (Ono and Stern. cP2 B2. 1980). tP4 Ll 0 .69 1. the elastic constants are constant up to about 150K.5 M n i.65 2. The substitutions of Co by Mn or Re.07 2. cP2 B2. 1969).97 0.10 1. Vo (Fleischer. cP2 B2.91 0.b. however. cP2 B2. The substitution of Ti on the Al site hardly causes an increase of the elastic constants. CoAl-. cP2 B2.06 1.67 2. and CoAl-base (B2. The elastic constants of a (nickel-rich) Ni76(Al15 Ti9) single crystal (Ll 2 .

shear modulus G. and . 1977). 1977). 1940) .S 1 2 for a Cu3Au single crystal (Reproduced by permission of the American Institute of Physics from Siegel. The elastic moduli decrease linearly with increasing temperature. A discontinuity in the modulus versus temperature curves Elastic moduli.000 00677 The elastic moduli are measured at temperatures above 500-600 0 C by the sonic resonant frequency for different vibrational modes using long cylindrical rods of circular cross-section (Schafrik. cP4 structure) (Siegel. Figure 13 shows the temperature dependence of the compliance constants S11. S44. Temperature dependence of the stiffness C44 for a Fe3Si single crystal (Reproduced by permission of the American Institute of Physics from Rausch and Kayser. and -S 1 2 for Cu3Au (Ll 2 . 1977) v = 0. S44. an anomalous temperature dependence of the elastic moduli is observed. and Poisson's ratio v are given at temperature T ( 0 C) by E= 173. and Young's modulus E. (10"2GPa-1) Temperature (0C) Figure 13. tP4 structure) (Schafrik.0141 T (25-935 0 C) (25-935 0 C) (25-847 0 C) (23) (24) (25) Temperature (K) Figure 12.03427 G = 70.0.9 .234 + 0.C44(IO2GPa) temperature dependence of the elastic moduli was measured from 25 to 940 0 C for polycrystalline TiAl (Ll 0 .59-0. 1940). Temperature dependence of the compliances su. The When alloys exhibit an order-disorder transition or phase transformation with a rise of temperature.

and rare earth-iron Laves-phase (Ho. 191 A) Laves-phase compounds. PrCo2 (C15. (Klimker et aL. Temperature dependence of the stiffness C44 for (a) LaAl2 and (b) GdAl2 single crystals. etc. A similar temperature dependence of the elastic constants is reported in CuZn (B2. in a review.e.2 to 300K for various polycrystalline rare earth-cobalt (Klimker et aL. etc. Figure 15 shows the temperature dependence of Young's modulus for the rare earth-cobalt Laves-phase compounds (Klimker et aL. 1974).. Rosen et aL. rare earth-iron (Klimker et aL. 1973) exhibit similar behaviors in the temperature dependence of the elastic moduli. CeFe2 and HoFe2 (C15.e. That is. a lattice softening accompanying the magnetic ordering and/or spin reorientation process. and has higher C44 values in the magnetically ordered phase than the extrapolation from the disordered phase. above which the compliance constants exhibit high values. The temperature dependence of the elastic moduli has been investigated from 4.5 0 C. etc. etc. 1973). for which the Curie temperatures are below room temperature. the magnetic ordering temperature) to room temperature. 1973). i. the elastic moduli are also dependent on magnetic field (Klimker et aL. 1974). GdAl2. (Rosen et aL. Elastic constant (10 GPa) Temperature (K) Figure 14. It has been noted (Figure 16) that. cF24 structure). 1979). showed that for intermetallic compounds like CuZn. YFe2. 1979). For the rare earthiron compounds RFe2. the elastic moduli decrease monotonically with increasing temperature as do those for LaAl2 (Klimker et aL. This is because Cu3Au has an order-disorder transition at 387. The broken curve represents an extrapolation of the behavior of the magnetically disordered phase (Reproduced by permission of the American Institute of Physics from Schiltz and Smith. for LaAl2. cF24 structure). 1979. GdFe2 (C 15. exhibits a change in slope of the C44 versus temperature curve near 170 K. cF24 structure) (Schiltz and Smith. The moduli increase rapidly with increasing magnetic field. exhibit anomalies in the elastic moduli in the temperature range from 227 K (i. and from 14 to 90 K. the order-disorder phase transformation is often accompanied by abrupt variations in Young's modulus. above which it becomes disordered. 1979). 1974) Temperature (K) Figure 15. followed by saturation. 1979) Young's modulus (10 GPa) . In comparison. cF24 structure). The intermetallic compounds that have a magnetic transition also exhibit an anomalous temperature dependence of the elastic moduli. Temperature dependence of Young's modulus for rare earth-cobalt Laves compounds (Reproduced by permission of Pergamon Press from Klimker et aL. C44 increases continuously with decreasing temperature. for the rare earth-iron or rare earth-cobalt Laves-phase compounds in the magnetically ordered state.. observed at 387. and this behavior is indicative of a lattice softening process preceding the magnetic phase change (Klimker et aL. like DyFe2. which has a magnetic transition from a disordered to an ordered phase with decreasing temperature. 1979). cP2 structure) by Young and Bienenstock (1971).5 0 C.Tb)Fe2 (Rosen et aL. respectively. however. Cu3Au. like NdCo2. The rare earthcobalt Laves-phase compounds RCo2. which has no magnetic ordering in this temperature range. Figure 14 shows the temperature dependence of the elastic constant C44 for Laves-phase LaAl2 and GdAl2 (Cl5. Guillet and Le Roux (1967). the elastic moduli decrease with decreasing temperature in the noted temperature range.

. Figure 17 shows the temperature dependence of the elastic constants for a Nb3Sn (A15. Chang and Barsch. EH and E0 are Young's modulus at magnetic field H and O.. Magnetic-field dependence of the modulus change (EN-E0)ZE0 for RCo 2 compounds at 77 K. 1967). 1967) ^ (102GPa) C^(IO2GPa) Hydrostatic pressure (108 Pa) Figure 18. cP8 structure) (Keller and Hank. 1980) and V3Si (A15. 1979) Contrary to the normal behavior in metals. Effect of hydrostatic pressure on the elastic constants C11. Nb3Sn exhibits a lattice softening as temperature is reduced. cP8 structure) (Testardi et al. Thus. 1980). A similar temperature dependence of the elastic constants is observed for V3Si (Al5. 1967). That is. In this compound the shear modulus (C11-C12)/2 also decreases with decreasing temperature. Below 32 K. respectively (Reproduced by permission of Pergamon Press from Klimker etal. one shear wave cannot be propagated below the temperature at which a crystallographic phase transformation may take place (Keller and Hank. the elastic constants C11 and C44 decrease and the shear modulus i (C11-C12) vanishes with decreasing temperature in Nb3Sn (A15. and vanishes near the critical temperature. 1965). respectively.. Temperature (K) Figure 17. cP8 structure) single crystal (Keller and Hank. Tc. 1981) -1 (C11-C12) (GPa) C44 (GPa) . C11 becomes equal to C12. These compounds have high superconducting transition temperatures. Temperature dependence of the elastic constants for a Nb3Sn single crystal (Reproduced by permission of the American Institute of Physics from Keller and Hank. of 17 and 18 K.Modulus change (%) Modulus change (%) Magnetic field. H (kOe) Figure 16. C44. 17 K. 1967. cP8 structure) single crystals (Chang and Barsch. and \ (C11-C12) for a HgTe single crystal (Reproduced by permission of Taylor & Francis from Miller et al. because a shear wave with the modulus (c n -c 1 2 )/2 cannnot be propagated.

the elastic moduli at room temperature have been related to melting temperatures for various intermetallic compounds (Fleischer. below which it has no phase transformation. For Nb3Sn (Al5. and (C11 . 1991. 1991) 6. C11. The values of C44 and the shear modulus 4(C11-C12) decrease linearly with increasing pressure. 6. 1988). Relationship between the shear modulus and the melting temperature (Reproduced by permission of the Japan Institute of Metals from Fleischer.55 GPa. cP8 structure) single crystals.4 Relation Between the Elastic Moduli and Melting Temperature The elastic moduli are considered to be related to the interatomic forces in inter metallic compounds. 1991). the pressure dependence of the elastic constants has been investigated from about 13 to 300 K (Chang and Barsch. 1980). Figure 19 shows the shear modulus G at room temperature versus melting temperature for polycrystalline intermetallics (Fleischer.9 1981). The correlation presented in Figure 19 may indicate that . Thus.Shear modulus (GPa) Melting temperature (K) Figure 19. but C11 goes through a maximum at about 1.33 GPa. This indicates that the probability of a high shear modulus increases remarkably with melting temperature. for single-crystal HgTe (Miller et al. Skinner and Zedalis. C44. Figure 18 shows the effect of hydrostatic pressure on the elastic constants. cF8 structure) has been investigated under hydrostatic pressure up to 1.c12)/2.3 Pressure Dependence of the Elastic Moduli HgTe (B3.

R. L. J. elastic properties of dislocations. R. (1991b).. RuSc. J. S.. Izumi). p. L. Clark. 215.. 8. R. 453. F. L. R. Appl.. 24A. (1990). 114. Metall. R. Res. L. Such information is also necessary to estimate or check the potential energy between different atoms. (1967). H. 909. (1987). and Zabala.. 1279. A. Rev... / . (1990a). 21A. J. etc. The elastic moduli of polycrystalline materials have also been summarized. H.. Fleischer. 971. and Barsch.. and Weizer. R. Cabarat... V. M. L. Taub.. etc. R. 239. Rev. A. (1988) / .. L. etc. C. Trans. B43. Fleischer. (1989).. 2276. Materials Research Society. Garber. the year in which the first edition of Intermetallic Compounds. Trans. In the present chapter. 20. In Proc. Field. L. 807. Fleischer. and the elastically anisotropic behaviors of single crystals have been discussed. Butler.. (1920). / . Solids. J. Trans.. fill. 2964. (1989).. Fleischer. L. Trans. Trans.. Metall. Fleischer. S.. The elastic properties of dislocations as calculated from the anisotropic elasticity have been shown. (1990). 163. and Strakana. Metall. 37. A. L. Gilmore. R. 24.. Soc. C. (1967). C. and Le Roux... S. Guenin. Jr (1969). Solids... 681. (1991). and Field. R. F12. 129. on the elastic moduli were described. H. B. Appl. O. Voter. Fleischer. R... 3990. M. Phys. K. O. Metall. In Intermetallic Compounds (ed. R. D. and Westbrook. W. O. and Palmer. (1986). R. 2801. M. J. Phys. J... D. R. J. 649.. Fleischer. L. S. R. 413. 38.. E. Fleischer. Belson. S. Res. and Kear. J. 2813. and Zabala.. 2913. L... on Intermetallic Compounds (JIMIS-6) (ed. Trans. 5. R. 977. L.. L. R. Wiley. V. I. 5. which is used for computer simulations like molecular dynamics. L. S.. and Zabala. ScL Paris. 7.. M. Phys. H. A. R. C . D. and Srolovitz. (1963). etc. Phys. 2.. Field. R. P. 1374. R. Guillet. although the elastic constants themselves can be calculated using various methods like EAM. Ignat. D. and Methfessel. (1989). C. and Zabala.. (1993a). Trans.. was published. the effects of various factors like composition. Paris. temperature. (1948b). N. 22A. Fleischer. A. Z. Dickson. R. Girifalco. S. and Wachtman. Phys.. Appl. A. 68. Acknowledgement The author would like to thank Dr R.. Guillet. etc. Japan Institute of Metals.. . Rev... Pittsburgh. R. 81. 2OA. J. Fleischer. Denike. Mater. / . J. L.. J. H. A. B. E. (1991a). Liu. (1973). Copley. R. Fleischer. Metall. S. and Zabala. Gilmore.. P. L. J. / . Soc. Symp. L. J. 40. and Le Roux. the elastic constants of single-crystal intermetallic compounds have been summarized. References Alouani. G. WRDC-TR-90-4046. L. R. In High-Temperature Ordered Intermetallic Alloys III (eds C. Intl. Field. Metall. of intermetallic compounds. (1991). Phys. AlIl. fall significantly below the scatter band in Figure 19. Phil. The temperature dependence of the elastic moduli may also be further used for the investigation of various kinds of transformations. (1975). and Zabala. Fleischer. Grivord. and Briant. 3242. and Neumann.. C. etc. Metall. F-LAPW. 1389. R. 687. Phys. 44. R. Mater. 6515. Phys. G. D. Trans. 49.. Appl. Knowledge of the elastic constants of a single crystal is necessary to calculate or estimate various elastic properties like elastic anisotropy. Acad. and Daw. Report. Scripta Metall. The elastic moduli of SbY. Sci. (1990). Westbrook). Cheng. Phys. R. B. A. Chem.. p. J.. and Briant. 2709. (1967). J. p. 9. A. 1951. Ada Metall... New York. / . 227. Res. C. Chang. BIl. Guard. and C. (1982). 21A.the melting temperature can be used to predict approximately the elastic moduli of intermetallic compounds. 6500. R. AIME. and Le Roux.. 8. Fu. R. T. L. Foiles. temperature. 21A. Sendai. 22A. Trans. P. L. Guillet. Fu. Y. J. 226.. Albers. Koch). 64. L. 28. Stoloff. Griffith. R.. and no explanation for these anomalous elastic moduli has been given.. Mater.. and Granato. R. (1980). L. R. AIME. D. C. Anderson.. H. R.. W. Finally. and Yoo. Chen. (1959). Summary and Conclusion The elastic properties of many intermetallic compounds have been investigated extensively since 1967.. M. (1959). 157. 403. C. since they are sensitive to composition. Himmel. (1948a). Phys. J.. Trans. L. Grivord. (1967). Chang. Fleischer for contributing many references on the elastic moduli of intermetallic compounds. R.. S. 3063. (199Ob)Me^r//.. R. and Thomas.. R. J. G. B. (1990). R. 227. (1993b). R. edited by Westbrook. Appl. Acad.. M. Phys. Cabarat.. M. Phys. K. Chem. Rev.

Oxford. Westbrook. 195. Nakamura. R.. and Klimker.. D. Chem. 37. Yoo.. and Liebowitz. M. H. M.. C.. 15. H. 52. 45. Appl. Rosen. A. Takasugi. M. Acta Metall.. /. Hickernell. 40. U. Osburn. W. Papaconstantopoulos.. Solids. Schreiber. Rev. Phys. C. and Calvert. S. (1972). R. J. p. Saunders. F. and Lothe. Jr. Stoloff.. (1940). M. and Fu. J. Japan Inst. K. T. and Danielson. Izumi). E. Proc. p. Intl. Nabarro. London. 2326. M.. Trans. High Temp. S. 256. L. W. H. 537. L... (1974).. Dariel. Chem. D. New York. Scripta Metall... (1982). Metall. Phys. Barbara. Rev. Sumer. U.. J. 48. E. Proc... M. Mater. E. S. W. Phys. 171. Villars. /. and Klein..-Q... 363. 4. This chapter was originally published in 1995 as Chapter 37 in Intermetallic Compounds.. Phys. 4049. (1970). Theory of Dislocations. 35. (1990). Schulson. N. 49. Prevender. (1992).. Appl. O. New York. D. and Smith. Phys. M. 42. and Chirba. (1988). 655. American Society for Metals. Introduction to Anisotropic Elasticity Theory of Dislocations.. L.. M. Schafrik. R. and Guntherodt. Symp. Klimker. Lett. W. Rev. (1969) /.) (1967). C. L. Metals Park.. H. J. X. (1971). ScL.. G.. Mater. H. Zirinsky. J. On Intermetallic Compounds (JIMIS-6) (ed. 26.... Westbrook and R. P.. Phys. M. 1. and Hank. 25.. R. Rayl. Shannette. Cohen. 26. M. (1966).. Steeds. S. TMSAIME. 2208. G. W. Podloucky. J. Anderson. and Koiwa. Rosen. 48.. Rev. D. Vol. H. R.Guo. Nakamura.. (1991) Mater. Sendai. J. (1979). T. J. (1973). (1968). London. (1944). Pergamon. L. F. Nakamura. Yasuda. 2487. R. R. W. BlO. p. 40. McGraw-Hill. 154. Phys.. Scripta Metall. Phys. 52. M. Davies. Atzmony. 466. (1974). F.. and Smith. A Handbook of Lattice Spacings and Structures of Metals and Alloys.. Wiley. O. J. G.. In Proc. Vol. Klimker. 1. X. T. Soc. T. and Johnston. T. 57.. C. New York. Young. A Treatise on the Mathematical Theory of Elasticity. Phys. 4680. A. (1947). M.. Appl. 15. F.. Phys. S. B. F. Phys. Res. Pearson. B41. H.. (1991). J. L. and Stern.. J. (1977). Phys. (1991). Appl.. Whitten. Love. (1966). (1971). andMiura. J. and Koiwa. Phys. A. 35. Metals.. and LaPlaca.. Rehwald. Nakamura. Phys. J. OH. Jr. New York. . Vol. R. Sih. Japan Institute of Metals.t 40. (1968). (1991). ScL Eng. and Koiwa. A. Peierls. 34. T. Gibson. H. A. 975. /. and Kimura. Ada Metall. Trans. Y. J. (1969). G.. 84. Phys. T. A43. (1973) Elastic Constants and Their Measurement. Keller. D. 25. King.. Butterworths. 1559. 207. 42. II (ed. K. 31. Yasuda. (1990). J. H. 1827. (1962). In High-Temperature Ordered Intermetallic Alloys II (eds N.. Rev. and Groves. W. and Bienenstock.) (1986). (1987b). 250. Phys. B. (1994)... M. Mehl. G. 4th Edn.. H. 1447. Takasugi. M. J. Al.1. J. J. (1990). M. Met. Yoo. p. Skinner. (1940). (1981). and Wolfenden. 2345. (1981) Phil. Schiltz. M. T.. 33. M.. Rausch. C17. M.. and Hirano. Kayser. J. E. H.. Clarendon. /. H. 8A. Kimura.. (1989)ISIJIntl. J. Wasilewski. Metals Park. M. and Gayton. J. in press. J. H.. (ed. Ada Metall. T. Materials Research Society. and Yogurtcu. Massalski. and Zedalis. (1973). McGraw-Hill. Solids. B. A. R. R. B.. Smithells. Matsumoto. status solidi. 3008. and O. OH. J. /. American Society for Metals. V. Miller. Reed. (1984). (1965). P. and Kayser. Appl. Jr. 5635. Appl. 324. R. 59. M. P. T. 22. and Rosen. L. 184. J.. 381. Yoo. Klimker. Metals. K. E. 164. 455. G.. 1003. In Fracture-An Advanced Treatise. M. 1199. P. R... P. 44. Thornton. (ed. Penney.. K. H.. edited by J.. 3309. p.. L. J. Trans. Prog.. Weil. Koch. Soc. O.. (1980). (1969). Dariel. /. Dover. /. B8. G. Bateman. 1: Principles. Schiltz. T. K. Intermetallic Compounds. Bull. Mater. Mock.. P. H. 1783. 207. 79. Y. and Atzmony. and Freeman. Mater. F. (1967).. (1977). and Dariel. Nadgornyi. J. M. E. Mag. Metals Reference Book. 2968. 5th Edn.. A64. Rev. F. (1958). S. 4. H. B8. and Soga. Mater... Rev. Phys. Hirth. Rev. C. A. Harmouche.. J. Binary Alloy Phase Diagrams. Phys. and Stassis. J. (1971). Trans. 32. 30. W. B. T. E. Pittsburgh. Chem. R. 335. X.. F. ScL. S. 100. J. Chung. M. 462. E. Appl. 195. Suzuki. W. Phys. 487.. C . p. Academic Press. (1991).. X. and Cody. Solid State Commun. 29. C. Mater. Metall. 628.) (1976). (1987a). (1986). Liu. and Smith. R. 723. B. 112. (1967). (1988). Liebowitz). Japan Inst. Appl. Plaskett. Rosen. P.. Phys. Testardi. Nakamura. P... M.. Trans. (1965). TMSAIME. 2283. Phys. Yasuda. Phys. R. Acta Metall. Phys. C. M. 1. (1991b)..t 25. 404. (1991a)... Sci. and Kayser. H. and Smith. (1956). Mater. R... A. Leamy. Siegel.. (1985) Pearson's Handbook of Crystallographic Data for Intermetallic Phases. 245. and Hirano. 6.. (ed.. Trans. R. Fleischer. 4681. Izumi). Solids. S. Nakamura. ScL.. F. G. J. Rev... Mishima. 236. New York.. A. Ono. 68. S. 10311.

762 1. tP4 Ll 0 .318 3. Table 10 shows the elastic moduli of polycrystalline NbCr2 and Hf25V60Nb15 (Cl5.625 Poisson's ratio V Debye temperature eD(K) 751 688 552 587 584 475 210 406 372 0.883 0.2798 0.880 0.037 3. hP6 References Tanaka et al (1997) Ctoiet al (1996) Chu et al (1996) Tanaka et al (1996a) He et al (1995) Tanaka et al (1996b) Eshelman and Smith (1975) Eshelman and Smith (1978) Butler et al (1980) 4.01 0. Villars and Calvert (1985) 6 RUS method 'pulse echo technique Table 9.857 0. TiAl single crystals have been prepared. Table 9 shows the elastic moduli of polycrystalline compounds estimated from the elastic constants in Table 8 using the Hill approximation. hP9 Ll 0 .. 1996a.612 0. not at the stoichiometric composition.74 0.220 0.283 0. tI6 C40.15(8) C\\ C 11 C 11 C1I C11 Cn C 11 C 11 c \\ 5.847 1.757 0. 1995).556 1.638 Shear modulus G (102 GPa) 1..510 0.5477 1.78 1.628 3. hP9 C40.390 0.744 0.189 0.748 0.157 1. Elastic constants (102 GPa) of intermetallic compounds at room temperature Intermetallics MoSi/ NbSi24 TaSi2" TiAF TiAY Ti3AY Zr 2 Ni' Al2Cu^ CeNi/ C11 C33 C44 C55 C44 C44 C44 C44 C44 C44 C44 C44 C44 C 66 C12 c!3 0.282 0.054 4.184 0.437 1.05 1. but at an Al-rich composition.624 0.802 3.948 r 6 6b C C h C66 0. Elastic moduli of polycrystalline compounds estimated from the elastic constants in Table 8 using the Hill approximation Intermetallics MoSi2 NbSi2 TaSi2 TiAl TiAl Ti3Al Zr2Ni Al2Cu CeNi5 Bulk modulus ^ (10 2 GPa) 2.532 1.86 1. elastic constants of single crystals have been measured for intermetallic compounds with tetragonal structures as well as cubic structures. 1990). tI12 D2 d .767 1.64(3) 1.83 1.753 1.680 4. Table 8.7966 2. of single crystals at room temperature for several intermetallic compounds.157 0.196 0.4480 1.44 0.925 1.045 1. Although TiAl is obtained over a large range of chemical composition from about 49 to 56 at% at room temperature (Massalski et al.098 1.7023 2.145 0.Addendum Al. Elastic Constants at Room Temperature Recently.42 0.829 1.759 0.0966 0. elastic constants of single crystals as well as polycrystalline materials have been determined for many intermetallic compounds using a resonance ultrasound spectroscopy (RUS) method (Maynard. cF24) measured by a RUS method at room temperature (Chu et al. 1996).2399 0.72 0.757 0.866 1. 1995).563 0.741 0.078 Cl 3 Cl 3 Cl 3 Cl 3 C13 C13 C13 C13 C13 "Pearson (1958).65(9) 2.76 2.34 0.22 0.100 1.590 1. Table 8 shows the elastic constants.453 1.099 1.915 1.784 0.784 0.. Thus a single crystal of Ti-56 at% Al was used for measurement of elastic constants (Tanaka et al.878 1. Since this method does not require large single crystals for measuring elastic constants.086 1. hP8 C16.7270 1.11 1. c^. He et aL.309 . TiAl-based intermetallic alloys are considered promising light-weight materials for high temperature structural applications.027 1. tP4 DO19.948 C23 Structure 0 C l l b .187 1. 1994.901 0.433 Young's modulus E (10 2 GPa) 4.77 c c 66 h 0. tI12 C16.4464 C 6 66 b 1.

1994). A3. decrease. 1994).759 1. with increasing temperature from 4 to 1373 K.Table 10.9 2.759 0. F-LAPW (First-principles full-potential linearized augmented-plane-wave total-energy method) ''theoretical.862 1.34 0.996 1.. tI6).98 0.77 0. C44 and C66 decrease from about 410.796 0. which is usually used to connect a sample with an ultrasonic wave transducer when using conventional ultrasonic methods. compared with those at room temperature and theoretical values.42 to 0. 1997). cF24) increase linearly from about 26 and 75GPa to 30 and 85GPa. With increasing temperature from 7OK to room temperature the shear and Young's moduli of polycrystalline Hf25V60Nb15 (Cl5. The temperature dependence of the elastic moduli for this intermetallic compound are remarkably different from those of the elastic constants for usual intermetallics like TiAl. hP8) single crystals in a temperature range from 4 K to room temperature (Tanaka et aL.612 °C66 = ^extrapolated value 'theoretical. (1991) Tanaka et al. (1996b) 3.812 0.87 1.294 1. 436.762 4.053 1. S. 206 and 200 GPa to 352.00 0. tP4 DO19. tP4 Ll 0 .90 0.037 1. cF24) has no phase transformation at low temperatures and that the anomalous temperature dependence can be understood from the electronic structure (Chu et aL..187 1. LAPW (linear augmented-plane-wave method) A2.461 Young's modulus Shear modulus £ (10 2 GPa) G(XQ2 GPa) 2.624 1.96 0. cF16) single crystal in Figure 12.880 0.141 0.405.20 1. F. B.744 0. which were estimated from the elastic constants of single crystals. Phys. L743.3007 Poisson's ratio V Structure References NbCr 2 Hf25V60Nb15 0. C: Solid St. Chu et aL reported that Hf25V60Nb15 (Cl5.875 0. MoSi2. Fe3Si.50 1.09 1. Elastic moduli of polycrystalline compounds measured by a RUS method Intermetallics Bulk modulus A: (10 2 GPa) 2. hP9 C l l b .453 2.468 1. Elastic constants (102 GPa) of intermetallic compounds at different temperatures Intermetallics NbSi2 MoSi2 TiAl TiAl' TiAl'' Ti3Al Temperature 120K 300K OK* 290K OK* 298 K theoretical theoretical 3.82 1. 174 and 163 respectively.26 0.88 1. 1996a. J. Table 11 shows the elastic constants of some intermetallic compounds at low temperatures.78 1.802 4.149 1.883 0.027 1. c n .890 0. That is. B. and Palmer.748 0. They decrease linearly with increasing temperature above room temperature.786 0.8442 0.73 4.948 0. Temperature Dependence of Elastic Constants Temperature dependence of elastic constants can also be measured using a RUS method. shear and Young's moduli of polycrystalline MoSi2 (Cll b . tP4) and Ti3Al (DO19.84 3. cF24 C15.90 1.680 5.83 1. (see Figure 12). (1980). D. References Butler.878 0.251 2. 1997).05 0. cF24 Chu et al.85 1.404 Cl5.442 0. The temperature dependence of the elastic constants was measured for a MoSi2 (Cll b . Givord. where the values at 0 K were extrapolated from the temperature dependence of the elastic constants.741 1. (1994) Table 11. because the method does not require glue. . hP8 References Chu et al. 13.641 0. the bulk. 1996b). Its bulk modulus is hardly dependent on temperature (Chu et aL.3 K 270K (CW-C\l)l2 c u C 33 C44 C l2 C I3 Structure C40. and its Poisson's ratio decreases linearly from about 0. (1997) Tanaka et al (1996a) Fu and Yoo (1990) Mehl et al.626 0. C12 and C13 hardly depend on temperature in this temperature range. 514.05 1. and Poisson's ratio increases with increasing temperature from 4 to 1373 K (Tanaka et aL.100 4. tI6) single crystal in a temperature range of 4 to 1373 K by a RUS method (Tanaka et aL. respectively. (1996) Tanaka et al. tI6 Ll 0 . although they are almost saturated at low temperatures as shown for a Fe3Si (DO3. C33.145 0. etc..748 0.. Phys. (1995) Chu etal.140 5.471 0. Givord. tP4 Ll 0 . A similar tendency was observed for TiAl (Ll0.065 2.

D. F. M. J. Tanaka. (1996a).. Osburn. Phil.. T. (1996). 867.. Chen. p. Petrovic. (1975).Chu. 49-1. Mehl. Inui. Appl. J.. 186. He. In Alloy Phase Stability and Design (eds G.. F.. J. (1996). Vol. D. Lett. 33. E. M.. 1295. Symp. J. T. R. A. and Mitchel. M. 10... 44. Binary Alloy Phase Diagrams... and Koiwa.. Tanaka. K. 73. F. and Koiwa. F. Maloy. Okamoto.. E. Res. J. Stock. T. Mater. (ed. H. and Smith.. 71. Mag. A. and Yoo.. M. Inui. 46.. Lett. T. Yamaguchi. Ada Mater. He. A. 1. M. Phil. F. F. and Migliori. F. 2nd edition. 3284. A. M. . Mag. Massalski. and Koiwa.. (1995).. Physics Today. (1978). 225. Appl.. J. J... C.. 26. D.. B. Soc. (1995). Metals Park. L. Mater. Eshelman. Y. S. Giamei). F. Res. Mater.. and Mitchell. 73. Chu. H. H. M. P. Mater. M. Phys. Yamaguchi. H. A. Lei. E. J. (1996b). (1997). 49. American Society for Metals. 5080. J. Pittsburgh. Onome. Lei. P. M. E. R. Phys. Migliori. 3035. Maynard.. Vol. J. A239-240. Eshelman. Shwarz. Mag.. and Klein. Mater. M.. 70. Thoma. 62. T. ScL Eng. Pope and A. Phil.. M. S. M. Tanaka.. 159.. p. Yamaguchi. B. 1475. (1994). H. Inui. Y. J. 1187. Fu. and Smith. Phil. Minonishi. Y. Chu. 277. Scripta Metall. Ohio. B. K. Res. Papaconstantopoulos.. K.) (1990).. Mag. and Mitchell.. (1991). (1990). Proc. Soc.. B. K... Ichitsubo. 188.

These compounds are often very lightweight. A review of the literature shows that the details of the fracture process and the attempts to improve ductility vary from system to system. H. Brown University. or intergranular. Rhode Island 02912-9104. but their brittleness prevents their use. 2000 John Wiley & Sons Ltd . New York 12301. and therefore should be attractive for many engineering applications. i. L.e. To bring all of this information together in a comprehensive way. Edited by J. We will first summarize the research that has been done on compounds with either the Ll 2 (cP4) structure or the B2 (cP2) structure. Box D. Examples of the two brittle fracture modes found in intermetallic compounds: (a) cleavage in NiAl. we have organized this chapter as follows. fracture along specific crystallographic planes in either a single-crystal or polycrystalline sample. Providence. and that there are many individual reports on ductility and fracture in specific compounds. as far as fracture and ductility are concerned. USA) 1. fracture along the grain boundaries in a polycrystalline sample.Chapter 2 Intergranular and Cleavage Fracture Clyde L. Fleischer ©1995. PO Box 8. Examples are shown in Figure 1. an additive that will segregate to the grain boundaries and increase cohesion across them. Introduction One of the most limiting properties of many intermetallic compounds is their brittleness at room temperature. (b) intergranular fracture in Ni3Ge Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. These efforts have focused primarily on trying to find alloying additions that will impart ductility to the matrix or. Research and Development Center. Briant General Electric Company. These are emphasized because. i. Significant research efforts in recent years have been directed at trying to ductilize these compounds so that they can be used. they have been investigated the most Figure 1. USA (now at Division of Engineering. remain strong at high temperatures. Westbrook and R. The fracture mode for these brittle compounds is either cleavage. specifically for compounds that fail inter granular Iy. 182 Hope St. Schenectady.e.

These common items are used in that chapter to discuss grainboundary chemistry and structure. . Mulford and Pope. A stress concentration builds up at the grain boundary from dislocation pile-ups and nucleates an intergranular crack. the crack propagates along them and failure occurs before yielding across the grain boundary takes place (Khadkikar et al. unlike the situation in many disordered alloys (Briant. Within this class Ni3Al has received by far the greatest amount of attention. 1967. These will be considered in the fourth section of this chapter. Ll 2 compounds have a crystal structure that is analogous to the face-centered cubic structure of disordered alloys. Al3Sc and Si3U fail by cleavage and all others fail by intergranular fracture. based on slip requirements. We will consider only single-phase alloys for which the fracture mode was determined in either a roomtemperature tensile. They also have a sufficient number of independent slip systems for polycrystals of these compounds to be ductile. 1988). It is important to note that. The fracture in polycrystalline samples occurs along the grain boundaries. We note that several of these topics are also discussed in Chapter 24 by Takasugi in this volume.. Figure 2. We will first examine the fracture modes in a number of Ll 2 compounds. by pointing out areas where further research would be valuable. 1987). We will then summarize the results for compounds from other crystallographic systems.. and by discussing general guidelines that should be followed in investigating fracture modes in these compounds. 1965). the results listed in Table 1 show that most binary polycrystals are brittle. A review of this work is best divided up into the general areas that reflect the types of research that have been carried out.) The few compounds that are ductile are usually composed of two transition metals. (There are other Ll 2 compounds that fail by cleavage and that are formed by making significant alloying additions to nonLl2-based compounds.thoroughly. We will almost exclusively present results on polycrystals. Among the brittle binary compounds. one must explain why so many of them are brittle. research has shown that intergranular fracture in these compounds occurs in the absence of any segregated embrittling impurity such as sulfur or phosphorus (Takasugi et al. We will then report which of those compounds that normally fail in a brittle fashion have been shown to gain ductility through alloying additions. we will look more closely at the improvement of ductility in Ni3Al caused by boron additions and discuss the possible mechanisms by which it occurs. Single crystals of Ni3Al are ductile (Copley and Kear. Aoki and Izumi (1979) reported that boron additions made polycrystalline Ni3Al ductile. as shown in Figure 2(a). since most studies in the literature are on samples of this type. and work in many laboratories began on this topic. bending. or notch impact test. Ll 2 (cP4) Compounds In recent years the causes of brittleness and the characterization of brittle fracture have been most extensively studied in the Ll 2 class of intermetallic compounds. However. These summaries will be followed with a description of the mechanisms that have been proposed to explain the brittleness of these compounds. Finally. This finding raised the possibility that a number of brittle intermetallic compounds might be made ductile through additions of boron or some other element. (a) The Ll 2 (cP4) crystal structure and (b) the B2 (cP2) crystal structure 2. They have the stoichiometry A3B with the A atoms located on the cube faces and the B atoms at the corners. 1973). Given that these compounds should be ductile. but polycrystals are very brittle (Davies and Stoloff. To make these summaries we have also had to define certain limits for what we will discuss. It is easy to explain how intergranular fracture might occur. We end the chapter by listing the general conclusions that we can make. If the grain boundaries are weak. whereas here they are used to discuss intergranular fracture.

and that elements must be present at a level greater than approximately OA0Io to be detected. Ni3Al and Ni3Ga in the as-cast condition are reportedly brittle over their entire stoichiometric range and the fracture is intergranular (Takasugi et al. TD.%) Figure 3. 1987). Huang et al. 1965). (1990) C George et al. Micrographs show that the structure almost always contains some second phases or casting voids (George et al. To our knowledge. 1987) IG. (1987) IG Briant et al. the test environment has been found to affect ductility and fracture mode. (1963) TD Takasugi and Izumi (1985a) IG -I. 1987). 1991). no research has been directed at identifying these nucleation sites in either Al 3 Sc or Si 3 U. However. It seems likely that these second phases might provide nucleation sites for cleavage in this compound. 1985. This interpretation is . Al3Sc is a line compound. Takasugi et al. it usually nucleates at a specific microstructural constituent (McMahon and Cohen.. The listings in Table 1 are for compounds that have compositions very close to the stoichiometric ratio A3B. Figure 3 shows that alloys containing 20% Ti fail with an elongation of 58% whereas alloys with 23% Ti fail with less than 20% elongation. the effect of stoichiometry on fracture mode has not been investigated in other binary Ll 2 compounds. The cracks form within this constituent at a low stress and then propagate through the matrix. The total elongation of Co3Ti plotted as a function of titanium content. all compositions will be given in atomic percent. but with these reservations.e.TD Takasugi and Izumi (1985a). 1990.. Figure 4 compares tensile elongation data for samples of Co3Ti tested in air and vacuum. The exact composition can vary over some range for most of these compounds without causing a change in crystal structure. which employed Auger electron spectroscopy to measure grain-boundary segregation.. When this fracture mode is observed in disordered compounds.) The fracture mode changes from predominantly ductile.. These changes in stoichiometry have been shown to affect the mechanical properties of many intermetallic compounds (Wood and Westbrook. It has been assumed that this environmental degradation in ductility is a form of hydrogen embrittlement. Co3Ti shows a very marked change in ductility as the stoichiometry changes (Takasugi et al. Fracture mode of Ll 2 binary compounds Compound Ni3Al Ni3Si Ni3Ga Ni3Ge Fe3Ga Zr3Al Cu 3 Pd Ni3Fe Cu3Au Ni3Mn Co3Ti Pt 3 Ga Al3Sc Si3U fl Ductile (D) or Fracture brittle (B) mode* References B B B B B D D D D D B B B Elongation (%) B Takasugi and Izumi (1985a) IG Takasugi and Izumi (1985a) IG Takasugi and Izumi (1985a) IG Takasugi and Izumi (1985a) IG Takasugi and Izumi (1985a) IG Schulson (1984) TD Takasugi and Izumi (1985a) TD Takasugi and Izumi (1985a) TD Vidoz et al. Liu et al. (1990. 1985. Data were taken from tensile tests run in the air (From Takasugi et al. i. In addition to changes in stoichiometry. cleavage. transgranular fracture in samples containing 20 and 21% Ti to predominantly intergranular fracture in samples containing 23% Ti. it cannot exist outside perfect stoichiometry. unless specifically stated.. In contrast. In all cases the ductilities measured in the tests run in air are lower than those measured in vacuum tests. one can conclude that the grain boundaries in these compounds are intrinsically weak. 1991) C Fleischer (personal communication) IG Ti concentration (at. and the fracture surfaces of the samples tested in air contained more intergranular fracture. C. (Note that. transgranular ductile. These studies. intergranular. 1962). Cleavage fracture requires more consideration. One must recall that Auger electron spectroscopy does not detect hydrogen.Table 1. 1985).. so one must ask whether the fracture mode and ductility of these compounds will also vary with composition. failed to reveal any elements present on the boundaries except the constituent elements of the compound.

1987b). Titanium additions have been shown to improve ductility in Ni3Si (Takasugi et al. Titanium additions as great as 9% provide little improvement in ductility as long as the nickel concentration is below 78%. 1990). An extreme example of this effect is the pest phenomenon described by Westbrook and Wood (1964) in which oxygen apparently penetrates the grain boundaries during the heat treatment and causes the sample to turn into a powder. These elements are usually added at a concentration well below 1% and often as low as 0. (1988b). This figure Elongation (%) . taken from Masahashi et al. Table 2 summarizes the studies that have examined the effects of additives on ductility. We now turn our attention to the work that has sought to find additives that will ductilize brittle. a hallmark of hydrogen embrittlement. The other group of additives we will term trace-element additions.. These additives can be grouped into two categories.05 to 0. One we will term major alloying additions. The effect of manganese on the ductility of Ni3Al is shown in Figure 5(a). 1993) have recently reported that samples of this composition prepared by cold-forging and recrystallization gave elongations of approximately 2. The final variable that can affect ductility is the processing history of the sample.. although the fracture remained completely intergranular. 1988b). In addition.8%. In the category of alloying additions. For example. the environmental constituent responsible for degrading ductility must do so by affecting the nucleation of a crack. Takasugi et al. as shown in Figure 5(b). or aid slip nucleation at the grain boundaries. so we will not dwell on it here. as shown in Figure 6. (1992. However. that these alloying additions primarily make the matrix more ductile. Ni76Al24 has been found to be brittle in as-cast samples (Takasugi et al. Once this crack has formed. so that yielding across a grain boundary will occur before a crack can nucleate. In all cases where the fracture mode was reported. One general point to note is that an element that is beneficial for one compound may not be beneficial for another. the elongations in air were only 3.. The concentration of the additive is usually greater than 1% and it is generally presumed. 1987) reasonable. an improvement in ductility is measured. 1991).5%. However. (1987) showed that variation in stoichiometry also affects the elongation in Nix(Al052Mn04S)10O-X alloys. and samples prepared by recrystallizing a worked single crystal gave elongations of 11 to 13%. The fracture mode for these Ni-rich samples was transgranular and ductile. these compounds have to be made with the nickel concentration above 75%. and other forms of environmental embrittlement. They reported room-temperature tensile elongations approaching 20% in the Ni-rich samples. Masahashi et al. stable crack growth (Briant. and in samples prepared by rapid solidification (Taub and Briant.1 to 4. Takasugi et al. 1985). These data.^Mn^.Ti concentration (at.. it was intergranular and appeared to be brittle. these latter alloys contained small amounts of zirconium. in samples prepared from powder consolidation and subsequent extrusion (Schulson et al. The total elongation of Co3Ti plotted as a function of titanium content for tensile tests run in air and in vacuum (Data taken from Takasugi et al. polycrystalline Ll 2 compounds. since these alloys tend to be susceptible to hydrogen embrittlement. The effect of iron on Ni3Al ductility has been studied much less than that of manganese. To obtain single-phase alloys. These data were obtained from tensile tests run in room-temperature air. and hydrogen is the one environmental constituent that could easily move into the lattice at room temperature. is the occurrence of slow. though not proven. Another processing variable that can affect elongation is the environment used for the heat treatment. One can observe that manganese additions cause an improvement in ductility. George et al.1 %. In this way they apparently improve cohesion at the grain boundaries and prevent crack formation from occurring before general yielding takes place. Therefore. 1985b.. it propagates rapidly and independently of the environment.. relative to pure Ni3Al. If the nickel concentration is allowed to increase to 80%. 1987. are for the stoichiometric alloy Ni75Al25 . When the zirconium was removed. In these tensile tests the fracture occurs quite rapidly. manganese and iron have been found to improve the ductility of Ni3Al (Takasugi and Izumi.%) Figure 4. These elements have been shown to segregate to the grain boundaries. 1987). the fourth column lists the elements that were found not to improve ductility.

1985). The general results from a number of studies of boron in Ni3Al are the following: • Boron additions can increase the ductility of polycrystalline Ni 3 Al from essentially zero to approximately 50%. 1988a. 1985). Takasugi Taub et al. boron additions do not produce ductility improvements (Taub and Briant... The amount of solid-solution hardening produced by boron is consistent with it occupying an interstitial site in the matrix. Liu etal. Masahashi et al. that has received the most attention by researchers.. (1988a) Cr. As the amount of aluminum decreases below 25%. • Boron additions increase the hardness and yield strength of the compound (Huang et al. as shown in Figure 8. 1984. It is also assumed that one reason why improvements in ductility begin to decrease when boron is added beyond some optimal concentration . 1985. (1985). Si. we will begin our discussion by describing this effect in some detail and then showing how results for other compounds fit into this description.Table 2. • Boron segregates to the grain boundaries in Ni3Al. • There is an optimal boron concentration for producing the ductility improvements in the Ni-rich compounds (Taub et al. Ti.. V. Takasugi et al. 1985. • The improvement produced by boron is very dependent on alloy stoichiometry. the segregation is required to produce the beneficial effect (Liu et al. palladium and zirconium have a small but beneficial effect in Ni 3 Al. (1987) also shows that the tensile elongation improves when the alloys are tested in vacuum. Taub and Briant (1987a) Taub and Briant (1987a) Ti C Masahashi et al. (1990) Takasugi et al. Ga. 1984). Liu et al. This effect can be observed from the data shown in Figure 7. A much greater effort has been devoted to studying the effect of trace-element additions on ductility in Ll 2 compounds. (1984). The effect of boron and other elements on the ductility of Ll 2 compounds Does boron improve ductility Compound Ni3Al Yes/No Yes Reference Aoki and Izumi (1979).. Liu et al.. 1987b.. (1987). provided that all boron is in solid solution... (1988a) Other elements that improve ductility Element Fe Mn References Elements that do not improve ductility Element References Be Pd Zr Ni3Ga Ni3Si Yes Yes Takasugi et al. This increase in ductility is accompanied by a change in fracture mode from intergranular fracture to ductile. (1985b) (1988b) Ge. the ductility improvements produced by boron additions increase. Masahashi et al. transgranular fracture (Aoki and Izumi. Therefore. Huang et al. Cu. But it has been boron. Masahashi et al. At aluminum concentrations of 25% and above.. Briant and Taub. (1986) (1985b). It is supposed that this segregation improves cohesion at the grain boundaries. Sc Chiba et al. 1988. (1991) Aoki and Izumi (1979) Takasugi et al. which significantly improves ductility in Ni3Al and Ni3Ga and provides smaller improvements in Ni3Si and Co3Ti. 1985. C Takasugi and Izumi (1988a). Takasugi and Izumi Masahashi et al. Takasugi efar/. Sb. The amount of segregation increases as the bulk boron concentration in the alloy increases. 1984). The results reported in Table 2 show that carbon improves ductility in Ni3Si and that beryllium. Liu et al. (1987). Huang (1985b) Takasugi and Izumi et al. as shown in Figure 7. (1990) Taub and Briant (1990a) Ni3Ge Pt3Ga Co3Ti No No Yes Taub and Briant (1987a) Briant et al. 1979. Taub et al. Since a large fraction of this work has been devoted to the study of boron in Ni3Al. 1987). Taub etal. (1986b) S et al. 1987. Takasugi et al. (1987). Co..

1990) Elongation(%) (AI. (a) The total elongation of Ni3(Al...%) Figure 5. The data used to make this point are shown in Figure 10(a).Mn) concentration (at.%) Ni concentration (at. In this figure the plastic strain to failure is plotted as a function of the boron grain-boundary concentration.. 1985).. (b) The total elongation of Nix(Al0 52 Mn 0M ) m _ x plotted as a function of the (Al. 1988a. 1985) . The open circles represent data for Ni-rich alloys. it is clear that there can be a significant change in ductility without a significant change in boron concentration on the grain boundaries. The data were obtained from roomtemperature tensile tests run in air (From Liu et al. The data were taken from room-temperature tensile tests performed in air and in vacuum (From Takasugi et al. If one plots these data for boron grain-boundary composition as a function of ductility. 1985).Mn) content. 1988b). IG denotes intergranular and TG transgranular fracture (from Takasugi et al. This result is demonstrated by the data in Figure 9 (Masahashi et al. Briant and Taub (1988) also examined this question and used hydrogen charging to increase the amount of intergranular fracture in Ni-rich samples. We will now consider these mechanisms. • The test environment affects ductility... These results show that. Their results are shown in Figure 11. One of the initial theories used to explain why the boron effect was observed only in Ni-rich compositions was that boron segregation was increased in the Ni-rich compounds (Liu et al. 1992.. for the same grain-boundary concentration of boron.Mn) plotted as a function of manganese content.Elongation(%) Elongation(%) Mn concentration (at.%) Figure 7. and the full circles represent data for stoichiometric or Al-rich compositions. The data are taken from room-temperature tensile tests (From Masahashi et al. 1987) is because the matrix has become so hard that ductility is inherently limited (Liu et al. They have addressed the questions of how boron exerts its beneficial effect and why this effect is so dependent on alloy stoichiometry. George et al.. The effect of changes in nickel concentration on ductility in Ni3Si compounds containing 9% titanium. The effect of boron concentration on elongation of polycrystalline Ni3Al. 1993). as shown in Figure 10(b). a Ni-rich alloy Elongation(%) B concentration (at.%) Figure 6. These fundamental observations have led to much additional research on the mechanisms by which boron provides this ductility improvement.

The grain-boundary boron composition was obtained from Auger electron spectroscopy. The elongation data were obtained from room-temperature tensile tests run in air (From Liu et al. 1985) change could possibly explain why boron improves ductility. This result holds for both boron-containing and boron-free compounds. The bulk boron concentration was 0. The alloys contained 0. (1989) have examined this question. This result is corroborated by atom probe/field ion microscopy measurements of boron grain-boundary segregation.05 wt. The data were obtained from room-temperature tensile tests run in air (From Liu et ai.% boron. Two other questions that have been raised are whether or not the nickel and aluminum concentrations at the grain boundary are different from the bulk concentration and whether they are affected by the presence of boron. A comparison of the room-temperature elongation of Ni3Al obtained from tests run in air and in vacuum plotted as a function of the boron concentration (From Masahashi et al. These numbers are not atomic percents but. In the Ni-rich alloys the grain boundaries have less aluminum than the bulk. as the boron on the grain boundaries increases. The effect of aluminum concentration on elongation in Ni3Al. Therefore.%) Figure 9. George et al.%) B/Ni Auger peak-height ratio Figure 10. (a) The boron on the grain boundary plotted as a function of the aluminum concentration in Ni3Al. we conclude that changes in bulk alloy composition may affect the grain-boundary composition of the constituent elements in the compound. but that this change is not affected by boron segregation.%. 1991). and their results are summarized in Table 3. which show that boron segregation does not depend on stoichiometry (Brenner and Hua. this ratio should increase. then that B/Ni Auger peak-height ratio Elongation(%) Al concentration (at..Al concentration (at. (b) The boron on the grain boundary plotted as a function of the tensile elongation. and is expressed as a ratio of the boron 180 eV peak height to the nickel 102 eV peak height.05 wt..y 1985) has more ductility than a stoichiometric or Al-rich alloy. 1988a) Elongation (%) .%) Figure 8. but this change again appears to be independent of the presence of boron. If boron causes a significant change in the grain-boundary composition of the constituent elements. This result Elongation(%) B concentration (at. The grain boundaries and the bulk have essentially identical compositions in the stoichiometric alloy.

Although for a given sample there appears to be some correlation between low aluminum concentration and a low total number of X-ray counts taken in the analytical electron microscope. it has been thought that if the material near the grain boundaries became disordered in these compounds. This possibility can be tested by measuring the yield strength as a function of grain size..Table 3. and their results are plotted in Figure 12. 1989) Al on grainboundary fracture surface (0Io) 25. 1988) is not in agreement with results obtained by analytical electron microscopy reported by Baker et al.2 Al on transgranular fracture surface (0Io) 25. (1989).8 + 2.3 ±3 24. Schulson and Baker.2 24. The boron on the grain boundaries plotted as a function of tensile elongation. These results do not agree with the Auger data discussed above. The open circles represent data for alloys that were Ni-rich. The full circles represent data for alloys that were on stoichiometry or were Al-rich. The next possible mechanism that we wish to consider is based on the following idea. who suggested that. King and Yoo (1987) formally demonstrated that disorder would allow many additional dislocation reactions to occur at . the fact that the authors used the same counting time for each condition should eliminate any effect of total counts on measured composition.4 + 2. when dislocations pile up at a grain boundary. Schulson and co-workers (Schulson et al. It was noted above that in disordered alloys grain boundaries are usually strong and that an embrittling impurity is required to cause them to be weak fracture paths. Another possible mechanism that has been proposed to explain why boron improves ductility is that it enhances slip transmission across grain boundaries. for both the Ni-rich alloy and the stoichiometric alloy.2A1-O. The picture that is presented is that.24B Ni-24A1 Ni-24A1-O.3 21. these concentrations were obtained using an Auger sensitivity factor that was too low. 1990b). When their uncorrected data were compared with those obtained by George et al.6 25. 1986. and at present there is no explanation for this discrepancy. whereas in the Al-rich compound the slopes are identical.7±3. This stress relief could occur because the boron facilitates the transmission of the dislocations through the boundary or because it aids in nucleation of dislocations in the grain boundary. Briant and Taub (1988) reported very high aluminum concentrations for the grain boundaries in the alloys that they studied. but that in stoichiometric and Ni-rich alloys the presence of boron produced very low aluminum concentrations at the grain boundaries. 1991.4 25. Consequently. They found that the presence of boron did not affect composition in Al-rich alloys. the presence of boron lowers the slope of the line drawn through the data points.4 Alloy composition Plastic strain to failure Ni-25A1 Ni-25. (1988. However. The Auger peak height ratio used was B 180 eV/Ni 848 eV (From Briant and Taub. 1987) have pioneered this work on Ni3Al. An increase in the slope indicates that slip transmission across grain boundaries is more difficult. if boron lowers the difference between different ordering configurations at the boundary.7+1.1 ±2.9 + 3 22.. yielding will take place across the grain boundary and relieve stress before fracture occurs.24B B/Ni Auger peak-height ratio Figure 11. A comparison of the aluminum concentration on grain boundaries and bulk for Ni3Al (Data taken from George et aL. then it could aid in the accommodation of stress concentrations by the motion of grain-boundary dislocations. Weihs et al. Schulson has interpreted these results to mean that boron enhances slip transmission across grain boundaries in the stoichiometric and Ni-rich compounds. the results from both studies were found to be in excellent agreement (Taub and Briant. The results suggest that. 1990). This interpretation has been described on a more atomistic scale by Frost (1988).9 ±1. and that this improvement in slip transmission is the mechanism by which boron improves ductility. The slope of this plot is known as the HaIlPetch slope. 1991. an improvement in ductility would be realized.

Work by Baker et al. This effect may well be related to the changes that boron causes in chemical bonding at the boundary (Frost. Mills (1989) and Mills et al. One possible way for disorder to be created might be through boron segregation. since it is not a hard sphere. (1988. and have carefully compared experimental images with calculated ones. there has been a tendency Yield strength (MPa) Yield strength (MPa) Yield strength (MPa) . or by changing its own segregation behavior with changes in alloy stoichiometry. 1988). Although this technique is very powerful. or by affecting the aluminum or nickel concentration at the grain boundary. Whether or not disordering occurs at the grain boundaries and whether or not boron affects it have been hotly debated issues. these samples were taken from bi-crystals that were then mechanically tested so that one could be sure that the grain boundary that was examined in transmission electron microscopy was also the one that was tested for ductility. and they reported that in some cases the width of the region was as great as 20 nm. Mackenzie and Sass (1988) first reported that such disordered regions exist with an observation on a single grain boundary in Ni3Al. One can conclude from this study that a grain boundary that is completely ordered to within one atomic plane of the interface can also be ductile. there is evidence that it affects dislocation transmission within and across the grain boundary or dislocation generation at the boundary. primarily because this work must be done with high-resolution transmission electron microscopy. They also showed z-contrast images to support their interpretation. 1990) and Baker and Schulson (1989) also suggested that this disordered region existed. (b) stoichiometric alloys. (1991) have also examined this problem in some detail. it would appear that there is no convincing evidence that boron produces its beneficial effect either by causing disorder over several atomic layers at the grain boundary. As a result of the significant improvement in ductility that boron produces in Ni3Al. However. In conclusion. The room-temperature yield strength of Ni3Al plotted as a function of grain size for (a) Al-rich alloys. but they did not compare calculated and experimental images. With either of these mechanisms one must try to understand chemical bonding at the grain boundary and the effect of boron on it. The changes in chemical bonding that boron must make. These ideas will be discussed further in Section 5 of this chapter. might also simply improve cohesion across the grain boundary.Grain size (urrr0-8) Grain size (jinr0-8) Grain size (nrrr0-8) Figure 12. and (c) Ni-rich alloys (Data taken from Schulson and Baker. In all cases they found that order existed to within one atomic layer of the grain boundary. Furthermore. 1991) the grain boundaries and thus aid in slip transmission across grain boundaries and improve ductility. the chance of observing artifacts in the images is very great unless extreme care is taken with the experiment and the interpretation.

In addition to boron. boron also increases the strength of this compound significantly. 1987a). The increased strength counteracts any potential improvement produced by the boron segregation and makes the material more brittle. Sulfur is a strong embrittling element at grain boundaries in many metallic systems (MuIford. and in all cases. Boron segregates to the grain boundaries in NiAl and causes the fracture mode to become transgranular cleavage (George and Liu. and the fracture mode of carbon-doped samples remains intergranular.. It segregates to the grain boundaries in Ni3Al and lowers cohesion across the grain boundary. even in the presence of boron. Briant and Taub (1989) showed that ductility will be destroyed if boron is added in concentrations beyond its solubility limit. We mentioned above that boron can increase the hardness of these compounds and that.tensile elongation. and gallium provide some additional ductility to the single crystals (Darolia. The grain-boundary borides help in either the nucleation and/or the propagation of the intergranular crack. ductility will begin to decrease because of this increased strength. Figure 2(b) shows that this structure is analogous to the body-centered cubic structure of disordered alloys. 1991)..%) also improve ductility in Ni3Si (Taub and Briant. our discussion will center on them. it simply destroys in some way the improved bonding produced by boron. Boron appears to have the same effect in Ni3Ga as it does in Ni3Al (Takasugi et al. it segregates to the grain boundaries in both compounds. and the material shows no. Finally. Wasilewski et al. work was performed to determine if it would produce a similar improvement in NiAl. and that any ductility possessed by pure NiAl is lost.0 Composition Ni-50Al NiAl +300 wppm B NiAl + 300 wppm C NiAl + 500 wppm Be 'Vedula et al. Because boron counteracts grain-boundary brittleness in Ni3Al. little boron segregates and no improvement in ductility is observed.01% borides form along the grain boundaries (Taub and Briant.0° 0 0 3. the fracture is intergranular (Vedula et al. both on and off stoichiometry.. where one atom type is at the center of the unit cell and the other is at the eight corners. Small additions of boron (<0. The solubility of boron in Ni3Ge is so low that even at concentration of <0. 1987a). . its beneficial effect can also be counteracted. 1988). Within this class of compounds. However. Although boron can improve ductility in several Ll 2 compounds. 3. The results in Table 4 show that carbon also raises the yield strength of NiAl. These compounds have the stoichiometry AB. Therefore. molybdenum. At boron concentrations greater than this amount. B2 (cP2) Compounds Another class of intermetallic compounds that has received considerable attention are those with the crystal structure designated as B2.. 1987. In summary. 1990).9 1986b). Its solubility in both compounds is similar. Beryllium additions have no effect on the fracture of NiAl. 1983). rather. 1991. 2. and some of the early reported results Table 4. Therefore. Carbon does not segregate to the grain boundaries. borides precipitate along the grain boundaries. Deviations from stoichiometry will induce extreme brittleness. 1990) Yield Ultimate tensile Tensile strength strength elongation (MRa) (MPa) (0Io) 154 178 229 329 336 307 2. no additive elements have been found that improve ductility of polycrystalline NiAl. It does not appear to compete with boron for grainboundary sites (Briant and Taub. 1989. 1987a). no boron segregation is observed in Pt3Ga. Polycrystalline samples of NiAl show a very small amount of ductility (1-2%) if the composition is precisely on stoichiometry. The results have been mixed. Room-temperature tensile properties of NiAl with and without additions of trace elements (except where noted data taken from George and Liu. Finally. Recent work has also shown that small additions of iron. George and Liu (1990) have investigated the effects of carbon and beryllium on the ductility of NiAl. and there is no improvement in ductility (Briant et put boron in all brittle intermetallic compounds to determine if it will improve their ductility. 1990). 1967). if boron is added above some optimal level.2. The results are summarized in Table 4. (1989). George and Liu. but this value is very dependent on orientation (Darolia. These borides provide sites for crack nucleation and counteract the beneficial effect of the segregated boron. FeAl can exist over a relatively wide range of stoichiometry. NiAl and FeAl have been the most thoroughly studied. so that borides begin to precipitate along the grain boundaries. Taub and Briant. Single crystals of NiAl have been reported to exhibit elongations on the order of 1-2%. and as long as the compound is Ni-rich it improves ductility. 1990). Sulfur additions have also been shown to counteract the improvements produced by boron (Taub et al.1 at.

2 1. Liu and George (1991) examined the effect of stoichiometry on fracture mode and found that. and tests in dry oxygen gave the greatest amount of elongation. A boron-doped alloy containing 40% aluminum. did show Room-temperature elongation (%) Oil Vacuum Oxygen Environment Figure 13. Crimp and Vedula (1986) and Liu and George (1991) have also found that boron additions in the higheraluminum-content compounds can change the fracture mode from intergranular to cleavage. Room-temperature tensile properties for FeAl tested in air (except where noted data taken from Liu and George.7B fl IG. At the two aluminum concentrations that fractured by transgranular cleavage in air. They attributed the detrimental effect of testing in air and water vapor to hydrogen..3 0 4. "From Crimp and Vedula (1986). and the results are shown in Figure 13. cleavage. which failed by transgranular cleavage in air. and vacuum gave intermediate elongations. oil.5% Al alloy. as the elongation increased. 1991) Yield strength (MPa) 347 360 390 229 391 Tensile strength (MPa) 442 412 406 229 530 577 972 Elongation (0Io) 3. They are quite similar to those reported for NiAl.4 2. but they also cause a significant increase in yield strength and very little improvement in ductility. intergranular. Their results are summarized in Table 5. Liu and George (1991) continued this work and examined four different aluminum concentrations. the fracture mode changes from cleavage to intergranular. the effects were similar to those described above. 1989) . for the binary alloy seemed to be contradictory. (1989) have reported that the test environment plays a significant role in determining the ductility of FeAl.2 <0. The effect of environment on the elongation of Fe-36. C. Boron additions cause the fracture mode to change to cleavage. The abscissa gives the test environment and the ordinate the room-temperature elongation (Data taken from Liu et al. Their results are summarized in Table 6.Table 5.5A1 Fe-4OA1 Fe-43A1 Fe-48. for the samples that contained more aluminum and failed by intergranular fracture in air. the effect of test environment on elongation was much less. Recently. as the aluminum concentration increases from 30 to 50%. Their first report was for an Fe-36. However. Tensile elongation was the least when the material was tested in air or water vapor.5A1 Fe-4OA1+ 300 wppm B Fe-46A1 + O. These results for the boron-containing compounds are also summarized in Table 5. Tests in Ar+ H2. Liu et al. They also found that.3 0 Fracture mode" C C IG IG IG* C C* Composition Fe-35A1 Fe-36. the fracture mode changed from transgranular cleavage to intergranular fracture.5% Al alloy.

