Professional Documents
Culture Documents
Delamination in Wood,
Wood Products
and Wood-Based Composites
123
Editor
Dr. Voichita Bucur
CSIRO
Clayton Laboratories
Materials Science and Engineering
Bayview Avenue
3168 Clayton Victoria
Australia
voichita.bucur@csiro.au
Cover image: Intra-ring internal checking in sample (100 × 50 mm – width × thick) of regrowth
Victorian Ash (Eucalyptus delegatensis or E. regnans). Photo taken by Philip Blakemore.
It is with great pleasure that I prepare this foreword. The senior author, Professor
Voichita Bucur, is one the preeminent wood scientists in the world today. She is
well known for her excellent research on acoustics, especially the acoustic prop-
erties of wood and wood-based materials. Her previous books Acoustics of Wood
and Nondestructive Characterization and Imaging of Wood are outstanding ref-
erence documents; they provide a summary of much of the world’s research and
development efforts in these two important technical areas.
Professor Bucur has contacted widely respected technical authorities and asked
them to prepare chapters dealing with various aspects of the formation and detection
of separations and delaminations in wood-based materials.
P. Blackmore – CSIRO Australia, S Blumer Holzinnovationzentrum, Austria,
G Daian Melbourne University, Australia, BSW Dawson – SCION New Zealand,
F Divos – Faculty of Wood Science Sopron, Hungary , L. Donaldson – SCION
New Zealand, T. Gereke ETH Zürich, Switzerland, P.J. Gustafsson Lund University
Sweden, N. Haque – CSIRO Australia, CL Huang – Weyerhauser USA, S. Kazemi-
Najafi – Tarbiat Modares University, Iran, C. Mueller – ETH Zürich, Switzerland,
J. Neuenschwander – Empa Switzerland, P. Niemz – ETH Zürich, Switzerland,
K. Persson – Lund University Sweden, M.S.J. Sanabria Empa, Switzerland,
U. Sennhauser Empa, Switzerland, A. P. Singh – SCION New Zealand, all gra-
ciously agreed and provided excellent technical contributions.
This book is organized into three parts. Part I, General Aspects, presents much
needed basic information, including terminology, the theoretical basis for evalu-
ation of delamination in wood and wood-based materials, and mechanical stress
development in the woody cell wall in response to various stressors. A vibration-
based approach is proposed to evaluate delamination with ultrasonics or with low
frequency vibrations. Crack initiation and growth of delamination is studied with a
fracture mechanics approach. A theoretical model for collapse recovery is proposed.
Part II, Methodology for Delamination Detection and Factors Inducing and
Affecting Delamination, begins by examining a variety of methods for detecting
delamination in wood products, then delves into discussion of the formation of
delamination or separations at several levels – from the microscopic, anatomical
level within solid wood sections to examination of the interface of wood and surface
v
vi Foreword
coatings. The techniques presented for observing separations include confocal laser
scanning microscopy, light microscopy, scanning electron microscopy and ultra-
sonics. Excellent discussions of delamination caused by moisture induced stresses,
including those that form during the drying of wood and lumber products and those
observed with weathered wood surfaces, are included.
Part III, Delamination in Different Products, focuses on practical aspects of
delamination in a wide range of wood products. An excellent discussion of the
industry’s perspective is presented. Practical discussions dealing with the forma-
tion, detection, and performance problems associated with delamination in trees,
logs, laminated panels, composites, glued laminated timbers, and parquet floors are
presented in detail.
The authors prepared this book to serve as a primary reference on subject of
delamination in wood-based materials and products. It was prepared to provide a
concise source of information on the topic to manufacturers and users of wood
products, as well as research scientists. It was made possible through the efforts
of dedicated scientists who spent countless hours in laboratories developing techni-
cal information on this important subject. This book is a tribute to their efforts and
a significant contribution.
This book is a significant contribution to the wood science and technology liter-
ature. Professor Bucur has completed another significant contribution to the wood
science literature.
Delamination occurs in all man made composite materials as well as in natural com-
posites like wood, bones or rocks. Many groups of specialists with widely different
backgrounds and interests need knowledge of factors influencing delamination in
wood, wood products and wood based composites. I was amazed with the lack of
information on the subject and particularly with the way in which the available infor-
mation is scattered in the literature. Out of this amazement arose the idea to write
and edit this book.
Part I of the volume deals with general aspects of delamination, the terms used
for defining delamination in wood science and technology and with the theoretical
aspects in the evaluation of delamination. Part II is directed at the methodology
developed for delamination detection. Factors that induce and affect delamination
are analyzed. Part III is a study of delamination in different products. Extensive
reference is made to the literature. An attempt has been made to select the most
important references for the corresponding chapter. Thus, for any given topic, it
should be easy for the reader to quickly acquaint himself with what has been done
by looking up the listed references. It is also the hope of the authors that this volume
will be a valuable source of information for the practitioner who mostly deal with
the design or evaluation of structures subjected to delamination.
In recent years manufacturers are becoming more aware of the importance of
delamination and other factors that affect the performance of their finished products.
Thus there is an evident need for this type of book.
Experts called upon to render opinions on structure safety are faced with not
only the daunting task of discovering and quantifying structural defects such as
delamination, but also translating those observations into the probability of failure
and determining levels of “unacceptable risk”. Even though the mechanics of wood
failure is better understood today than two decades ago, and the tools for nonde-
structive identification of defects are more accurate and powerful, the fact remains
that deciding what level of defect represent an “unacceptable risk” continues to be a
subjective judgment. This is particularly true for structures with significant but not
severe defects such as delamination and on sites that present high levels of risk (i.e.
snow).
The bibliography of this book is intended to be comprehensive and we hope, an
important contribution of this book (near 1000 references) is to accurately identify
vii
viii Preface
the evolution of ideas in the last decades. All references cited in the text are included
in the reference section at the end of each chapter.
At present, no comparable book exists covering the entire field of delamination
in wood, wood products and wood based composites.
The editor would like to firstly acknowledge the contributions by colleagues act-
ing as authors of the individual chapters, who gave their time and energy to prepare
this excellent text.
I would like to express my sincere thanks to all colleagues and organizations that
have made possible the publication of this volume, to the CSIRO – Commonwealth
Scientific and Industrial Research Organisation – Australia and SCION- Forest
Research Institute, New Zealand, who supported this idea. In preparing such a text
it is very difficult to acknowledge all the help given to the editor. I am indebted
to the three main scientific communities, wood science, mechanical and acousti-
cal communities who have undertaken research and development that is reflected in
the cited publications. This book encompasses a variety of recent research result, a
number of unpublished results and refinement of older material.
This book would certainly not have been possible without the help of my col-
league Nick Ebdon, CSIRO – Clayton, who work very hard on the preparation and
formatting all figures.
Last but not least, I would also thank my family and my Australian friends who
followed with interest and enthusiasm the progress of the manuscript of this book.
Working for this book was for me an extraordinary opportunity to discover the
natural splendors of Australia and the atmosphere of this country, which is a proud
modern civilization.
Permission for the figures cited in this book have been granted by Copyright
Clearance Center, (http://www.copyright.com), by different organisations and col-
leagues cited in the corresponding chapters of this book. The authors are very
thankful for their kind permission to reproduce figures.
As editor of this book, I own special thanks to Ms Danila Durante, Information
Specialist, CSIRO Australia, Information Management & Technology Division,
in Melbourne for numerous hours spent together for copyright permissions with
the new electronic system required by Copyright Clearance Center. Many, many
thanks are also addressed to Ms Bee Thia, Information Specialist, CSIRO Australia,
Information Management & Technology Division, for her continuous and enthusi-
astic help in collecting documents and books cited in this volume.
ix
Contents
xi
xii Contents
xiii
xiv Contributors
Saeed Kazemi–Najafi Wood & Paper Science & Technology Department, Tarbiat
Modares University, Noor, Iran, skazemi@modares.ac.ir
Christian Müller Institute for Building Materials, Wood Physics, ETH Zürich,
Schafmattstrasse 6, CH-8093, Zürich, Switzerland, muellerc@ethz.ch
Jürg Neuenschwander Electronics/Metrology/Reliability Laboratory, Swiss
Federal Laboratories for Materials Science and Technology, Empa,
Überlandstrasse 129, CH-8600, Dübendorf, Switzerland,
juerg.neuenschwander@empa.ch
Peter Niemz Institute for Building Materials, Wood Physics, ETH Zürich,
Schafmattstrasse 6, CH-8093, Zürich, Switzerland, niemzp@ethz.ch
Kent Persson Division of Structural Mechanics, Lund University, P.O. Box 118,
SE-221 00 Lund, Sweden, kent.persson@construction.lth.se
Robert J. Ross Forest Products Research Laboratory One Gifford Pinchot Drive
Madison, Madison, WI 53726, USA, rjross@fs.fed.us
Sergio J. Sanabria Electronics/Metrology/Reliability Laboratory, Swiss Federal
Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129,
CH-8600, Dübendorf, Switzerland, sergio.sanabria@empa.ch
Urs Sennhauser Department of Electronics/Metrology/Reliability Laboratory,
Swiss Federal Laboratories for Materials Science and Technology, Empa,
Überlandstrasse 129, CH-8600, Dübendorf, Switzerland, urs.sennhauser@empa.ch
Erik Serrano University of Vaxjo, Lucklings plats 1 SE 35195 Vaxjo, Sweden,
erik.serrano@vxu.se
Adya P. Singh Wood and Biofibre Technologies, Scion Te Papa Tipu Innovation
Park, Rotorua 3010, New Zealand, adya.singh@scionresearch.com
Part I
General Aspects
Chapter 1
Introduction
Voichita Bucur
Contents
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Solid Wood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Wood-Based Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.1 Background
In order to improve the quality of mass produced wood-based composites and in
order to undertake quality assessment of adhesive interfaces in these materials it
is first necessary to develop the theoretical basis describing both qualitatively and
quantitatively, the quality parameters of the composite, and secondly to develop new
non-destructive techniques for their testing and evaluation.
Mechanical integrity of interfaces in wood-based composites plays a major role
in determining the serviceability of structures and their components. New advanced
materials (i.e. parallel-strand lumber, laminated veneer lumber, etc.) are designed
with specialty interfaces to increase fracture resistance of wood-based compos-
ite materials and to accommodate residual stresses. Of particular note is that the
mechanical properties of wood-based composites, used mainly in civil engineering,
may degrade severely in the presence of damage, often with tragic consequences.
Therefore damage detection is a very important issue in the context of structural
health monitoring for mechanical engineering infrastructure with elements in wood
and wood-based composites.
Wood-based composites are complex materials exhibiting important anisotropic
properties. Commonly observed damage in these materials are: delamination
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
Fig. 1.1 Delamination observed on cross sections of Douglas-fir laminated lumber. Note delam-
ination occurs predominantly in wood elements in the direction of medulary rays, frequently
starting or finishing at the interface between the earlywood and latewood (Vick and Okkonen 2000,
Figure 5a)
Fig. 1.2 Delamination in spruce latewood tracheids between S1 and S2 layers (Zimmermann
et al. 1994, Figure 3)
1 Introduction 5
As noted by Garg (1988) many years ago, the prediction of delamination in com-
posites is a challenge for both scientists and manufacturers. This is due to the large
number of parameters involved in the design of composites and, on the other hand,
to the complexity of the stress state which leads to the initiation and propagation of
delamination.
For the initiation aspect of delamination, the tolerance prediction is based on
semi-empirical criteria, such as point-stress or average stress criteria. Due to the
use of such criteria, industries are led performing numerous tests in order to ensure
the safety margins for delamination failure are not exceeded. The non-propagation
certification relies on fracture mechanics analyses, which are very complex and
introduce difficulties for the characterization of the initial delamination pattern
(Srinivasan 1996; Murata and Masuda 2006).
The last 30 years there have been several important advances toward a better
understanding of the mechanics of laminated composites and of the damage mecha-
nisms, because of their intensive utilisation in aerospace engineering. This progress
concerns the analysis and identification on the micro, macro and meso scales, as
well as the development of advanced anisotropic material models. To be able to rely
on computational models, both academics and manufacturers recognize that a pre-
requisite is to develop a detailed material model with a clear identification procedure
and to validate this model by means of representative experimental tests.
The physics of delamination is governed by interactions among different damage
mechanisms, such as fibre breakage, transverse microcracking and debonding of
the adjacent layers of the cell wall. To understand the physics of delamination in
composite biological materials and more specifically in wood, wood based products
and wood-based composites, it is necessary to have detailed knowledge about the
microstructure of these materials.
As noted by Kelly (1989) in the Concise Encyclopedia of Composite Materials,
“plant cells are a good example of laminated composite material; the shape of the
cells is roughly tubular with various laminae of cellulose microfibrils glued together
to form a wall. Each lamina has a characteristic fibre orientation which can be
random, cross-helical or single-helical. . . . . . These biomaterials are grown under
stress; this means that the loading conditions of the structure as a whole can be
used effectively as blueprints for the most efficient use of fibre reinforcement. By
their very nature, natural fibrous composites are better materials in tension than in
compression and their use in many applications is often limited by this fact. The
excess of tensile strength available can be profitably used to pre-stress in tension
the regions of the structure which are more vulnerable into compressive loads. Also
the presence of water as compression members will result in lighter structures”.
a complex aggregation of cells, of tubes shape, which during the life of the tree had
biological function. The structural features of wood are oriented following the prin-
cipal directions of growth of the tree, namely longitudinal-parallel to the axis of the
tree, radial and tangential – versus the annual rings”
Several models have been proposed to represent wood structure in relation to its
mechanical behaviour, starting with Price (1929), who modelled the cell structure
as an array of parallel cylindrical tubes, of isotropic structure, oriented in the stem
direction.
Another version, proposing also a tubular model, useful for modelling the cell
wall as a laminated composite material is presented in Fig. 1.3. A softwood or
conifer wood cell is essentially a hollow tube of about 30 μm diameter with a multi
layered laminated wall composed generally from four layers – primary wall, S1 , S2
and S3. The S2 layer, is the principal load bearing component of the cell wall and
is close to 80% of the total cell wall area. It contains cellulose components in the
form of microfibrils of about 10–20 nm in diameter. In most cases the microfibrils
lie at an angle to the cell axis and form a steep helix at an angle, ranging between
0◦ and 25◦ and 0◦ and 50◦ for hardwood and softwood respectively. Fibres with low
microfibril angle (10◦ ) posse high tensile strength (400 MPa) and low elongation
(1%). The cells are parallel to the grain direction and are bonded to each other by
an amorphous matrix containing mostly lignin. Nearly 90% of the cells are aligned
in one direction forming a honeycomb structure with highly anisotropic mechanical
properties. The alternation of spring and summer growth (earlywood and latewood
layers in the annual ring) in softwood and ring porous hardwood species from tem-
perate climates produces well known ring patterns which introduce a further element
of complexity.
Fig. 1.3 Layered structure of the cell wall modelled as a laminated composite material
(Mark 1967, Figure 1-7)
1 Introduction 7
– the plate-type bending and shear deformations of the cell walls which are dom-
inant in tangential direction, when the transverse shear loading and longitudinal
compression straining are applied on solid wood specimens.
– the solid-type deformations are dominant in longitudinal and radial directions
when longitudinal shearing loading straining are induced on wood specimens.
8 V. Bucur
Fig. 1.4 Schematic representation of the degradation of mechanical properties of wood (expressed
by Young’s modulus) compared to those of pure cellulose (Jeronimidis 1980, Figure 2)
At a cellular scale the plate-like deformation modes were studied combining ran-
dom/periodic multi-step homogenisation with corresponding values obtained from
continuum micromechanics modeling. The average predictive capacity of this model
is low, about 8%, with very large variations depending on the value of the elas-
tic constants. The highest errors were observed on GRT (error can be as high as
290%) and on Poisson’s ratios (error of about 75%). It is very likely that the
1 Introduction 9
Fig. 1.5 Transverse section of spruce parallel-strand lumber which contains fractured and delam-
inated zones. The arrows indicate the zones tested via nanoindentation (Konnerth and Gindl 2006,
Figure 1)
The behaviour of wood cement composites has been reported from the beginning
of there presence on the market, over 70 years ago as low-density and insulation
material. Today the cement bonded structural flakeboards offer high, fire, insect and
fungal resistance. In addition the quality has improved resulting in better weather-
ability and acoustic insulation (Lee et al. 1987; Mosemi and Pfister 1987; Fan
et al. 1999).
References relating to the modelling of mechanical behaviour of fibre-based
composites are as abundant as those for laminated wood-based composites, but only
several are cited here (Smulski and Ifju 1987; Claisse and Davis 1998; Lopez-Anido
et al. 2000; Moulin et al. 1990; Ogawa 2000; Pirvu et al. 2004; Rowlands et al. 1986;
Tascioglu et al. 2003; Tsai et al. 2005; Xu 2002; Xu et al. 2005; Chakraborty et al.
2006). Mechanical properties of fibre-based composites are influenced by factors
such as: fibre geometry, orientation and distribution, fibres packing in flake of differ-
ent orientation, random distribution of flakes, moisture content, adhesive-type, etc.
Single layer flake models and multilayer mat structures were suggested (Bodig and
Jayne 1982; Steiner and Dai 1993; Dai and Steiner1994; Lenth and Kamke 1996)
to explain the mechanical behaviour of fibre based composites. Several authors
(Ogawa 2000; Tascioglu et al. 2003) reported successful utilisation of hybrid fiber-
reinforced polymer composites – glulam products for structural applications in civil
infrastructures such as beams for bridges stringers, panels for bridge and pier decks.
It was noted that these composites are very resistant to delamination tests during
accelerated exposure to wetting and drying (Pirvu et al. 2004)
Mechanical defibering action produces important structural modifications such
as: internal fibrillation observed as a helical wraps of fibres, cell wall delamination,
external fibrillation which is the peeling off of the fibrils from the fibre surface, with
formation of fines, fibrils or fibrillar lamellae attached to the exterior fibre surface
and fibre shortening, depending on the refining conditions, the fibre type – hard-
wood or softwood – and the pulp type – mechanical or chemical. It is appropriate
to mention here that the hydroxyl groups available on the surface of the cellulose
molecule are the prime means by which fibres and cement, or other material used as
matrix, bond together.
The increasing environmental concern about the wastes from wood, wood prod-
ucts, forest waste and construction waste materials has given rise to the development
of new or improved technological processes such as the water vapour explosion pro-
cess. This process rapidly defibrates wood wastes producing a new raw material for
novel wood cement composites (Wei et al. 2004). Figure 1.6 shows the interfacial
zone between cement and wood fibres, with a delamination of the cell wall near the
wood-cement interface.
As noted by Schneider (1994) the development of fibre-based composites test-
ing methodology was encouraged as part of the efforts being made to control the
performance of low cost building materials for use in developing countries.
The renewed interest in producing new composites with wood fibre began almost
inadvertently in 1960, and Australia was a leading country in this field as noted by
Coutts (2005). In the 21st century a great need still remains to improve the durability
of fibre-based products and to study new, cheaper methods of fibre production and
12 V. Bucur
Fig. 1.6 Interface zone between cement and wood fibres, with delamination of the cell wall near
the interface wood-cement. (Wei et al. 2004, Figure 1C)
1.4 Summary
Commonly observed damage in wood products and wood-based composites are:
wood fibre fracture, delamination between plies or debonding of wood–adhesive
layers. Delamination which is probably the most frequently observed damage, may
be produced during manufacturing or, during in service loading such as accidental
excessive loading produced for example by snow or, by fatigue in highly variable
environmental conditions of temperature and humidity. Damage detection in general
and delamination in particular is a very important issue in the context of struc-
tural health monitoring for mechanical engineering infrastructure with elements in
wood and wood-based composites. The development of computational techniques
in the last 25 years, and the progress achieved in mechanical characterisation of
solids in general and of composites in particular, affected positively the develop-
ment of the modelling of wood mechanical behaviour in function of its structure.
Related studies clearly suggest that delamination in solid wood can occur between
different layers of the cell wall at submicroscopic, microscopic and macroscopic
structural levels.With respect to wood-based composites, the behaviour of two
groups of products has been analysed: the laminated products (plywood, laminated
veneer lumber, parallel-strand timber, structural particleboard, oriented strandboard,
etc.) and the fibre-based products (fibreboards, fibres-cement composites, carbon
fibre-reinforced plywood, particleboard, wood and glass-fibre composites). Finite
element modelling of laminated wood composites as a multilayer system was
1 Introduction 13
proposed. More recently analysis of the multiaxial stress states in parallel wood-
strand composites, has been proposed using a 3D non-linear stochastic finite element
model and Monte Carlo simulations. The development of fibre-based composites
testing methodologies must be encouraged as part of the efforts being made to
control the performance of low cost building materials.