8 C C C a significant improvement in ductility when tested in oxygen. unlike the case with NiAl and FeAl. for compounds with the Al-rich compositions.2 2..) The main point to note is that.intergranular. Other Intermetallic Compounds In addition to the research on Ll 2 and B2 compounds described above.3 <0. 1991). 1991) Composition Fe-35A1 Fe-35A1 Fe-35A1 Fe-35A1 Fe-36.4 7.3 8. (Note that in the table a crystal structure that is given in parentheses is the structure of the base compound but alloying additions have caused a change in the structure to the one that is listed first. More detailed mechanistic ideas need to be worked out to explain these results.6 1. just as with the Ll 2 and B2 compounds.4 7. the aluminum concentration at the grain boundary is greater than its bulk composition.5A1 Fe-4OA1 Fe-4OA1 Fe-4OA1 Fe-43A1 Fe-43A1 Fe-43A1 Fe-4OA1 + 300 wppm B Fe-4OA1 + 300 wppm B Fe-4OA1 + 300 wppm B "IG.5A1 Fe-36. Furthermore. the improved ductility is associated with a transgranular fracture. and this improvement is greater for the Al-rich compositions of this compound.2 <0. when boron is present in the higheraluminum alloys. boron additions improve toughness. these compounds tend to be brittle at room temperature and the fracture is either cleavage or intergranular. cleavage. since once the crack is propagating it seems unlikely that the environmental constituent causing the loss in ductility could keep up with it and continue to cause its damaging effect.1 2. We have chosen not to discuss these alloys in depth because the research on them has not been as broad and because most of the important features of brittle fracture in these compounds have already been described by considering the other systems. ductility and fracture mode have been reported on compounds with other crystal structures. Test environment Air Air/oil coating Vacuum Oxygen Air Air/oil coating Vacuum Oxygen Air Air/oil coating Oxygen Air Air/oil coating Oxygen Air Air/oil coating Oxygen Yield strength (MPa) 347 344 341 345 360 356 351 360 390 402 402 229 313 327 391 398 392 Tensile strength (MPa) 442 570 558 785 412 524 496 805 406 466 537 229 313 327 577 751 923 Elongation № 3.6 5.3 <0.5 16. Attempts to alloy these compounds have in some cases produced significant .3 Fracture mode* C C C + IG C C + IG Mostly IG IG IG IG IG IG IG 4. Another alloy that has been investigated in some detail is AlRu (Fleischer et al. C. but.3 18.4 17. The binary compound is very brittle. the environment must affect crack nucleation. one must explain why this improvement in the lower-aluminum alloys is associated with a change from transgranular to intergranular fracture. Table 7 summarizes these studies. whereas in the higher-aluminum alloys the fracture is brittle intergranular in all environments and the ductility is unaffected by the environment. 4.Table 6. These results are certainly intriguing.2 5. The effect of test environment on room-temperature ductility of FeAl (data taken from Liu and George. but in contrast to the results on Fe-36. as opposed to intergranular fracture in the boron-free alloys. As mentioned above.5A1 Fe-36.4 3. but they are also puzzling.5A1 Fe-36.5% Al the fracture mode remained transgranular. Given that assumption. Auger electron spectroscopy has shown that.

Taub and Briant. (1991) Murata et al. Eberhart et al. 1991). at least in their polycrystalline. improvements in ductility (McKamey et al.0 0 0 0. and that this electronic charge transfer should lead to a directionality in the bonds. Intermetallic compounds. Examples of these impurities are sulfur. Messmer and Briant. Kear (1986).Pd. and that it might contribute electronic charge to the gain boundaries and thus enhance cohesion (Messmer and Briant. Mendiratta et al. and. Ductility and fracture mode data for various intermetallic compounds Compound Fe3Al Fe69Al31 Fe70Al28Cr2 ^e 6 6 8 5 Al 28 Cr 5 Zr 0 jB 005 TiAl Al66Ti23Fe6V5 Al3Ti + Ni. Briant. 1980. 1991. 1982). Also. Painter and Averill. because the bonding in these compounds should be more directional than in a disordered alloy. the composition of the boundaries can be identical to that of the bulk. It was also suggested that the beneficial effects of boron and changes in stoichiometry could be explained through these general ideas. intergranular. Models for Brittle Fracture In all of the results that were presented above. ^IG. one would have regions of electronic charge depletion at points where two electropositive elements attempted to form a bond across a boundary and regions of electronic charge excess at points where two electronegative elements tried to form a bond. Briant and Messmer. Although these models suggest that electronegativity differences and electronic charge transfer are at the heart of this embrittlement.Table 7. so we will begin our discussion with it. 1979. Both Izumi and Takasugi (1988) and Briant (1990) suggested that in ordered intermetallic compounds there is electronic charge transfer between the two elements making up the compound. Izumi and Takasugi. Existing tables for electronegativity would not be useful in predicting embrittlement. cleavage. and this bond weakening gives rise to the grainboundary fracture. 1989. Taub etaL. especially for transition metals.4 16. (1987) Mendiratta et al. Research addressing the question of why these impurities cause embrittlement suggests the following mechanism (Losch. but the fracture mode still appears to be brittle. alloying has caused a change to the structure listed first. The question that we now wish to consider is why brittle fracture modes are observed. Fracture will occur along them in the absence of impurities. or M n + V Al 3 Ti+ Cr.5 Fracture mode* C C C C C C C C C IG + C References McKamey et al. The actual tensile elongations associated with these fractures may be greater than 20%. 1982. 1990). tend to fail with a brittle fracture mode. as shown by the data in Table 3. one clear trend seems to stand out. As mentioned earlier in the text. (1991) George et al. 1988. C. (1987) McKamey et al.8 4 8 5.. the geometric distortions at the grain boundary would make the bonds between the electropositive and electronegative elements less favorably oriented. 1987a. (1989) McKamey et al. (1991) Schneibel and Porter (1989) A crystal structure in parentheses is that of the base compound. Calculations have also suggested that boron would not be electronegative with respect to nickel. Most of the work that has sought to explain intergranular fracture in intermetallic compounds has drawn on these models (Takasugi and Izumi. quantum-mechanical . at a grain boundary where the bonding pattern must be disrupted. tin. This electronic charge transfer weakens the metal-metal bonds at the grain boundaries. and antimony. (1990) George et al. Most of the modeling studies have dealt with intergranular fracture.4 >3. As mentioned above. Therefore. 1985b. 1984. However.6 9. phosphorus. (1991). The embrittling impurities are electronegative with respect to the host metal.Fe A-W5Fe55Zr25 a Crystal structure* DO3 DO3 DO3 DO3 Ll 0 Ll 2 (DO22) Ll 2 (DO22) Ll 2 (DO22) Ll 2 (DO23) Elongation 4. 5. the models have had to be adapted to the fact that intermetallic compounds have grain boundaries that are intrinsically weak. 1986a.. single-phase form. (1992) George et al. inter granular fracture in disordered alloys can almost always be attributed to the presence of an embrittling impurity that segregates to the grain boundaries and weakens them. it was pointed out in this work that these concepts were developed based on full-scale quantum-mechanical calculations. All of these factors should lower cohesion across the grain boundary. 1987). They draw electronic charge off the host metal and onto themselves.

in particular. They found that there was a directional component to the bonding with a build-up of electrons along the transition metal-aluminum axis. and results show that changes in stoichiometry that increase the transition-metal content of the alloy usually increase ductility. Instead. With this new information we should now rethink the issue of grain-boundary embrittlement. one could argue that the grain-boundary interstices might be depleted of electronic charge and that this depletion would tend to cause embrittlement. (1986. Schulz and Davenport (1992) have recently reported the results of quantum-mechanical calculations on FeAl. and calculations that specifically consider grain-boundary geometries are required. 1991) and Frost (1988). However. it could not be used to distinguish a more brittle compound from a less brittle one. there could be added repulsion because of the excess electronic charge on them. and that there was no correlation between what little charge transfer occurred and brittleness of the compound. probably by their movement through interstitial regions of the lattice. Ni3Mn. some of the concepts that emerged from their study help us to reconsider the problem of grain-boundary embrittlement. Schulz and Davenport (1992) found that there was little electronic charge transfer in these compounds from aluminum to any of the three transition metals. both the electropositive aluminum and the electronegative transition metal appeared to draw electronic charge out of the interstitial regions of the structure. It should be noted that these ideas are also completely consistent with boron allowing more movement of grain-boundary dislocations or making the generation of dislocations in the boundary occur more easily and thus aiding in the transfer of slip across the grain boundaries as proposed by Schulson et al. If ordering pulls electronic charge out of these interstices. Ga.calculations have shown that boron contributes electronic charge to the grain boundaries in nickel. where X = Al. Taub et al. they do hold out the only possibility for giving us insight into these problems. so their results would be specifically applicable to single crystals. and least in Ni3Ge. more ductility should be observed. The other factor that could play a role here is that the boundary. First. Transition metals. and the reason for this is quite clear. Among the three compounds that they investigated. 1991) found that electronegativity differences did not explain very well the brittleness differences that they observed. Given that both elements may tend to pull electronic charge out of the interstitial region. Cu3Au. Though they are timeconsuming and expensive.9 the materials became ductile. They focused on bulk properties rather than grain boundaries. Most of the ductile binary compounds in the Ll 2 series are those which are composed of two transition metals (Co3Ti. (1986a) and Taub and Briant (1987a) used the Pauling electronegativity scale to explain the effect of boron in Ni 3 X. They found that valence differences explained their results better. Their results suggested that at an average electronegativity value of 1. CoAl. However. can take on different valences and bonding characteristics as their environment is changed. Other schemes have also been proposed but exceptions can be found in all cases. less in Ni3Si. As noted above. or Ge. In regions where atoms are found to take up positions with less than the equilibrium distance. Takasugi and Izumi (1985b. All of this discussion has been based on the idea that boron would act primarily by improving cohesion across the grain boundary. The results of the calculations were very surprising. when they included compounds that were composed entirely of transition metals. so the idea of using a table of numbers that would work in all cases is unreasonable. to describe a complex phenomenon such as chemical bonding. because they give us general information about the electronic charge transfer between aluminum and the transition-metal atoms. and NiAl. In fact. The reason that metals are so ductile derives from their electronic configuration and the fact that electrons can be easily shared between atoms. as the difference in valence between the two elements decreased. However. one must resort to full-scale quantummechanical calculations. FeAl is the least brittle in its single-crystal form and CoAl is the most brittle. then it would decrease ductility. . Ni3Fe. They used linear augmented Slatertype orbitals and solved for the bulk electronic structure within the local-density-functional approximation. One cannot use the idea of electronic charge transfer as originally thought. Cu3Pd). Several attempts were made to quantify these arguments using various atomic properties. boron provides the greatest improvement in Ni3Al and Ni3Ga. One would expect little charge transfer between compounds made up of these elements. these are only general statements. with its breakdown of geometric order. Si. Thus we might expect to see some changes in bonding as we proceed from FeAl to NiAl to CoAl. It is impossible to assign a single number. and they suggested that. such as a value of electronegativity or valence. would cause a breakdown of directional bonding and become weak relative to the matrix. and therefore its presence might supply electronic charge in regions where two electropositive elements meet at the grain boundary.

M. 133. S.. B42. This rather cavalier practice has been all too common in the past. 39. Res. Briant. L. A. and Taub.. 7. A. Mag. 25. (1988). and Hua. Hamada. Nippon Kinzaku Gakkaishi. Finally. in compounds where one component is a transition metal. References Aoki. M.. (1990). and Kear. Symp. Mater..26% Zr was added. ActaMetall. Briant. E. R. Amsterdam. 23. I.. These will be essential for us to describe fully the fracture processes in these alloys. L. 281. In general. L. C. Taub. . The most successful studies of fracture in disordered alloys have taken a microstructural approach to the problem. 569. 5. C.. L. and Watanabe. The second is the type of work that needs to be done. E. 36. E. Schulson. Mag. The results described above suggest that much of the data reported in the literature may have been obtained under poorly controlled conditions. B62. E. 239. S. One surprising point that comes out of reviewing this field is how few compounds have been investigated over their entire range of stoichiometry. Phil. it is suprising that essentially no studies of this type have been performed on intermetallic compounds. H. (1988). They determine through various techniques the initiation sites for fracture and then determine how the crack propagates through the matrix. for the calculations. B. L. Copley. if the moisture content of the laboratory air can have the effect that these results imply. it is clear that quantum mechanics will be required. The third is the appropriate way to carry out this research so that the results will be meaningful in the broad context of intermetallic compounds. Given that there is a very complete literature on this topic for disordered alloys. 1271.-J. (1979). C. To make comparisons between alloys one needs to run tensile tests and to measure elongations. based on the work that has been published.An element that supplies electrons and segregates to the grain boundary might contribute electrons to the grainboundary interstices and thus aid in grain-boundary dislocation motion. and Pennycock. Physique Coll. 21A. Res. The primary conclusion to be drawn from the many studies that have been published is that the fracture mode in most intermetallic compounds tends to be one that we associate with brittle fracture and that the ductility of these compounds is often quite limited. Metall.. Elsevier. the elongation may be as high as 50%. C. and Izumi. 379. J. we need to have this fundamental information as a baseline for many of the other studies that have been carried out. TMSAIME. L.. L. The role of environment on test results needs to be investigated much more thoroughly. Phil. R. I.. 6. C. examples are reported in this chapter where the fracture appears to be brittle but the tensile elongation can be greater than 20%. Finally. Briant. J. Trans. 2339. A.. and Taub. /. Baker. (1992b). Baker. S. Such studies are not difficult to perform and usually involve careful optical and scanning electron microscopy on fracture surfaces and on samples tested to just below the fracture point. Phil. underscore this point. we consider the proper way in which experiments should be carried out. and Hall.. L. 1190.. 3. P. (1985) Metallurgical Aspects of Environmental Failures. and that through these calculations we can eventually describe the fracture process at a new level of microstructure. I. 2841. and Michael. S. C. Brenner. Chiba. However. S. Given that stoichiometry can have a major effect on ductility. as well as the effect of processing history.. 1883.. Stoichiometry and test environment can have a major effect on the measured ductility. We will consider each of these in turn. (1989). L. Individual reports of ductility on individual compositions are often hard to interpret and put into the correct context. J.. (1990). Briant.. Mater.. M. 922. L. Proc. The results presented by George et al. Baker. Michael. With alloying. (1988). Acta Metall. (1967). R. C. (1991). Next we address areas for future research. 659.. 1799.. (1991). Mater. Schulson... M. O. Trans. 43. Mater. ScriptaMetall. K. Samples should be tested in a vacuum to eliminate any environmental effects. (1989). Mag. Briant. J. 57.. and Messmer. 2761. Conclusions There are three different categories of summary statements that can be made concerning this work.. (1990). M. A. 49.. Soc. and Schulson. The first is the specific conclusions that can be drawn. C5. Further quantum-mechanical calculations are also required to obtain a better description of fracture based on differences in chemical bonding. Briant. Scripta Metall. S. Briant. I. increasing that component increases ductility. These alloying additives are often specific to particular compounds. which show the great improvement in the ductility of Ni3Al when 0. (1980). One should also test alloys of the composition specified and not add small amounts of third elements to improve processing and then forget about them in the discussion.

. P. Lett.. R. 1995. Liu. H. J. Chang. A. M. and Horton. K. T. T.. Huang. O. Mackenzie.. MetalL Trans. Mater. Field. McKamey. Takasugi. Pope. H. and Takasugi. (1991). Goods. 1357. Foiles. L. Eberhart. /. Darolia. and McKamey. . Symp. 35. Proc 5th Intl. 1247.. Vol. D... 33. (1990).... K. L. 38. Ada MetalL. I. Hahn. R. 27. 33. Baker. Zinoviev. De Van. P. 33. (1991). J. and Briant. 19. 11. Izumi. I. S. H. (1989). O. (1989). Scripta MetalL. 43(3). Res.-C. R. 23. (1987a). (1988). 1639. 25. S. A. O. Mulford.. (1989) Mater. A. Huang. M. and Izumi. J. T. F.. and Hanlon. Res.. P. 17. Mater. J...-M. Soc. Munroe. A. H. J. (1985). L. M. Viens. P... K. I.. E. Res.. E. Takasugi. (1991) J. Briant. AIME. PA... George. 18A. (1965). (1988).. 2199. In High Temperature Aluminides and Intermetallics (eds S. Y. P. J. and Sass. Scripta MetalL. I. (1987b). Res. N. E. Takasugi.. Rev. Briant. (1986). Y. and Shabel.. 955. Int. Schneibel. D.. and Briant. 23. Proc. E. 1779. King. Warlimont). Fu. 299. Vedula. George. (1986a). Losch. Metals Park. /. (1987). Chang. I. Mater.-M. M. T. L. (1987). L. S. E. C . (1986). Kear. and Stoloff. 335. 3. Masahashi. L. Scripta MetalL. E. C. 381. Fleischer. and Takeda. A. P. Huang. H.-M.. S. 426. 133. (1986b).. T. Takasugi.. Soc. Res. MetalL Trans. Mills. P. Y. K. and Hall. and Jackson. T. I. Xu. Mater. Frost. Takasugi. T. and Briant. C . J. (1992). Metals Rev. 36. J... J. Acta MetalL. and Padget. C. 1253. 12. 58. C. and Briant. L. Symp. A. and Porter. L. D.. 1703.. 22A... C.. Acta MetalL. Mater. 18A. American Society for Metals. Guha. K. George. Ada MetalL. Res. 129. Messmer. 2061. Res. 365. (1987).. Warrendale. Butler. Chatterjee. B. M. Taub. (1988a). 32. D. (1982). L. H. Vedula. and Izumi.. George. H. J. (1987). 2OA. J. R. Westbrook and R... J. Acta MetalL. T. A. I. 2025. 5.. Japan Inst. A.. and Westbrook. O. 1063. Taub. E. E.. 234. Mater. R. E. (1983). Metals.... M... McMahon. C. Scripta MetalL Mater. and Izumi... C. and Briant. C. T. Symp. p. 21. Res. Mulford. and Horton. (1979). Metals. Mater. R. L. A. C. Whang. 297. S. H. and Sikka. ScL Eng.. Latanision. Wasilewski.. C. (1993). Takasugi. (1991).. 1243. J. J. C.. and George. Liu.. and Izumi. (1986). (1986). C. and Pope. and Cohen. 133. H. Khadkikar. MetalL Soc AIME.. P. Porter. Acta MetalL. 208. and J. S. 195. Nagashima. D.. D. MetalL Trans.. J. and Boulogne.. C. R. E. M. Wood. P. Steigler). 25. 13.. and Johnson. 78. M. (1984). 1115. 15A. W. C. This chapter was originally published in 1995 as Chapter 38 in Intermetallic Compounds.. Liu.. and Izumi. M. I. S. K.. NucL Mater. T. (1990). 213. 823. Takasugi. 20. R. W. Vidoz. 754... TMS AIME. (1964). E. Scripta MetalL Mater. Briant. G. Mater.. Scripta MetalL Mater. R.. M. S... Acta MetalL. 23. O. White. L. 629.. J. 27. and Pope. E. Horton.. 288. Liu. and Chang. (1965). Taub. L. G. R. 30. (1990a). C. 14A. and Schulson.. G. 1024.-C. 419. and Wood. (1989). Ehlers. H.. (1990b). C. Weihs. G. Taub..-M.. T. 1283. p. 4. W. (1987). I. (1984). T. P. S.. D. T.. J. 1.. (1992). Steeb. 865.Crimp... E. 1807. Trans. Pope. I. 343. E. D. George. Amsterdam. A. Mendiratta. 1109. MetalL Trans. and Izumi. Painter. T.. R. and Pope. P. H. on Rapidly Quenched Metals (eds S. Proc. Trans. 34. 81. Chang... Lee. B . E. Acta MetalL. T. Taub. N. Schulson. 979. 35. Lazarevic. C. Taub. (1992).. W. A.. 153. 239. MetalL Trans. P.. 1156. MetalL Trans. C. 5. and Baker. and H. L. R. Horton. MetalL Trans. Jr. 22. Acta MetalL. K. Mills. H. and Liu. Mater. Fleischer.. W. (1988). Taub. 44. O. (1984).. 213. V.. C. W. Acta MetalL. McKamey. C . C. P. and Lipsitt. 6. and Briant. and Vedula.. and Chang. D. Soc. P. Takasugi. S. C. 399. Scripta MetalL. K. L. D. D. OH. 129. A. (1991). (1987). and Averill. (1985). Trans. 2298. Soc Symp. M. P. and Huang. 714.. M. 1407. L.. 19A. 29. Phys.1375. L. A. H. and Cahn. E. Masahashi. Scripta MetalL. R. 25. 941. (1991). T. A. K. H. 27. Murata. A. M. H. N. P. 224. A.-M. C. E. and Liu. Res. P. Acta MetalL. Weihs.. D. Taub. O. D. 591. and Schneibel. and Davenport.. 1597. L. F. George. L. M. S. Scripta MetalL. K. Scripta MetalL Mater. Schulson. Acta MetalL. Res. R. M. D. S. K.. Masahashi. J. 60. Proc. p.. J. on Rapidly Solidified Material. S. E. C. H. 17A. (1990). M. /. 33. Westbrook. Acta MetalL.. D. 875. and Izumi. E. M. (1985b). 1395. 36. 1259. P. ISIJ JntL. C. V. edited by J. (1984). K. Taub. Conf. A.. P. Davies. (1989).. Acta MetalL.. J.. Trans.. Huang. S. and Izumi. (1988b). A. Chang. Acta MetalL. (1962). Proc. and Schneibel. and Baker. H. Morinaga.. Huang. Schulson. Mater. K. (1991) Scripta MetalL Mater. and Hall. L. L... Liu. 193.. 457. Scripta MetalL. Liu. Acta MetalL Mater. Proc. 39. A.. Symp. L.. T. Mater. J. 1: Principles.. 425. A.. G. G. 527.. V. C. (1985). North-Holland. Schulz. 21. J. 2971. J. George... 233.. Schulson. Acta MetalL. O. 551. Tortinelli. and Yoo. R. P. Soc. Pope. Taub. N. and Whetstone. 283. 345. M. (1991). S.. C. T. Proc ASM Conf. (1963). 31.. P. O. (1991). K. O. R. Mater. 32. (1985a). T.. (1967).. TMSAIME. 35. T. MetalL Trans. S. (1989). A. A.. Liu. Res. 747. R.. Frost. P. (1973).. Acta MetalL.

However. Introduction Much of the present general and still-growing interest in intermetallics results from their potential for applications as structural high-temperature materials (see e. Edited by J. Thus intermetallic phases with high crystal symmetry— possibly cubic phases—and small unit cells are preferred for developments of structural materials. This ratio is called 'rupture length' and is about 15 km for advanced superalloys. as is exemplified by the familiar NiAl. A closer look indeed shows that the deformability of intermetallics is between those of metals and ceramics. or b. 1991. Phases that contain light elements—e. CuAu type). however. L. however.g. that the deformability Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. For conventional high-temperature materials.8Tm for Si and Al2O3. indeed.Chapter 3 Plastic Deformation Gerhard Sauthoff Max-Planck-Institut fur Eisenforschung GmbH. as will be discussed later. The intermetallic NiAl. 1989. by Schulson.. as is exemplified by Figure 1. by Miracle and Darolia. which is about the range of covalent Si and is higher than that of the ceramic Al2O3. The major problem of strong intermetallics is their brittleness. by Das. which means a sufficient creep resistance. by Yamaguchi and Inui.g. For such applications. Cu2AlMn type)—which are produced by atomic ordering in the f. there are some whose melting points exceed those of the constituent elements.c. D-4000 Dusseldorf Germany 1. Mg—may have such a low density that they compare favorably with conventional alloys in spite of lower strength or restricted service-temperature range. 2000 John Wiley & Sons Ltd . The creep resistance scales inversely with the diffusion coefficient and directly with the shear modulus. candidate phases should have melting temperatures at or above 1600 0 C for service temperatures of 1100 0 C or higher. Candidate phases of interest (in particular. Al3Ti type). the maximum service temperature is about (rarely above) | r m . and both the diffusion coefficient and the shear modulus scale with the melting temperature Tm. titanium aluminides. too. i. H. BiF3 type). Experience indeed shows that the brittleness of intermetallic phases increases with decreasing lattice symmetry and increasing unit-cell size (Paufler. which. 1976). L2t (cF16. the brittleto-ductile transition of NiAl is at about 0ATm.c. For moving parts. Ti. softens with rising temperature at about half the melting temperature. should be less severe than that of ceramics because of the different types of atomic bonding.c.c. by Banerjee. Ll 0 (tP4. There are many intermetallics with high melting points. Sauthoff. Sauthoff. by Vedula. and by Kumar in Volume 2). and their possibly less symmetric crystal lattices (Paufler. the ratio of rupture strength and specific weight is decisive. Density is an additional parameter of high importance. and iron aluminides) show such simple crystal structures—Ll2 (cP4. the ordered atomic distribution. a candidate phase must have sufficient strength at the service temperature. CsCl type). Thus NiAl bridges the gap between metals and ceramics. Plastic deformation is more difficult in intermetallics than in metals because of the stronger atomic bonding. B2 (cP2. It has to be noted. Cu3Au type). Liu et al. 1985).. Fleischer ©1995. by Huang and Chesnutt. DO22 (tI8. 1990. Engell et al. whereas this transition only occurs above 0.g. Al. nickel aluminides. Si. Westbrook and R.e. lattices and are cubic or only slightly tetragonal (see e. 1994). which is a candidate phase for high-temperature applications. and Chapters 2-10 by Liu and Pope. DO3 (cF16.

1967. Finally the preparation of components with reasonable quality must be possible for the intermetallic alloy in question. Kumar. The processing of high-strength intermetallics is difficult because of their limited ductility. melt spinning and gas atomization with compaction by hot isostatic pressing or hot extrusion (the latter is also used frequently for thermomechanical processing) (Fischer and Weber. 1991. Then strength is studied primarily in compression. Thadhani. and thus they may preclude any application. alloys for mechanical testing have usually been produced only in small quantities. Intermetallics for high-temperature applications must also be oxidation-resistant in most cases (see Chapter 43 by Doychak in this volume). B2) that exhibit insufficient deformability. cracks. Novel methods of alloy synthesis are used to try to circumvent the brittleness problems. Chapter 27 by Martin and Hard wick in this volume). Si. and indeed there are phases with highly symmetric and small unit cells (e. 1990. and therefore phases for applications at 11000C and higher should be inherently oxidationresistant.Yield stress (MPa) T'Tm Figure 1. i. In view of these difficulties. and rapid solidification. Sen and Stefanescu. 1982). 1993. e. which can be done with small specimens of simple shape that are less sensitive to internal cracks.g. hardness tests are used for studying Deformability (%) . However. Thus aluminides are strongly favored for hightemperature applications.e. including directional solidification. However. shock synthesis. reactive consolidation. Furthermore. with a sufficiently low content of inherent macroscopic defects—in particular pores. Hindam and Whittle. Likewise ductility and toughness are studied by bending tests with small simple specimens instead of by tensile tests. In cases of insufficient oxidation resistance. 1991. mechanical alloying. the stability of such coatings decreases with increasing temperature. silicon oxide may form low-melting silicates. 1965) of intermetallics is a complex function of various other materials parameters besides crystal structure. Inherent defects increase the apparent brittleness and decrease the fracture strength. Yield stress (—) and deformability (—--) for various materials as a function of homologous temperature T/Tm (Tm = melting temperature) (From Westbrook. and thus the development of the necessary processing techniques is a very demanding task. chromium oxide and silica are volatile at low oxygen pressures above 1000 0 C. Al. protective coatings may be applied (see Chapter 22 by Nichols in Volume 2). spray casting.g. e. and inclusions. Such elements are first of all Cr. orthorhombic AlNbTi2. Oxidation resistance is provided by the presence of elements that can form protective oxide layers. and Be is toxic. and Be (Aitken.g. as well as phases with complex crystal structure that can be deformed plastically.

the plastic deformation of intermetallics usually occurs by the movement of dislocations. the crystal symmetry of even the intermetallics with simple crystal lattices is lower than that of conventional alloys because of the particular atomic order—see Section 4. 1979). as well as for phases with other crystal structures (Stoloff and Davies.e. 1990) for a large number of intermetallic phases with various crystal structures—including the f. 1994). which control the strength and ductility of intermetallics. and which has been proposed to be due to the formation of very stable dipoles between superdislocations (Marcinkowski. Sauthoff. 1974. MgCu2 type). The indentations can be evaluated with respect to strength at low temperatures and creep at high temperatures. 1974. Umakoshi. Cu3Au type). and the tetragonal structures Cll b (tI6. Westbrook. the hexagonal structures A3 (hP2...e. 1991. Davis et al.c.. NaCl type). and thus Figure 2 shows the strengthening behavior of this phase for various states of atomic order. the atomic order across the slip plane is destroyed by the leading partial and restored by the trailing partial of the super dislocation. which shows an order-disorder transition at about 900 0 C. 1984. Stoloff. The presence of the more complex superdislocations in intermetallics with their ordered atom distribution affects the strengthening behavior significantly compared with disordered alloys. Mg type). (cF4. and DO19 (hP8. C15 (cF24. MoSi2 type). B3 (cF8. Another outstanding effect of atomic order and the presence of superdislocations is the possibility of a positive (i. structures Bl (cF8. CsCl type). However. The resulting complex dislocations are known as superlattice dislocations (or superdislocations) and their energies are composed of the elastic energies and core energies of the partials and of the energies of the antiphase boundaries and/or stacking faults. Deformation Mechanisms 2. In the following the important features of the plasticity of intermetallics are discussed with respect to the basic mechanisms of plastic deformation. 1931). behavior.. B2 (cP2. 1985). 1991.e.c. Baker and Munroe. it exhibits the disordered f. BiF3 type). a behavior that has also been observed for another Ll2-phase Cu3Au (Sachs and Weerts. This is illustrated by the behavior of the ternary phase (Co0 7Fe0 3)3V. MgZn2 type). and Ll 2 (cP4. Cu3Au type) structure at lower temperatures—apart from intermediate a-phase precipitation between 890 and 940 0 C (Liu and Inouye. 1990. 1991. The observed dislocation types and glide systems have been listed (Marcinkowski. i.g. This effect is attributed to the presence of short-range order in the disordered or partially ordered state. C14 (hP12. 1991). and Ll 0 (tP4. 2. If such a super dislocation moves. the reader is referred to Sauthoff (1994) as well as to the other chapters of this book. Field et al. 1960. Cu type)structure at higher temperatures and the ordered Ll 2 (cP4. CuAu type)—and have been discussed in detail (Yamaguchi and Umakoshi. Fleischer et al. Paufler. Li et al. ZnS type) Cl (cF12. 1984). The disordered state at high temperatures can be retained at room temperature by quenching. Chapter 22 by Veyssiere and Douin in this volume). W3O type).2 and the crystallographic chapters in Part III. C16 (tI12. the initial flow stress is lower for the ordered state than for the disordered or partially ordered state in Figure 2.the material behavior locally. 1977. The particular type of dislocation is determined by the symmetry and energetics of the intermetallic in question. 1989. CuAl2 type). Ni3Sn type). It is noted that in some B2 (cP2. Engell et al.1 Slip As in conventional disordered metallic alloys. 1990.c. which makes dislocation movement more difficult than in the fully ordered state (Stoloff. Sargent and Ashby. which does not affect the order of atoms. Clearly the atomic order increases the strain-hardening rate. CaF2 type). 1993). 1990.. and in particular the deformation behavior has been the subject of recent reviews (Yamaguchi and Umakoshi. i. Al content and alloying additions (Baker and Munroe. 1976. In spite of this strengthening effect. 1989. Marcinkowski. anomalous) temperature dependence of . For a detailed discussion of the various intermetallic phases of interest with respect to properties. and the cracks that are produced by the indentations can be used to obtain estimates of the fracture toughness (Liang et al. Thus in such an intermetallic a perfect single dislocation. 1990. Intermetallics have been reviewed repeatedly in the past as well as recently (e. has a longer Burgers vector— and thereby a much larger energy—than a dislocation in a disordered alloy. the simple cubic structures Al5 (cP8. 1964a. and applications.. Consequently such dislocations usually dissociate into two or more partial dislocations with antiphase boundaries and/or stacking faults in between in order to reduce their energy.c. 1992). and DO3 (cF16. 1985). CsCl type) phases—in particular in FeAl and NiAlboth perfect single dislocations and superlattice dislocations have been observed depending on deformation conditions and composition.

The splitting on (0 1 0} is favored energetically because the energy of the antiphase boundaries (APBs) between the partials is lower on {0 1 Oj. 1970). Izumi. which determine the plastic deformation of Ni3Al. i.Stress (MPa) Strain (%) Figure 2. 1985. and curve C. Stress-strain curves of polycrystalline (Co07Fe0 3)3 V at room temperature for various states of atomic rder: curve A.e. Pope and Ezz. This has been most studied and best understood in the case of the Ll 2 (cP4.c. 1981.e. a superdislocation on {010} is sessile because the dislocation cores of the partials spread outside the plane of the APB. 1989. and only at higher temperatures does normal softening occur. Paidar etal.1990. With rising temperature.. Plastic flow of Ni3Al has been studied by many authors in much detail. ordered Ll 2 . and slip is confined to {1 1 1} at low temperatures.e. resulting in immobilization of the dislocations (Kear-Wilsdorf mechanism). Umakoshi.c. 1984. Liu et al. The sessile dislocations act as obstacles and give rise to rapid strengthening with increased flow stress. The details of these mechanisms and reactions have been and still are the subject of specific studies.. Suzuki et al. 1989. partially ordered. and indeed the minimum strain for an observable flow-stress anomaly was found to be about 10"5 (Thornton et al. as-quenched. Stoloff. the superdislocation on {1 1 1} (with higher energy) is glissile because the core spreading is better confined to the slip plane. This glissile state is metastable since it is thought that the partials must first be coalesced into a single dislocation before cross-slip back to (0 1 0} can occur. 1984. and the theoretical description becomes more and more elaborate (Pope . The results have been reviewed and discussed intensively and the basic mechanisms are thought to be well understood now (Stoloff and Davies. However. At temperatures above the flow-stress maximum the immobilized dislocations are mobilized again by enhanced thermal activation. 1993). 1986). disordered f. which leads to the familiar softening. 1979) the flow stress. On the other hand. i. Cu3Au type) phase Ni3Al and is illustrated by Figure 3.. aged 16 000min at 700 0 C. curve B. Between room temperature and about 700 0 C the flow stress increases with increasing temperature to reach a maximum. As has already been discussed (Sauthoff. The screw dislocations can split on {1 1 1} planes and on (0 1 0} planes. Liu and Stiegler. Liu and White. cross-slip to {0 1 0) becomes possible by thermal activation. Thus it is clear that the anomalous temperature dependence of flow stress is observed only with sufficient strain. i. aged 8 min at 700 0C (From Liu and Inouye. Hence dislocations are generated primarily on {1 1 1} planes on loading. 1990. the anomaly results from the anisotropy of the energy and mobility of the super lattice screw dislocations.. 1966. 1984. Yamaguchi and Umakoshi.

Suzuki et al. Khantha et al. Ll 0 ). Westbrook. 1991. can be obscured by other concurrent strengthening effects.% Ti. 1988. Stoloff. e. It has been found for such metals that twinning is preceded by slip and the resolved shear stress for twinning is substantially higher than that for slip.b. but also at lower strain rates and higher temperatures. Al3Ti type) structures (Mishima et al. gradually increasing the low-temperature flow stress (Jung and Sauthoff. 1992.2% flow stress (MPa) . TiAl (tP4. 1985. The lattice in the twinned region—separated from the matrix by the twinning plane—is usually the mirror image of the parent lattice.Temperature (0C) Figure 3. 2. Paidar et al. 1991.. Veyssiere. whereas it is weaker in Co3Ti. a flow-stress maximum is no longer discernible (Schulson. Vitek and Sodani. Sodani and Vitek.c.C16)—and various mechanisms have been proposed for these phases (Stoloff and Davies. Suzuki et al. the close-packed hexagonal metals. An analogous effect is produced when C is dissolved in Ni3Al. 1985). Mg3Cd (hP8. as is exemplified by Figure 3. Pope.Fe)3Ge when going from Ni3Ge to Fe3Ge. Such flow-stress anomalies have also been shown by other phases with different crystal structures—e. in particular the DO19 (hP8. and the propensity for twinning increases with increasing deformation rate and decreasing temperature. 1983. DO19). 0. This means that deformation twinning is prevalent not only at high strain rates and/or low temperatures. 1989. 1991.. i. where slip is frequently limited to the unique basal system. the grain size can be reduced to such an extent that the low-temperature flow stress is increased to a level higher than the original flow-stress maximum at high temperature.g.% Ti + 2 at. 1993). Molenat and Caillard. (o) 2 at. Twinning has also been observed in b.c. 1992).% Cr (From Thornton et al. 1968). the flow-stress maximum occurs just below the critical temperature of disordering and is related to the decreased degree of long-range order. Saada and Veyssiere. FeCo-2V and Fe3Al). but it does not alter the symmetry or structure of the crystal. Likewise the flow-stress maximum of Ni3Al can be increased and shifted with respect to temperature by ternary alloying additions. Vitek et al. Yoo and Fu. 1991. 1991..g. metals. ( A ) 10. and Cu3Au. there is a continuous transition from a positive temperature dependence to the familiar negative one in the Ll 2 ternary phase (Ni. Yoo et al. and the differences are closely correlated with the differences in stability between the Ll 2 (cP4. CuZn (B2). Fe2B (til2. Twinning is a familiar deformation process in metals where the possible slip systems are severely limited. The positive temperature dependence of the flow stress. 1976. Indeed the anomaly is exhibited by various Ll 2 phases to different extents. DO3). Heredia and Pope. twinning is the other principal mode of plastic deformation: it is more important for intermetallics than for disordered alloys since dislocation slip is usually restricted because of the higher energy and lower mobility of dislocations in intermetallics (Yoo. Kawabata et al. 1989. 1989). 1991.. From this discussion it is clear that the flow-stress anomaly of Ni3Al is a function of the binding forces between the atoms. of Ni 3 Al. Fe3Ga. 1970) and Ezz.. ( • ) 10. 1992. 1991. 1984. Hirsch. Ni3V (tI8.5 at. The formation of twins during deformation often results in serrations in the stress-strain curve. and f . Cu3Au type) structure and other possible structures. 1984.c. In the case of phases with an order-disorder transition (e. In Ni 3 Si and Ni 3 Ge the anomaly is still more pronounced than in Ni3Al.5 at. as for disordered alloys. The process of twinning is a cooperative movement of atoms producing a macroscopic shear (Honeycombe. As an example. Fe3Al (cF16.2 Twinning Besides dislocation slip. 1989b. e. Ni3Sn type) and DO22 (tI8.c. 1964a. Zr3Al. Paufler.g. DO22). B2).. The twinning transformation changes the orientation of the twinned region. Yoo et al.g.% Cr. 1991.. Haasen. 1965. Temperature dependence of flow stress for cast polycrystalline Ni3Al (•) and Ni3Al with various alloying additions: (0) 6 at. 1994). 1985).. The crystallography of twinning has been discussed in detail by Christian (1975). Wunnike-Sanders and Sauthoff. 1991a..e. which leads to a less strong coupling of the partials of the superdislocations. FeCo-2V (cP2.% Nb.

T is temperature. Al3Ti type). and DO22 (tI8. 1990). been the subject of various summarizing reports (e. Chapter 26 by Schetky in Volume 2). 1989. 60% of the melting temperature or higher—with creep rates between about 10~8 and 10~6 s 1 (Rudy and Sauthoff.. A is a dimensionless factor. Ll1 (cF16. 1991. and this may be supposed to be the case for intermetallic phases.g. W3O type) structure (Khantha et al. Rudy. as is illustrated by the so-called deformation maps (Frost and Ashby. B2 (cP2. DO3. 1985.Fe)Al exhibits a power-law behavior at high temperatures—e. and hexagonal DO19 (hP8. However. here the creep is controlled by the viscous glide of dislocations. Sauthoff. 1979). Important practical examples are the shape-memory phase NiTi with B2 structure (Goo et al.6. 1991a.. 1993). and the findings have . Otherwise. D is the effective diffusion coefficient. and the stress exponent varies between 3 and 3. In both cases only <10 0> dislocations have been observed. 1986. and the stress exponent is 4 or 5. 1975. which refer to the steady-state creep stage. The creep behavior of NiAl and NiAl-based alloys has been studied systematically. 1990b). behavior characteristic of the class I alloys (Sherby and Burke. The following discussion of the main characteristics of the creep behavior of intermetallics refers to NiAl-based intermetallic alloys as typical examples since the understanding obtained there of the rate-controlling mechanisms is thought to be valid for other intermetallics. 1991. are described by the familiar Dorn equation for dislocation creep (Mukherjee et al. 1985. and the Laves phase HfV2 with cubic C15 (cF24. Chapter 35 in this volume). 1989) as well as for the cubic Al5 (cP8. viscous dislocation glide is rate-controlling. the energetics and kinetics of deformation twinning of intermetallic phases have not yet been studied adequately. CsCl type).e. Chapter 7 in Volume 2) and Al3Nb (Schechtman and Jacobson. as in pure metals. Cu3Au type). The dislocation density remains high (about 10 10 cm~ 2 ).g. 1991.. which are used during typical creep experiments. Dislocation creep of conventional disordered alloys is produced by gliding and climbing dislocations..Important contributions of deformation twinning to plasticity have been reported for various intermetallic phases with various crystal structures—cubic Ll 2 (cP4.e. the DO3 phase Fe3Al (Park and Goo. 1988. 1989). Jung et al. Alloys with this characteristic are termed class II alloys. stress exponents between 4 and 4. and a is the applied stress. Chapter 4 by Huang and Chesnutt in Volume 2). i. Reip. k is Boltzmann's constant.. DO22. The secondary creep behavior of (Ni. which are used during short-term tests for flowstress determinations.e. Chapter 9 in Volume 2). however. tetragonal Ll 0 (tP4. 1987). these phases behave like class II alloys with dislocation climb controlling the creep. The creep of the ternary B2 phase (Ni. 1967. and DO3 (cF16.Fe)Al shows analogous characteristics. a well-defined substructure is found after creep. In agreement with this.3 Dislocation Creep High-temperature deformation of metallic and other materials is controlled by various deformation mechanisms depending on temperature and deformation rate. 2. too (Sauthoff. If climb is the slower step. Yamaguchi and Inui. 1992). 7-TiAl with Ll 0 structure (Kim and Dimiduk.. The comparison of such maps for various materials with different crystal structures and types of atomic bonding shows that the deformation behavior at temperatures above one-third of the melting temperature (order of magnitude) is controlled for many materials by creep processes not only at deformation rates of 10~7 s"1 and lower. G is the shear modulus. but also at higher rates of say 10~~3 s' 1 . B2. b is the Burgers vector. and the dislocation density within the subgrains is about 10 7 cm~ 2 . the observed secondary creep rates. Vedula. with the exponent n usually between 3 and 5. see also Wayman and Inoue. which leads to dislocation tangles without subgrain formation and a stress exponent of 3. Thus any discussion of high-temperature deformation behavior has to be centered on creep mechanisms.. no subgrain formation has been observed even after long creep times. Cu2Mg type) structure (Livingston and Hall. i. The subgrain size is of the order of 10 /an. Obviously the driving force and the atomic mobility that are necessary for subgrain formation are sufficient only in the Ni-rich phases. and correspondingly the effects of deformation twinning on strength and ductility are not yet well understood (Yoo et al. CuAu type). the creep rate is controlled by dislocation climb. 1991). Nix and Ilschner. the DO22 phases Al3Ti (Yamaguchi et al. This indicates class I behaviour. Ll 0 .5 have been found for the Ni-rich phases. which gives rise to a well-defined subgrain structure. The crystallography of twinning—in particular the conjugate relationship between the order twinning and the active slip system at elevated temperatures—has been discussed in detail for the crystal structures Ll 2 . Cu2AlMn type). 1982). BiF3 type). In the Fe-rich phases and in FeAl. In the Ni-rich phases and in the binary NiAl. 1969): € = A{DGb/kT){o/Gf where e is the secondary strain rate. Ni3Sn type) (Yoo. and DO19 (Yoo. Shimokawa et al. i.

1990b). grain-boundary sliding was observed in (Ni. This deviation indicates the contribution of diffusional creep. D is the effective diffusion coefficient.Fe)Al (Rudy.4 Diffusional Creep At lower stresses. 1982) and has also been observed for (Ni. Still such models do not yield the correct prediction for the stress dependence of the total primary strain—i. 1968). as has been shown for a NiAl-base alloy (Klower and Sauthoff. i. 1987). Sauthoff. With decreasing grain size. 1969. 1982). Even the observed dislocation densities correspond to those in conventional disordered alloys. even for conventional metallic alloys. and the total creep rate is given by the sum of the partial rates.These findings show that dislocation creep of such intermetallic alloys and of conventional disordered Ni-base or Fe-base alloys is controlled by the same mechanisms.. 1986. as has been observed for the ternary B2 phase (Ni. hardening with decelerating creep rate.e. by dislocation creep of the grains in the stress-temperature range concerned. which results from the stress-induced diffusion of atoms (or vacancies) and is a linear function of stress: edm = Adiff(QD/kTd2)(j where ^4diff is a dimensionless factor (usually A diff = 14). Gottstein and Argon. 1992). which is dislocation creep for not-too-small grain sizes. Because of the stronger stress dependence of dislocation creep. which considers the diffusion both through the grain (Nabarro-Herring creep) and along the grain boundaries (Coble creep).Fe)Al.. a fact that must be considered when comparing such thermally activated processes. 2. 1987). i. MgZn2 type) structure (Machon. Besides dislocation creep. Estrin and Mecking. The process is. i.Fe)Al (Jung et al. Mecking and Estrin.g. it is finally to be noted that the creep resistance of Ni3Al does not show any anomalous temperature dependence (Schneibel and Hazzledine.g. the contribution of grain-boundary sliding increases. the initial transient creep stage exhibiting normal primary creep.e. the observed stress-strain rate relationship may deviate from the power-law behavior. and d is the effective diffusion length. from which a decrease of the total creep resistance results. which is usually approximated by the grain size (Frost and Ashby. the contribution of diffusional creep becomes more prominent with decreasing stress. Early models have been based on simple phenomenological approaches (e. both positive and negative stress dependences have been found for intermetallic alloys (Reip. 1973.e. Another example is the ternary Laves phase NbNiAl with the hexagonal C14 (hP12. 1992). the apparent stress exponent is smaller than 3 and decreases with decreasing stress. 1991. The theory of diffusional creep that results in the above constitutive equation is well established on the . in particular the activation of octahedral slip and cube slip at different stresses (Schneibel and Hazzledine. since the resulting grain shifts lead to stress concentrations at grain-boundary junctions and must be accommodated by deformation processes within the grains. 1987. 1991). the strain rate-temperature range of creep corresponds to that above the peak temperature in Figure 3. Derby and Ashby. as is well known for disordered alloys (Frost and Ashby. Primary creep has been studied only in a few cases. In view of the anomalous temperature dependence of the flow stress of Ni3Al (see preceding section). 1969. The reason for this is that creep occurs at much lower strain rates than used in flow-stress tests. the total creep rate is controlled by the slower process. which is known as inverse creep (Hemker et al.e. 1991b). For such coupled deformation processes. Schulze and Sauthoff. Davies et al. Klower and Sauthoff. 1992) as well as for conventional metallic alloys (Ilschner. Only recently have creep models been developed that take into account microstructural changes (e. Alexander and Haasen.e. 1984. This close similarity is surprising since the elementary atomic diffusion processes and the types and numbers of dislocation slip systems are quite different because of the different crystal structures. softening with accelerating creep rate. Inverse creep is always observed when the number of mobile dislocations is initially insufficient and increases during creep. Secondary creep is preceded by an initial transient primary creep. Webster etal. 1992). Inverse creep has been studied and analyzed in detail in the case of Si (Alexander. Dislocation creep and diffusional creep are independent creep processes that act in parallel in the grains. i. as has been shown (Rudy.. of course. Indeed. i. The reason for this behavior is again the complexity of the available slip systems. the strain for reaching the secondary stage. which produce secondary creep rates below 10~8 s" 1 . is succeeded by a second transient stage with anomalous primary creep. 1986. Q is the atomic volume. not an independent deformation mechanism. 1983). 1992). Sauthoff.e. Ahmadieh and Mukherjee.. In the case of Ni3Al. 1987. for which a detailed quantitative physical understanding is still missing. 1986. 1975) and are not related to microstructural features of plastic deformation.

and it has been used to consider effects of grain shape and grain growth (Burton and Greenwood. 1962). 1991. Herring. 1950.9 1987) shows the high-temperature part of a deformation map for the ternary intermetallic phase (Ni0 8Fe02)A1 with cubic B2 structure that results from the A2 structure by atomic ordering. and (Ni. which is known as Harper-Dorn creep. and microstructure. Obviously the normalization of applied stress by the shear modulus and of temperature by the melting temperature is sufficient for a unifying description of the global deformation behavior of various materials including intermetallics. 1991).e. its strength. the refractory carbide ZrC with cubic Bl structure. 1984. 1989. 1948. Clegg. 1980). 1977). Burton. to high-temperature deformation. but not yet for intermetallic phases..3 Tm to 0. which correspond to service conditions. 1982. Arzt et al. there is a transition to dislocation creep (Greenwood et al. Nevertheless. 1969. This understanding is based on detailed studies of the low. Greenwood. Figure 4 (Jung et al. do not differ too much for these dissimilar materials in spite of the differences in atomic bonding and crystal structure. Much more work is necessary with respect to the various intermetallic alloys being developed. Vehoff. Complete deformation maps are available for many metals and alloys as well as for some non-metallic materials (Frost and Ashby.g. It supposes grain boundaries to be ideal sources and sinks of vacancies.g. Besides grain boundaries. 1972. Again the behavior of the intermetallic does not differ too much from that of the other two examples. there is a threshold stress for diffusional creep (Ashby.. Diffusional creep is of great practical importance since it controls the long-term creep deformation at very low deformation rates. climbing dislocations may act as sources and sinks for vacancies. when drawn to normalized scales. Coble.. As to ductility and toughness. Wolfenstine et aL9 1991). is supposed to occur in the temperature range 0. with a map for an only partially metallic interstitial compound. 1985. the transition from low-temperature deformation. Stevens. but again the experimental verification of proposed models is still in discussion (Mohamed et al. but not yet for intermetallics. 1972. i. b. Langdon and Yavari. 3. Arzt. The diffusion data that are needed to analyze the creep behavior are rare or not available at all for the intermetallic phases of interest (see Chapter 32 by Larikov in this volume). which is controlled by thermally activated creep processes. NiAl (Vandervoort et aL9 1966). The resulting diffusional creep. in Figure 1 for NiAl and Figure 5 for TiAl.e. shows that these maps. 1977). Most intermetallics exhibit low or no ductility at low temperatures and become ductile only at elevated temperatures by the thermal activation of deformation processes. 1977.g. Fiala et aL. 1979. analogous fracture maps are available only for some metals and alloys and a few other materials (Ashby et al.basis of thermodynamics (Nabarro. Such deformation maps show the material's response to applied stress and temperature with respect to deformation rate. e. Nabarro. but the experimental evidence is still in discussion (Harris etal. 1977). Mishra et al.5 Tm for intermetallics. 1972. As to ordered inter metallic alloys. 1973. 1982). 1989. Fracture of intermetallics has become a subject of systematic research only recently (see e. With increasing stress. where Tm is the melting temperature. Crossland and Clay..Fe)Al (Sauthoff. Burton. 1983.. Stress and Temperature Dependence of Strength and Ductility (Deformation Maps) The preceding discussion has shown that plastic deformation is controlled by various different mechanisms depending on stress. Burton. the transition metal Cr with A2 structure (i.c). 1963. Mishra et al. A quick overview on the prevailing deformation mechanisms is given by the deformation mechanism map for a given material (Frost and Ashby. The brittleto-ductile transition by thermally activated processes . If grain boundaries do not act as ideal sources and sinks of vacancies. experimental studies of diffusional creep are rare even for conventional disordered alloys. Gandhi and Ashby. diffusional creep has been studied in the case of an advanced Ni3Al alloy (Schneibel et al. A comparison of one such map for a pure metal. becomes rate-controlling for very large grain sizes.1988) as well as the coupling of diffusional creep with grain-boundary sliding (Ashby.9 1975. 1982. for example. 1979). The resulting brittle-to-ductile transition temperature in such cases corresponds to the temperature of rapid softening by creep as is visible.c.and high-temperature deformation studies of various intermetallics.9 1986). in particular the B2 phases AgMg (Wood and Westbrook. which is controlled by the conservative motion of dislocations. Models are available. 1982). This means that the mechanical properties of intermetallic phases at high temperatures are determined by creep processes and depend sensitively on temperature and deformation rate. Chapter 38 by Briant in this volume). 1992. temperature. Clegg and Martin. In the deformation mapping concept just described. e. 1990b).

e. and the diffusion coefficients. High-temperature deformation map with curves of constant compressive strain rate as a function of normalized compressive stress and temperature (G = shear modulus. which control not only the kinetics of phase reactions but also the kinetics of high-temperature deformation. 1986. 1982). Vehoff. Plasticity-Controlling Factors 4. polycrystalline TiAl with 54 at. Stress (MPa) Temperature (0C) Figure 5. and thus the melting temperature is a rather complex Elongation (%) leads to a strong strain-rate dependence of the brittleto-ductile transition temperature and of toughness.Next Page T(0C) o/G T'Tm Figure 4.e. 1992). ultimate tensile strength (UTS). Jung et al. and 0. 1990b.2% yield stress—as well as plastic deformability—i. Examples are the elastic moduli. most materials parameters that characterize the deformation behavior are well correlated with the melting temperature (Frost and Ashby.e.1 mm grain size (From Rudy.% Al (From Lipsitt et al.e.y 1975) However. which not only control the elastic deformation but also are important parameters for describing the plastic deformation. a phase melts when the total energy of the solid phase is higher than that of the liquid phase. Tensile strength—i.1 Strength and Character of Bonding It is generally supposed that a high melting temperature indicates high phase stability and strength of bonding. i. creep. i. . Tm = melting temperature) for (Ni 08 Fe 02 )Al with 0. 1987) 4. tensile elongation—as functions of temperature for single-phase. fracture stress. Indeed.. the brittle-to-ductile transition temperature increases and toughness decreases with increasing strain rate (Sauthoff.

Westbrook and R. Fleischer ©1995.g. Parks Road. A and B sublattices in Figure 1. energetics. Urbana. is referred Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. one of which is occupied by B atoms (B sublattice) and the remaining three by A atoms (A sublattices).Chapter 4 Structure of Antiphase Boundaries and Domains Yong-Qian Sun Department of Materials.c. in which the superlattice (unit cell marked by full lines) is made up of two sublattices (unit cells marked by broken lines). derived from the body-centered-cubic (b.) structure. APBs are also effective barriers to dislocation propagation. there are two A sublattices. e. L. University of Oxford. APBs also play an important role in phase transformations involving the generation or elimination of APBs. This chapter is concerned with the basic crystallographic. USA) 1. one B sublattice. ordered from the face-centered-cubic (f. A vector.c. and energetic properties of APBs and APDs. in the Ll1 (cF16) superlattice (composition A2BC). The boundary marked x is a non-conservative APB where there are no B atoms of atoms that occupy them. The ordered arrangement of atoms in an intermetallic compound is usually represented in terms of a superlattice and the constituent. consists of four interpenetrating simple-cubic sublattices.c. Similarly. Many of the interesting mechanical and electromagnetic properties of ordered alloys have been interpreted on the basis of the presence of APBs and APDs and their crystallography. Oxford OXl 3PH. Electromagnetic properties are affected by APBs and APDs because of their effects on the band structure.c. In ordered alloys the mobility of dislocations is often governed by their core structures. as well as in the disorder-order transition. A twodimensional analog is shown in Figure 1. which in many cases can be controlled through heat treatment and alloying. UK (now at Department of Materials Science and Engineering. The two sublattices are marked by broken lines and p is the APB vector. The Ll 2 (cP4) superlattice (composition A3B). Edited by J. microstructural. within the superlattice unit cell. Introduction Antiphase boundaries (APBs) and antiphase domains (APDs) are microstructural features that are unique to ordered alloys and are not present in disordered alloys. and chemistry. identical sublattices. their segregation to the APB. APB and APD in a two-dimensional AB alloy with a square superlattice unit cell. University of Illinois. which in most cases involve APBs. 2000 John Wiley & Sons Ltd . The sublattices are usually named after the types Figure 1. Illinois 61801. The role of alloying elements in intermetallics is sometimes reflected in their effects on the energetics and chemistry of APBs through.) structure and also having four simple-cubic sublattices. H. for example. and as a result the mechanical properties are influenced by the morphology of APDs. and one C sublattice.

Sanchez et al. When p. Popov et al. See Tables 1-6 for a list of p vectors in six of the most common superlattice structures. and B2 (cP2). Tichelaar et al. Ngan. as first pointed out by Flinn (1960).n = 0. In the latter case the composition in the boundary is different from the bulk if the removed layer does not contain atoms in the stoichiometric proportion. such a vector. details of which can be found in Chapter 22 by Veyssiere and Douin in this volume. 1962. is a lattice vector of the disordered structure. there is no crystallographic restriction on the plane normal n of the interface. Deviations from ideal APBs may also appear in the forms of local disordering (Brown. When p. Kikuchi and Cahn. 1970. 1966. Ll 2 (cP4). the APB can be formed by a pure shear without a change of stoichiometry in the boundary. (cF16)) or more (e. 1992) and segregation of vacancies. however. 1963. In dislocation mechanisms.g. and followed by the growth of the ordered regions. In some superlattices. Lasalmonie etal. the widely considered circumstance has been that ordering takes place in an existing disordered or short-range-ordered crystal grain. Leamy et al.n value. which separates two rigidly displaced crystals with uniform composition and degree of order. Beeler. the atom positions. and such APBs are termed nonconservative. 1959. Ll 0 (tP4). 1966.n 9^0. 1960) and higher-order neighbors (Marcinkowski and Brown. A non-stoichiometric.. between nearest neighbors (Flinn. 1979. 1989. 1987. Hazzledine and Sun. Deviations from ideal lattice positions are referred to as structural as an antiphase vector if it translates one type of sublattice into coincidence with another type. p and n are useful in assessing.. Veyssiere.. for example. such APBs are called conservative and are involved in the glide dissociation of dislocations. these will be referred to as chemical effects. Dislocation mechanisms have also been found to generate a special kind of APD. Yamaguchi and Umakoshi. the anisotropic nature of the surface energy of the APB by examining the bond changes. the local composition. the reader is referred to Chapters 11-20 in this volume. 1989. an antiphase boundary is an interface separating two identically oriented crystals that have been displaced with respect to each other by p. There are also others in which two (e. and subsequently frequently observed by transmission electron microscopy (TEM) (Marcinkowski. minimum APB energy is expected to be on the (10 0) planes in which. For descriptions of crystal structures in intermetallic compounds. Figure 1 shows an APD and the enveloping APB in the two-dimensional superlattice. In ordered alloys there are two distinctly different processes in which APBs and APDs may be formed. denoted p. non-conservative APB is shown marked by x in the two-dimensional superlattice in Figure 1. In the Ll 2 structure.. and as a result there are no B atoms on either side of the APB. and the degree of order are all expected to deviate from those for an ideal APB. For a real APB. The p vector in the two-dimensional superlattice is shown in Figure 1. which become APDs when they eventually come into contact. Cupschalk and Brown. 1992).e. an APB appears in the form of a usually narrow ribbon coupling two or more partial dislocations. An antiphase domain within an ordered crystal is a region that is enveloped by a continuous piece of APB.. e. According to the p.g. DO22 (tI8). after the APB displacement. known as an APB . 1963. on a qualitative basis. 1981a. DO3 (cF16). The number of APB or APD variants available in a superlattice is given by the number of antiphase vectors. bonding among the nearest neighbors is not changed. 1988. Latest experiments using a combination of weak-beam and lattice-resolution TEM have demonstrated that APBs may be involved in more complicated dissociation modes.5). 1970. Ray et al. 1991). i. this APB can be considered as being formed by removing a layer of B atoms parallel to the boundary. 1987. 1992. Such observations have formed the basis for the experimental measurement of the surface energy of the APB (Section 2. there is just one type of antiphase vector. 1990). DO19) (hP8)) types of p vectors are present. namely disorder-order transformations and the dissociation of dislocations. solute. 1990. Thus.n|. occurring during cooling through a certain transition temperature. Schoeck and Korner. Wu et al. such an arrangement defines an ideal APB.. owing to the different bonding environment near the boundary from the interior. when the displacement vector lies within the interface. and L2. and in general the APB vector p is equal to the Burgers vector of the dislocation that produces the APB. the APB connects two identical superpartial dislocations. By definition. Korner and Schoeck. the APBs can be classified into two basic types (Marcinkowski. and excess atoms to the APB (Popov et al. 1990. 1968. 1985.. The arrangement of atom species across an APB is fully characterized by p and n... but not of the superlattice. across the APB. or residual or supplementary displacements (Yamaguchi et al. the APB can be considered as being formed by a combination of a pure shear and the removal or insertion of a layer of material with a thickness equal to |p. 1991). Paidar. Beauchamp et al. proceeding first by the nucleation of small ordered regions or nuclei. 1967). In disorder-order transformations. Fu and Yoo.g. In the classical example first proposed by Koehler and Seitz (1947).

g. The contrast originates from the interference between the waves diffracted into the g reflection from the crystals on both sides of the APB. Before quantum-mechanical calculations for many-body problems became available in the early 1980s. and the composition and the degree of order are uniform throughout the domains.. Although such formulations have been largely replaced by modern ab initio quantum-mechanical . Theory and technique for the imaging of APBs have been well-established.p.p ^ integer and this is possible only if g is a superlattice reflection (i. Lasalmonie et al. Paidar.e.t 1969. when the fault plane is inclined to the foil surface. As for APBs coupling partial dislocations.. Leamy et al. The contrast formation of an APB is therefore similar to that of ordinary stacking faults. 2.2. 1988). Early calculations considered only the interaction energies among the nearest neighbors (e. Since in most cases the extinction distances of superlattice reflections are comparable to.Figure 2. this subject will be dealt with in Section 4. thermally formed APBs are often imaged as a single bright or dark band. The image vanishes when g. The Structure and Energy of APBs 2. theoretical estimates for the APB energy were based on the Ising model.. (b) An APB on (001) in which the bonding with the nearest neighbors is not disturbed. followed soon by more elaborate treatments that included higher-order interactions (Marcinkowski and Brown. Contrast is formed when g.p = integer.. p = y<l 10)» represented by the arrow. 1962. Yamaguchi et al.f 1987. 1981a. a reflection forbidden in the disordered counterpart). Transmission electron microscopy (TEM) has been used extensively for the observation of APBs and APDs. The reader is referred to Chapter 22 by Veyssiere and Douin in this volume for examples of TEM observations of APB-dissociated dislocations. and the reader is referred to the review by Marcinkowski (1963) for the imaging of thermal APBs and to the papers by Cockayne and coworkers (Cockayne et al. (c) An APB on (11 Ij in which a nearest B-B bond is created tube. (a) The crystal structure of Ll 2 . One of the key assumptions in such calculations is that the pairwise interaction energies are independent of the presence of other atoms. and the phase difference between the waves is given by 27rg. APBs are usually imaged using the method of dark-field imaging in which only one diffracted beam (denoted g) passes through the imaging lenses. it is usually impossible to separate the fault contrast from the interference from the strain field of the coupled dislocations. the images of which are sharp bright lines under the weak-beam condition. The atoms sit at ideal lattice positions right up to the boundary. a condition that is satisfied when g is a fundamental reflection or normal to p. the image consists of an oscillation of dark and light bands. 1970) for the weak-beam method. Ray et al. in which the increase in the internal energy due to the presence of an APB was derived from the changes in the interaction energies between atom pairs. which is a straight. or larger than. in this case the APB is defined as the space between the dislocations. usually trailing APB-dissociated superdislocations containing jogs. 1970. tubular APD. Beauchamp et al. 1985.1 Ideal APBs and APB Energy Anisotropy An ideal APB is a sharp geometrical plane that separates two identically oriented crystal domains that have been displaced rigidly relative to each other by p. Flinn. 1960). the foil thickness.

Yoshida and Takasugi. Korner..KAB is the ordering energy of the nearest neighbors. There are also Ll 2 alloys. which is replaced by an A atom. which is substituted by a B atom. In the B2-ordered structure. In Ll 2 . notably Ni3Fe (Calvayrac and Fayard 1972. e. In most of the existing models. However. (1984) observed in /3-CuZn that the APB energy on [1 10) was in fact higher than on (1 12). T>ou\n etal. the rise of the yield stress originates from the formation of KearWilsdorf locks on screw dislocations (Kear and Wilsdorf. the driving force of which derives partly from the lower APB energy on (0 0 1}. every atom has 12 nearest neighbors: an A atom has eight nearest B atoms and four nearest A atoms. The simple estimates based on the pairwise nearest neighbor interactions have been found satisfactory in accounting for the morphological features of thermal APDs observed in some Ll 2 alloys in which the APD structure has been described as maze-like with the APBs lying preferentially on {0 0 IJ planes: Cu3Au (Fisher and Marcinkowski. Such a configuration of bonding between the nearest atoms is maintained by a shear APB (p = < 1 1 0>/2) on the (0 0 1) plane (Figure 2(b)). which we shall use in later discussions. Saka et al. Dimiduk. twothirds of the A atoms are still surrounded by eight A atoms and four B atoms. Flinn's calculation (Flinn. which often showed comparable values of APB energies on (001) and {1 1 1) in Ni3Al (Stocks et al. the density of the same bond is given by /. Paidar et al.g. Sutcliffe and Jaumot. the newly created B-B bond is indicated in Figure 2(c). is the interaction energy between second nearest neighbors. 1989). while a B atom is completely surrounded by 12 A atoms. which by definition is the APB energy. 1960) predicted a higher APB energy on (1 1 2} than on (1 1 0) by a factor of 2/V3 . On this basis Flinn predicted that in Ll 2 the APB energy should be highly anisotropic with a maximum on [1 1 1) and a minimum on [0 0 1). 1980). Cu3Pt (Yodogawa et al. 1953). 1984.. in which the morphology of thermal APDs is 'swirl-like' with the APBs adopting random orientations.) In general. . t/. Here we take the Ll 2 (A3B) structure as an example. In a general plane n. 1987.. and in Co3Ti (Oliver. Fu and Yoo. can be shown to be given by Tlll =2F/(V3* 2 ) where V= \( Vp1^ + KBB) . Ni3Fe (Korner and Karnthaler. For an APB on the (111) plane. argued that the APB energies should be associated principally with nearest-neighbor interactions since second-neighbor interactions had been estimated to be only one-tenth as strong as between nearest neighbors (Cowley. see Pope and Ezz (1984) and Suzuki et al. L1) across an APB between atoms connected by a vector t. 1975). however. 1950. Recent weak-beam TEM observations have shown that even in Ni3Al. 1992). (Here Vtj. but one-third of the A atoms lose a B atom. and at the same time all the B atoms lose an A atom. 1973. Ni3Mn (Reynaud et al. 1989) (see Table 1). The APB energy anisotropy also forms the basis for the interpretation of the yield stress anomaly in some Ll 2 alloys (e. 1985). and a is the lattice constant. Ni3Al and Ni3Ga) in which the yield stress exhibits a significant increase with temperature (for reviews. 1976). 1986.. the [0 0 1} APB energy is often comparable to that on {1 1 l)(Veyssiereeftf/. This gives L1 = N0(I1M)ZM Tables 1-6 contain a list of energies of shear APBs expressed in terms of the ordering energies of the nearest neighbors (V) and the second nearest neighbors (W) (the formulas are taken from Yamaguchi and Umakoshi (1990)). Similar findings have also been made in Ni3Si (Tounsi.calculations. Paidar (1985) more recently included the interaction among second neighbors in Ll 2 and found that Flinn's result was correct for most of the boundary planes except those with low indices. 1991). who first derived analytical solutions for Ll 2 and B2 structures. shown in Figure 2(a).. with iJ = A9B9 is the interaction energy between nearest neighbors. times the cosine of the angle between n and t. the number of bonds per unit area (bond density. and Ni3Al (Horton and Liu. 1980). In these alloys the APB energies are thought to be largely isotropic. 1961). there is thus no nearest B-B bond. (1989)). and Pd3Fe (Yodogawa et al.. the bond density I1 in that plane is given by Ii = N0It1] where N0 is the number of atoms per unit volume. 1985).n/lnllt/l. Consider first the plane perpendicular to the bond direction t. 1988. Wti..g. in contradiction to the predictions from the nearestneighbor model. Flinn (1960). 1992). The experimental findings have been further substantiated by the latest quantum-mechanical calculations. Morris et al. which has a very pronounced yield stress anomaly. 1962. their merit still remains in that they provide a clear physical insight into the anisotropic nature of the APB energy. The energy increase per unit area. 1985.. Takeuchi and Kuramoto... can be calculated on the basis of the following consideration. Hirsch. 1988.

the compound is predicted to be always stable in the Ll 2 structure (Carlsson and Meschter.. 1985).. the {0 01} APB energy in Ll 2 would be negative (Kikuchi and Cahn. Structural relaxations have also been shown to contribute to reducing the degree of APB energy anisotropy in Ll 2 alloys (Beauchamp et al. 1979). for Al3Ti. is often associated with the formation of long-period tetragonal superlattice structures formed by introducing an APB each M[O 0 1} atom layer. In Al3Sc. alloys with low {001} APB energies are less stable with respect to DO22 or DO23 than those with high {001} APB energies (Wee and Suzuki. was accompanied by a decrease in the {001} APB energy. by increasing the tetragonality ratio (c/a). . If the interatomic interactions are included up to the fourth neighbor. Using potentials generated from the generalized perturbation method. Mishima et al. namely. 1991. Ll 2 are reflected in the {001} APB energy in Ll 2 . 1987). It has therefore often been argued that the relative stabilities of DO22 and DO23 vs. 1989). (1989) found that.2 APB Energy and Structural Stability Some superlattice structures are related to each other by introducing APBs periodically on specific crystallographic planes. The DO22 structure is derived from Ll 2 by forming an {001} APB on every two {001} atom layers. 2. Fu. (1980) observed in Cu 3 Au-Ni alloys that the occurrence of maze-like patterns of thermally formed APDs. The DO23 structure is related to Ll 2 by {001} APBs on every four {001} atom layers These departures are usually taken as indications for much stronger interactions among high-order neighbors. The DO22 structure is obtainable from Ll 2 by creating an APB on every second {001} atom layers. increasing the structural stability of DO22 against Ll 2 .L12 DO 2 2 DO23 Figure 3. the tetragonality being given by c/a = 2M. if the DO22 structure is stable. Nicholson et al. In both cases the compound composition remains A3B but the crystal structure has become tetragonal. the {001} APB energy has been shown to be equal to the difference in the internal energies between Ll 2 and DO22 (Nicholson et al. 1979. In the framework of pair-potential calculations. in which the calculated {001} APB energy is very high (-45OmJm" 2 ). suggesting a lower {001} APB energy. DO23 is derived from Ll 2 by forming an APB on every fourth {001} atom layer (Figure 3). 1990). Yodogawa et al.

from the linear relationship can be quite substantial in some alloys. 2.3 Deviations from Ideal APBs— Structural Effects y (mJ rrf ) Owing to the different bonding environment around an APB, atoms near the APB are expected to deviate from ideal lattice positions, as a result of which both the direction and the magnitude of the APB displacement vector p differ from that for an ideal APB. In general, the deviation vector 5p should vary as a function of the position relative to the APB. In most cases, however, a simplification is made in which the APB still separates two rigidly displaced crystals but with a displacement vector in the form p ' = p + 5p with 5p being a constant. Yamaguchi et al. (1981a) considered for the Ll 2 structure the supplementary displacement lying in the plane of shear APBs on [1 1 1) and [0 1 0). They showed that a rigorous determination of dp would require the construction of the 7 surface, which gives the variation of the (general) fault energy as a function of the displacement vector. The equilibrium displacement of the real APB corresponds to a local minimum in the 7 surface. However, as demonstrated by Yamaguchi et al. (1981a) the directions of such supplementary displacements can be predicted from the symmetry of the atom arrangement in the plane of the fault. In Ll 2 on the {1 1 1} plane, the introduction of a |<1 1 0> APB destroys the threefold rotational symmetry and the deviation vector 6p was shown to be always parallel to the <1 1 2> direction perpendicular to the ideal APB shear p. Figure 4 shows the section of the {1 1 1) 7 surface that contains the concerned < 1 1 2> direction; the curves were calculated using different A-B central-force potentials with C-I to C-6 representing increasingly larger ordering energies. The ideal APB is represented by R' at which there is no deviation in the < 1 1 2> direction. The figure shows that, even with moderate ordering energies, such as those given by C-I and C-2, the local minimum is already shifted towards Q', which marks the position of the stable super lattice intrinsic stacking fault (SISF). When the ordering energy is further increased, larger shifts are formed and eventually the {111} APB becomes completely unstable for C-6. In the {0 1 Oj plane, the fourfold rotation symmetry is maintained after the APB shear and there is thus no supplementary shear lying within the fault plane. Beauchamp et al. (1987), using

Figure 4. Section of Ll 2 {111) 7-surface containing the supplementary displacement parallel to <112>. R' is the position of ideal APB shear. C-I to C-6 are from potentials representing increasing A - B ordering energies (Reproduced by permission of Taylor & Francis Ltd from Yamaguchi et al., 1981a)

The APB energy has also been proposed to correlate with the stability of the ordered structure vs. the disordered structure, being reflected, in the Ising model, in the relationship between the APB energy and the critical temperature Tc of the disorder-order transition. Models for the disorder-order transformation based on the nearest-neighbor interactions have yielded in some structures a proportional relationship between the ordering energy V and the transition temperature r c , i.e. V=kTc/w where k is the Boltzmann constant and w is a constant that depends on the structure and on the model used (for a review, see Guttman (1956), and also see Tables 1-6 for a list of the relationships between Tc and V and W in six of the most common ordered structures). The ordering energy thus obtained can be directly substituted into the analytical expressions for the APB energies (e.g. Flinn, 1960; Marcinkowski and Brown, 1962), leading to a proportional relationship between 7 and Tc of the form ya2<xTc (see Section 2.5 for more details). Recently, Sparks et al. (1991) have used the ordering energy derived from Tc as an indication for the relative phase stability among Ll 2 , DO22, and DO23. It is on this basis that it is generally believed that, the higher the order-disorder transition temperature, the higher the APB energy. Experimental measurements (see Section 2.5 and Figure 7) show a general trend that is roughly in keeping with the above relationship, but deviations

Figure 5. TEM dark-field images of thermal (010} APB in Fe3Ge. Electron-beam direction is [001]. (a) Superlattice reflection g= 100, and the [010] APB (indicated by arrow) is in strong contrast, (b) Fundamental reflection g= 111, and the APB is in residual contrast (From Ngan, 1992)

pair potentials for Ni3Al similar to those used by Yamaguchi et al. (1981a), calculated the supplementary displacements parallel to shear APBs lying on a range of crystallographic planes and < p was found at 5 maximum on (1 1 1) with a magnitude approximately equal to 0.06a. Yamaguchi et al. (1981b) also considered the supplementary shears lying within the fault planes in b.c.c. ordered structures (B2, DO3, and Ulx). In the B2 structure, there is no supplementary displacement in the |< 1 1 1 > APB on either (1 1 0) or {1 1 2) planes. In DO3 and L2j the supplementary shears are parallel to <00 1> on {1 1 0) and to <1 1 1> on (1 1 2). In the Ll 0 structure, deviations along < 1 1 2> directions have been predicted for APBs on (1 1 Ij planes and, similar to the Ll 2 structure, no deviation is expected in APBs on (10 0) (Yamaguchi et al., 1985). In general, a supplementary displacement lying within the fault plane is expected if the crystal rotation symmetry is destroyed by the introduction of the APB, while supplementary

displacement normal to the fault plane should occur in all cases. Beauchamp et al. (1987) found that the effect of relaxations was to reduce significantly the degree of APB energy anisotropy estimated from the nearestneighbor interactions (e.g. Flinn, 1960). Recently, more sophisticated theoretical calculations have been made for the APB energies, but there has so far been little application of the modern techniques to the extent of structural deviations from ideal APBs. Fu and Yoo (1992), using the full-potential linearized augmented plane-wave (F-LAPW) method, found in NiAl that allowing for relaxation normal to the APB would significantly reduce the APB energies. In NiAl, the (1 1 0) APB energy (7110) is 1000 mJ m~2 without relaxation, but decreases to 81OmJm" 2 after allowing for relaxation. Supplementary displacement at the APBs has been observed experimentally in TEM in a number of alloys. In TEM an APB is visible if g.p ^ integer and the

contrast vanishes when g.p = integer, the latter happens when g is a fundamental reflection or perpendicular to p. Owing to the supplementary displacement 5p, g.p ^integer even in fundamental reflections or in those normal to the ideal APB shear p, thus enabling the identification of the direction and magnitude of 5p from the residual contrast. Figure 5 shows an example of the residual contrast of thermal (0 1 O APB observed in J Ll 2 Fe3Ge (Ngan, 1992). Figure 5(a) is a dark-field image taken with the 10 0 superlattice reflection in which the (0 1 0) APB (indicated by an arrow) is in strong contrast (the other fringe patterns are due to superlattice stacking faults). In Figure 5(b), taken with the 111 fundamental reflection, the APB contrast is much weaker but still visible (i.e. showing residual contrast). Comparisons with simulated images show that the residual contrast is formed by a supplementary displacement normal to the APB with a maximum amplitude of the order of 0.3 A. Lasalmonie et al. (1988) studied the residual contrast of thermal APBs in rapidly solidified Ni3Al and found that for a non-conservative APB on [1 0 Ij the residual contrast could be accounted for by a supplementary displacement containing components both parallel and normal to the fault plane. The component normal to the fault plane was found to be of the order of 10~2a with the component parallel to the APB being much smaller. Thermal APBs in rapidly solidified B2-ordered Fe3Al have also been found to exhibit residual contrast (Prakash et al., 1991), but in this case the additional displacement was thought to be caused by impurities segregated to the APB. 2.4 Deviations from Ideal APBs— Chemical Effects Even after allowing for structural relaxations as shown above, an APB is still thermodynamically unstable since it involves an increase in the internal energy but with very little extra entropy to compensate. Therefore, when thermal fluctuations permit, thermodynamic equilibrium configurations would invoke an increase in the degree of randomness near the boundary, which may appear either in a reduction in the local degree of order or in the segregation of impurity or excess (when composition is not exactly stoichiometric) atoms to the boundary. The first treatment was given for a stoichiometric B2-ordered alloy by Brown (1959), who considered the situation in which the entropy increase was effected by a decrease in the degree of order at the APB. On the basis of the Bragg-Williams approximation (Bragg and Williams, 1934, 1935), the

degree of order was found to increase with the vertical distance from the boundary, and the thickness of the APB was defined as the distance at which the degree of order becomes essentially the same as that in the bulk. The thickness of the APB is very small except very close to the transition temperature Tc. At 0.9757;, the thickness of the APB is about 10 interplanar spacings. The disordering at APBs in the Ll2-ordered structure was investigated by Popov et al. (1966), also using the Bragg-Williams approach. They found that disordering within the first atomic layer adjacent to the APB started at a temperature as low as 0.2JTC. Similar to the finding of Brown in B2 alloys, disordering was found to extend to approximately 10 atomic layers only when the temperature is just a few degrees below Tc. A more elaborate investigation of the equilibrium structure of APBs in the Ll2-ordered structure was made by Kikuchi and Cahn (1979) using the tetrahedron approximation of the cluster-variation method. They found that a distinct disordered phase may coat the APB, the effect (termed 'wetting') being more pronounced in alloys of non-stoichiometric compositions. The above findings have been further substantiated by Sanchez et al. (1987), who, using a treatment similar to Kikuchi and Cahn, showed in more detail how the APBs get more diffuse when the temperature approaches that of the transition. In addition, local disordering around the APB was shown to involve an increase near the APB in the concentration of the majority atoms (an increase of 1.2%, from 75 to 76.2%). For off-stoichiometric compositions the disordering is more pronounced, with greater segregation of majority atoms to the APB (increase of up to 6.9%). Indirect experimental evidence for the thermal disordering at APBs has been provided by TEM observations in which the separation of the APBcoupled partial dislocations was found to increase with annealing temperature, suggesting a lowering of APB energy with temperature. Models for the disordering at the APBs (Popov et al., 1966; Kikuchi and Cahn, 1979) predicted that a decrease in the local degree of order should bring about a corresponding drop in the surface energy. Popov et al. (1966) showed that for the Ll 2 structure the {1 1 1) APB energy would decrease to about 50% of the low-temperature value when the temperature was raised to -0.6T c ; at this temperature the [0 0 IJ APB energy was shown by Kikuchi and Cahn (1979) to exhibit a similar decrease. The decrease of the APB energy with temperature has been observed in a number of ordered alloys, using in situ TEM observations in which the width of the APB is monitored during dynamic



Figure 6. Variation of APB energies with temperature observed in some Ll 2 compounds. For Ni3Al the measurements were made for {001} APB. Curve 1 was obtained using weak-beam TEM (Veyssiere et al., 1989) and curve 2 was a measurement using lattice-resolution TEM (Crimp, 1989). For other compounds the APB is on {11 Ij. The APB energies are in fractions of the value at room temperature (70)

annealing, or using post mortem examinations. Morris and Smallman (1975) measured the width of (1 1 1) APB in Ll 2 Cu3Au (7; = 663 K) samples that were deformed and subsequently annealed at 623 K (0.94Tc) for 5-15 min. From freshly deformed samples, the APB energy was measured at about 48mJm~ 2 . In subsequently annealed samples, however, the APB energy was found to have decreased by about 56%, to 27mJm~ 2 . Veyssiere et al. (1989), using the weakbeam method, observed the variation of the (001) APB energy in Ni3Al in samples deformed at three different temperatures, and the energy was found to have decreased from 126mJm~2 at room temperature (0.187;) to 72mJm~ 2 at HOOK (0.667;). (Ni3Al is known to remain ordered up to the melting point and here Tc is taken as the melting temperature.) The decrease of (0 0 1) APB energy with temperature in Ni3Al was soon confirmed by Crimp (1989) using lattice-resolution TEM. In Ni3Fe (7; = 776 K), Korner and Schoeck (1990) observed the lowering of {1 1 1) APB energy with temperature using the TEM in situ annealing

technique, and found that the [111) APB energy was halved at 0.847;. In situ annealing experiments have also been made in Ll2-ordered Al3Ti containing Fe and Mn by Morris (1992a,b), who reported similar decreases in the (1 1 1) APB energy. Results of the equilibrium APB energies as a function of temperature in some Ll 2 alloys are plotted in Figure 6, in which the APB energy at room temperature has been taken as unity and the temperatures are represented in fractions of the transition temperature Tc. Segregation of vacancies to APBs has also been shown to contribute to the increase in the compensating entropy as well as to a lowering in the internal energy. In quenched 0-CuZn with the B2 ordered structure, Cupschalk and ,Brown (1968) observed in TEM a large density of dislocation loops decorating thermal APBs. The dislocation loops were thought to be formed by the collapse of disks of vacancies that had preferentially migrated to the APBs. In addition to a contribution to an increase in the entropy term, Cupschalk and Brown showed that vacancies sitting at the APB in the B2

whereas the third one is fixed. (1961) used the APB energy obtained from T0 to calculate the equilibrium separation of the superpartials in Cu3Au and found that the predicted separation (78-124A) was just slightly lower than the experimental measurement using bright-field TEM imaging (90-135 A).) in some of the tables indicates that all permutations of the first two indices are allowed. In Tables 1-6 values derived from the transition temperatures are referred to as being obtained by method T0. 1991). Marcinkowski et al. 1970). Umakoshi et al.0 (Peierls. 1971). In Ll 0 . the APB energy y is related to the ordering energy V via y = F(h. and grain boundaries.. 1949. There has been little direct experimental observation of segregation of alloying elements to APBs. using the atom probe method.73 (Li. Crawford. Vacancies preferentially concentrated at the APB were shown by Cupschalk and Brown to form an important mechanism controlling the kinetics of the migration of thermal APBs. T0 = wV/k. in Ll 2 . 1961. In the DO3 structure (A3B) the possibility of an additional transition of DO3 to B2 (e. but at the APB there are m dangling bonds between AA or BB atoms and 8 — m dangling bonds between AB atoms. 1945). giving w = 0.k. The major finding of the work is that the concentration of the elements at the APB depends on the type of sublattice that the ternary elements occupy in the bulk alloy. 1971). there is a depletion of B atoms and an enrichment of A atoms if the ternary elements preferentially occupy A sublattices. implying that the predicted energy is somewhat higher than the true value.. to 1. We thus have the following relationship between y and T0: k y = —.kJ)S2^ T0 w In Ll2-ordered Cu3Au. Cu3Au and CuAu (Marcinkowski et al.l) gives the orientation dependence.* in which APB energies in compounds with the six most common ordered structures are listed. .0 (Murakami et al. 1966). Potter. For the Ll 2 structure the variation of yina2 with T0 is represented by the two straight lines in Figure 7 corresponding to w = 0. Leamy et al. At the (1 1 Ij APB there is a depletion of A atoms and an enrichment of B atoms if the ternary elements preferentially occupy the B sublattice. 1986) and FeAl (Mendiratta and Law. WBC = kT0/3. In the nearest-neighbor approximation. (1989. the transition temperature T0 is also related to V (and to W in DO3 and L2^ through a proportionality factor. 1991). 1974).1 to 4. a similar relationship was obtained. The relationships of V and W with T0 are summarized in Tables 1-6. 1950).0 (for a review. the critical temperature is related to the second-nearestneighbor ordering energy between B and C atoms.g. in Fe3Al) enables both V and W to be determined (Rudman. .structure had lower self-energies. 1960. Co2AlTi (Umakoshi and Yamaguchi. 1969/70). In the L2j structure with A2BC composition. the agreement between the above theoretical estimate and the experimental measurement from the separation of the APB-coupled superpartials is reasonable.82 (Yang. 1956). 2. observed in B-doped Ni3Al the segregation of B to APB as well as to dislocations. conversely.) In the Ising-type models for the disorder-order transition.. NiAl (Ball and Smallman. where k is the Boltzmann constant and w is a constant that depends on the structure and on the models used. and S is the long-range order parameter (Marcinkowski.k.82 *The notation <. 1956. Sastry and Ramaswami. a is the lattice constant..l)^S2 a2 where F(h. (In DO3 and \2X < 1 0 0> APB energies are related to the ordering energy W of the second-nearest neighbors. see Guttman.. w varies from 0. 1987).5 (Cowley. 1936). stacking faults. A vacancy in B2 in the bulk is associated with the elimination of eight nearest AB bonds (dangling bonds). Segregation of alloying elements to {1 1 1) APBs in a model Ll 2 (A3B) compound containing ternary elements has been investigated theoretically by Wu et al. In the B2 structure w varies from 3. 1976). a1 F(h. Strengthening models based on the above segregation study were found consistent with the strengthening of Ni3Al by ternary elements measured experimentally (Wu et al. Horton and Miller (1987). Similar findings have also been made in Cu3Au by Sastry and Ramaswami (1976) using the weak-beam dark-field method with a higher resolution limit. This approach has also been used more recently: Ni2AlTi. . and TiAl (Shechtman et al. m being dependent on the orientation of the APB and the specific location in the APB at which the vacancy sits. see Tables 5 and 6.] and {. . 1985. 1. amounting to a reduction of m V in the vacancy selfenergy. .5 Energy of APBs Early estimates for the surface energy of the APB in ordered alloys were based on the ordering energies V and W obtained from their linear relationships with the transition temperature T0: B2 (Rachinger and Cottrell. 1950). 1963..

8 Al.9Sn 21. (1989) Crimp (1989) Dimiduk (1989) Dimiduk (1989) Dimiduk (1989) Dimiduk (1989) Dimiduk (1989) Dimiduk (1989) Dimiduk (1989) Baluc (1990) Neveu (1991) Neveu (1991) Chen et al. SISF. 0. no relaxation allowed at APB.9 Al 24.)c V=IcT0Zw.0 (Peierls. References Veyssiere et al. *n.7 Al. d Deduced from instability condition of APB vs.0 V 24. 1. (1989) 112-205 310 270 >430 100-184 670 Morris (1992a) Morris (1992b) Inui et al.24 B 23.5 (Cowley. (1992) 450 Fu (1990) See text and/or list of acronyms. fault energy determined using isotropic elasticity.8 Au 24. 10.Table 1.7Sn 23.9 Si. (1986) Foiles and Daw (1987) Stocks et al.) WB(iso.6Hf Ni3Al Ni3Al Ni3Al Ni3Al Ni3Al Ni3Ga Ni3Ga Ni3Mn 22.%) Ni3Al -Ni 3 Al -Ni 3 Al 17. 1.4 Al. 1.1 Al..82 (Yang. 4.7 Al. (1961) Morris and Smallman (1975) Sastry and Ramaswami (1976) Tounsi (1988) Tounsi (1988) Yoshida and Takasugi (1991) 119-218 270 155 210 130 282 rc WB WB LMTO Tc WB WB Ded/ T0 F-LAPW(n) Oliver (1992) Oliver (1992) Pei et al. 1. (1985) Douin et al. ym= -2W/a\ Method a Tc WB WB WB WB HREM WB WB WB WB WB WB WB WB WB WB EAM(n)* EAM(n) ^ » LMTO(n) F-LAPW(n) WB(iso. 1945). (1987) Fu and Yoo (1989) Tm (mJm 2) 139-255 180 ±30 (350 0 C) 250 ±30 169+19 163 + 21 190 ±26 166+18 174 ±20 192 + 26 198 + 32 165 150 + 20 190 + 20 142 96 7010 (mJ m 2) 140 (350 0 C) 90±5 250 + 30 126 149± 12 104 + 8 122+11 170±21 129 ±12 146± 15 155 + 17 201+32 155 120 ±20 170 + 20 83 28 197 230 140 220 123-226 110±30 62-114 75 65-119 92 50-92 80 + 5 48 + 5 Suzuki et al. .16Ti 24 Al.85 Au -Cu 3 Au Ni3Si 23.1 Al. w = 0. 3 Ni 22Ti Co3Ti Al3Ti 28 Ti. 1950) Composition (at.0 V 20. 1.5 Si 10.7 Ti Co3Ti 23 Ti. 1936). c iso.9 Al 24. (1986) Korner (1988) Veyssiere et al.26 Hf 22.1Si 24.7 Al.) Tc WB Marcinkowski and Miller (1961) Korner and Karnthaler (1985) 55 T0 39±5 123-225 220 250 220 250 124 TEM(iso.) Tc WB WB WB Marcinkowski et al.0Ta 22. 8 Fe -Al 5 Ti 2 Mn -Al 3 Ti(Fe) Al3Sc Al3Sc a 7lll = (2V-6W)/V3a2. 0. 0.5 Al 22. 6.7 Mn Ni3Fe -Ni 3 Fe Cu3Au 24. 3.4 Al. APB energies in Ll 2 compounds: p = KH0>.) WB(iso. (1987) Stocks et al.2 Al 25. (1979) 17±5 Tc TEM(iso.

yim= and Umakoshi. 1949. (1991) Fu and Yoo (1992) Fu and Yoo (1992) Rudy and Sauthoff (1986) Clapp et al.73 (Li. c Tc WB See text and/or list of acronyms.W)/45a2y ylm= -2W/a2. Pn = KOl I ] .396. 4. V=kTc/wy w = 0. APB energies in DO22 structure: P1 = W H ] . (1991) Fu and Yoo (1992) Fu and Yoo (1992) Rao et al (1991) Fu and Yoo (1990) Fu and Yoo (1990) Fu and Yoo (1990) Fu and Yoo (1990) Freeman et al (1991) Rudy and Sauthoff (1986) Crawford and Ray (1977) T110 (mJ m 2) 82-107 40 50+10 7ll2 (mJm. ^n.12W)ZVIa2. c . 1956) Composition (at.6 Zn 47. T111 =(4 K . no relaxation allowed at APB. Table 4.0 (Guttman. T i n = 2 ( K . Method* Tc TEM WB WB Tc WB Est. * Calculated using pair potential. relation with Tc not known Composition (at.Table 2. 1990). no relaxation allowed at APB. (Yamaguchi and Umakoshi. APB energies in Ll 0 compounds: p = K 0 1 1 ] . Ordering energy estimated from Monte Carlo simulation.W)/V6a2 (Yamaguchi and Umakoshi. 1950) Composition (at.%) CuZn 48.%) -Al 3 Ti "See text.W)/\lla\ V=kTc/wy w = 3.%) TiAl -TiAl TiAl TiAl TiAl TiAl a -AW/a2 (assuming c/a=\) (Yamaguchi T111 ( m J m 477 145 + 15 520 670 510 306 2 ) T1JOO) ( m J m 2 ) Method 0 Tc WB CaIc* LKKR(n) c F-LAPW(n) EAM(n) References Hug et al (1989) Yamaguchi et al (1985) Woodward et al (1991) Fu and Yoo (1990) Rao et al (1991) 100 (600 0 C) 435 347 430 66 See text and/or list of acronyms. = (2V-4W)/y/3a2 V111 (mJm 2) 200 (400 C) 0 T1QOi ( m J m 2 ) Method 0 WB Reference Hug et al (1989) 25 (400 C) 0 T112 = 8 ( K . c EAM LMTO(n) EAM(n) LMTO(n) LMTO LMTO(n) LMTO(n) LMTO(n) LMTO(n) LMTO(n) LMTO References Head et al (1967) Saka et al (1984) Dirras et al (1992) Crawford and Ray (1977) Rudy and Sauthoff (1986) Freeman et al. c F-LAPW(n)* F-LAPW(n) F-LAPW Tc Est. 1990).076. Table 3. c n.0 Zn 46Zn FeAl FeAl FeAl FeAl FeAl FeAl NiAl NiAl NiAl NiAl NiAl NiAl NiAl Ni2MnAl3 Ni2CrAl3 Ni2VAl3 Ni3Al2Cr Ni3Al2V AlRu AlRu CoAl CoAl Fe3Al Fe3Al fl b T". (1989) Freeman et al (1991) Rao et al.2 ) 95-124 57±7 182-237 230* 160 500 650 300 224-291 400 240 1000 425 1000 810 740 510 250 250 550 248-322 690 228-297 360 45-60 15" 39±7 62 ± 210-273 1040 820 260-336 380 950 460 1230 990 286-372 263-343 52-68 Te Est. 3. 3. Extrapolated value. 7 l l 0 = 4 ( K . 1990).278. APB energies in B2 structure: p = |<l 10>.

x^ 0. Crawford. 83 mJ m~2 by Chen et al. however. W=}kTc when TC=TC\ T'c'\s the temperature of transition of DO3->B2 (Rudman. (Yang. y\n = {2E{-4E2)/<6a2. 7J 10 = V/yfla2y y\n = 2V/V6a\ y\Q = 2W/<2a2. P11 = <111> «100».. (A2BC) structure: P1 = JO11>. 1950) respectively. E1 = 2F AB + 2K A C . 1991).) References Umakoshi and Yamaguchi (1985) Umakoshi et al. the {111} APB energy in Ni3Al was calculated to be 142 mJm. 197 mJ m~2. 1977). the APB energy is affected by the long-range order parameter S via y = y0S2. 1971) Composition (at. this is to be compared with values obtained experimentally (full circles) and with those determined by atomistic calculations (open circles). 1960.5.2 ) 132 115 80 Method 0 WB(iso. Crawford. 24. a similar discrepancy is also found in the {001} APB energy. Dimiduk (1989) measured the APB energies in binary Ni3Al at three compositions containing 22. Much higher values of {0 0 1} APB energy in Ni3Al were obtained by Stocks et al.e. Tables 1-6 contain results of APB energies calculated by these methods.F BC .Table 5. The reasons for the discrepancies in the theoretical calculations are not yet . *iso. the first two compositions were found to have similar (111) APB energies. In theoretical treatments based on the Ising model. using the interchange potential method. 1986. x ^ 0. For example. For details of atomistic calculation.n T i 0 8 6 7J10 (mJ m 2) 7! 1 J 0 (IIiJm. 1990). the APB energy decreases with decreasing concentration of the minority component. and linear muffin-tin orbitals (LMTO) method (Hong and Freeman.2 ) 43 100* 169 225 Method 0 WB Tc TEM(iso. in which the maximum order parameter SM depends on the composition. employing a number of techniques: embedded-atom method (EAM) (Chen et al.03 Al i. APB energy in L2. 7I10 = OE1 -2E2)/\f2a2. both using the embedded-atom method.% Al to 3OmJm" 2 at 27at.%) Fe3Al Fe3Al Fe3Si Fe3Si a b c Yi10 ( m J m .2 by Chen et al. more sophisticated theoretical calculations of APB energies have been made on computers. 1970) and thus y=^^2.) c References Crawford and Ray (1977) Lakso and Macinkowski (1969) See text and/or list of acronyms. Such a relationship can be used to estimate the effect of composition on the APB energy in nonstoichiometric alloys. E2 = 2WAB + 2WAC. 1971).2 ) 60 70 b 126 7J1J0 ( m J m ..) 6 WB(iso. and 25. Foiles and Daw. iso.5 (Cowley. see Chapters 3 and 4 by Carlsson and Meschter and by Voter in this volume. calculated using isotropic elasticity. In the Ll 2 structure with composition A^ x B x . (1986) and 96mJm~ 2 by Foiles and Daw (1987). 1990). WBC = kTc/3. the maximum degree of long-range order is given by SM = 2x and thus y = 4yox^ (Leamy et al. 1990. There is as yet. (1986) -Co 2 AlTi "See text.25. 7Ji2 = 4^BC/V6flr2. V=}k(T(!+2Tc). 1945) and w= 1.% Al. 7J110 = 2WBC/V2flr2. 1970. Extrapolated value. 1971) Composition (at. In the B2 compound with composition A 1 ^B x .. 1987). 7^12 = 4W/V6a2 (Yamaguchi and Umakoshi.%) Ni2AlTi NI2.. (1987). full-potential linearized augmented plane-wave (FLAPW) method (Fu. W=$kTc' when T'C<TC\ V=}kTc. the APB energy is expected to decrease with the decrease in the degree of long-range order.9 at. where y0 is the APB energy for the case of perfect order. Such a trend was found experimentally in B2-ordered Fe3Al in which the {1 1 Oj APB energy was found to decrease from 11OmJm -2 at 36at.0 (Murakami et ai. and still higher at 230 mJ m~2 by LMTO. Table 6. 1990)..°7o Al (Crawford and Ray. Fu and Yoo.9. the maximum degree of long-range order is given by SM = 4x (Leamy et al. Recently. and 28 mJ m" 2 by Foiles and Daw. i. but the APB energy in the third composition is higher by about 20-30 mJ m~2. fault energy determined using isotropic elasticity. P11 = <111> «100».WBC (Yamaguchi and Umakoshi. Discrepancies of similar magnitude have also been found in calculated APB energies in B2 compounds. no consistent agreement between the results given by the different techniques or by different workers using the same technique.2. APB energy in DO3 compounds: P1 = I d 11>.

5 clear. which has a rather low resolution limit (Marcinkowski. Early measurements were made by TEM bright-field or darkfield imaging under the two-beam condition. If relaxation is the primary source of the discrepancies. The (11 Ij APB energies in Ll 2 alloys plotted against the disorder-order transition temperature Tc. which is a function of the dislocation character and the elastic constants.. The weak-beam technique developed by Cockayne et al. * The experimental determination of the APB energies is usually made using TEM by measuring the separation between APB-coupled superpartials.g. (1969) has a resolution limit of the order of 10-15 A and has been widely used for the measurement of APB energies. Head et al. Fu and Yoo (1992) considered the effect of structural relaxation normal to the fault plane on APB energies in FeAl and NiAl. 1963. structural relaxation within the fault plane is not expected on either (1 1 0} or (1 12) owing to crystal symmetry in these two planes. 1967). which is related to the APB energy y via K(6)b2 T=-1 2irr where b is the Burgers vector of the superpartials bounding the APB and K(O) is the energy factor. Full circles represent values determined experimentally using TEM. They found that substantial reductions in the APB energies were obtained as a result of allowing for relaxation. 1987. One possible reason could be that very often structural relaxations at the APBs have not been allowed for in the calculations (e.. while in general relaxation occurs both normal and parallel to the fault plane. the lower values calculated by including relaxation are the more realistic ones. 1991). The key assumption in this technique is that the APB-coupled superpartials are in equilibrium separation r. Chen et al. Stocks et al.. 1986. The {1 1 2} APB energies ^ 112 ) decreased from 1230 and 104OmJm-2 to 990 and 82OmJm"2 for FeAl and NiAl respectively.82 and w= 1. In general the line image of a dislocation in . It should be noted that for |< 1 1 1 > APB in the B2 Z (10"20J) rc do 2 K) Figure 7. Hong and Freeman. Open circles represent values given by atomistic calculations. as well as surface energies of ordinary stacking faults. The two straight lines are energies obtained from the nearest-neighbor interaction with the ordering energies derived from Tc using w = 0.

When there are four or more APB variants available.g. there are two basic types of metastable three-dimensional domain structures. e.e. their line images are expected to be shifted in the same direction by approximately the same amount. multiply connected 'pipes' of one domain weave through the primary domain. (1988). Since. Depending on the number of APB variants available. and Horton and Liu (1985). a case that has been known since Bragg (1940). the merging of randomly nucleated ordered domains would lead either to isolated 'islands' of one domain embedded in another domain. screw and edge dislocations on (0 0 1} in Ll2 compounds (Sun et al. . Both of the above two types of three-dimensional domain structures have been observed experimentally: e. Experimental measurements using the weak-beam method are often subjected to about 10-20% fluctuations. the transformation is termed 'neostructural' (Tanner and Leamy. The morphology of APDs in isostructural ordering transitions is usually well-defined and the preferential orientation of the APBs. for which several factors may be responsible.g.' in Tables 1-6). One is that the dislocations may not be always in the equilibrium configuration. 3. the post-ordering microstructure can be described as a three-dimensional mosaic or foam built up by the domains. Lattice-resolution TEM.g. but they did not extend their treatment to the double-curved pipes in three dimensions.1 APD Morphology and Crystallography Owing to the surface energy of the APB. Fisher and Marcinkowski (1961). but some of the early work was based on isotropic elasticity (marked 'iso. in which intertwining. and Marcinkowski and Brown (1962). but it is possible to produce a metastable threedimensional domain structure in which the annihilation of the domains becomes very slow. i.b. 1968). The low annihilation rate results from the surface tension from the two principal curvatures driving the APB in opposite directions. in most cases. in the ordering of many Ll 0 and DO22 structures. Lasalmonie et al. Cupschalk and Brown (1967). in which the dislocations are imaged end-on in a thin foil. although a pentagonal dodecahedron is a close approximation. may render the domain structure metastable since the surface of the connecting pipes is double-curved or saddle-shaped. in this configuration the surface tensions are fully balanced.TEM does not coincide exactly with its true position and. which would cause the fault energy to vary with position. known as the 'Swiss-cheese' structure (English. 1974). the partial dislocations bounding the APB have the same Burgers vector. 1966). and at corners in groups of four (Bragg. The latter structure. and Crawford (1971). the two principal radii of curvature are opposite in sign (Cupschalk and Brown. schematically illustrated in Figure 8 (English. Another factor that may cause fluctuations in the separation between partial dislocations is inhomogeneous element distribution. Cupschalk and Brown (1968) examined the rate of the flattening of wavy APB and found that it was much slower than the shrinkage of an isolated spherical domain. provides a much higher resolution power (< 2 A) than the weakbeam technique.9 1991)). When there are only two APD variants available. Another crystallographic factor that affects the morphology of thermal APDs is whether or not the crystal system changes during the disorder-order transition. and this is particularly likely for dislocations that are liable to transform into locked structures (e. In a three-dimensional space there is no regular polyhedron that can be stacked with the plane faces meeting at the above angles. or the lack of it. The 'island' structure is highly unstable with a relatively high shrinking rate (Cupschalk and Brown. or to a configuration. analogous in morphology to a polycrystalline aggregate of grains. In most of the measurements the energy factor K(B) was calculated using anisotropic elasticity. Grown-in APBs and APDs 3. If ordering results in a change in the crystal symmetry. in B2 with two variants. The difficulty of the lattice imaging method lies in that specimen preparation is more difficult and that in general a very high dislocation density has to be introduced by heavy deformation. e.g. The structure becomes metastable when the domains meet along lines in groups of three at mutual angles of 120°. the true separation can therefore be measured directly from the image. in Ll 2 compounds with four variants. The advantage of the weak-beam method is that the APB width can be observed over a wide range of dislocation characters (see Chapter 22 by Veyssiere and Douin in this volume for examples of TEM observations of dislocations). 1966). shifts of the order of 10 A are predicted. Tables 1 -6 contain results of APB energies measured using the weak-beam (WB) method. where g is the reflection used for the imaging. the direction of the shift being dependent on the sign of g. 1940). This technique was used by Crimp (1989) to measure the (0 0 1) APB energy in Ni3Al. 1968). in the ordering of most B2 and Ll 2 alloys from corresponding cubic disordered systems. The transformation is termed 'isostructuraF if the crystal system remains the same upon ordering. under the weak-beam condition. the thermally formed APDs should vanish when true equilibrium is attained.

and finally the disordered phase diminishes and the ordered domains come into contact. followed by a stage in which the ordered and disordered phases coexist. e. twinning in Ll0-ordered CuAu and TiAl (Tanner and Leamy. 1974. Often the lattice strain caused by the miscrientation in neostructural changes can be accommodated by deformation. e. 1990).g. can have different orientations in different domains. a publication of the Minerals. Yamaguchi and Umakoshi.2 APDs and Disorder-Order Transition The presence of APDs as a result of cooling through the disorder-order transition temperature is usually taken as an indication that the transition is a first-order phase change involving a latent heat. 236. the domain structure is more complicated because unique axes.Figure 8. 14-18 (1966).g. Pennsylvania 15086) can usually be interpreted on the basis of the APB energy anisotropy. Warrendale. the tetragonal axis in Ll 0 . Such a transition starts with the ordered domains being nucleated independently in an existing solid disordered crystal. 3. Metals & Materials Society. The * Swiss-cheese' structure of thermal APDs in compounds with two APB variants (English 1966) (Reprinted with permission from Transactions of the Metallurgical Society. . In neostructural transitions. as a result of which neighboring domains may be misoriented.

Cupschalk and Brown (1968) proposed a mechanism for the growth of APD in which the atom diffusion controlling the domain migration is essentially the same as the bulk diffusion of substitutional atoms. 1990). and the disordering starts with the APBs being coated (wetted) by a layer of the disordered phase. in which the activation energy for domain growth was found to be very close to that for bulk diffusion (Poquette and Mikkola. The vacancies preferentially segregated to the APB. Thermal APDs form as a result of the suppression of domain growth by the rapid cooling rate. 4. The main difference is that a misorientation exists at a grain boundary while crystallographic directions are maintained upon crossing an APB. at least in the case of an ideal APB. Ordering occurring in an existing crystal was termed 'sequential ordering' by Cahn (1987) to contrast with 'direct ordering' in which the alloy freezes directly into the ordered state. for example. Horton and Liu (1985). appears to contradict the observation that. Cahn emphasized that this distinction was particularly important when the disorder-order transition temperature Tc is very close to the melting point. and that thermal APD should form only in sequential ordering. b is the Burgers vector of the bordering superpartials. This view. Compositional inhomogeneity brought about by. the role of the existing APBs is to act as preferential nucleation sites for the disordered phase. which grew when the temperature was brought very close to Tc. The investigation of Kikuchi and Cahn (1979) using the cluster-variation method showed that the wetting of the APB with the disordered or nearly disordered phase would start well before reaching the true transition temperature. and the migration of the APB is therefore expected to be much slower than grain growth. Morris. The reduction in the self-energy due to APB dissociation is given by E=In(T0) -yr0 where K(O) is the energy factor. 1974). based their interpretation of APDs in rapidly solidified Ni3Al alloys on the swiftness of domain growth in this material. 1969). the selection of the active dislocation system is based on competitions in both the dislocation self-energy and dislocation mobility. In continuous transitions the domain structure is expected to disappear upon reaching the critical temperature. APDs may also be formed but the physical picture is not as clear (Tanner and Leamy. The growth of APDs is in many ways similar to the growth of grains in that both are driven by the reduction in the curvature of the boundary. which has its Tc near to or above the melting point and in which thermal APDs are not normally observed. and in this sense the role of the APB is not very different from other interfaces. in Ni3Al. It is possible that the absence of a three-dimensional domain structure in a given material may result from both strong ordering and the rapid domain growth when the transition temperature is very high. on the other hand. grain boundaries or free surfaces. Wetting was found particularly pronounced at nonconservative APBs. e. namely an atom makes a jump only if it is adjacent to a vacancy. enable the boundary migration with short-range diffusion. In first-order or discontinuous transitions. According to Cahn (1987). and r0 is the equilibrium . wetting with the disordered phase at APBs and other interfaces has been directly observed using TEM (Tichelaar et al.g. the domains always retain a certain degree of long-range order even at the transition temperature.the boundary between them being APBs. (1992) observed the wetting of APBs in Cu3Au by the TEM in situ annealing technique and found that the wetted regions became nucleation sites for the disordered phase.1 Stability of APB Dissociation In a given material. 1992. In disordering. 1968). Leroux et aL. which were observed experimentally in quenched /3-CuZn in the form of collapsed vacancy disks at the APB (Cupschalk and Brown. In higherorder or continuous transitions. Morris (1992b) has shown recently that the bulk composition of the alloy may significantly change the APD structure in rapidly solidified Ni3Al in which the element distribution was found to vary substantially across the grain structure. The relatively loose structure at the grain boundary enables easy interchange of atoms during diffusion. dendrite formation during solidification has also been found to affect the domain structure. 1985). while at the APB the atomic registry is maintained. APBs in Dissociated Dislocations 4. 1975. Tichelaar et al. thermal APDs are formed most readily in rapidly solidified samples (Horton and Liu. the undercooling brought about by the rapid cooling should push the transition further toward the category of direct ordering in which the disordered phase tends to be retained and APD formation should be even more strongly suppressed. A rigorous determination of the order of the transition requires that the 'lever rule' be checked from microstructural observations. In experiments. however. This model was supported by the experimental measurement of domain growth in Cu3Au.

metals. but they were not observed until the early 1980s (Chou et al. e. and this requires that the APB energy is lower than a certain level. in which the APB energies are very high. and this is consistent with the operative dislocations being APB-dissociated < 1 1 1 > superdislocations.. In /3CuZn and FeAl. < 1 1 1 > dislocations are formed only if the resolved shear stresses on < 1 O 0> dislocations are close to zero (Rachinger and Cottrell. In the mechanism of Vidoz and Brown (1962) the crosssection of the APB tube is a parallelogram. In this mechanism. but prohibited at low temperatures since on (00 Ij the dislocations are locked along both screw and edge orientations (Hazzledine and Sun.. which. 1982. In the Ll 2 structure APB-coupled dissociation also competes with the dissociation into SISF-coupled |<112> super-Shockley partials. dislocations in deformed samples are usually < 1 O 0> type. Loretto and Wasilewski. Takasugi et al. Ni3Al and Ni3Ga.2 APB Tubes The possibility of APB tubes being produced by dislocation mechanisms was proposed in 1962 (Vidoz and Brown. the (1 1 Ij APB is thought to be stable or metastable and the screw dislocation may (temporarily) assume a planar mobile structure on (1 1 Ij. (1981a) showed (Figure 4) that. quite similar to the core structure of < 1 1 1 > screw dislocations in b.c. a <1 1 0> dislocation dissociated into APB-coupled superpartials on a (0 0 1) plane is energetically more favorable than the same dissociation on {1 1 Ij owing partly to the lower (0 01) APB energy. 1962).g. 1956. this may happen even when 7SISF is not sufficiently low to satisfy the self-energy criterion based on elasticity. In the latter case the role of the thermal fluctuations is to free the dislocation from the locked structures. not only is the dissociation mode made into SISF but also the core of the bordering |<112> partials is always non-planar when the dislocation is lying along a < 1 1 0> direction. Sun et al. 1986). and the material therefore becomes softer as the temperature rises. But the < 1 1 1 > dislocation may become energetically favored if it can reduce its selfenergy sufficiently by dissociating into APB-coupled partials. who treated the relative stability using isotropic elasticity. the (111) APB can be made unstable. by increasing the ordering energy in the Ll 2 structure. becomes more likely to be locked into the Kear-Wilsdorf configuration. the APB dissociation may also be made unstable if the y surface does not contain a minimum at the position corresponding to an APB shear. 1991. 1983). In the B2-ordered alloys the competition is usually between APB-dissociated < 1 1 1 > superdislocations and <1 0 0> ordinary dislocations. An unstable (111) APB has been thought to account for the normal yield stress in Pt3Al and Pt3Ga (Tichy et al.separation. In the Ll2-ordered alloys. when a superdislocation consisting of two APB-coupled superpartials is intersected by a * forest' dislocation or a dislocation in a secondary slip system. Formation mechanisms of APB tubes. with increasing temperature. a <1 00> dislocation has a lower self-energy than an undissociated < 1 1 1 > superdislocation. According to Suzuki et al. When the (111) APB is unstable. Chou and Hirsch. (1979). . (b) The cross-slip and annihilation mechanism ' z?7siSF where r0 is the equilibrium APB width. In NiAl and CoTi. 1990). the APB energies are relatively low. and the favored dissociation mode depends on the relative magnitudes of the APB and SISF energies. In addition to the self-energy.c. 4. Yamaguchi et al. An unstable (111) APB has been shown to alter the yield behavior of Ll 2 alloys significantly. APB dissociation on screw dislocations is favored if 7lll<1 9 Figure 9. 1971. 1991). in which case the stable configuration is the SISF dissociation. In alloys that exhibit the anomalous yield stress. (a) The VidozBrown non-aligned jog mechanism.. On the basis of the b1 criterion.

contrast is formed if g. which requires that g be a super lattice reflection. is imaged . one on each of the two superpartials. As a result. since tube contrast was observed in some fundamental reflections. in this reflection the dislocations are in strong contrast and the APB tubes are invisible. On the basis of this contrast selection rule. (a) The reflection 102 is parallel to the tube axis and the tubes are out of contrast. TEM weak-beam images of APB tubes in Ni3Ga. with the contrast being strongest in reflections perpendicular to the tubes axis. Figure 10(a) is a weakbeam image taken with 2 0 2 reflection parallel to the Burgers vector of dislocations in the primary slip system. (b) The reflection 020 is perpendicular to the tube axis and strong tube contrast is observed two jogs are created. Chou and Hirsch (1983) carried out systematic calculations of the image contrast of APB tubes in TEM. Chou et al.p 5* integer criterion was often violated. In their model an APB tube is treated as the equivalent of two overlapping APBs and the tube produces contrast because there is a phase difference of 2irg. The key factor in this mechanism is that the forest dislocations must be produced through the activation of the secondary slip systems. The geometrical feature of the APB tubes thus created is that they are connected to superdislocations containing non-aligned jogs. The contrast criterion is therefore essentially the same as that of the imaging of ordinary APBs. the APB created by the leading partial is not eliminated by the trailing partial and an APB tube is formed (Figure 9(a)).Figure 10. The two jogs are not aligned along the direction of the Burgers vector of the primary dislocation if the forest dislocation keeps moving during the intersection. Figure 10(b). (1982) and Chou and Hirsch (1983) successfully observed APB tubes in FeAl (B2) and Ni3Al (Ll 2 ).p 5* integer. in which APB tubes are visible in the form of thin lines. More recently Sun (1992) observed APB tubes in lightly deformed Ll2-ordered Ni3Ga and found that the contrast formation based on the g.p between waves diffracted from inside and outside of the tube.

instead of to jogged dislocations. lattice parameters and elastic constants. Ngan for permission to use unpublished work (Figure 5).p criterion. the strengthening by ordered precipitates in which the structure and energy of APBs are the controlling parameters (Chapters 2 and 12 by Liu and Pope and by Ardell in Volume 2). The structure of an ideal APB can be fully characterized by the APB displacement vector p and the plane normal n. shows that during cross-slip the partials do not follow one another along the same path. but clearly further developments are required in the field of fault-energy calculations. strain contrast.using 0 2 0. This observation suggests that there are lattice distortions normal to the tube axis. (1987) based entirely on the dislocations in the primary system. judging merely from the fact that different techniques have given very different results and that their agreement with the experimental measurements is usually subject to large variations. i. found to be inconsistent with the mechanism of Vidoz and Brown (1962). the surface tension causes a hydrostatic pressure inside the tube and a pure-shear strain field outside. and interesting variations occur when the APB is allowed to relax to reach thermodynamic equilibrium. in this sense the tube contrast formation is essentially the same as for dislocations. the attractive elastic interaction drives the two screw dislocations toward each other through cross-slip and they eventually annihilate. which is consistent with the experimental observations. viewed along the tube axis. It is hoped that the reader will find discussions related to the above subjects in these other chapters. in addition to the APB displacement.. Indeed. First. APB tubes were observed in slightly deformed samples in which no secondary slip system was activated. and as a result the annihilation is not complete and an APB tube is created. and to Wolf son College. The latest ah initio atomistic methods are promising in that they employ more rigorous treatments and that some have been very successful in predicting some of macroscopic properties. A simplistic situation is illustrated in Figure 9(b). arises as a result of the local deviation (as a result of the shear distortion) from the diffraction condition in the bulk. the dipole marked D in Figure 10. the work-hardening mechanism involving APB tubes. The traditional treatment based on pair wise interactions has the advantage of offering a clear physical picture about the factors governing the structure and energy of APBs. in which the screw partials cross-slip independently during the annihilation. The overall arrangement of the APB tubes in deformed samples observed in TEM was. The key feature in this model is that the tubes are connected to both ends of near-edge dipoles. It has also not been possible to cover several other topics closely connected with the structure and energetics of APBs and APDs: these include dislocation locking mechanisms involving APB transformation and disordering (Chapters 22 and 38 by Veyssiere and Douin and by Briant in this volume). Secondly. e. The computer simulation of Chou et al. 5. however. 1992) that.g. and the effect of the presence of APBs and APDs on electromagnetic properties (Chapters 40 and 41 by Kouvel and by Braunovic in this volume). An ideal APB has been shown to be not necessarily stable. strengthening by thermal APDs (Chapter 39 by Sauthoff in this volume). the APB energies obtained using the ordering energies derived from their relationship with the disorder-order transition temperature Tc have been found to be in order-of-magnitude agreement with the experimental measurements made by the TEM method (Tables 1-6 and Figure 7). which is also a fundamental reflection but perpendicular to the tube axis. . when two APB-dissociated screw superdislocations having opposite signs pass each other on elevated planes.e. Owing to the limitations of space. the tubes were often found to be connected to dislocation dipoles. H. In this mechanism. Strain contrast of APB tubes has since been observed in Fe3Ge by Ngan (1992). The additional contrast. in this case the tube cross-section is a simple parallelogram. to the SERC for financial support. 1992) for a circular APB tube. W. Oxford. For a circular tube in an isotropic material. Simple analytical solutions for the displacement field were obtained (Sun. and it led to the proposal (Sun. 1987). an APB tube also possesses a continuous distortion field that arises as a result of the surface tension of the APB. it has not been possible to cover in detail the theoretical treatment. e.g. while in general the cross-section is more likely to be an irregular polyhedron (Chou et al. and instead the aim has been to highlight the basic physical origins for the deviations from ideal APB structures. Acknowledgements The author is grateful to Dr A. Concluding Remarks Much of the discussion in this chapter is developed on the basis of the clear distinction between ideal APBs and real APBs. 6. Images were simulated and were found to be in qualitative agreement with the experimental observation. A new mechanism was proposed by Chou and Hirsch (1983) and Chou et al. which leads to the violation of the g. for a Junior Research Fellowship.

Hong. H.. J.. Beauchamp.... 81. and Miller. A145. and Yoo. I. Mag. Calvayrac.. H. (1971). (1967). Izumi).. Universite de Lausanne. Crawford. Liu. In Intermetallic Compounds (ed. Charpency. J. M. Korner. H. J. Phil. and Anstis. Phil Mag. (1940). A. and O. C. C. 245. T. 81. P. J. Voter. AlSl9 540. Carnegie-Mellon University. 4. Y. L. 58. T. 236. G. 18. Cupschalk. A. Broddin. Stoloff. and Schoeck. and Hondros. Res. 2... Horton. Fisher. A. Riflkin. M. A. (1966). (1991). S. (1985). C. N. 5. (1961). Phil. P. Douin. Chen. Rubins. E. W. 213. L. P. and Veyssiere. Pope. M. Proc. and O.. Phil. W. 22.. and Law. A. 565. Trans. 55A. Proc. B. C. P. Johnson. J. Hug. J. Stiegler). Res. C. L. Lin. 543. (1949). Mag. (1991). and Yoo. Res. T. 699. Soc. 2. G. Phil. and Warlimont. Appl Phys. Ada Metall. Stiegler). Taub. E. J. 60.. P. 62. Hazzledine. Mater. W. P. (1971). 300. Mater. (1960). Loretto. 209.. H. Soc. and Van Tendeloo. Soc. L.. Soc. (1950). and Cahn. and Veyssiere. Soc. 607. 81. 15.. Phil Mag. Thomas and J.. (1990). 133. Luzzi. G.. S. English. G. and Xu. Mag. In High Temperature Ordered Intermetallic Alloys IV (eds L. 33. M. G. 3. Y... Acta Metall. Y. Acta Metall.. Hirsch. Y. (1947).. Phil. 13. Stoloff. and Humble. Kikuchi. Wiley. In High Temperature Ordered Intermetallic Alloys IV (eds L. O. A.. p. S. Ada Metall. 799.. Marcinkowski. C. L. P. and Yoo. Proc.. Liu. / . P. Taub. C. 31. (1977). 159. A. (1990). (1966). 1517. M. J. and Brown. 333. P.. Stoloff. 145. M. E.. Mater. Dimiduk.. P. and Williams. Koch). S... In High Temperature Ordered Intermetallic Alloys III (eds C T . Proc. J. M. Z. PhD Thesis. M.. M.. and Whelan. and Williams. G. (1934). O. 847. J. Douin. Phil Mag. Stiegler). Res. 621. M.. (1986). O. J. Phys. D. Y. 14. and J. Koch. N.. A. Proc. G. In High Temperature Ordered Intermetallic Alloys II (eds N. T. 657. Phys. Phil. Leroux. and Miller. M. Yu. C. 45. Phil. (1961). 29. C. V. J. D. 300. Scripta Metall. J. Matter. R.. Washburn). and Sun. (1992). G. (1969). E. H. Marcinkowski. Rev. 569. M. R. J. A. 909. AIME. 20. Freeman. P. B. Mag. 507. AIME. J. A. Lasalmonie. Solid State Phys.. S... R. Lett. Baluc. Kear. and C. I. Cowley. K. 27. Appl Mech. 56. A. P. 871. 65A.7.. N. L... R. P. R. Mater. Phil Mag.. A. A. 58A. In Electron Microscopy and Strength of Crystals (eds G. Johnson. M. 1265. (1968). G.. Inui. Ada Metall. (1992).. Korner. and Daw.. W. Res. Pope. University of Oxford. M. F. (1989). Brown. (1935). Lett. I. Bragg. Carlsson. (1988).. A. Foiles.. 218. D. M. Douin. T. A387. Soc. 1311. A. 129. Soc. D. Izumi). M. L. (1987). 3. Rev. Phys. Soc.. Phil Mag. T. H. Meschter. Guttman. J. In High Temperature Ordered Intermetallic Alloys ///(eds C. 375. 33. 971. 891. and Wasilewski. 693. A. Z. K.. (1959). Stoloff. and Seitz. and Veyssiere.. C.. (1988). Crimp. Koch). 91. Liu.. p.. Trans. P. Korner. 1389.. (1989). 815.. Mag. and J. P. Veyssiere. Phil Mag. M. Taub. 105. S. 76. M. 133. H. 61. Vitek. and Wilsdorf.. Chou. J. AIME. C. Chou. E.. Res. Soc. P. ScI. 65A. 125.y B43.. Res.. N. Li. 52. and Brown.. M. Beeler. C. Nature.. C. and Ray. and Fisher. / . Schwellinger. 51. D. (1986). T. M. Li. A. B. PhD Thesis. L. P. Fu. 1385. New York. Bragg. Soc Proc. Phil. R. Porter. 9. Mag. S. 972. (1962). Head. J. Johnson. S. 213. T. N. 145.. J. Hug. (1950). (1987). Proc. 16.. Mag. D. Proc. McLean. Koch). Marcinkowski. S o c .. Res. Mater. Proc. 521. Loretto. Dirras. (1987). Trans. (1985). Stoloff. (1990). Status Solidi. E. Res. M. 549. (1990). W. 23. M. (1989). J. (1967).. (1970). Ray. C . C. I. P. / . (1979). R. N. Y. C . M. W. Hirsch. 54. 1337.. 224. H. P. S. J. L. References Ball. Ada Metall. P. M. J. Beauchamp. (1987).. J. and C. and Beauchamp.. P. Mater... Hong. C . P.. Bragg.. (1989). A. C. C. 19. J... H. . Clapp. Chou. A. 14A. R. Proc. and Hirsch. R. and C. A. D. H. 20.. Leamy. C. Phil. and Voter. J. Phil. Proc.. (1992). Mendiratta. Mater. C . 11.. (1987)... Trans. E. In High Temperature Ordered Intermetallic Alloys IV(eds L.. R. Marcinkowski. N. F. Ada Metall. J. Phys. and Smallman. Liu. 77. I. L. and Freeman. S.. Mag.. Interscience. Mag. S. New York. Chenal. Koch. Mater. C. / . M. P. Flinn. R. 245. (1987). G. In High Temperature Ordered Intermetallic Alloys III (eds C. A. Mater. 65A. 537. T... 669. Phil Mag. 2191. (1956). (1967).. F. Hirsch. Pope. Mater. (1982).. J.. 27. H. 40. Acta Metall. N. Fu. and Srolovitz. (1963). S. Cadeville. (1991). Mater. Proc. Loiseau. Cockayne. H. Cahn. and Yamaguchi. Soc. A. A. Phil Mag. 6. P. Mater. In High Temperature Ordered Intermetallic Alloys II(Qds N. H. C. and Fayard. Pope.. J. W. 217. Phys. Hazzledine. F. J. 6446. Lakso. 1. G. C. Res. and Karnthaler. 5. 382.. F. (1990). 29.: Condens. D. (1962). Brown. P. Horton. 105. and Liu. M. B. H. Fu. 6. (1969). J. Fu.. (1992). 1111. 703... 233. T. (1961).. (1972). and Marcinkowski. (1989).. B. Westbrook). M.. PhD Thesis. Q. Phys.. D. B. Koehler. Rev. (1983).. R. 20. Crawford. 3479. and J. Cupschalk. Res. and Beauchamp. R. 213. AIME. R. 133. S. Mag. J. D.. N. Liu. T. A.. D. Rev. Bull. and Brown. 35. M. and Marcinkowski. (1991).