References
Astley RJ, Stol KA, Harrington JJ (1998) Modelling the elastic properties of softwood. Part II: the
cellular microstructure. Holz Roh Werkst 56:43–50
Bejo L, Lang EM (2004) Simulation based modelling of the elastic properties of structural
composite lumber. Wood Fiber Sci 36:395–410
Bodig J, Jayne BA (1982) Mechanics of wood and wood composites. Van Nostrand Reinhold
Company, New York, NY
Bucur V (2003) Ultrasonic imaging of wood structure. Proceedings of 5th world confer-
ence in ultrasonics, Paris, pp 299–302. http://www.sfa.asso.fr/wcu2003/procs/webside/artickes.
Accessed 7 September 2004
Castro G, Paganini F (2003) Mixed glue laminated timber of poplar and eucalyptus grandis clones.
Holz Roh Werkst 61:291–298
Chakraborty A, Sain M, Kortschot M (2006) Reinforcing potential of wood p[ulp – derived
microfibres in a PVA matrix. Holzforschung 60:53–58
Claisse PA, Davis TJ (1998) High performance jointing systems for timber. Constr Build Mater
12:415–425
Clouston P (2007) Characterization and strength modelling of parallel strand lumber.
Holzforschung 61:394–399
Clouston P, Lam F (2001) Computational modelling of strand-based wood composites. ASCE J
Eng Mech 127:844–851
Clouston P, Lam F (2002) A stochastic plasticity approach to strength modelling of strand-based
wood composites. Compos Sci Techn 62:1381–1395
Clouston P, Lam F, Barrett JD (1998) Incorporating size effects in the Tsai-Wu strength theory for
Douglas –fir laminated veneer. Wood Sci Techn 32:215–226
Coutts RSP (2005) A review of Australian research into natural fibre cement composites. Cem
Concr Compos 27:518–526
Dai C, Steiner PR (1994) Spatial structure of wood composites in relation to processing and per-
formance characteristics. Part 3. Modelling the formation of multi-layered random flake mats.
Wood Sci Techn 28:229–239
Fan M, Dinwoodie JM, Bonfield PW, Breese MC (1999) Dimensional instability of cement bonded
particleboard : Behaviour of cement paste and its contribution to the composite. Wood Fiber
Sci 31:306–318
Fritsch A, Hellmich Ch (2007) ‘Universal’ microstructural patterns in cortical and trabecular,
extracellular and extravascular bone material: Micromechanics – base prediction of anisotropic
elasticity. J Theor Biol 244:597–620
Garg CA (1988) Delamination. A damage mode in composite structures. Eng Fract Mech
29(5):557–584
Gerrard C (1987) The equivalent orthotropic elastic properties of plywood. Wood Sci Techn
21:335–348
Gibson LJ, Ashby MF (1988) Cellular Solids. Structure and properties. Pergamon, Oxford
Harrington JJ, Booker R, Astley RJ (1998) Modelling the elastic properties of softwood. Part I:
The cell – wall lamellae. Holz Roh Werkst 56:37–41
Hata T, Umemura K, Yamauchi H, Nakayama A, Kawai S, Sasaki H (2001) Design and pilot
production of a spiral winder for the manufacture of cylindrical laminated veneer lumber. J
Wood Sci 47:1105–1123
14 V. Bucur
Park HM, Fushitani M (2006) Effects of component ratio of the face and core laminae on static
bending strength performance of three – ply cross – laminated wood panels made with sugi
(Cryptomeria japonica). Wood Fiber Sci 38:278–291
Persson K (2000) Micromechanical modelling of wood and fibre properties. Ph D thesis. University
of Lund
Pirvu A, Gardner DJ, Lopez-Anido R (2004) Carbon fiber – vinyl ester composite reinforcement
of wood using the VARTM/SCRIMP fabrication process. Compos Part A 35:1257–1265
Price AT (1929) A mathematical discussion on the structure of wood in relation to its elastic
properties. Phil Trans Royal Soc A 228:1–62
Rowlands RE, van Deweghe RP, Laufenberg TL, Krueger GP (1986) Fiber – reinforced compos-
ites. Wood Fiber Sci 18:39–57
Schneider MH (1994) Wood polymer composites. State of the Art review Paper. Wood Fiber Sci
26:142–151
Schniewind A (1981) Mechanical behavior and properties of wood. In: Wangaaed FF (ed) Wood:
Its structure and properties. Educational Modules for Material Science and Engineering Project.
Pennsylvania State University, University Park, Pennsylvania, USA, pp 225–270
Schniewind AP (1989) Concise encyclopedia of wood & Wood-based materials. Pergamon, Oxford
Shaler SM, Blakenhorn PR (1990) Composite model prediction of elastic moduli for flakeboard.
Wood Fiber 22:246–261
Smulski SJ, Ifju G (1987) Flexural behaviour of glass fiber reinforced hardboard. Wood Fiber Sci
19:313–327
Suo S, Bowyer JL (1995) Modeling of strength properties of structural particleboard. Wood Fiber
Sci 27:84–94
Srinivasan AV (1996) Smart biological systems as models for engineered structures. Mater Sci Eng
C 4:19–26
Steiner PR, Dai C (1993) Spatial structure of wood composites in relation to processing and
performance characteristics. Part I. Rationale for model development. Wood Sci Techn
28:45–51
Tascioglu C, Goodell B, Lopez – Anido R (2003) Bond durability characterization of preservative
treated wood and E – glass/phenolic composite interfaces. Compos Sci Techn 63:979–991
Triche MH, Hunt MO (1993) Modelling of parallel-alligned wood strand composites. Forest Prod
J 43(11/12):33–44
Tsai M, Chou HC, Xie YM, Li YF, Lin LD (2005) Study on the accelerated aging of CFRP – wood
composites. Forest Prod J 24(3):237–246
Vick CB, Okkonen EA (2000) Durability of one-part polyurethane bonds to wood improved by
HMR coupling agent. Forest Prod J 50(10):69–75
Vincent JFV, Currey JD (Eds) (1980) The mechanical properties of biological materials.
Cambridge University Press, London
Wang K, Lam F (1998) Robot – based research on three – layer oriented flakeboards. Wood Fiber
Sci 30:339–347
Watanabe U, Norimoto M (2000) Three dimensional analysis of elastic constants of the wood cell
wall. Wood Research. Bull. Wood Res. Institute, Kyoto, 87:1–7
Watanabe U, Norimoto M, Morooka T (2000) Cell wall thickness and tangential Young’s modulus
in coniferous early wood. J Wood Sci 46:109–114
Watanabe U, Fujita M, Norimoto M (2002) Transverse Young’s moduli and cell shapes in
coniferous early wood. Holzforschung 56:1–6
Wei YM, Fujii T, Hiramatsu Y (2004) A preliminary investigation on microstructural characteris-
tics of interfacial zone between cement and exploded wood fiber by using SEM-EDS. J Wood
Sci 50:327–336
Wu Q, Lee JN, Han G (2004) The influence of voids on the engineering constants of oriented
stranboard: A finite element model. Wood Fiber Sci 36:71–83
Xu J, Widyorini R, Kawai S (2005) Properties of kenaf core binderless particleboard reinforced
with kenaff fiber – woven sheets. J Wood Sci 51:415–420
16 V. Bucur
Xu H (2002) Structural characterization of hybrid fiber reinforced polymer – glulam panels for
bridge decks. J Comp Constr 6(3):194–203
Yamamoto H (1999) A model of the anisotropic swelling and shrinkage process of wood. Part I:
Generalisation of Barber’s wood fiber model. Wood Sci Techn 33:311–325
Yamamoto H, Abe K, Arakawa Y, Okuyama T, Grill J (2005) Role of the gelatinous layer on the
origin of the physical properties of tension wood of Acer sieboldianum. J Wood Sci 51:222–233
Zimmermann T, Sell J, Eckstein D (1994) SEM studies on traction – fracture surfaces of spruce
samples. Holz Roh-Werkst 52:223–229
Chapter 2
Terms for Delamination in Wood Science
and Technology
Voichita Bucur
Contents
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
– Check – a separation of wood along the fibre direction that usually extends across
the rings of annual growth, commonly resulting from stress set up in wood during
seasoning.
◦ End check – a seasoning check occurring on the end of a board or other piece
of wood.
◦ Heart check – a check that extends across the growth layers in one or more
directions from the pith toward, but not to, the surface of a piece of wood. A
synonym is pith check
◦ Roller check – a crack in the wood structure caused by a piece of cupped lumber
being flattened between machine rollers
◦ Star check – a heart check in which the separation extends in more than one
direction from the pith
◦ Surface check – a check occurring on the surface of a piece of wood, usually
on the tangential face not extending through the piece.
◦ Through check – a check that extends through a piece of wood, or from a
surface to the opposite or to an adjoining surface.
– Collapse – the flattening of single cells or rows of cells during drying or pressure
treatment of wood, characterized by a caved or corrugated appearance
– Cracks see shake
– Cross Break – a separation of the wood cells across the grain. Such breaks may be
due to the internal stress resulting from unequal longitudinal shrinkage or external
forces.
– Honeycombing – in lumber and other wood products, is the separation of the
fibers in the interior of the piece, usually along the rays. The failures often are
not visible on the surface, although they can be the extensions of surface and
end checks.
– Shake – a longitudinal separation of the wood. Generally two forms of shake
are recognized, although variations and combinations may be used in industrial
definitions
◦ Heart shake – a shake that starts out at or near the pith and extends radially.
Synonyms are heart cracks, rift crack. A heart shake in which several radial
cracks are presented is termed a star shake
◦ Ring shake – shake occurring in standing trees, in the plane of the growth
rings in the outer position of the latewood for partial or entire encirclement
2 Terms for Delamination in Wood Science and Technology 19
The cell wall has a typical layered structure composing three main layers – S1 ,
S2 , S3 – of variable thickness, in the micron (μm) range, composed of cellulosic
microfibrils embedded in an amorphous matrix. Delamination can occur between
layers as well as inside the same layer, and can be produced by growth related
defects in living trees or can be a defect which develop in wood after it has been
cut. Table 2.1 synthesises the terms related to the cell wall structure, describing
wood delamination at the submicroscopic level.
The spectrum of terminology that has been used in profusion in the numer-
ous articles cited in this table need to be put in concordance with the mechanical
approach proposed in Chapters 3 and 4 of this book, for the description of phenom-
ena related to the delamination in wood and wood – based composites. On the other
hand, as noted by Wilkins (1986) the future nomenclature “needs to remain flexible
and include further terms derived from the development of the tools for wood struc-
ture inspection”. One can speculate about the contribution of new technologies for
higher resolution microscopy in relation to wood ultrastructure which influence its
mechanical behaviour.
Table 2.1 Nomenclature of cell wall deformation as referred in publications until 2007
Stress lines or thrust Kucera and Bariska Observable with Wrinkling in the cell wall
lines (1982) SEM towards the lumen, affecting
“cell wall lining”
Telescopic Kucera and Bariska Mainly thick A diversion of cells from their
shortening (1982) walled cells natural axial orientation
Thrust line Kisser and Frenzel Common in most Slight local thickenings of the
(1950) wood species cell walls due to small
deformation of the fibrils.
Pre-slip plane
Thrust lines, slip Wilkins (1986) Observable with May be considered
planes, light morphologically similar
compression microscope or
creases with SEM
Thrust-line or stress Wilkins (1986) Observable only Pre-slip plans deformation,
lines describe only with SEM which are not distinguishable
those slip planes from slip planes when using
not observable SEM .
with light
microscopy
Transverse fracture Sell and High resolution Poly-laminated concentric
surface Zimmermann FE -SEM structure of the cell wall
(1998) layers observed in transverse
surface
Wrinkling of cell Scurfield et al. Involves only the covering lining
wall (1972) the lumina of fibres. It is a
stage after multiple slip plane
formation in the sequence of
events occurring during axial
compression
The standard ASTM D 1101 refers to the integrity of glue joints in structural
laminated wood products for exterior use and employs the term delamination such
as: “Delamination is a term used to express separation of the wood surfaces at the
glue joints. When the separation takes place in the wood member, even though very
close to the glue joint, it is termed wood failure or checking”. Furthermore it is noted
that since glue joints at knots and knotty areas in general are not detectable under
severe exposures, development of delamination at knots should be disregarded and
not included in the measurements or calculations.
Quantification of the delamination effect in laminated panels is noticed in the
following standards:
Fig. 2.1 Position of delamination in layered composite materials. (Bolotin 1996, Figure 1)
Legend: (a) internal delamination, situated within the bulk of the material, can be studied with
conventional fracture mechanics (b) near – surface, or crack – like defect, very often accompanied
by their buckling, can be studied with the theory of elastic stability (c) multiple cracking – crack
like flow affecting the load carrying capacity of the member and the safe life of the structure
2 Terms for Delamination in Wood Science and Technology 27
Fig. 2.2 Delamination in poplar wood, between the middle lamella and the other cell wall layers
or between the middle lamellae LM, S1 and S2 or between S3 and G due to drying. Bar _________:
10 μm (Clair 2001, Figure 69)
28 V. Bucur
Fig. 2.3 Delamination in fibreboards is observed as large void between the fibers Murmanis
et al. (1986, Figure 4) Legend: V = vessel, P = parenchyma, F = fibers, ML = middle lamella
(a) Wet formed hardboards, high density with 0.5% phenol – formaldehyde. The dark granular
material is scattered between cells. White zones are voids. Arrow shows the softened mid-
dle lamella (ML). Microphotograph×4760. (b) Wet formed hardboards, high density with 0.5%
phenol – formaldehyde. Because of the pressure ML is in the fiber lumen. White zones are
voids. Microphotograph×5300. (c) Dry-formed hardboard, high density with 0.5% phenol –
formaldehyde. Parenchyma (P), vessel (V) and fibres (F) are present. White zones are voids.
Microphotograph×3040.
Near surface delamination is situated just near the member surface and is
always accompanied by buckling such as blisters in plywood originating from the
manufacturing process. Its growth is observed as interlaminar damage.
Delamination producing multiple cracking through the whole thickness of the
member, without separation of the layers is typical for seasoning checks in solid
2 Terms for Delamination in Wood Science and Technology 29
wood. In fibrous composites this defect is very frequent and is induced during
the manufacturing process by thermal factors. In structural layered wood-based
composites multiple cracks can be observed in LVL or in glulam structural mem-
bers in service, used in civil engineering (i.e. houses, bridges, sport halls, etc.).
Technological instabilities in the fabrication process, shrinkage produced by thermal
and chemical factors, biological degradation, etc., are sources of initiation of mul-
tiple delaminations. Local instability and crack growth in glued laminated timber
may produce the global instability of large structural components which in extreme
cases may even result in failure of the whole structure with the potential to cause
loss of life.
The implementation of damage detection strategies must be a constant object
of preoccupation for engineers and wood technologists for permanent structural
health monitoring of buildings and structures incorporating wood or wood-based
composites. Reliable information regarding the integrity of the structure can help
in the prognosis of these structures under current environmental conditions and
estimate the remaining useful life of the system.
2.7 Summary
References
American Society for Testing and Materials (2007) Standard terminology relating to wood and
wood-based products. ASTM D 9 – 05. Philadelphia, PA
American Society for Testing and Materials (2007) Standard test methods for evaluating properties
of wood - base fibre particle panel material. ASTM D 1037-06a. Philadelphia, PA
30 V. Bucur
American Society for Testing and Materials (2007) Standard terminology relating to veneer and
plywood. ASTM D 1038- 83 (2005) Philadelphia, PA
American Society for Testing and Materials (2007) Standard terminology relating to wood-based
fibre and particle panel material ASTM D 1554 - 01 (2005) Philadelphia, PA
American Society for Testing and Materials (2007) test methods for structural panels in shear
through the thickness. ASTM D 2719 – 89 (2007) Philadelphia, PA
American Society for Testing and Materials (2007) Standard test method for shear modulus of
wood-based structural panels. ASTM D 3044 – 94 (2006) Philadelphia, PA
American Society for Testing and Materials (2007) Standard test method for toughness wood-based
structural panels. ASTM D 3499 – 94 (2005) Philadelphia, PA
American Society for Testing and Materials (2007) Standard practice for establishing allowable
properties of structural glued-laminated timber (glulam). ASTM D 3737- 07 Philadelphia, PA
American Society for Testing and Materials (2007) Specification for evaluation of structural
composite lumber. ASTM D 5456-06 Philadelphia, PA
American Society for Testing and Materials (2007) Standard test method for surface bond strength
of wood-based fibre and particle panel material ASTM D 5651 – 95a (2002) Philadelphia, PA
American Society for Testing and Materials (2007) Standard guide for evaluating mechani-
cal and physical properties of wood-plastic composites products ASTM D 7031 -04 (2004)
Philadelphia, PA
ASTM D1101 - 97a (2006) Standard Test Methods for Integrity of Adhesive Joints in Structural
Laminated Wood Products for Exterior Use
Bienfait JL (1926) Relation of the manner of failure to the structure of wood under compression
parallel to the grain. J Agri Res 33:183–194
Boatright SWJ, Garrett GG (1983) The effect of microstructure and stress state on the fracture
behaviour of wood. J Mat Sci 18:2181–2199
Bolotin VV (1996) Delaminations in composite structures: its origin, buckling, growth and
stability. Composites: Part B, 27B:129–145
Brush WD (1913) A microscopic study of the mechanical failure of wood. U.S. Depart Agri Rev
Forest Serv 2:33–38
Chafe SC (1977) Radial dislocations in the fiber wall of Eucalyptus regnans trees of high growth
stress. Wood Sci Techn 11:69–77
Clair B (2001) Etudes des proprietes mecaniques et du retrait au sechage du bois a l`echelle de la
paroi cellulaire . PhD thesis Universite de Montpellier II. France
Côté WA, Hanna RB (1983) Ultrastructural characteristics of wood fracture surfaces. Wood Fiber
Sci 15:135–163
Dadswell HE, Langlands I (1934) Brittle heart in Australian timbers: a preliminary study. J Couns
Sci Ind Res Australia 7:190–196
Dinwoodie JM (1966) Introduction of cell wall dislocations (slip planes) during the preparation of
microscopic sections of wood. Nature 212:525–527
Dinwoodie JM (1968) Failure in timber. Part I. Microscopic changes in cell wall structure
associated with compression failure. J Inst Wood Sci 4:37–53
Dill-Langer G, Lutze S, Aicher S (2002) Microfracture in wood monitored by confocal laser
scanning microscopy. Wood Sci Technol 36:487–499
Donaldson LA (1995) Cell wall fracture properties in relation to lignin distribution and cell
dimensions among three genetic groups of radiate pine. Wood Sci Techn 29:51–63
Fruhmann K, Burgert I, Stanzl-Tschegg SE, Tschegg EK Mode I (2003) Fracture behaviour on the
growth ring scale and cellular level of spruce and beech loaded in the TR crack propagation
system. Holzforschung, 57:653–660
Green HV (1962) Compression caused transverse discontinuities in tracheids. Pulp Paper Mag
Canada 63(3):T 155 – T 168
Jacard P (1910) Etude anatomique des bois comprimés. Mitt Schw. Centralanstalt. Forst.
Versuchwessen 10:53–101
2 Terms for Delamination in Wood Science and Technology 31
Keith CT (1971) The anatomy of compression failure in relation to creep – inducing stresses. Wood
Sci 4:71–82
Keith CT (1974) Longitudinal compressive creep and failure development in white spruce
compression wood. Wood Sci 7:1–12
Keith CT, Côté Jr. WA (1968) Microscopic characterization of lip lines and compression failures
in wood cell walls. Forest Prod J 18:67–74
Kisser J, Frenzel H (1950) Mikroskopische Veränderungen der Holzstruktur bei mechanischer
Überbeansprucging von Holz in der Faserrichtung. Schr Österr. Ges. Holzforschung 2:3–27
Kisser J, Frenzel H (1952) Makroscopische und microsckopische Strukturänderungen bei der
Biegebeanspruchung von Holz. Holz Roh- und Werkstoff 10:415–421
Kucera LJ, Bariska M (1982) On the fracture morphology in wood. Part I: A SEM - study of defor-
mations in wood of spruce and aspen upon ultimate axial compression load. Wood SciTechnol
16:241–259
Meyer RV, Leney L (1968) Shake in coniferous woods – an anatomical study. Forest Prod J
18(2):51–56
Morris C (ed) (1992) Dictionary of science and technology. Academic, Sandiego, p 604
Murmanis L, Youngquist JA, Myers GC (1986) Electron microscopy study of hardboards. Wood
Fiber Sci 18(3):369–375
Reiter A, Sinn G (2002) Facture behaviour of modified spruce wood: a study using linear and non
linear fracture mechanics. Holzforschung 56:191–198
Reiter A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood species
under mode I loading perpendicular to the grain. Mater Sci Eng A 332:29–36
Robinson W (1920) The microscopical features of mechanical strains in timber and the bearing of
these on the structure of the cell wall in plants. Phil Trans R Soc 210 B:49–82
Scurfield G, Silva SR, Wold MB (1972) Failure of wood under load applied parallel to grain. A
study using scanning electron microscopy. Micron 3:160–184
Sell J, Zimmermann T (1998) The fine structure of the cell wall of hardwoods on transverse fracture
surfaces. HolzRoh Werkst 56:365–366
Thuvander F, Berglund LA (2000) In situ observations of fracture mechanisms for radial cracks in
wood. J Mat Sci 35:6277–6283
Tschegg EK, Fruhmann K, Stanzl-Tschegg SE (2001) Damage and fracture mechanisms during
mode I and mode III loading of wood. Holzforschung 55:525–533
Vasic S, Stanzl-Tschegg SE (2007) Experimental and numerical investigation of wood fracture
mechanisms at different humidity levels. Holzforschung 61:367–374
Wardrop AB, Dadswell HE (1947) The occurrence, structure and properties of certain cell wall
deformations. J Coun Sci Ind Res Aust 221(5):14–32
Wilkins AP (1986) The nomenclature of cell wall deformations. Wood Sci Technol 20:97–109
Chapter 3
Delamination Detection – A Vibration-Based
Approach
Voichita Bucur
Contents
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Delamination Detection with an Ultrasonic Technique . . . . . . . . . . . . . . . 33
3.3 Delamination Detection with a Model-Based Method . . . . . . . . . . . . . . . 38
3.3.1 Linear Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.2 Nonlinear Behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.4 Some Practical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.1 Introduction
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
bonds. The defects expected to be present at the interface are cracks at the interfaces
of different oriented plies, inter-ply delamination, adhesion weakness at interfaces
between plies or between a ply and an adhesive layer. In all cases the basic pur-
pose of the nondestructive evaluation methods is the determination of the integrity
of bonds. The efficiency of ultrasonic methods is related to the understanding of the
relationship between the measured parameters and the interface mechanical proper-
ties, which is dependent on the theoretical approach used to predict the behavior
of the interface, according to the specific kind of defect expected to be present
Combining the experimental data with the theoretical knowledge (Hirsekorn 2001)
it is possible to gain important information about the linear or non linear interface
behavior (Krohn et al. 2002; Solodov et al. 2002).
In the years 1970–1990 stress was put on the damage identification and health
monitoring of laminated composites through the overall mechanical characteris-
tics of the structure by measuring the stiffness matrix, the viscoelastic parameters,
etc. determined with ultrasonic waves. Theoretical models for plane wave propaga-
tion in layered anisotropic composites were developed in a very impressive amount
of articles and reference books. Because space limitation only several references
has been selected (Green 1985–2006; Chimenti 1981–2006; Bunsell 1988; Nayfeh
and Chimenti 1988; Hosten et al. 1987; Rose et al. 1990; Alleyne and Cawley
1992; Deschamps and Hosten 1992; Rokhlin and Wang 1992; Potel and de Belleval
1993a, b; Saravanos et al. 1994; Lavrentyev and Rokhlin 1998).
In that follows our attention is focused on the ultrasonic method based on Lamb
waves. Lamb waves are defined as mechanical waves corresponding to vibration
modes of plates having the thickness of the same order of magnitude as their
wavelength. Lamb waves are suitable for the nondestructive evaluation of large
structural elements, due to their prominent characteristic - the long range propa-
gation, with low dispersion energy, even in materials with high attenuation ratio.
The Lamb waves are able to put in evidence the presence of defects, as noted
in a very extensive body of literature from which several references has been
extracted (Rokhlin 1979, 1980; Pilarski and Rose 1987; Auld 1980, Chimenti and
Martin 1991; Nagy 1992; Ogilvy 1995; Huber et al. 1997; Cawley and Alleyne
1996; Wright et al. 1996; Kazys R and Svilainis 1997; Maslov and Kundu 1997;
Singer 1997; Delsanto et al.1998; Delsanto and Scalerandi 1998; Kundu et al.1998;
Rokhlin and Wang 1998; Royer and Dieulesaint 2000; Hayashi and Kawashima
2002; Kessler et al. 2002a; Stoessel et al. 2002; Su et al. 2002; Sohn et al. 2004;
Simonetti 2004; Shkerdin and Glorieux 2004, 2005; Toyama and Okabe 2004;
Beadle et al. 2005; Fritzen and Mengelkamp 2005; Giurgiutiu et al. 2005; Hera
et al. 2005; Konstantinidis et al. 2005; Lucero and Taha 2005; Nieuwenhuis et al.
2005; Raghavan and Cesnik 2005; Sundararaman et al. 2005; Terrien et al. 2007).
Lamb wave characteristics such as dispersion curves, phase velocity, attenuation,
reflection and transmission coefficients has been used to detect delamination, poros-
ity, matrix cracking, and other surface defects. Interaction of Lamb wave modes with
defects is an extremely valuable tool in providing quantitative information on the
interface flaws and bond quality. Under different propagation modes Lamb waves
generate high normal and shear stresses at different plate depth and consequently
3 Delamination Detection – A Vibration-Based Approach 35
some modes should be more sensitive to the interface defects and its stiffness vari-
ation than other modes. Terrien et al. (2007) investigated the interaction of Lamb
modes with micro-defects with a simulation combining a finite element approach
and a modal decomposition method. The region around the defects is described by
the finite element mesh. The numerical simulation required first, the finite element
modeling with an explicit algorithm for solving the transient wave propagation, sec-
ond, the modal decomposition which allows to plot dispersive curves and to define
the real, the evanescent and the leaky Lamb modes that exists at a given frequency
and third, the analytical propagation of Lamb waves which are phase velocity and
frequency dependent.
The experimental setup for Lamb wave generation and detection on an aluminum
plate of 1 m long, 300 mm wide and 2 mm thick, with notched of different sizes is
shown in Fig. 3.1. The measured ultrasonic signals at different times and distances
from the source are shown in Fig. 3.2 in which the A1 are Lamb modes transmitted
by the notch, and S0 and A0 are incident modes produced by mode conversion.
(Note : Si – symmetric modes and Ai antisymmetric modes). The reflections from
the notch are clearly visible on Fig. 3.2a. The velocities of different Lamb modes
transmitted by the notch can by identified as can be seen from Fig. 3.2b–d. All
the modes which can propagate at different frequencies are shown in Fig. 3.3. (i.e
green rectangle for excitation window at 2.25 MHz frequency , with a tone burst of
5 cycles at 66◦ incidence angle). In Fig. 3.4 are represented the incident waves, the
transmitted waves, the reflections and the mesh used to identify the mode conversion
with 2D Fourier transform technique. Figure 3.5 shows the modes A0, A1 and S0 of
Lamb wave at 2.25 MHz in a 3 mm thick steel plate in a sound zone and in a zone
with 1.5 mm deep notch.
Fig. 3.1 Experimental setup for Lamb wave generation in a plate with two main defects, a large
notch and a defect assimilated to a crack produced by 5 thin notches. (Terrien et al. 2007, Figure 18)
36 V. Bucur
Fig. 3.2 Signals measured at different points: (a) 80 mm in front of a notch of 500 μm depth and
700 μm width, (b) 20 mm, (c) 45 mm, (d) 165 mm from the notch (Terrien et al. 2007, Figure 19)
As noted by Terrien et al. (2007), “knowing the modal expansion of the wave
propagating on the right of the notch and the waveform of the displacement normal
to the plate” it is possible to predict the waveform at any distance from the source.
The method described here is elegant and has evident advantages such as the pos-
sibility to extract the mode conversion produced by the defects, and to predict the
waveform quite far from the damaged area, if the depth of the defects is smaller then
one half of the plate thickness.
3 Delamination Detection – A Vibration-Based Approach 37
Fig. 3.3 Corresponding dispersion curves of symmetric (solid lines) and antisymmetric (dashed
lines) propagating Lamb modes. Legend: the excitation window is shown by the rectangle is for
2.25 MHz tone burst of 5 cycles and 66◦ incidence angle). (Terrien et al. 2007, Figure 21)
Fig. 3.4 Incident, transmitted and reflected waves and the mesh used to identify the mode con-
version of Lamb waves with 2D Fourier transform technique. (Terrien et al. 2007, Figure 15)
Despite of the evident advantages of the ultrasonic method described here for
the nondestructive inspection and evaluation of structural elements, drawbacks and
limitations are evident, when this method is applied to real – time health monitoring.
This method is local in nature, passive and labor intensive.
However, it is to note that the development of the time reversal concept in modern
acoustics (Fink 1992, 1997; Cassereau and Fink 1992; Wu et al. 1992) brings new
prospective for the utilization of guided Lamb waves for the aerospace structures
(Sohn et al. 2005) and for different civil and medical applications.
38 V. Bucur
Fig. 3.5 The modes A0, A1 and S0 of Lamb wave at 2.25 MHz in a 3 mm thick steel plate in a
sound zone (a) and in a zone with 1.5 mm deep notch (b) (Terrien et al. 2007, Figure 16)
Fig. 3.6 Experimental equipment for the excitation of flexural vibrations in a cantilever beam
using a hammer. The beam response is detected by the laser vibrometer. (Berthelot and Sefrani
2004, Figure 1)
40 V. Bucur
Fig. 3.7 Experimental equipment for testing a free-free beam and measurements of modal fre-
quencies and damping. The excitation can be induced by a shaker or by a piezoelectric actuator
(PZT5) glued on the surface of the specimen, near the free end. (Chrysuchoidis and Saravanos
2004, Figure 2)
Paolozzi and Peroni 1990; Petyt 1990; Hanagud et al. 1990, 1992; Pandey et al.
1991; Tenek et al. 1993; Luo and Hanagud 1996; Messina et al. 1998; Sampaio
et al. 1999; Wahl et al. 1999; Lestari and Hanagud 1999; D’Ambrogio and Fregolent
2000; Brandinelli and Massabo 2002; Kessler et al. 2002b; Lee et al. 2003; Berthelot
and Sefrani 2004; Chrysochoidis and Saravanos 2004; Della and Shu 2005; Ghoshal
et al. 2005; Coutellier et al. 2006; de Borst and Remmers 2006; Ladevèze et al. 2006;
Lestari et al. 2007).
Because of the fact that the damage is a typical local phenomenon, several
difficulties can arise in its detection and location such as:
– higher frequency modes are able to capture local responses, whereas lower
frequency modes capture the global response of the structure
– for the excitation of higher modes more energy is required than for the excitation
of lower modes and loss of information can result from the reduction of time
history measurements
– shifting from the linear to nonlinear response.
Nonlinear damage is observed in the case when the initially linear-elastic structure
behaves nonlinearly after the damage has been produced. Nonlinear normal modes
3 Delamination Detection – A Vibration-Based Approach 41
Fig. 3.8 Modeling of the delamination effects in a representative composite beam (Luo and
Hanagud 2000, Figure 1). Legend: b is the beam width, H is the beam height, L is the beam
length, and respectively mi , Di , Si , Ai (i = 1, 2, 3, 4) the mass density per unit length, bending
stiffness, cross sectional shear stiffness and extensional stiffness of four beams. H2 and H3 are the
distances between the neutral axis of delaminated beam and the neutral axis of intact beam
42 V. Bucur
Fig. 3.9 The nonlinear spring model describing the behavior of the effects between the delam-
inated sublaminates as shown with the dashed line (Luo and Hanagud 2000, Figure 2). Legend:
do – relative displacement; w2 - displacement of the beam 2, w3 – displacement of the beam 3
the delamination, the delaminated sublaminates have the same flexural displace-
ment and slopes. Thus, the effects between the delaminated sublaminates may be
described by a nonlinear spring model as shown qualitatively in Fig. 3.9 by a dashed
line. Furthermore this nonlinear model was reduced into a piecewise linear model
depending on the relative displacement, expressed as w2 − w3 . Three situations can
be observed:
(a) w2 − w3 > 0 the delamination tends to open in vibration, the distributed contact
force is zero. The spring model is represented by the solid line OA.
(b) w2 − w3 = at a fix value, the delamination is completely closed during the
vibration. The spring model is represented by the solid line BC
(c) - do < w2 − w3 < 0 the delamination beam is vibrating in a small amount of
relative displacement. The spring model is represented by the solid line OB.
With the above considerations and from the solutions of the governing equations
of motion of delaminated structures in different stages of vibration it was possible
to synthesized the nonlinear dynamic response, through a nonlinear modal analysis
technique developed by Luo and Hanagud (1997c). Figure 3.10 shows a typical
mode of a transverse isotropic beam with interface 3, 3-inch delamination, and it is
to note that the prediction is closed to the model. In conclusion, it is to note that the
nonlinear dynamic response of the studied structure is precisely predicted with the
proposed piecewise-linear model by Luo and Hanagud (2000).
The reader interested in the case of multiple delaminations is invited to read the
articles published by:
Fig. 3.10 Comparison between the experimental data and the prediction data based on the non-
linear model (Luo and Hanagud 2000, Figure 13). Legend: (a) vibration of the composite beam
with interface 3, 3-inch delamination, experimental data provided by Shen and Grady (1992)
(b) prediction based in the nonlinear mode
Figure 3.11 shows the nonlinear response of a smart composite cross – ply cantilever
beam with delamination at the first mode of vibration.
Another group of methods used for the implementation of nonparametric- models
methods and based on the identification of the nonlinear response of the structure
are the neural-network-based methods (Luo and Hanagud 1997c).These methods
are not commented here.
Prognosis with statistical model development for feature discrimination is also a
group of methods recently developed for structural health monitoring and damage
detection (Montalvao et al. 2006). These methods are not commented here.
– the damage detection (existence of damage in the system), the damage location
(where is the damage), the type of damage (what kind of damage), the extent
of damage (how severe is the damage) and the prognosis (how much useful life
remains).
– the operational and the environmental condition which referees to the safety and
economic motivations for performing the monitoring, and on the other hand which
are the limitation on acquiring data.
After catastrophic events such as earthquakes, excessive snow loading, etc, the
structural health monitoring process must provide reliable information about the
integrity of the structure.
The review of the theoretical ideas proposed in this chapter where expressed
in order to perceive and identify for the future, the research directions able to
identify the damage detection induced by delamination in wood products and in
wood-based composites using ultrasonic and vibration measurements, for a practical
implemented technology. This imply three main aspects : the understanding of the
theoretical aspects related to the physical phenomena for delamination initiation and
growth , the development of models and testing procedures, and the developments
and validation of specific codes.
3.5 Summary
In this chapter the damage detection studies in composite materials were sum-
marized in the context of structural health monitoring, which is referred as the
process of implementing a damage detection strategy for mechanical engineering
infrastructure (Allix and Blanchard 2006).
The review of the theoretical aspects related to the detection of damages induced
by delamination in composites was oriented in two main directions:
References
Abhyankar NS, Hall EK, Hanagud S (1993) Chaotic vibrations of beams: Numerical solution of
partial differential equations. Trans ASME J Appl Mech 60, March:167–174
Adams RD, Cawley P, Pye CJ, Stone BJ (1978) A vibration technique for non-destructively
assessing the integrity of structures. J Mech Eng Sci 20:93–100
Alleyne DN and Cawley P (1992) The interaction of Lamb waves with defects. IEEE Trans
Ultrason Ferroelectr Freq Control 39:381–396
Allix O, Blanchard L (2006) Mesomodeling of delamination: towards industrial applications.
Compos Sci Technol 66:731–744
Auld BA (1980) Acoustic field and waves in solids. vol 1. Krieger, Malabar, FL.
Beadle BM, Hurelaus S, Jacobs LJ, Gaul L (2005) Detection and localization of small notches
in plates using Lamb waves. Proceedings of the 23rd international modal analysis conference.
(IMAX XXIII), Paper no 96
Berthelot JM, Sefrani Y (2004) Damping analysis of unidirectional glass and Kevlar fibre
composite. Compos Sci Technol 64:1261–1278
Borst R de, Remmers JJC (2006) Computational modelling of delamination. Compos Sci Technol
66:713–722
Brandinelli L, Massabo R (2002) Free vibrations of through – thickness reinforced delami-
nated beams. 15th ASCE engineering mechanics conference – EM 2002, June 2–5, Columbia
University:1–8
Bunsell AR (ed) (1988) Quality and damage control in composite materials. Elsevier Applied
Science Publishing, London
Caron JF, Diaz Diaz A, Carreira RP, Chabot A, Ehrlacher A (2006) Multi- particle mod-
elling for the prediction of delamination in multi-layered materials. Compos Sci Technol 66:
755–765
Cassereau D, Fink M (1992) Time reversal ultrasonic field. Part III. Theory of the closed time
reversal cavity. IEEE Trans Ultrason Ferroelectr Freq Control 39:579–592
Castro E, Garcia-Hernandez MT, Gallego A (2007) Defect identification in rods subject to forced
vibrations using the spatial wavelet transform. Appl Acoust 68(6):699–715
Cawley P (1990) Low frequency NDT techniques for the detection of disbands and delaminations.
Br J Non-Destr Test. 32:454–461
Cawley P, Adams RD (1987) Vibration techniques of NDT. In Summerscales J (ed) Nondestructive
testing of fibre – reinforced plastics composites, Elsevier, London, pp 151–200.
Cawley P, Adams RD (1979) The location of defects in structures from measurements of natural
frequencies. J Strain Anal 14, 2:49–57
Cawley P, Alleyne D (1996) The use of Lamb waves for the long range inspection of large
structures. Ultrasonics 34:287–290
Chattopadhyay A, Dragomir-Daescu D, Gu H (1999) Dynamics of delaminated smart composite
cross – ply beams. Smart Mater Struct 8:92–99
Chimenti DE (ed) (1981–2006) Review of progress quantitative nondestructive evaluation. Plenum
Press, New York, NY
Chimenti DE, Martin RW (1991) Nondestructive evaluation of composite laminates by leaky Lamb
waves. Ultrasonics 29:13–20
Chrysochoidis NA, Saravanos DA (2004) Assessing the effects of delamination on the damped
dynamic response of composite beams with piezoelectric actuators and sensors. Smart Mater
Struct 13:733–742
Coutellier D, Walrick JC, Geoffroy P (2006) Presentation of a methodology for delamination
detection within laminated structures. Compos Sci Technol 66:837–845
D’Ambrogio W, Fregolent A (2000) The use of antiresonances for robust model updating. J Sound
Vibr 236:227–243
Della C N, Shu D (2005) Free vibration analysis of composite beams with overlapping delamina-
tions. Eur J Mech A Solids 24:491–503
3 Delamination Detection – A Vibration-Based Approach 47
Delsanto PP, Scalerandi M (1998) A spring model for the simulation of the propagation ultrasonic
pulses through imperfect contact interfaces. J Acoust Soc Am 104:2584–2591
Delsanto PP, Romano A, Scalerandi M, Moldoveanu F (1998) A genetic algorithm approach to
ultrasonic tomography. J Acoust Soc Am 104:1374–1381
Deschamps M, Hosten B (1992) The effects of viscoelasticity on the reflection and transmission of
ultrasonic waves by an orthotropic plate. J Acoust Soc Am 91:2007–2015
Doebling SW, Farrar CR, Prime MB (1998) A summary review of vibration – based damage
identification methods. Shock Vibr Dig 30, 2:91–105
Doebling SW, Hermez FM, Peterson LD, Farhat C (1997) Improved damage location accuracy
using strain energy based mode selection criteria. AIAA J 35(4):639–699
Fink M (1992) Time reversal of ultrasonic field- Basic principles. Part 1. IEEE Trans Ultrason
Ferroelectr Freq Control 39:555–566
Fink M (1997) Time reversed acoustics. Physics Today 20:34–42
Friswell MI (2007) Damage identification using inverse methods. Phil Trans R Soc A 365:393–410
Fritzen CP, Mengelkamp G (2005) In situ damage detection and localization in stiffened structures.
Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 268
Gammadi LNB, Hanagud S (1995) Vibration characteristics of beams with multiple delaminations.
Proceedings of the 36 AIAA/ASME/ASCE/ASC Structures, structural dynamics and materials
conference – adaptive structures forum, New Orleans , LA, pp 140–150
Ghoshal A, Kim HS, Chattopadhyay A, Prosser WH (2005) Effect of delamination on transient
history of smart composite plates. Finite Elem Anal Des 41(9–10) :850–874
Giurgiutiu V, Buli X, Cuc A (2005) Dual use of travelling and standing Lamb waves for structural
health monitoring. Proceedings of the 23rd international modal analysis conference (IMAX
XXIII), Paper no 361
Green RE Jr (Ed) (1985–2006) Nondestructive characterization of materials. Vol. 1– Vol. IXV,
Plenum Press, New York, NY; Springer, Heidelberg
Gummadi LNB, Hanagud S (1995) Vibration characteristics of beams with multiple delaminations.
Proceedings of the 36th AIAA/ASME/ASCE/AHS/ASC – structures, structural dynamics and
materials conference. New Orleans, LA, pp 140–150
Hanagud S, Luo H (1994) Modal analysis of a delaminated beam. Proceedings of the 10th
international. conference experimental. mechanics, Lisabon, June 18–22, pp 880–888
Hanagud S, Luo H (1997) Damage detection and health monitoring based on structural dynam-
ics. Structural health monitoring: current status and perspectives proceedings of international
workshop on structural health monitoring, pp 715–726.
Hanagud S, Nagesh Babu GL, Roglin RL, Savanur SG (1992) Active control of delaminations in
composite structures. Proceedings of .33rd AIAA/ASME/ASCE/AHS/ASC SDM conference,
pp 1819–1829
Hanagud S, Nagesh Babu GL, Won CC (1990) Delamination in smart composite structures.
Proceedings of the 1990 SEM spring conference on experimental mechanics, Bethel, CT, Soc
Exp Mech Inc:776–781
Hayashi T, Kawashima K (2002) Multiple reflections of Lamb waves at a delamination. Ultrasonics
40:193–197
Hera A, Shinde A, Hou Z (2005) Issues in tracking instantaneous modal parameters for structural
health monitoring using wavelet approach. Proceedings of the 23rd international modal analysis
conference. (IMAX XXIII), Paper no 338
Hirsekorn S (2001) Nonlinear transfer of ultrasound by adhesive joints – a theoretical description.
Ultrasonics 39:57–68
Hosten B, Deschamps M, Tittmann BR (1987) Inhomogeneous wave generation and propagation
in lossy anisotropic solids. Application to characterization of viscoelastic composite materials.
J.A.S.A.M. 82:1763–1770
Huber RD, Mignogna RB, Simmonds KE, Schechter RS, Delsanto PP (1997) Dynamic full – field
visualization of uktrasound interacting with material defects : Experiments and simulation.
Ultrasonics 35:7–16
48 V. Bucur
Nayfeh AH, Chimenti DE (1988) Ultrasonic wave reflection from liquid – coupled orthotropic
plates with application to fibrous composites. J Appl Mech 55:863–870
Nayfeh AH, Chin C, Nayfeh SA (1995) Nonlinear normal modes of a cantilever beam. J Vib Acoust
177:477–481
Nieuwenhuis JH, Neumann JJ, Greve DW, Oppenheimer IJ (2005) Simulation and testing for Lamb
wave generation. Proceedings of the 23rd international modal Analysis conference (IMAX
XXIII), Paper no 216
Ogilvy JA (1995) A model for the ultrasonic inspection of composite plates. Ultrasonics 33:85–93
Pandey AK, Biswas M, Samman MM (1991) Damage detection from changes in curvature mode
shapes. J Sound Vibr 145:321–332
Paolozzi A, Peroni I (1990) Detecting of debonding damage in composite plates through natural
frequency vibrations. J Reinforced Plastics Compos 9:369–389
Perel VY (2006) A new approach for dynamic analysis of composite beam with an interplay crack.
Nonlinear Dyn Syst Theory 6(2):171–186
Petyt M (1990) Introduction to finite element vibration analysis. Cambridge University Press. UK
Pilarski A, Rose JL (1987) A transverse – wave ultrasonic oblique incidence technique for
interfacial weakness detection in adhesive bonds. J Appl Phys 63:300–307
Potel C, de Belleval JF (1993a) Propagation in an anisotropic periodically layered medium. J
Acoust Soc Am 93:2669–2677
Potel C, de Belleval JF (1993b) Acoustic propagation in anisotropic periodically multilayered
media: A method to solve numerical instabilities. J Appl Phys 74:2208–2215
Raghavan A, Cesnik CES (2005) Analytical models for Lamb waves based structural health moni-
toring. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper
no 289
Rokhlin S (1979) Interaction of Lamb waves with elongated dalaminations in thin sheets. Int Adv
Nondestr Test 6:263–285
Rokhlin S (1980) Diffraction of Lamb waves by a finite crack in an elastic layer. JAcoust Soc Am
67:1157–1165
Rokhlin SI, Wang YJ (1992) Analysis of boundary conditions for elastic waves. J. Acoust. Soc.
Am. 91:1875–1887
Rokhlin SI, Wang W (1989) Critical angle measurement of elastic constants in composite materials.
Journal of Acoustical Society of America. 86:1876–1882
Rose JL, Pilarski A, Huang Y (1990) Surface wave utility in composite material characterization.
Res Nondestruct Eval 1:247–265
Royer D, Dieulesaint E (2000) Elastic waves in solids. Springer, Berlin
Salawu OS (1997) Detection of structural damage through changes in frequency: A review. Eng
Struct 19:718–723
Sampaio RPC, Maia NMM, Silva JMM (1999) Damage detection using the frequency response
function curvature method. J Sound Vibr 226:1029–1042
Saravanos DA, Birman V, Hopkins DA (1994) Detection of delaminations in composite beams
using piezoelectric sensors. Proceedings of the 31th AIAA/ASME/ASCE/AHS/ASC structures,
structural dynamics and materials conference, pp 181–191
Shen MMH, Grady JE (1992) Free vibrations of delaminated beams. AIAA J 30(5):1361–1370
Shkerdin G, Glorieux C (2004) Lamb mode conversion in a plate with a delamination. J Acoust
Soc Am 116:2089–2100
Shkerdin G, Glorieux C (2005) Lamb mode conversion in an absorptive bi- layer with a
delamination. J Acoust Soc Am 117:2253–2264
Simonetti F (2004) Lamb wave propagation in elastic plates coated with viscoelastic materials. J
Acoust Soc Am 115:2041–2053
Singer L (1997) Bond strength measurements by ultrasonic guided waves. Ultrasonics 35:305–315
Sohn H, Farrar CR, Hemez FM, Shunk DD, Stinemates DW, Nadler BR (2003) A review of
structural health monitoring literature : 1996–2001. Los Alamos National Laboratory Report,
LA-13976 MS
50 V. Bucur
Sohn H, Park G, Wait JR, Lomback NP, Farrar CR (2004) Wavelet – based signal processing for
detecting delamination in composite plates. Smart Mater Struct 13:153–160
Sohn H, Park H, Law KH, Farrar CR (2005) Instantaneous online monitoring of unmanned
aerial vehicles without baseline signals Proceedings of the 23rd international modal analysis
conference (IMAX XXIII), Paper no 259
Solodov IY, Krohn N, Busse G (2002) CAN: an example of nonclassical acoustic nonlinearity in
solids. Ultrasonics 40:621–625
Sridharan S (Ed) (2008) Delamination behavior of composites. Woodhead Publlishing, Cambridge,
England
Stoessel R, Krohn N, Pfleiderer K, Busse G (2002) Air-coupled ultrasound inspection of various
materials. Ultrasonics 40:159–163
Su Z, Ye L, Bu X (2002) Evaluation of delamination in laminated composites based on Lamb waves
methods: FEM simulation and experimental verification. Proceedings of the 1st European
workshop on structural health monitoring, pp 328–335
Sundararaman S, Adams DE, Rigas EJ (2005) Characterizing damage in plates through beam-
forming with sensor arrays. Proceedings of the 23rd international. modal analysis conference.
(IMAX XXIII), Paper no 249
Tenek LH, Henneke EG II, Gunzburger MD (1993) Vibration of delaminated composite plates and
some applications of nondestructive testing. Compos Struct 23:253–262
Terrien N, Osmont D, Royer D, Lepoutre F, Déom A (2007) A combined finite element and modal
decomposition method to study the interaction of Lamb modes with micro-defects. Ultrasonics
46:74–88
Toyama N, Okabe T (2004) Effect of tensile strain and transverse cracks on Lamb wave velocity
in cross – ply FRP laminates. J Mat Sci 39:7365–7367
Tracy JJ, Pardoen GC (1989) Effect of delamination on the natural frequencies of composite
laminates. J Comp Mat 23:1200–1215
Tseng WY, Dugundji J (1971) Nonlinear vibrations of a buckled beam under harmonic excitation.
J Appl Mech 38(6):467–476
Vakakis AF (1996) Normal modes and localization in nonlinear systems. Wiley, Chichester
Wahl F, Schmidt G, Forrai L (1999) On the significance of antiresonance frequencies in
experimental structural analysis. J Sound Vibr 219:379–394
Wang BS, He ZC (2007) Crack detection of arch dam using statistical neural network based on the
reductions of natural frequencies. J Sound Vibr 302:1037–1047
Wang JTS, Liu YY, Gibby JA (1982) Vibration of split beams. J Sound Vibr 84(4):491–502
Wang SS, Yu TP (2006) Nonlinear mechanics of delamination in fiber – composite laminates:
asymptotic solutions and computational results. Compos Sci Technol 66:766–784
Wright WMD, Hutchins DA, Hayward G, Gachagan A (1996) Ultrasonic imaging using laser
generation and piezoelectric air-coupled detection. Ultrasonics 34:405–409
Wu F, Thomas JL, Fink M (1992) Time reversal of ultrasonic fields Part II Experimental results
IEEE Trans Ultrason Ferroelectr Freq Control 39:567–578
Zou Y, Tong L, Steven GP (2000) Vibration – based model – dependent damage (delamination)
identification and health monitoring for composite structures – a review. J Sound Vibr 230:
357–378
Chapter 4
Initiation and Growth of Delamination in Wood
and Wood-Based Composites, a Fracture
Mechanics Approach
Voichita Bucur
Contents
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2 Links with Fracture Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.1 Linear Elastic Fracture Mechanics . . . . . . . . . . . . . . . . . . . . . 60
4.2.2 Nonlinear Fracture Mechanics . . . . . . . . . . . . . . . . . . . . . . . 72
4.3 Micro-structural Aspects in Wood . . . . . . . . . . . . . . . . . . . . . . . . 79
4.4 Micro-structural Aspects in Wood-Based Composites . . . . . . . . . . . . . . . 84
4.5 Fracture Mechanics Parameters for Ecological Relevance . . . . . . . . . . . . . 91
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.1 Introduction
Fracture Mechanics concept has been applied to wood material as reported during
more than fourthly years, by numerous references, review articles and books. Some
of them are cited below. (Attack et al. 1961; Porter 1964; Boyd 1973; Schniewind
and Centeno 1973; Schniewind and Lyon 1971, 1973; Schniewind and Pozniak
1971; Leicester 1971, 1973, 1974; Pearson 1974; Jeronimidis 1976, 1980;
Schniewind 1977; Barrett 1976, 1981; Schniewind et al. 1982; Valentin and
Morlier P 1982; Jung and Murphy 1983; Petterson and Bodig 1983; Boatright
and Garrett 1983, Triboulot et al. 1982, 1984; Tschegg 1986; Patton – Mallory
and Cramer 1987; Gustafsson 1985; Boström 1988, Akande and Kyanka 1990;
Valentin et al. 1991; Aicher 1992; Aicher et al. 1993, 1998; Stanzl-Tschegg et al.
1994, 1995, Zink et al. 1994, 1995; Renaud et al. 1996; Gibson and Ashby 1997,
Bodner et al.1997; Thuvander and Berglund 1998; Tschegg et al. 2001; Sippola and
Frühmann 2002; Cotterell 2002; Reiterer and Sinn 2002; Smith et al. 2003; Vasic
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
and Smith 2002, 2003; Conrad et al. 2003; Nairn 2006; Vasic and Stanzl-Tschegg
2007; Nairn 2007a, b, c; Hofstetter et al. 2007).
The increasing interest on the physical phenomena which lead to the onset of
delamination, its development and its interaction with other damage mechanisms
is determined by the expected economical benefits for wood products and various
wood-based composites structures. In order to avoid over dimensioning in struc-
tural elements design, it is necessary to understand as deep as possible, the physics
behind the damage mechanisms and to develop theories and tools (analytical or
numerical) able to take into account the onset and growth of delamination from
the earliest phases of design. The applications of wood-based laminated composites
are limited by delaminations which can be introduced during the fabrication pro-
cess or later in service life. The presence of delaminations degrades the stiffness,
strength and fatigue characteristics of structural elements and has the potential to
cause catastrophic failure of the structures.
In this chapter are analysed the basic concepts related to fracture mechanics
which allow the understanding of initiation and growth of delamination in wood
and wood-based composites. Basic theoretical approaches and the state of the art
for characterization and predicting delamination are outlined.
Fig. 4.1 Schematic representation of a crack in a plate and in an infinite solid (Triboulot et al.
1984, Figure 7)
4 Initiation and Growth of Delamination 53
The definition of the damage zone ahead of a crack tip is crucial for the studies
of wood fracture (Vasic et al. 2002).
Because of the anisotropic nature of wood, defined with three principal axes L,
R and T (longitudinal, radial and tangential) six different fracture system orienta-
tions can be defined such as: for crack propagation in L direction, the systems RL
and TL, for crack propagation in direction R the systems LR and TR, and for crack
propagation in T direction, the systems RT and LT. Note that the first letter indicates
the direction normal to the crack plane and the second letter indicates the direction
of crack propagation. In practice the crack’s path is very complex. The crack path
for the systems RL and TL propagates always parallel to grain. However, the crack
path for transverse directions TR and RT could propagates in any direction. When
cracks propagation in R direction, two situations were observed, the path toward
the bark or toward the pith. As reported by Attack et al. (1961) the toughness in
green spruce in TR was 100 J/m2 and 180 J/m2 in RT direction. Schniewind and
Centeno (1973) reported no differences between both directions in the stress inten-
sity factor in air-dried Douglas –fir (0.35 MPam–2 ). Dill – Langer et al. (2002) noted
that in softwoods crack growths in TR system in tension perpendicular to the grain
is not steady and rupture of earlywood cell walls was observed. Another mecha-
nism of rupture was observed when the crack growths in the RT system, namely
the rupture between adjacent tracheids. Thuvander and Berglund (2000) observed
the crack arrest in earlywood. Ashby et al. (1985) noted that in low density wood
such as balsa the fracture propagates by cell wall rupture, while in high density
wood species the fracture between cell walls, by peeling the middle lamellae was
observed.
Most studies on wood fracture mechanics rely on the concept of linear elastic
fracture mechanics (LEFM), because of the simplicity of this approach.
The concept of linear elastic fracture mechanics (LEFM) is based on the rela-
tionships existing between the stress in the vicinity of a crack tip and different
characteristics of the structure such as: the nominal stress applied, the material
mechanical and physical properties, the size, shape and orientation of existing flaws.
This theory stipulates that the stress level in the vicinity of the crack tip tends toward
infinity. In real materials, obviously there is a zone where the elastic solution breaks
down. The size of the plastic zone dp at the crack tip in a material with σ Y yield
strength, can be written such as:
2
1 K
dp =
nπ σY
54 V. Bucur
– the stress intensity factor (K) which is based on the local stress distribution around
a crack tip. Critical intensity factor (KC ) is considered a material parameter that
defines the resistance to crack growth (referred also as fracture toughness of the
material).
– the strain energy release rate (G), which is based on the global energy balance
The stress intensity factor (K) and strain energy release rate (G) will be described
in that follows.
The stress field around a crack tip has been documented in many reference books
and we cite only the most recent (Sandford 2003; Anderson 2005). Using the nota-
tions from Fig. 4.1, the stress field in the immediate vicinity of a crack tip, for an
isotropic solid, can be written such as:
KI θ θ 3θ
σx = √ cos 1 − sin sin
2π r 2 2 2
(4.1)
KII θ θ 3θ
−√ sin 2 + cos cos
2π r 2 2 2
KI θ θ 3θ
σy = √ cos 1 + sin sin
2π r 2 2 2
(4.2)
KII θ θ 3θ
−√ sin cos cos
2π r 2 2 2
KI θ θ 3θ
τxy = √ cos sin cos
2π r 2 2 2
(4.3)
KII θ θ 3θ
+√ cos 1 − sin sin
2π r 2 2 2
σz = 0 for plane stress (4.4)
4 Initiation and Growth of Delamination 55
and
σz = ν σx + σy for plane strain conditions (4.5)
KIII θ
τyz = √ cos (4.6)
2π r 2
KIII θ
τzx = √ sin (4.7)
2π r 2
The constants KI , KII , KIII are termed stress-intensity factors for the Modes I, II
or III and describe the intensity of the stress field as a measure of the severity of
the crack. The stresses are singular at the crack tip for r = 0 and have a square root
singularity.
The displacements (u, v) under stress conditions can be written as:
KI r θ θ
u= cos k − 1 + 2 sin2
2μ 2π 2 2
(4.8)
KII r θ 2 θ
+ sin k + 1 + 2 cos
2μ 2π 2 2
KI r θ 2 θ
v= sin k − 1 + 2 cos
2μ 2π 2 2
(4.9)
KII r θ θ
− cos k − 1 − 2 sin2
2μ 2π 2 2
2KIII r θ
w= sin (4.10)
μ 2π 2
where
μ is the shear modulus,
ν is the Poisson’s ratio,
k = (3 − v)/(1 + v) for plane stress and k = (3 − 4v) for plane strain.
The strain energy release rate G is related to the work required to close a crack
of length a + a to a length a, and is based on the Irwin’s crack closure concept
(Irwin 1957). The total strain energy release rate G is expressed such as:
KI2 K2 K2
G = GI + GII + GIII = + II + (1 + ν) III (4.11)
E E E
where G I , G II , GIII are strain energy release for the modes I, II, III and E = E
in plane stress and E = E/(1 − v2 ) in plane strain, E = Young’s modulus of the
isotropic material. For orthotropic materials these parameters must be corrected with
the corresponding elastic constants.
56 V. Bucur
For cracks in orthotropic materials, the stress distribution is much more complex,
as described by Wu (1967), Walsh (1972), Wang (1984), Tada et al. (2000), Raju and
O’Brien (2008), and Sridharan (2008) .
It was generally admitted, that under short load duration the dry wood exhibits
brittle fracture and linear elastic behaviour. This statement implicitly requires the
theory of linear elastic fracture mechanics (LEFM) for the description of wood
behavior.
The main limitations of (LEFM) are:
– the necessity of assuming the existence of a crack,
– the effects of fracture process zone are in the vicinity of the crack tip,
– the available energy goes into the creation of a single new fracture zone.
However the linear elastic assumption is not suitable for examining the viscoelas-
tic behaviour of wood, the mechano-sorptive effect, the scale effect, the mechanical
behaviour under long term loading, the microstructural phenomena, etc. For these
cases the quasi – brittle fracture is assumed and the phenomena are studied with the
nonlinear fracture mechanics.
The nonlinear fracture mechanics (NLEFM) introduces the notion of planar pro-
cess zone where cohesive stresses are assumed to occur (Boström 1988, 1992;
Patton – Mallory and Cramer 1987; Gustafsson 1985, 1988; Vasic and Smith 2002).