65A. G. A. Zschack. Izumi. (1970).. A154. 57. 743.-G.. Acta Metall. Johnson. F. and Kuramoto.. Phil. University of Birmingham. G. Int. and Rao. H. M. E... Sci. ISI J. P.. P. O. AIME.. I.. Sanchez. X. Tsurisaki. U. G. J.y 133. Trans. and J. D.. Tichelaar. PhD Thesis. Takasugi.. and Ezz. T. Soc. 4. E. Wu. Nicholson. D. Liu. I. SpringerVerlag. Yoo. Eng. 97. Y. M. Stoloff. (1990). Mag.. Umakoshi. M. Piller.. N. Phil. P. R. Soc. Woodward. (1953). S. 415. 525. Phys. H.. M.. R. N. and Turchi. In High Temperature Ordered Intermetallic Alloys IV (eds L. Gonis. G. 201.. Blackburn. 61. Proc. Eng.y 39. Liu. (1985). S. A. C. Sci. O. D. and Tien. Stoloff). K. Status Solidi. (1975). S. and Yamane. and Yamaguchi.. Sanchez.. D. C. 569. 1. Poquette. J. D. Rachinger. (1989). D. S.. In High Temperature Ordered Intermetallic Alloys III (eds C.. 469. P. C. 13. S. Eng. Acta Metall. Res. C. C. Acta Metall. A. G.... F. P. 21. Phil. Pei. and Jaumot. H. Int.. Schoeck. Proc. (1991). Paidar. Sutcliffe. Sparks. Liu. Soc Proc. and Sauthoff.. Pope. 1803. Oya.. 64. Proc. Pinski. D. (1960). T. N. M. S. 65A. M.. A. T. (1985).. and O. D. 61.. E. T. Kawase. Stoloff.. (1984). 207. Dordrecht. Acta Metall. Mag. Mag.. Phil. T.. P... 123. Proc. H.. Johnson. Mag.. F. 7. G. 266. 5OA. J.. (1974). and Smallman. P. H... P. Y. Temmerman. and Brown. Soc Proc. 125. and Liu. (1979). Koch. Mater.. Morris. Phil. J. Ray. Massalski. A. R.. (1973). S.. Pope. Y. Y. M. Mater. 221. Buckley. Koch). Koch.. M. Brown. Crimp. (1992). T. Sun. S. In High Temperature Ordered Intermetallic Alloys IV (eds L. Crawford. Proc.. J.. Neveu. D. Stiegler). Mag. and C.. G. 73. H. Phil. C. F. E. C.. Zhang. Pope. A. and Tien. 363. Mishima. 136. Soc Proc. M. Morris. Stoloff. Warlimont). Acta Metall.. (1989). 245. Mag. I.Mishima.. 213. Soc... and Schneible. (1976). Jones. 29... D. p. Ice. Ngan... Res. P. Trans. Mater. C. D. 145. and Li. (1987). PhD Thesis. Specht. Acta Metall. Y. 65. Johnson.. Kozlov. 357. J. 1. E.. G. Mater. Gyorffy. Phil. R. 912. J. and Lipsitt. 23. 52A. Kachi. P. M.. P. (1984). Res. 389. (1988). Temmerman. S. Rao. Liu. Res.. and Golosov. R. 715. Phil. Saka. Reynaud. D. T. P. G. 1027. B. P.. 180. Wu. M. a32. and Miura. Nicholson. 435. M. F. W. In High Temperature Ordered Intermetallic Alloys II (eds N. Veyssiere. Mag. M. G.. H. (1992). Q. T. Mag. C. and Tien. M. Izumi). P.. D.. S. In High Temperature Ordered Intermetallic Alloys III (eds C.. M. Y. 321. and Cottrell. 634. Y.. D. (1969/70). S. and Mikkola. P. 311.. A. and J. Morris. Vitek. Tounsi. Mag. 81.. Suzuki. K. 59. A. D. Phil.. 33. 5. Koch). E... A. (1962). R. Q. (1986). Mater.. and Smallman.. T. Acta Metall.. Phil. A. (1992b). and Pope. 32. Schapink. J. M. Mag. 785.. D.. Murakami... V. 17. N. Mag. T. and N. Res.. A. J. Scripta Metall. P. A.. B. L. Sterne. A. Mater. (1989). 193. W. L. J. (1979). H. G. IA. (1966). Y. In High Temperature Ordered Intermetallic Alloys IV (eds L. and Vitek. A. 917. B. Temmerman.. (1976). S. Tichy.. W. C . (1984). 21. I. Johnson. Shechtman. Coulomb. In High Temperature Ordered Intemetallic Alloys (eds CC. 5767. 9.. D. A. New York. (1990). NATO Advanced Research Workshop).. 287. H. L. (1975). and Takehara. K. P. Sanchez. Umakoshi. Y. Takeuchi. and Stocks. P. F. p. and O. M. Soc. S. 51. In High Temperature Ordered Intermetallic Alloys HI (eds C T . (1985). L. (1991). D. Pope. C. 19. S. N.. B.. (1992a). E. Prakash. Universite de Poitiers. G.. and Korner. Stoloff. Taub. G. Paidar. K. Stocks. M. Mater. Taub. Phi. Wu. A. O. T. D. W. A. T. Stiegler).. Sun. Phys.... Y. F. 53A. (1974). 29. (1936). H.375. 175. (1986).. Phil. Veyssiere... C . 11. Soc Proc. Res. and Ono. D. S. Peierls.. P. (1969).. and C. H. Sastry. Wee. 2249. (1956). 109. Metall. E. In High Temperature Ordered Intermetallic Alloys IV (eds L. 25.. (1989). 87. Phys.. Stoloff.467. J. and Ramaswami.. 5. and Sellars. A. (1991). B39(9). W. M. (1991).. Mag. 61. J. Stiegler). (1986). Sterne. R. In Ordered Intermetallics—Physical Metallurgy and Mechanical Properties (Proc. 1167. Suzuki. E. Stiegler). S. 725. Morris.. 133. Res.. Rev. 213. In Order-Disorder Transformation in Alloys (ed. Johnson. Soc. Potter. and Leamy.. H. A. V. 65. Res. (1991). Met. G.. Rev.. and Pettifor. KIuwer Academic. (1989). Trans. E. L. Mater. Stocks. Acta Metall.. (1991). and Cockayne. J. C. Hazzledine. and Suzuki. (1975). 81. Phil. and Partharathy. O. 20. Rudy. M. Metals. A. M. PhD Thesis. Liu. E. P. Pope. Soc. P.. Butler. Maclaren. (1987). Phil. Yamaguchi. Soc. S. (1971). PhD Thesis. Morris. C. D. Mag. 53. 213. 21. Liu. Mag. . and C. P. and Beauchamp. 33. (1992). W. P.. T. C. Universite de Paris-Sud. Mater. Y. Nohara. D. N. S. J. A. 23. E. V. (1985). and J. Nakanishi. 213. G. M. and Suzuki. Lett. D. 81. O. 8. Tanner. J. Koch. Mater. A. L. and Imura. Sterne. Universite de Paris XI. K. A. Ichihara.. Popov. Proc. Mater. T. M. S. In High Temperature Ordered Intermetallic Alloys II (eds N. H. Res. Koch). Japan Inst. (1992).. C . Lasserre.. 1373. 15. Pope. Proc.. Woodward. C . Acta Metall. O. Izumi). (1989). I. and Takeuchi. and Couret. J. Acta Metall. Acta Metall.. 119. M. P. Veyssiere. Horton. W. Status Solidi. M.. Mag. Rudman. 133. E. (1991). Oliver. X. C.. Mater. Phil. T. I. and J.. Res.. P. Douin. Taub... Vidoz. S...

Yamaguchi. H.. Yodogawa. Philosophical Magazine. lower than the respective APB energies (140 and 220mJ/m 2 ) given by the earlier work (Fu and Yoo. and Umakoshi.. and Sun. T. 77..g. Umakoshi. P. Suzuki. the first-principles calculation suggests the true ground-state structure of Ni 3 Ge to be DO22 which is derivable from Ll 2 by forming {001} APBs on Table 7 APB energies in Ll 2 compounds (Paxton and Sun. Wee. T. C. 1. ScL. Pt3Al. Dimiduk. For Ni3Al. Vitek. V. APB energies in Cu3Au. and S. p. 78. Pope. (1981a). The calculated {111} and {001} APB energies were used in predicting the formation of Kear-Wilsdorf locks by the cross-slip of (101) screw super dislocations. (1998). Ni3Si.. Vitek. O. Vol. The APB energies given by the new calculation are lower than those given by earlier works in which the same method was used but only one mode of relaxation was allowed (e. Y. Mag. Yamaguchi. and Umakoshi. University of Tokyo Press. 1989). In Intermetallic Compounds. Japan Institute of Metals.. the calculated APB energies are 115 and 188mJ/m2 on {001} and {111} planes. Addendum On the subject of APB energies. M. 17. the FLAPW method was recently used by Paxton and Sun (1998) to calculate the APB energies in several Ll 2 intermetallic compounds. p. D. Q. Westbrook and R. M. Y. Y. both normal and parallel to the interface planes. 1998) Composition Ni3Al Cu3Au Pt3Al Ni3Si Co3Ti Ni3Ge Y111 (mJ/m2) 188 129 349 682 692 83 y010 (mJ/m2) 115 47 523 683 412 -40 Method FLAPW FLAPW FLAPW FLAPW FLAPW FLAPW alternating planes (Figure 3). 85. 43... Ninomiya. (1985).. 1991). Phil... The Kear-Wilsdorf locks in Ll 2 intermetallic compounds and their effects on mechanical properties are discussed by Veyssiere and Douin in Chapter 5 in this volume. Structural and Mechanical Properties (ed. Izumi). (1990).Yamaguchi. Mater. Chem.. Sendai. 1265. M. 1989) and closer to the results measured by weak-beam electron microscopy (e. (1945). Phys. 1989.g. Y. V. In this work. J. edited by J. Prog. N. M. and Yamane. 43. and Suzuki.. Phil. Neveu. A.. Yang. M. Oya. (1991). Scripta Metall. 849. K. D. and Pope. 1: Principles. Reference Paxton. This chapter was originally published in 1995 as Chapter 21 in Intermetallic Compounds. (1981b). 14. L Fleischer. In Dislocations in Solids (eds H. Notable is the negative {010} APB energy in Ni 3 Ge resulting from Ll 2 being its metastable crystal structure. Mag. 403.. (1980). Y. Yamaguchi. Y. atomic relaxation. 34. P. Tokyo. and Takasugi. Takeuchi). 1027.. T. Ni 3 Ge were also calculated with the same approach and are listed in Table 7. 66. Sumino. Fu and Yoo.. . T. M. was allowed in calculating the final interface energy after full equilibration. Yoshida.. D. T. Co3Ti.

whose complexity and variety have become increasingly apparent over the last two decades. to a lesser extent. optical. H. This chapter focuses on problems that have attracted the attention of the scientific community over the last decade. Dislocations in intermetallics constitute an extremely wide field that cannot be reasonably covered in this chapter. they may not all have been reliably answered yet. Unite Mixte CNRS/ONERA. though many of the most important questions on the plasticity of intermetallic compounds were raised during the late 1950s and the early 1960s. and it is worth realizing that. knowledge of individual dislocations not only provides information on the variety of plastic behaviors of crystals. and they can reproduce all sorts of core rearrangements. 92322 Chatillon Cedex. The organization of dislocation networks has been demonstrated to be fractal in some circumstances. which makes the detailed study of their core* structure so important and attractive in this category of alloys.Chapter 5 Dislocations Patrick Veyssiere and Joel Douin Laboratoire d'Etude des Microstructures. L. others are dictated by the detailed structure of dislocation cores.1 Definitions and Tools It is a fact that a number of macroscopic mechanical properties of intermetallics are governed by the intrinsic mobility of dislocations. it should be read in close connection with Sun's contribution in Chapter 21 in this volume. Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. the reader is thus invited to refer "Here the term core refers to the fine structure of a dislocation regardless of its extent. grain-boundary structure—with deep implications for the electrical properties of semiconducting compounds—and. Some physical properties depend on the long-range elastic stresses and strains of dislocations. including the effect of dopants. Physicists interested in electronic structures can now make more reasonable atomistic studies of dislocations in metals and alloys as well as in ionic-covalent solids. Introduction Individual dislocations. Beside technologically oriented studies. the magnetic. UMR 104. their organization. Nevertheless. where the reader may find complementary specialized information on topics with which the present authors are not particularly familiar. and the influence of these on the physical properties of crystals constitute an open field in which investigations have been very active over the years. be addressed in Chapters 24 and 39 by Takasugi and by Sauthoff in this volume and in Section I in Volume 2. Fleischer ©1995. but it provides some predictive tools in a wide range of domains covering crystal growth. This chapter on dislocations will focus on plasticityrelated analyses. dislocation studies in ordered structures have significantly complemented our overall understanding of dislocation processes for themselves or in relation to plasticity. 29 Avenue de Ia division Leclerc. 2000 John Wiley & Sons Ltd . and thermal properties of crystals. at the beginning of studies of the mechanical properties of this class of materials. These problems are often complex. In addition. 1. Edited by J. dislocations have retained the attention of physicists as examples of solitons and of singularities of an order parameter. however. Westbrook and R. France 1. the chemical effects that take place in their vicinity. Some of the other topics mentioned above will. Dislocations have been and will probably remain primarily studied in connection with plasticity. because of the abundance of available mechanisms.

(a) Ll 2 (cP4).Figure 1. compared to Ll 2 . one of the unit translations represented in this structure. (c) Ll 0 (tP4). terminates on an atom located out of the unit cell - . (cF16). (g) DO3 (cF16). (i) ClIb (tI6). (f) L2. ^ [ 3 3 1 ] . (h) phase O (oC16). (d) DO19 (hP8). the two cube faces perpendicular to the c-axis contain only like atoms. The various ordered structures examined in this review. (e) B2 (cP2). The thick arrows represent some of the shortest unit Burgers vectors of each structure. (b) DO22 (tI8). note that the unit cell consists of two Ll 2 unit cells shifted by | [110] every other (001) plane.

Under both WB and HREM.2 Observing the Fine Structure of Dislocations Macroscopic mechanical tests provide limited direct information on dislocations that consist of essentially the operating slip system and a few microscopic parameters of the deformation such as the activation volume and the activation energy. 1979. The symmetry of an alloy is reduced upon ordering (Figure 1) and the shortest unit translations in the disordered lattice are not in general perfect Burgers vectors of the ordered alloys. This is exemplified in the following results obtained under HREM and earlier reviews of interest (Marcinkowski. where some of the aspects that are discussed below may be developed from different points of views. (1991a) showed that. 1984. Understanding the fine structure of dislocations requires. 1970). it was not anticipated that under WB one could ever routinely resolve between dislocations separated by. the spinel structure. in their study. Nowadays. Brown. with the consequence that the sample surface is a favored nucleation site for cross-slip (George et al. their line images are not shifted by the same amount (Baluc et al. Yamaguchi and Umakoshi. that direct analyses are conducted at the appropriate resolution level. and dislocation studies can now be conducted very accurately. and this notation has survived over the years. Though companion partials have the same Burgers vectors and are therefore viewed under the same g. Long ago (Marcinkowski. Afterwards. although it is certainly superfluous. 3 to 4 nm. which is specific to ordered alloys. 1963. 1991a. <100> dislocations remain perfect upon ordering in both the B2 and the Ll 2 structures. 1966.. the imaging of lattice columns. Amelinckx.. where the dissociation and subdissociation schemes are exactly the same as in Ll 2 alloys. the observed separation between superpartial cores differs from its actual value and that the correction depends upon the dissociation plane. which. the analysis of dislocation images can be uncertain because of problems associated with contrast artifacts and of difficulties in interpreting images.. in Ni3Al. for instance. contrast may not always be trivial to understand. Stoloff and Davies. 1963).b. In the last 30 years. (ii) Weak-beam technique. It should be kept in mind that since a dislocation image results from the interaction between an electromagnetic wave and a distorted lattice.f For this reason. as shown by the many similar situations encountered in ceramics. and observations are even being conducted at resolutions better than 1 nm. intermetallics have been studied under high-resolution electron microscopy (HREM). powerful observation techniques have emerged that have benefited from the improved resolution of transmission electron microscopes (TEM). the expression superlattice dislocation was introduced in order to help to identify in ordered alloys these dislocations whose nature would be changed if the crystal became disordered. 1991a. 1992). Pope and Ezz. On the other hand. that is. Difficulties that are currently encountered in the course of studies of intermetallic alloys have been reviewed elsewhere (Clement et al. 1992) that. 1989. 1. 1967. say. it has been shown recently (Oliver. respectively). For instance. WB experiments are used to pinpoint details separated by about 2 nm without the need of drastic precautions.b. 1992).b condition. is the so-called Eshelby twist and its implications for the contrast of end-on screw dislocations: Mills (1989) used this effect to point out subtle contrast differences between the pair of y<l 10> partials in Ni3Al depending upon whether they form a super dislocation or a dipole (same sign or opposite sign. in fact. on the other hand. 1991a) and their intensities differ in general (Veyssiere and Morris. It was only in the early 1980s that materials scientists started to apply WB extensively to dislocation analyses in intermetallics. a pair of peaks could originate from double ^ee. this convention will not be respected too carefully in the following. It is worth recalling that one of the early applications of the weak-beam (WB) technique was aimed at demonstrating the fourfold dissociation in Fe3Al (DO3 structure) (Crawford and Ray. 1973. WB images of pairs of partials with collinear Burgers vectors had been used quite carelessly for years until Baluc et al. however. Less well-known. for particular deviations from the Bragg condition. Moller and Haasen. coincided with the projection plane. 1976). the expression superlattice was simplified into the prefix super. 1975). . as well as in deriving physical information from these. (i) High-resolution electron microscopy. Veyssiere. Simultaneously. At that time. more accurate studies require some expertise with the instrument. 1990) as well as to the other chapters in Section IV in this volume and Chapter 39 by Sauthoff. It is wellknown that boundary conditions at free surfaces are liable to induce peculiar relaxation properties such as the widening or narrowing of the separation between companion Shockley partials (Hazzledine et al.. }<111> is no longer a perfect Burgers vector in the B2 lattice.

and Hemker and Mills (1993). A splitting appears on the upper superpartial under the g . This may explain why such screw superpartials have been occasionally reported as subdissociated into Shockley partials with separations that are unexpectedly large when compared to the same dislocations viewed under HREM. (b) Simulation of image profiles showing that. stands for arbitrary units).Intensity (a. under g-2. For additional information on the practice of the WB technique. knowledge of the mobility of individual dislocations under load is important in analyzing corecontrolled mechanical properties.Hg. (a) Screw superdislocation in Ni3Al. see Cockayne (1972). with two beams. the splitting appears only when four beams are included in the calculation whereas. no extra image arises: the splitting is an artifact that originates from a multibeam effect diffraction in the core of a y<l 10> dislocation imaged under conditions such that g. two . no extra peak appears on either one superpartial (a.u. Finally. in the appropriate orientation. The simulations demonstrate that.) ^ 9 "<• Figure 2.3g orientation.7g. Experimental evidence of the splitting of the peak associated with a y <110> superpartial under conditions where g. n is the abcissa of the intersection of the systematic row by the Ewald sphere).b = 2 and selected deviations from exact Bragg orientation are shown (g . Veyssiere (1991a. For this purpose. 1992). Veyssiere and Hug (1990).u.3g conditions. (c) Same as in (b) but under g .b = 2 (Figure 2).

Finally. 1991). Nohara (1991). see also Chapter 4 by Voter in this volume)) that are fitted to a number of crystal parameters. Caillard et al. Veyssiere. from continuum elasticity (see Section 2. Baker and Schulson (1989).Next Page techniques that provide direct information should be mentioned: (i) in situ straining in the electron microscope. 1991a). On the other hand. since in some systems. This can be used either to analyze dislocation properties directly or as an input for image simulation of HREM pictures. Core simulations in intermetallics have been reviewed by Vitek (1985).1. in addition. Most simulations have been conducted using empirical central-force pair potentials and N-body potentials (Finnis-Sinclair method (Finnis and Sinclair. In the latter. 1992). for instance. On the other hand. it is still unclear from an experimental standpoint whether a strong negative temperature dependence of the flow stress (TDFS) in Ll 2 alloys undergoing octahedral slip at low temperature. Si. The former has been applied quite extensively to a variety of solids. The description of the elemental motion of |<110> partials on the cube plane (Section 2. which could introduce some serious limitations in dislocation core analysis (Section 2. but not much more than other TEM techniques of investigation (Clement et al. 1991a. (1977a. In B2 alloys. (1991). As with most techniques.3 Simulating the Fine Structure of Dislocation Two levels of simulation are used to analyze dislocation cores. 1991). and Suzuki et al. the etch-pitting technique has not been applied to intermetallics studies at least to a level comparable to that of in situ straining. and other ionocovalent crystals where the etch-pitting method has been successfully applied. 1. and in recent years to intermetallic compounds. (1991a..b). on dislocation-solute interactions (Dimiduk et al. Simulations based on empirical and semiempirical potentials cannot take the angular dependence of the bonding into account (Pettifor. the details of the flow-stress anomaly could not be explained from computer simulations. (1987a. they remain useful to check dislocation core properties based on crystal symmetry.b. on y surfaces and surface defects (Khantha et al. and (ii) the Johnston-Gilman etch-pitting techniques.1). they have strong limitations. atomic positions in the vicinity of a dislocation are determined using empirical or first-principles calculations of atomic interactions.2). They have shed light on the influence of crystal symmetry on dislocation cores (Pasianot et al. 1992). There seems to exist a potential of applications for etch-pitting measurements of dislocation mobilities in intermetallics (Nadgorny.1) to atomistic simulation. Simulation methods based on first principles are becoming available.3). Vitek et al. 1992).3.2% (Section 5. 2. Pasianot et al. except for specific crystal directions . Vitek et al. embedded-atom method (Daw and Baskes. not all difficulties encountered during in-situ experiments are. such as in Ll 2 alloys.. implies that the flow stress is governed by the core of |<112> partials (Section 5.. Illustrations of such studies can be found in the papers of Baker et al. 1984). Etch-pitting experiments could be significantly hindered in intermetallics. 1992. 1991). 1991) is not verified experimentally (Minonishi et al. Nevertheless. it can be noted that the prediction of a yieldstress increase with decreasing temperature in Ti3Al alloys deformed under prism slip (Vitek et al. They have provided a number of clarifications.. and the situation does not improve when dealing with more complicated defects such as dislocation cores. simulations are not in general fully reliable (see antiphase-boundary (APB) energy calculations in Ni3Al and in NiAl (Chapter 21 by Sun in this volume). especially in Ll 2 alloys. The Fine Structure of Superdislocations in Relation to Crystal Structure The fact that most ordered alloys are elastically anisotropic (see Chapter 37 by Nakamura in this volume) prevents one from comparing forces or energies analytically and. but the number of atoms that is tractable at present is significantly fewer than what dislocation studies require (De Fontaine. Pettifor.2. and this for stacking-fault and APB energies which can be adjusted. 1992). Moreover.b). (1991). see Vitek. Conclusions based on atomistic simulations of defects in intermetallics are not definitive.. clearly identified. the flow-stress anomalous behavior occurs only at permanent strains of the order of 0. because dislocation densities are larger in intermetallics than in LiF. it is clear that dislocation densities are then much too large to be tractable by means of surface etch pits. 1983. (1991).b). implying approximations at various levels. but they will evolve with the rapid refinement of approximations to describe atomic interactions from first principles and with the improvement of computational capabilities. MgO. 1985) is probably the most spectacular success of atomistic simulations in intermetallics.. 1991. Simulations may nevertheless be helpful in solving problems that would otherwise require intuition. and Parthasarathy et al. Its limitations are numerous. However.

(CEA-CEREM. such as atomic diffusion. (d) bound triple defect consisting of two vacancies and one antisite atom. we have therefore identified HF and EF. As in pure metals.9 1964. one should consider formation and migration enthalpies HF and / / M . interstitials. (O unbound wrong pair. (b) split interstitial.9 1971). the values of pVp and pVM are typically 10" 5 eV. at atmospheric pressure p. and can form thermally or accommodate non-stoichiometry. negligible compared to EF and EM. Koczak et al.1) compounds formation energy* and k is the Boltzmann constant. and their combinations. Edited by J. Strictly speaking. Much less work has been done in the case of intermetallic compounds. HM and EM. sintering. in long-range ordered (LRO) alloys. high-temperature creep and other mechanical properties. the role of impurities has scarcely been studied.Chapter 6 Point Defects Charles de Novion Laboratoire des Solides Irradies. H. and we shall only mention a few examples in this review. Introduction Point defects have a significant role in intermetallic compounds. But. In the following. behavior under irradiation. (e) vacancy-impurity ( • ) bound pair. HD and ED. (c) bound pair of antisite defects. A large literature exists concerning point defects in pure metals. L. or produced by irradiation or cold working. the latter consist. interstitials. Abromeit and Wollenberger. where EF is the (a) (b) (C) (d) (e) (f) Figure 1. impurities. They can be retained by quenching. Many other types of defects are possible. and in particular irradiation-induced crystalline-to-amorphous transitions. Fleischer ©1995. *The formation energy £"F and the formation volume VF are defined as the change of energy (volume) of the crystal when displacing an atom from the bulk to the surface (vacancy) or from the surface to the bulk (interstitial) without changing the surface energy. France 1. as they control many properties of technological importance. for example in the case of a binary AB alloy. The rate of disappearance of excess point defects is proportional to exp(-EM/kT). Section 5. In particular. Westbrook and R. 2000 John Wiley & Sons Ltd . Introduction of point defects by irradiation has even allowed one to obtain an ordered phase (FeNi) in a system where it was hindered by the low atomic mobility (Neel et al.3) and A15 (Nb3Sn. 1987). and simple covalent and ionic crystals (Wollenberger. as will be discussed for example for BIl (CuTi. Section 9. where EM is the migration energy. In pure metals. Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. point defects are not found thermodynamically at O K: they form at finite temperature T. antisites). 91128 Palaiseau Cedex. of B atoms wrongly placed on A sublattice sites (see Figure 1). focusing on intrinsic point defects (vacancies. URA CNRS No. Typical point defects in a binary ordered alloy: (a) vacancy ( • ) . 1972. 1380) Ecole Poly technique. 1983. dilute alloys. but also. Crawford and Slifkin. point defects in intermetallic compounds include vacancies. antisite defects. In most solids. with concentration C—Qxp(-E¥/kT).

vacancies have a formation energy E^ of the order of 0.5 to 2eV, leading to a thermodynamic vacancy concentration at the melting point of a few 10~4, and ^M = 0.5 to 1.5 eV. The self-diffusion coefficient is equal to exp(-£' D /A:r), with E^ E^+ E^. On the other hand, in some intermetallic compounds that permit off-stoichiometry (see Section 3.2), one finds structural ('constitutional') vacancies, even at low temperatures: the concentration of such defects may be important (several percent) and in some cases the defects order. Very generally, point defects distort locally and induce electronic perturbations in the crystal: these effects lead to elastic and electronic interactions between them, as well as with other defects (dislocations, grain boundaries, etc.). The complex defects so formed may introduce more distortion into the lattice than simple point defects, and therefore have greater effects on the mechanical properties. With increasing temperature, point defects become more and more randomly positioned, and complex defects (e.g. divacancies, vacancy-impurity or interstitial-impurity bound pairs) dissociate. From the fundamental point of view, vacancies and interstitials in intermetallic compounds present special features and a much more complex behavior, as compared to pure metals. For example, in a pure metal of simple crystal structure, all the bulk lattice sites are equivalent for the formation of a vacancy. On the contrary, in an LRO concentrated alloy, the various sublattices are not equivalent. Moreover, atoms of a given sublattice may have various atomic environments: in an LRO alloy as well as in a concentrated solid solution, many types of vacancies can be produced. Also, the vacancy formation energy depends on the extent of off-stoichiometry, and the thermal vacancy concentration may be much larger than in pure metals. In the same way, the migration of vacancies is more complex than in pure metals, because sites are inequivalent; for example, in an LRO alloy, the jump of a vacancy between two neighboring sites belonging to different sublattices changes the LRO parameter, and this will increase the migration energy. The same arguments apply to interstitials. In pure metals of simple (f.c.c. or b.c.c.) crystal structure, the structure and properties of interstitials are well-known; for example, in many f.c.c. metals, it consists of a dumbbell oriented in a <100>-type direction, with a large formation energy (several eV) and a small migration energy (typically 0.2 eV). In intermetallic compounds, the structure of interstitials is generally not known: in some cases, several types of complex defects occur

simultaneously. In concentrated solid solutions, the migration properties of interstitials are typical of transport in disordered media, requiring specific concepts such as percolation, localization, and trapping: this leads to a slowing down of the interstitial, which can in some cases show a larger migration energy than the vacancy.

2. Experimental Techniques for Studying Point Defects 2.1 Production of Point Defects Because their formation energy is generally lower than that of interstitials, the quenching of a stoichiometric intermetallic compound from high temperature, if fast enough, will give isolated vacancies and antisite defects. The best way to obtain self-interstitials is by lowtemperature electron irradiation, which in fact creates isolated vacancy-interstitial pairs (Frenkel pairs) (Wollenberger, 1983). Ion or fast neutron irradiation directly gives defect clusters, so-called 'cascades.' Electron irradiation can be performed either in a highvoltage transmission electron microscope (HVTEM) with direct observation of defect clustering, or in a Van de Graaff accelerator; in the latter case, measurement of defect production and recovery is generally made by electrical resistivity. By varying the incident electron energy, one can determine the threshold displacement energies Ed for various types of atoms. For example, in Cu3Au (Ll2 (cP4)) and CuAu (Ll0 (tP4)), the minimum electron energies for displacement of (light) Cu and (heavy) Au were found respectively to be 400 and 900 keV (Ed ^ 18 eV). By irradiating between 400 and 900 keV, one creates only Cu Frenkel pairs; with an electron energy above 900 keV, gold atoms are also displaced (Alamo et al., 1986).

2.2 Determination of Point-Defect Formation Properties Structural vacancies in non-stoichiometric intermetallic compounds have been detected by coupled density and lattice-parameter measurements. The measured density is compared to calculated densities in order to choose between several models: antistructure defects, vacancies within a sublattice, interstitials, etc. The application of this method to CoAl alloys is shown in Figure 2 (Fleischer, 1993). Care must be taken to check the absence of porosities and to determine precisely the chemical content of the sample.

Relative density

Atoms per cell

Aa/a0 (• ) and AUL0 ( o )

Lattice parameter (nm)


C (Al-rich)

C (Co-rich)

Figure 2. Relative density (specific gravity) and lattice parameter (measured at room temperature) and deduced cell occupancy as a function of composition for CoAl B2 alloys (from Fleischer, 1993). Each excess Al atom produces a vacant lattice site on the Co sublattice. Excess Co leads to lesser vacancy concentrations

Figure 3. Variation of lattice parameter a ( • ) and macroscopic length L (O) as functions of temperature in B2 (cP2) Co48Ga52. Also shown are the deduced vacancy concentration for Co48Ga52 (broken curve 1) and Co56Ga44 (broken curve 2) (From Van Ommen et al., 1981)

(Simmons and Balluffi, 1960). The extra thermal vacancy content is given by ACW(T) = (1 + Cv0){ [ 1 + AL(TVL0] V [1+Aa(TVa0V-I] where Cvo, L0 and a0 are given at room temperature. A typical set of curves, obtained for Co48Ga52, where deviation from stoichiometry is mainly accommodated by Co structural vacancies (see Section 3.2), is given in Figure 3. Above 560 0 C, the formation of thermal vacancies leads to a splitting between L(T) and a(T) and to a deviation from the normal thermal expansion (Van Ommen et al, 1981). The study of vacancies is now mainly performed by the positron annihilation technique (see for example Corbel, 1987). Because the electron density is weak in a vacancy, the lifetime r of a positron trapped in it is

The progress of X-ray and especially neutron diffraction techniques now allows one to obtain information on structural vacancies and antisite defects (in the percent concentration range) from the precise determination of structure factors (see for example Kogachi et al., 1992); neutron measurements are possible up to very high temperatures. To create thermal vacancies, atoms may be thought of as taken out of lattice sites in the interior of the specimen and placed on its surface. This changes the number of sites and the length of the specimen (although the vacancy formation volume is smaller than the atomic volume because of relaxation effects). Therefore, thermal vacancy formation can be studied by the comparison of the relative change of length L and lattice constant a, measured in situ at high temperature

Peak counts (arb. units)

Temperature (0C)
Figure 4. Coincidence counts at the peak of the angular distribution of annihilating photons, measured as a function of temperature in /3-brass CuZn, B2 (cP2). Broken curve: fit with a single exponential. Full curve: fit with a T-dependent formation energy (From Kim and Buyers, 1980)

noticeably larger than if itinerant in the bulk of the crystal (typically 180 and 100 ps, respectively). Also, the value h at the peak of the correlated angular distribution of the annihilating photons, or the parameter S of the Doppler broadening line, are very sensitive to the concentration of vacancies (above a threshold of 10~7). The parameters r, h or S, measured either in situ at high temperature, or after quenching, are generally analyzed in a two-state trapping model, where they are assumed to depend linearly on the positron annihilation rate at the defect, which is proportional to the vacancy concentration. In the simplest model, one has, for example hB + L(T) = l+Awexp(-E%/kT) (where B stands for 'bulk' and V for 'vacancies'). Figure 4 shows, for CuZn (B2 (cP2)), the annihilation rate h (T) plotted as a function of temperature, from which the formation energy £p of an average vacancy has been obtained (Kim and Buyers, 1980) (see Section 3.1). When lattice defects are generated thermally, their formation energy gives an extra contribution to the heat capacity of the crystal. To our knowledge, such measurements have not been applied in intermetallic compounds. A frequently used technique to study point-defect creation is electrical conductivity measurement, because of its simplicity and great sensitivity. Indeed, in metallic materials, point defects are scattering centers for electrons: they increase the electrical resistivity. In semiconductors, they introduce localized levels in the hvAvexp(-E^/kT)

gap and change the number of carriers. The measurements should preferably be performed at low temperature to avoid thermal contributions: for example, determination of the vacancy formation energy on samples quenched from various temperatures or study of the defect production during irradiation. Information on the structure of interstitials can be obtained from diffuse X-ray scattering (Dederichs, 1973). In particular, the long-range part of the strain field of point defects leads near the Bragg peaks in reciprocal space to a diffuse intensity that diverges as Q2 (Q = Thki-Q, where Q is the scattering vector, and Thkl the position of the Bragg peak) and is labeled 'Huang' scattering. Its intensity is directly related to the elastic constants and to the formation volume of the defect (in fact, the interstitial in the case of electron irradiation). The symmetry of Huang scattering in reciprocal space is dependent on the defect symmetry in real space. Of course, the information is less clear when several types of defects coexist, which is the general case in alloys. As an example, the X-ray Huang intensity of ordered Fe-40 at% Al is given in Figure 5: it is shown to increase after electron irradiation at 20 K. Analysis of the data suggests that the most probable configuration is the <110> split interstitial with orthorhombic symmetry. The subsequent increase after annealing at 300 K is due to the formation of very small clusters (10 interstitials) (Beaufort et al., 1986). 2.3 Determination of Point-Defect Migration Properties The most popular technique for studying the migration of vacancies in a binary alloy AB is the measurement of

q/h Figure 5. Diffuse X-ray intensity (arbitrary units) measured at 77 K in a single crystal of B2 (cP2) Fe-40 at.0Io Al near the (200) reflection in the direction q = [100] (q in units of h = 27r/lattice parameter) (From Beaufort et al., 1986). • , data before irradiation; A , data for irradiation at 20 K and annealing at 77 K; • , data after annealing at 300 K

(i) If one applies an instantaneous small temperature jump T-*- T± AT, the original vacancy concentration will differ by ACV(7) from its new equilibrium value and will evolve toward it, with a rate depending on the vacancy mobility. Generally d(ACv)/dt is proportional to -AC V , whence CV(T+AT, t) = CW(T+ Ar, t = oo) + ACy exp( - VT). Many properties that depend on the vacancy content will evolve linearly with ACV(T, t) and therefore exponentially with time: electrical resistivity, magnetic susceptibility, length, etc. (see Figure 6). The relaxation time r generally depends exponentially on temperature: T=T0 QXp(EZkT), from which the activation energy E of the process can be determined, (ii) The detailed analysis of isochronal and isothermal recovery of defects created or retained at low temperature (by quenching or irradiation) allows one to obtain the characteristic activation energy of a given migration stage.


atomic self-diffusion coefficients D% and Dg by tracer techniques (see Chapter 32 by Larikov in this volume). Atomic diffusion is generally due to vacancies, and the migration energy of A-type vacancies is estimated as the difference between the activation energy for selfdiffusion E^ and the vacancy formation energy E?A. More precise information on migration energies has been obtained from experiments that require only a small number of vacancy jumps:

I (arb. units)

t (min) Figure 6. Typical semilogarithmic plot of the length variation [L(oo)-L(0] of CoGa measured as a function of time at 773 0C after a 20 0C temperature step increase (From Van Ommen and de Miranda, 1981)

Point-defect migration energies can also be deduced from the growth or shrinkage of dislocation loops during or after irradiation in a HVTEM (see Chapter 22 by Veyssiere and Douin in this volume).

75 eV in the LRO state. CuZn): their concentration is directly related to the LRO parameter value.* The positron trapping increases with boron content. This unusual shape is probably due to the trapping of two different vacancytype defects: either (i) Cu and Zn vacancies (Chabik and Rozenfeld. positron annihilation (the lifetime of a positron in a three-dimensional vacancy cluster increases with the size of the cluster and can reach 500 ps). 1987).4 eV corresponds roughly to the increase of Zn diffusion activation energy between the disordered and the ordered phases (Kuper et al. and 0. density.e. 1987).6±0. and long-range order parameter S (determined by X-ray diffraction) measurements performed at room temperature have shown that the departure from stoichiometry is mainly accommodated by antisite defects.. This has interesting consequences for the interpretation of mechanical properties: trapping of boron atoms near vacancies might explain the lack of ductility of B-doped stoichiometric or Al-rich Ni3Al. deviation from the stoichiometric composition is accommodated by antisite atoms: CuZn. whereas in Ni-rich alloys this does not occur. 3.. where the vacancy formation energy is related to breaking of the first-neighbor bonds (isp(ordered)-!^(disordered) T h i s result is in apparent contradiction with the conclusion of a diffusion study. 1980). the Cu and Zn diffusion coefficients are very near to one another (Kuper et al. and TiAl (Shirai and Yamaguchi. it has to be remarked that it is much larger than predicted from a simple Bragg-Williams model. roughly a few 10~6. increasing their concentration and forming stable complex defects. Ni3Al. 1990). This value is in agreement with recent theoretical calculations for formation of a Ni vacancy (1. boron enriching the grain boundaries and increasing the ductility (Das Gupta et al. The first interpretation seems unlikely as. and Al atoms on Ni sites (labelled AlNi) in Al-rich alloys. In the second interpretation. a precise fit of the lowtemperature data (T<430 0C) requires a temperaturedependent vacancy formation energy (the full curve in Figure 4): Zsv=O. 1980). Coupled lattice-parameter. in the LRO state..2eV (Wang et al. and strongly increase the strength of the alloy (Huang et al.3 at.2.1. 1984). Vo. 1985a. Ni atoms on Al sites (labelled NiA1) in Nirich alloys.. 1988). 1984).4 Recovery and Clustering of Point Defects Recovery of point defects created at low temperature occurs by several intermediate stages.. 3. The homogeneity range of this alloy extends from 23 to 27. very similarly to metals: from the positron trapping rate plotted as a function of the quenching temperature. positron annihilation measurements on polycrystalline samples. vacancies form only thermally. Cu3Au (Doyama et al. The absence of structural vacancies has been checked by positron annihilation in the case of CuZn (Kim and Buyers. either pure or doped with boron. f In rapidly quenched stoichiometric Ni3Al. Positron measurements (angular correlation technique) show an annihilation rate versus temperature with a double S shape (see Figure 4). 1975).1. NiTi. boron atoms remain in solid solution up to 1. but stoichiometric and Al-rich alloys contain a small but detectable (Ni) vacancy concentration. . 1956).% Al. 3.98) (Aoki and Izumi. one obtains the formation energy of a vacancy. This variation of 0.47 eV) (Foiles and Daw. E^ = 1.1 Alloys with No Structural Vacancies For most of the intermetallic compounds. 1992). Nevertheless. In all cases. and Huang X-ray scattering (it increases when interstitials cluster into dislocation loops). Nevertheless.5 at.. Experimental Studies of the Formation of Vacancies and Antisite Defects 3. transmission electron microscopy (when the size of defect clusters is larger than 2 nm). thermal vacancies form at high temperature.. 1981). 1956). i. and has been studied by various techniques: electrical resistivity. which remains LRO up to the melting point..1 Alloys With High Ordering Energies High ordering energies are exemplified by the case of Ni3Al. The stoichiometric compound Ni3Al is nearly perfectly ordered at room temperature (S = 0. b).35eV in the disordered state. or (ii) vacancies and divacancies (Kim and Buyers. Fe3Al (Schaefer et al. Cu3Au. which suggests a small Ni vacancy concentration (~ 10~6) on both sides of stoichiometry (Hoshino et al. differential scanning calorimetry. this suggests that boron atoms or clusters are trapped near constitutional Ni vacancies.5 at.* In Ni-rich Ni3Al (76. show that at room temperature Ni-rich alloys contain no detectable vacancies. Paired antisite defects are also formed thermally in high concentration for those alloys which present an orderdisorder transition in the solid state (Cu3Au. generally involving clusters. tend to occupy interstitial sites.% Ni).2 Alloys with Low Ordering Energies The influence of long-range ordering on vacancy properties has been studied in some detail in /3-brass (CuZn).

Van Ommen et al. On the other hand. 1939. 1975. 1964.10eV in CuAu (Kim. CoGa (Berner et al. determined by composition and not temperature. are they stable at OK?).. An important question is the following: are these defects true structural vacancies (i. Thus the observed vacancy concentration . It can be seen that the number of defects is higher in alloys quenched from 1173 K than in slowly cooled alloys (Berner et al. and that the full curves represent the total (structural plus thermal) defect concentrations. 1985b). assuming two atoms per unit cell. 1973. 1968).05 ±0. Generally. showing that a quenched foil of thickness 0. The concentrations of vacancies and antisite atoms versus Co content are shown for this compound in Figure 7. the positron lifetime of the defects in the quenched Cu3Au sample is 240 ps instead of 200 ps for the single vacancy (obtained by electron irradiation). Fleischer. AgMg has been classified by Chang and Neumann (1982) as an antistructure defect alloy. or are they frozen-in thermal vacancies? It is generally believed that the broken lines of Figure 7 represent the structural vacancies. on the Ni-rich side. In fact. Seybolt and Westbrook. 1981). and NiGa (Donaldson and Rawlings. The vacancy content at room temperature has been shown to depend on the thermal history of the Defect concentration (at.2 Alloys with Structural Vacancies Bradley and Taylor (1937) observed that the density of Al-rich NiAl was smaller than calculated from the lattice parameter.00±0. The vacancy formation energy has also been determined in gold-copper alloys by electrical resistivity measurements in quenched samples (isF=1.04eV (Girifalco.. 1975. Cobalt concentration (at. Concentration of vacancies and antisite defects in CoGa alloys as a function of composition and of thermal treatment (From Schwab and Gerold. are called structural (or constitutional). 500-900 0C (E^= 1.. Certainly more experimental and theoretical effort must be made in this direction. The same behavior was found for a number of other group VIII-IIIA intermetallic compounds with the B2 (cP2) structure (CsCl type) (Neumann et al. 1987) and detailed X-ray diffraction studies (Kogachi et al. in Ni45Al55. 1. 1982). in particular FeAl (Ho and Dodd. 1963). 1969)) and by positron annihilation (angular correlation-type measurements in the disordered state. 1980) sample.54 = 0. 1976. 1978). whereas Hiinecke et al (1987) have detected in this compound a small structural-vacancy content. the extra nickel atoms substitute normally for Al atoms on the Al sublattice.%) =kT c /l.06eV in Cu3Au (Benci et al. Wachtel et al. Wasilewski et al.00±0.%) Figure 7. At room temperature. 1987)). as well as for a few other B2-structure compounds (see review by Chang and Neumann. the lattice parameter peaks at or close to the stoichiometric composition (Cooper. 1992).3. Van Ommen et al. This may be the explanation for some unresolved cases: for example. They concluded that the excess Al atoms do not substitute on the Ni sublattice. CoAl (Bradley and Seager. 1993). The determination of cell occupancy as a function of composition from measurements of specific gravity and lattice parameter is shown on Figure 2 in the case of CoAl. The CoGa alloy has been the most studied experimentally. 1976).11 eV in Cu3Au. 10% of the Ni sites are vacant. but that vacancies must be created on that sublattice. The composition dependence of the vacancy content at room temperature in the NiAl intermetallic compound was checked by positron annihilation (Doyama et al. These vacancies. 1964))..e.4 mm retains small threedimensional vacancy clusters ( — six vacancies) (Doyama et al. 1981).

for compositions that are intermediate between NiAl and Ni2Al3. 1972. or amorphization by implantation. the energy transferred by MeV electrons to nuclei via Rutherford scattering is small and of the order of the displacement threshold energy Ed (~20eV). 4. Jacobi and Engell. purely thermal vacancies accounting for the observed behavior. 1992). Coupled length and lattice-constant measurements (see Section 2.. NaTl-type structure).on the Co-rich side would correspond to thermal vacancies. One can see that. nuclear spin-lattice relaxation) (Brun et al. and C6 (hP3. BS1 (NiAs type).06 < x< 0. 1981). for example.. in ordered or disordered fashion. the thermal-vacancy content is much larger than in a metal: 3. Kogachi et al. the increase of vacancy content below 50% Al at room temperature is very sudden (Doyama et al. In our opinion. even for Co-rich alloys.. the compound contains Ni vacancies.. which is shown as broken curves on Figure 3 for Co48Ga52 and Co56Ga44 (Van Ommen et al. 1987. 1986) (see Section 4). • The fact that.06. It is generally proposed that the creation of two thermal vacancies always occurs in combination with the formation of an antisite atom (Co atom on Ga site): this is the triple-defect structure proposed first by Wasilewski (1968). and LiAl (Brun et al.. A similar phenomenon is found for NiAs and NiSb. and when A:>0. 1983).2) were performed between 800 and 1200 K to determine the thermal dependence of vacancy concentration. 4. 1980). this is contrary to diffusion data. 7-brass phases in the Cu-Al and Cu-Ga systems (above 35% Al or Ga) (Hume-Rothery et al. much larger than in Ga-rich alloys.%). 1971). This has been found in the Ni-Al system: the Ni2Al3 phase is produced from NiAl (B2) by subtracting a Ni plane every three. 1952). 1939. because of the small electron mass. for example. Production and Structure of Interstitials It is important to know the elementary properties of selfinter stitials.. where no structural vacancies exist. it has been supposed that..Ni)Al based on the above B2 alloys (Lipson and Taylor. 1986). if we want to understand in detail and model the behavior of intermetallic compounds under technological irradiation conditions. disordering and dissolution of precipitates in steels used in nuclear reactors. the Ni1+ArSb phase exists at room temperature with . This is. performed . at least in some compounds.. If we suppose that the thermal vacancies are Co vacancies Gust like the structural vacancies). vacancies tend to order and prepare the Ni2Al3 phase (Delavignette et al. then this creation must be accommodated by another type of defect to maintain a constant composition. 1983). This vacancy ordering is responsible for anomalies observed in several physical properties around 100 K (electrical resistivity.Ni)Al and (Cu.5% at 1200 K for Co56Ga44 (see broken curve 2 in Figure 3). Ni2In type).. electron irradiation. where respectively one-third and one-quarter of the cation sites are unoccupied (see Chapter 14 by Parthe in this volume). e. Ni interstitial atoms (Leubolt et al. Co and Ga selfdiffusion coefficients being weakly compositiondependent (Stolwijk et al. because the B82 (hP6. strong arguments for the existence of structural vacancies at low Ni or Co content in B2-structure aluminides or gallides are the following: • The large vacancy content retained at room temperature in Co-rich alloys would require a very large migration energy. Intermetallic compounds with the B82 (hP4) structure (NiAs type) also contain structural defects. NiTe2 type) structures form a continuous transition: this can be made by subtracting Ni atoms from the Ni sublattice. whereas those on the Ga-rich side would be mainly structural vacancies. the case for Ga2S3 (derived from hexagonal wurtzite structure) or CdGa2S4 (derived from cubic zinc-blende structure).g. data from Van Ommen et al. On the basis of thermodynamic calculations (Kim. Therefore. Some tetrahedral compounds contain structural vacancies on the cation sublattice. • CW(T) for Co48Ga52 (measured between 800 and 1200 K) extrapolates to a finite value of 3-4% at 0 K (broken curve 1 in Figure 3.1 Production As mentioned in Section 2. structural vacancies are unlikely.1. When JC<0. 1981). The structural vacancies sometimes order. including CoGa.0. vacancies on the Li sublattice are the dominant defect for Lideficient compositions (48 to 50 at. Taylor and Doyle. Neutron diffraction measurements show that these vacancies order on every tenth (840) plane below 97 K. In /3-LiAl (B32 (cF16). Structural vacancies have been found in other alloys: ternary alloys such as (Fe. • There exist vacancy-ordered phases in the Ni-Al system (see below). in the Ni-Al system.. 1972). specific heat.

I n t n e case of disordered FeCo. 1987). This suggests that there is a unique interstitial..2). 1983b). one must also take into account radiation-induced disordering (Schulson. Riviere et al. 4. The large damage rate in ordered FeCo is then ascribed to the creation of antisite defects. The latter process is extremely efficient. taking into account interatomic (A-A. Defect production by electron irradiation is more complicated in intermetallic compounds than in pure metals: apart from the existence of different types of interstitials (see Section 1).. and o% and o* are the displacement cross-sections of A and B atoms. the data are consistent with a specific resistivity of 0. is the most suitable for investigation of isolated Frenkel pairs. the (average) interstitial was found to be very similar to that in f. 1983a). or successive replacement collisions along crystal directions where A and B atoms alternate «110> in Ll 2 .oTc Therefore from the plot of dc/d<t> versus c for disordered FeCo.. B-B) collisions. the partial filling of which depends strongly on the state of order (Riviere et al. In Cu3Au.2 Experimental Determination of the Self-Interstitial Structure Structures of self-interstitials can be determined by measuring the X-ray Huang scattering (see Section 2. E$u and E%u have been determined experimentally from the dependence of damage rate on incident electron energy. and Ti) have also been shown to have complicated effects in CoAl (Fleischer. Cu3Au. and Cu3Au (Urban and Ehrhart. whence p F = 13 ±3 [iQcm/°/o Frenkel pair (assuming E%° = Ele = 22eV). 1986). metals: a <100> dumbbell with a relaxation volume of about 1. three ordered alloys have been studied by this technique: the Ll 2 compounds Ni3Fe (Bender and Ehrhart. Re. This is due (i) to the existence of a recombination volume V0 within which any vacancy-interstitial pair will spontaneously recombine and (ii) to subthreshold energy transfers inducing athermal migration and annihilation of interstitials (with a cross-section ar). A-B.5 to 2 in ordered compared to disordered copper-gold alloys (Alamo et al. because vanadium forms a resonant virtual bound state at the Fermi level. This can occur by several mechanisms.5 MeV electron irradiation (Riviere et al. 1981). 1983b). ppe-PFO. the average interstitial relaxation volume is much smaller: 0. To our knowledge. The Frenkel-pair resistivity is probably larger by a factor of 1. The addition of a third element to an ordered binary alloy can modify considerably the point-defect properties: for example. in particular uncorrelated mutual recombination between a vacancy on the a sublattice and a B low temperature to avoid point-defect migration. Figure 8 shows the resistivity increase of ordered and disordered FeCo (B2 cubic structure) during 2. 1986). 1986). at low fluence.8Qa. and CuAu (Alamo et al. as the number of replacements per displacement can be typically of the order of 100. In intermetallic compounds.V0c)2 . as discussed in Section 2.. In the latter alloys.. <111> in B2-ordered alloys). the behavior of FeCo-2% V is very complex. one sees that the defect production rate shows a slight tendency to saturate at high fluence.1. Solutes (specifically Mn.c.. if p is assumed to depend quadratically on the LRO parameter S (Muto.c. where both types of atoms are of similar size.5fia. and the B2 structure Fe-40% Al (Beaufort et al. the small Cu-Cu dumbbell. the displaced larger gold atoms transferring to a copper site according to the reaction: Cu-Au dumbbell-*Cu-Cu dumbbell + antisite Au defect . > / (dp/d0) 0=o = a^p^+o§pP where pp and P? are t n e average resistivities of A and B Frenkel pairs. This has the advantage that the specific resistivity of a Frenkel pair is one to two orders of magnitude larger than that of an antisite defect. it is reasonable to assume that the damage comes only from Frenkel pairs. B-A. 1993). the rate of increase of electrical resistivity p with fluence < is. the production of point defects by electron irradiation has generally been studied by electrical resistivity measurements. It can be shown that the Frenkel-pair production rate obeys the following relation: dc(<t>)/d<t> = Pf l dp((t>)/d<t>= a d (l . (1983b) obtain values for the saturation resistivity and F0^250Qa (atomic volumes). On Figure 8. Similar studies have been performed on Fe-40% Al (Riviere et al. In Ni3Fe. The values of o% and o* depend on the incident electron energy and on the threshold displacement energies of A and B. 1936). In a binary alloy A-B. and can be calculated analytically for isotropic ('amorphous') alloys (Lesueur. 1982). which are quite similar to the pure-metal values.07 fiQ cm/% antisite defect and about 30 replacements per displacement. Because of the very similar electronic structure and size of Fe and Co. 1979).

Resistivity change (\iQ. cm)

Fluence (1019Cm"2)
Figure 8. Resistivity increase Ap during 2.5 MeV electron irradiation at 21 K as a function of the fluence in ordered and disordered FeCo (From Riviere et al, 1983a)

This is confirmed by the isochronal recovery curves, which are very similar for incident electron energies of 0.66 and 2.36 MeV (without and with Au displacements) (see Section 7) (Alamo et al., 1986). In Fe-40% Al, the asymmetry of the Huang scattering suggests that the most probable interstitial configuration should be a <110>-oriented dumbbell with orthorhombic symmetry as in b.c.c. Fe or Mo. 5. Theory of Point-Defect Formation Properties 5.1 Empirical Prediction of Structural-Vacancy Compounds An explanation for the occurrence of structural vacancies has been given from the Hume-Rothery rules. One knows that the ordered B2 compounds exist only if the number of conduction electrons per atom ne is near to and does not exceed 1.5. This is the case for stoichiometric NiAl, if one admits that Ni, a transition metal, contributes zero electrons. Then, for an aluminum-rich alloy N i 1 - ^ 1 + J n if non-stoichiometry occurs via AlNi antisites, «e = 1.5(1 +x); the Fermi surface will enter the second Brillouin zone, strongly increasing the internal energy. The alloy energy can be minimized by creating vacancies in the Ni sublattice, ne then remaining equal to 1.5. The qualitative validity of this model is confirmed by the behavior of AgMg, where Ag brings one electron and Mg two electrons, ne being maintained at 1.5 if non-

stoichiometry occurs by an equal number of Ag vacancies and MgAg antisites, as found experimentally (Hunecke et al., 1987). The same explanation applies to the y phase of CuGa, for which ne should not exceed 21/13. Although the Hume-Rothery rules explain many experimental facts, other authors have tentatively explained the occurrence of structural vacancies by sizeeffect arguments (Lipson and Taylor, 1939) or by interionic repulsion (Jacobi and Engell, 1971). Diamond-like iono-covalent tetrahedral defect structures are obtained when the valence-electron concentration (VEC) per atom ranges from 4 to 4.8. Because each anion-cation bond is saturated by two electrons, and each non-bonding anion orbital pointing toward a cation vacancy contains also two electrons, the number y of structural vacancies per atom is simply related to VEC: .y = VEC/4-l (e.g. VEC = 4.8 and y = 0.2 for Ga2S3) (Parthe, 1964). 5.2 Empirical Prediction of Vacancy Formation Energies In pure metals, empirical relationships have been noted between the monovacancy formation enthalpy and either the heat of vaporization, the Debye temperature, or the electron density at the boundary of an atomic cell (Miedema, 1979). In particular, the analogy with a macroscopic hole has suggested the fairly accurate Miedema relationship

where 70 is the surface energy and Qm the molecular volume, and where the effective value of the constant Q, larger than for a macroscopic hole, includes both electronic and elastic relaxation effects. Such simple correlations cannot be directly extended to ordered alloys, as we have seen that vacancies on different sublattices have very different properties. In fact, because electronic long-range interactions are screened in metals, the properties of a vacancy are mainly controlled by its immediate neighborhood. Using this argument, and the analogy with a macroscopic hole previously applied successfully to pure metals, Miedema (1979) suggested that, in a dilute binary alloy A x B 1 .^ (x<l, e.g. LaNi5) with all A atoms surrounded by B atoms, a vacancy on an a site will be equivalent to a mono vacancy in pure metal B: Z^Va _ V £ F - £^F B. Corrections were then introduced in this model (i) to include size effects QA?*QB (the larger the A atom, the larger its vacancy formation energy £p A , see formula (I)) and (ii) to extend it to arbitrary concentration values, a vacancy on an a site then being partly surrounded by A and B atoms. The generalized formula may be written as £?" = (1 ~fa)ErA +/S«Vn B ) 5/6 £F B where £"FAanc* EFB are pure-metal values, and the factor /JJ characterizes the degree to which an a site is surrounded by B atoms (/^ = 0.75 in B2 LRO cubic alloys) (Miedema, 1979). Qualitatively, the Miedema model explains a certain number of experimental facts, such as the occurrence of a very large vacancy concentration on the Co sublattice of CoGa (the formation enthalpy of vacancies in pure Ga being only = 0.5 eV, correlated to its low melting point, 30 0C). But quantitatively, the agreement with the scarce experimental data is not good: for example, in Ni3Al Miedema estimates E?1 = 0.4 6V9 whereas the experimental value (see Section 3.1.1) is 1.6 eV. Reasons for these disagreements might be that the Miedema model neglects specific charge transfers near the vacancy as well as the possibility of strong atomic relaxations. 5.5 Theoretical Computation of Point-Defect Formation Properties Several attempts have been made recently to calculate from first principles the electronic structure of point defects in intermetallic compounds (and from this their

formation and interaction energies), generally the B2 ones (FeAl, NiAl,. . .). Local partial densities of states are calculated for each type of defect from the band structure of the stoichiometric compound, either by the tight-binding method (Masuda-Jindo, 1987), or in the framework of the linear muffin-tin orbital method, using Green's function formulas (Koenig et aL, 1986; Koch and Koenig, 1986, 1987, 1988; Stefanou et aL, 1987) or defect-specified supercells (Gu and Fritsche, 1992). Up to now, these calculations have allowed only qualitative predictions: for example the antisite FeA1 defect in FeAl is correctly predicted to possess a localized magnetic moment (contrary to iron atoms in stoichiometric FeAl) (Gu and Fritsche, 1992). On the other hand, the calculated sign of the interaction energy between Fe vacancies in FeAl is in contradiction with experimental observations (Koch et aL, 1987). Difficulties are found when limiting the defect potential to a simple site. These calculations have to be improved to take correctly into account the charge transfer and charge redistribution near the point defect, as well as atomic relaxations (the latter have been shown to decrease considerably the formation energy of antisite defects in Cu3Au, see Rey-Losada et aL, 1993). Up to now, the main progress in simulating pointdefect properties in intermetallic compounds has been made by semiempirical methods. The first detailed papers in this sense concern the A15-structure superconductors (Moseev et aL, 1983, 1986; Welch et aL, 1984). Although these authors used very simplified pair potentials, qualitative results were obtained, which are discussed in more detail in Section 9.1. The best results have been obtained by embeddedatom-type methods, applied first with good success to many metallurgical properties of pure metals: surface energy, point-defect properties (see for example Foiles et aL, 1986; Chapter 4 by Voter in this volume). In these methods, the energy of each atom is computed from the energy F1(P1) needed to embed it in the localelectron density p, provided by the other atoms of the alloy (approximated by the superposition of atomicelectron densities: P1=T1J^iPj(R0)), plus an additional electrostatic short-range core-core repulsion <t>0(R0) = Zi(R0)Zj(R0)ZRiJ. The total energy is then written as ^tOt=E Wi)+T S *(,•(*<,•) The function F is deduced from the equation of state of the pure metals, known from their sublimation energies, atomic volumes, and bulk modulus. Parameters describing the effective charges Z1(R0) and the numbers of s, p, d electrons of each type of atom involved in the Pj(R0) are computed from the fit of

selected properties of the pure metals (elastic constants, vacancy formation energy) and of the ordered alloy. Two types of calculations are made: (i) energy minimization at OK, (ii) Monte Carlo simulations to produce the thermal equilibrium distribution of the atoms at a given temperature. This can be applied to any atomic distribution, disordered or ordered, with or without defects or discontinuities. Calculations are typically performed on box-like volumes of ~ 1000 atoms. Foiles and Daw (1987) applied the method to Ni3Al. They could correctly construct the Ni-Al phase diagram, and calculated various point-defect properties: good agreement with experimental data (Wang et al., 1984) was found for the formation and migration energy of the Ni vacancy; the Al vacancy formation energy was found to be lower than that of the complex consisting of a Ni vacancy and a NiA1 antisite defect. A very detailed work was published recently on CuTi and CuTi 2 body-centered tetragonal structure alloys (Shoemaker et al., 1991), unfortunately systems with no experimental information on point-defect properties. In both compounds, the removal of a Cu or Ti atom results in a vacant Cu site, with an adjacent Ti Cu antisite defect in the latter case. Interstitials have complicated structures of the crowdion type, on a Cu (111) row, which involves seven Cu for six sites (CuTi case) or five Cu for four sites (CuTi 2 case) and the creation of two or three antisite defects in the respective cases of Cu or Ti displacements. The embedded-atom technique was applied to the study of inter stitials in NiAl (Caro and Pedraza, 1991) and Ni3Al (Caro et al., 1990; Pedraza et al., 1991). In Ni3Al, the enthalpy differences between dumbbell, octahedral, and crowdion Ni interstitial configurations are small, typically 0.1 to 0.2 eV. The <100> Ni-Ni dumbbells in pure Ni planes (Ep= 3.63 eV) are favored: Ni-Al and Al-Al dumbbells have larger formation energies (4.45 and 6.22 eV respectively, which is understandable on the basis of size arguments) and hence convert into Ni-Ni dumbbell + Al Ni antisite defects. This behavior is very similar to that studied experimentally in Cu 3 Au (see Sections 4 and 7). Threshold displacement energies were computed for Ni3Al (Caro et al., 1990); these values should not be considered as quantitative because of their sensitivity to interatomic repulsive potentials, which were determined only from the elastic constants. Nevertheless, it is found (i) that <100> and <110> are easy directions for defect production as in f .c.c. pure metals ( £ d ^ 2 0 to 3OeV), (ii) that <110> N i - A l - N i - A l collision sequences are efficient, and (iii) that the

displacement process can lead to various metastable interstitial configurations (octahedral, crowdions, . . .). The spontaneous recombination volumes were calculated for ordered and for disordered Ni3Al, and were found to be much smaller than in pure metals, which should lead to larger point-defect concentration at saturation (i.e. at high fluence) and have consequences for radiation-induced segregation and amorphization.