In such materials, the fracture is preceded by localized phenomena in the plastic
zone, the damage is assumed to occur on a surface, and a nonlinear region can be
detected prior to the peak load, followed by strain softening region after the peak.
The crack tip opening displacement (CTOD) can integrate these phenomena and can
be used to model fracture under conditions of large plastic deformation. For fracture
to occur there must be a critical crack tip opening (δ) which can be calculated as:
4 K2
δ= (4.12)
π EσY
The stability of a crack depends on the interaction of the applied loads and the
material toughness. When unstable the cracks can growth with different velocities.
However, it is the whole system which has the property of stability and not the crack
itself. Crack initiation and crack propagation are best characterized by the fracture
energy, whereas the stress intensity factor only gives information on crack initiation.
In wood like in other solids, the fracturing under mechanical loading takes place
in three steps namely, crack initiation, crack propagation and fracture. During crack
initiation a process zone is formed in front of the crack tip, with numerous micro-
cracks. The microcracks constitute the delamination front which profuse micro
cracking ahead of the delamination front. The coalescence of existing microcracks
forms macrocracks which propagate. During crack propagation, in the weak zone,
behind the crack tips bridging effect takes place, which becomes gradually weaker
until rupture occurs, as the complete separation of fracture surfaces. In solid wood
and wood-based composites bridging process induces energy dissipation which
strongly influences their fracture behaviour.
4 Initiation and Growth of Delamination 57
Smith and Vasic (2003) noted that in wood mechanically loaded, “cracks start
to grow from microscopic defects in the cell walls and cell boundaries. As these
small cracks accumulates, the compliance of the material increases. Prior to peak
load there is a localisation process in which the damage that causes failure becomes
more confined to a narrow region. By the time peak load is reached, a critical crack
accompanied by a fracture process zone has been established, and strain softening
can occur. The reason the fracture is not sudden, is that toughening mechanisms have
been mobilised near the crack tip, causing energy to be dissipated more gradually”.
The concepts developed with NLEFM are:
a b
Fig. 4.3 J-integral paths and softening curves. (a) J-integral analysis along the path 1 . . . 6 . The
bridging zone develops from the crack tip at x0 to the notch root at x root . (Nairn 2009, Figure 1); (b)
in bridging modelling, the crack opening displacements normal and tangential to the crack surface
can be described by different softening functions such as : A – linear elastic, B – linear elastic brittle
C – triangular with initial linear regime followed by a linear softening regime (Dourado et al. 2004),
D – arbitrary traction, often approximated with a cubic function, E- linear softening, F-nonlinear
softening function (Schmidt and Kallske 2007, Figure 2); (c) crack opening with microcracking
and bridging components (Stanzl-Tschegg et al. 1995, Figure 2)
58 V. Bucur
and can be analysed with J-integral paths noted i (Fig. 4.3a). The J-integral
[energy/unit area, or J/m2 ] is the path contour integral at crack tip which takes into
consideration the stress vector acting perpendicular to the contour, the displacement
vector and the strain energy density.
It is to note that tractions act normal and tangential to the crack surfaces. The
traction forces depend on the corresponding crack opening displacements, increas-
ing to a peak (or cohesive stress) and decreases to zero when the tractions fails
(corresponding to critical opening displacement). Materials can exhibit softening
behaviour as shown in Fig. 4.3b. In case of wood, most frequently the bilinear
and the polynomial functions were used. Figure 4.4c shows the bilinear soften-
ing model which explains the development of the microcracking component and
the bridging component as suggested by Stanzl-Tschegg et al. (1995). Some other
functions were used in finite element simulation of crack growths such as bilinear
and trilinear (Douardo et al. 2004; Coureau et al. 2006a) or nonlinear (Schmidt and
Kallske 2007).
(c) energy release rate expressed by J integral is the energy that is extracted through
the crack tip singularity.
Fig. 4.4 Theoretical behaviour of materials exhibiting bridging zone, with J integral paths
(Coureau et al., 2006). zone 1 – onset of softening behaviour at GR (a0 ), the resistance GR defining
the onset of the crack propagation of the equivalent elastic crack; zone II – progressive increase
of the resistance to crack growth au. R-curve depends on the sharp of the softening behaviour, the
ultimate load depends on the slope of the softening curve; zone III – crack propagation at constant
resistance GR (a > ac ) = GRc ; zone IV – propagation at constant resistance to crack growth, when
successive failures of interface element located ahead of the crack tip. (Note that the experiments
were with spruce, Figure 12)
4 Initiation and Growth of Delamination 59
Figure 4.4 illustrates the influence of different parameters describing the soften-
ing behaviour on three curves, namely COD –load curve (Fig. 4.4a), Rcurve − GR
(GR is energy release rate from LEFM) curve (Fig. 4.4b) and, w (relative displace-
ment in tangential direction of the interface obtained from the upper and lower
substrates) – σ curve (Fig. 4.4c).
Coureau et al. (2006) described four zones:
zone I – onset of softening behaviour at GR (a0 ), the resistance GR defining the
onset of the crack propagation of the equivalent elastic crack
zone II – progressive increase of the resistance to crack growth au . R-curve
depends on the sharp of the softening behavior, the ultimate load depends on
the slope of the softening curve.
zone III – crack propagation at constant resistance GR (a > ac ) = GRc
zone IV – propagation at constant resistance to crack growth, when successive
failures of interface element located ahead of the crack tip. (Note that the
experiments were with spruce)
From COD –load curve (Fig. 4.4a) one can see the evolution of compliance (λ)
as a function of initial crack length (a0 ) and critical crack length (ac ). Crack propa-
gates at G = GRc ∀a. The levelling of Rcurve (Fig. 4.4b) might indicate that in wood,
the toughness mechanism do not tends to infinite, where crack bridging requires
sufficient deformation to produce closing forces (Smith, Landis et al. 2003). The
softening behaviour of the cohesive crack is shown in (Fig. 4.4c). The normal stress
transmitted by the interface decreases progressively from the interfacial normal
strength (ft ) to 0, when critical opening displacement (wc ) is generated.
Numerical methods can be used to evaluate the J- integral for any crack, type
of loading and body configuration (Atluri 1986; Anderson 2005). Since numerical
analyses are time consuming, simplified approaches for engineering calculations
have been developed (Berto and Lazzarin 2007).
The limitations of NLFM are related to J integral. Theoretically, the utilisation of
this parameter is based on the elastic response of the material. However it is assumed
that the nonlinear elastic material will not have permanent deformation. J integral is
appropriate for monotonic loading conditions (where material unloading behaviour
is not significant) and for small newly form process zone when crack advances due
to the creation of stress free surfaces.
A compromise between the LEFM and NLEFM has been proposed through the
development of Damage Mechanics which is a phenomenological approach for
material that do not exhibit plastic deformation and can not de characterized by
brittle rupture. In such materials the formation of microcracks, defined as damages,
induced stiffness decreasing which can be quantified by a damage variable which
express the magnitude of this stiffness decreasing. Using Damage Mechanics (DM)
approach Daudeville (1999) simulated the fracture in wood, by treating the problem
of crack initiation in “originally uncraked “ structure of spruce specimens loaded in
bending and by comparing the load displacement curves obtained with LEFM and
DM. Both approaches correctly predicted the load-displacement curves. Moreover,
the critical energy rate (parameter of LEFM) and the fracture energy (parameter of
60 V. Bucur
DM) where compared with the experimentally determined dissipated energy to frac-
ture of the specimens. It was demonstrated that in both approaches fracture energy
is the major parameter that governs crack propagation in wood.
In that follows concepts related to LEFM and NLEFM will be discussed in more
details in view of application to solid wood and wood-based composites.
The development of the concepts related to the application of LEFM required sev-
eral hypotheses (Stanzl-Tschegg et al. 1995; Tschegg et al. 2001; Vasic et al. 2002;
Vasic and Smith 2002; Jensen 2005a, b, c; Keunecke et al. 2007) such as:
– the homogeneity of the linear elastic material in which fracture takes place
– the pre-existing crack propagates always along one direction
– crack-tip displacement is associated with three principal pure modes of fracture,
Mode I, Mode II, Mode III.
– the intensity of stress distribution in the vicinity of the single crack tip is fully
characterized by the stress intensity factors by three intensity factors, KI , KII , KIII ,
associated with three principal pure modes of fracture
– crack surfaces are traction free at all stages of loading
– the crack propagates dynamically at a certain velocity once the critical fracture
toughness (KC ) or strain energy (GC ), release has been reached
– the inelastic process zone is limited to a small volume at crack tip.
The experimental conditions that influence the fracture process in wood are:
the geometry of the specimens, the loading orientation and rate, and the moisture
content. Wood fracture toughness is also strongly dependent on wood species and
density.
The most common geometry of specimens used for the measurements of fracture
toughness in Mode I and Mode II are shown in Fig. 4.5. The specimens can be tested
in tension, bending, or shear.
The effect of loading rate on wood fracture toughness has been studied by Conrad
et al. (2003) and Vasic, Ceccotti et al. (2009). Conrad et al. (2003) noted that sub-
stantial crack growth can take place at low strain rate, whereas at high strain rates
higher toughness values were measured. In this late case, the dissipation of energy
is slow down because of the relatively short time of the process. Vasic, Ceccotti
et al. (2008) noted that the fracture resistance curves at deformation speed between
0.05 and 200 m/min is influenced by the structural inertial effect. The twice-as-high
fracture resistance at 200 m/min deformation rate proves the existence of a critical
deformation rate above which the viscoelastic response of wood is suppressed. This
phenomenon can characterize the ductile brittle transition limit for wood.
As regards the loading orientation Table 4.1 gives some experimental values of
fracture toughness in Mode I and Mode II determined for different species. As can
be seen from this table, wood anisotropy is well expressed by the values of KIc . For
example, for Mode I, for Douglas fir the values are such as:
LR
KIc· > KIc·
LT
> KIc·
RL
> KIc·
TR
= KIc
RT
> KIc·
TL
4 Initiation and Growth of Delamination 61
b1 b2
c1
c2
Fig. 4.5 Specimens geometry for fracture testing (a) fracture testing in mode I (Figure 4.5.a1)
and mode II (Figure 4.5.a2) Yoshihara (2006, Figures 3 and 4). (b) Splitting test for macroscopic
studies (Figure 4.5b) (Tschegg 1986, patent AT 390328) (c) splitting test for micromechanical
studies in SEM chamber (Vasic et al. 2002, Figure 2) Figure 4.5.c1 loading device. Figure 4.5 c2
specimen
4 Initiation and Growth of Delamination 63
Table 4.1 Experimental values of fracture toughness in Mode I and Mode II determined for
different species
Plane kNm–3/2
Mode I
Douglas fir TL 309 Schniewind and Centeno (1971)
RL 410
TR 355
RT 355
LT 2417
LR 2692
Western white pine TL 190 Johnson (1973)
Western red cedar TL 185 Johnson (1973)
Hoop pine TL 494 Walsh (1971)
Hard maple TL 492 Johnson (1973)
Paper birch TL 564 Johnson (1973)
Red oak TL 407 Johnson (1973)
Lauan TL 478 Johnson (1973)
Messymate stringbark TL 505 Walsh (1971)
Maiden’s gum TL 681 Walsh (1971)
Balsa TL and RL 112 Wu (1963) cited in Wood
handbook (1999)
Mode II
White spruce TL 1890 Barrett (1981)
Lodgepole pine TL 2187 Barrett (1981)
Amabilis fir TL 1626 Barrett (1981)
Douglas fir TL 2143 Barrett (1981)
RL 2230 Wood handbook (1999)
Red spruce TL 2190 Wood handbook (1999)
RL 1665 Wood handbook (1999)
Balsa TL 159 Wu (1963) cited in Wood
handbook (1999)
The highest value of KLR Ic is explained by the fact that when crack propagates
in R direction in LR plane, the transversal section of rays and tracheids is the major
obstacle for crack propagation. Moreover KTL Ic has the smallest value because of
the weakest split behaviour of wood in this plane; in this case it is suppose that
the crack is initiated and then propagates in the middle lamella rich in amorphous
lignin and poor in cellulose. As noted by Boatright and Garrett (1983) because of
anisotropic and heterogeneous structure of wood, the TL system is “weak” and the
LT system is “tough”. Similar remarks can be pointed out when fracture energy to
failure (GC ) in tension perpendicular to grain is calculated for softwoods (Table 4.2).
It was observed that always GRL C > GC and the ratio GC /GC is between 1.26
TL RL TL
and 1.55.
It is generally admitted that Mode I cracks propagate in a brittle manner
with low energy consumption, whereas for Mode II cracking much energy is
consumed in creating and breaking the hairy fragments that have been seen on the
microfractographic images on the crack surfaces.
64 V. Bucur
Table 4.2 Mechanical and fracture characteristics of some species determined with splitting
technique and FEA (data from Reiterer et al. 2002)
√
Young’s moduli (GPa) Kinitial (N/m) KIc (MPa m ) Gf (J/m2 )
EL ER ET RL TL RL TL RL TL
It is to note that for real structures the Mixed Mode is dominant and conse-
quently this mode is of major interest for the studies related to health monitoring
of structures. In wood very often a Mixed Mode I/II is possible because of the fact
that cracks propagates along the fibres, irrespective of original crack orientation.
Jernkvist (2001) proposed a theoretical model for a Mixed Mode I/II based on the
fact that ” the Mixed Mode loading is supposed to displace the microcrack zone to
one side of the main crack plane, and the coalescence of the microcracks with the
parent crack may in this case require transverse cutting of tracheids walls. This pro-
cess will create a rough crack surface which does not follow the fibre directions as
can be seen in the simulation shown in Fig. 4.6.
The quality of the surface observed in-situ with ESEM for spruce specimens
loaded in Mode I in TR system by splitting technique is shown in Fig. 4.7. The
wood structure depicted in this figure is perfectly localized on the load-displacement
diagrams. The arrow at position 3a indicates the crack tip at – 20 N shortly before
loading. The crack was located in the early wood zone with a razor blade. The crack
front is widened, but no propagation occurred. The profile of the crack mouth open-
ing is parabolic, wider in earlywood than in latewood. This image corresponds to
the initial step – no crack propagation. At position 3b, in spruce the first propagation
event occurs, the load dropped, the crack penetrated the latewood and stopped in the
earlywood zone of the next ring. For beech specimens, the initial position is shown
at the position 3c, corresponding to – 52 N. The profile of the crack is parabolic.
For beech, the first propagation occurred at – 65 N at position 3d. In TR system and
Mode I the behavior of spruce is different than that of beech.
The behavior of different species (ash, oak, alder and spruce) related to the crack
propagation in RL and LT, Mode I is shown in Fig. 4.8 with load displacement
curves obtained by the wedge splitting test. In hardwoods a macrocrack initiation
takes place at the maximum splitting force, followed by unstable crack propagation
and several steps for crack arresting. The spruce specimens behaved very differently,
showing a continuous load-displacement curve, with a maximum load peak related
to a deviation from the linear behaviour. It was noted that “spruce displays more
ductile and the hardwoods more linear elastic and brittle behaviour. Table 4.2 gives
some fracture mechanics parameters deduced for ash, oak, alder and spruce with
FEA. For all species the fracture RL parameters are higher then TL parameters and
this is explained by the higher proportion of medulary rays. Table 4.3 gives the value
4 Initiation and Growth of Delamination 65
Fig. 4.6 Crack growths along the fibres in pure Mode I and in Mixed Mode I/II – theoretical
models (Jernkvist 2001, Figure 5)
Fig. 4.7 Load-displacement diagrams in situ and specimens of spruce and beech loaded in Mode I,
inside the chamber of an environmental scanning electron microscope loaded in Mode I in the TR
crack propagation system (Frühmann et al. 2003, Figures 2 and 3). The arrows show the position
a, b, c, d on the load displacement diagrams when the corresponding images were taken such as: a
and b for spruce and c and d for beech
ten times higher crack growth resistance compared to Mode I. “Under pure Mode
III load, crack initiation takes place under Mode III in beech as well as in larch.
More advanced cracks, however, propagate predominantly as Mode I. The change
of the fracture mode takes place preferentially in RL orientation in beech and in TL
orientation in larch” (Tschegg et al. 2001). This behaviour is related to the pres-
ence of medullary rays, much more numerous and important in size in beech than
in larch.
The influence of wood moisture content on fracture characteristics was thor-
oughly reviewed by Wang et al. (2003). The maximum fracture toughness was
4 Initiation and Growth of Delamination 67
reported at 17% wood moisture content. King et al. (1999) noted that the mode I
fracture toughness was lower for wet wood, in all fracture directions, (Table 4.4)
than for dry wood (radiata pine specimens) tested in bending (three point bend-
ing and single edge notched). In situ examination with environmental scanning
Table 4.3 Fracture energy Gf in tension perpendicular to grain and in three point bending test, for
specimens with constant width (b = 45 mm) [data from Daudeville 1999]
Fig. 4.9 √Influence of density on fracture behaviour. (a) Variation of fracture toughness, Mode I
KIc (MPa m) versus density (Conrad et al. 2003, Figure 5) (b) fractured zone in middle lamella
region (Donaldson 1997, Figure 5) (c) Microdensity variation in a fractured zone (Donaldson 1997,
Figure 6)
microscope has shown that in green wood water droplets moved away from the cell
lumen around the crack tip. During drying microcracks were observed. Crack bridg-
ing is part of toughening mechanisms. Based on in-situ experiments with ESEM,
Vasic and Stanzl-Tschegg (2007) have shown the influence of moisture content
on fracture toughness and fracture energy (Fig. 4.10) on several European species.
Three main regions of moisture content can be observed, in which the influence of
wood structure is obvious.
Table 4.4 Fracture toughness in Mode I and Mode II for dry and wet Pinus radiata specimens
(data from King et al. 1999)
√
Fracture toughness [MPa m]
Fracture Wood RL RT TL TR LT LR
Fig. 4.10 Influence of moisture content on some fracture parameters in RL crack propagation
(Vasic and Stanzl-Tschegg 2007, Figure 7). (a) total fracture energy Gf (N/m) versus moisture
content; (b) fracture toughness KIc (kNm−3/2 ) versus moisture content
From this data it seems evident that the moisture content has the ability to change
wood fracture mechanism from brittle to ductile.
Vasic and Stanzl-Tschegg (2007) noted that the stress gradient at the crack tip
might have a significant effect on the local moisture distribution, free water flow
and vapour diffusion in the vicinity of the crack. Experiments on green wood and
modelling with discrete finite elements (Frühmann et al. 2003; Vasic and Stanzl-
Tschegg 2005, 2008; Sedighi – Gilani and Navi P 2007) has shown that the process
zones are confined to one or only a few cell rows, and the lattice fracture model
shown distributed damage in the most stressed regions between the area where a
concentrated force is applied and, the notch plane where the fracture is initiated.
The aspects discussed previously have proved the limitations of LEFM concepts
(synthesized in Table 4.5) and the necessity to introduce new concepts.
Table 4.5 Limitation of LEFM for wood fracture studies (data from Vasic et al. (2002)
Fig. 4.11 SEM micrograph (Eastern Canadian spruce) of a crack tip (Vasic and Smith 2002,
Figures 2, 4) (a) the crack tip coexists with a bridging zone behind the tip crack, towards the end
of the experiment. The bar line is 100 μm. (b) Fibers bridging. The bar line is 100 μm
4 Initiation and Growth of Delamination 73
Fig. 4.12 Energy release rate Gfc versus crack size for end-tapered DCB specoimen (Vasic and
Smith 2002, Figure 11)
Table 4.6 synthesizes the crack models suggested for wood fracture studies with
nonlinear fracture mechanics at overall macroscopic level. As underlined by Landis
and Navi (2009), these models break away from classic continuum framework,
referred to the cross grain fracture of wood, represent material heterogeneity and
used FEA with different stress – crack opening functions (linear, bilinear, trilinear,
non linear). All these models ask for high computational expenses.
Fig. 4.13 Fracture toughness versus crack length for end-tapered DCB specimen (Vasic and Smith
2002, Figure 12)
74 V. Bucur
Table 4.6 Crack models, with a crack interface element, for wood NLEFM studies
Morphological based models – lattice models and material point model – has
been developed to understand the structural complexity of wood and to relate micro
and macro mechanical behavior. Lattice models have been developed by Landis
et al. (2003), Davids et al. (2003), Wittel et al. (2005), Vasic and Stanzl-Tschegg
(2007), Mishnaevsky and Qing (2008), and Landis and Navi (2009).