6. Thermodynamics of Point Defects In order to model kinetic properties, it is important to know the nature and concentration of point defects in given experimental conditions. Up to now, there has been no satisfactory theory of the equilibrium number and distribution of point defects in intermetallic compounds as a function of the (long-range or shortrange) ordered state. The available models, limited to binary alloys, are all based on the Bragg-Williams mean-field theory of order, assuming (somewhat naively) pair interaction energies, generally limited to first neighbors. 6.1 General Formalism Let us consider a binary ordered alloy with two sublattices, labeled a and /3, preferentially occupied by A and B atoms, respectively. Because of their low equilibrium concentration, interstitials are neglected. We shall adopt the following terminology due to Kim (1984): Na9 N0 NA, NBi NY NKJ 6 rf One has NA = NAa + NAp Na = N^ +N1ta + Nya NR = NBa + NB0 Ng = N^N1V +Ny1, number of a or /3 sites number of A atoms, B atoms, vacancies number of Kth species on yth sites W w N^, NWa, NAS, / V 7VV/3 deviation from stoichiometry long-range order parameter

In the Bragg-Williams approximation, the configurational entropy is given by / S = kBln( \



Np\ ? NA,\ NB0\ Nw,\

\ (2) /

We shall restrict discussion to the case of B2-structure alloys, where a sites are at the corners and /5 sites at the centers of the cubic lattice cells, each a (resp. 0) site being coordinated to eight 13 (resp. a) sites. One then has Ak = A/^= 1/2, and


7VA + 7VB


Introducing the first-neighbor energies eAB, e ^ , and eBB, and the ordering energy e = eAB - ( e ^ + e BB )/2, the internal energy is then equal to

(NAaNm + NBaN^)zAE


if one assumes that bonds between a vacancy and its neighbors are broken (eVA = eVB = 0). Some authors (Neumann et ai, 1976; Edelin, 1979) have refined this model, taking into account 'ghost' interaction energies between vacancies and A and B atoms: eVA, eVB. 6.2 Calculation of the Vacancy Concentration in Ordered Alloys A first approach to this problem is given by the vacancyenergy model of Cheng et al. (1967). Although, as discussed below, the model is incorrect in its original form, at least if the vacancy formation energies on a. and /3 sites are very different, it gives interesting information. The formation free energy g for a vacancy on a given site in an alloy is mainly dependent on its local atomic environment. Let us assume that g depends only on the nature of atoms directly coordinated with the vacancy. Distributing at random NlWa and N'Wp vacancies on a or (3 lattice sites coordinated with / B atoms and (z-i) A atoms, and giving to each of them a free energy gia (resp. g/j8), one finds by minimizing the total free energy relative to the vacancy concentrations N'Va and
AT' • * /Vy^.


= ^^0+^-2^-/(l-^25)'exp(-||)

> y i ) vy





*For disordered dilute alloys A1 _ C B C (TJ = 0, c< 1), this reduces to the model developed by Lomer (1958) for vacancy-impurity interactions.

",a ^,a ^Vf 7 ZpFigure 9. Distribution of the various vacancy concentrations on a (Cu-rich, ) and /3 (Zn-rich, ) sites of the stoichiometric CuZn alloy calculated from the modified vacancy-energy model. Each 'curve' corresponds to a value of the LRO parameter rj (and of temperature T). The abscissa shows the number of Cu atoms coordinated to the vacancy. Pair interaction energies are given in the text (From Cheng et al., 1967) As up to now the 2(z+ 1) gia and g//3 coefficients cannot be experimentally determined or theoretically calculated, for numerical calculations a simple bondbreaking model has been applied (modified vacancyenergy model of Cheng et al. (1967)), where the bondenergy change when displacing an A or B atom from the lattice to the surface is - [ ( z - / ) e A A + 'eAB]/2 or - [ (z - /)e AB + /eBB ] / 2 respectively. Applying this model to CuZn, with eAA= eznZn = 0.34 eV, eBB = eCuCu = 0.88eV, and eAB = 0.62eV, it is found that, above the critical ordering temperature TCi the concentrations of vacancies on a and /3 sites are indistinguishable, but that, below Tc, NWa and 7VV/3 become different, with fewer vacancies on a sites. The average vacancy formation energy increases with the LRO parameter TJ. The distribution of the various vacancy concentrations is given in Figure 9. It is seen that the most populous types of vacancies change with t] or temperature. For example, for 7 = 0.5 (7Vr c ^0.85), 7 vacancies on a. sites (Zn) are mostly surrounded by one


+ (Nm-NA0)

it is found that K1 varies as exp(\6e/JcT).Al/1) = (NYa + NV$)/2N which is proportional to exp(. and the concentration of a vacancies xyct = NYoc/2N obeys the equation (deduced from the above massaction law): (xVa)3/2 + 6(jcVa)2 . • For non-stoichiometric alloys. vacancy and /3 antisite-defect concentrations obey the mass-action law N^(NyJ2 = K4. the defect concentrations depend only on the activation energy Q. the formation energy of a pair of antisites (B in a and A in 0) in the stoichiometric compound is 16EA8-SeAA-SeBB. • At the stoichiometric composition (6 = 0). 19^7. Nw(NvJ = K4. and those on 0 sites (Cu) by eight Zn. This simple model was fitted to the vacancy and antisite-defect concentration data of CoGa and NiGa. it can be shown (Edelin. and antisite defects on the A-rich side (see Section 3. 1987). This low Q value explains the high defect concentrations. This model is altogether general and simple (Bragg-Williams approximation. Of course. and e allowed the calculations to reproduce very satisfactorily the total vacancy concentration measured in these alloys for several compositions and temperatures. pair interactions. NAQ. because it does not take into account local rearrangements of atoms around a vacancy after its creation. 1986. 1986. The triple-defect model. eBB. FeAl. In particular. Bakker and Van Ommen. 1979). and 0. The following relations have been found at low defect concentration: • The a.1 for configurational entropy S and bond energy E. the criterion for preferential disorder by vacancies (triple defects) is e BB <0 (repulsive B-B energies). but a linear combination of the latter with the antisitedefect formation energy.3 The Triple-Defect Case As strong experimental arguments exist that in B2 alloys non-stoichiometry is accommodated by vacancies (triple defects) on the B-rich side. it is assumed that N^ = Ny0 = O.Cu and seven Zn. 1979) that the defect concentrations obey mass-action laws: NA0NBa = Kly NVaNW0=K2. 1979. many pure-B surface layers will be formed when creating = 10~3 vacancies if E^<E?a. described in Section 3.89 atomic volumes for Co replacing Ga. Starting from equations (2) and (3) of Section 6. bond-breaking model for describing vacancy creation). 6.2). . Therefore positron annihilation or Simmons and Balluffi-type experiments do not give the vacancy formation energy.3QZkT) = 0 As long as 5< 1. The fit of density data gives the formation volume of the defects in CoGa: 0. Another process must occur simultaneously to maintain the nominal superficial composition. one antisite in /3) is 16eAB-8eAA. as noted by Kim (1984). Cu3Au. In the case of small deviation from stoichiometry (8<\). NvJFVfP = K3. S = (N^-Ny01Zl)ZN. the vacancies were shown to be entirely on the transition-metal sublattice. the agreement is fairly good with Q = 0. where 3Q = 2E?a + E^ is a linear combination of the a vacancy and /3 antisite-defect formation energies (E^01 = 4eAB.69 atomic volumes for the Co vacancy. he shows that the free energy F=E-TS depends on only three independent variables: N^. where the A^s aie composition-independent. Let us recall that this vacancy-energy model is incorrect.29 eV for NiGa (Edelin. N^019 and n = (NWa+NW0)/2. Therefore. NWa = 2/VA/?. In the case of the B2 compounds. Indeed. and it was suggested that they are thermal vacancies with low formation energies rather than structural vacancies (see discussion in Section 3.265 eV for CoGa and 0. Edelin. NBa. 1987) applied his model to NiGa. and the vacancy concentration is thermally activated: NYa varies as exp(-QZkT). In this model. The curves of Figure 9 are asymmetric (between a and 0) because eCuCu>eZnZn. several simplified models have been developed to describe this behavior (Neumann.Q/KT). CoGa. and CuAu. Kim (1984.2). due to the creation of vacancies (measured in Simmons and Balluffi-type experiments) is 3(Aa/a . 1980. as the ratio of the number of surface sites to the total number of atoms in a typical crystal is of the order of 10~8. and that of a triple defect (two vacancies in a. Van Ommen et aL9 1981). Minimizing F with respect to these three variables gives a set of equations that have to be solved numerically to obtain the defect concentrations: NVa. the crudeness of these models requires that they should not be taken too quantitatively. This is taken into account in the model developed by Kim (1984.3QZkT). • The difference between the relative change of lattice parameter and the linear expansion of the crystal. Reasonable input parameters e ^ . satisfies this requirement. 1978. NW(3. Is^ = SeAB-SeAA in the bond-breaking model).exp( . and K4 is proportional to exp(. In these models.2 and depicted in Figure 1. The same criterion is found for accommodating B-rich non-stoichiometry by vacancies on a rather than by B on a antisite defects.

and in others it is clear that interstitials are much less mobile. 1987).\ eV). Its recovery stages. The positron lifetime is maximum at 390K and is consistent with clusters containing more than six vacancies. Other alloys such as Ni3Al (Dimitrov et al. Recovery of Point Defects After Irradiation or Quench In pure metals. Chang and Neumann. The most-studied compound is certainly LRO Cu3Au. the analysis of the Huang scattering shows that the interstitial clusters remain very small (single and di-interstitials) (Urban and Ehrhart.. which is consistent with an interstitial migration energy of the order of 1 eV. 1981.36 MeV electron irradiation at 21 K for ordered Cu3Au (From Alamo et al. the interstitial freemigration stage seems to be around 500K. Isochronal recovery of electrical resistivity versus annealing temperature after 0. the copper selfinterstitial (Alamo et al. • The positron lifetime increases suddenly. much more than in pure metals (Riviere et al. the vacancy migration leads always to an increase of the . 1983a). In FeCo and Fe60Al40. these Ap (T)ZApj T(K) Figure 10. Good agreement is found with experimental data (Neumann. 1983b) behave similarly to Cu3Au. strongly suggesting a single type of defect. evidenced by the study of isochronous annealing of irradiated materials. • The resistivity shows a large annealing stage and becomes lower than the unirradiated value (Alamo et al.. 1986) clusters evaporate and single vacancies order the sample during their long-range migration toward sinks (Doyama et al.36MeV electrons (see Figure 10). 1986). indicating that interstitial clusters have disappeared. or bound to impurities. where resistivity decreases only slightly. 1982). the recovery of point defects created or retained at low temperature follows a general scheme. Irradiated CuAu and CuAu3 have not been studied in detail. During stage II (100-250 K). Some behave very similarly to metals.. 1977. 1986). and at 100 K have either recombined with vacancies. Van Ommen et al. this is consistent with a spectrum of energy barriers for the movement of the interstitial (Alamo et al. Electron-irradiated disordered Cu3Au shows a similar behavior concerning vacancy stages. 1985a). but do not behave as does Cu3Au: CuAu shows a more complex stage behavior (Alamo et al. interstitials become mobile around 50-100 K (ElM = 0. Instead of definite stages.The model was also used to calculate thermodynamic activity and partial enthalpy. indicating vacancy clustering (Doyama et al.. and increasing the long-range order of the alloy during their migration. Typically. 1985a). in a metal such as copper. measured by electrical resistivity. are practically identical after irradiation by 0. The recovery curves are very similar to that of Cu pure metal. or clustered in small dislocation loops.. as in Cu3Au and CuAu. In the case of Fe60Al40. 1986). These three observations are consistent with vacancy migration starting from 250K. vacancies either annihilating on interstitial loops or forming vacancy clusters. and CuAu 3 shows no vacancy clustering (Doyama et al. but the lowtemperature recovery is very different from that of LRO Cu3Au. This is because the mobility of interstitials is much higher than that of vacancies. 1992) and FeCo (Riviere et al. Vacancies become mobile above — 250K and form clusters that dissociate at high temperature (^500K). 7. 1985a). above 410 K.. Information on intermetallic compounds is fragmentary. one observes a continuous recovery from 50 to 250 K. At 250 K: • The Huang signal disappears. suggesting that the 40 and 70 K stages are due respectively to close Frenkel-pair recombination and free interstitial migration.66 or 2. 1986).66 and 2.

pure metals.. For example. owing to its heterogeneous character or to competing mechanisms. This is because (i) vacancies of the considered element have a small formation energy and predominate. The kinetics of vacancy elimination. including short.2 and 1.5eV. Cu vacancy in Cu3Au. 1975. For random disordered binary alloys. one has to take into account jumps within each sublattice as well as between different sublattices. the similarity of these values with those of pure nickel is somewhat fortuitous. Migration Properties of Point Defects 8. /A(B) the correlation coefficients. respectively..e. The mean-field expression for self-diffusion coefficients has been extended to several LRO intermetallic structures by Bakker and Westerveld (1988). and presence of inclusions (which help nucleation of vacancy loops). Of course.g. shows either a simple (single activation energy) or a complex behavior. In the case of A-type vacancies in A3B compounds with the Ll 2 structure (e. (10 O) or (111) dislocation loops. which have been extended to short-range ordered alloys (Stolwijk. when an A atom and a vacancy trap each other and exchange their positions frequently without long-range migration. Cupschalk and Brown.35 ±0. and helices. In some cases. 1988). the vacancy mobility is very similar to that in a pure metal. but its successive jumps are correlated. the mean-field approximation is particularly insufficient to describe vacancy migration when CA< 1 and WA/WB>1. 1973. In Ni3Al. 1967). and by small-angle X-ray and neutron scattering in NiAl-based alloys (Epperson et al.. Various types of defects are observed: voids. including numerical simulation. Dimitrov et al. Hoshino et al. As remarked by Bakker (1987). one has CA L where W°£ is the exchange rate between a vacancy on a /3 site and an A atom on an a site and ph% is the probability of finding a vacancy on a j8 site near neighbor of an A atom on an a site (W™ andp%% are defined similarly). i..c. the atomic selfdiffusion coefficients have been calculated in a meanfield approximation by Manning (1971): ^A(B) = ^S ^VZA(B)^A(B) . for the f. and (ii) three-dimensional transport of A atoms is possible via first-neighbor jumps of vacancies on the a sublattice as in f. observed by electrical resistivity (Riviere et al. Wang et aL. we shall only discuss transport at the atomic level. In this case. 1992). Vacancy clusters formed during annealing of quenched samples were observed by transmission electron microscopy in Cu3Au (Condat and Fayard. b. and the different types of jumps with different jump lengths and different exchange rates of the vacancy with its neighbors. 1968. 1986). Its analytical form depends on the existence of defect cluster nuclei in the quenched samples. measured by the time dependence of defect cluster density and size. 1977. 1981). the sum of Ni vacancy formation and migration energies. Very different behaviors are observed. 1957). (111) loops are dissociated into two concentric partials separated by a stacking fault (Junqua et al.6 ±0. CuZn (Cupschalk and Brown. Eibner etal... Junqua et al. depending on the crystal structure and on the type of atom considered: (i) Ll2 structure. A correct description of vacancy migration in an alloy must take into account its true defect structure. atomic diffusion being treated in detail in Chapter 32 by Larikov in this volume. Ll 2 structure A3B. Weber et al.15 eV. Epperson and Loomis..c. 1972) and in a number of B2-structure alloys: FeAl (Wechsler. in most cases.and long-range order.1 Vacancies where s is the first-neighbor jump distance. as the formation of a Ni vacancy in Ni3Al Owing to the values of formation and migration energies.c. determined experimentally (1. 1957. 1977). 1975).. and WA(B) the exchange rates of the vacancy with an A or B neighbor (assumed independent of the local surroundings). NiAl (Ball and Smallman. The movement of a vacancy is no longer a random walk (as in pure metals). CoGa (Schwab and Gerold. it is generally assumed that. Ni vacancy in Ni3Al). Their occurrence seems to depend on several factors: quenching temperature. 1981).c.. 1983a. 1978. 1980). and AuCd (Wechsler. assuming near-neighbor jumps. surface oxidation. 1983) and in disordered Cu3Au (Doyama et al. atomic transport in intermetallic compounds occurs via vacancy migration... 1985b). 8. 1967). Weber et al.LRO. is 15% lower than the Ni selfdiffusion activation energy (iiD —3. Recovery of vacancies retained by quenching has been studied by positron annihilation in Ni-rich NiAl (Wang et al. Alamo et al. Marshall and Brittain. 1984. 1973. The correlation coefficients have been calculated by various methods. Here.

BIl (tP4) CuTi. This can be due to experimental uncertainties. In highly ordered alloys. and its migration to near-neighbor Ni sites involves a saddle point with two Ni and two Al first neighbors. numerical simulations confirm that a Ni vacancy jumps more easily into a nearneighbor Ni site (1. self-diffusion has been measured in a few ordered B2 alloys. (c. • The transport of matter can occur preferentially on the sublattice that contains fewer vacancies. (iii) LlOt BIl3 and CIl6 structures..02 eV) than into a near-neighbor Al site (1. it is assumed that NN jumps are allowed for Co. Various atomic mechanisms have been proposed for diffusion of vacancies in B2-ordered alloys (see Figure 11): (a) Vacancy nearest-neighbor (NN) jump between an a and a /3 site. (g) Two correlated NN jumps through a divacancy (pair of vacancies of both sublattices). • The correlation factor /deduced from isotope-effect measurements is shown to be small (AuZn. see Fishman et al. see Gupta et al. One will have many forward and backward jumps of the vacancy. see Hilgedieck and Herzig. Let us summarize the main results of these studies: • In a given alloy. the correlation effects for diffusion will be high (fA. or (iii) different competing mechanisms (Stolwijk et al. which preserves the local order. inefficient for mass transport. the self-diffusion coefficients of the two atomic species generally only differ by less than one order of magnitude. see Hagel and Westbrook.28 eV) (Foiles and Daw. • The diffusion coefficients are often minimum and the activation energy maximum at (or near to) the stoichiometric composition. which are made by stacking pure A and pure B atomic planes along the (001) direction. Nevertheless. In the B2 structure. (e.9 1980). 1984). 1991) have confirmed this for CuTi. (1961). a sites are entirely surrounded by (eight) /3 near neighbors and vice versa. where the migration energy of the Cu vacancy (via NN jumps) is calculated to be 0. which does not preserve the local order. Recent computer simulation studies (Shoemaker et al. which favour diffusion: the composition dependence is then non-symmetrical. This has been tentatively explained by (i) an increase of the vacancy migration energy with the LRO parameter S (Girifalco.. Numerical Monte Carlo simulations have shown that this mechanism is not very efficient because in fact the various potential barriers oblige the vacancy to perform a large number of backward jumps (Arnhold.9 1956). • In alloys presenting an order-disorder transition (case of CuZn. which allows NN jumps without altering the local order. This can be understood (i) by the fact that EF and EM increase with S9 which is maximum at stoichiometry. as CoGa antisites are allowed. 1987). 1965). (1983). 1981). but also to the fact that these energies are averages and apparent energies deduced from a fit within a narrow temperature range.19eV within Cu .d) Six-jump cycle proposed by Elcock and McCombie (1958) and Huntington et al. the Arrhenius plot of InD versus \/T shows a change of slope at the transition and a curvature in the LRO state. as shown by computer simulation of the correlation factor/(Bakker et al. see Kuper et al.f) Triple-defect jump proposed for CoGa by Van Ommen et al. (ii) B2 structure. In the tetragonal structures of Ll 0 (tP4) CuAu. and (ii) in some cases by the existence of structural vacancies in nonstoichiometric alloys. the difference is somewhat larger in alloys containing structural vacancies.9 1970). the reverse jump that reestablishes the chemical order will be very easy. / B <1). but that Ga diffuses via NNN jumps into the divacancy. but GaCo are forbidden. (ii) a decrease of the correlation factor in the LRO state (Bakker. 1982) and to decrease with T below the order-disorder critical temperature (FeCo. Here the intermediate configuration consists of a NN divacancy (VCo + VGa). 1967).involves the breaking of eight Ni-Ni bonds and four Ni-Al bonds. model proposed for PdIn by Hahn et al. Threedimensional migration involves creation of antisite defects and disorder. • The activation energy for self-diffusion can be different from the sum of vacancy formation and migration energies (case of AuCd. 1964). (b) Vacancy next-nearest-neighbor (NNN) jump between sites of the same sublattice. Experimentally. diffusion of an element in its own sublattice via NN jumps can only occur in two dimensions. which suggests that their migrations are coupled. because the corresponding migration energy is very low (case of AgMg where DAg<DMg. 1976). and Cll b (tI6) CuTi2.. Thermal disorder (antisite defects) is then important to help atomic diffusion in B2-ordered alloys. If an A atom jumps on a (5 (vacant) site. (1981).

Jump mechanisms for ordered B2 alloys: (a) nearest-neighbor jump.f) possible jumps for a triple defect.d) sixjump cycles. (c. (O) A atoms. ( • ) vacancies (From Bocquet et al. ( • ) B atoms.Figure 11. (e. (b) next-nearest-neighbor jump.t 1983) .

Jennaneef al. 5 . 1991). Sitaud et al.2eV. assigned to interstitial-defect migration: 0. 1981): £ j ^ 0 . Ti migration requires first creating a vacancy on the Ti sublattice. measurements were performed over a relatively large temperature interval (600-800 0 C) and showed the existence of at least two processes with different activation energies ( £ ^ . • In stoichiometric Ni3Al. E ^ = 1. • Resistivity and positron trapping rate measurements in Ni3Al (120 0 C stage. see Figure 6).2 Interstitials The migration energy of interstitials has been measured by various techniques in a few irradiated ordered or disordered concentrated alloys: • In disordered f..Cu Ti Vacancy • Electrical resistivity measurements in FeAl (610 0 C stage. The latter value is lowered to 0. sheets and 1.c. (b) migration path of Cu vacancy through the Ti (001) planes. In the latter experiment. Doyama et al. In a few cases. 1974) and Cu3Au (250 0 C stage. 1987) and TEM measurement of the depleted zone at the edge of a thin foil (Housseau and Pelissier.. 1981. Cu3Au by electrical resistivity (Gilbert et al.2 and 3.. 1983a) and TEM study of the growth of dislocation loops (Kinoshita et al. 1981).02eV (Sitaud et al. .75eV.. Fe-Cr-Ni austenites with various Ni and Cr contents by electrical resistivity (Dimitrov and Dimitrov.1 .. 1991) 8. by Zener relaxation (Beretz et al. two activation energy values were determined in the resistivity recovery stage. Fe-40% Al by electrical resistivity (Riviere et al.. Self-diffusion activation energies (Stolwijk et al. • In ordered b.c. • In ordered f. 1966).c.. (c.c. 1991).3 eV. 1984. 1981) measured in CoGa form a consistent set of data. Dimitrov et al. as described in Section 2. vacancy migration energies (Van Ommen and de Miranda.d) the same. AgZn^.c. Figure 12.32 eV across Ti sheets..75 or 0. 8 eV.5-4 eV) attributed respectively to NN cobalt-•vacancy and NNN gallium-• vacancy jumps.c.16 ±0.71 +0.c. The increase of E^1 with long-range order parameter S is in qualitative agreement with the theoretical predictions of Girif alco (1964). 1987) and NiSb (285 0 C stage. Vacancies in CuTi: (a) the perfect crystal. and formation energies (Van Ommen et al.5 to 0.02 eV and 0. • Length after-effect in CoGa (Van Ommen and de Miranda. El1= 1.. 1982): E]4^l to 1.61 S2. assisted by one or two CuTi antisite defects (From Shoemaker et al. iE^«1.9 eV. 1973): ElM~0. Benci and Gasparrini.c. 1980).2: • Magnetic after-effect in CoGa (Van Ommen et al..36 S2. Riviere and Grilhe.. 1977).20 + 0.66 eV if one or two Cu antisite defects are present on the Ti sublattice (see Figure 12).15 eV. £ ^ = 0.28±0. 1992). As the Ti vacancy is calculated to be unstable relative to the complex (Cu vacancy) + (CuTi antisite). • In disordered f. Vacancy mobility was also studied by isothermal after-effect measurements following small temperature increments. E^ IASeV. vacancy migration energies have been determined by isothermal variations of a physical property measured during the thermal recovery of a quenched or electron-irradiated sample: • Differential scanning calorimetry experiments on NiAl (480 0 C stage..35 ±0. 1983): Elu « 0.

explaining therefore . . indicating a strong slowing of the interstitial. by the same mechanism as in pure f. no damage. Models have been developed for the migration of <100> dumbbells in disordered f. of considerable interest: the Al5 superconductors. causing a decrease in the density of states at the Fermi level N(EF). where fragmentary Huang scattering data suggest a similar interstitial as in b. we have discussed mostly 'model' intermetallic compounds with simple crystal structures (generally cubic: Ll 2 . Configuration and migration of the Cu-Cu split interstitial proposed for ordered Cu3Au.2 eV) and of the order of the vacancy migration energy. typically for x<0. 1986) Up to now. for £"e>0. of the electron-phonon coupling. and centered on the cube face perpendicular to the X axis (From Alamo et ah. For example.1 A15 Compounds The superconducting properties of Al5 compounds are extremely sensitive to point defects. the superconducting critical temperature Tc is maximum at the stoichiometric composition and decreases after a quench (Bakker.5 <E e < 0.g.3. we have seen (Section 4) that the interstitial probabjy has a unique structure. We shall now present briefly some properties of point defects in more exotic systems. I r [010]-»I z [001]). Beyond a percolation threshold X0-0. i. . crowdion type) have been proposed for the interstitial. and HI-V semiconductors (e. Fe or Mo «110> dumbbell). If the B-B interstitial configuration is of lowest energy. both p and Tc change during irradiation. or A-B <* B-B. one encounters three domains: for £'e<0. only the lighter Ge atoms are displaced. suggesting a high activation energy for threedimensional migration.8MeV. for 0. the migration of this interstitial requires changes of configuration such as A-A ** A-B. transition-metal carbides and nitrides. complex structures (i.c. in an effective-field approximation (Bocquet. 9. and of the critical temperature J 0 . 9. When increasing the incident electron energy EQ.e. where the superconducting properties are due to narrow bands formed with the 3d orbitals of the transition-metal chains. This can be understood in the frame of the model of Labbe and Friedel (1966). B interstitials are trapped in the clusters and their long-range migration requires detrapping. this activation energy is much larger than in pure metals (Elu = 0. 1981).c. by a rotation of 90° (for example.c. the interstitial mobility greatly increases toward typical pure-metal values. The labelling Ix [010] means the dumbbell is oriented along the <010> direction.c. when the transition metal Nb is also displaced. Specific Systems Figure 13. GaAs).8 MeV. This is direct proof that point defects in the Nb chain With the exception of LRO Cu3Au and Ni3Al.e. metals. such as Fe-40% Al.e. in V3Ga. Point defects in the chains decrease the electronic relaxation time and. In B2-ordered alloys. solid solutions A 1 ^B x . B2. as long as the B clusters do not extend to the whole lattice (i.5MeV. In other ordered alloys.y 1986). . the electrical resistivity p increases but the critical temperature remains unchanged.c.1 to 0. according to the uncertainty principle.) and containing two metal species. can migrate toward a configuration of the same energy centered on one of its eight Cu first neighbors. and = therefore a high activation energy consistent with experimental data. with [010] orientation. the <100> Cu-Cu dumbbell.the existence of a well-defined recovery stage at 70 K (Alamo et al. broaden the narrow bands. We can see in Figure 13 that the interstitial centered in the middle of the cube face Y.c. 1987).3). This was clearly demonstrated by low-temperature electron irradiation experiments on Nb3Ge (Rullier-Albenque and Quere. 1986). In the case of LRO Cu3Au.

(1984). 1990). Nb2C. 1983. carbon sublattice and each Nb atom tends to be surrounded by five carbons and one vacancy. because they interact with free carriers. which gives rise to diffuse streaks in the electron and neutron diffraction patterns (de Novion et al. limited to a few neighbor shells.c. 1983). and have a large effect on electronic transport. 1987. V2C. 9.are responsible for the decrease of Tc. etc. Surface and interface defects due to processing are also important. 1990. (de Novion and Landesman. very short-range local order remains. with a much lower migration energy along the Nb chain (<1 eV) than between chains (~4eV). the Nb vacancy migration is found to be nearly one-dimensional.. the techniques for studying point defects in these materials are quite different from those presented in Section 2: they consist mainly of electrical measurements (conductivity. focused collision sequences along <120>type crystal directions (• • • Nb-Ge-Nb • • • NbGe-Nb • • •)• Similar effects were observed on V3Si. (1985) from a theoretical analysis of the band structure.e. 1988) experiments have also allowed local structural distortions around a carbon vacancy to be determined. long-range order for special compositions (V8C7. between Sn first-neighbor sites (distant by #V3/2).. For example. where the Tc decrease is ~ 2K/at. The structure calculated by Welch et al.3 GaAs and Other III-V Semiconductors Bulk point defects have a considerable importance in HI-V semiconductors such as gallium arsenide GaAs. acting as scattering centers. Hall effect. A large fraction of these octahedral sites (up to 50% in the cases of TiC0 5 and TiN0 5) are unoccupied and may be considered as * structural' vacancies. in TiCx. vacancies tend to remain third neighbors on the f. irradiation induces large lattice distortions. de Novion et al. the activation energy for self-diffusion is a pure migration energy ED = EM ( = 3 to 4 eV). in NbC083. Ti2C. Owing to their non-metallic character and to the very low defect concentrations. and V6C5 cubic carbides. even at very low concentration. . with the carbon or nitrogen atoms in the center of the octahedral interstices. 1985). as several types of jumps must be taken into account: within the Nb chain (jump distance a/2).c. the C atomic diffusion coefficient is simply proportional to the vacancy content. 1986). 1987).^B x . with superlattice reflections. between Nb chains (J um P distance oV6/4).2 Transition-Metal Carbides and Nitrides Many refractory transition-metal carbides and nitrides can be described as a close-packed metal lattice (f. Chemical bonding and transport properties depend strongly on the structural-vacancy content. 1990). In the split vacancy model of Welch et al. the main interaction is a repulsive one between secondneighbor carbon vacancies that are distant by <100)a. In fact. Both atom species (metal and carbon) diffuse by independent.c. The Sn vacancy is found to be metastable and to convert into a split Nb vacancy plus an antisite Nb8n. and from Nb to Sn site or vice versa (jump distance aV5/4). antisite defects being mostly created by short. 1987. Welch et al. (1984) for the Nb sub lattice vacancy in Nb3Sn is interesting: the Nb vacant site is suggested to split into two * partial vacancies' separated by a segment of Nb chain translated by a/4 (half the Nb-Nb distance) and forming a sort of stacking fault. The atomic configuration of point defects and the surrounding lattice distortions have been computed numerically in Nb3Sn and Mo3Si (Moseev et al. 9. In the disordered state. the low value of the metal diffusion coefficient and the high activation energy (ED .c. of Bl (cF8) NaCl type or h. This certainly plays a role in the decrease of N(EF) and Tc.c. i. Effective interatomic ordering energies have been deduced from the elastic neutron diffuse scattering spectra and are in qualitative agreement with those deduced by Landesman et al.p. A detailed analysis shows that both vacancies and antisite defects contribute to this. Vacancy migration is complex in an A15 compound such as Nb3Sn. channeling (Kaufmann and Meyer. 1986. Atomic diffusion has been extensively studied in some of the cubic carbides in the disordered state (see review by Davis. Metal diffusion requires a vacancy-creation term as in simple f. For carbon. and Ta2C hexagonal hemicarbides) below a critical temperature.).. random migration of defects on the respective sublattice of each component. metals. and extended X-ray absorption fine structure (EXAFS) (MoisyMaurice and de Novion. 1984) using pair interactions.. Epicier. Vacancies and metalloid atoms form a pseudo-solid solution with order-disorder phenomena analogous to those encountered in metallic solid solutions A 1. antiphase domains.c. owing to the large concentrations of carbon vacancies. Detailed X-ray scattering (Dunand et al..c.% V displaced (Rullier-Albenque and Senateur.6 to 7 eV) are related to the strong bonding in these compounds (the melting point of TaC is 4000 0C). as shown by the increase of the static Debye-Waller factor. traps and recombination centers.

1992). cubic semiconductor such as GaAs. VAs does not trap (or only very weakly traps) positrons. computer simulation supports defect structures not found in metals: the split Nb vacancy in Nb3Sn. 1988). (iii) The effect on plastic properties is very important: point defects may strengthen or weaken the compound. . 10. The charge state of the As vacancy shows two transitions controlled by the position of the Fermi level: V A . In fact. the strength is greatest at stoichiometry. at high temperature.1: generally.c. is very fragmentary or even completely lacking in many compound families (e. induced optically at low temperature. At low temperature (compared to the melting point). On the contrary. Increases of hardness and yield strength were also observed after vacancy quenching (as predicted by Rudman. impurities. The major defect. but complexes involving vacancies. or the seven-Cu-atom crowdion for the interstitial in CuTi. and As-rich clusters (up to 100 A diameter) have been detected in GaAs.c. This has been studied experimentally in detail in the case of transition-metal carbides (Toth. (i) The effect on atomic diffusion has been discussed in Section 8. the antisites GaAs and AsGa. or AsGa +As interstitial. each with a specific relaxation of its atomic surroundings. because the positron trapping depends on the state of charge of the vacancy. A detailed review of native defects in GaAs has been published recently (Bourgoin et al. In a case such as Cu3Au. which can change with temperature and doping (Corbel et ai. (ii) Point defects generally soften the elastic constants. and migration properties.1 and Figure 8). because mobile point defects associated with offstoichiometry favor dislocation climb. 11.-V As and V AS -V As . 1971). and electron paramagnetic resonance. which recently received considerable interest and many interpretations (simple AsGa antisite. Our knowledge of the type of defects. formation.1). we have mostly emphasized the structure and energetics of point defects. The vacancy-type defect concentration is typically 10 to 100 ppm in bulk materials. but presents an instability VGa-*VAs + AsGa in semi-insulating materials. In some alloys with low crystal symmetry. Some point defects have been clearly identified and characterized. antisites. In the present section. There is a common agreement that the Ga vacancy is stable when the Fermi level is high in the gap. Williams. Influence of Point Defects on Physical Properties Up to now. (iv) Finally.deep-level transient spectroscopy. This is partly due to the lack of single crystals of good quality and sufficient purity. point defects increase the electrical resistivity (as discussed in Section 4.) is the socalled EL2 defect.g. the hardness of NiAl and CoAl is a minimum at the stoichiometric composition (Westbrook. 1956). interstitial and Frenkel pair. with only occasional reference to their effects on properties. only a few compounds of simple crystal structure (mainly B2. in particular because each type of defect can have several charge states.or Ll2-ordered alloys) have been studied experimentally. the geometry of the Cu sublattice allows rather simple structures and migration mechanisms of vacancies and interstitials. Strong interactions of native defects with dislocations. structural vacancies and antistructure defects induce point-defect hardening via strain-field effects or elastic-modulus effects (Fleischer. and perhaps interstitials. . . In a simple. and decrease the thermal conductivity and the superconducting critical temperature (Section 9. 1962) or irradiation. in particular in electron-irradiated single crystals: the As vacancy. optical measurements. similar to those in f . one finds already a much more complicated situation. and small defect clusters (such as the divacancy VGa + VAs). 1971. The inhomogeneous distribution of native defects in the material has a great influence on its electrical properties. Conclusion Point defects in intermetallic compounds present a great variety of behavior. This defect shows an interesting metastable behavior.. the diffusion coefficient is a minimum at the stoichiometric composition.). . we briefly summarize the latter. Positron annihilation is of much more complex interpretation than in metals. and of their structure. Native defects in GaAs depend strongly on the cooling rate after processing. which is a donor largely responsible for the semi-insulating behavior of undoped GaAs. . 1987): for example. They are not simple intrinsic defects (except the antisites GaAs and AsGa which accommodate non-stoichiometry). Laves phases) (see discussion by Wever. the origin of which has not yet achieved a consensus—see discussion by Dabrowski and Scheffler (1992). . 1988). . metals. depending on temperature.

Acta Metall. de Novion.. Acta Metall. H. p. B145. Bakker. K. (1972). Zurich.. and Gasparrini. Bakker. 1045. p. and Van Ommen. and Siegel.The possible point-defect structures. American Ceramic Society. ScL Forum.. 657. 8. For example. M. /. Y. Brebec. Haasen). Legnini. Caro. Berner. Davis. (1983). Stucky. H. Susman. Phys.. and Riviere. (1987). 847. Geibel. N. 221. 15-18. P... Solids. A. Phys. Acta Metall. Bakker. Bradley. C . Metals. (1983). Mater. J. D. /.. and Izumi. Phys. (1975). Vacancies and Interstitials in Metals. W. J.. 15-18.. 21. Cahn and P. are so numerous in inter metallic compounds that experiment alone cannot solve the problem. Prog. Acta Metall. 64. J. (1984).. and Fayard. Rad. Acta Metall. 3rd Edn. P. A. (1966). F3. (1976). 25. 1. OH. 23. (1991). A. M. D. (1981). Doyama. Takamura. Kluwer Academic. Unfortunately the corresponding experimental data are most often lacking. It is only recently.. C. P. The point-defect properties calculated from the electronic structure will have to be integrated in a proper thermodynamic theory. and Wachtel. R. (1990). A32. . and Moser. H. Appl. Inst. T. W. (1988). Beuneu. StatusSolidi. T. O. 471. C . C. 10211. (1968). Von Bardeleben.. Beaufort.... Graneli. M. F14. P. Miinster. Nitrides andBorides (NATO ASI Series E. M. /. (1987). L. Status Solidi. Ball. and Scheffler. 64. H. Chang... (1987). S. ScL Forum. Acknowledgements The author wishes to thank Dr O. 793.. Gasparrini. (1987). and Limoge. Y. J.. Dabrowski.. Vol. (1981). (1984). and Zuurendonk.. and Neumann. Dimitrov. References Abromeit. P. (1988).. and consequences on the mechanical properties. 1047. Freer). L. E. Benci. Dimitrov for a critical reading of the manuscript.. (1990). P.. 42. de Novion. G. Saarinen. Priem. Dimitrov. D. E. North-Holland.. O.. and Slifkin. ScL Forum. P. N. Columbus. J. Chem. L. and Wollenberger. L. In Advances in Ceramics. and Seager. In Physical Metallurgy. R(1978). Mater. American Ceramic Society. 319. C. J. Appl. Bocquet. J. 20. and Taylor.. Phys. A. Nucl.. Kiritani). J.. 747. B. 16. /. Dederichs. H. University of Tokyo Press. and Brown. C . and Ehrhart. In The Physics and Chemistry of Carbides. (1969).. V.. and Rozenfeld. 8192. Cheng. Benkaddour. Mater. S. Aoki. and Landesman. 81. and Skold. (1988).. and Averback. 880. S. Appl. and Pedraza.. Physique Coll. migration properties. and Finel. ScL Forum. Scanning Microsc. D. theory here is still in its infancy. /. M. Phil. (1967). Phys. Dimitrov. Soc. Dedek.. Mater. Columbus. Guillot. G. 4. 143. 56. Solids. 11. (1937). A159. 1213. and Dorn. Mag. Richel. Phys. Wynblatt. Zhang. Dordrecht. (1972). 545. Effects. Phys. J. St... Soc. M. 14. C .. C5. Mater. K. Smedskjaer. 20. Rev. J... Caro.. 639. Some good papers have been published recently. and Moser. point defect-dislocation interaction.. O.. (1987).. M.. B59/60.. Bocquet. S. G. P. J. Crawford. Phys. 221. Phys. 45.. 1409. (1982). L. R. (1963). H. Brun. J. (1939). H. 12.. V. Mater. Plenum. that more sophisticated theoretical methods have been developed and seem to be able to predict point-defect structures and properties with some accuracy. G. 34. H. 96. Vol. 31. Symp. R.. E. R. 185) (ed. Chem. 115. Proc. P. M. Alamo. H. 1165. Solid State Chem. 721. Sci. Status Solidi. A. C . C . Bourgoin.. (1975). since the 1980s. Jr. (1986). p. 26. 15. Phys. the commonly used Miedema and bond-breaking semiempirical models to estimate pointdefect formation energies are quite contradictory. 263. Great progress can be expected from the combined use of Monte-Carlo and moleculardynamics simulations (Rey-Losada et al. O. J. B. Lorenzelli. 88. (1973). 5. Stolwijk. Westfalishe Wilhelms Universitat. Bakker. Amsterdam. Corbel. P. Cupschalk. T. Res.. 15. J. Benci. 69. ..... Dimitrov. Such knowledge will also allow study in important fields that are practically unexplored up to now in intermetallic compounds: point defect-impurity interaction. 1993). A.. Germany. R. Hautojarvi. Vol. Considerable work is still required. 1035. Unfortunately.. Beretz. B38. Res. A. Sitaud. Bradley. Res. J. P. O. and Westerveld. .. and Dimitrov. B. R.. (1986). Victoria.. 83-87. X. (1992). (1987). N. Dejus. Arnhold. G. Y. 545. A.. A..10199.. and Desarmot. (1992). (eds R. Bakker.. J. F. R. and Dworschak. C. (1986). Condat. J. C . F. R65. H. Solid State Commun. Rad. 233. Mater. G. Cooper. etc. J. Corbel. 36.. E. Non-stoichiometric Compounds. Bender. and Amelinckx. 735. (1981). /. .. Point Defects in Solids. OH.. p. (1972). and Halbwachs. 1155. (1987). J. 499. In Advances in Ceramics... D. Das Gupta. Chabik. A. In Point Defects and Their Interaction in Metals (eds J. A. J. A. G. C. 409.: Condens. A.. H. C .. (1967). K.. 23. 529. de Novion. NonStoichiometric Compounds... F. H.. Tokyo. M. and Germagnoli. Phys. Nucl. V. Effects. Metall. Methods Phys. 571. S. 15-18. New York. and M. S. and Stievenard.. and Smallman. A13. Nuovo Cimento. Instrum. M.. Hillairet.. 1275. G. 385. Delavignette. H. F.. Proc. Gerold. B. 805. O. P. (1982). Dimitrov. C. Van Der Meij.. J. Acta Metall. 100. P. Thesis. 329. Corbel. A. p. Trans Tech Publications.

. C. V. 36. G.. (1967). Phys. Mater. Doyama. 21. (1981). (1987). 7983. A79. Jennane. Doyama. T... C . Jain. In Positron Annihilation (eds P. J. Nitrides and Borides (NATO ASI Series E.. Trans. and Damask. E. M. A. Status Solidi. Rev. 455. (1952).. A. 1329. W. Landesman.: Condens. 85. H. and Engell. Rad. and Protasov... Hiinecke. Epperson. 1905. Huguenin. D. Hagel. Lipson. Kogachi. Kim.. London. Kuper. J. Koch. D. 609. Nucl. Manning. A. M. T. (1982). J.. J. L. W. M. Mag. Minamigawa. 65. S. S. (1971). Lett. Miedema. Goshchitskii. Baskes. C. I. (1983). Mag. N. Sci. B. Scripta Metall.. 38. Mater. /. and Koenig. (1987). Rev. L. Herman. (1966). V.B2&. Foiles. N. 11. H. Y. and Jacobi. T. Goshchitskii. (1978). and Dautreppe. and Pelissier. 20. V. P. M. Hahn. Gu. (1976). M. (1939). Dunand. Physique. A. C. Turchi. Mater. S. J. 70. J. Kostorz. G. H.. 15-18. Z. Mag. C. 177. K. E. R. 99. Tokyo. J. 2356. L. (1990). 1299. 25. S. Lazarus. and Daw. Moisy-Maurice. (1992). In Positron Annihilation (eds P. Rothman. Edelin. R... W. G. J. Phil. (1987).. G... J. H. 1305. Kim.... B33. Metallk. M. M. 67.. S. (1987). 109. H. Singru.. A. Forum. 905. M. J. Moseev. Physique. 35. Muto. Acta Metall. J. 11. S. (1987). B. Acta Metall. Elcock.. Gerstenberg. R. A.. Sci. Phys. (1975). and Tomizuka. 19.. /.. and Ducastelle. Herman. and Yvon.. (1986). and Loomis.. and Komarek. M. 1055. Phys.. 4. (1986). and Reynolds. C . D. 873.. R. Res. H. J. (1971). 557. Betterton. Rev. Metals... Acta Metall. p.. Bernardini. V. Taub.. Phys. Institute of Metals. and Corbel.. and M. D. B29. 1271. D. and Brittain.. R. D. W. R. W. 359. and K. Forum. 104. and Averback. (1970). P. Schultz. 476. 153. 8. Acta Metall. K. C. Housseau. M. Rev. C... P. K. 71. (1986). J. L. B33.. Diffusion in Body Centered Cubic Metals.. Rev.. A79. Shimotomai... M. Forum.. Status Solidi. 49. and Wever. K.. p. E. Singru. 1737. W. N.. C.. Kiritani). Rev. S. (1988). Phil. (1978). /. 215. G. W. A18. Forum. 739. Pauleve. Foiles. P. O. Nucl. (1992). Hume-Rothery. P. (1984). Acta Metall. Phys.. B33. and Wever. Fishman. and Moine.. Mater. Mater.. Scripta Metall. T. 2. (1958). Rev. Status Solidi. I. and Kitajima. 232. and Friedel. S. P. Hoshino. M.. J. Chudinov. (1993). J. H. (1964). Z. K. V. T. 969. 45. Sci. C .. H.. 46. (1956). D. 15-18. Neel. Epicier.. (1983).. Metallic. World Scientific. H. Proc. (1961). N. 1311.. 40. Eibner.. Gupta. (1992). Phys. and Herzig.. M. 57. A. J. and Dodd. N. R. Lett. Masuda-Jindo. D. 921.. 303. and Lieberman. Tokyo. Koch.. Koch. Sci. Ho. Phil. R. 27. C. 5319. R. D. Epperson. (1980). F. Fromberg. and Ortiz. Gilbert.. J. (1964). R. Mater. V. S.. Sci. (1985). Marshall. A94. Berner. Manning. J. J. P. Fleischer. J. H. I. Rev. Scripta Metall. V.. Lomer. J.. B... Inst. B31. Fleischer. 1257. Kinoshita. 80. 1509. 1703. (1981). Rev. H. /. Koenig.. 387. Acta Metall... J. Mag... Phil. and Koenig. 153. Mater. . 24. Kluwer Academic. M.. L. Phys. G.. T. I. (1987). H. J. C.. 74. 31. G. Stefanou. Physique.. and Hillairet. S. B4. Kim. Phys. Phys. 19. M. Phys. J. Sci. Metals Park. Phys. p. Treglia.. Kaufmann. Phil. J. Dordrecht. Phys. Hatem. Frohberg.. Leubolt. Appl. M.. A. University of Tokyo Press. OH. M. and Chang. 1536. N. p. M. and Taylor.. and Iwata. C. W. M. 147. Alamo.Donaldson. (1988). Moseev. Phil. Girifalco. S. and de Novion. Phys. Z. Doyama. 323.. 9. Solids... V. and Nerses. 197. Effects. 1001. L. (1986). Forum. Matter. 863. Singapore.. and Rawlings. 284. (1987). A38. S. Jacobi. (1975). Mater. A. M.. Phys. J. R. Rev. 605.. 857. and Fritsche. Chem.. Phys.. Gopinathan). R. p.. Appl.. and Damask.. 749.. 15-18. B. Moser. In Vacancies and Other Point Defects in Metals and Alloys (Institute of Metals Monograph and Report Series. 1111. (1985b). Chudinov. H. Laugier. S. 701. Rev. (1985a). 15-18. /. A. (1985). Doyama. 185) (ed. A6. Kim.. Phys. 5. Acta Metall. K. Koch. 903. (1983). and Westbrook. N. Engell. Koczak. 345. J. and Protasov. J. H. In Point Defects and Their Interactions in Metals (eds J.. Mater. 1083. K. 383. M. Stefanou. 37. Huntington. Rev. B2.. L. Huang. J. 119.. Phil.. Gupta.. Soc. Gopinathan). Takamura. 30. C . J.. 317. D. Junqua. (1958).. Flack. G. and Koenig.. N.. (1986). Status Solidi. Inst. Papers. B. P.. Isper. and Nakahigashi. G. H. 811. (1983). Lazarus. M. and Koch. 54. N. M. Singapore.. 6216. T. C. S.. Mukai. S.. Mag. M. and McCombie.. and Buyers. J. 12. M. Pauthenet. A44. (1984). 23). American Society for Metals. Mater.. O.. I. (1973).. and Lieberman. Neumann. 1113. J.. M. (1936). Jain. (1987). 11. (1977). p. O. Wang. C . K. In The Physics and Chemistry of Carbides.. 15-18.. (1971). Chem. S. Moser. C .. and K. and Daw. A. J. G. 108-109. Mag. Vol. 49. Lesueur. C. Labbe.. and Moya. Phys.. 55. and Meyer. J. Metallic. M. (1988). 2299. J. Sci.. World Scientific.. 906. D. G. Res. Metall. A. (1979).. Kim. 559. A. A173. (1979). Hilgedieck. Res. Acta Metall. (1973). Phys. R. Miller. J. D. (1986). C. 529. 1451. J. Phys. E. Freer).. R.. Mag. C.. Desoyer. Phys. A.. R. and Koenig. no. (1982).. 32. R.. (1965).

. Rad. Status Solidi. Phys. P. and Pontikis... A. Intermetall. Crystallogr.. S.. J.... O.. Wang.. (1991). In Diffusion in Solids: Unsolved Problems (ed... R. (1979). 1140. 45. Y. (1981). Symp. P. 29. p. Dinhut. S. Chem. H. 71.. S. M. Williams. O. Wachtel. A42. 69. edited by J.. (1964). Lutton.. F14. 6. B35. R. Res.. L. D. (1957). Z. 433. Gordon and Breach. 449 Shirai. J..... (1968). and Eckert. R. Proc. Y. J. J. L5. 17. P. A.. Mater. Rev. 103. W. H.. Rev. J. Van Ommen. A25. Zak. F. Schwab. M. A. Rullier-Albenque. 5. and Bakker. Amsterdam. W.. A. A48. Soc. Pedraza. Intl. Wasilewski.. Rad. M. Transition Metal Carbides and Nitrides. Rad. Wever. 11869. (1983a). O. A. 81A. Tohoku University. Wesley. H. 1225. J. Solids. M. L. (1973). W. 6. 69.. (1976).. (1960). (1984). ScL Forum. T. Wharton. 473. Mater. and Gerold. and Doyama. Sitaud. Acta MetalL. Simmons. 54. M. J.. 150. 239. C. F. 1165. Phys. M. Schaefer. A152. (1983).. and Senateur. Rullier-Albenque. R. Y. 117.. L. H. E. 83. J. (1986). A. 2705. P. Westbrook. 10. 88. and Hanlon. New York. Seybolt. Zurich. and Lesbats. (1981). Acta MetalL. 37. E. Symp.Neumann. Phil. (1974). Acta MetalL. P.. Butler. Cl. B41. (1983). J. (1980). J. and Yamaguchi. R. J.. 549. U. 28. Wurschum. Izumi). Rudman. Mater. p. J. Haasen).. Solids. Prog.. /. M. (1962). . Acta MetalL. M. G. M. and Hatcher. (1992). 12. M. E. Van Feggelen. Japan. Moleman. 201. M.. 38. 123. P. H.. (1977). M. J.. 173. Phys. 5. Mag. and Bakker. p. Appl. Wechsler. and Desarmot. Dehrli. R. Phys. Phys. Acta MetalL. (1987). J. Solid State Chem... Urban. 38. Waegemaekers. R. 55. E. M. S. Mag. Mater. Phys. and Farkas. Fourdeux. Moser. Meurtin. H. Rey-Losada. J. J. M. T. Solids. (1992). H. 47. 29. 337.. Effects. 118. H. Soc. Mag. Dai.. B. A.. Cahn and P. C7. North-Holland. Academic Press. 39. Compounds. (1990).. A... N. P. J. Riviere.. Rev. 34.. W. 1251. F. 593. and Westbrook. 195... H. H. 429. Phys. H. Effects. /. D.. J. R. N. T. Herte. and Bakker. Schlatter. H. 291. Appl. Van Ommen. (1983b). Physique Coll. G. Westbrook and R. 57. and Doyama. A. Shimotomai. Sendai. and Ehrhart. Linse. A43. A. F. F. 387. (1981). C . D. Reckman.. and Gerold. 223. (1987). 330. Phil. J. O. New York. J. Effects Defects Solids. P. J. Taylor. F. Stolwijk. D. A... Stolwijk. Caro. 39. M. Phys. Acta MetalL. ScL Eng. M.. T. Res. Chang.. C .. 28. and Lee. Proc. J. Dinhut. and Dimitrov. (1991). C . (1977). J. (1968). 232. J. 3rd Edn (eds R. Status Solidi. J. 1461. R.. Hayoun. Chem. H. Keller. Nucl. 24... H. E. M. P. Shimotomai. A. Electrochem. Lett.. 1: Principles. B105. (1993). V. and Balluffi.. 137.. (1971). E. Physique Coll. (1964). /. Weber. P. Vol. and Dural. H. This chapter was originally published in 1995 as Chapter 23 in Intermetallic Compounds. R.. N. and Grilhe. G. V. Wollenberger.. N. M. A. Effects. S. N. (1980). Neumann. Acta MetalL. V. D. R. Murch). J. D. Riviere. Shoemaker. Yu.. 15-18. Parthe. Dienes.. Wasilewski. M. Chem. In Physical Metallurgy. R. Wang. Stefanou. Paris.. J.. G.. W. R. Trans Tech Publications. (1980).. (1981). Mat. W. (1971). (1972) J. Phys. Fleischer. 4234. (1956).1. O. J. Schulson. Q. Phys. Dimitrov. and Doyle. Structure and Mechanical Properties (JIMIS-6) (ed. A. 783. E. Phil... Rad. and Lam.. Welch. Toth. Sob. D... and de Miranda. W. Van Gend. V. and Dederichs. Lazareth. P. Phys. 143. Riviere. Van Ommen. Jr. J. Crystal Chemistry of Tetrahedral Structures. (1984). P. (1991). and Quere. Sabochik.

grain-boundary sliding. there are the crystallographic parameters that describe the orientations of the neighboring grains and the interface. and field ion microscopy (FIM). though because of the energetic requirement these defects are minimal. in the grain-boundary region of intermetallic compounds. extensive progress has been made on experimental observations and theoretical understanding of grain boundaries in intermetallic compounds. we review direct observations using transmission electron microscopy (TEM). Johnson and Blakely. In perfect lattices of intermetallic compounds. For example. direct Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. the concentration of component atoms at a grain boundary can vary depending on the stoichiometric range in the bulk. Furthermore. electronic conductivity. atoms of one component prefer to bind with atoms of the other component atom as their nearest neighbors owing to their chemical bonding nature. Thus. There are some indications that grain boundaries in intermetallic compounds behave uniquely and thus present some very interesting properties. L Fleischer ©1995. recrystallization. In this chapter. H. and corrosion resistance (Chadwick and Smith. However. Johnson and Blakely. i. concerning the chemistry on grain boundaries of intermetallic compounds. Sendai 980. Therefore. wrong bonds between like atoms may be introduced. grain boundaries in intermetallic compounds may involve the influence of bond defects in addition to the lattice distortion. 1977). normal bonds between unlike atoms may not be conserved or may be distorted. 1977). During the last decade. However. Tohoku University. Aoba-ku. This electronic structure and the nature of the chemical bond at grain boundaries can influence the cohesion of a grain-boundary plane. grain boundaries in intermetallic compounds play important roles in various properties of polycrystalline materials. and the electronic conductivity of intermetallic compounds. such as mechanical properties. and the distributions of the primary compositional atoms and the impurity atoms (Chadwick and Smith. 1976. and can deviate from its stoichiometry in the matrix. many problems still remain. which is closely related to the mechanical properties of intermetallic compounds. Katahira 2-1-1. is particularly important. also. 1976.e. These compositional perturbations at grain boundaries can affect many material properties. which essentially consist of multicomponents in contrast to pure metals. and these techniques and theories have also been applied to grain boundaries of pure metals and disordered alloys. Next. Introduction Just as with grain boundaries in pure metals and disordered alloys. There are numerous variables that can affect the structure and properties of a grain boundary. such as grain-boundary fracture. mostly ordered alloys. and also the component atoms can interact with the impurity atoms. 2000 John Wiley & Sons Ltd . which are technologically and scientifically important. In grain boundaries of intermetallic compounds. It has been recognized in pure metals and alloys that many properties of a grain boundary depend strongly upon the chemical composition at grainboundary planes. Edited by J. understanding of the chemical composition at grain boundaries of intermetallic compounds. grain-boundary corrosion.Chapter 7 Structure of Grain Boundaries Takayuki Takasugi Institute for Materials Research. Japan 1. high-resolution transmission electron microscopy (HRTEM). the atomic structure and chemical composition of grain boundaries in intermetallic compounds result in a unique charge distribution of the electrons. Westbrook and R. the nature of the atomic accommodation.

Sasaki et al. 1989. FIM observation is free from this geometrical restriction and offers spatial information on the atoms at or near a grain boundary.. Many observations by HRTEM. and then the possible atomic configurations can be constructed. when the grain-boundary plane is parallel to the electron beam. 1991a. In contrast. The atomic positions consisting of all the Ni atoms are imaged in this figure. Some unique characteristics of grainboundary structures in intermetallic compounds are mentioned.e. 1992. In other words. tilt and mixed boundaries. It was found that small-angle twist. 1976. 1992. wrong bond pairs across the boundary (see Figure l(b)). These are called coincidence-site lattices. it involves to some extent regions adjoining the grain-boundary plane because it has been shown that these regions also affect the properties of intermetallic compounds. geometrical and atomistic computer models for the structure of grain boundaries of intermetallic compounds are presented.. and were apparently replaced with vacancies.b. and FIM combined with atom-probe (FIM-AP) techniques are presented. it is also noted that distinct columnar cavities occur in grain boundaries owing to the depleted bond pairs. i. Extensive HRTEM observations have been performed on grain boundaries of Ni3Al. Westbrook. 2. When two cubic lattices with the same lattice parameter. Mills. and large-angle CSL boundaries were primarily ordered up to very close to the grain-boundary plane. the terminology * grain boundaries' is used in a wide sense.e.. 1991a. b. 2. TEM combined with energy-dispersive X-ray (TEMEDX).. 1977). the reader is referred to textbooks (Chadwick and Smith. and these can then be discussed in relation to the grain-boundary structure. 1991a. i.observations using Auger electron spectroscopy (AES). 67. and FIM have been reported. Finally. Direct Observations of Structure HRTEM and FIM can offer the most direct information for the structure of grain boundaries. they can introduce a three-dimensional superlattice of common lattice points. 1991) or solidified polycrystals (Mackenzie and Sass. such as low.1 HRTEM and TEM Observations Concerning the atomic configuration at a grain boundary. and the fraction of lattice points in common is usually written as 1/E.e. HRTEM observation is generally applicable only when lattices in either or both of the grains adjoining the grainboundary plane are parallel to the electron beam.Ti) compound. Johnson and Blakely. special CSL and random (general) boundaries have been observed.e. the expected two Al(Ti) atoms. In the last part of this chapter. i. Johnson and Blakely. The major interest in these studies has been directed to determination of the existence of structural disordering (or ordering) at or near the grain-boundary plane. were not detected in this image.. which have one lattice point in common. This observation reveals that the grain-boundary structure attained a symmetrical configuration without a substantial rigid translation along x and y directions (i. The lattice points (or fringes) at or near a grain boundary can be imaged. which is supposed to have a high ordering energy.and high-angle tilt and twist boundaries. However. because this feature was postulated to relate to the mechanical response of the grain boundary in this type of alloy. Figures l(a) and (b) show an atomic image and the corresponding atom distribution observed in the £ = 13. where E is an odd integer (Chadwick and Smith. Mills et al. In this chapter. with and without boron. Various types of grain boundaries. .. Williams etal. 1992. 1990). These textbooks contain numerous references not only on the terminology of grain boundaries but also on the terminology of the microscopy technique described in this chapter. by which contrast images of features such as geometrical defects or grain-boundary dislocations are observed. wherever the terminologies commonly used in the field of grain boundaries and their physical meanings are not described in sufficient detail. which were prepared as oriented bicrystals (Kung et al. Consequently. 1992). are related by any of a series of special rotations through the common point. 1989. 1988. Krzanowski. 1975. conventional TEM. 1976. the objective in these observations was to understand the intrinsic brittleness of Ni3Al in its polycrystalline form and the beneficial effect of boron addition on grain-boundary brittleness. Baker etal. (1980) showed an HRTEM image of a grain boundary prepared from a bicrystal of the Ll2-type Ni3(Al. Also. In this figure. along the interface) and also that an almost ideal Ll 2 order is preserved up to the grain-boundary plane. Less-direct observations are done using conventional TEM. it is suggested that chemical ordering is primarily sustained without locally accommodating the atoms with a possible chemical disordering. regardless of the addition of boron (Kung et al. leading to the energy minimum.38° coincidencesite lattice (CSL) boundary with [10 0] rotation axis. Kung et al. mostly on alloys with Ll2-type structure. the nature of the chemical bonds and the associated electronic structure are discussed in relation to the atomistic structure and the composition of grain boundaries. 1977.

Note that large and small circles represent atoms distributed on the first (10 0) atomic plane and the second (10 0) atomic plane. superlattice fringes. 1988. However. was impossible from these observations. There are no HRTEM observations of grain boundaries for other intermetallic alloys. Tichelaar and Schapink (1986) demonstrated that these structures can be generated by . the structure and the associated composition at grain boundaries of Ni3Al depend on the material processing. i. (a) Atomic lattice image and (b) the corresponding atom distribution observed in the £ = 13. Here. 1989). and this phase was found to cover all the boundary regions in boron-doped Ni3Al that was prepared by powder processing (Baker et al.Ti) bicrystal. for a weakly ordered Ll 2 alloys. stoichiometry.e. it was found that a disordered phase up to 20 nm wide. Thus. For example. for example. Ti atom Figure 1. how much translation^ and local lattice accommodation takes place at the interface.1417. 67. is present up to close to the interface plane in both cases. Therefore. which is a discrete phase with a distinct interface with the matrix. were demonstrated for coherent (Tichelaar and Schapink.. respectively (Reprinted with permission from Sasaki et at. corresponding to the (100) and (0 1 0) planes.38° CSL tilt boundary with [100] rotation axis of Ni3(Al. 1992). the lattice images at the interface itself were not clear. Also. Kung et al. and also the orientation parameters. This is mostly due to the grain boundary being inclined to the beam direction. exists along portions of grain boundaries in boron-free Ni3Al. Mills etal.. To get a generalized conclusion for grain boundaries in Ll 2 structure.5 nm thick.36 nm lattice fringes. more studies are needed. a symmetric configuration is defined as the structure where each sublattice occupied by Cu and Au atoms in one component crystal is fully symmetrical to the corresponding sublattice occupied by Cu and Au atoms in the other component crystal. There are other examples: a compositionally disordered region —1. Copyright (1990) Pergamon Press Ltd) Mills. Figure 2 shows HRTEM images of two largeangle twist boundaries (with L = 5 orientation) in Ni3Al bicrystals with boron and without boron (Kung et al. In these observations. The atomic configuration at a grain boundary and the existence of its multiplicity. e. whether the lattice is ordered just at the interface is inconclusive. arc-melted and wellrecrystallized Ni3Al materials did not have such a thick disordered phase and regardless showed very high ductility. 1989. 1991. 1986) and incoherent (Tichelaar and Schapink.38. variants. i.. A square array of 0. Also. alloying of boron. discussion of the detailed atomic configuration at the interface.9 1992). was found to be present in the vicinity of a large-angle general boundary in borondoped Ni3Al (Mackenzie and Sass.Ni atom Al. HRTEM observations for grain boundaries in other kinds of crystal structure are also needed. 1988) E = 3 twin boundaries of Cu3Au-ordered alloy using the conventional TEM method of a fringes. an asymmetric configuration is defined as a structure where one sublattice occupied by Cu atoms in one component crystal is symmetrical to a sublattice occupied by Au atoms. ActaMetall. On the other hand.9 1990). which was not an imaging artifact. Mater. Krzanowski. Figure 3 shows that two structures consisting of symmetric (denoted S in Figure 3(b)) and asymmetric (denoted A in Figure 3(b)) configurations were separated by an antiphase boundary (APB) on one side of the boundary plane.g.e..

Grain boundary Grain boundary Figure 2. Also.. Mater. . 1992.y 1989. Figure 4 represents the diffraction image of a [0 0 1 ] small-angle boundary in Ni3Al with a twist component of 1. HRTEM images of two large-angle twist boundaries (with L = 5 orientation) in Ni3Al bicrystals (a) with boron and (b) without boron (Reprinted with permission from Kung et al. 40. 81. it was found that boron additions to Ni3Al reduced the APB energy at the interface (Kung and Sass. The latter translation is generated by a dislocation in the boundary. (1972). Copyright (1992) Pergamon Press Ltd) rigid translations of the type \{l 1 0> or the type \(2 11> parallel to the boundary plane.4°.0° and a tilt component of 0. Buis et al. 1992).. Ada Metall. It was shown that small-angle [ 0 0 1 ] twist boundaries and small-angle [ O i l ] title boundaries in Ll 2 -type Ni3Al contain dislocations with Burgers vectors corresponding to APB-coupled \[ 1 1 0] superpartials. 1991). The dislocation structures of grain boundaries in intermetallic compounds have also been observed using the conventional TEM technique (Kung and Sass. The structures of these boundaries were thus shown to be in agreement with theoretical models proposed by Marcinkowski (1968) and also with the CSL theory summarized by Balluffi et al.

. S and A indicate symmetrical or asymmetrical structures. (b) Grain-boundary structures corresponding to TEM images.1991). The spacing of the superdislocations distributed at the interface was larger than the value calculated from anisotropic elasticity theory. A represents an A atom at level fl/2j2 below the paper and B atom at height a/2$ above the paper. accompanied by ordinary dislocations (i. Numbers are assigned to different domains.e. This discrepancy was attributed to changes in the APB energy due to . The m and t correspond to 'matrix' and 'twin^ orientations. 3 represents a B atom at height zero and A atom at height a/2Jl above the plane of the paper. It was shown that a symmetrical low-angle tilt boundary consisted of a regular array of <1 1 lHype superdislocations in large areas. C represents an A atom at height zero and B atom at height a/'$ above the plane of the paper. Projections (ii) and (iii) contain a dislocation (marked D) with b = j [2 1 1 ] and line direction u= [1 1 0].Twin boundary Twin boundary Figure 3. 1989. respectively. 1986). y<l 1 1» in small areas. while the center (ii) and right (iii) are projections along [0 1 I ] . (a) TEM bright-field electron micrographs showing two structures consisting of a symmetric and an asymmetric configuration separated by an antiphase boundary (APB) on one side of the (1 1 1) boundary plane (indicated by arrows) in ordered Cu3Au (Reproduced by permission of Taylor & Francis from Tichelaar and Schapink. The left (i) is a projection of a (1 1 1) twin boundary in an Ll 2 A3B ordered alloy along [110]. O and A represents A atoms at heights zero and a/2Jl above the plane of paper respectively. i \ represents a B atom at level a/2$ below the paper and A atom at height cr/2j2 above the plane of the paper The dislocation structures of grain boundaries in B2 (cP2)-type ordered Fe3Al alloys have been observed (Buis et al.

was shown to consist of a hexagonal dislocation network accommodating the deviation from the CSL misorientation. Various types of grainboundary structures and the associated compositions were observed. Disordered phases were occasionally detected in boundaries of these alloys. Also. 1987a. Copyright (1992) Pergamon Press Ltd) segregation of solute atoms to the APBs associated with the superdislocations. 1986.g. Brenner and Hua. (a) Diffraction (g = 0 2 0) image of [0 0 1 ] small-angle boundary in Ni3Al with 0 (twist) = 1. 1992). 40.. Acta Metall. depending on material processing and thermomechanical treatment. However. Mater. the asymmetrical E = 3 boundary. 1987a. 1990. b.4°. differences in the transition temperature for the two types of twin boundaries were not explained. 1986. c. e. However. (b) Schematic diagram showing the structure modeled using APB-coupled superpartial dislocations (Reprinted with permission from Kung and Sass.2 FIM Observations FIM observations have been performed on grain boundaries of Ni3Al polycrystals by two laboratories: Oak Ridge National Laboratory (ORNL) (Miller and Horton. 1986a.0° and 0 (tilt) = 0. Tichelaare/tf/. Horton and Miller. The temperature difference between the transition temperature of both twin boundaries and the bulk was demonstrated to range from 0. 1967) but the lattice vectors. Sieloff et al. 99. b. 2.. introduced by powder processing. The kinetics of the order-disorder phase transition at twin boundaries (of E = 3. 1991). 1990a. b) and the University of Pittsburgh (Brenner et al. Some of these dislocations were dissociated into ordinary dislocations. The major interest in the FIM observations as in the HRTEM observations was the detection of the state of order (or disorder) in the boundary region and the effect of boron on it. It was shown that the order-disorder phase transition occurs in these boundaries prior to the phase transition in the bulk. the Burgers vectors of these dislocations were not the expected smaller DSC (displacement shift complete) lattice vectors (Bollmann.. On the other hand. but most of those phases were probably non-equilibrium phases. b.APB region Figure 4. asymmetrical tilt boundaries consisted of two types of <1 1 O) dislocations. which involves an extra 4° misorientation..5 to 2 K. In the . [1 1 Ij and {211} types) has been observed in the Ll2-type Cu3Au alloy using an in situ TEM technique (Antonopoulous et al.

1987a. Scripta Metall. Brenner et al.24. Takasugi et al. Brenner and Hua. 1988. In this image.. White. 1989). Choudhury et analysis.. giving strong evidence et al. Three direct observation techniques have been 1990a. b. 1986. enrichment or depletion Figure 5. 1989. Note configurations with different atomic composition but with nearly equal energy.1 AES and TEM-EDX observations analysis. 3. concerning the atomic configuration and linear defects. 1986. 1986a. segregations are considered.. 1985. Therefore. Direct Observations of Chemistry The chemical composition of grain boundaries in intermetallic compounds as well as in ordinary metals strongly affects material properties. 1991a. Ogura et al. On the other hand. which showed the successive superlattice planes al. This result indicates the matrix may take place owing to a McLean (1957) ordering of the matrix right up to the boundary (Reprinted form of grain-boundary segregation or variants for with permission from Brenner and Ming-Jian..% boron showing bright and dim contrast features of edges of superlattice planes on both sides. It was demonstrated microscopy combined with energy-dispersive X-ray from these observations that the limiting thickness of techniques (TEM-EDX) (Baker and Schulson. 1985. Horton and evidence of ordering was given in an atom-probe (AP) Liu. George and Liu. . Copyright (1990) Pergamon Press Ltd). such as dislocations at grain boundaries of intermetallic compounds. The stoichiometry at grain boundaries may be different from the stoichiometry of the matrix.35 wt. For both kinds of segregation the near-vicinity of the interface in Ni3Al polycrystals species. More quantitative Taub. At least two kinds of atomic species. 1990). the ordered Ll 2 structure persisted in boundary or in the matrix. have the possibility of a disordered structure. 1991. 1990a.% boron (Brenner and Hua.35 wt. the constituent atoms. It is generally known that It was discovered by Aoki and Izumi (1979) that Ni3Al the resolution obtainable by FIM is insufficient to polycrystals were ductilized by addition of a small .. 1990. 1992. Since intermetallic compounds by definition consist of components of two or more elements. c. 1986.e.. FIM observations were indeed successful in the determination of grainboundary chemistry when they were combined with AP 3. Briant and for ordering right up to the boundary. Mater. the chemical compositions at grain boundaries are not so simple. Suzuki and Masahashi. 1991). b. two types of equilibrium and non-equilibrium without boron and also even in those with boron. alternating bright and dim adopted for the chemistry at grain boundaries of lattice-plane edges were visible at the two (00 2) superintermetallic compounds: Auger electron spectroscopy lattice poles that are in close proximity to the boundary (AES) (White et al. 671. 1990a). and solute or impurity atoms are considered for the chemistry of grain boundaries in intermetallic compounds. Horton and Miller. transmission electron on the two sides of the boundary. 1987a.determine the exact site of atoms because of geometrical and field-effect constraints. Sieloff et al. b). b. an ordered layer that can be resolved is approximately Krzanowski. b. there are no reliable FIM observations at the present time. Liu at opposite sides of the interface. 1988.. 1984. Figure 5 shows an FIM image of a grain boundary of Ni3Al with 0. the that the black circle corresponds to the area where atom-probe (AP) analysis has been done segregation of solute atoms and impurity atoms to the grain boundaries of intermetallic compounds may be influenced by the alloy stoichiometry at the grain equilibrium case. As will be described later. FIM image of a grain boundary of Ni3Al with of the constituent atoms at grain boundaries relative to 0. and field ion microscopy combined 1 nm (Brenner and Hua. i. meaning that only with atom-probe methods (FIM-AP) (Miller and a few atomic layers in the vicinity of a grain boundary Horton. but were not successful in the determination of the local atomic positions. 1985. George et al.

.7 at. Each spectrum was measured from each grain-boundary facet. 25 at.% Al) and (b) Ni3Si (water-quenched. Scripta MetalL. 551. 19. AES spectra from intergranular fracture facets of (a) Ni3Al (water-quenched. Note that signals of carbon and oxygen detected in a few spectra were due to contamination from the vacuum pump system (Reprinted with permission from Takasugi et al. 23. Copyright (1985) Pergamon Press Ltd) .Electron Electron energy IeV) energy IeV) Figure 6.% Si) compounds.

as shown in Figure 6 (Takasugi et al. Briant and Taub. Liu et al. it was shown that the diffusion coefficients at 973 and 773 K were 10"16 to 1 0 1 7 Hi2S"1 and 5x 10~21 Hi2S-1.480. 1990.. which were prepared from high-purity raw materials. of course. 1985). Mater. Takasugi et aL. the binding energy of boron atoms to a grain boundary was estimated as 14-43 k J mol"1. (O) 0. Briant and Taub.144.. the first inquiries using AES were directed to determine whether or not the grain boundaries of Ni3Al are contaminated by a harmful species and thereby embrittled. 1992). • . George et aL.. As a result.e. Suzuki and Masahashi. 1991a.3wt. 1992). 1984. 1985).%): ( • ) 0. The solubility of boron in the matrix (i. Concerning the segregation of unintentionally included impurity atoms to grain boundaries. (O) 0. depending on the bulk concentration of boron. except for carbon. Choudhury et aL. which similarly fails in a predominantly intergranular manner (George and Liu. Ogura et aL. depending on temperature and bulk boron content. The equilibrium type of segregation of boron in Ni3Al was also demonstrated by the experimental results that the boron content at grain boundaries increased with decreasing temperature and with increasing bulk boron content.%) Temperature (K) Figure 7. indicating that the sulfur enrichment is due to equilibrium-type segregation (Takasugi et aL. AES results on fractured grain boundaries of Ni3Al (Takasugi et aL. thus indicating the orientation dependence of the grain-boundary segregation of the impurity atoms. Takasugi et aL. 1992).%) (Sieloff et aL. 1984. During the last decade. Therefore. Copyright (1992) Pergamon Press Ltd) Bulk boron content (at. Ni3Al compounds fail in an extremely brittle manner along their grain boundaries. 1990). for example. an enthalpy of adsorption of boron to a grain boundary of BkJmol" 1 was calculated (Choudhury et aL. 1987a) and the enrichment factor of boron at grain boundaries of Ni3Al ranged between 5 and 50. the kinetics of boron segregation to grain boundaries of Ni3Al was evaluated based on the experimental data. extensive efforts have been made using AES to understand the behavior of the segregation of boron into grain boundaries in Ni3X. 1985). and ( A) 0. White. ( • ) 0. 1985). are likely to be embrittling elements in intermetallic compounds as well as in ordinary metals and alloys. 57. These atoms.% (0. Many AES results showed that the signals of boron were removed by sputter etching of the fracture surfaces. oxygen. The equilibrium grain-boundary boron content vs. Based on these AES measurements. temperature. Without boron. • and A were slowly cooled . Horton and Liu.480. showed no impurity atoms on these grain boundaries.amount of boron. AES results showed that the grainboundary brittleness of Ni3Al cannot be associated with the grain-boundary segregation of impurities. and Ge) and NiAl with and without boron. and carbon into the grain-boundary facets of Ni3X has been reported (White et aL. ActaMetall. Si. as shown in Figure 7 (Choudhury et aL.4at. ( • ) 0. 1986.240. as shown in Figure 8. Also. respectively. Consequently. 1985.048. 40. b. The segregation of sulfur. (Reprinted with permission from Choudhury et al.240. as well as on heattreatment temperature and orientation of the grainboundary facets (Liu et aL. Ogura et aL. 1992). Liu et aL. Liu et aL. No impurities were detected on the fractured grain boundaries of NiAl (George and Liu.. 1988. particularly in Ni3Al. it has to be noted that the elimination or control of these unintentionally included impurity elements is still an important technological method to improve the mechanical properties of intermetallic compounds. and ( A) 0. and the activation energy for diffusion Typical error bar Grain-boundary boron (at. Ga. Hence. 1985. 1985. 1985. revealing equilibrium-type segregation. 1990). a number of AES results have been accumulated on fractured grain boundaries of Ni3X (X = Al. clean grain boundaries were observed on stoichiometric NiAl (with B2 (cP2) structure). The sulfur peak in Auger spectra in the grain boundaries of Ni3Al and Ni3Si disappeared immediately upon the start of sputtering. 1985. 1985. 1985. 1988. Also. 1985) and Ni3Si (Takasugi et aL. 1991.048. 1989. in relation to its ductilization effect (White et aL. Thus. Choudbury et aL. The distribution of the segregated atoms was not uniform from boundary facet to boundary facet. the perfect lattice) of Ni3Al was approximately 1.

being independent of the stoichiometry of the bulk (Sieloff et al. with increasing bulk Ni (decreasing Al) concentration. thus indicating that boron did not attract so much excess Ni to grain boundaries and also did not introduce a disordered phase at grain boundaries. 1985). More work is needed to provide a general conclusion for the . that the grainboundary enrichment factor is proportional to the inverse of the solid solubility of the species in the matrix (Seah and Hondros.e. ppm).1 and 142MmOl-1 for carbon in 7Fe (Ananyn et al. more than 50 (Briant and Taub. the general form of behavior observed in conventional alloys.3 In 2 S. 1985). with and without boron. However.% boron (100 wt. 40.g. more recent observation indicated no clear competition of boron with sulfur for grain-boundary sites (Briant and Taub. For example. B.. 1985). and 7-Fe). it was demonstrated by FIM-AP (Sieloff et al.e.1 and 15OkJmOr1 for carbon in Ni (Massaro and Petersen. 1984. b) and TEM-EDX (Baker and Schulson.. On the other hand. (Reprinted with permission from Choudhury et al.. 1973). sulfur tends to segregate more strongly to free surfaces than to grain boundaries (White et al. Liu et al. the diffusion coefficient at 773 K and the associated activation energy were reported to be 4. 2. On the other hand. 1985). The diffusion coefficient and the associated activation energy for boron in Ni3Al are considerably smaller and larger. The solubilities of boron in Ni3Si and Ni3Ge were shown to be very much lower than that in Ni3Al (Briant and Taub. Thus. This result suggests the co-segregation of B and Ni. Cu.1 and 88 kJ mor 1 for boron in 7-Fe (Busby et al. the alloy stoichiometry may strongly influence the segregation of the boron atoms either through its bulk concentration or through its grainboundary concentration. 1988). 1984. but by about the same amount in the two alloys. 1988. 1989. The intensity of boron segregated to grain boundaries increased. respectively. Copyright (1992) Pergamon Press Ltd).Grain-boundary boron (at. 1987a.7XlO. and 6. Calculated curves were drawn on the assumption of three values of the diffusion coefficient of boron in Ni3Al was between 200 and 300 kJ mol"1 (Choudhury et al. also holds in the Ll2-type intermetallic compounds.. However. the enrichment factor of boron at grain boundaries of these compounds was very large. i. Krzanowski. 57. than those for interstitial solutes in typical face-centered cubic metals (e. 1992). 1984... and may not be generally applicable.. Liu et al. were only slightly Ni-enriched relative to the bulk (George et al.%) Annealing time (s) Figure 8. That is. The most interesting result for boron segregation obtained by AES observations is the correlation of segregation between the constituent atom. as shown in Figure 9 (Liu et al. 1970). 1989) observations that grain boundaries in Ni3Al without boron had almost the bulk composition. 1953). Also. it was recently claimed that their results may have been influenced by rapid solidification or insufficient equilibration. Liu et al.048 at.9XlO.. 1987a. boron has a strong tendency to segregate to grain boundaries but not to free surfaces (White et al.15 In 2 S. Thus.. Recent observations using AES indeed showed that grain boundaries in similarly processed Ni3Al alloys.. 1988). b). Suzuki and Masahashi. and also the amount of grain-boundary Ni increased significantly. i.15 In 2 S. Ada Metall. 1988). site competition between sulfur and boron at grain boundaries of Ni3Al was detected (White et al.3XlO. The variation of boron concentration on the grain boundary with annealing time at 973 K for an Ni3Al alloy with 0.. and the dopant..... 1991). Correspondingly. Mater. Ni (or Al). 1971). Ni.

ppm boron. ActaMetall. The degree of hardening is generally more severe in alloys containing an excess of the more electropositive element. . Interpretation of Various Property Measurements by Grain-Boundary Structure and Chemistry It is considered likely that the grain-boundary structure and chemistry play significant roles in the various properties of many intermetallic compounds. including Ni3Al. thereby resulting in the suppression of the intergranular brittleness (Fleischer et al. FIM-AP observation indicated that boron segregates to grain boundaries both in Ni-rich as well as in Ni-poor Ni3Al. 1993). the relative area of the grain boundary to the matrix in the probe was taken into the calculation and then a grain-boundary enrichment factor for boron ranging from 2000 to 3000 was obtained (Takasugi et al. grain-boundary hardening of up to 30% relative to the matrix (the grain interior) has been detected by means of microhardness measurements and has been summarized by Westbrook (1965). 1990).05 wt. 1967).Al concentration (at. microhardness at or near grain boundaries. the solid solubility of boron in which has been shown to be less than 50 wt. Peak-height ratio (B/Ni) Peak-height ratio (Al/Ni) correlation of the segregation of boron with the alloy stoichiometry at the grain-boundary plane in these intermetallic compounds. They will be briefly discussed in the following sections and also in detail in Chapter 38 by Briant in this volume. Also. there was no attendant improvement in tensile ductility because boron is an extremely potent solid-solution strengthener in NiAl. 3. Careful AP examination of boundaries that are oriented perpendicular or parallel to the direction of analysis is required to get the most accurate results. ppm (Takasugi et al. Variations of boron and Al concentrations with Ni (or Al) concentration at grain boundaries of Ni3Al0. Boron segregation to grain boundaries can be generally identified as the bright 4. In early work. Concerning the boron segregation and the alloy stoichiometry in the matrix (or at grain boundaries). 1990).. Among these properties. ppm) and was shown to suppress intergranular fracture (George and Liu.2 FIM-AP Observations FIM combined with AP analysis provides a large number of results for the chemistry and for the associated grain-boundary structure of Ni3Al with and without boron. a large number of properties associated with the grain boundaries of many intermetallic compounds (for example. 213.. The co-segregation of Ni and boron to grain boundaries was difficult to ascertain because of the considerable concentration fluctuations.. a part of which has already been described in the previous section. However. electrical conductivity. and AgMg. Figures 10(a) and (b) show the FIM image and the AP result on grain boundaries of Ll2-type Ni319(Si0 53Ti0 47)081 containing 16 wt. the consensus for the results obtained by FIMAP was that there was an enrichment of Ni at least at some of the boundary planes and its zone was of the order of 2nm wide. 33. during recent decades.%) Figure 9.12 at. intergranular fracture has been well-interpreted based on a number of experimental data described in the previous sections and also with help of theoretical calculation. NiGa..% (300 wt. and corrosion behavior) have been observed to be very interesting and specific (e. Mater. Westbrook. However. Copyright (1985) Pergamon Press Ltd) images at grain boundaries and also can be quantitatively measured by AP analysis. 1991). In this case. In many intermetallic compounds. it was shown by the compressive test and the fracture test with chisel and hammer that boron has a tendency to segregate to the grain boundaries in stoichiometric and Al-rich B2 (cP2)-type AlRu. NiAl.% boron (Reprinted with permission from Liu et al..g. Boron segregation to the grain boundaries was detected in stoichiometric B2 (cP2)-type NiAl polycrystals at a bulk level of 0.

alternately. ppm boron in the bulk (Reprinted with permission from Takasugi et al.g. as shown in Figure 11 (Westbrook and Wood. 1965). have been successfully employed to restore ductility (Westbrook.g. Mg-rich AgMg. the intermetallic compounds that exhibited grain-boundary hardening underwent disintegration at intermediate temperature. the so-called 'pest effect. 1993) e.GB GB Number of ions detected mln Figure 10 (a) FIM image and (b) the AP result on grain boundaries of Ll 2 -type Ni 319 (S 0 53Ti0 47 ) 081 containing 16 wt. 225 0 C for AgMg and 900 0 C for NiGa) or. Also.' .. drastic quenching treatments from above a minimum temperature (e. Grain-boundary hardening has been eliminated by several means. 1963). adding third elements such as Zr to FeAl.

for example. indirect insight into the chemistry of the grain boundaries can be obtained by referring to early work by Westbrook (1967). where it reacts to form MoO3 and SiO2. At intermediate temperatures harmful elements like oxygen segregated in the region near the boundary. unfortunately much time has passed without any direct observations of the structure and compositions at the grain boundaries of these intermetallic compounds. the susceptibility to pest disintegration increases with increasing Mo content.8% Mg 50. a recent study for polycrystalline MoSi2 showed that 'pest' disintegration occurs through transport of oxygen into the interior of the specimen along pre-existing cracks and/or pores (grain boundaries). At high temperatures. Westbrook and Wood (1963) proposed a model based on preferential grain-boundary diffusion of a reactive gaseous element and segregation of the element in the lattice structure near the grain boundary. Grain-boundary hardening in AgMg. the controlling element would be identified and also the correlation between these chemical properties and the chemistry at the grain boundary would be understood.g. NiAl. in which electrochemical potentials and the grain-boundary corrosion properties are affected. Silicon-rich compositions did not disintegrate. Extensive studies beyond those directed toward the boron effect on the intergranular fracture of Ni3Al are required to solve these phenomena entirely. . direct measurements for the chemical composition of the grain boundaries of these intermetallic compounds had not been made at that time. or complex defects (or compounds) with vacancies (or component elements).Grain boundary Grain boundary Hardness (kg mm-2) 49. hardening and embrittlement are relieved. at low temperatures the oxygen diffusion from the environment is too low to penetrate to the grain boundaries. and other impurities. nitrogen. Thereby. instead.3% Mg Typical size of indentation Distance from grain boundary (//m) Figure 11. nor did they form the pest oxide (MoO3). these previous experimental observations for grain-boundary hardening indicate that there is a strong correlation between grain-boundary hardening (and also the associated grain-boundary brittleness) and alloy stoichiometry and impurity atoms (such as gaseous elements or impurities at the grain-boundary planes). segregation of oxygen. MoSi2. Westbrook (1965) has demonstrated that these phenomena were associated with the impurity distribution at the grain boundaries by. and ZrBe13. However. because the bulk diffusion and grain-boundary diffusion rates for oxygen become comparable. Also. illustrating hardening only with a stoichiometric excess of the more electropositive element (Mg) (From Westbrook and Wood. Thus. lead to hardening and embrittlement. Also. in spite of Westbrook's (1965) claims. These results could be explained by direct observation of the chemical compositions at the grainboundary plane. The internal stress produced during the formation of MoO3 results in disintegration to powder (McKamey et al. a continuous dark-gray protective oxide (SiO2) formed on the surface. it was shown that near stoichiometry. e. Indeed.f 1992). On the other hand. However. 1963) It was shown by several observations that there is a close correlation between the upper temperature limit for the disappearance of grain-boundary hardening and the temperature limit for the disappearance of the disintegration effect in intermetallic compounds.

. type S is the fully symmetrical configuration. and FIM. The conclusions of these theories regarding the existence of special coincidence boundaries and the determination of possible Burgers vectors of dislocations in these boundaries have been confirmed in a number of experimental studies (Buis et al. it was shown that there are two types of configuration (i. These theories do not succeed in describing the energetics and thus the detailed atomic structure of grain boundaries. The parameter describing the 'position of the grain boundary' similarly yields several configurations. forming with the same kind of atoms as each other. The former are the 'relative orientation* and the 'grainboundary inclination*. The CSL theory can give physical and geometrical meaning to the results obtained and has the possibility of showing any crystallographic nature that might be unique in grain boundaries of intermetallic compounds.5. i. where one of the three sublattices forming the A atoms in one half of the crystal is symmetrical to the sublattice forming the B atoms in the other half. while type III consists of the plane where the lattice positions are all occupied by A atoms. Considerations of symmetry in the lattice and of the geometric degrees of freedom in the grain boundary generally require two 'macroscopic' parameters and two 'microscopic* parameters to describe grain boundaries.' associated with the atomic 'core' structure of grain boundaries. the tilt boundary (Takasugi and Izumi.. 1985. 1970. Modeling of Grain-Boundary Structures In spite of a large number of direct observations by HRTEM. Brandon et al.13° tilt boundary with [100] rotation axis of an A3B Ll2-type intermetallic compound. the grain-boundary structure in intermetallic compounds is not well-understood. the favored configuration of the grain boundary as determined by the CSL theory can give us a model of the periodic boundary. Therefore. Grimmer et al. Foiles. Farkas and Rangarajan. Frost.e. Indeed. 1987. Pond and Smith. 1970.. The early investigations of the structure of grain boundaries in intermetallic compounds were based on geometric and crystallographic concepts (Takasugi and Izumi. type S and type PS) about the 'relative translation' and also two types of configurations (i. 1987. which will be used in computer calculations in order to accommodate atoms. Also. geometrical consideration of the grain-boundary structures of intermetallic compounds reveals the possibility of the existence of variants for the atomic configurations if their energies are not too different. In this case. 1988). Vitek and Chen. 1991).. However. 1989a. Recently a number of atomistic modeling studies have been carried out (Chen et al. Brandon. 1988.e. 1987. Schober. From these analyses. 1987). interaction of lattice dislocations with grain boundaries was studied (Pestman et al. The CSL grain-boundary structures of the A3B Ll2-type and the AB B2-type intermetallic compounds have been considered for two extreme cases of geometry.. 1967. 1989. 1966. 1990. 1972. 1983. Kruismanefa/. Figure 12 shows eight configurations deduced for the L = 5. Ranganathan. For the latter type configurations. 1986. 1964. type II and type III) about the 'position of grain boundary. modeling of the grainboundary structures and the implications deduced from such modeling become very important to the understanding of the properties of grain boundaries in intermetallic compounds. while the latter are the 'relative translation' and the 'position of the grain boundary. Thus. Farkas. it was also deduced that . As an example. 1974). Schapink and Tichelaar. where the sublattices in one half of the crystal are really symmetrical to the sublattices in the other. Also.. owing to experimental limitations. this limitation does not diminish the significance of the application of CSL and O-lattice theories to grain boundaries of intermetallic compounds. 1977). 1970. 1987. type II consists of the plane where A atoms and B atoms are evenly distributed. the 'relative translation' parameter yields two different types of configurations for the CSL grain boundary in the case of a pure tilt misorientation. as described in an earlier section. as they have been quite useful in disordered alloys. 5. Schober and Balluffi. 1989. Vitek et al. 1983) and the twist boundary (Takasugi and Izumi. Ackland and Vitek.' For the former type configuration. TEM. One is a fully symmetrical configuration and the other is a pseudosymmetrical configuration. 1986. 1990. 53. and then produces different atomic compositions of A and B atoms at the grainboundary plane. First.e. while type PS is the pseudosymmetrical configuration. On the other hand. 1991). Bollmann et al...1 Geometric and Crystallographic Models A quantitative description of grain boundaries in pure metals and disordered alloys has been achieved in the framework of the coincidence-site lattice (CSL) and the O-lattice theories (Bollman. computer simulation permits the minimization of the energy of an assembly of a large number of atoms and therefore gives a stable structure in the grain-boundary region. 1987. 1989. 1988). including limitations in resolution. TEM observations for twin boundaries of Ll2-type Cu3Au indicated the existence of variants of the atomic configuration at grain boundaries (Tichelaar and Schapink.

Depending on the portion of the two types of wrong pair bonds. the existence of different near-neighbor bonds at the grain boundary may produce different electrochemically related properties. respectively. and again the possible ordering configurations were the number of configurations created depends upon the type of crystal structure. The excess energies of the different configurations due to the wrong bonds were estimated by counting the wrong pair bonds across the grain boundary (Takasugi and Izumi. 1990). Thus. It was shown that: • The excess energy does not systematically depend on the misorientation (or the value of E). respectively.13° tilt boundary with a [10 0] rotation axis as an example of the A3B Ll 2 -type intermetallic compounds (Reprinted with permission from Takasugi and Izumi. Also. These are the antiphase boundary (APB) fault. 1983. and misorientation. 1987). The variations for different configurations in atomistically flat or stepped grainboundary planes in the Ll 2 structure were discussed. and PS-B-III configurations contained a SF. and the excess energy differences between these types become less in the high-angle (general) boundaries. grain boundaries in intermetallic compounds may involve the incidence of bond defects in addition to lattice distortion. 1987. Also. For example. the wrong bonds are visualized by the full lines between the two atoms in Figure 12. shows widely different values relative to those of the AB B2 structure. • The excess energy in grain boundaries of the A3B Ll 2 structure. rotation axis. the S-II and PS-C-III configurations did not contain any planar faults. distinct columnar cavities will occur in grain boundaries owing to the depleted bond pairs or undistorted but misoriented lattices meeting at the grain-boundary plane. Wrong pair bonds across the boundary were usually recognized in these geometrical analyses. it was predicted that. This prediction agrees with direct observation of grain . if the chemical ordering is dominant. suggestive of multiple structures with nearly the same energy. 53. An extended analysis of geometrical considerations for grain boundaries in ordered structures has been carried out by Frost (1987).e. three types of faults were defined in analogy with the planar faults in the perfect lattice of the ordered structure.and structure-related properties even for the same grain-boundary facet. open and full symbols correspond to A atoms and B atoms. PS-A-II. Thus. In the grain-boundary region of intermetallic compounds. PS-A-III. in Figure 12. On the other hand. the existence of different ordering configurations at grain boundaries may create different chemistry. reflecting a large anisotropy of the APB energy on several crystallographic planes in the perfect lattice. as described in an earlier section.Figure 12. proper bonds with unlike atoms may sometimes not be conserved or can be distorted. For example. and the complex stacking fault (CSF). A . also. Note that large and small circles represent atoms distributed on the first (10 0) atomic plane and the second (10 0) atomic plane. • The excess energy usually depends on the type of configuration. Thus.y 31. PS-B-II. wrong bonds between like atoms may be introduced. unless wrong bonds are permitted at grain boundaries.. the stacking fault (SF). the SIII. 1983. i. Copyright (1983) Pergamon Press Ltd.A and B-B pairs. and the PS-C-II configuration contained an APB. 1987). Eight atomic configurations constructed on the E = 5. Geometrical analysis gave good insight into this problem (Takasugi and Izumi.Ti) compound (Figure 1) (Sasaki et at. although energetic requirements dictate that these defects are minimal. The linkages shown are wrong-atom pairs between two atoms boundaries of an Ll2-type Ni3(Al. Ada Metail.

1989). Indeed. 1991). Previous hard-sphere models indicated that the difference in size between the constituent atoms and in the associated ordering energy is a significant factor controlling the grain-boundary structure of Ll 2 alloys (Farkas and Rangarajan. metals and Cu3Au.. the grain-boundary structures obtained were similar to those reported for pure f. In Cu3Au and. 1992). 1989). A more sophisticated geometrical crystallographical method. (b). it was demonstrated that the results shown in Figure 13 are applicable to other [001] symmetrical tilt boundaries. respectively (Ackland and Vitek. were analyzed. dependence on alloy system). rather than the misorientation dependence.e.c. In both intermetallic compounds. 1989a. it was demonstrated that a step in the boundary is associated with an interfacial glide dislocation and also that larger steps in the boundary can only occur for such disolocations if additional shuffling of atoms occurs (Schapink and Tichelaar. those in pure f. 1990). Grain boundaries of pure f. was used to analyze possible configurations in the Ll r type intermetallic compounds (Farkas and Rangarajan. 1987). 1988).c.. and Ni3Al alloys. being practically the same as that found in Ni. Also. such as the selection of component elements (i. consisting of atoms of different sizes. and the atoms in the boundary region cannot be uniquely assigned to either one grain or the other. local-volume potentials (Chen et al. the stoichiometric effect. in Cu and Ni. metals. in Ni3Al. while in the case of the weakly ordered alloy (Cu3Au) and also for pure f. metals like the constituent Cu atoms.. Two distinctly different intermetallic compounds.c. 5. These potentials were all empirical and were fitted to reproduce the properties of the alloys studied. Figure 13(a) shows the energetically most favorable structure in Cu. and (c) show the structures of the E = 5 (3 1O)[OOl] symmetrical tilt boundaries in pure Cu. 1987. However. thus suggesting the existence of variants of the atomic configurations at the grain boundary. Based on these facts.c. Concerning the stoichiometric effect on the grainboundary structures of A3B Ll 2 alloys with high ordering energy. On the other hand. Pestman et al. consisting of atoms of similar size.. for the E = 5 (1 3 0) tilt boundary. Cu3Au and Ni3Al.e.presented (Frost. i. Indeed. the symmetrical E = 5 (3 10)[001] and (21O)[OOl] tilt boundaries and other CSL boundaries were studied in detail using appropriate pair . respectively) (Schapink and Tichelaar. The major interests of these calculations have been directed toward the compositional effect.c. and less dense than. 1989). several boundaries were found to present two possible structures. all atoms can be regarded as uniquely attached to either one or the other grain. For example. Geometrical analysis in the frame of the CSL theory also indicated that the possible atomic structure in Ll2-ordered alloys can be generally altered by interface dislocations with DSC° or DSCd vectors as Burgers vectors (where DSC° or DSCd vectors are ordered or disordered DSC vectors. the HRTEM observation on Ni3(Al1Ti) alloy described in an earlier section suggested that there was little relaxation of atoms. the shortest DSC0 parallel to the (1 3 0) boundary plane was given by y [3 1 0]. metals and of Ll 2 alloys with low and high ordering energies were calculated using Af-body potentials (Ackland and Vitek. This structural difference was attributed to a distinct difference in ordering energy: in the case of the strongly ordered alloy (Ni3Al). and the atoms in the grain-boundary region persisted in the ideal ordered structure by removing wrong bonds and also by preserving right bonds (Sasaki et al. 1990. of course. in which a modified hard-sphere model and CSL theory were involved. Cu3Au. 1987).. using pair potentials (Kruisman et al. 1988). the configurations of grain boundaries that are most dense were found to be generally different from. Ni3Al. Figure 13(a).c. Ackland et al. Thus. for the strongly ordered alloy. For the weakly ordered alloy. As alloys with low and high ordering energies. and the alloying effect in crystallographically identical boundaries. Quite a similar result was obtained by computer calculation (Ackland and Vitek. 1990). so that the ideal Ll 2 structure is practically undisturbed on either side of the boundary up to the boundary plane. the relaxation of atoms in the boundary region is substantial.c. Vitek et al. Cu3Au and Ni3Al were calculated as examples. 1987). the chemical order is the principal controlling factor of the energy of the system. a more relaxed grainboundary structure with a possible chemical disordering is energetically favored. these interfacial dislocations were observed to exist in the Cu3Au alloy in agreement with theory (Tichelaar and Schapink. and N-body potentials (Ackland and Vitek. it was demonstrated that the segregation of elements of smaller size would be strongly favored. 1988)..2 Computed Atomic Models Most of the calculations of atomic structure of grain boundaries in intermetallic compounds have been made on the Ll 2 structure. Cu3Au.c. 1991). Calculations for still other tilt and twist boundaries reveal their general validity (Ackland and Vitek.

. 1988).. away from stoichiometry the structures resemble more closely the corresponding structures in singlecomponent systems. inter metallic compounds. extended calculations for other types of grain boundaries (i. 1989) potentials (Kruisman et al. in fact. i.. Sutton.e. 1984). that Al-rich grain boundaries have higher energy (Chen etal. Wang et al. and with A-atompoor composition. 1988). 1984.Figure 13. was still applicable to the grain-boundary structure of ordered alloys. These figures clearly indicate that the structural units in the stoichiometric boundaries were more distorted than those of either A-atom or B-atom-rich boundaries. with A-atom-rich composition. and (c) Ni3Al alloys. Furthermore. 1986. Also. Calculations showed that the concept of structural units.e. other CSL boundaries) showed that the multiplicity of the grain-boundary structure types was much more .1989a. Energetic considerations for the segregation of surplus atoms to the grain boundaries of Ni3Al answered this inconsistency.1987. respectively (Reproduced with permission from Ackland and Vitek. it has been shown. the structural changes associated with the deviation from stoichiometry toward Ni-rich compositions can explain the fact that the ductilization of Ni3Al was observed in Ni-rich compositions. Here. Consequently. originally developed for pure metals (Sutton and Vitek. 1983. in Figure 15 a surplus of two A atoms was added per period and in Figure 16 two A atoms were removed per period.. (b) Cu3Au. but cannot explain the fact that the same effect was not observed in Ni-poor compositions. while this is not the case for Al enrichment at the grain boundaries (Kruisman et al. respectively. Thus. 1990). Ni enrichment at the grain boundary can be reached without any significant difficulty. 15 and 16 show the relaxed atomic structures at grain boundaries in compounds with stoichiometric composition. Structures of the E = 5 (3 1 O)[OO 1] symmetrical tilt boundaries in (a) pure Cu. Figures 14.

In Ni-rich alloys the segregation energy was 13 kJmol" 1 . 1990). The fracture strength of boundaries that consist of various compositions involving surplus of Ni or Al and boron in interstitial sites was calculated. 36. as has already been predicted from geometrical analysis of intermetallic compounds. 2729.Figure 15 D = 5 (3 1 0) boundary structures in A3B Ll 2 with a surplus of two A atoms per period relative to the stoichiometric composition (Reprinted with permission from Kruisman et al. where the maximum stress needed to break a bicrystal consisting of a L = 5 (2 1 0) symmetrical tilt boundary is plotted for various chemical compositions of the grain boundaries.. Two structures with nearly equal energy are shown (Reprinted with permission from Kruisman et al... though not very strongly. c. Copyright (1988) Pergamon Press Ltd) .. 1989a. 36.c. 1988). The effects of boron doping on the grain-boundary structures and energies of Ll2-type Ni3Al alloys have been calculated using appropriate potentials (Chen et al. and thus Ni segregation is favored. This was attributed to the fact that the grain-boundary structures of ordered alloys differ not only topologically but also compositionally (Kruisman et al. b. with the results shown in Figure 17. ActaMetall.. However.. showing the enhanced Figure 16.c. Copyright (1968) Pergamon Press Ltd) segregation propensity. the segregation energy of Ni was found to be 44 kJ mol"1. Copyright (1988) Pergamon Press Ltd) Figure 14. 2729. ActMetalL. The stress was the highest in the case of the Ni-rich boundary with boron and the extensive compared to pure metals. the fitting of which has been done by linear muffin-tin orbital (LMTO)-based quantum calculations for a number of Ni-B and Al-B compounds and f. boron. 2729. Acta Metall. £ = 5 (3 1 O) boundary structures in A3B Ll 2 with stoichiometric composition. E = 5 (3 1 0) boundary structures in A3B Ll 2 with a deficit of two A atoms per period relative to the stoichiometric composition (Reprinted with permission from Kruisman et al. when boron is present in grain boundaries. 36.

has the same beneficial effect as in Ni3Al. the interactions (e. attraction. In a strongly ordered A3B Ll 2 alloy involving the contribution of Maximum stress (GPa) . 1989b.e.c. in this section only a brief description is given. and (5 0/5 0) or Al-rich lowest in the case of the Al-rich boundary without boron.Bulk GB 100/100 GB 100/50 GB 50/50 (Al-rich) Figure 17. A first qualitative consideration was based on geometrically constructed grain boundaries with the Ll 2 structure and some principal features were demonstrated (Izumi and Takasugi. 1987). Cu3Au) and high ordering energy (i. 1990).y 1990).e. • The structure and energy of grain boundaries are very much affected by the stoichiometry of the component atoms and by the addition of interstitial atoms. c).y 1991). then grain-boundary fracturing was suppressed (George and Liu.g. such that a sufficient amount of boron segregates to the grain boundaries without introducing significant solid-solution hardening. and transmission) between y<l 1 0 > screw dislocations and grain boundaries with symmetric [110] tilt boundaries were calculated under an applied shear stress or with no applied shear stress (Pestman et al.. Few calculations have been done for grain-boundary structures and energies in intermetallic compounds with other than Ll 2 crystal structures. substitutional sulfur was shown to be a very deleterious element in reducing the fracture stress (Chen et al. several principal grain-boundary structures and their properties in Ll2-type intermetallic compounds are recognized from these computer simulations: • The grain-boundary structures with least energy are not necessarily similar to those of the corresponding boundaries in pure f. thus creating a higher stress concentration near the grain boundary than that in disordered alloys (Pestman et al. Since a detailed discussion of this subject will be given in the chapter by Briant in this volume. it was shown that segregated boron appears to decrease the intergranular cohesion and therefore does not have a beneficial effect on the ductility. This experimental result showed that boron in NiAl behaves in a similar way. Also. Some arguments for the chemical bonding and the electronic structures in grain boundaries of intermetallic compounds have been made in relation to the intrinsically brittle/ductile properties of Ll 2 alloys. i. Electronic Bonding Aspects The chemical bonding nature and the corresponding electronic structure in grain boundaries.c. Also. metals.e. Thus. Maximum stress needed to break a bicrystal of Ni3Al consisting of a E = 5 (2 1 0) symmetrical tilt boundary plotted for various chemical compositions of the grain boundary (Reproduced with permission from Chen et ai. • Variants with different ordering configurations exist in ordered alloys and are much more extensive than those in disordered alloys. It was thus suggested that Al-rich grain boundaries are more likely to fail than those which have the stoichiometric composition or the constituent Ni or Al metals. the experimental result is somewhat different from this prediction: when boron is added to stoichiometric NiAl.% 1989a). strongly affect not only grainboundary properties but also bulk properties. Only one example is presented: the grain-boundary structures of B2-type NiAl were calculated using many-body potentials (Chen et al. thereby resulting in columnar cavities. The chemical compositions at the grain boundary are denoted as (1 0 0/1 0 0) or Ni-rich when both sides of the grain boundary comprise 100% Ni layers. It was shown that transmission of the dislocation through the boundary occurs at high stress levels. However. and the superpartials constituting the arriving dislocation decrease their separation in response to the applied stress. while the stoichiometric composition was more homogeneous and less distorted. Calculation showed that both Ni-rich and Al-rich compositions led to grainboundary vacancies. 6. absorption. Using many-body potentials representing an Ll 2 ordered alloy with low ordering energy (i. The other two cases are denoted as (1 0 0/5 0) or stoichiometric. Ni3Al).y 1991). which are primarily associated with their crystal structure and chemical composition. E = 5 (2 1 0) and (3 1 0) symmetrical tilt boundaries were studied in terms of the stoichiometric effect and the boron doping effect.

A-B covalent bonds perpendicular to the grain-boundary plane. the structures of grain boundaries of intermetallic compounds observed by transmission electron microscopy (TEM). This idea originally stems from the mechanism by which grain-boundary brittleness has been proposed to arise in a grain boundary to which impurity atoms are segregated (Losh. 1979.. resulting in a lower grain-boundary cohesion. and the doping of boron) on the grain-boundary strength (and resultant fracture) of Ll2-ordered A3B intermetallic compounds can be interpreted by the electronic modification of the electronic structure and atomic bonds through compositional modification at the grain boundary (Izumi and Takasugi.b. the bond may be characterized by the * directionality' and the 'heteropolarity' of the electronic charge distribution. there was a markedly enhanced hybridization with s orbitals on both Si and Ni throughout the bands for the grain-boundary structure of Ni3Si (Figure 19(a)). This situation is schematically shown in Figure 18. should be limited. Cluster calculation of the density of states was done for Ll2-ordered A3B intermetallic compounds (Eberhart and Vvedenski. the 'heteropolarity' of electron charge density due to A-B bonds at the boundary plane introduces something like a penny-shaped cavity into the locality. and field ion microscopy (FIM) were presented. 1986. in contrast to the diffuse sp levels of the parent crystal. addition of substitutional atoms. Losh. 1986. Some attempts to understand the electronic nature of the grain boundary have recently been performed based on theoretical calculations (Chen et al. Thus. Consequently. From the point of view of the 'directionality' (Gelatt et al. Next. Wakeyama et al. high-resolution transmission electron microscopy (HRTEM). the angle between A-B bonds from a given B atom will always be close to the desired 90° value.e. 1983. the relative s-orbital electronegativity of the constituent atoms was predicted to provide a reliable indication of the grain-boundary cohesion. stoichiometry. 1986). for which the non-central contribution to the energy of formation is at a minimum (i. 1988. 7. Evidently. preservation of the chemical ordering without significant local atomic . the grain-boundary polyhedron developed directional Pd levels above the Fermi energy. from the point of view of the 'heteropolarity. preservation of the ordering close to the boundary plane. bond directionality) concomitant with grain-boundary structure than are those of Ni3Al and Ni3Si. which is ductile. Based on this calculation. It seems unlikely that. being consistent with experimental data. A-B bonds (i. the HRTEM technique especially provided very significant information: for example. Hashimoto et al. Figure 19 shows the density of states calculated for crystals in the bulk and polyhedra at a grain boundary in Ni3Si and Cu3Pd alloys. 1976.. Eberhart and Vvedenski. 1988). B atoms have a greater tendency to pull electron charge out of A-A bonds. 1979). the grain boundaries of Cu3Pd are better able to accommodate the bond misorientation (i. 1980. Schematic representation of the bond pairs at a grain boundary of the A3B Ll 2 alloy (Reproduced with permission from Izumi and Takasugi. Of these direct observations. 1984a. Miedema. atom A). When the valency difference between A and B atoms is larger (or component atom B is more electronegative with respect to the other component.strong covalent bonds. Messmer and Briant. Summary Grain boundary Figure 18. Thus. 1984). It was then demonstrated that the reduction of charge density in AA bonds results in reduction of the grain-boundary cohesive strength. most stable). 1988).e. across a grain boundary. covalent bonds) prefer to be at a 90° angle. it was demonstrated that alloying effects (such as component atoms. Masuda-Jindo. 1980. Shao and Machlin.. which are supposed to sustain the grain-boundary cohesion. 1987) In this chapter. However. For brittle Ni3Si. Briant and Messmer. where the atomic configuration and bond nature in a geometrically constructed 12 = 5 CSL boundary with [100] axis is drawn. for Cu3Pd.' the B atom withdraws electron density from the A-A (metal-metal) bonds and forms ionic A(metal)-B bonds. Based on this concept. Takasugi and Izumi. the valency difference and the electronegativity criterion can generally be justified as measures of transfer of electrons between constituent A atom B atom atoms and thus as predictors of grain-boundary brittleness. 1983. 1987..e.

studies of grain boundaries of intermetallic compounds have been limited to the Ll 2 (cP4) structure. particularly that Ni3Al. 1986) accommodation. A number of other types of intermetallic compounds have their own attractive properties. However. and FIM combined with atom-probe method (FIM-AP) were presented. the associated plastic deformation. e. Density of states (DOS) calculated for crystals in the bulk (-—) and polyhedron at a grain boundary (—) in (a) Ni3Si and (b) Cu3Pd alloys. Specific grain-boundary structures have been observed by the HRTEM technique. Particularly. However. These models predicted that variants exist. Therefore. The structure and chemistry of APBs are discussed in Chapter 21 by Sun in this volume. and the existence of variants of the atomic configurations. i. and also established that the grain-boundary strength and the associated brittleness depend on the ordering energy.g. intercrystalline interfaces similar to grain boundaries. Thus far. Understanding of grain-boundary structures in intermetallic compounds has progressed much since the early 1980s.e. and the doping with interstitials. which has high potential as a high-temperature structural material. Among these direct measurements. a detailed description is beyond the scope of this chapter. a multiplicity of possible atomic configurations. APBs are twodimensional defects. Not only experimental work but also basic theoretical approaches were required to understand the associated mechanism. and also a number of sophisticated methods of calculation have been provided to determine structural and electronic properties.DOS (arbitrary units) DOS (arbitrary units) Energy (Ry) Energy (Ry) Figure 19. These findings stimulated many researchers. the electronic structures in grain boundaries of intermetallic compounds must be studied. A significant body of work from both experimental and theoretical points of view has been done on grain boundaries in covalently bonded III-V and II-VI compounds. i. Next. . a number of problems still remain unresolved and therefore many more studies are needed before well-founded conclusions can emerge. TEM combined with energy-dispersive X-ray technique (TEM-EDX). and microstructures of many intermetallic compounds. alloy stoichiometry. Finally. the phenomenology of APBs has many parallels with that of grain boundaries. It is well-known that the antiphase boundaries (APBs) in intermetallic compounds affect many material properties. for which more general grainboundary information is needed. was ductilized by boron doping. respectively (Reproduced by permission of the American Physical Society from Eberhart and Vvedenski. the structure and the associated electronic structures of which strongly affect electronic properties of devices.e. Then geometrical and atomistic computer models for the structure of grain boundaries of intermetallic compounds were presented. for the interrelation between alloy stoichiometry and doping atoms such as boron. This progress is due to some exciting findings. the nature of the chemical bonds and the associated electronic structure were discussed in relation to the atomistic structure and the composition of grain boundaries. the chemistry of grain boundaries of intermetallic compounds observed by Auger electron spectroscopy (AES). such as the dislocation structures. AES and FIM-AP techniques provided very distinct results.

. M. Sieloff. (1986). (1990). A. W. Schapink. D.. J. J. (1990a). Mater. H.... and Vitek. /. . 35. Y. Horton. T. Boring. Bollmann. Symp. 667.. and Srolovitz. Soc. K. G. 1. Hashimoto. A. and Blakely. C A. Scripta Metall. (1992). and Takasugi. 1807. E.. Glakov. J. A. V. 375. M. O. Bollmann.. Rev. Res. P. Sendai. Kruisman. S. 29. A. Mater... Surf. Y. (1979). A. (1979). London.. S... S. Th. Acta Metall. and Vvedenski. 24. King.. (1987). 467. V. Albers. (1990b). (1970). Balluffi.. (1990). Res. 32. and Pennycook. (1990). D. A. Wakayama. /. Rasmussen.. M. R.. 36. Phil. Brenner. Boring. Phys. L. S. M... Komem. and Schapink. Choudhury. and Srolovitz.. (1992). Mackenzie... Proc. Baker. (1991b). Voter. J. 379. C. Schulson. 353. (1984a). T.. 35. D. F.. White. D. 20. Academic Press. O. 1277. V.. Scripta Metall. A.. A.. Lett. R. and DeHosson. (1987a). Doyama. Mag. F. 363.. and Burke. M.. and Schulson.. M. Interfacial Segregation. Grain Boundaries in Metals. B62. P. Proc. Yamamoto. F. J. Kung.. References Ackland. Mag. Mater. D.-J. S. 36. (1983). 25.. B27. H. Farkas. 35.. B57. (1990). J. A. Lett. 51. Metals. 735.. J. C T. George. Hashimoto. Gelatt. W. A. S. J. H. S. and Liu. T. Ishida. and Vitek.. E. George. S. Foiles. G. 81.. and Liao. Res. Williams. Res.. Metals Park. 754. Res.... Liu.. A. 61. (1989c). (1980).. H. Rev. ScriptaMetall. and Sass. Proc. (1985). C. V. P. (1957). (1988). Mag.. Res. Brenner. and Briant. Baker. 519. R. J. P. C L.. Srolovitz. L. C. 5. 2729. Devan. A. R. S. M. Liu. Symp. J. M.. Brandon. M. Ackland. and Messmer. E. J. Mater. J. Symp. Proc. P. Mater. Voter...-J. C . Res. Krzanowski.. Soc. Mag. ASM International. F. Ralph. Phys. 40. Mater. H. T. Briant. and Horton. Ranganathan.. R. (1990). Voter. D. Scripta MetalL Mater. L. A. Farkas. 1. M. 220. S. OH. Tichelaar. M.. and Warrington. 16. C. P. (1992). J. 27. Res. Kung. and Brooks. (1984b). Scripta Metal. S. . A. H. (1970). Symp. Chen. F.. 861. Chen. M. J. W. 10324. L. and Horton. 197.. 97-119. V. 25. Voter. D. R. 1271. (1992).... Walter. R. (1988). and Takasugi... Brenner. 23. (1989). and Moruzzi. A13. P. Trans. and Fujiwara. Scripta Metall. 99. A56. and Hua. Mater. Kung. and Sainfort. ScL. Izumi). 403. 12.. and Tichelaar. P. Yamamoto. Zotov. Johnson. A. (1987). A. (1988). V. J.. (1991). T. 1251.13. J. S. 81. (1987). Phil. Scripta Metall. M. A. Acta Metall. M. M. (1986). (1989). A30. Mater. pp 347-70. W. J. 61. 19. Losh. Proc. G.. Aoki. Energy... D.. Scripta Metall. M. and Hua. V. pp. W. S. A. A. R.. F. E.. (1991).. G. W. F. D. and Izumi. 40. Japan Institute of Metals. C. Rasmussen. and Skorov. 637.. Rasmussen. L. George. Y.. G. Mag. Mater. Scripta Metall. 57.. ActaMetall. Mag. Michaut. M. P. Crystal Defects and Crystalline Interfaces. 58. Liu. H. 24. 81. 23. (1976) Grain Boundary Structure and Properties. Tangri). 22A. Symp. Tichelaar. E. and Carmichael.. I. H. Soc. 3.. 979. Res. and WaId. F. 105.. Ananyn.. D.. S. Chen... and Rangarajan... P.. 671. C T. Soc. R. Vitek. Fleischer. R. D.. G. D.. and Finnis. In Structure and Property Relationships for Interface (eds J. P. S. and Voter.. Buis.. and Hua.. M. B. F. C L . V. F. (1987).. and Padgett. L. White. 1190. Soc. A. G. D. and K. 215. L. D. G. A.. P. Antonopoulous. (1989a). and Smith. (1953). Metall. Acta MetalL Mater. Acta Metall. I. Mater. L. Mater.. 33. R... Albers. Acta Metall.. F. and Doyama. T. F. M... S. Voter. D. R. 2005.. Scripta MetalL.. A.... Grimmer. 47. 1389. Berlin. M. McKamey. Buis. and Schober.. A. E. (1989). Status Solidi..133. 383. Acta Metall. Tichy.. Phys. R. E. J. A. (1972). In Intermetallic Compounds—Structure and Mechanical Properties (JIMIS-6) (ed. Acta MetalL. T. Acta MetalL Mater. C . Physique.. P... Busby. 81... Springer. Soc Symp. Phil.. C. Japan Inst. 45. Ishida.. F. S. and Taub. D. Brandon. Tortovelli. Oxford University Press. and Sass. Mag. L. H. and Sass. Horton. and Sass. D. 813. (1977). O. C. (1989). /. W. Oxford. J. R. E. Acta MetalL. 25. Brenner. Izumi.8. 195.. 11... Proc. J. Symp.. S. and Hay. W. 23. 217. C G. Jr. Res. L. J. J. (1991a). Phys. Phil. R. 186. Chen. 40.. (1972). Mater. M. 659. J. L.. and Wells. 1219. L. 5. C . Izumi. S. At. S. E. C . Scripta Metall. P. 45. S. and Hay. 1463. (1991b). and Miller. 105. (1986). and Liu. M. 2747. Acta Crystallogr. Mater. Wargu. J. and Schapink. (1987). D. (1991). American Society for Metals.B41. (1988) Phil. Chadwick. D. 149. G. (1985)... Phil. (1974). O. J. P. (1987). McLean. D. C. 955. Ackland. (1987b). D. Eberhart... S.. Horton. Phil. ch. Kung. S.. J. J. K. White. 31. A59. Chen. 4. S. P. (1991a). G. Briant. Metals. 24. D. Mater. J. Res. Mater. and Sass. D... George. Mater. Boring. 426... A.-J. 5. C L . Field. Soc.. Mater. 1259. Frost. E. Chen. (1964).. 62. 569. (1967). and Hay. Rev. 213. C2. B42. B. P. (1988). Acta Metall. Scripta Metall. J. (1989b). PmC1 133. (1990). 32. Vitek. Michael. 2761. Bollmann.. and Miller. M. 68. 81. C. W. 133. M. Acta Metall. P. Albers.. J. V. R... Res.. J... W. 173. 7. 1885. Acta Metall. Mater.

R. (1971).. (1988). Proc.. S.. Ada Crystallogr\.. (1988). Phys. 21. 263. (1991). and Sakuri. J. /. R. and Machlin. Mater.. 739. F. Masumoto. J. Metals Rev. W. Proc. T. and Petersen. Status Solidi. 237. Scripta Metall. Westbrook. Schapink.. 557. Physique... (1986). Sieloff. J. Proc. Ranganathan. 67. and Izumi. J. K. and Schapink. Images of Materials. Appl. Westbrook. H. Tichelaar. (1984).. Soc.. Proc. Goods. Mater. 47. C. (1976). R. 32. Schober. 353. New York.-F. R.. 1093. K.. Metall. A. Vitek. C . (1984).. H. L. T. 1417. A54. P. 433. D. J. Symp. Williams. Cl. 44. T. and Masahashi. Hanada. 429. Takasugi. Phil. 311. 913. P.. 441. S. Bruemmer). K. Phil. Alan.. B. S. Mater. F. O. Shao. J. A106. Technol. S. Grain Boundaries in Engineering Materials (Proc. Mater. (1965). F... Phil. (1990). (1986b). . p. J. T. L. Soc.. Mater.. Physique.. 21. 289. Res.. 38.. Less-Common Metals. Phil Mag. and Yalisove.. Vitek. Ogura. (1986). and Tichelaar. Trans. 29. Trans. Metals. Physique. W. Metall. and Schapink. Cl. A. and Cserti.. 31. A335. (1987). Sutton. G. J. F. Physique. T.. Grain Boundary Chemistry and Intergranular Fracture (eds G. V. edited by J. J. 1283. and Vitek. N. J. Mater. (1987a). S. Wiley. (1993). A. Acta Metall. Miller. V. Pestman. 209. A16. and Schapink. S. Scripta Metall. J. V. K. Acta Metall. D. H. M. Kruisman. S. Vitek. D. (1977). 133. Soc Symp. /. and Izumi.. 19. J. D. 30. J.. 46. Mag. P. 42.. (1966).. T. M. M. J. Mag. D. J. 1633. V. Izumi... and DeHosson... A. D. Westbrook. W.. 823. Mater. White. C. 191. Vac Sci. Hiraga. S. S. Acta Metall. O.. A4. 32. Voter. 1063. Sieloff. J. R. P. M.. R. and Chen. and Hirabayashi. W. Miedema. D.. Forum. Phil. Suzuki.. 91. Soc.. C. S. C. A. 49. Suzuki. F. Symp. M. V. Res. Mater. Schapink.. T. and Horton. (1973). 789. T. Schober. Acta Metall. M. 12.. H. Brenner.. 174.. and Vitek.. Foiles.. F. Intl.. E. Soc. Sieloff. D. 1417. J. Japan Inst.. A. 671-88. (1980).. Phil. 186. Padgett. M. Soc. 1811. and Burke. 551. 8. Was and S. S. 238.. P. 213. Mag. W. A. L. 289.. S. M. (1990). C5. (1963). (1983). Th. Vitek.. Pope. S.. 155. H. M. Res.. Rev. Mag. D. Miller. and Burke. P. P. R... (1967). (1991). Mag. T. H.. Ackland. 1587. L55. T. Scripta Metall. (1991). 145. 1237.. M. M. 377... (1988). D. and Schapink. New York. and Li. Intermetallic Compounds (ed. C2. 47. Physique. Symp. M. E. /. D. D. (1985). Chen. S. Hashimoto. Westbrook. Res. Acta Metall.. Wakayama. Y. O. Scripta Metall. Metals. S. pp. 5534. H. F. W. Fleischer. (1991).. Vitek. D. P. 36. D.. 21. P. Wang. and Smith. 415. K.. L.. Proc. T. D. F. 457. and Horton. Masuda-Jindo. Soc Symp. 993. DeHosson. F. Scripta Metall. (1968). Forum. A. T. M. (1989). Tichelaar. Baton Rouge. /. DeHosson. (1985).. 25.-J.. E. 20. E. Hanada. Sci. D. C5. F. Phys. Mater. D. and Wood. H. (1983). O. 1. (1986a). S. J.. Th. 1187. 237. B.. Shindo. Brenner. M. Sutton. 23. Miller. Messmer.. Mater. and Balluffi. (1983).. Proc. Ishida.. Tichelaar. J.. Brenner. E. Liu. A.Marcinkowski. D. S. Mills. Takasugi. Westbrook and R. Chem. R. 159. T.. A65. 51. R. Hono. Scripta Metall. J. X. 293. Takasugi. R.. B.. J. 17. W. J. R. Oxford University Press. S. Mater. Scripta Metall. T. Physique. Pond.. V. White. 109. and Hondros. and Horton. (1990). J. Westbrook). D. and Izumi. and Izumi.. O. W. (1987b). Seah.. (1989). Acta Metall. Takasugi. Fourth Bolton Landing Conf.. Tichelaar. and Takasugi. M. George. (1986). Hirabayashi.. 38. and Izumi. Acta Metall. 18.. P.). Bull. (1970). (1991). J. G. Phys. Mills. Takasugi. and Briant. 81. A309. M. Pestman. Sasaki. D. A. 47. Claitor's. 35... and Schapink. (1984).. 46. Phil.. 9. (1975). G.. Cl. Vol. 87. Sutton. (1986c). 25. M. This chapter was originally published in 1995 as Chapter 24 in Intermetallic Compounds. and Gronsky. 29.. M.. /. Inst. and Whetstone.. (1988). (1984). Shindo. J. 30. 197.. and Doyama. Massaro. Th. (1992). Yamamoto. Takasugi.. 49.. G. /. A. J. F. Solids. 1. M. and Ming-Jian.. O. 1: Principles. K. (1992). V. G.. 22. (1970). Mater. L. Res.. 2061. /. (1992). S..