Material point model (MPM) is a very recent and promising model that
discretized the solid in an array of points, developed by Nairn (2006–2009),
Guo and Nairn (2006). Figure 4.14 shows the numerical modeling of wood
structure when fracture occurs in TR plane. Figure 4.15 shows a digitized
image of Douglas fir specimen with a notch (a), the corresponding MPM con-
verted image for radial direction on a scale of 0◦ –90◦ from white to black
(b), the crack growth started with an initial kink (c) and the simulated crack
growth (d).
With MPM to each point specific properties such as stiffness and toughness can
be assigned. In a numerical study of the transverse modulus of wood as a function
of grain orientation and properties including heterogeneity and anisotropy Nairn
(2007b, c) demonstrated in a very elegant manner the feasibility of the material
point model using different degrees of complexity for the mechanical behaviour
of wood, ranging from the simplest transverse isotropic hypothesis to the more
76 V. Bucur
complex heterogeneous cylindrical orthotropy (Table 4.7). The material point model
requires data for reliable mechanical parameters of wood structural element that
can be obtained through the development of new methodology such as acous-
tic microscopy (Bucur et al. 1995; Clair et al. 2000; Bucur 2003, 2005) or other
ultrasonic techniques as demonstrated by Bucur et al. (1994).
As a conclusion of this section, it can be noted that analytical and numerical
models have been successfully developed for wood structure and fracture mechanics
studies.
Gibson and Ashby (1997) derived an analytical model for wood structure
described as a regular array of hexagonal cells and derived results for initiation of
fracture by either elastic or plastic buckling. This is a 2D model and mimics only
the softwood structure.
The numerical modelling of wood structure is more complex and includes the
finite element analysis, the lattice method and the material point method.
Finite element analysis reduces the analysis to an idealized structure. The
limitations of this approach are described by Smith, Landis et al. (2003) such as:
– the wood structure is very complex, difficult to discretized into an FEA mesh
– the common practice of reducing analysis to a small idealized structure limits its
value for numerical modelling of the details of failure mechanisms,
– the number of elements required to accurately mesh realistic wood morphology is
computationally expensive
– the difficulty to consider the contact between cells and the large deformations
Fig. 4.15 Crack growths with Material Point numerical modeling (Nairn 2007c, Figure 2,
Figure 3) (a) digitized image of Douglas fir specimen with a notch (b) the corresponding MPM
converted image for radial direction on a scale of 0–90◦ from white to black (c) the crack growth
started with an initial kink (d) the simulated crack growth
The lattice methods – for which the wood structure is replaced by a model of
rod and spring elements is limited to linear elastic material properties. Variations
in wood structure have been introduced by allowing strength and/or stiffness prop-
erties of the elements to be statistical quantities. Lattice models have focused on
longitudinal properties of wood where the rods are wood fibres and springs repre-
sent transverse properties. In principle lattice models could be applied to transverse
properties or 3D modelling, but that capability has not been demonstrated.
78 V. Bucur
Table 4.7 Some hypotheses for wood structure modelling and fracture mechanics studies at
annual ring level with material point method
The material point method for numerical modelling of wood structure is capable
of modelling wood anatomy in more details than the methods described previously.
The advantages of this method can be such as:
starting on the outside they traverse groups of lamellae or the entire wall”. Thus, to
identify the specific behaviour of hierarchical microstructure of wood is essential
for further developments of advanced models of damage and fracture.
During the 1980s a new step in understanding wood behaviour and delamina-
tion was achieved with fractographic studies using scanning electron microscope,
ex-situ (Borgin 1971; Kucera and Bariska 1982; Bariska 1994; Bodner et al. 1996;
Zimmermann et al. 1994; Donaldson 1997; Seel and Ziemmermann 1998; Ando
and Ohta 1999). Some micrographs ex-situ are shown in Fig. 4.16 for spruce and in
Fig. 4.17 for beech. In spruce loaded on Mode I and impact bending, brittle frac-
ture was observed in latewood tracheids as well as delamination between S1 and S2 .
Ductile fracture was observed in fracture in long term bending with specimens at
20◦ C and 65% relative humidity. The microfibrils are pulled out of the secondary
Fig. 4.16 Fracture morphology in spruce. (Zimmermann et al. 1994, Figures 3, 6, 8, 11) frac-
ture in impact bending, with specimens at 20◦ C and 35% relative humidity. Latewood tracheids,
brittle fracture with S2 clean surface. Delamination between S1 and S2 . (a) fracture in long term
bending with specimens at 20◦ C and 65% relative humidity. Latewood tracheids, ductile fracture
the microfibrils are pulled out of the secondary wall (b) fracture in impact bending, with speci-
mens at 20◦ C and 35% relative humidity. Delamination of middle lamella (matrix) and secondary
wall composed from microfibrils. (c) fracture in impact bending, with specimens at 20◦ C and 35%
relative humidity. Latewood tracheids. “A fast and very brittle fracture led to a partially smooth
fracture surface whereas the remaining part of the fracture was more ductile and exhibits a rough
surface with a certain separation of microfibrils and matrix”
4 Initiation and Growth of Delamination 81
Fig. 4.17 Fracture morphology in beech in static bending and in impact bending (Seel and
Zimmermann 1998, Figures 1, 2) (a) libriform fiber after short time static bending, strong deforma-
tion in tension zone of the cell wall with radial agglomerations (arrows) on S2 (b) libriform fiber
after impact bending. Brittle fracture with delamination in fibril/matrix structure. Some microfibrils
are oriented radially and some others are arranged in a layered structure
wall. Delamination was observed between the middle lamella (matrix) and S2 . The
ductile fracture led to a relatively rough fracture surface. In beech libriform fiber
in fracture in impact bending, brittle fracture with delamination between microfibril
and matrix was observed. Some microfibrils are oriented radially and some oth-
ers are arranged in a layered structure. Donaldson (1997) reported the aspects of the
ultrastructure of transwall fracture surfaces in Radiata pine wood using transmission
electron microscopy. The fracture initiation and growth was studied under tensile
stress parallel to the cell wall layers. Figure 4.18 shows a tangential fracture of two
adjacent cells in Pinus radiata loaded in tension. A delamination is observed where
level the crack tip propagates by separating cell walls at the middle lamella in a
splitting or peeling mode. At the annual ring level “stick-slip type” of crack growth
was observed. Because of non-uniform stress distribution, the cracks deviate from
the pure radial direction namely in earlywood zone. The latewood fracture mostly is
without plane deviation. When crack propagation in earlywood approaches the late-
wood zone, its growth rate decreases and could be arrested in earlywood. Latewood
failure occurred mostly by cell splitting because of weak middle lamella. Dill-
Langer et al. (2002) studied the in-situ the damage mechanisms of crack propagation
in tension perpendicular to the grain, in spruce micro specimens (12 mm3 ) with
initial notch. For spruce at 12% moisture content two different mechanisms were
identified: the rupture of earlywood cell walls when crack propagation is in tangen-
tial direction, and debonding between adjacent tracheids, when crack propagates
in radial direction. The cell wall rupture is related to the meso-scale behaviour of
annual ring structure while the debonding mechanism is very brittle and related to
the micro-scale wood behaviour.
The development of in-situ techniques will serve to the modelling approaches and
for implementation of non linear and anisotropic laws in different fracture models
of wood and wood-based composites.
84 V. Bucur
Fig. 4.21 Fracture morphology for radial cracks in green sapwood of Pinus sylvestris observed in
situ with under tension loading (Thuvander and Berglund 2000, Figures, 4, 7, 8, 10) (a) TR crack
arrested in latewood (b) Two cracks are linked and the bridging zone is torn (c) Crack alignment
in R direction because of the rays (d) TR crack tip in the middle lamella of earlywood. (mode of
crack growth: cell splitting or peeling)
The structure of wood-based composites is spatially much more complex than that
of wood as can be seen from Figs. 4.22 and 4.23 for the fracture surfaces of wood-
based composites tested in tension fracture Mode I (Niemz and Diener 1999).
The failure of adhesive layers introduces new problems in old and new structures.
A delamination test for structural wood adhesives used in thick joints has been
proposed by Lavisci et al. (2001). While the technology to produce wood-based
composites has advanced significantly in last decades, the theories for predicting the
behaviour of these materials advanced less. The industry needs reliable and specific
4 Initiation and Growth of Delamination 85
Fig. 4.22 Fracture surfaces of wood-based composites on ASTM E 399 – 1994 specimens type
tested in tension fracture Mode I (Niemz and Diener 1999, Figure 1, Figure 2, Figure 3) (I) Front
view of: (a)- OSB split parallel to the particle orientation; (b) a- OSB split perpendicular to the
particle orientation; (c) MDF ; (d) plywood – 7 layers (II ) Fracture surfaces of (a)- OSB split
parallel to the particle orientation; (b) a- OSB split perpendicular to the particle orientation; (c)
MDF ; (d) plywood – 7 layers (III) ASTM E 399 Specimen used for delamination testing in wood
based composites
86 V. Bucur
Fig. 4.23 MDF specimens size and orientation (Matsumoto and Nairn 2008, Figure 1); (a) com-
pact tension specimen a= 30.48 mm w= 76.2 mm = 31.75 mm. Note = 0 for the ASTM
specimen (b) orientation of specimens in a panel. The first letter indicates the normal to the crack
and the second one the crack propagation direction
The physical phenomena behind delaminations growth are induced by the appli-
cation of any type of load – compression, tensile in joints, bending or fatigue.
Fracture mechanics is the best tool for the identification of a threshold level for
the growth of delaminations and have a fundamental importance in understanding
the real mechanical behaviour of delaminated composite structures. When design-
ing with wood-based composites the causes associated with the delamination failure
88 V. Bucur
must be taken into account. Delamination management approaches can also strongly
influence the lifecycle costs and the maintenance costs of structures.
In order to efficiently exploit the potential of wood-based composite materials in
structural applications the delaminations initiation and growth must be adequately
predicted and controlled. Multi-scale approaches simulating delamination related
physical phenomena at different levels of detail with different degrees of accuracy
were suggested by Ladis et al. (2002), Smith and Vasic (2003), Moses and Prion
(2004), Smith et al. (2007), and Stanzl-Tschegg and Navi (2009).
Since delamination and fracture process in wood-based composites is with fibers
bridging across the crack plane, the general preferred approaches are based on
energy release rate and not on stress intensity factor.
Crack resistance curves (Ehart et al. 1996, 1998, 1999) were determined for some
wood-based composites (particleboards, MDF, Parallam, etc) with wedge splitting
technique, under the assumption of linear elastic material behavior. Difficulties
determined by the frontal process zone and bridging zone and the measurements of
crack length required the calculation of an effective crack length by normalization
and comparison with an equivalent linear elastic material with no crack tip process
zone. Two models were derived, the plastic energy model and the microcracking
model which relies an effective crack length.
Matsumoto and Nairn (2008) developed an original new energy based method
for crack growth detection in MDF. For crack growth under continuous loading,
detection image correlation method has been developed with simultaneous optical
detection of crack length. In Fig. 4.23 are shown the specimens for four orthogonal
crack directions in a MDF panel. The increment of crack growth Δa was mea-
sured, between two successive images from the shift in the strain profile (Fig. 4.24).
In the case of MDF the unloading curves after crack propagation do not return to
the origin probably because of residual stresses, plasticity or crack-plane interfer-
ence. “Crack-plane interference means the bridging material left in the wake of
the crack cannot be unloaded back to the original specimen configuration. Instead,
Fig. 4.25 R-curve revised method (Matsumoto and Nairn 2008, Figure 4) (a) – integral transfor-
mation of force and crack length data as a function of displacement (b) – energy area to crack
length (c)– R-curve as found from the slope of energy area
Table 4.8 Average values for the initiation toughness (Gc ), the slope of rising R – curve for MDF
specimens and σ c for MDF (Matsumoto and Nairn 2008)
good agreement between discrete and revised analyses of the R-curve for LT frac-
ture for MDF specimens, 19 mm thickness and 609 kg/m3 density. Table 4.8 gives
the initiation toughness (Gc ) and the slope of rising R – curve for MDF specimens
in which the effect of panel density and thickness and of the crack orientation has
been demonstrated.
The originality of the model “material point model” proposed by Nairn and
co-workers compared with previous approaches, started with machined notch, for
which the subsequent process zone does not influence the initial crack growth
(Niemz et al. 1997, 1999, Morris et al 1999, Ehart et al. 1996) is related to the
following points:
4.6 Summary
in safety critical applications. This has led to composite structures being perceived
as expensive to fabricate and needing frequent inspection and repair. The recog-
nise approach to study delamination has been the fracture mechanics. The theory
of fracture mechanics has been successfully applied to wood, wood products and
wood-based composites since more than 50 years and provided valuable concepts
for evaluation of the influence of cracks, notches or other stress raisers in struc-
tural elements. The space – time multi-scale nature of the delamination process in
wood can be related to the prediction of crack nucleation, growth and arrest. Crack
tip displacement is related to crack growth and propagation. The definition of the
damage zone ahead of a crack tip is crucial for the studies of wood fracture. If
the fracture process zone is small compared to the length of the crack, linear elas-
tic fracture mechanics (LEFM) methods yield an accurate prediction of the load
level at which a crack in a structural component will grow. Any deformation of the
crack can be described through a combination of three fracture pure modes: Mode
I – opening mode in tension, Mode II – the in plane shear mode and, Mode III –
the out of plane shear mode. However, mixed fracture modes can be recognised
also. The anisotropic nature of wood allows the development of six different frac-
ture system orientations. For the situations where the fracture process zone is not
small compared with the length of a crack, the energy methods and the concepts
of nonlinear fracture mechanics (NLEFM) can be used. This approach can be used
to accurate prediction of wood fracture behaviour through laboratory tests and in
reliable interpretation of the mechanical capacities of notched small dimension tim-
bers, or structures with mechanical connections made with fastenings (nails, bolts,
shear plates, split rings), etc. A range of failure criteria have been developed based
on the physics of delamination fracture in wood and in wood-based composites.
These criteria included parameters that relate to the influence of loading, material
characteristics and environmental factors. Experimental investigation and predic-
tive (analytical and numerical) modeling are linked through microfractographic
studies.
The fundamental knowledge on fracture behaviour of wood can have relevance
for structural use of timber, in pulping industry, for wood drying technology, or in
processes of machining and cutting.
References
Aicher S (1992) Fracture energies and size effect law for spruce and oak in mode I. Determination
of fracture properties of wood especially in mode II and mixed modes I and II. Proceedings of
the RILEM TC 133 Meeting, Bordeaux
Aicher S, Reinhardt HW, Klöck W (1993) Nichtlineares bruch-mechanik-massstabsgesetz für
Fichte bei Zugbeanspruchungsenkrecht zur Fasserrichtung. Holz als Roh Werkst 5:385–394
Aicher S, Dill Langer G, Ranta – Maunus A (1998) Duration of load effect in tension perpendicular
to the grain of glulam in different climates. Holz Roh-Werkst 56:279–284
Akande JA, Kyanka GH (1990) Evaluation of tensile fracture in aspen using fractographic and
theoretical methods. Wood Fiber Sci 22:283–297
4 Initiation and Growth of Delamination 93
Anderson TL (2005) Fracture mechanics: fundamentals and applications. 3 rd edition, CRC, Boca
Raton
Ando K, Ohta M (1999) Variability of fracture toughness by the crack tip position in an annual
ring of coniferous wood. J Wood Sci 45:275–283
Ashby MF, Eastering KE, Harryson R, Maiti SK (1985) The fracture and toughness of wood. Proc
Roy Soc London A 398(1815):261–280
ASTM E 399 (1994) Standard test method for plain strain fracture toughness of metallic materials
Annual Book of Standards, American Society for Testing Materials, Philadelphia, PA
Atluri SN (1986) Computational methods in the mechanics of fracture. North Holland, Amstrerdam
Attack D, May WD, Morris EL, Sproule RN (1961) The energy of tensile and cleavage fracture in
black spruce. Tappi J 44(8):555–567
Bariska M (1994) Fracture mechanics and wood anatomy. IAWA J 15:199–200
Barrett JD (1976) Effect of crack – front width on fracture toughness of Douglas fir. Eng Fract
Mech 8:711–717
Barrett JD (1981) Fracture mechanics and the design of wood structures. Phil Trans Roy Soc
London 299(8):217–226
Beismann H, Schweingruber F, Speck T, Körner C (2002) Mechanical properties of spruce and
beech wood grown in elevated CO2 . Trees 16:511–518
Beismann H, Wilhelmi H, Baillères H, Spatz HC, Bogenrieder A, Speck T (2000) Brittless of
twig bases in the genus Salix: fracture mechanics and ecological relevance. J Exp Bot. 51:
617–633
Berto F, Lazzarin P (2007) Relationship between J – integral and the strain energy evalu-
ated in a finite volume surrounding the tip of sharp and blunt V-notches. Int J Solid Struct
44(14/15):4621–4645
Boatright SWJ, Garrett G (1983) The effect of microstructure and stress state on the fracture
behavior of wood. J Mat Sci 18:2181–2199
Bodner J, Grüll G, Schlag MG (1996) In-situ fracturing of wood in the scanning electron
microscope. Holzforschung 50:487–490
Bodner J, Schlag MG, Grüll G (1997) Fracture initiation and progress in wood specimens
stressed in tension. Part I: Clear wood specimens stressed parallel to the grain. Holzforschung
51:479–484; Part II: Compression wood specimens stressed parallel to grain Holzforschung 51:
571–576
Bondietti EA, Momoshima N, Shortle WC, Smith KT (1990) A historical perspective on divalent
trends in red spruce stemwood and the hypothetical relationship to acid deposition. Can J For
Res 20:1850–1858
Borgin K (1971) The cohesive failure of wood studied with the scanning electron microscopy. J
Microsc 94(1):1–11
Boström L (1992) Method for determining of the softening behaviour of wood and the applicability
of a nonlinear fracture mechanics model. PhD Thesis LUTVDG/TVBM 1012 Lund Institute of
Technology, Lund, Sweden
Boström L (1988) The fictitious crack model – A fracture mechanics approach applied on wood.
International Conference on Timber Engineering, Seattle DC vol 2:559–565
Boyd JD (1973) Helical fissures in compression wood cells: causative factors and mechanics of
development. Wood Sci Technol 7:92–111
Bucur V (2003) Imaging of microscopic structure of wood. Proceedings of the 5th world congress
on ultrasonics Paris, pp 299–302. http://www.sfa.asso.fr/wcu 2003/procs/webside articles/.
Accessed 20 December 2007
Bucur V (2005) Acousics of wood. Springer, Heidelberg
Bucur V, Rasolofosaon PNJ (1998) Dynamic elastic anisotropy and nonlinearity in wood and rock.
Ultrasonics 36:813–824
Bucur V, Saied A, Attal J (1995) Identification of wood anatomical elements by acoustic
microscopy, 22nd International. Symposium on Acoustical Imaging, Florence, Italy
94 V. Bucur
Bucur V, Herbé C, Nosei G (1994) Annual ring characteristics of Pinus taeda measured by
ultrasonic and X-ray techniques. IAWA J 15(2):121–132
Clair B, Gilles Despaux G, Chanson B, Thibaut B (2000) Utilisation de la microscopie acoustique
pour l étude des propriétés locales du bois : étude préliminaire de paramètres expérimentaux.