1990).. 1986.5Nb alloys) in water-cooled nuclear reactors. Howe AECL Research. 1986) changes have been observed during irradiation of various alloys that are in service or have potential use in an irradiation environment.Chapter 8 Irradiation Damage Lawrence M. 1990. and ion-beam simulations of various kinds. The irradiation behavior of the various intermetallic compounds of zirconium containing Fe. From the applied point of view. However. 1989. This accelerated swelling phenomenon was attributed to an irradiation-induced crystalline-to-amorphous transformation associated with an enhancement of diffusion and plastic flow in the amorphous material (Rest et al. Ontario. behavior of amorphous materials. 1991). Canada) 1. Sn. 1988. Ontario KOJ UO. is of considerable interest. 1986). Cr. Edited by J. Cr)2 in the a-recrystallized matrix and the dissolution of the metastable precipitate Zr2(Fe. The ordered alloy Zr3Al (Ll 2 (cP4) structure) has some very attractive properties (low absorption cross-section for thermal neutrons.Fe)2 and Zr2(Ni. and acceptable corrosion behavior) for its use in a nuclear reactor. Birtcher et aL. Povolo and Hillairet. significant mechanical (Chapter 6 by Schulson in Volume 2. Deep River. Piercy. Zircaloy-2. Ni. 2000 John Wiley & Sons Ltd . L. which included the amorphous transformation and the dissolution of the intermetallic precipitate Zr (Fe. Schulson et aL.Fe) precipitates in Zircaloy-2 and Zircaloy-4 (Gilbert etal. and nature of the bombarding projectile. Zircaloy-4. b). this interest was associated mainly with nuclear reactor programs but it now extends to the fields of ion-beam modification of metals. Yang et al.. 1991. Motta and Lemaignan. 1985. Rullier-Abenque and Senateur. Chapter 6 by Schulson in Volume 2). and microchemical (Griffiths. 1989. physical (Rest et al. 1993a. particularly for the CANDU (Canadian deuterium-uranium) system (Schulson. 1989. candidate reactor fuel materials such as U3Si (D0c (til6) structure) and U6Fe (D2C structure) reveal extraordinary large voids at low and medium fuel burn-up (Rest et al.g. 1980. Canada (Present address: 2 Frontenac Crescent. as an irradiation-induced ordered -* disordered -• amorphous phase transformation occurs below ~ 700 K (Howe and Rainville. (1986) also reported on the phase instabilities in neutron-irradiated Zircaloy-4. its use in a nuclear reactor is somewhat questionable.. 1979). 1990). the transformation could occur with or without a concurrent preferential depletion of Fe from the precipitates. Griffiths. P. Initially. Fleischer ©1995. 1992a. 1990. Reactor Materials Research Branch.. good mechanical properties at elevated temperatures. 1990.. H. Hillairet. Some of the above effects appear to be associated with the macroscopic irradiation Basic Mechanical Properties and Lattice Defects of Intermetallic Compounds. which are contained in the fuel cladding and pressure tubes (e. 1988. Nb. Lefebvre and Lemaignan. Box 697. 1963. Motta et al. Yang et al. Chalk River. Yang. For example. and Si.. Further information on the potential use of Zr3Al as a structural material in nuclear reactors can be found in Chapter 6 by Schulson in Volume 2). 1979. Hillairet. Depending upon the precipitate. Chalk River Laboratories. and Zr-2. Westbrook and R. 1988. Introduction There is considerable interest in irradiation effects in intermetallic compounds from both the applied and fundamental aspects. 1977. temperature of irradiation.. Yang (1990) and Yang et al. ion-beam processing of electronic materials.O. KOJ 1P0.Cr) in the jS-quenched matrix. An irradiation-induced crystalline-to-amorphous transformation was found for Zr(Cr. 1974.

nickel-based solid solution). (ii) intermetallics in superalloys under cyclic loading where there is sustained shearing (and sometimes dissolution) of the precipitates (Brechet et al... See Chapter 16 by Stekly and Gregory in Volume 2 on superconducting applications. 1987). 1984). The development and refinement of metallic glassforming techniques.. attention will be focused on some specific areas and. ordered alloys) allow for some rather sophisticated studies of irradiation damage (Piercy. 1960. Consequently there is a strong impetus to obtain a detailed understanding of the factors controlling the crystalline-to-amorphous transformation. during irradiation. and it is essential.. Sabochik and Lam. and (iv) ordered compounds when formed by vapor-phase deposition (Bellon et al. mechanical alloying. particularly using molecular-dynamics simulations (Caro et al. 1993. 1992a. The field of irradiation damage in intermetallic compounds is rapidly expanding. Full exploitation of the above techniques requires a detailed understanding of phase stability during irradiation as well as the properties of the irradiationproduced defects. to understand fully the mechanisms involved. 1986). 1985). Other examples of where intermetallic compounds are sustained in non-equilibrium conditions by external forcing are (i) compounds produced by ion-beam mixing (Banerjee et al. However. the critical temperature Tc for the transition from the superconducting to the normal state can be reduced appreciably due to irradiation-induced disordering (Brown et al. This frequently involves the formation of an intermetallic compound in the implanted region. 1954). 1985. vapor and sputter deposition (Biickel and Hilsch. special features of the intermetallic compounds (e. and the rapid quenching of alloy melts (Klement et al.c. Irradiation damage studies on Ni-Al alloys provide pertinent information on the defect properties (Dimitrov et al. such as electrodeposition (Brenner et al.c.. Improvements in the determination of interatomic potentials has also resulted in more detailed calculations of point-defect properties in intermetallic compounds. 1988). 1967). The good structural stability and creep resistance at high temperatures of nickel-based 7/7' superalloys result from the presence of a hardening 7' phase (an Ll2type intermetallic compound. Antisite defects (atoms occupying wrong sites) need to be considered in addition to Frenkel defects (interstitial atoms and vacancies). 1991a. dispersed in a shortrange-ordered f. therefore. and fatigue properties. 1992). hydrogenation. (iii) intermetallics undergoing phase transformations during ball-milling (Martin and Gaffet. On the other hand. Even though amorphous alloys are metastable materials. Crystalline materials can undergo solid-state amorphizing transformations by interdiffusion reactions. Devanathan. The specific areas that will be covered are: . This procedure can give rise to a considerable improvement of the corrosion. Devanathan et al. Nb3Sn. 1990). they often exhibit properties that are superior to those of the corresponding crystalline material (Johnson. The atomic mobilities (via vacancy diffusion) in the two phases control the high-temperature properties (Voorhees. see Chapter 1 by Anton in Volume 2. or the application of pressure (Johnson. hopefully. 1978a). Ion-beam mixing can also occur... Various A15 compounds (e. 1993a. Kirk et al. and Okamoto and Meshi (1990). and the challenge of obtaining more comprehensive understanding in the future. 1977. 1986)... Instead. and in general there is a more complex behavior of the irradiation-produced defects than in elemental metals. 1963. the study of irradiation damage in intermetallic compounds offers an interesting challenge. 1950).response of these alloys in service. 1990. 1993. the creep rate is mainly controlled by dislocation climb either in the 7 phase or in the 7' phase. Adequate shielding from the irradiation would be needed to prevent appreciable depression of Tc if the above Al5 compounds were used in practice. and no attempt will be made in this chapter to cover all of the various aspects.g. have resulted in extensive investigations of the properties of metallic glasses and a considerable growth in applications (Steeb and Warlimont.. For example. wear. Compound formation using ion implantation and annealing is a powerful technique for forming multilayer structures (Celler and White.c). Investigations of the nature of the damage produced during irradiation have proven to be quite useful in understanding the behavior of intermetallic compounds that are driven far from equilibrium.. Duwez. some insight will be given into the physical processes involved. through these.b). Crystalline-to-amorphous transitions are discussed further in Chapters 31 and 29 by Greer and by Johnson in this volume. Rullier-Albenque and Senateur. the present state of our knowledge. see Figure 1.b. Lam et al.g. From a fundamental point of view. and thus depends on the diffusion rate in the corresponding phase. 1952. 1988). particularly in the near-surface region. and Nb3Ga) have possible applications in superconducting magnets in controlled thermonuclear reactors. Ion implantation is now used quite extensively for modifying the properties of metals. V3Si. Wagner and Wright.b. 1986).

In disordered specimens irradiated with 1. Schematic of a lattice of an intermetallic compound containing Frenkel defects (vacancies and interstitial atoms) as well as regions of chemical disorder in which there are antisite defects • Point defects in intermetallic compounds. The purpose of selecting electron energies of 1.Vacancy Interstitial atom Vacancy Region of chemical disorder (containing) antisite defects) produced by a focused collision sequence Interstitial atom Anti-site defect Figure 1. i. 1979). m0 is the rest mass of the electron. The principal objective of many of these studies was to attempt to elucidate the mechanisms responsible for the irradiation-induced disordering or irradiation-enhanced ordering. Electron irradiations were used in order to produce fairly simple defects. Consider. 2. as will be discussed in Section 3 (also see the review by Schulson. In many cases. Ordered and disordered Cu 3 Au samples were irradiated with 1.5 MeV was that at the former energy only Cu atoms were expected to be displaced in the Cu 3 Au lattice. Point Defects in Intermetallic Compounds 2. changes in electrical resistivity were monitored during the irradiation and subsequent annealing.5MeV electrons at — 2OK and subsequent isochronal anneals were performed from 30 to 450 K. c is the velocity of light. isolated pairs of vacancies and interstitial atoms.0 and 1.e. For electron irradiation. four annealing stages were . • Irradiation-induced amorphization. for example. The most extensive studies of point defects in intermetallic compounds have been performed in ordered alloys. the maximum energy transferred per primary collision is given by £ T>max = 2E0(E0 + 2An0C2JZM2C2 (1) where E0 is the energy of the incident electron. In such alloys both irradiation-induced disordering and irradiation-enhanced ordering can occur.0 and 1.1 Experimental Investigations A variety of experimental methods have been used to elucidate the properties of point defects in intermetallic compounds. and M2 is the mass of the struck atom in the sample. whereas both Cu and Au atoms should be displaced at the higher energy. (1973). the experiments on the electron irradiation of Cu 3 Au by Gilbert et al. • Irradiation-enhanced ordering and irradiationinduced disordering of ordered alloys.0 MeV electrons. More detailed information on point defects in intermetallic compounds can be found in Chapter 23 by de Novion in this volume. The specific examples given below are relevant to understanding the irradiation behavior of these compounds.

centered at about 90. which appeared to be the dominant vacancy species. b) using electrical resistivity measurements. and migration of vacancies initially trapped at antiphase domain boundaries. Other ordered compounds having this structure are Zr 3 Al. Minimum displacement threshold energies for Cu atoms in ordered and disordered Cu3Au and CuAu were determined at 20 K by Alamo et al. Computer studies of irradiation damage in ordered Fe3Al (Jackson et al. respectively. The Ni vacancies in Ni3Al. The Au interstitial is believed to take the place of a Cu atom and to produce a Cu interstitial associated with an antisite defect.c.36MeV electron irradiations and electrical resistivity measurements. Cu atoms are nearest neighbors in <1 1 0> directions on {1 1 1) planes. a disordered specimen exhibited the four above stages. respectively. and Pt 3 Fe cases. Labbe and Friedel (1966) have developed a model of A15 compounds in which the onedimensionality of transition-metal atoms gives rise to . and this was correlated with the large ideal resistivity of these materials. Lines of Cu atoms can therefore serve as focusons to dissipate the energy given to a primary knock-on atom. and break-up of the vacancy clusters occurs at 400-500 K. For both 1. In the Cu3Au superlattice (Ll 2 structure) as shown in Figure 2. self-interstitial atoms were found to be mobile above HOK and vacancies to be mobile above 330K.0 and 1. V migration. long-range defect migration resulted in increases of local order. Damage production by electron irradiation and its subsequent annealing have been investigated in solid solutions of NiAl and Ni(Al. Positron annihilation experiments by Doyama et al. The critical temperature Tc of Al5 (cP8) compounds like Nb3Ge or V3Si is very sensitive to lattice disorder and decreases under irradiation by neutrons. In the Ni3Al intermetallic compounds the Frenkel-pair resistivities ranged from 97 to 109 ^G cm/%.5-9. Ni3Fe. interstitial cluster growth at . 1972) show such a mechanism operating.Ti) as well as (Ll 2 ) Ni3Al intermetallic compounds by Dimitrov et al. Frenkel-pair resistivities deduced from the comparative analysis of the initial damage rates in the solid solutions and in pure nickel were 6. Ni3Mn. and vacancy migration above 300K in both alloys. respectively.. Schematic of the ordered Cu3Au lattice (Ll 2 (cP4)structure). The assignment of the other stages shown in Figure 3 was as follows: recovery of close vacancy-interstitial pairs at —40 K in Cu3Au and CuAu. the ordered Cu3Au exhibited only three stages at 75. V migration. These were assigned to recombination of an interstitial gold atom IAu with a copper vacancy VCu close pair. Post-irradiation annealing studies in the ordered compounds showed the same recovery stages regardless of the energy of the incident electrons (see Figure 3). The mobility of selfinterstitials (corresponding mainly to Ni-Ni dumbbells) occurred at a higher temperature (75 K) than in Ni (50 K) and did not induce significant ordering.66 and 2. It was proposed that focusons dissipate the energy in the ordered specimens such that Au interstitials cannot form (a focuson is a focused energy packet in which energy is only transferred along a particular direction). a particles.4 /iQ cm/ 0Zo9 which are similar to that for pure nickel.1 4 0 K in Cu3Au and . (1986) using 0. plus ones at 55 and 190K that were attributed to IAu-VAu close-pair recombination and IAu migration. and 410K. ICu migration. Following irradiation with 1. 240. 330.5 MeV ion irradiations. (1992a. It was argued that a single type of interstitial (Cu) was migrating freely at 70K in both Ll2-ordered Cu3Au and Ll 0 -ordered CuAu.5MeV electrons. 130.1 3 0 K in CuAu. (1985) indicate that in ordered Cu3Au vacancies migrate and form three-dimensional clusters at 260-400 K. heavy ions or electrons. It was found that the minimum displacement threshold energy is ^j u m i n -18eV in all Figure 2. were slightly less mobile than in Ni and promoted some increase of long-range order. and trapped V migration.detected. Proposed <1 00> Cu-Cu split interstitial configurations and their migrations are depicted in Figure 4. and 410K that were assigned to ICu migration. lattice and the V (or Nb) atoms are arranged in linear chains along the <10 0> directions. In the above Al5 compounds the Si (or Ge) atoms constitute a b. The recovery spectrum of the irradiated Ni3Al compounds was similar to Ni but with a smaller amount of fine structure.c. Ni3Al. In the solid solutions.

36 MeV (—).% for Si and V respectively.2±O. . but have different scales (From Alamo et aL. the minimum energy that has to be transferred to an atom to produce a point defect) for specific atoms in the lattice and can also be used to obtain specific Frenkel-pair resistivities. In general. For example. Combining measurements of electrical resistivity p and critical temperature Tc can provide information on the displacement threshold energies Ed (i. leading to an increase in p. essentially no change of Tc should occur when disorder is created in the b.e. but Tc was only affected when V atoms were displaced at .c. at Ee~0.5^Q cm/at. the most efficient mechanism for producing antisite defects under electron irradiation is the replacement-collision sequence.(MAp1) [d(Ap)/dT](%/K) T(K) (1/Ap. and specific Frenkelpair resistivities were /0pP = 4. Rullier-Albenque and Senateur (1986) have measured p and Tc in V3Si irradiated with electrons at —21K and analyzed their results within the framework of the linear chain model of Labbe and Friedel (1966). leaving the transition-metal rows unaltered. Values of Ed were 25 ± 1 eV for both Si and V.)[d(Ap)/dr](%/K) T(K) Figure 3. sublattice.66 MeV (—) and 2. % and PFP = 27 ± 3 fiQ cm/at.. a condition favorable to a high Tc.c. In the A15 structure. On the basis of this model. 1986) a high density of electronic states near the Fermi level. the left and right vertical axes are for (1/Ap1)Id(Ap)ZdT]. Recovery stages in (a) ordered Cu3Au and (b) ordered CuAu following electron irradiation at 20 K.20 MeV. The production of antisite defects occurred at Zs > 0. For both graphs.E=OJSMeV.50 MeV. The results indicated that Si atoms were displaced at an electron energy £ = 0.

electron irradiations were employed and isolated Frenkel pairs were produced.Cu3Au Cu Au Cu3Au (b) CuAu (c) Figure 4. nd. The maximum energy ETmax that can be transferred to a lattice atom of mass M2 by an incident particle of mass M1 and energy E0 is given by ^T. Proposed <1 0 0> Cu-Cu split interstitial configurations before and after migration in ordered Cu3Au and CuAu (From Alamo et al.. In all of the investigations discussed above. For bombardment with ions and neutrons. 1969): nd = 0.Sv(E)/2Ed (3) . a more complex damage cascade is produced. 1986) the diatomic rows along which atom displacements can produce antisite defects are the <1 2 0> rows (see Figure 5). usually with a vacancy-rich core region and interstitial atoms (either single or in small clusters) in the peripheral regions. Long-range collision sequences are unlikely in these directions because of the 'void' between the V-Si-V atomic sequences.max = 4M1M2E0Z(M1 + M 2 ) 2 (2) The number of displacements per cascade. is given approximately by (Sigmund.

0 MeV an interstitial atom that is constrained to move in one neutrons. Only when bombardment. and (b) the atomic examples are given below. and for Dl a (tllO) Ni4Mo peaks centered at —37. as shown in Figure 6. The distance for a fhience of ~1. while uncorrelated long-range migration was detected above 30 K. The quantity DPA is given by to the loss of long-range order under irradiation when their mobility is low. for zirconium irradiated by 1. due to recombination events the number site was the second lowest-energy configuration. the Ni interstitial will tend to be quite of bombarding projectiles. 1986) Baskes (1984) and Foiles and Daw (1987). an atom V1 (or Si) has to go through two Caro and Pedraza (1991) have used the embeddedbarriers ((1) and ( I ' ) ) and an atom V2 has to pass the barrier atom potentials for Ni and Al. direction). Schematic sketch illustrating (a) the (O O 1) plane of detailed theoretical information is now emerging on the Al5 structure of V3Si and the <1 2 0> direction along which point defects in intermetallic compounds.Information on the long-range migration of selfinterstitial atoms (SIA) produced in collision cascades in Ni4Mo and Pt3Co was obtained by Aidelberg and Seidman (1987) using the field-ion microscope (FIM) technique. where v(E) is the portion of the energy that goes into and of various possible interstitial configurations in the nuclear collisions (remainder gives rise to electronic ordered B2 (cP2) alloy NiAl. that. Several atom replacements can produce disorder. an adjacent site becomes vacant will the extra Ni atom for E> 1.2 Theoretical modeling With the establishment of suitable potentials for interatomic compounds (see Chapter 23 by de Novion in this volume) and large memory and fast computers Atom Si (particularly for molecular-dynamics calculations). The <1 1 1 crowdion > excitations) and Ed is the threshold displacement energy incorporating an extra Ni atom was found to be the required to displace an atom from its normal lattice site. It should be noted.e. either when they recombine with D P A = ndos<t>t (4) nearby vacancies that are not of their own sublattice or when they occupy a Ni site upon driving the Ni atom where as is the scattering cross-section. the lowestnumber. and t is the time of stable since all its neighbors are Al atoms. If an extra Al atom is incorporated. and the arrival of the SIA was detected when it reached the surface.04. of point defects that actually survive in the initial although it contains an Al atom. fluences where considerable overlap of the displacement The displaced Al atoms may be the main contributors cascades occurs. it still contains an extra cascade event is considerably less than the calculated Ni atom. particularly at high (that the Al atom moves into). In order to make a replacement. v(E) ~ 0. more Figure 5.5 K were attributed to long-range migration of interstitial atoms. <t> is the flux into a crowdion configuration. 1 DPA corresponds to move in and create an antisite defect. developed by Daw and (2) (From Rullier-Albenque and Senateur. to calculate the energy of formation of antisite defects. configuration around the <1 2 0> direction. The specimens were irradiated in situ with either 30 keV Kr ions or 30-60 keV Xe ions to fluences of (2-1O)XlO12 ions/cm2. i. For neutron irradiation of many metals. The surface of the FIM specimen served as a sink for the SIA. as in Figure 6. It is also convenient to consider the number of energy stable configuration is a <1 1 1> crowdion that times each atom has been displaced during irradiation.0x 1025 neutrons/m2. nd = 615.5 and -77. In the case of Ll 2 (cP4) Dl a (tllO) Pt3Co a broad peak centered at -37. 2. In the ordered alloy. No long-range migration of SIA was observed between 6 and 30 K.0 MeV and £ d ~25 eV. incorporates an extra Ni atom plus a Ni antisite defect or displacements per atom (DPA). The Ni-Al <1 1 O) mixed dumbbell at an Al however.5 K was attributed to long-range SIA migration. of vacancies. Atom V . lowest-energy interstitial configuration (a crowdion is As an example.

six possible dumbbell configurations. Open circles show initial > positions.5 ps). and <1 1 1> directions.e. interstitial trapping occurred in the vicinity of a vacancy and thereby enhanced the survival of Frenkel pairs. Besides two types of vacancies and two types of antisite defects. The Al atom at A moves into site B. have been computed in ordered Ll 2 (cP4) Ni3Al by Caro et al. displacing the Ni atom at D. The number of Frenkel Figure 7. several types of interstitials are possible (i. (1990). The most mobile dumbbell appears to be the Ni-Ni(B). Threshold displacement energies for atomic displacement along <1 1 0>.. 1992) .Figure 6. i. which is a Ni sublattice site. and four proper crowdions). In the Al sublattice.. which has eight Ni atoms as nearest neighbors regardless of its orientation along any of the three major crystallographic axes. . a Ni atom can belong to either an A or B plane. the Ni(B) had the lowest enthalpy. It was argued that this may have important implications for irradiation-induced amorphization of intermetallic compounds. Vi). Molecular-dynamics calculations provide valuable insight into the evolution with time of defect structures created in the collision cascade. Among the octahedral interstitials. for example. the molecular-dynamics simulations of low-energy displacement cascades in the Bll-ordered compound CuTi (Figure 7) by Zhu et al. 1991) spontaneous recombination of Frenkel pairs was found to be third-nearest neighbor provided the chemical order was maintained. Figure 8 shows the number of Frenkel pairs produced by a Cu primary knock-on atom (PKA) as a function of recoil energy at the end of the collisional phase (0. however. four octahedrals. where A is a plane containing 50% Ni and 50% Al and the B plane contains 100% Ni. since each Frenkel pair makes a large contribution to the total enthalpy of the crystalline phase. even though antisite defects were created. an extra Al or Ni atom generates a dumbbell. Vi. . filled circles final ones (From Caro and Pedraza. whereas an Al interstitial is immobile since it converts into an antisite defect. The results also predict crowdion formation at the end of replacement-collision sequences along <1 1 0> for A planes as well as the usual <100> dumbbell for B planes. and formation enthalpies of several symmetrical interstitial atom configurations. The Bl 1 (tP4) structure of ordered CuTi (From Zhu et al. depending on the direction of stacking. In disordered alloys. Schematic for defects in the (B2) NiAl lattice showing the relaxed atomic positions attained when an Al atom is added at site A (14. and causing the Al atoms at E and F to be displaced. The Ni atom initially at B moves into position C. <1 0 0>. The stacking of the Ll 2 structure (Figure 2) along the <100> direction is ABAB .2 ps) and at the end of the cooling phase (2. For the Ni sublattice. (1992). whereas the A l Al(A) dumbbell had the highest enthalpy. also using embedded-atom potentials developed by Daw and Baskes (1984) and Foiles and Daw (1987). in ordered alloys the recombination of close pairs was favored.e. The dumbbell interstitial with the lowest enthalpy was found to be Ni-Ni(B). The simulations also showed that the spontaneous recombination volume for some defects in Ni3Al can be much smaller than in pure metals. generating an antisite defect. Consider. A crowdion with an extra Ni atom along a <1 1 1 direction is produced. as indicated.

the threshold energy with Ti as a PKA was greater than that with Cu as a PKA by a factor of 1.2 ps (o) and 2. using embedded-atom potentials. All of the interstitial configurations contained antisite defects and probably have large migration energies. 3. The energies and configurations of interstitials and vacancies in the B2 (cP2) ordered compounds NiTi and FeTi have been calculated by Lutton et al. On average. Consequently the interstitial atoms in these compounds may be much less mobile than those in pure metals.. The Fe interstitial atom in FeTi was of the Fe-Fe <1 1 1 > dumbbell type. this behavior was not observed in <1 1 0> and <1 1 2>. An increase in the degree of order in a disordered or partially ordered alloy corresponds to an increase in the . whereas for complete disorder F^ = XA and P*=XB and 5 = 0.e. In NiTi. Ed was small in low-index directions).Number of Frenkel pairs Recoil energy (eV) Figure 8. whereas if the PKA was a Cu atom. P* = pP=\ and S = I . For complete order. removal of a Ti atom led to the creation of a vacant Ni or Fe site and an adjacent antisite defect. It can be seen that there is an appreciable reduction in the number of Frenkel pairs during the cooling phase. about half the number of replacement atoms that had been produced in the collisional phase returned to their original lattice sites in the cooling phase. In both compounds.e. and the Ti interstitial formed a <1 1 0> Fe-Fe dumbbell. 1934) of long-range atomic order S in a binary alloy is given as follows: S = (P^ -XA)/(\-XA) = (P»-Xh)/(\ -XB) (5) where the parameters P^ and P^ are the probabilities of an A atom residing on an a site and a B atom residing on a £ site. see Shoemaker et al. the stable configuration of a vacancy after the removal of a Ni atom was a vacant Ni site. Irradiation-Enhanced Ordering and IrradiationInduced Disordering of Ordered Alloys 3.1 Irradiation-Enhanced Ordering The Bragg-Williams description (Bragg and Williams. Average number of Frenkel pairs produced by a Cu PKA as a function of recoil energy at 0. For example. and in FeTi the removal of an Fe atom resulted in a vacant Fe site.5 ps (•).67. Interstitial atoms in NiTi formed split <1 1 1 configurations consisting of a Ni> Ni dumbbell with one or two adjacent antisite defects. It was also found that when the PKA was a Ti atom. The Kinchin-Pease function (equation (3)) for E6= 15 eV is shown for comparison (From Zhu et al. The effective threshold energy calculated by averaging over all the directions and both Cu and Ti PKAs was 62 eV. and XA and XB are the atomic fractions of the respective species. respectively. (1991). At the end of the cooling phase. half of the replacements are antisite defects. (1991) using atomistic simulation. For detailed information on the computation of the energies and configurations of vacancies and interstitial atoms in CuTi and CuTi2. the directional dependence of Ed was similar to that in pure metals (i. the modified Kinchin-Pease function) is also shown for Ed= 15 eV. considering both Cu and Ti PKAs and a recoil energy of 500 eV. 1992) defects predicted by equation (3) (i.

At all irradiation temperatures a steady-state degree of order was developed. and Dl 2 structures. irradiation-enhanced ordering and irradiation-induced disordering. as can also be seen in Figure 11. S = (Is/I{)p(Is/If)*=l (6) where /s and /f are the intensities of the super lattice and fundamental reflections. and Banerjee and Urban's model has been developed for alloys with B2.. The most comprehensive models of irradiationinduced ordering and disordering. Butler (1979). the degree of order was obtained from the relation. In Butler's model.number of correct atomic bonds. Zee and Wilkes (1980). The irradiation-induced disordering is expressed by Zee and Wilkes (1980) as follows: Time of irradiation (min) Figure 9. and Banerjee and Urban (1984). Variation of the degree of long-range order S for initially ordered Cu3Au for various irradiation temperatures as a function of time of irradiation with 600 keV electrons (From Hameed et al. In the above models. to an increase in the probability that an A atom occupies an a site and that a B atom occupies a 0 site. Ll 2 . the study of ordering and disordering in Cu3Au during 600 keV electron irradiation (in a high-voltage electron microscope (HVEM)) by Hameed et al. The phenomenon can be suppressed if the temperature of irradiation is sufficiently low and accelerated if the temperature is raised above ambient. are those due to Liou and Wilkes (1979). It also occurs under irradiation conditions where isolated Frenkel pairs are created (electron irradiation) as well as when there are more complex displacement cascades (heavy-ion and neutron irradiation). i. 1979) using a variety of experimental techniques that this increase in S during irradiation involves the thermally activated migration of irradiation-produced point defects. only AB alloys are considered. The rate of ordering and disordering was determined between 140 and 400 K by measuring the ratio of the intensities of the superlattice and fundamental reflections in electron diffraction patterns. which contains measurements obtained from both initially ordered and initially disordered specimens.e.e. Figure 9 and 10 show the way S varies for initially ordered (Ll2) Cu3Au and initially disordered Cu3Au samples during electron irradiation at various temperatures. for example. respectively. Specifically. (1982). i. which have been used to analyze data of the type shown in Figures 9-11. Liou and Wilkes. and Zee and Wilkes consider AB3 alloys (Cu3Au in particular). In general. 1982) . it is assumed that interstitial atoms do not contribute significantly to the ordering and that disordering takes place athermally. It has been demonstrated in numerous experiments (see the review by Schulson. in these models two competitive processes are occurring simultaneously. Consider. The final state of the alloy at any irradiation temperature is independent of the initial condition.

S Time of irradiation (min) Figure 10.S)2 .E0JknT) ] {XAXB(\ .t 1982) (dS/d/)irr=-^5 (7) (dS/dOor£l = [ (Z01 + Z^ . XA and XE are the fractions of A and B atoms. whereas for fast-neutron irradiation Zee and Wilkes (1980) use e = 80. Cv is the steady-state Se Temperature of irradiation (K) Figure 11. vs is the vacancy jump attempt frequency. Variation of the steady-state degree of order Se developed during irradiation at various temperatures with 600 keV electrons at a displacement rate of ~ 10~3 DPA/s (From Hameed et aLy 1982) .V0SZk^T)[S +X^Xn(I -S)2]} (8) where e is the number of replacements per displacement. and S is the instantaneous degree of order. Variation of the degree of long-range order S for initially disordered Cu3Au for various irradiation temperatures as a function of time of irradiation with 600 keV electrons (From Hameed et al.e x p ( .2)(^V/2)(Z/^B)CV exp( . Under electron irradiation and near threshold e~ 1. To describe the rate of change of order due to thermally activated processes during irradiation. Zee and Wilkes (1980) use the equation: Where Za is the number of A atoms around /3 sites. Z^ is the number of B atoms around a sites. K is the displacement rate.

Specifically. P is the atom displacement rate (DPA/s). S is the long-range order parameter.60 eV is considerably less than that expected for vacancy migration in Cu3Au. % Zn (Halbwachs and Beretz. and (dS/dt\ is the thermal ordering rate. E?m is the ordering jump energy. there are some ordered alloys in which the migration energy for interstitial migration may be quite high.2. along with corresponding experimental investigations. T is the absolute temperature. (dS/d0r is the total disordering In the above equations. C A . in order to apply models of the above type to study the irradiation behavior of ordered compounds in general.84 eV values used by Zee and Wilkes (1980). as pertinent data for Ni4Mo were not available. It is obvious that. and may be even higher than that for vacancies (i. The parameters used in fitting the theoretical model to the experimental data were ism = 0.8 and 0. respectively. In the model by Banerjee and Urban (1984). S is the degree of long-range order. e = 80 for fast neutrons and 20 for thermal neutrons. the net ordering rate under irradiation is given by dS/dt = (dS/d0c + (dS/dOr + (dS/d0t (H) where (dS/d0c is the rate of disordering due to collision sequences. In the analysis of the Cu3Au and Ni4Mo results described above. me is the effective number of pair exchanges produced in a collision sequence. as shown in Section 2. (1982) applied the model of Zee and Wilkes (1980) to their experimental data. detailed information is required on the configurations and migration energies of the vacancies and interstitial atoms. kB is Boltzmann's constant. A value of Em = 0. it was assumed that only the migration of vacancies is contributing to the ordering. were used in place of the 0. and a survival rate for vacancy production of 5% for fast neutrons and 100% for thermal neutrons.Xm6PSZCACB (dS/dt)T=-PS (12) (13) (dS/d0t = K+ Cfcl . the reversal of that normally found in most elemental metals).2. Computations of the type described in Section 2. and CA. and F0= V/S (V is the ordering energy). E?m for ordering vacancy jumps.62 eV. the role of the nearby surfaces acting as sinks for the migrating defects (which might affect the above analysis) still has to be assessed.K_ ClCi The steady state of order Se is developed at any temperature when (dS/d0irr + (dS/d0ord = 0 (10) (14) Zee and Wilkes (1980) applied their model to data that had been obtained by various investigators on fastand thermal-neutron irradiations of Cu3Au. shown in Figures 9-11. and claimed that there was reasonably good agreement between their model and the experimental results. \ is a geometrical factor. The calculated temperature dependence of the steady-state degree of long-range order was in good agreement with the experimental observations.8eV for vacancy motion. However. In the case of electron irradiations in a HVEM. Hameed et al. This effect was attributed to the marked slowing-down of the selfinterstitials by the undersized solute atoms and may also . The above model was applied by Banerjee and Urban (1984) to analyze the results obtained by Banerjee et al. are certainly required. and C^ refer to the concentrations of either A or B atoms on a or /3 lattice sites.e. CA and CB are the concentrations of A and B atoms in the alloy. There is also evidence for this reversal from studies of mechanical relaxation effects in the irradiated substitutional alloy Ag-24 at. The actual ordering rate dS/dt is a balance between the irradiationinduced disordering rate (dS/dOirr an<3 the irradiationenhanced ordering rate (dS/dt)ord. (1984) on the electron irradiation of Ni4Mo (Dl a (tllO) structure). respectively (these are related to the concentration of vacancies on the a and /3 sublattices of the ordered alloy). on the rates of ordering and disordering in electron-irradiated Cu3Au.vacancy concentration. C^. Reasonable agreement with these experimental data was obtained if minor modifications were made to the theory. Thus: (dS/dt) = (dS/dt)iTT + (dS/dOord (9) rate by random defect annihilation.6 and 0. which may be quite a reasonable assumption for those particular alloys for the temperature regime used for the analysis. and may indicate that point-defect mobility is considerably increased during irradiation due to irradiation-enhanced diffusion or by the formation of divacancies. particularly for irradiation times beyond the initial transient stage. 1980). K+ and K_ are rate coefficients for ordering and disordering. It also may be true for many other ordered alloys. Values of Em and EPn of 0. (dS/d0c = . although the values used for the diffusion behavior as well as the displacement threshold energies were those for Ni.

3. -0. Various mechanisms have been proposed for the rearrangement of atoms from their correct positions in an ordered lattice to a random distribution of the atoms as irradiation proceeds. LRO persisted for samples initially in this state and SRO was only preserved up to 550 K for samples initially in that state. Nb3Sn. Nb3Ge. disordering) and lattice dilations arising from the presence of point defects and defect clusters. This behavior is characteristic of an ordering transition of the first kind below and above the coherent instability temperature.1-0.25% in these materials (see tabulation by Schulson. at temperatures below which irradiation-enhanced ordering can occur. and which have the form a. where S0 is the initial degree of long-range order. In investigations on the Ll 2 compounds Ni3Mn (Aronin. For a general review of some of these aspects. 1979). In general. Nb3Pt. Ni3Mn. A more detailed comparison. and Mo3Os. collapse of cascades to vacancy loops by Jenkins and Wilkens (1976).1.t 1978b) as well as Nb3Al (Sweedler and Cox. whereas in Zr3Al. the degree of longrange order S decreased exponentially with the irradiation fluence.3 DPA was required with heavy-ion and neutron irradiation to reach S ~ 0. (1984) on the effect of electron irradiation on the orderdisorder transformation in (Dl3) Ni4Mo. V3Ga. In general. as disordering occurs in an ordered alloy. the alloy completely disorders. and at temperatures above 720K a complete transition to SRO was obtained. In Zr3Al and Ni3Al. Cu3Au. and Nb3Al. see Schulson (1979). It is believed that maintenance of LRO requires a vacancy mechanism. the A15 compounds Nb3Sn and Nb3Ge (Brown et al. 1975). is consistent with the concentration-wave description of the SRO structure and supports the concept of spinodal ordering. At temperatures below 800 K the SRO-LRO transition occurred in a continuous fashion. which occurs via the completely disordered state.( P ' A B / C B ) = 1 . Hence under irradiation. Pt 3 Fe. Ni3Al. Banerjee and Urban .that describe the conditional probability PAB ( P ^ ) of finding B (A) atoms on the successive coordination shells around A (B) atoms. consider the study by Banerjee et al. ions. 1963). only SRO was observed. 1979). plastic spikes by Seitz and Koehler (1956). At 200-450 K. and thermal and fast neutrons. and <f>t is the irradiation fluence. and the transition between LRO and SRO. 1954). .1 (see Schulson. Irradiation also increased the lattice parameter a. however. lowering of the transition temperature for superconductivity as well as the magnetic saturation in ferromagnetic materials and decreased intensity in the diffraction intensity in superlattice reflections.= 1 . The disordering occurs for bombardments with electrons. Many of these increases were considerably larger than those observed in intermetallic compounds that can be disordered thermally in the absence of irradiation. while above 800 K a nucleation and growth mechanism was operative. complete disordering occurs at lower calculated DPA levels (see equation (4)) for cascade-producing irradiations (ions and fast-neutron irradiation) than for irradiations where only Frenkel pairs are produced (electrons and thermal neutrons). 1984).2%.1-0.occur in other solid solutions where a marked size difference exists between the constituent elements. the Dl a compound Ni4Mo (Banerjee et aL. Some examples will now be given of experiments that have been undertaken in an attempt to elucidate some of the mechanisms of irradiation-induced disordering. 1949). The state of SRO is usually given by the Cowley-Warren parameters a. Nb3Al.1 to 1. by amounts ranging from 0.2 Irradiation-Induced Disordering From measurements on the changes in electrical resistivity. it has been shown that irradiation-induced disordering occurs in alloys that were initially ordered. Between 550 and 720 K. where Aa/a~ 0. this dependence can be expressed as S = Soe-K*< (16) As an example. a mixed SRO-LRO state occurred. It is also of interest to obtain information on changes of short-range order (SRO) during irradiation as well as of long-range order (LRO). and Pt3Fe (Piercy. Electron micrographs and diffraction patterns were obtained during in situ electron irradiations at 50-1050 K in a HVEM. Nb3Sn. These include thermal spikes by Seitz (1949). Cu3Au (Siegel.e. At temperatures below 200K. 1978). It is believed that an interstitial mechanism is responsible for maintaining the SRO. —1-3 DPA was required with electron irradiation to reduce S to —0. the observed volume changes arise from the creation of wrong bonds (i. and random recombination by vacancies and interstitials by Carpenter and Schulson (1978). Zr3Al (Carpenter and Schulson. as indicated by studies on Zr3Al. should also take into account the large amount of spontaneous recombination that occurs promptly in dense cascades. Above 450 K. K is a parameter expressing the effectiveness with which the incident particle disorders the material. replacement collisions by Kinchin and Pease (1955).( n A / C A ) (15) 1984.

It was concluded that an average of 2. According to the model of ferromagnetism in Ni3Mn by Marcinkowski and Poliak (1963).7) recoil with an average energy of 492 eV. Pt3Fe was chosen because it is normally paramagnetic above the antiferromagnetic Neel temperature of 70 K but becomes ferromagnetic when disordered. The results also indicated that there was a region of partial disorder surrounding each completely disordered region. Unit cell of ordered Ni3Mn and (1 0 0) Ni-Mn planes before and after occurrence of a replacement collision sequence RCS (From Kirk et al. was used for calculating the number of jumps per atom for the short time interval that the spike was heated. whereas the passage of a <1 1 0) RCS in a partially ordered alloy will have some probability of placing a Mn atom into a position with three nearest-neighbor Mn atoms. In one series of experiments.C. with a frequency factor of 1014S"1 and an energy of 3 eV for interchanging two atoms. It was concluded that 112-150 <1 1 0) replacements occur. From an estimation of the temperature (~ 1. before an RCS (1 0 0) plane. A normal Arrhenius equation for the atomic jump frequency.25r0. Detailed information on the role of replacement collision sequences in producing disordering was obtained by Kirk et al. per (n. The atomic structure of ordered Ni3Mn is shown in Figure 12. along with a representation of a Ni-Mn (100) plane before and after the passage of a <1 1 0> replacement-collision sequence (RCS).7) recoils. 1978) and Kirk and Blewitt (1982) from magnetic saturation experiments performed on the Ll 2 alloy Ni3Mn.C.5 x 104K) and duration (~4x 10~n s) of the thermal spikes produced during each neutron collision. after an RCS Figure 12. thus causing a spin flip and a decrease in saturation magnetization. they irradiated Ni3Mn samples F. They found 56 <1 10) replacements at a 450 eV recoil . The magnetic changes were related to the production of <1 1 0) RCSs by (n. The passage of a <1 1 0> RCS in the fully ordered alloy will not change the state of magnetic order.4x 1019 fission neutrons/cm2) at -333 K. A comparison of the experimental data with a computer analysis of the magnetic model is shown in Figure 13.97 nm. Ni3 Mn. ordered Ni Mn (1 0 0) plane.Piercy (1963) performed magnetic measurements on (Ll2) ordered Pt3Fe that had been neutron irradiated (3. The above calculation was based on an average energy of 8.3 disordered regions were produced per fission neutron and that approximately 3300 atoms were disordered on average in these regions. in which case the central Mn atom will align antiferromagnetically.t 1977) containing different degrees of order S and measured the decrease in magnetization during irradiation with thermal neutrons at 5-6 K. This is to be compared with the results of molecular-dynamics computer calculations by King and Benedek (1982). each Mn atom will align ferromagnetically with its nearestneighbor Ni atoms unless three or more nearest-neighbor Mn atoms are present. Piercy (1963) concluded that purely thermal effects could account for the disordering of about 3100 atoms contained within a radius IAlSr0 and the partially disordering of a further 1200 atoms contained within a radius of 1. within one or two sequences. Hence the size of small disordered regions ( 5 nm) in an ordered — matrix could be detected since they were ferromagnetic regions with a low magnetic anisotropy contained in a non-ferromagnetic matrix. through the unique dependence on sample order (as discussed above). (1977.4 x 1017-6. this energy being initially distributed uniformly over a radius r0 of 1.6 keV being available in a thermal spike that disordered 3300 atoms.

Their results also suggest that in the cascades produced by the fission -A4nMs (G) Fluence (1017 n / cm2) Figure 14.Nunber of spin flips per 80-atom sequence Long-range order.7). Furthermore.. The absence of long-range RCSs in even the peripheral regions of the cascade suggests that thermal agitation of the atoms within the cascade may prevent this from occurring. Kirk et al. while random replacements produce an increasing magnetization change per unit neutron fluence with increasing order. they argued that long <1 1 0> RCSs produce a decreasing magnetization change per unit neutron fluence with increasing order (S > 0.72 (A) (From Kirk et al.4%) of the replacements that were measured can be attributed to long-range <1 1 0> RCSs. they obtained dAS/d<f>t = K(S0-AS) (17) where K is now related to the number of random replacements per neutron. . (1978) and Kirk and Blewitt (1982) also studied the effect of fission neutrons on the saturation magnetization of Ni3Mn.93 (O) and S0 = 0. The number of Ni-Mn interchanges per highenergy primary knock-on (average energy 30 keV) was 3400. thus implying that very few ( .. The ratio of the slopes (KS0) of the two curves in Figure 14 was within I0Io of the ratio of the degrees of order.1 MeV) for two simultaneously irradiated samples of Ni3Mn with long-range order parameters S0 = 0. Equation (17) shows that the initial disordering rate (KS0) is an increasing and linear function of the initial degree of order. Decrease in saturation magnetization with fissionneutron fluence (is > 0. 1977) energy and much branching of sequences to yield an average maximum number of 4-5 replacements per sequence. 1978) neutron irradiations. An irradiation experiment on two samples of different degrees of order (as shown in Figure 14) was performed to try to separate the effects of the two types of disordering. Experimental results obtained for thermal neutron irradiation (A) are compared with a computer model calculation (o) (From Kirk et al. From equation (16). and there were 60 replacements per displacement (as determined experimentally). The dependence of the number of Mn atom spin flips per 80-atom sequence on the degree of long-range order in Ni3Mn. the distribution of the interstitial atoms is spatially quite compact just outside of the vacancy-rich core. S Figure 13.

. 1980. and the Al5 compound Nb3Sn. Cu3Au. which corresponded to Bv varying from 1. giving rise to zones of reduced long-range order within the ordered matrix. Winterbon. Zone axes near [0 0 1] (From Howe and Rainville.. Zr3Al. Described next are some specific results on Zr3Al. 1991) . Based upon the Monte Carlo damage simulations of Walker and Thompson (1978). for various implants in Si and Ge. (1976) and Jenkins and Wilkens (1976). and the results of the latter study will be described here.e. a value of KR = 0. 1979. 1977. The principle of the method has been described by Jenkins et al. and (113) fundamental reflections. Dark-field electron micrographs taken of the same area under two-beam dynamical conditions with a fundamental (2 2 0) reflection (a) and a superlattice (1 1 0) reflection (b). and Nb3Sn.09ev/atom (i. 1977. and VR is the volume ratio defining the fraction of the statistical cascade volume filled on average by an individual cascade.9 1970. under various irradiation conditions. Cu3Pd. Zr3Al is an Ll 2 alloy that under various irradiation conditions can be disordered and then eventually amorphized (Howe and Rainville. 1991). Using the (1 1 1). (2 0 0). The average diameter Dsc of these regions was 5.33 to 0. Figure 15. The strain-contrast features indicate that they are probably three-dimensional clusters of interstitial atoms.1 nm. Schulson et al. although the possibility of amorphous regions being produced directly in the cascade cannot be ruled out. Within individual cascade regions the long-range order is disrupted. 1979. the damaged regions exhibited black-white features (see Figure 15(a)) that indicate that these regions have a spherically symmetrical strain field. Centres of dilation would be consistent with the 5% volume increase observed (Schulson et al. Ni3Si. 1979). It is also of interest to describe these results in the context of the average deposited-energy density within a collision cascade.1-8. Ion bombardments were performed at 295 K with 15-120 keV ions.The damage produced within a reasonably dense collision cascade in ordered alloys can be observed directly in the electron microscope. and Ni3Mn.I x I O 1 1 ions/cm 2 ) bombardments with Ar ions (Howe and Rainville. Damaged regions produced in Zr3Al by 100 keV Cu ions. 7VV is the number of lattice atoms contained within a spheroid whose axes are determined by the longitudinal <AA^>1/2 and transverse (Y2)1'2 straggling components of the statistical damage distribution (Winterbon et al.2p(E)/NvVR (18) where v(E is the portion of the monatomic ion energy that is lost in nuclear collision events. the DO3 compounds Fe3Al and Fe3Si. Bv decreases with increasing ion energy). 1980) as well as Cu ions (Howe and Rainville. 1975).18 was chosen by Howe and Rainville (1991) for 63 Cu ion implants in Zr3Al. (2 2 0). Bv is defined as follows for monatomic implants: dv = 0. This technique has been used for studying the disordering of the Ll 2 compounds Cu3Au. The nature of the damaged regions formed in individual collision cascades has been investigated using low-fluence ( .

They represent regions within the damage cascade in which sufficient disorder of the ordered lattice has occurred to give rise to a visible region using a superlattice reflection. Their average diameters Dd were in the range 2. the ratios Dsc/2(Y2y/2 (strain-contrast regions) and Dd/2(Y2)Vl (disordered regions) both decreased with increasing ion implantation energy. and Kr ions (50 and 100 keV) at 30 and 300K.1979) for the crystalline-to-amorphous transformation in Zr3Al. 1991) . many damaged regions were present in addition to those revealed using the fundamental reflections. 1982). It was found that all dislocation loops were associated with disordered zones. When the same areas of the foil previously imaged with fundamental reflections were imaged under dynamical conditions with (1 10) superlattice reflections. The disordered zones were imaged using superlattice reflections and dislocation loops using fundamental reflections. These additional damaged regions appeared as black spots and were considerably smaller than the strain-contrast regions.7 nm. The contrast arises from the difference in structure factor (for superlattice reflections) between the disordered zone and the ordered matrix. but which do not contain a sufficiently high concentration of vacancies or interstitials to give rise to a strain-contrast effect. as shown in Figure 15(b). but not 5/2<Y2>V2 Ion energy (keV) Figure 16. (1987) have performed a systematic transmission electron microscopy (TEM) study of the conditions leading to the formation of vacancy dislocation loops at the sites of defect cascades produced by irradiating with Ar. Cu. The number of observed regions per incident ion increased with increasing ion energy (decreasing Bv) for the defects having a spherically symmetrical strain field as well as for the disordered regions. i. At the lower ion energies (15 and 30 keV). due to a local reduction in the long-range order parameters. As shown in Figure 16. Even taking into account the increase in the number of damaged regions per incident ion with increasing ion energy. whereas at higher energies there were multiple disordered regions produced within a single cascade.4-3. there was essentially one disordered region produced in each collision cascade. the fraction of the theoretical collision cascade volume occupied by the total volume of the damaged regions still decreased appreciably with increasing ion energy (decreasing #„). Black et al.e. The collapse of defect cascades to dislocation loops in (Ll2) ordered Cu3Au has been studied quite extensively (Jenkins and English. Variation of DJIiY1Y1 (strain-contrast regions) (o) and D^/liY2}* (disordered regions) (•) with energy of implantation for 63Cu ions in Zr3Al (From Howe and Rainville.

Irradiation-Induced Amorphization 4.Next Page vice versa. with perfect order being recovered within shells of 0. (1993) found evidence for vacancy clustering in the core of the displacement cascade. suggests that directed motion of vacancies toward the cluster nucleus occurs during the collapse process.5 and 1.5 x 1013 Kr ions/cm2. the martensitic structure in Nb3Sn was observed at 12-30 K and then images of the irradiation-produced defects were obtained using either fundamental or superlattice reflections. 4. along with the high values of defect yield at 30K. The spatial and temporal balance between irradiation damage and annealing can . whereas the fast quenching occurring in the core of the cascade did not allow for a comparable loss of the chemical SRO. 1981). It was postulated that the centers of strain can be associated with pinning sites of magnetic flux and the corresponding increase in Jc. Kirk et al. there appeared to be evidence for the onset of amorphization.5 nm for Ni3Al and Cu3Au. 1993). Experiments by Brown and Blewitt (1979) and Guinan et al. the characteristic contrast of the martensitic structure had disappeared. the collapse is occurring during the phase of collision cascades when extensive vacancy motion is possible. Upon irradiating to 3. Diaz de Ia Rubia et al. At higher ion doses of either Xe or Kr ions. Following the 30K irradiation. Displacement cascades were observed to collapse with a significantly higher probability at 300 K than at 30 K. The transition between the disordered region produced by the cascade and the surrounding ordered region was sharper for Ni3Al than for Cu3Au. Support for this view also comes from experiments on the irradiation-produced change of Jc in pure and alloyed Nb3Sn conductors and their comparison with the alloying effect in unirradiated samples (Suenaga. (1987) in their TEM studies of Cu3Au bombarded with 50 keV Cu ions at 30 K. overlap of the collision cascades produced regions of stronger disorder and possibly even amorphous zones. By a dose of 5 x 1012 Kr ions/cm2. i. as it entails the complete loss of measurable crystalline order. Since vacancies do not undergo long-range thermal migration at 30 K. This observation.g. in the thermal spike phase that occurs within about 10~n s of cascade initiation (Guinan and Kinney. Within a thermal spike model. However. It is an extreme response of the solid to a combination of a large external input of energy and the presence of kinetic constraints that preclude the formation or the maintenance of the stable intermetallic phase. The disordered or amorphous zones may be associated with decreases in Jc and the critical temperature Tci as observed under neutron irradiation at comparable damage levels.5 MeV Kr-ion irradiations on Nb3Sn at 13-15 K and observed the damage evolution in a high-voltage electron microscope (HVEM). but the basic mechanism responsible for these irradiation-induced changes (e. Irradiation-induced disordering and defect production in Cu3Au and Ni3Al (both Ll 2 structures) were studied by molecular-dynamics simulation for a 5 keV recoil atom at 10K (Diaz de Ia Rubia et al. thus suggesting that a revision to the pre-transformation cubic structure had occurred due to the strain associated with the defects produced by the irradiation. For all of the bombardments. antisite disorder). The effects of fast-neutron irradiation on some bulk properties of Al5 superconductors are fairly wellestablished.. an increase in ambient temperature might be expected qualitatively to decrease the cooling rate within the cascade region and therefore to increase the thermal spike lifetime. The crystalline order parameter decreased to . no additional cascade collapse was observed during warming to 300 K.e. there is also considerable evidence that the enhancement of Jc is related to the irradiation-induced disorder.0x 1011 Xe ions/cm2 visible strain centres were observed that appeared to be regions of high concentration of point defects but which contained relatively little atomic disorder (i. (1984) suggest that defects produced by collision cascades in Al5 compounds are capable of pinning magnetic flux lines and thus initially increasing the critical current densities during ion or neutron irradiation at low fluences. They observed that the loss of the underlying crystalline structure and the loss of the chemical short-range order were controlled by different timescales.e. (1990) performed 50 keV Xe-ion and 1. At ion doses of 8. 1986). The kinetic path taken by the material to arrive at this large departure from equilibrium depends not only on the characteristics of the material but on the irradiation conditions as well. 1985). collapse of the displacement cascades to dislocation loops occurred with high probability at temperatures as low as 30 K. and the effect was more pronounced the higher the mass of the incident ion. respectively.0 within 1 ps.1 Experimental Observations The irradiation-induced amorphization of intermetallic compounds is of scientific interest and technological importance. but there were no signs of interstitial cluster production in the intracascade regions. critical current density Jc) are not well-understood (Weber. The size of these disordered regions was consistent with those found by Black et al. Initially.

c. 2000 John Wiley & Sons Ltd . scheme for f. most notably the Hume-Rothery electron compounds as found. brief mention is made of optical applications in connection with phase transformations. it is considered in this chapter because of its intimate relationships with crystallography. in 0-brass of near-equiatomic composition. University of Washington. austenite to transform into b. Martens) was the name given by Osmond in 1895 to the microstructure observed in quenched steels. P.Chapter 9 Crystallographic Transformations C. University of Illinois. A brief survey of shape-memory phenomena within the framework of martensite crystallography is presented. however. Twinning is a shear process without phase transformation. the prototype of which occurs in quenched steels.t. these transformations are also considered. alloys. Crystallography 2. Westbrook and R. ceramics. Edited by J. martensite (Figure 1). now known to be associated with many metals. Some mechanical aspects of martensitic transformations have been considered for their insight into stress-induced transformations. martensite (in honor of Professor A. Martensitic transformations and shape-memory phenomena are common to many intermediate phases of the /3-brass type. The Bain distortion relates corresponding unit cells in the two structures and specifies the 'upsetting' strain. and Au. Illinois 61801. but in more modern times the word 'martensite' designates a transformation mechanism.1 The Lattice Correspondence and Bain Deformation A great scientific and atomistic step was taken by Bain (1924) when he proposed a rather simple. L. Scope By 'crystallographic transformations* one generally means that. Washington 98195. USA Han R. These transformations occur in a variety of intermetallies. 2. Fleischer ©1995. Actually. Seattle. This chapter begins with a general consideration of the crystallographic features of martensitic transformations. and many similar alloys of Cu. even though massive transformations are non-crystallographic. which also undergo massive and bainitic transformations. Inoue Department of Materials Science and Engineering. The most widely studied crystallographic transformation is the martensitic transformation.c.c. when a new phase (product) forms from the old (parent). by means of which martensite is formed. Urbana. and even some polymers. for example. albeit elegant. this subject and the various martensite crystal structures are presented in detail in Chapter 26 by Schetky in Volume 2. The bainite transformation is highly controversial despite its martensitic crystallographic features. Finally. Ag. Basic Mechanical Properties and Lattice Defects of Inter metallic Compounds. USA 1. H. Marvin Wayman Department of Materials Science and Engineering. and thus detailed descriptions of the crystal structures and substructures for individual alloy systems such as Ni-Al versus Cu-Sn are avoided. This picture is too simple in itself but has remained an integral part of the more sophisticated theories that have ensued. it bears certain definite geometrical relationships. The principles are general.

3 The Phenomenological Theory of Martensite Crystallography (PTMC) Modern crystallographic theories suppose that. so we will proceed to discuss these BM and WLR theories.c. Modern crystallographic theories of martensite formation such as those of Bowles and Mackenzie (1954) (BM) and Wechsler et al (1953) (WLR) rectify these apparent inconsistencies. In an IPS the displacement of any point is in a common direction and is proportional to its distance from the undistorted and unrotated invariant plane. 1951). 2. Such a distortion resembles a shear but also includes a volume-change component. which from Figure 1 is (001) M ll(001)p [001 ]MII [001 ]p [010] M ll[110] P etc. however. •Subscripts P and M designate the parent phase and martensite respectively. Nishiyama. the rotation taken aside. instead. Careful analysis of the tilted martensite platelets shows. Actually this simple orientation relationship is not observed.2 The Shape Change Y-Fe a-Fe It has been known for some time that the formation of martensite causes roughening (in Bain's words) of a prepolished surface.c.t.c. as known from X-ray diffraction (Kurdjumov and Sachs. the Bain distortion precisely accounts for the f. no temporal sequence is implied by these operations). unit cells is both compressed and 'rigidly' rotated during the course of the austenite-martensite transformations.c. the above planes and directions are not exactly parallel. 2. the measured IPS is inconsistent with the correct structural change as given by the Bain strain.t. lattice of iron produced by the upsetting of the former (After Bain. 1924) Note that the Bain strain implies a lattice orientation*.c. as shown in Figure 2.prism delineated in the two f. which are fundamentally identical but differ in mathematical order. that an invariant plane strain (IPS) has occurred (Bowles. the * upsetting' produced by the Bain strain is not an IPS. as shown in Figure 2. 1934).c. austenite structure and (b) the b. to b. Bain's model for the formation of martensite in steels showing (a) a b. Clearly.t. We now face a dilemma: in the case of iron alloys or steels (or in general).c. the f. or in matrix form P = SRB or equivalently P = RBS and P S 1 = RB Figure 1.c. 1930. the invariant plane is the habit plane of the martensite plates. . a simple shear (S) occurs simultaneously (actually.c.c. In other words. These three operations taken together are equivalent to the IPS shape deformation (P). structural change. on the other hand.c. in addition to the Bain strain (the 3x3 matrix B) and rotation (R) alluded to earlier. lattice delineated in the f. Incidentally.

Out of this.4 What is a Martensitic Transformation? The 'shear-like'. More recently Cohen et al. 1965) characteristics of martensitic transformations are universally recognized. • Assumed plane p and direction d for S. In other words. As late as the 1950s some refused to recognize that a martensitic transformation could occur in materials other than steels (Crussard and Philibert. the habit plane. by predictable interactions with mechanical stresses.Figure 2. but few explicit definitions exist. involving cooperative atomic movements. The ILS reduces to an eigenvalue problem with input data: • Lattice parameters of the phase (hence strain magnitudes). Both phases are ordered Since S"1 is also a shear (but in the opposite direction). its combination with the IPS P produces an invariant line strain (Bowles and Mackenzie. There will. 1954). therefore. • Assumed Bain correspondence (hence the strain axes). an example of which is shown in Figure 4 for internally twinned martensite. We may. The surface upheavals caused by martensite formation are analyzed to be an invariant plane strain. The idea of a shape change and lattice (Bain) correspondence is also universal. (1979) concluded that: . and magnitude of the shape deformation P (surface tilt) are predicted. If the theory has any physical basis. lattice orientation relationship. the combined rotation and Bain strain RB are also the identical invariant line strain (ILS). and 'military' (Christian. then one would expect a fine-scale corrugation within the martensite as shown schematically in Figure 3. 2. the martensite contains a substructure consisting of adjacent domains that are either slipped or twinned relative to each other. At this point we must address the third item listed above. conclude that martensitic transformations are shear-like (displacive) in nature. 1956). the simple shear. accordingly. The basic PTMC presented above and the notion of an internally modulated martensite phase applies to all known martensitic transformations. Optical micrograph showing an Fe-24Pt alloy partially transformed to martensite. as will be seen later. That is. 'cooperative'.

In view of this difficulty. i. In other words the kinetics are vastly different although the crystallography is unchanged. it appears that an earlier description (Clark and Wayman. 1973). 1957) . Such a definition may be disputed on the grounds that it is too broad by attempting to mix both crystallography and kinetics. virtually diffusionless structural change having a dominant deviatoric component and associated shape change such that strain energy dominates the kinetics and morphology during the transformation. Transmission electron micrograph showing internal transformation twins in an Fe-Ni-C alloy martensite Figure 5. an irrational habit plane and orientation relationship. and the presence of an internal inhomogeneity such as twinning. all of these features being uniquely related in accordance with the phenomenological crystallographic theory. showing the hysteresis of the martensitic transformation on cooling and the reverse transformation on heating. a lattice correspondence between substitutional atoms which implies the absence of long-range diffusion. (b) Internally twinned martensite plate Figure 3.5:47. The alternate serrations at the habit plane average the strain out to zero on a macroscopic scale Resistance ratio Temperature (C°) Figure 4. the shape strain is the same in both cases. like the Fe-Ni alloy shown in Figure 5. yet the transformation in ordered alloys is completely thermoelastic while that in disordered alloys is non-thermoelastic. Electrical resistance changes during cooling and heating Fe-Ni (70:30) and Au-Cd (52. As a case in point.e.(a) Internally slipped martensite plate A martensitic transformation is a lattice-distortive. Schematic sketch showing (a) internally slipped and (b) internally twinned martensite. Fe-Pt alloys near the composition Fe3Pt undergo an austenite-martensite transformation irrespective of whether or not the parent phase is ordered (Dunne and Wayman. for nonthermoelastic and thermoelastic transformation. No differences in the transformation crystallography can be detected. 1969) may be more appropriate: A martensitic transformation is in general described by an invariant-plane strain relief effect at a free surface.5) alloys. respectively (After Kaufman and Cohen. according to which this body of information is completely internally consistent.