Ann For Sci 57:335–343
Coureau JL, Gustafsson Persson K (2006) Elastic layer model for application to crack propagation
problems in timber engineering. Wood Sci Technol 40:275–290
Coureau JL, Gustafsson PJ, Persson K (2006a) Elastic layer model for application to crack
propagation problems in timber engineering. Wood Sci Technol 40:275–290
Coureau JL, Morel S, Gustafsson PJ, Lespine C (2006b) Influence of the fracture softening
behaviour of wood on load-cod curve and R-curve. Mater Struct 40:97–106
Conrad MPC, Smith GD, Fernlund G (2003) Fracture of solid wood: a review of structure and
properties at different length scale. Wood Fiber Sci 35:570–584
Cotterell B (2002) The past, present and future of fracture mechanics. Eng Fract Mech 69:533–553
Davids WG, Landis EN, Vasic S (2003) Lattice models for the prediction of load – induced failure
and damage in wood. Wood Fiber Sci 35:120–134
Davims JP, Rubio JC, Abrao AM (2007) Delamination assessment after drilling medium-density
fibreboard (MDF) by digital image analysis. Holzforschung 61:294–300
Daudeville L (1999) Fracture in spruce: experiment and numerical analysis by linear and non linear
fracture mechanics. Holz Roh Werkst 57:425–432
DeBaise GR (1972) Morphology of wood shear fracture. J Materials 7(4):568 –572
DeBaise GR, Porter AW, Pentoney RE (1966) Morphology and mechanics of wood fracture. Mater
Res Stand 6:493–499
Dill – Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic emis-
sion. In Aicher S (ed) Wood and wood fiber composites. University of Stuttgart, Otto Graff –
Institute, Department of Wood and Timber Engineering, Stuttgart, pp 93 –104
Dill – Langer G, Lütze S, Aicher S (2002) Microfracture in wood monitored by confocal laser
microscope. Wood Sci Technol 36:487–499
Dinwoodie JM (1966) Induction of cell wall dislocations (slip planes) during preparation of
microscope sections of wood. Nature 212:525–527
Dinwoodie JM (1968) Failure in timber. Part I. Microscopic changes in cell wall structure
associated with compression failure. J Inst Wood Sci 21:37–53
Dinwoodie JM (1974) Failure of timber. Part II. The angle of shear through the cell wall during
longitudinal compression stressing. Wood Sci Technol 8:5–67
Donaldson LA (1995) Cell wall fracture properties in relation to lignin distribution and cell
dimensions among three genetic groups of radiate pine. Wood Sci Technol 29:51–63
Donaldson LA (1996) Clonal variation in the fracture properties of radiate pine wood. In
Donaldson LA, Singh AP, Butterfield BG, Whitehouse L (eds) Recent advances in wood
anatomy. New Zealand Forest Research Institute Ltd, Rotoruapp 283–291
Donaldson LA (1997) Ultrastructure of transwall fracture surfaces in Radiata pine wood using
transmission electron microscopy. Holzforschung 51:303–308
Donaldson LA, Singh AP, Butterfield BG, Whitehouse L (1996) Recent advances in wood anatomy.
New Zealand Forest Research Institute Ltd, Rotorua
Dourado N, de Moura MFSF, Morel S, Valentin G, Morais J (2004) A finite element simulation
of the SENB test with crack propagation. Proceedings of the 3rd international Conferenceon
European Society for Woods Mechanics, Vila Real, Portugalpp 67–76
Dourado N, Morel S, de Moura MFSF, Valentin G, Morais J (2008) Comparison of fracture
properties of two wood species through cohesive crack simulations. Compos Part A: 39:
415–427
Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1996) Characterization of crack propagation in
particleboard. Wood Sci Technol 30:307–321
Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1998) Fracture characterization of PARALAM R
–
PSL in comparison to solid wood and particleboard. Wood Sci Technol 32:43–55
4 Initiation and Growth of Delamination 95
Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1999) Mode III fracture energy of wood composites
in comparison to solid wood. Wood Sci Technol 33:391–405
Forest Products Laboratory (1999) Wood handbook – wood as an engineering material. FPL GTR
113, USDA, Forest Service, Madison, WI
Frühmann K, Tschegg EK, Dai C, Stanzl-Tschegg SE (2002) Fracture behaviour of laminated
veneer lumber under Mode I and III loading. Wood Sci Technol 36:319–334
Frühmann K, Burgert I, Stanzl-Tschegg SE, Tschegg EK (2003) Mode I fracture behaviour on the
growth ring scale and cellular level of spruce (Picea abies L. Karst.) and beech (Fagus sylvatica
L.) loaded in the TR crack propagation system. Holzforschung 57:653–660
Gardiner B (1995) The interaction of wind and tree movement in forest canopies. In Coutts MP
Grace J (eds) Wind and trees, Cambridge University Press, Cambridge, pp 41–59
Gibson LJ, Ashby MF (1997) Cellular solids: structure and properties. Cambridge University Press,
Cambridge
Gordon JE, Jeronimidis G (1980) Composites with high work of fracture. Philos Trans R Soc A
294:545–550
Griffith AA (1920) The phenomenon of rupture and flow in solids. Phil Trans R Soc A 221:163–169
Grosser D, Schulz H, Utschig H (1985) Mögliche anatomische Veränderubgen in erkrankten
Nadelbäumen. Holz Roh-Werkst 43:315–323
Guitard D, Castera P (1995) Experimental and mechanical modelling of wind induced trees
sways. In Coutts MP, Grace J (eds) Wind and trees, Cambridge University Press, Cambridge,
pp 195–203
Guo YJ, Nairn JA (2006) Three-dimensional dynamic fracture analysis using the material point
method. CMES 1(1):11–25
Guo YJ, Nairn JA (2004) Calculation of J-integral and stress intensity factors using Material Point
Method. Comput Model Eng Sci 6:295–308
Gustafsson PJ (1985) Fracture mechanics studies of non-yielding materials like concrete. Technical
Report TVBM – 1007 Lund Institute Technology, Sweden
Gustafsson PJ (1988) A study of strength of notched beams. Proceedings of CIB W18 Meeting 21,
Vancouver, September 1988
Hillerborg A (1991) Application of the fictitious crack model to different types of materials. Int
J.Fracture 51:95–102
Hillerborg A, Modeer M, Petersson PE (1976) Analysis of crack formation and crack growth in
concrete by means of fracture mechanics and finite elements. CemConcr Res 6:773–782
Hoffmeyer P, Hanna RB (1989) Electron beam damage during testing of wood in the SEM Wood
Sci Technol 23:211–214
Hofstetter K, Hellmich C, Eberhardsteiner J (2007) Micromechanical modelling of solid-type
and plate-type deformation patterns within softwood materials. A review and an improved
approach. Holzforschung 61:343–351
Homberg S, Persson K, Petersson H (1999) Nonlinear mechanical behaviour and analysis of wood
and fiber materials. Comp Struct 72:459–480
Irwin GR (1957) Analysis of stresses and strain near the end of a crack traversing a plate. J Appl
Mech 24:361–364
Jensen JL (2005a) Quasi-non-linear fracture mechanics analysis of the double cantilever beam
specimen. J Wood Sci 51:566–571
Jensen JL (2005b) Quasi-non-linear fracture mechanics analysis of the splitting failure of single
dowel joints loaded perpendicularly to the grain J Wood Sci 51:559–565
Jensen JL (2005c) Quasi-non-linear fracture mechanics analysis of splitting failure in sim-
ply supported beams loaded perpendicular to grain by dowel joints. J Wood Sci 51:
577–582
Jeronimidis G (1980a) Wood, one of nature’s challenging composite. In Vincent JFV, Currey
JD (eds) The mechanical properties of biological materials Cambridge University Press,
Cambridge, pp 169–182
Jeronimidis G (1980b) The fracture behaviour of wood and the relations between toughness and
morphology. Proc R Soc B-Biol Sci 208:447–460
96 V. Bucur
Jeronimidis G (1976) The fracture of wood in relation to its structure. Leiden botanical Series No
3: 253–265
Jernkvist LO (2001) Fracture of wood under mixed mode loading. Eng. Fract. Mech. Part I
Derivation of fracture criteria 68(5):549–563; Part II Experimental investigation of Picea abies.
68(5):565–576
Johnson JA (1973) Crack initiation in wood plates. Wood Sci 6:151–158
Jung J, Murpjy JF (1983) An investigation of the fracture of butt joints in parallel – laminated
veneer. Wood Fiber Sci 15:116–134
Keith CT, Côté WA Jr. (1968) Microscopic characterization of slip lines and compression failures
in wood cell walls. For Prod J 18(3):67–74
Keunecke D, Stanzl-Tschegg S, Niemz P (2007) Fracture characterisation of yew (Taxus baccata L)
and spruce (Picea abies L.Karst) in the radial-tangential and tangential radial crack propagation
system by a micro wedge splitting test. Holzforschung 61:582–588
King MJ, Sutherland IJ, Le-Ngoc L (1999) Fracture toughness of wet and dry Pinus radiata. Holz
Roh- Werkst 57:235–240
King MJ, Vincent JFV (1998) Fracture energy during clearing of Pinus radiata. Holz Roh Werkst
56:259–265
Koch G, Bauch J, Dünisch O, Seehann G, Schmitt U (1996) Sekundäre Veränderungen im Holz
akut belasteten Ficten in Hochlage des Osterzgebirges. Holz Roh Werkst 54:243–249
Kucera LJ, Bariska M (1982) On the fracture morphology in wood. Part I. A SEM – study of
deformation in wood of spruce and aspen upon ultimate axial compression load. Wood Sci
Technol 16:241–259
Landis EN, Navi P (2009) Modelling crack propagation in wood and wood composites. A review.
Holzforschung 63:150–156
Landis EN, Vasic S, Davids WG, Parrod (2002) Coupled experiments and simulations of structural
damage in wood. Exp Mech 42(4):389–394
Lavisci P, Berti S, Pizzo, B, Triboulot P, Zanuttini R (2001) A delamination test for structural wood
adhesives used in thick joints. Holz Roh Werkst 59:153–154
Lee SH, Quales SL, Schniewind AP (1995) Wood fracture, acoustic emission, and the drying
process. Part 2. Acoustic emission pattern recognition analysis. Wood Sci Technol 30:283–292
Leicester RH (1971) Some aspects of stress fields at sharp notches in orthotropic materials.
Division of Forest Products, Technical Paper No 57, CSIRO Australia
Leicester RH (1973) Effect of size on the strength of structures. Division of Buildingg Research
Tech Paper No 71 CSIRO Australia
Leicester RH (1974) Fracture strength of wood. Paper at 1st Australian Conference Eng. Mater.
University of New South Wales, Sydney
Lespine C (2007) Courbe R et propriétés de fissures cohésive dans la rupture quasi-fragile.
Proceedings of the « XXX Rencontres Universitaires de Génie Civil, Prix René Houpert »
Loidl D, Tschegg EK, Stanzl-Tschegg (2008) Fracturing of wood under superimposed tension and
torsion. Wood Sci Technol 42:61–74
Mark RE (1967) Cell wall mechanics of wood tracheids. Yale University Press, New Haven, CT
Matsumoto N, Nairn JA (2008) The fracture toughness of medium density fibreboard including the
effects of fiber bridging and crack-plane interference. Eng. Fracture Mech 76:2748–2797
Mattheck C (2004) The face of failure in nature and engineering. Verlag, Forschungszentrum
Karlshrue
Mattheck C (1996) Trees – the mechanical design. Springer, Heidelberg
Mattheck C, Bethge K (1991) Failure of trees induced by delamination. Arboricultural J 15:
243–253
Mattheck C, Kubler H (1997) Wood – the internal optimization of trees. Springer, Heidelberg
Mattheck C, Bethge K, Albrecht W (1995) Failure modes of trees and related failure criteria. In
Coutts MP, Grace J (eds) Wind and trees, Cambridge University Press, Cambridge, pp 195–203
Mishnaevsky L, Qing H (2008) Micromechanical modelling of mechanical behaviour and strength
of wood: state of the art review. Comput Mat Sci 44:363–370
4 Initiation and Growth of Delamination 97
Reiterer A, Sinn G (2002) Fracture behaviour of modified spruce wood: a study using linear and
non linear fracture mechanics. Holzforschung 56:191–198
Reiterer A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood species
under mode I loading perpendicular to the grain. Mater Sci Eng A 332::29–36
SAA (1975) AS 1720 – 1979 Timber engineering code. Sydney Standards Association of Australia
Sandford RJ (2003) Principles of fracture mechanics. Pearson Education Inc., Upper Saddle
River, NJ
Sedighi – Gilani M, Navi P (2007) Micromechanical approach to wood fracture by
three-dimensional mixed lattice-continnum model at fiber level. Wood Sci Technol 41:
619–634
Scheffler M, Niemz P, Diener M, Lostig V, Hardtke HJ (2004) Holz Roh Werkst 62(2):
93–100
Schmidt J, Kallske M (2007) Simulation of cracks in wood using a coupled material model for
interface elements. Holzforschung 61:382–389
Schniewind AP (1977) Fracture toughness and duration of load factor. II. Duration factor for cracks
propagating perpendicular to grain. Wood Fiber Sci 9:216–226
Schniewind AP, Centeno JC (1973) Fracture toughness and duration of load factor. I. Six principal
systems of crack propagation and the duration factor for cracks propagating parallel to grain.
Wood Fiber 5(2):152–158
Schniewind AP, Lyon DE (1971) Tensile strength of redwood dimension lumber. I. Relation to
grade and working stress. Forest Prod J 21(7):45–49
Schniewind AP, Lyon DE (1973) A fracture mechanics approach to the tensile strength perpendic-
ular to grain of dimension lumber. Wood Sci Technol 7:45–59
Schniewind AP, Ohgama T, Aoki T, Yamada T (1982) Effect of specific gravity, moisture content,
and temperature on fracture toughness of wood. Wood Sci 15(2):101–109
Schniewind AP, Pozniak A (1971) On fracture toughness of Douglas – fir wood. Eng Fract Mech
2:223–233
Sedighi-Gilani M, Navi P (2007) Micromechanical approach to wood fracture by three-
dimensional mixed lattice continuum model at fiber level. Wood Sci Technol 41:
619–634
Seel J, Ziemmermann T (1998) The fine structure of the cell wall of hardwoods on transverse
fracture surfaces. Holz Roh Werkst 56:365–366
Silva MAL, de Moura MFSF, Morais JJL (2006) Numerical analysis of the ENF test for mode II
wood fracture. Compos Part A:1334–1344
Sippola M, Frühmann K (2002) In situ longitudinal tensile tests of pine wood in an environmental
scanning electron microscope. Holzforschung 56:669–675
Smith I, Chui YH (1994) Factors affecting mode I fracture energy of plantation grown red pine.
Wood Sci Technol 28: 147–157
Smith I, Landis E, Gong M (2003) Fracture and fatigue in wood. Wiley, New York, NY
Smith I, Snow M, Asiz A, Vasic S (2007) Failure mechanism in wood based materials: a review of
discrete, continuum and hybride finite-element representation. Holzforschung 61:352–359
Smith I, Vasic S (2003) Fracture behaviour of softwood. Mech Mater 35:803–815
Smith J, Kallske M (2007) Simulation of cracks in wood using a coupled material model for
interface elements. Holzforschung 61:383–389
Sridharan S (2008) Delamination behaviour of composites Woodhead Publishing, Cambridge,
England.
Stanzl-Tschegg SE, Navi P (2009) Fracture behaviour of wood and its composites. A review.
Holzforschung 63:139–149
Stanzl-Tschegg SE, Filion L, Tschegg EK, Reiter A (1999) Strength properties and density of SO2
polluted spruce wood. Holz Roh Werkst 57:121–128
Stanzl-Tschegg SE, Tschegg EK, Teischinger A (1994) Fracture energy of spruce wood after
different drying procedures. Wood Fiber Sci 26:467–478
Stanzl-Tschegg SE, Tan DM, Tschegg EK (1995) New splitting method for wood fracture
characterization. Wood Sci Technol 29: 31–50
4 Initiation and Growth of Delamination 99
Stanzl-Tschegg SE, Ehart RJA, Tschegg EK (1997) Fracture behaviour of glued wood laminate
compounds. Proceedings of the 9th international conference on fracture, Sydney, Australia
Swinehart DE, Broek D (1995) Tenacity©, fracture mechanics, and unknown coater web breaks.
Tappi Journal 79(2):233–237
Tada H, Paris PC, Irwin GR (2000) The stress analysis of cracks handbook. 3rd edn. ASME, New
York, NY
Tan DM, Stazl-Tschegg S, Tschegg EK (1995) Models of wood fracture in Mode I and Mode II.
Holz als Roh- und Werkst. 53:159–164
Tschegg EK (1986) Equipment and appropriate specimen shapes for tests to measure fracture
values. (in German). Patent no 390328 Österreichisches Patentamt
Tschegg EK, Frühmann K, Stanzl-Tschegg SE (2001) Damage and fracture mechanisms during
mode I and III loading of wood. Holzforschung 55:525–533
Thuvander F, Berglund LA (2000) In situ observations of fracture mechanisms for radial crack in
wood. J Mater Sci 35:6277–6283
Thuvander F, Berglund LA (1998) A multiple fracture test for strain to failure distribution in wood.
Wood Sci Technol 32:227–235
Thuvander F, Wallström L, Berglund LA, Lindberg KAH (2001) Effects of an impregnation
procedure for prevention of wood cell wall damage due to drying. Wood Sci Technol
34:473–480
Thuvander F, Berglund LA (2000) In situ observations of fracture mechanism for radial cracks in
wood. J Mater Sci 35(24):6277–6283
Triboulot P, Asano I, Ohta M (1983) An application of fracture mechanics to the wood cutting
process. Mokuzai Gakk 29:111–117
Triboulot P, Jodin P, Pluvinage G (1984) Validity of fracture mechanics concepts applied to wood
by finite element calculation. Wood Sci Technol 18:51–58
Turkulin H, Holzer L, Richter K, Ssell J (2005) Application of the ESEM in wood research. Part
II. Comparison of operational modes. Wood Fiber Sci 37:565–573
Valentin G, Boström L, Gustafsson PJ, Ranta-Maunus A, Gowda S (1991) Application of frac-
ture mechanics to timber structures. RILEM State of the art report. Res. Note 1262, Technical
Research Centre of Finland, Espoo, Finland
Valentin G, Morlier P (1982) A criterion of crack propagation in timber. Mater Struct 15:88–95
Vasic S, Ceccotti A, Smith I, Sandak J (2009) Deformation rates effects in softwoods. Crack
dynamics with lattice fracture modelling. Eng Fract Mech 76(9):1231–1246
Vasic S, Stanzl-Tschegg S (2008) Softwood/hardwoods fracture at different humidity levels: ESEM
in-situ real time experiments. Holzforschung 62
Vasic S, Stanzl-Tschegg S (2007) Experimental and numerical investigation of wood fracture
mechanisms at different humidity levels. Holzforschung 61:367–374.
Vasic S, Stanzl-Tschegg S (2005) Fracture mechanisms and properties of green wood subjected to
opening Mode I. In Tschegg S, Sinn G(eds) Proceedings of the COST Action E35, Rosenheim
Workshop September 29– 30.
Vasic S, Smith I (2002) Bridging crack model for fracture of spruce. Eng Fract Mech 69:745–760
Vasic S, Smith I, Landis E (2002) Fracture zone characterization–micro-mechanical study. Wood
Fiber Sci 34:42–56
Vasic S, Smith I (2003) Contact – crack problem with friction in spruce. Holz Roh–Werkst
61(3):182–186
Vasic S, Smith I (1996a) On the influence of ultrastructure and fibres bridging in Mode I fracture
of wood. Proceedings of the 2nd international conference on the deevelopment of wood science
/technology and Forestry ICWSF’96, Sopron, Hungary
Vasic S, Smith I (1996b) The brittleness of wood in tension perpendicular to the grain: micro-
mechanical aspects. Proceedings of the COST 508 Wood Mechanics Conference, Stuttgart,
Germany, pp 555–569
Vasic S, Smith I (1998) Bridged crack model of wood fracture: analysis and numerical modelling.
Proceedings of the world timber conference Montreux, 17–20 August, Swiss Federal Institute
of Technology, Lausanne, Suisse, pp 1818–1819
100 V. Bucur
Vasic S, Smith I (1999a) The effect of bridging stress on fracture toughness of wood. RILEM
Symposium Timber Engineering, Stockholm
Vasic S, Smith I (1999b) Failure analysis of tensile strength perpendicular to the grain. RILEM
Symposium Timber Engineering, Stockholm
Walsh PF (1971) Cleavage Fracture of Timber. Div. For. Prod. Tech. Pap. No. 65, CSIRO,
Melbourne
Walsh PF (1972) Linear fracture mechanics in orthotropic materials. Eng Fract Mech 4:533–541
Walsh PF (1973) The interaction of butt joints. J Inst Wood Sci 6(2):22–27
Wang SS (1984) Edge delamination in angle –ply laminates. AIAAA J 22(2):256–264
Wang L, Lu Z, ZhaoG (2003) Wood fracture pattern during the water adsorption process.