Nucleation of Martensite It is implied from earlier discussion that martensitic transformations occur during quenching. growth continues until the plates eventually impinge upon a grain boundary or another plate. and accordingly various dislocation-type nucleation models have appeared. Accordingly. they usually occur over a temperature range given by M8 (the martensite start temperature) and Mf (the martensite finish temperature). but instead must commence at some lattice defect or inhomogeneity. where A refers to the parent phase generally termed austenite (following steels). Simply put. It is observed that the critical stress to form martensite increases linearly with an increase in temperature from M8 up to a temperature Msa above which martensite forms only after plastic deformation of the parent phase. 1972). 5. A good analogy here would be to imagine filming the transformation on It is generally believed that martensitic transformations cannot be homogeneously nucleated (Cohen and Wayman.. On the other hand. complete crystallographic reversibility. The Au-Cd alloy with narrow thermal hysteresis exhibits thermoelastic behavior. 1957). and quite specific nucleation models have appeared (Olson and Cohen. no matter what the stress. This behavior is shown in Figure 6 (Olson and Cohen.and Strain-Induced Martensitic Transformations As early as the 1930s Scheil (1932) predicted the formation of martensite above M8 by the application of a stress. With reference to Figure 5 it can be said that the Fe-Ni alloy exhibits non-thermoelastic behavior whereas the Au-Cd alloy shows thermoelastic behavior. In non-thermoelastic transformations during cooling. When the specimen is heated the reverse transformation occurs by the 'backward' movement of the martensite/parent interface. Kinetics of Martensitic Transformations cooling. dislocation clusters. In an oversimplified view. Thus.e. transformation proceeds by the continuous growth of the plates upon cooling. and because of the plastic deformation new nucleating sites are introduced. In thermoelastic martensites. in the above description one uses all aspects of the crystallography and PTMC to define martensite. etc.In effect. above which martensite cannot be mechanically stimulated. a martensite plate usually springs full size into existence. The literature abounds with cases where martensite is nucleated at grain boundaries. If the cooling process is stopped. While this may be the case for steels. At Msff the stress surpasses the parent yield strength. Thus. According to him. the nucleation is stress-assisted and existing nucleating sites are simply aided mechanically. the shear stress required to activate the transformation decreases with decreasing temperature (being zero at M8) whereas the shear stress required for austenite slip increases with decreasing temperature. and the reverse transformation occurs by the separate nucleation of the austenite phase. 1981). Stress. It has been convincingly shown (Saburi and Nenno. however. and then its interface becomes sessile. as is shown in Figure 5 for Fe-Ni and Au-Cd alloys (Kaufman and Cohen. Although in steels when the martensite is heated it decomposes by tempering. Furthermore. M d . 1986) that dislocations act as preferred sites for martensite nucleation. The martensite plates revert completely to the parent phase and to the original lattice orientation. one can imagine that the relaxed lattice at the core of a dislocation permits the Bain distortion to 'get started easier'. 4. The magnitude of the hysteresis is. and then running the film backwards to depict the reverse transformation on heating. in this case the nucleation of the parent phase requires a higher-energy barrier than the shrinkage of plates already there. at temperatures near M8. not the only difference. Indeed. in general a reverse martensitic transformation occurs given by the As and A1 temperatures. But the thermal hysteresis can vary widely from case to case. martensitic transformations in many materials can occur during 'slow' cooling. the Fe-Ni alloy exhibits wide thermal hysteresis. i. but. applied stresses should induce plastic deformation by the martensitic mode rather than by slip. Other differences are found in characteristics of both the forward (P->M) and reverse (M->P) transformations. it is not so in general. The evidence for nucleation at imperfections is overwhelming. 1976). if it is resumed. 3. There is another temperature. where the reverse transformation is essentially the inverse of the forward (cooling) transformation. a distinction between elastic stress-induced martensite and plastic strain-induced martensite can be made. twin boundaries. . growth ceases. Note that between M8 and Msa the stress is below the yield strength of the parent and is thus an elastic stress.

At present.Applied stress manner. The SME behavior is basically a consequence of a martensitic transformation. 1958). It is to be noted that the 'one-way' memory just described is a one-time-only occurrence. Because of crystallographic restrictions and the necessity to maintain ordering. Between the SME discoveries in Au-Cd and Ni-Ti came similar findings for /3-brass (Reynolds and Bever. As the martensite phase is deformed. activity in the SME field did not really become extensive until Buehler and collegaues (Buehler et al. there are 12 possible combinations of stacking planes and shear directions. There are six {110} planes and two possible shear directions for each {110} plane.and two-way shape memories is given in Figure 8. In other words. 1979). deforming it. Each combination leads to two possible crystallographically equivalent unrotated and undistorted habit planes. In addition to the familiar 'one-way' memory. shapememory alloys are found to have common characteristics such as atomic ordering. Cu-Zn-Al. there are numerous variants of the Bain strain during the forward transformation but only one during reversal.. and a 'mechanical' shape memory resulting from the formation and reversal of stress-induced martensite. some variants grow at the expense of others. See also Chapter 26 by Schetky in Volume 2 on applications of shape-memory alloys. ab. Chang and Read studied a Au-47. Completely recoverable strains of over 100Io have been observed.3 Two-Way Memory A comparison between the one. the original specimen shape and parent single crystal are regenerated. Shape-Memory Effects 6. When this resultant single crystal of martensite is heated between As and Af. the basal plane of martensites (say 18R (or 9R)) can be known to originate from one of the {110} basal planes of the parent DO3 (or B2) phase. The surviving variant is that whole-shape strain direction that is most parallel to the tensile axis. showing no relief effects. and ac are discussed in the text (After Olson and Cohen. This inhomogeneous shear (of the phenomenological theory) during transformation is also known to occur on the basal plane. In the two-way shape •From the observed parent-martensite orientation relationship and other features such as stacking faults in the martensite. Upon cooling a single crystal of the parent. 1972) 6. They form in self-accommodating groups of four variants in a diamond-like morphology. 1952) and Cu-Al-Ni alloys (Rachinger. etc. the single crystal of martensite has only one way to undergo the reverse transformation. Hence. typically 24* variants of martensite form. Ni-Ti (Nitinol). This diagram is of universal applicability.2 One-Way Memory The shape-memory process is schematically illustrated in Figure 7 (Saburi et al. and eventually only one variant persists. resulting in 24 martensite variants in a parent grain Temperature Figure 6. and Ni-Ti alloys are all being commercially exploited for their SME. the specimen surface is featureless. 1963) found similar behavior in a less exotic alloy. and a martensite phase that forms in a self-accommodating . At this point. <ra. Although discovered in Au-Cd by Chang and Read (1951). 6.1 Overview The shape-memory effect (SME) has brought renewed interest that has led to recognition of diverse martensitic transformations. shapememory alloys can also exhibit a 'two-way' memory.. Some shape-memory alloys also exhibit 'rubber-like' behavior. a thermoelastic martensitic transformation that is crystallographically reversible.5Cd alloy with a view to the general features of its cubic-orthorhombic martensitic transformation. Stress-temperature diagram showing the critical stress for martensite formation in a typical iron alloy as a function of temperature. Cu-Al-Ni. The explanation of the shape-memory phenomenon is now universal and well in hand. thus permitting maximum elongation of the specimen. but can be revived by re-forming the martensite. When compared. 6.

way memory Below Af1 (b) Two .Single crystal of parent phase Self. The two-way memory is associated with built-in microstresses. and (ii) a shape change during cooling. This differs from the one-way memory in two respects: (i) infinite repeatability. These microstresses * program' the object so that during cooling between . no macroscopic shape change 24 variants of martensite Reverse transformation on heating.accommodating martensitic transformation on cooling. Schematic illustration of the various processes involved in the shape-memory effect (After Saburi et al.which progressively accumulate after appropriate thermomechanical treatment (Schroeder and Wayman. 1977). An example using coil springs to compare the differences between the one-way and two-way shape memories memory the object changes shape upon both cooling and heating. recovery of the original shape Deformation by variant coalescence on stressing (below M f ) Single crystal of martensite No shape change on unloading Single crystal of martensite Figure 7. 1979) (a) One .way memory Below M f Above Af Below M f Below M 1 Above Af Figure 8.

it is interesting to note that the rubber-like behavior is found only after the martensite is aged at room temperature (M8 = 60 0C) for at least a few hours. First. Freshly transformed specimens exhibit shape-memory behavior and bent rods will not spring back to their initial shape. Broadly speaking. When the stress is released. the parent phase undergoes ordinary plastic deformation. and heated above Ax again. 1979) M8 and Mx only one variant of the martensite forms— hence the shape change. A stress-induced martensite then forms and in reverse generates the microstress sources. and then heated to above Af. but one should use care in the interest of preciseness.1Zn shape-memory alloy deformed in tension at about 50 0C above its M8 temperature (Schroeder and Wayman. 60 years later. Olander's (1932) report of rubber-like behavior in Au-Cd was the first indication of the existence of a shape-memory effect. At the extreme rear the stress-strain curve shown in the o-e plane corresponds to the deformation of martensite below Mx. 1979). i. These two types of behavior collectively fall in the category of 'pseudoelasticity'. But rubber-like behavior is characteristic of a fully martensitic structure whereas superelastic behavior is associated with formation of martensite under stress. Both of these systems show shape-memory behavior. who studied an In-20. It is then cooled below Mx and deformed again in exactly the same manner to the same shape. or a combination. and by Burkhart and Read (1951). That is. Instead.125 0C and transformed at -77 0C (After Schroeder and Wayman. the strain follows the lower plateau and fully recovers as the SIM reverts to the parent. the middle o-e plane. The two-way memory can be brought about by either.8Zn) deformed at about 50 0 C above the Ms temperature. 6. Yielding at an essentially constant stress (upper plateau) corresponds to the formation of 9R stress-induced martensite (SIM) from the B2 parent. This phenomenon was later confirmed by Chang and Read (1951) in a Au-47. It is ironic that even today. of two procedures. This process is repeated a few times and then the specimen is fully programmed. 6. At about 9% strain the specimen becomes fully martensitic.induce the martensite decreases with temperature and falls to zero at M8.7Ti alloy.4 Superelasticity Figure 9 is a stress-strain curve for a single-crystal specimen of a Cu-39.5Cd alloy. a specimen can be superelastically stress-cycled above M8. Superelastic stress-strain loop for a Cu-Zn alloy (Cu-39. The stress necessary to . Considering the Au-Cd alloy. At a temperature above M s (and Af) SIM is formed. the specimen is subjected to deformation in the martensitic condition. Stress (MPa) Stress (ksi) 6. M5= .e. The shape-change process upon heating is the same as with the one-way memory: a single crystal of the martensite reverts to a single crystal of its parent. leading to a superelastic loop with an upper and lower plateau. the alloy deformed like rubber with low effective modulus when loaded but it sprang back to the initial shape when unloaded.5 Rubber-like Behavior Rubber-like behavior was first observed for a Au-Cd alloy by Olander (1932). as seen in the e-T plane. A stress-strain relationship such as that shown in Figure 9 is frequently referred to as a superelastic stress-strain loop.6 Summary Figure 10 is a stress-strain-temperature diagram for a Ni-Ti shape-memory alloy that summarizes its mechanical behavior. The induced strain. At a still higher temperature and above M d . This behavior corresponds to a mechanical (as opposed to a thermal) shape memory. no SIM is formed. the origin of rubber-like behavior remains obscure. Strain (%) Figure 9. Rubber-like behavior also constitutes a mechanical type of shape memory as does the process of SIM formation. based on microscopic observations. the front o-e plane. about 4%. Alternatively. Chang and Read (1951) attributed the rubber-like behavior to favorably oriented * regions' growing at the expense of others. recovers between As and Af after the applied stress has been removed and the specimen heated.

substitutional alloys including ordered intermetallics. 1930). C. 8. a Stress.9 1972. they also exhibit surface relief (Ko and Cottrell. indicative of diffusioncontrolled processes. 1975). e (%) Figure 10. Many researchers have been attracted to the former model but some believe in the latter because similar surface relief also accompanies a higher-temperature product. Bainite also forms during continuous cooling.. Ohmori and Maki (1991) have pointed out recently that Widmanstatten ferrite can form displacively if the magnitude of the shape strain and the morphology of surface relief are taken into account. 1988). it should be noted that martensitic transformations occur in numerous ceramics.1. 1952. and the product exhibits a plate-like morphology. the transforming particles in effect de-energize the cracks by absorbing energy and slowing cracks (Sakuma. By proper doping (alloying). o (MPa) Stress. a characteristic of a diffusionless (or displacive) reaction. 8. e Strain. and is termed 'bainite* in honor of E. 1962). Lorimer et al. An especially interesting case involves zirconia. Bain who first observed such microstructures in collaboration with Davenport (Davenport and Bain. However. This product is a mixture of ferrite and carbide.Below /Wf Stress. 1973). Bainite Transformations Bainite forms isothermally as an intermediate transformation product above M s but below the temperature of eutectoid decomposition in ferrous alloys. Widmanstatten ferrite (Watson and McDougall. Martensitic Transformations in Ceramics Although beyond the scope of the subject matter at hand. thereby an incubation period is present during isothermal annealing. Speich. When the matrix is stressed. e Strain. Because of a lack of decisive evidence. Because of these contradictory characteristics. which is distinctly different from pearlite (a eutectoid product). mechanisms are mainly categorized into two distinctly different models: a diffusionless martensitic transformation (Garwood.1 Bainite Transformations in Ferrous Alloys Bainite transformations in steels are characterized by a C curve on a time-temperature-transformation (TTT) diagram. it will become ductilized by the stress-induced tetragonal-to-monoclinic martensitic transformation in the stress field of advancing cracks. Stress-strain-temperature diagram for a Ni-Ti (Nitinol) shape-memory alloy showing shape-memory and superelastic characteristics and the deformation behavior of the parent phase above the Afd temperature (above which no martensite can form regardless of the magnitude of the stress). Bainite transformations occur with compositional changes. which will be described in Section 11. o Strain. whether bainite transformations in ferrous alloys are displacive or diffusion-controlled remains unsolved. 1954) and a diffusion-controlled precipitation phenomenon (Hehemann et al. Temperature increases from upper right to lower left 7. Attempts to improve ductility of intermetallic compounds by utilizing stress-induced martensites have recently been made. tetragonal precipitates of zirconia will form in the cubic matrix. Bainite transformations occur not only in ferrous alloys but also in non-ferrous .

Ag-Zn (Kubo and Hirano. As seen in Table 1.Table 1. and then transforms into a f. 1968). Cu-Zn-Al (Wu and Wayman.2 Bainite Transformations in Ordered Intermetallic Alloys Bainite transformations occur in many Cu. Takezawa and Sato.-based. 1989). Wu et al.3Zn Cu-28. 1974).0Al Cu-40Zn-4Au Cu-40Zn-9Au Ag-45Cd Ag-27. (cF16) B2 (cP2) B2 (cP2) Bainite Disordered 9R Disordered 9R Ordered 9R Ordered 9R to disordered 9R Ordered 18R to ordered 9R Ordered 9R Ordered 9R to disordered 9R Ordered 18R Disordered 9R 9R(?) References Cornelis and Wayman (1974) Sato and Takezawa (1968) Takezawa and Sato (1986) Takezawa and Sato (1990) Wu et al. Ordered intermetallics exhibiting bainite transformations include Cu-Zn (Cornelis and Wayman. The 9R and 18R structures are layered structures and are similar in stacking sequence. and Ag-Cd (Wu et al. For example..c.1Al Cu-29.1Zn-6. region.4 Shape Change A shape change occurs in Cu-Zn and Cu-Zn-Al alloys when aged at temperatures above Ms under stress. the change in composition during a bainite transformation can be clearly observed by electron diffraction and microanalysis techniques. while the latter is from a DO3 parent with a stacking sequence AB'CB'CA'CA'BA'BC'BC'AC'AB'. (3R) structure of the equilibrium a phase. cI2) phase of binary and ternary Cu-Zn alloys and binary Ag-Cd and Ag-Zn alloys transforms when cooled or quenched into a metastable ordered b. 1989.and Ag-based alloys Alloy Cu-37. in a Cu-29. Cu-Zn-Au (Doig and Flewitt. This crystal structural difference in terms of atomic ordering is due to the difference in diffusion rate of solute atoms in bainite plates.c. in addition to the ax phase.1Zn-6. Tadaki et al. the Ci1 bainite becomes the disordered f.c. . The ax phase is a plate-shaped bainite and reveals surface relief. Upon prolonged aging. 1986).7Al alloy the B2 phase transforms initially into the bainite phase having an ordered 9R structure and then to a disordered 9R structure during thickening of the plates. the C 1 bainite is either ordered or x disordered at early stages of growth.c.8Zn-9. or L2j is determined by heat-treatment conditions and alloy compositions (Wu et al.c.c. Whether the parent phase is of B2. a rod-shaped Widmanstatten product of the equilibrium.c.8 to 43. DO3 (cF16) or Ll1 (cF16) structure.0to 28. This metastable phase undergoes a bainite transformation during isothermal aging into a f. 1983). 8. 8.c. and thereby better understanding of the bainite transformation is achieved. Crystal structures of parent and bainite phases at early stages of transformation in Cu.c. the former is derived from a B2 parent having a stacking sequence ABCBCACAB.c. 1988. (1989) Tadaki et al. and Ti alloys when a metastable phase is aged isothermally at a moderate temperature between the Ms point and the eutectoid temperature after being quenched from a high-temperature 0-b. both being related with the stacking of the parent [110} planes.c. while such disordering occurs even at an early stage in Cu-40Zn and Ag-45Cd alloys (Cornelis and Wayman. (1989) Kubo and Hirano (1972) 8. 1988). (1988) Tadaki et al. (1991) Wu et al. phase that has either a B2 (cP2). Because ordered intermetallics offer structural advantages over disordered alloys in terms of atomic arrangements and do not contain carbon atoms that cause intricate phenomena related with carbide precipitation as in steels.8Zn-9. a phase forms during isothermal aging in a higher temperature range. (A2.c.3Zn Cu-39. similar to that of martensites formed by subzero cooling. A TTT diagram study by Flewitt and Towner (1967) has revealed that. Cu-Al (Moon and Garwood. Ag.9Zn-6.. 1972).-based ^1 phase (also called ab phase) having either a 9R or 18R structure. The metastable phase becomes ordered when quenched.0Al Cu-24.c. (1989) Hamada et al. DO3.c.3 Crystal Structures of Parent and Bainite Phases The high-temperature 0-b. In the case of martensites. 1967). as shown in Figure 11 and 12. (1991) Tadaki et al. disordered f. 1990). Cu-Sn (de Bondt and Deruyttere.1Zn Parent B2 (cP2) B2 (cP2) B2 (cP2) B2 (cP2) L2. although the parent phase is ordered.7Al Cu-24. (cF16) B2 (cP2) B2 (cP2) L2. close-packed ordered structure. 1974. (1991) Wu et al.

8. 1974). with an activation energy corresponding to that for the diffusion of Zn in the matrix.1Al alloy upon isothermal aging at various temperatures (above M s ) at a tensile stress of 125 MPa. 1973) Takezawa and Sato (1986. Optical micrograph of Ot1 plate-shaped bainite in a Cu-39.5% in a Cu-28. 1974.3Zn alloy transformed at 450 0C for 90 s (After Cornells. (a) Optical and (b) interference micrographs of a. that of shape recovery with the shape-memory effect of the same alloy.1988) observed elongations up to 5.Figure 11. During prolonged aging. from their work. The ultimate shape change is roughly equal to. They conclude that diffusion is necessary to prepare suitable surrounding conditions for the nucleation and growth of bainite crystals to be formed by a shear process. but not higher than. showing surface relief (After Cornelis.5 Microstructure The microstructure and crystal structure of bainite crystals in Cu-Zn alloys are quite similar to those of diffusionless martensites forming at subzero temperatures in the same alloys (Cornells and Wayman. The tips of bainite plates are defect-free (Takezawa and Sato. They attribute the elongation to the formation of particular variants of bainite. The density of stacking faults is higher at higher aging temperatures. Warlimont and Delaey.9Zn-6. Figure 13.6Zn-l. 1939). 1988) think that the shape change results from a martensitic process. 1974). plate-shaped bainite of a Cu-39. Takezawa and Sato (1986.2Ag alloy transformed at 430 0 C for 5 s. The shape change occurred only after an incubation period and followed a curve that represents the kinetics of a diffusion-controlled transformation. similar to very thin martensite plates at an early stage of growth (Saburi et al. 1973) contain a fine internal structure consisting of stacking faults parallel to the basal plane of the 9R structure at early stages of transformation (Cornelis and Wayman. Bainite plates Figure 12. shows the incubation period as well as the good agreement of elongations upon the bainite transformation with a theoretical curve based on a diffusion-controlled Austin-Ricketts process (Austin and Ricketts. the 9R stacking sequence of the bainite is gradually destroyed and thin bainite plates increase their width concurrently with annihilation . The incubation period and the time necessary to reach the ultimate elongation decrease and the growth rate increases with increasing isothermal temperature.% 1986). 1986).

c.6 Habit Plane The habit plane of the ^1 bainite is (2 11 12J B2orL2 at early stages of transformation in both binary and ternary Cu-Zn alloys (similar to P1 '-9R martensite in binary Cu-Zn alloys) and changes to {123} B 2 o r L 2 l at later stages where the bainite eventually becomes disordered f. the density of stacking faults becomes very low where the crystal structure of the bainite becomes the disordered f.8. nearly parallel to {2 11 H]12 (Takezawa and Sato. variants having the highest Schmid factor do not always occur predominantly. (b) a rod.c. In a Ag-27.c. Macroscopic shape change during bainite transformation in a Cu-40Zn alloy subjected to a tensile stress of 98 MPa at 473 K. (c) grain boundary (After Takezawa and Sato. grain boundaries.9Zn-6. Flewitt and Towner. resulting in the difference in the ultimate shape change under stress comparing bainite and martensite formation. The curve is theoretical (After Takezawa and Sato. Stress-assisted bainite plates form under stressing at moderate temperatures in Cu-Zn and Cu-Zn-Al alloys. (Cornelis and Wayman. 1988). structure of the equilibrium a phase (Cornelis and Wayman. 1986). In Cu-Zn-Al alloys the habit plane is reported to be {155j L2 . On further aging.1 Al alloy: (a) oxide particle. and inclusions.1Zn alloy the habit planes are reported to be both {235} B2 and {134} B2 (Kubo and Hirano.7 Nucleation and Growth Similar to martensite. Figure 14 is a set of electron micrographs showing small ax bainite plates Figure 14. 1972). 1967). Several variants of stress-assisted bainite occur but all have the {2 11 12)L2i habit plane (Takezawa and Sato. 1986) of stacking faults. precipitate rods. 1974). bainite plates nucleate at or near structural defects such as dislocations. 8. Elongation strain (%) Time (x 103S) Figure 13. Different from 9R martensite.c. 1974. Electron micrographs showing various nucleation sites of bainite plates in a Cu-28. 1986) .

Substantial changes in composition occur during the bainite incubation period by stress-induced diffusion of solute atoms at dislocations.8 CmS" 1 at 2500 C . from Hamada et al. (1991).6 CmS 1 at 250 0 C and that the growth ceases at later stages. Figures 15 and 16.Figure 15.8 Growth Kinetics The Ct1 bainite plates in Cu-Zn and Cu-Zn-Al alloys grow anisotropically. The former assumption.5XlO. These nucleation sites are structurally inhomogeneous and may have Zn concentrations that are almost equal to that of a bainite plate. 8. in that there is no semicoherent interfacial boundary between an ax plate and the matrix at an early stage . 1991) formed at various nucleation sites in a Cu-41. and their results indicate that the rapid lengthening can only be rationalized under the two assumptions of a semicoherent interfacial boundary and supersaturation of Zn concentration.0Zn alloy aged at 2300 C (Takezawa and Sato. suggesting that the growth of «! plates is controlled by Zn diffusion. Transmission electron micrograph of an C 1 bainite plate formed after aging at 423 K for 175 ks. rapid lengthening and slow thickening. is not supported by high-resolution transmission electron microscope (HRTEM) studies of the interface. while plate thickening occurs slowly at a rate of 0. however.6XlO. because such inhomogeneity causes a localized stress gradient that assists diffusion of solute atoms. Lengthening kinetics were studied by Simonen and Trivedi (1977).e. Hornbogen and Warlimont (1967) studied growth rates and determined that ^ 1 plates in a Cu-Zn alloy lengthen linearly at a high rate of 1. The activation energy obtained from the lengthening rate is nearly equal to that obtained from diffusion coefficients of Zn in the 0 phase. Microanalyses along K lines a-b and c-d were performed (After Hamada et aL. show that the Zn concentration near dislocations is almost equal to that of a bainite plate in a Cu-Zn-Al alloy. 1986). i.

. 8. and it is nearly equal to that of the a. which has improved resolution. who reported that the composition of an C 1 bainite plate was similar to that of the x matrix near a tip of the Cs1 plate in a Cu-Zn-Au alloy. 1989) alloys. 1991) of transformation in a Cu-Zn-Al alloy (Hamada et al. The slow plate-thickening kinetics caused by Zn diffusion also may not agree with recent compositional observations. 1991). d (nm) Al content (mass%) Dislocation Al content (mass%) Dislocation Distance. but when diffusion actually occurs has been in debate. This discrepancy. d (nm) Distance. probably due to the difference in diffusion rates of solute atoms in these alloys. d (nm) Distance. Contradictory results are found by Doig and Flewitt (1983). 1990). however. .Zn content (mass%) Zn content (mass%) Dislocation Dislocation Distance. the composition in the bainite plate in all alloys remains almost constant upon further aging. phase.. 1992).. d (nm) Figure 16. Recent studies with analytical electron microscopy. and Ag-Cd (Tadaki et al. is probably due to low resolution. Although the ^1 bainite shows slightly different aging behavior with respect to disordering. bainite plate in Figure 15 (After Hamada et al. Zn and Al concentrations (a) along line a-b and (b) along line c-d of microanalyses of the a.9 Compositional Alteration It has been accepted that a compositional change occurs in association with bainite transformation. 1989. Cu-Zn-Al (Nakataefa/. have revealed that the compositions of an Ci1 bainite plate and the surrounding matrix are distinctly different even at a very early stage of growth in Cu-Zn (Cliff et al. Cu-Zn-Au (Tadaki et al. 1991). as described in the following section. Hamada et al.

structure during prolonged aging.. resulting in a rise of the Ms temperature.8.0Al alloy at different stages of the transformation. Tadaki et al.12 Models of Bainite Formation There are basically two different mechanisms describing the bainite transformation. and either ledges or lattice misfits appear at interfacial regions having stacking irregularity.c. 1989.. although in the former case the central issue remains as to when the two processes occur.10 Atomic Disordering Whether mechanisms of bainite transformations involve a diffusionless and/or a diffusion reaction can be assessed using the unique structural advantages that ordered structures possess. This basal plane maintains complete coherence with the (022) L 2 plane of the parent phase. In inner regions. A mostly descriptive model of the C 1 bainite x transformation has been proposed by Takezawa and Sato (1986. The boundary of both NN and NNN APD structures matches perfectly at interfaces between the matrix and C 1 plates.c. especially at the tip of the bainite plate. As such a nucleus forms by a martensitic shear process. At later stages. resulting in lengthening of the plate.. x C 1 plates thicken and the NNN order is lost in outer x regions. 1975). At temperatures below the B2->L2 1 ordering temperature. Hamada et al. Recent detailed microanalysis and electron microscope studies appear to rule out the latter and support the former. Wu et al. (1992) have concluded that the transformation is initiated martensitically after diffusion of solute atoms but eventually becomes dominated by diffusion kinetics.. Since the stable phase of the alloy is disordered f. The parent 0 phase of the alloy becomes an ordered Ulx structure via an intermediate B2 structure. (1992) studied inter facial structures between an C 1 bainite plate and the Ul1 matrix in a x Cu-Zn-Al alloy at different growth stages. (1991) have studied antiphase-domain (APD) structures of C 1 bainite plates and the matrix in x a Cu-26. although the NN order and the APD structure are preserved. 1991). These observations do not fit the rapid lengthening kinetics based on a semicoherent interfacial boundary as suggested by Simonen and Trivedi (1977).. the NNN order remains. 1967). 8. the periodicity loses its regularity. 1967). followed by further diffusion to approach the composition of the equilibrium phase (Takezawa and Sato. This thermodynamic approach and the microanalysis results appear to rule out model (1). (2) Diffusion occurs first at a region of the matrix where the chemical composition is close to that of the equilibrium ex phase and then a shear process is triggered by a rise of Ms. although a strain field appears at the interface where these two planes meet at an angle of a few degrees.7Zn-4. (3) A martensitic transformation occurs after some compositional change and then diffusion of solute atoms occurs along with structural and morphological changes (Nakata et al. 8. They observed that there are periodic stacking faults at every third (0018) basal plane of the 18R-Cx1 plate in early stages of growth. Thermodynamic approaches by Hsu and Zhou (1989) suggest that a composition-invariant bainite transformation is energetically unfavorable in Cu-Zn alloys. (1989) and Hamada et al. both by introduction/annihilation of stacking faults parallel . their calculations suggest that bainite transformation occurs by a shear process if diffusion of solute atoms occurs either before the transformation or concurrently with it. These observations indicate that substantial diffusion occurs at later stages of bainite growth. solute Zn atoms diffuse by a localized stress gradient around defects such as dislocations.c.c. while in the other it occurs as a diffusion-controlled precipitation process (Lorimer et al. The stress promotes diffusion of solute atoms. 1989).11 Inter facial Structure Hamada et al. structure (Flewitt and Towner. a new stress field is created in the matrix.c. In other words. followed by diffusion of solute atoms to approach the composition of the equilibrium a phase with a structural change into a f. Figure 17 is a high-resolution electron micrograph showing an interface between an C 1 plate and the x L2j matrix. At later stages. The models proposed are as follows: (1) A diffusionless martensitic transformation occurs first in the matrix without any compositional alteration. 1988. the bainite plate further transforms into a f. solute atoms diffuse out to the matrix. Hamada et al. and a nucleus of C 1 bainite that x is martensitically triggered. During thickening. Taking compositional alterations into account. In one type a bainite transformation involves both diffusionless and diffusional processes (Flewitt and Towner.c. 1990). Wu et al. at aging temperatures. In their model. 1986. The B2 phase possesses only nearest-neighbor (NN) order while the Ulx phase possesses both NN and next-nearest-neighbor (NNN) order. both NN and NNN order and the associated APD structures are inherited by C 1 plates during the initial stage of plate x formation and maintained at a subsequent stage of plate thickening.

the composition of bainite plates approaches that of the equilibrium phase (except for bainite in Cu-Sn alloys. At later stages. a lack of understanding of theoretical aspects remains. High-resolution electron micrograph of the interface between a{ 18R-bainite plate and L2. It is interesting to note that bainite transformations occur not only in metallic alloys but also in some ceramics such as ZrO2-Y2O3 and ZrO2-CeO2 (Nakanishi and Shigematsu.c.c. a phase. the observation that the initial composition of bainite plates in Cu-Zn-Al and Cu-Zn-Au alloys is almost equal to that of the equilibrium f.g. the periodicity of the stacking faults loses its regularity. matrix in a Cu-26. showing ledges and misfits (indicated by arrows) formed at interfacial regions where the 18R stacking periodicity loses its regularity (After Hamada et al. e. and ledges appear at some inter facial regions. However. 8. where no stable phases have the same structure as the bainite (de Bondt and Deruyttere. or simultaneously with. there are some unexplained phenomena. 1991). Bainite plates form martensitically either immediately after. However. phases upon isothermal aging above Ms and below eutectoid temperatures. Detailed structural and chemical analyses of bainite forming in ordered intermetallics have upgraded experimental understanding of bainite transformation.t 1992) to the basal plane of the 9R or 18R structure and by atomic disordering to reduce the chemical free energy.c.0Al alloy aged at 423 K for 864 ks. This model appears most plausible in describing certain observations associated with bainite transformation.13 Summary Bainite transformations occur in many ordered b. whose localized stress fields assist solute atoms to diffuse. some compositional change at a local region of the matrix that has lattice defects.7Zn-4. At an early stage of bainite growth. .Next Page Figure 17. Upon prolonged aging. periodic stacking faults are present and the fault plane maintains complete coherence with the parent phase.c. 1967)).

Index Note: Figures and Tables are indicated [in this index] by italic page numbers Index terms A A15 compounds irradiation effects point defects in superconductors as superconductors vacancy migration Ab initio calculations. antiphase-boundary energies Adiabatic elastic constants compared with isothermal elastic constants Ag-Cd system Ag-Cu-Zn system Ag-Zn system Al-Co system Al-Cu-Ni system Al-Cu-Zn system Al-Fe-Ni system Al-Fe system composition dependence of elastic moduli crystallographic transformation Alloying additions ductility affected by major trace-element Al-Mn-Ni system Al-Nb-Ti system Al-Ni-based alloys dislocations in ductility 99 240 This page has been reformatted by Knovel to provide easier navigation. Links 182 150 182 198 151 69 1 4 230 231 230 14 226 226 235 53 14 239 28 28 28 28 238 184 198 86 240 240 230 236 56 231 240 232 233 29 30 249 .

dislocation cores affected by Antiphase boundaries (APBs) amorphization at chemical effects classification of in D022/D023/L12 relation and dislocations and disorder-order transformations dragging of APBs and grain boundaries ideal deviations due to chemical effects deviations due to structural effects nearest-neighbor models residual contrast segregation of alloying elements segregation of vacancies structural effects surface energy in two-dimensional AB alloy Antiphase-boundary (APB)-coupled dipoles Antiphase-boundary (APB) energies ab initio calculations anisotropy calculations 181 201 198 207 210 112 67 206 68 68 71 83 68 118 159 68 68 68 72 73 76 75 55 76 67 123 76 69 69 79 This page has been reformatted by Knovel to provide easier navigation. Ruthenium…. Zirconium aluminides Amorphization irradiation-induced dose required experimental observations mechanisms models for Anisotropic elasticity.250 Index terms Aluminides crystal structures see also Cobalt…. Titanium…. Nickel…. Links 41 182 202 198 205 215 113 74 115 72 161 69 74 72 74 177 86 68 72 124 86 89 . Iron….

251 Index terms Antiphase-boundary (APB) energies (Continued) and deformation mechanisms and dissociation of dislocations effect of long-range order parameter experimental determination listed B2 compounds D03 compounds D022 compounds L10 compounds L12 compounds L21 compounds plot vs. diffusivity affected by Atomistic models. grain boundaries Atomic packing. grain boundaries Atom-probe (AP) analysis. . and point defects Atomic images. disorder – order transition temperature and structural stability temperature dependence Antiphase-boundary (APB) tubes Antiphase domains (APDs) in bainite phases crystallography disorder-order transition isostructural transformation morphology ‘Swiss-cheese’ structure in two-dimensional AB alloy Antisite defects formation of Antistructure atoms Atomic diffusion. grain-boundary chemistry studied using Au-Cd system 44 97 79 80 77 78 79 78 78 77 79 80 71 75 68 67 235 81 81 81 81 81 67 182 136 52 146 158 51 172 163 224 177 167 225 152 159 183 185 82 82 112 84 89 98 Links 103 122 168 This page has been reformatted by Knovel to provide easier navigation.

252 Index terms Au-Cu system twinning in see also Copper-gold alloys AuCu3 antiphase-boundary energies variation with temperature elastic properties Au-Cu-Zn system Auger electron spectroscopy (AES) limitations 76 75 3 230 27 27 16 234 31 77 238 Links 89 163 B B2 type compounds APB energies crystal structure defects/dislocations in dislocations in flow-stress anomaly in fracture of grain-boundary structure model vacancy migration vacancy migration in B11 structure irradiation effects vacancy migration in Bain deformation Bainite compositional alteration crystal habit plane crystal structure growth kinetics interfacial structure microstructure nucleation and growth shape changes 78 26 107 95 64 34 175 148 147 188 188 147 221 229 234 232 230 233 235 231 232 230 This page has been reformatted by Knovel to provide easier navigation. 92 114 107 95 108 188 149 222 .

amorphous phases studied using Brittle fracture B2 compounds L12 compounds models for Brittleness Brittle-to-ductile transition 112 This page has been reformatted by Knovel to provide easier navigation. plasticity affected by Boron ductility affected by adverse effect of sulfur fracture affected by interstitial boron in AlNi3 segregation at grain boundaries solubility in Ni3X Boron-doped Al-Ni-based alloys ductility grain-boundary structures Bragg-Williams approximation Bragg-Williams LRO model 29 159 74 189 108 134 204 34 27 37 25 41 48 109 136 35 160 142 29 34 26 29 29 165 31 166 35 176 112 49 65 229 235 236 229 235 230 192 107 Links 36 54 34 165 166 54 174 177 β-Brass [CuZn] dislocations in vacancies in Brillouin scattering.) alloys. dislocation in Bond directionality at grain boundaries dislocation cores affected by Bonding.253 Index terms Bainite transformations and atomic disordering in ceramics in ferrous alloys models describing in ordered intermetallic alloys Banerjee-Urban [irradiation-induced ordering/disordering] model Body-centred cubic (b. .c.c.

254 Index terms Brittle-to-ductile transition temperature. factors affecting Bulk modulus calculated values composition dependence for polycrystals estimation from single-crystal constants 13 15 13 11 13 65 Links C C11b structure dislocations in elastic properties of compounds point defects in CANDU nuclear power reactor Carbides. protective Cobalt aluminides elastic properties 13 This page has been reformatted by Knovel to provide easier navigation. transformations in C-Fe-Ni system Chalmers-Martius criterion violations Charge transfer. 92 111 5 147 181 151 2 3 229 224 114 107 37 111 14 52 207 244 25 25 25 83 42 14 65 208 247 115 236 . point defects in Cauchy relationships listed for various intermetallics Ceramics. and irradiation effects Clapeyron-Clausius equation Cleavage fracture examples meaning of term Cluster variation method (CVM) Coatings. embrittlement caused by Chemical composition and dislocation cores elastic properties affected by plasticity affected by Chemical disordering.

Nabarro-Herring creep Creep resistance and diffusion coefficient 53 139 184 132 145 159 238 137 147 57 211 193 149 161 145 185 139 197 193 201 151 56 210 118 203 199 15 133 137 47 43 227 158 170 158 172 49 44 Links 64 This page has been reformatted by Knovel to provide easier navigation. Dislocation…. Diffusional. vacancy formation in Coble creep Co-Fe-V system Coil-spring behavior. Harper-Dorn…. . vacancy migration in Corrosion Coulomb explosion model Cowley-Warren SRO parameter Creep see Coble…. irradiation-induced amorphization Copley-Kear model Copper-gold alloys electrical resistivity irradiation effects production of point defects in recovery of point defects TEM/HRTEM observations of grain boundaries twinning in vacancy formation in Copper-titanium alloys. shape-memory effects compared Coincidence-site lattice (CSL) boundaries Coincidence-site lattice (CSL) model Coincidence-site lattices Collapse of cascades to dislocation loops to vacancy loops Collision cascade parameters Collision cascades Composites Computer models. Inverse….255 Index terms Cobalt aluminides (Continued) effect of alloying elements point-defect formation properties Cobalt-gallium alloys.

. irradiation-induced Crystallographic transformations Crystal structures of aluminides of bainite phases illustrated Crystal symmetry.256 Index terms Creep resistance (Continued) effect of second phase temperature dependence Cross-slip Crowdions Crystalline-to-amorphous transformation. plasticity affected by Cu-Ga system Cu-Sn-Ni system Cu-Zn system bainite phases/transformations superelasticity 230 228 232 41 230 92 51 237 237 65 240 56 52 97 142 181 221 101 187 199 Links 101 188 201 119 206 215 233 235 D D03 structure APB energies dislocations in alloys fracture of compounds D019 structure dislocations in alloys elastic properties of compounds D022 structure APB energies dislocations in alloys elastic properties of compounds fracture of compounds relation to L12 structure D023 structure fracture of compounds 37 92 79 108 37 92 106 7 92 78 101 6 37 71 92 102 102 114 115 101 106 This page has been reformatted by Knovel to provide easier navigation.

collapse of Defect concentrations. jump mechanisms Directional bonding. boron in AlNi3 Diffusional creep Diffusion coefficients. boron in AlNi3 Diffusion mechanisms. requirement for amorphization Deformation. role of Kear-Wilsdorf configurations in Deformation maps Deformation twinning Density of states (DOS). at grain boundaries Directional solidification techniques Dislocation climb Dislocation cores and alloy composition effect of anisotropic elasticity effect of bond directionality factors affecting mechanical properties affected by properties in various structures and slip systems Dislocation creep Dislocation debris Dislocation locks conservative locks 111 112 112 111 118 99 113 46 122 97 117 This page has been reformatted by Knovel to provide easier navigation. Links 71 13 99 99 111 197 208 116 46 45 177 51 51 166 47 166 148 176 42 46 64 48 64 65 238 105 115 107 108 109 113 49 101 64 115 117 . at grain boundaries Diffusion and creep resistance factors affecting Diffusion activation energy.257 Index terms D023 structure (Continued) relation to L12 structure Debye temperature Decomposition reaction [for dislocation] examples for various structures Defect cascades.

alloys in C11b phases choice of slip directions in D03 alloys in D019 alloys in D022 alloys definitions dissociation of in D022 alloys in L12 alloys static equilibrium configuration in various crystal structures elastic properties affected by in L10 alloys in L12 alloys in L21 alloys in Laves phases mobility experimental determination of non-conservative dislocation observation of fine structure in ordered structures and point defects in quasicrystals sessile simulation of fine structure type-a superdislocations Disorder effects. and antiphase boundaries Disordering. antiphase boundaries formed Dislocations APB-caused in B2 alloys in b.258 Index terms Dislocation mechanisms. irradiation-induced Disorder-order transition antiphase boundaries formed 68 72 68 91 83 95 107 111 114 108 106 101 91 83 103 100 96 99 8 104 95 108 109 91 94 118 93 91 152 109 115 95 106 74 181 193 107 110 11 114 99 114 93 107 Links 114 188 115 12 117 114 117 This page has been reformatted by Knovel to provide easier navigation. .c.c.

factors affecting Ductility enhancement Dumbbell structure [of interstitials] 81 80 139 187 139 220 118 46 27 54 139 28 239 150 142 142 82 Links 184 184 187 187 188 48 186 53 188 54 E EL2 defect [in AsGa] Elastic-compliance-constant matrix Elastic constants ab initio/first-principles calculations calculated values. 152 4 50 4 17 16 152 2 1 3 5 18 4 11 12 . binary intermetallic compounds effect of magnetic transition effect of order-disorder transition effect of point defects first-principles calculation of single crystals Elastic moduli dynamic modulus magnetic-field dependence and melting temperatures of polycrystalline intermetallics static modulus Elastic properties anisotropy in single crystals composition dependence of dislocations 1 17 19 10 1 1 2 14 8 This page has been reformatted by Knovel to provide easier navigation.259 Index terms Disorder-order transition (Continued) antiphase domains APB energies vs transition temperature Displacement energies Displacements per atom Displacement threshold energies Dissociation climb Dorn [dislocation creep] equation Ductility.

260 Index terms Elastic properties (Continued) orientation dependence pressure dependence temperature dependence Elastic stiffness constants calculated values listed measured values listed Electrical resistance. changes during martensitic transformation Electrical resistivity effect of point defects point defects and point defects studied Electronegativity difference. Links 5 9 18 15 2 4 3 224 140 134 140 37 201 203 132 184 203 193 184 185 216 217 152 139 149 38 204 216 139 145 150 183 5 10 19 23 6 7 8 93 33 68 2 142 95 187 65 108 158 76 4 176 93 79 108 141 . embrittlement affected by Electron irradiation amorphous phases produced amorphous phases produced by defects produced effects compared with those of ion irradiation order-disorder transformation affected by recovery after Electron microscopy high-resolution dislocations grain boundaries weak-beam (WB) technique Embedded-atom method (EAM) applicability dislocations simulated by irradiation effects modeled using Embrittlement environmental by hydrogen 57 57 This page has been reformatted by Knovel to provide easier navigation.

A15 superconductors affected by Fe-Ni system Fe-Pt system Ferromagnetic martensite phase effect of magnetic field on transformation temperatures magnetic field induced Ferrous alloys. effects Flinn’s model Flow stress at low temperatures temperature dependence Flow-stress anomaly in B2 alloys in B2 phases and Kear-Wilsdorf locking in L10 alloys in L12 alloys locking of screws This page has been reformatted by Knovel to provide easier navigation. dislocation mobility measured using 57 27 93 95 65 35 34 50 37 Links 38 F Fast-neutron irradiation.261 Index terms Embrittlement (Continued) by impurities Enthalpy of phase formation Environmental factors ductility affected by fracture behavior affected by Eshelby twist Etch-pitting techniques. 198 224 223 244 244 229 201 167 162 95 194 118 121 43 44 95 64 116 104 44 99 118 45 64 108 53 95 215 168 245 246 . irradiation-induced amorphization of Field ion microscopy (FIM) observations with atom-probe analysis (FIM-AP) grain boundaries Finnis-Sinclair method Fission-neutron irradiation. bainite transformations in Fe-Zr compounds.

applications 2 79 2 4 89 4 120 119 184 25 34 26 27 32 26 28 37 27 48 1 96 182 209 139 183 210 34 28 35 33 37 Links 36 188 73 77 78 G Gallium arsenide. energy dissipated by Fracture B2 compounds effect of boron and environment and grain size L12 compounds and processing history and quantum-mechanical calculations and stoichiometry Fracture maps Fracture strength Frank criterion Frenkel defect-pairs irradiation-induced amorphization affected by Frenkel-pair resistivity Full-potential linear augmented plane-wave (FLAPW) method.262 Index terms Flow-stress anomaly (Continued) by elongated KW segments by pinning-point formation Focusons. 151 77 85 29 42 34 34 27 . point defects in GaNi3 antiphase-boundary energies APB tubes factors affecting ductility Gas atomization [processing technique] GeNi3 antiphase-boundary energies factors affecting ductility fracture of 89 29 25 This page has been reformatted by Knovel to provide easier navigation. applications Full-potential linear muffin-tin orbital (FLMTO) method.

. elastic properties Giamei dislocation lock Glass-forming techniques Grain boundaries amorphization at atomic configurations bonding at dislocation structures in disordered alloys slip transmission across vacancies at Grain-boundary chemistry AES observations direct observations FIM-AP observations TEM-EDX observations Grain-boundary embrittlement Grain-boundary enrichment factor Grain-boundary hardening elimination of and pest effect Grain-boundary sliding Grain-boundary structure direct observations by FIM by TEM effects of boron doping modeling of computed atomic models geometric/crystallographic models properties interpreted stoichiometric effects Grain size tensile properties affected by 55 163 163 163 166 34 166 167 168 168 47 157 158 162 158 174 170 172 170 167 172 177 177 176 176 169 55 37 177 177 167 206 170 175 160 32 32 175 171 8 117 182 Links 177 38 57 This page has been reformatted by Knovel to provide easier navigation.263 Index terms GeNi3 single crystal.

elastic constants affected by 93 33 11 120 2 65 42 134 139 140 57 57 18 19 136 145 158 12 176 13 32 167 172 64 134 211 176 215 216 55 I Initial flow stress.264 Index terms Grain size (Continued) yield strength affected by Griffith’s relation Growth kinetics. grain-boundary Hard-sphere models Harper-Dorn creep Heat capacity. elastic constants estimated using Hirsch [dislocation-locking] model Hooke’s law Hot forming processes Hot isostatic pressing (HIP) Huang X-ray scattering self-interstitial structure determined by Hume-Rothery rules Hydrogen embrittlement plasticity affected by Hydrostatic pressure. . effect of order Intergranular fracture examples meaning of term Interstitial clusters 43 25 25 25 145 This page has been reformatted by Knovel to provide easier navigation. at grain boundaries High-resolution electron microscopy observations dislocations grain boundaries Hill’s approximation. bainite 33 1 233 Links H Hall-Petch relationship Hardness. effects Heteropolarity. point defects affecting Heavy-ion irradiation.

. elastic properties Irradiation alloy stability during amorphization induced by dose required experimental observations mechanisms temperature effects collision sequence resulting disordering induced by effects of theoretical modeling ordering enhanced by Ising model Isochronal annealing treatments 211 181 201 198 207 203 151 181 181 187 189 72 183 76 193 205 198 202 8 161 34 16 13 199 203 196 216 131 132 132 132 189 228 223 222 47 182 182 149 138 139 Links 150 217 215 205 217 218 79 This page has been reformatted by Knovel to provide easier navigation.265 Index terms Interstitial defects migration properties production of structure Interstitial energies In-Ti system Invariant line strain Invariant plane strain Inverse creep Ion-beam mixing Ion implantation Ion irradiation effects on amorphization compared with electron-irradiation effects on disordering Iron aluminides elastic properties grain boundaries in tensile properties Iron silicides.

.266 Index terms Isothermal elastic constants compared with adiabatic elastic constants 1 4 Links J Johnston-Gilman etch-pitting techniques 95 K Kear-Wilsdorf (KW) configuration generalized properties role in deformation Kear-Wilsdorf (KW) locks/mechanism and flow-stress anomaly locking of screw dislocations by Khantha-Cserti-Vitek (KCV) [dislocation-locking] model Kinchin-Pease function Kinetics amorphization bainite growth martensitic transformation Kurnakov compounds 202 233 225 54 101 116 115 116 44 101 116 120 119 189 84 89 97 100 115 L L10 type compounds APB energies crystal structure dislocations in debris resulting flow-stress anomaly in fracture of vacancy migration L12 type compounds antiphase boundaries in APB-coupled dipoles 69 123 70 124 71 78 92 104 123 104 37 147 104 114 117 This page has been reformatted by Knovel to provide easier navigation.

applications Linearized muffin-tin orbital (LMTO) method. . for various irradiation temperatures 109 17 209 2 38 77 10 112 190 78 27 117 191 4 79 92 108 56 18 68 75 26 71 95 123 7 44 26 27 158 175 170 54 146 8 99 37 69 92 99 77 Links 89 92 102 114 117 118 79 This page has been reformatted by Knovel to provide easier navigation.267 Index terms L12 type compounds (Continued) APB energies variation with temperature crystal structure relation to D022 and D023 structures dislocations in debris resulting elastic properties flow-stress anomaly in fracture of fracture modes grain boundaries bonding at grain-boundary structure models trialuminides vacancy migration in L21 type compounds APB energies crystal structure dislocations in Lamellar microstructure Lattice softening Laves phases dislocations in temperature dependence of elastic properties Lindemann [melting] criterion Linear augmented plane-wave (LAPW) method. applications Line compounds Lomer-Cottrell dislocation locks Long-range degree. applications Linear augmented Slater-type orbital (LASTO) method.

grain-boundary hardening in Magnetic field induced martensite Magnetic ordering. synthesis using Mechanical properties Mechanical testing 229 221 223 224 225 237 226 225 222 238 237 237 237 240 237 42 1 42 This page has been reformatted by Knovel to provide easier navigation. Links 168 244 17 246 246 175 222 239 244 65 221 225 224 169 256 226 228 .268 Index terms M Magnesium-silver system. elastic moduli affected by Magnetostatic effect Magnetostriction Many-body interactions Martensites crystallography and ductility formation of magnetic field induced stress-induced meaning of term nucleation of Martensitic transformations in ceramics crystallography definitions electrical resistance changes during kinetics relationship with massive transformations and shape-memory effects stress.and strain-induced theories Massive-like transformations Massive transformations definition in intermetallics and optical-memory applications relationship with martensitic transformations Mechanical alloying.

269 Index terms Melting temperatures. antiphase boundaries Neutron irradiation. and elastic constants Melt spinning [processing technique] Memory effects Microhardness testing Microstructures properties affected by effect of second phases Miedema’s model Miedema’s vacancy-formation relationship Models amorphization bainite formation brittle fracture dislocation-locking flow-stress anomaly grain boundaries irradiation damage triple-defect Molecular dynamics (MD) calculations defect structures formation irradiation-induced amorphization Molybdenum disilicide dislocations in elastic properties 111 11 13 189 207 210 210 235 37 119 118 170 190 131 138 177 54 55 141 140 65 19 42 226 167 Links 144 147 22 N Nabarro-Herring creep Nearest-neighbor models. effects Nickel aluminides antiphase-boundary energies variation with temperature boron at grain boundaries brittle-to-ductile transition 77 75 165 41 78 89 47 72 193 194 203 49 64 This page has been reformatted by Knovel to provide easier navigation. .

point defects in Nuclear fission gas bubble formation. .270 Index terms Nickel aluminides (Continued) commercial developments creep behavior creep resistance dislocations in ductility improvement elastic properties factors affecting ductility fracture of irradiation effects plastic deformability point defects in TEM/HRTEM observations of grain boundaries tensile properties vacancies in vacancy formation in Nickel silicides. factors affecting ductility Niobium disilicide. factors affecting Nuclear fuel applications Nuclear reactor structural materials Nucleation bainite martensite 232 225 58 46 56 94 28 7 28 25 184 42 187 158 34 136 138 29 22 28 237 226 226 151 205 181 181 206 34 23 30 240 229 229 206 45 188 159 8 99 64 Links 100 11 111 13 123 160 O O (orthorhombic) phase Al-Nb-Ti system dislocations in structure 110 110 92 This page has been reformatted by Knovel to provide easier navigation. elastic properties Ni-Si-Ti system Ni-Sn system Nitinol alloy Ni-Ti system Nitrides.

271 Index terms O-lattice [grain-boundary structure] model One-way shape-memory effect Optical-memory alloys Order-disorder transition at grain boundaries elastic moduli affected by Ordering irradiation-enhanced spinodal Ordering energy. locking of screw dislocations by Plastic deformation controlling factors composition and stoichiometry crystal symmetry environmental factors microstructural features point defects factors controlling bond characteristics 52 51 57 54 152 49 49 65 65 65 65 65 119 55 1 210 28 50 50 222 119 41 57 168 169 120 This page has been reformatted by Knovel to provide easier navigation. . grain-boundary structure studied using Orowan equation Orthorhombic phase see O…phase Oxidation resistance. oxide layers providing Oxide dispersion-strengthened (ODS) alloys 42 55 189 193 172 119 162 16 170 226 240 227 Links P Paidar-Pope-Vitek (PPV) [dislocation-locking] model Particulate-reinforced alloys Peierls stress Percolation model for amorphization Pest degradation effect/phenomenon Phase enthalpy and Young's modulus Phenomenological theory of martensite crystallography Pinning.

272 Index terms Plastic deformation (Continued) composition and stoichiometry crystal symmetry environmental effects microstructure and effects of second phases mechanisms diffusional creep dislocation creep slip twinning Point defects in A15 compounds accumulation during irradiation clustering of concentration required for amorphization and dislocations and electrical resistivity experimental investigations formation properties computation of experimental determination of theory and lattice parameter low-temperature retention migration properties experimental determination and ordering energies physical properties affected by production of recovery of after irradiation or quench in semiconductors theoretical modeling theoretical modeling of thermodynamics 141 132 140 133 145 146 134 136 152 132 136 145 151 187 141 142 This page has been reformatted by Knovel to provide easier navigation. Links 52 51 57 54 43 47 46 43 45 131 150 207 136 208 152 149 132 183 64 64 64 64 137 .

amorphization caused by Pseudoelasticity Pseudosymmetrical [atomic] configurations 42 42 46 18 47 42 28 207 228 170 171 53 19 151 131 131 4 15 5 13 12 16 10 133 42 136 13 23 6 138 Links 144 147 8 152 64 Q Quantum-mechanical calculations boron bonding phase-formation enthalpies Quasicrystals. temperature dependence of elastic properties 181 17 This page has been reformatted by Knovel to provide easier navigation. vacancies studied using Powder metallurgy techniques hot isostatic pressing reactive sintering Power-law creep Pressure effects.273 Index terms Point defects (Continued) in transition-metal carbides and nitrides triple-defect model typical examples Poisson's ratio composition dependence orientation dependence of polycrystals estimation from single-crystal constants temperature dependence Polycrystalline intermetallics. elastic moduli Positron annihilation techniques. dislocations in 37 50 109 110 R Radiation damage Rare-earth compounds. . elastic constants Primary creep Processing ductility affected by Proton irradiation.

274 Index terms Reactive consolidation/sintering Rearrangement of atoms. 195 44 119 47 29 75 147 146 147 99 31 34 165 166 94 . in shape-memory alloys Rupture length Ruthenium aluminides antiphase-boundary energies ductility elastic properties effect of alloying elements Ru11Ta9 elastic properties effect of alloying elements RuTi elastic properties effect of alloying elements 13 15 15 13 15 15 78 36 13 15 15 42 193 193 139 193 185 22 11 228 41 23 12 194 Links 23 S Saturation magnetization. effect Replacement-collision sequence mechanism Resonance ultrasound spectroscopy (RUS). elastic constants determined by Reuss's approximation Rubber-like behavior. mechanisms for Recombination mechanisms Relaxation volumes. interstitial defects Replacement collisions. effects of fission neutrons Screw dislocations locking of Secondary creep Segregation at grain boundaries. of boron to antiphase boundaries Self-diffusion activation energy Self-diffusion coefficients Self-interstitials migration of 187 This page has been reformatted by Knovel to provide easier navigation.

point defects in Sessile dislocations Shape-memory alloys ferromagnetic compounds crystal structure effect of magnetic field on transformation temperatures magnetic field induced martensite magnetization behavior transformation temperature transformation temperatures stress-strain diagrams for superelasticity Shape-memory effects and martensitic transformations one-way memory processes involved rubber-like behavior superelasticity two-way memory Shear modulus changes during amorphization composition dependence and melting temperature of polycrystals estimation from single-crystal constants temperature dependence Shock densification processes Shockley partials Simulation methods. Links 139 151 115 226 242 242 244 244 246 247 242 228 228 226 226 226 227 228 228 226 4 204 15 19 13 12 16 42 93 95 4 99 100 101 13 23 227 229 229 245 246 .275 Index terms Self-interstitials (Continued) structures determined Semiconductors. dislocation fine structure Single crystals anisotropy of elastic properties elastic constants 2 1 This page has been reformatted by Knovel to provide easier navigation.

Links 42 114 43 113 9 119 53 133 193 131 106 104 27 43 225 55 55 55 57 225 239 228 44 45 228 229 226 71 226 228 105 52 55 226 65 150 186 13 64 . shape-memory alloys Stress-temperature diagrams. and antiphase-boundary energies Structural vacancies detection/determination techniques empirical prediction formation of 132 140 137 This page has been reformatted by Knovel to provide easier navigation. ductility affected by Strain-hardening behavior Strain-induced martensitic transformation Strength effect of APBs effect of second phase and grain size Stress corrosion cracking Stress-induced martensitic transformation and ductility improvement Stress-strain diagrams Cu-Zn system effect of order serrations in superelasticity loop Stress-strain-temperature diagrams.276 Index terms Sintering. elastic properties Sodani-Vitek (SV) [dislocation-locking] model Solid-solution hardening Specific gravity. reactive sintering Slip directions Slip mechanisms Slip systems SnTi3 single crystal. martensitic transformation Structural stability. and lattice parameter Spinodal ordering Split interstitials Stacking faults. in L10 alloys Staircase-configuration dislocations Stoichiometry.

in dissociation of dislocations Superlattice stacking fault (SSF) dipoles Superplastic forming Superplasticity Supertwinning Surface energies. segregation to grain boundaries Superalloys. effect of point defects Superconductors. antiphase boundaries Swelling behavior irradiation-induced Symmetrical (atomic) configurations Symmetrical tilt boundaries Synthesis techniques 84 97 123 55 65 239 68 181 205 170 172 42 171 173 76 65 98 46 34 182 150 182 43 95 228 43 106 72 100 93 123 152 198 93 55 Links T Takeuchi-Kuramoto [dislocation-locking] model Takezawa-Sato [bainite transformation] model Tensile strength Al-Fe alloys Al-Ni alloys brittle-to-ductile transition Thermal expansion. vacancy-affected 35 34 49 133 119 235 This page has been reformatted by Knovel to provide easier navigation.277 Index terms Subgrain formation Sulfur. nickel-based Superconducting critical/transition temperature. . A15 compounds as Superdislocations fine structure in relation to crystal structure Superelasticity Super-lattice dislocations Superlattice extrinsic stacking-fault (SESF) dipoles Superlattice intrinsic stacking fault (SISF) and Superdislocations Superlattice intrinsic stacking-fault (SISF) energies and antiphase boundaries and APB energies.

point defects in Threshold displacement energies see Displacement threshold energies Tilt boundaries. atomic structures Time – temperature-transformation (TTT) diagrams Titanium aluminides antiphase-boundary energies commercial developments creep behavior dislocations in elastic properties massive-like transformations plastic deformability TEM/HRTEM observations of grain boundaries tensile strength Toughness improvement Trace-element alloying additions Transformations applications to materials ductility improvement optical-memory applications bainite crystallographic martensitic massive massive-like Transition-metal compounds bonding characteristics point defects in Transmission electron microscopy (TEM) observations antiphase boundaries/domains 68 76 This page has been reformatted by Knovel to provide easier navigation.278 Index terms Thermal spike model Thermoelastic martensitic transformation III-V compounds. Links 193 225 151 172 229 77 58 64 102 117 6 238 49 159 49 57 28 239 239 240 229 221 223 237 238 50 151 69 80 73 73 74 29 104 13 105 22 106 23 115 173 230 78 198 211 .

175 136 145 148 149 149 137 140 . elastic properties Twin boundaries Twinning Two-way shape-memory effect 166 158 196 131 229 5 159 45 226 197 138 230 13 161 64 Links 206 144 215 147 221 238 U Ultrasonic wave velocity Ultrasonic wave velocity measurement.279 Index terms Transmission electron microscopy (TEM) observations (Continued) with energy-dispersive X-ray analysis (TEM-EDX) grain boundaries irradiation effects Triple-defect model TTT diagrams Tungsten silicides. irradiation effects 1 1 205 10 V Vacancies at grain boundaries dislocations as sources/sinks formation of energies predicted volume change and Hume-Rothery rules migration properties A15 compounds B2 compounds B11 compounds L10 compounds L12 compounds mobility and off-stoichiometry structural vacancies thermal vacancies 131 48 48 132 140 133 140 132 151 147 147 147 146 146 144 132 133 This page has been reformatted by Knovel to provide easier navigation. elastic properties determined by Uranium intermetallic compounds.

grain-boundary structure affected by Vidoz-Brown dislocation mechanism Voigt approximation Voigt notation Von Mises criterion 146 143 143 138 176 84 11 2 51 86 12 144 Links W Widmanstätten ferrite Wrong-pair bonds 229 171 Y Yield anomaly. irradiation effects 246 190 181 196 203 199 209 200 207 This page has been reformatted by Knovel to provide easier navigation. and APB energy anisotropy Yield strength Al-Fe alloys Al-Ni alloys Young's modulus composition dependence orientation dependence and phase enthalpy of polycrystals estimation from single-crystal constants temperature dependence 35 34 4 14 5 50 13 12 16 13 17 23 15 6 8 70 Z Zeeman effect Zee-Wilkes [irradiation-induced ordering/disordering] model Zircaloy alloys Zirconium aluminides. .280 Index terms Vacancy clusters Vacancy concentration Vacancy-energy model Vacancy-ordered phases Valency difference.