Holzforschung 57:639–643
Watanabe K, Landis EN (2007) An acoustic emission –based study of energy dissipation in radi-
ally loaded spruce. In Proceedings of the 3rd international symposium on wood machining.
Lausanne, Switzerland, pp 179–182
Williams JG (1989) The fracture mechanics of delamination tests. J Strain Anal 24(4):207–214
Wittel FK, Dill-Langer G, Kröplin BH (2005) Modelling of damage evolution in softwood
perpendicular to grain by means of a discrete element approach. Comput Mater Sci 32:594–603
Wu EM (1967) Application of fracture mechanics to anisotropic plate. J Appl Mech 34:967–974
Yoshihara H (2001) Influence of span/depth rate on the measurement of mode II fracture toughness
of wood by end – notched flexure test. J Wood Sci 47(1):8–12
Yoshihara H (2003) Resistance curve for the mode II fracture toughness of wood obtained by the
end – notched flexure test under the constant loading point displacement condition. J Wood Sci
49(3):210–215
Yoshihara H (2004) Mode II R-curve of wood measured by 4-ENF test. Eng Fract Mech 71:
2065–2077
Yoshihara H (2005) Mode II initiation fracture toughness analysis for wood obtained by 3 ENF
test. Compos Sci Technol 65:2198–2207
Yoshihara H (2006a) Estimation of the 4–ENF test for measuring the mode III R -curve of wood.
Eng Fract Mech 73(1):42–63
Yoshihara H (2006b) Characterization of fracturing properties of wood and wood based materials
on fracture mechanics. Mokuzai Gakk 52:185–195
Zimmermann T, Sell J, Eckstein D (1994) SEM studies on tension – fracture surfaces of spruce
samples. Holz als – Roh und Werkst 52:223–229
Zho H, Smith I (1991) Influence of drying treatment on bending properties of plantation – grown
white spruce. For Prod. J. 41(3):8–14
Zink AG, Pellicane PJ, Shuler CE (1994) Ultrastructural analysis of softwood fracture surfaces.
Wood Sci Technol 28:329–338
Zink AG, Pellicane PJ, Anthony RW (1995) A stress transformation approach to predicting the
failure mode of wood. Wood Sci Technol 30:21–30
Chapter 5
A Theoretical Model of Collapse Recovery
Philip Blakemore
Contents
5.1 Introduction
The theory that is thought to best explain the recovery of collapse reconditioning
supposes that the S1 and S3 layers are largely responsible for providing restoring
the cells to the un-collapsed shape. This is because these two layers are particu-
larly important in providing circumferential stiffness to each individual cell lumen.
Hence, it is the potential energy stored in these layers that principally provides the
force to restore the cell shape. In contrast, the S2 layer is considered to be the most
important for providing the inelastic material properties required to hold the cell in
the collapsed or deformed state. While moisture content is important for its effect on
the cell wall material properties (i.e. stiffness, creep, mechano-sorptive creep), the
uptake or movement of moisture within the cell walls is not thought to be critical for
collapse recovery. In this sense, the recovery phenomenon can largely be attributed
to a thermal effect (Blakemore and Langrish, 2008), and hence it is the relationships
with temperature for the various material properties which are critical for this mod-
elling work. The effect of heat then is to soften the S2 layer, which is holding the
cell in the deformed shape, allowing the stored mechanical energy in the S1 and S3
layers to restore the cell shape.
P. Blakemore (B)
Department of Materials Science and Engineering, CSIRO, Clayton South, VIC 3169, Australia
e-mail: philip.blackmore@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 101
DOI 10.1007/978-90-481-9550-3_5, C Springer Science+Business Media B.V. 2011
102 P. Blakemore
The three dimensional ultrastructure of the secondary and primary layers in the cell
wall, and the corresponding microfibril orientations, are shown in Fig. 5.1 The FEM
developed tries to incorporate as many of the basic features of this structure as pos-
sible. One of the primary limiations of the Innes (1995) was that it was only for a
single three layered cell in isolation. For the model developed here it was important
to incorporate a double cell wall (S3 , S2 , S1 , CML, S1 , S2 , S3 ).
As introduced above, the important characteristics for collapse recovery are the
effect of temperature on the stiffness properties of the S1 and S3 layers, and the
effect of heat on the plastic properties of the S2 layer. The orthotropic orientation of
these properties within the layers is also crucial. For these reasons, an orthotropic
viscoelastic material model is used for the S1 and S3 layers, while an isotropic
elastic–plastic material model is used for the S2 layer and the CML. Given that
in the S2 layer the radial and tangential properties are thought to be very similar,
this simplification is thought to be acceptable.
In that follows we discuss the viscoelastic material model (S1 , and S3 ) and the
non-linear (elastic plastic) material properties (S2 , and CML).
N
−t
E(t) = E∞ + En exp (5.2)
λn
n=1
where
En = Modulus constant for series n
λn = Relaxation time constant for series n – h
t = Time – h
−C1 (T − Ts )
log [αT ] = (5.3)
C2 + (T − Ts )
where
αT = Shift Factor
C1 = Constant (Specific to Ts )
C2 = Constant (Specific to Ts )
T = Measurement temperature – K
Ts = Reference temperature – K
λnT
αT = (5.4)
λnTs
104 P. Blakemore
where
λnT = Relaxation time constant for series n at the temperature T
λnTs = Relaxation time constant for series n at the reference temperature Ts
5.2.1.2 The Non-linear (Elastic plastic) Material Properties (S2 , and CML)
Innes (1995, 1996) used a non-linear elastic stress-strain relationship in his three
layer, orthotropic, single-cell collapse model. The relationship was essentially bro-
ken into two parts: an initial linear relationship, and then a non-linear relationship.
While MSC.Nastran can handle this form of non-linear relationship reasonably eas-
ily, MSC.Marc had no such capacity. However, given the large strains involved,
in reality some of that strain is likely to be plastic strain and so using an elastic-
plastic relationship, which MSC.Marc can employ readily, seemed appropriate.
This is despite the fact that this material type is mainly used to model smaller
strain behaviour in metals. The MSC.Marc elastic-plastic relationship is by default
isotropic. So unfortunately no orthotropic behaviour could be modelled. The match
between the MSC.Marc elastic-plastic relationship and Innes (1996) used a non-
linear elastic stress-strain relationship can be found in Chapter 6 of Blakemore
(2008).
Fig. 5.2 Multi–cellular collapse model in MSC.Mentat showing areas of different material
properties as developed in previous sections. (Blakemore 2008)
given time period. The increments had to be small enough to allow the solver to con-
verge on the intermittent solutions, particularly to be able to calculate the non-linear
plastic strain in the S2 layer and the CML. Given the temperature dependence of
the properties in both the viscoelastic and elastic-plastic material models, a coupled
solution was undertaken where the node temperatures were fixed at a given temper-
ature for each increment. The internal negative pressure was also loaded linearly in
the same loadcase. Once the full load had been applied, a second loadcase, which
lasted for 16 hrs in total, was undertaken to observe the stress relaxation that occurs
because of the viscoelastic behaviour of the S1 and S3 layers.
In the first instance, a small-strain solver solution was obtained (Fig. 5.3) high-
lighting the first problem with this model, which is, the large amount of plastic shear
strain occurring in the CML between the central cell and the four outer quarter cells.
Using a large strain (Total Lagrange) solver, which is a more realistic method, the
problem becomes even more evident (Fig. 5.4) since there is a large amount of shear
occurring along the axis joining the centre of the four outer quarter cells and in the
centre of the central whole cell, related to these geometric weak points (Fig. 5.2).
Even if the geometry was changed so that the S1 layers shared at least one node
along the centre joining lines, there would still be a large area of CML filling in the
106 P. Blakemore
Fig. 5.3 Cylindrical cell model obtained with the small strain solver. Highlighted in the close up
section is the point between the cells where a large amount of plastic strain is predicted to occur in
the CML. (Blakemore 2008)
gaps between any three adjoining cells. It should be noted that this sharing of nodes
in the above model was not done as it would produce a long thin element of CML
on either sides of the shared node. As a general rule, elements are more likely to
provide a good solution if they are composed of approximately even-length sides.
The next improvement to the model then was to base the cells more on a square
shape, as shown in Fig. 5.5. This square shape is also possibly more realistic
of the type of lumen shape that occurs in the collapse prone group of eucalypts
(Fig. 5.6). One of the reasons that a cylindrical model was attempted first was that
the orthotropic orientations for the viscoelastic model are most easily assigned in
terms of a cylindrical co-ordinate system. Fortunately, in meshing the central cell
shown in Fig. 5.5, the quad elements were generated in a cylindrical pattern such that
a given edge was always on the inside. The orientations of the orthotropic properties
were then transformed to be relative to that edge. To apply a simple compressive dis-
placement to the edge nodes in a similar manner to the previous model would still
produce a similar problem, possibly even worse, of shear planes in the four corners,
5 A Theoretical Model of Collapse Recovery 107
Fig. 5.4 Cylindrical cell model obtained with large strain (Total Lagrange option) solver.
(Blakemore 2008)
and a non-realistic collapsed shape. Instead of this then, a shear displacement in the
y-axis was forced onto the model. A shear force in the x-axis would result in similar
problems to that of applying a compressive force in this direction; except it would
only occur in two opposing corners and not in all four. Given the longer unsup-
ported edges at the top and bottom of the central cell, these edges are inherently
more likely to buckle or collapse and hence the y-axis shearing is the more realistic
for this base model.
The shear displacements were applied to the highlighted nodes (black circles –
Fig. 5.5) on the y-axis mid-plane. A shear displacement, instead of a shear force,
was applied for similar reasons as in the previous stage of the model development
(Fig. 5.4), where a compressive displacement was used instead of a compressive
force. The compressive displacement keeps periodic symmetry, such that all cells
at the boundaries move inward by the same amount. In this instance, while there
was no requirement to maintain a straight edge, there were still problems with how
realistic the resultant stress distribution from the applied shear forces would be. The
displacements of the two outer nodes were constrained to be twice those of their
respective neighbouring inner nodes. The two black circled nodes on the x–axis
mid–plane also had their x-displacement fixed to zero to prevent any rotation of the
model caused by the shear displacements.
108 P. Blakemore
Fig. 5.5 Grid layout, regions of the cell wall and forces for the square-cell based model showing a
much reduced area of CML. The arrows indicate the negative hydrostatic pressure on the inner S3
layer. The other red lines between nodes show the tie and servo links used to enforce the symmetry
considerations for this model. Black dots indicate important nodes for shear displacements and
related symmetry conditions. (Blakemore 2008)
To maintain the cyclical constraints requirement in the y-axis direction, the nodes
along the straight edge, at the top and bottom of the square central unit, were tied in
the y-axis degree of freedom. To ensure the cyclical constraints were met in both the
x and y directions, servo links were used on the nodes on the mid–planes of the four
quarter cells, so that their displacements matched the pattern for the equivalent plane
line on the central cell. A limitation in the software being used meant that the y-axis
displacement of the two side edges could not be enforced, but this was not pursued
further as the x-axis servo links on these nodes did a reasonable job of ensuring that
the collapsed shape in the four quarter cells matched the equivalent section of the
central whole cell. The locations of the servo and tie links are indicated by the red
lines on Fig. 5.5.
From the initial attempts to run this model, Fig. 5.7 shows the final increment
of the model shown in Fig. 5.5 before the solver failed to converge on the next
increment. This problem was largely a meshing issue in the corner areas, one of
which is highlighted in Fig. 5.7 which was experiencing a high level of stress and
5 A Theoretical Model of Collapse Recovery 109
Fig. 5.6 Scanning electron micrograph of E. regnans cross section. C = collapsed fibres, U =
uncollapsed fibres, V = vessel, R = ray cells. (Chafe et al. 1992)
Fig. 5.7 Square-cell based model in collapsed state that highlights a convergence issue in large
strain corners of the S3 layer. (Blakemore 2008)
110 P. Blakemore
became too distorted for the next incremental solution to be found. Figure 5.8 shows
the mesh refinement in this region that was used to overcome this difficulty.
The collapsing period of this model was run with all of the nodal temperatures
fixed at 25◦ C. The internal negative hydrostatic tension and shear (y-axis) displace-
ments were chosen iteratively to just initiate contact on the internal walls (Fig. 5.9).
Fig. 5.9 Final shape of the collapsed square-cell based model with the stress distribution shown
after the 16 hrs of stress relaxation were allowed to occur. (Blakemore 2008)
5 A Theoretical Model of Collapse Recovery 111
More extensive contact started to cause convergence problems for the solver. These
problems may have been related, at least partially, to the artificial shear boundary
conditions that were being applied in this model. The internal pressure applied in
this instance was –4.9 MPa. The shear displacements applied to the four circled
nodes on the horizontal mid-plane (Fig. 5.5) were –3.5, –1.75, 1.75 and 3.5 μm
from left to right respectively.
The internal pressure value of –4.9 MPa is not too dissimilar to the –5.33 MPa
(at 25◦ C) that Innes (1996) obtained to strain the inner edge of his model to 95%
of the value assumed to result in collapse. However, both values are high compared
with the pressures that Kauman (1964) estimated were likely in collapsing cells
which he estimated to be in the range of 1–2.35 MPa, based on the liquid meniscus
having a radius in the range of 600–1000 Å. The high negative pressures in both
this model and the Innes (1996) model could at least partially be explained by the
artificially regular geometry and uniformity of material properties used in both of
these models. In reality, shape and material irregularities are likely to act as weak
points where collapse is initiated at lower negative pressures than required in the
models here. Figure 5.10 shows the shape of the solutions after the negative internal
Fig. 5.10 Final shape of the collapsed square-cell based cell model after the negative hydrostatic
pressure and the shear displacements have been removed. (Blakemore 2008)
112 P. Blakemore
Fig. 5.11 Heat up profile of temperature against time used to simulate reconditioning in the FEM.
(Blakemore 2008)
pressure and the shear displacement boundary conditions have been removed. This
figure shows that the model, despite a little relaxation, essentially maintains the
deformed shape. Figure 5.11 shows the approximation to the measurements of the
central internal board temperature shown in Blakemore and Langrish (2008) that
was used as the basis for the application of heat to simulate reconditioning in this
model. This was applied by using a table of these temperatures to change all of the
nodal temperatures uniformly as a function of time. Figure 5.12 shows the shape and
stress distribution of the model after the nodal temperatures have been increased to
100◦ C to simulate steaming. The recovery of cell shape is barely discernible from
the shape shown in Fig. 5.13.
The next main reason for the lack of collapse recovery in the model relates to the
material properties being used.
As discussed earlier, Eq. (5.1) was central to many of the material properties,
and up to this point a density of 673 kg m–3 was used to be consistent with Innes
(1996). This density is very high compared with most of the experimental mate-
rial that has been used here. For this reason, the model was run again using a more
moderate density value of 500 kg m−3 . The viscoelastic parameters used for this
model are shown in Table 5.1 and the elastic-plastic model was also reconfigured
for this density value. Figure 5.14 shows that the change of basic density to 500 kg
m−3 had little effect on the recovery of the cell shape. The main differences were
that a negative internal pressure of only 3 MPa was required to just initiate internal
contact, and the highest residual stress in the model at the end of the steaming was
predicted to be reduced from ∼55 to ∼36 MPa. Again, this might just highlight how
important the effect of temperature is in this model, and how poorly it is currently
understood at the cell wall scale. All of the material properties used in this model
were also almost entirely obtained by analogy with measured material properties
on small samples or boards. This is largely because of the difficulties in measuring
5 A Theoretical Model of Collapse Recovery 113
Fig. 5.12 Shape and stress distribution after nodal temperatures were increased to 100◦ C.
(Blakemore 2008)
Fig. 5.13 Shape and stress distribution after nodal temperatures were increased to 100◦ C. In this
instance the temperature effect on the elastic moduli in the S1 and S3 layers has been removed and
the values remain fixed. (Blakemore 2008)
Table 5.1 Single term de Prony-series values fitted to E(overall) as a function of time for a density
value of 500 kg m–3
Fig. 5.14 Shape and stress distribution after nodal temperatures were increased to 100◦ C. Model
run assuming a basic density of 500 kg m–3 . (Blakemore 2008)
Table 5.2 Calculated elastic constants for the cell-wall layers at 12% moisture content in Pinus
radiata (Harrington et al. 1998)
NB: “l” refers to a direction parallel to the length of the microfibres, and not the longitudinal
direction of the cell
116 P. Blakemore
the recovery model is that it was attempted with green material properties main-
tained throughout. It is possible that the change of mechanical properties occurring
upon drying, with the cell in the collapsed and stressed state, could be critical for
collapse recovery. Based on analogy again with the mechanical properties of whole
boards, it could be assumed that the modulus of elasticity and shear strength in
the dried state, at around 12% moisture content, are approximately 1.5 times the
green values. However, this board analogy is for the board at two different mois-
ture contents, in which the board is essentially drying stress-free in both states.
Clearly this is not the case at the cellular level in the collapsed state, and hence it
is likely that this analogy would be even more tenuous than the similar assumptions
used up till this point. Nevertheless, given that the viscoelastic properties appear
to have minimal effect on the model attempted here, it may be possible to formu-
late a different time-dependent behaviour to mimic the change in elastic moduli as
moisture is removed. For this to be successful, it would be necessary to include
some component of mechanosorptive strain, which is dependent on some form of
drying model for the change in moisture content. Another initial response to this
might be to try and incorporate a simple drying model. After all, the finite element
method is very suited to analysing this sort of diffusion problem. Unfortunately,
most of the commonly used standard finite element modelling packages typically
only include a diffusion-based heat transfer capability, and have no capability for
modelling moisture content in its own right. The main reason that only a heat trans-
fer capability is included is that these type of software packages are mostly used to
design and test metallic, laminate and elastomeric parts where heat transfer proper-
ties are often very important. The underlying governing equations for heat and mass
diffusion are very similar and it is possible to obtain solutions for simple moisture
diffusion problems by reformulating and analysing them as a heat conduction prob-
lem. However, even if a combined heat transfer and moisture diffusion model could
be easily implemented, the collapse and recovery model will not be progressed sig-
nificantly until a reliable relationship between the mechanical properties, at the cell
wall scale, and (changing) moisture content can be better established. At the moment
there are no good data, apart from at a board scale level again, for these relationships
(Table 5.4).
5
Table 5.4 Results from FEM solution. Three layer orthotropic case (Blakemore 2008)
S3 S2 S1
A Theoretical Model of Collapse Recovery
5.3 Summary
A finite element model was attempted here to demonstrate the theoretical basis for
collapse recovery. The model built on a simple single-cell based model of collapse
that was developed by Innes (1995, 1996). The theoretical mechanism for recov-
ery of collapse by steam reconditioning has essentially been that proposed by Ilic
(1987, personal communication) (cited by Oliver, 1991), which assumes that the S1
and S3 layers are largely responsible for providing an elastic component of the cell
walls and that the S2 layer is considered the most important for providing the inelas-
tic material properties required to hold the cell in the collapsed or deformed state.
To this end, a viscoelastic material model was developed for the S1 and S3 layers,
while an elastic-plastic model was developed for the S2 layer. The model based
on these material properties was not able to clearly demonstrate collapse recovery.
This was largely attributed to the similarity of the dependence of the elastic moduli
as a function of temperature in all cell wall layers. By removing the temperature
dependence in the S1 and S3 layers, a much more significant, although still incom-
plete, recovery of cell shape was demonstrated. The lack of realistic behaviour for
the model predictions has highlighted the paucity of knowledge about mechanical
properties at the cell wall scale. Obviously, direct measurements at this scale are
extremely difficult, if not impossible. The most successful approaches so far to esti-
mate these properties has been to use a range of homogenisation and finite element
modelling techniques based on the generalised knowledge of the cell wall ultra-
structure and the properties of extracted chemical constituents. While not directly
comparable, published values (Harrington et al. 1998) for Pinus radiata at 12%
moisture content suggest that values used in this model may have been rather low.
To some extent this may have compensated for the stiffening effect of the sim-
ple, but inherently stable, geometry used in this model. Even though the method
used by Harrington et al. (1998) could have been used to determine better elastic
moduli for the different cell wall layers than those used, it was not attempted here
because there are still several critical limitations with this approach. These include
that the method is possibly less reliable at high moisture content states and that it
provides no additional information on critical behaviours such as non-linear large
deformation stress-strain relationships, time or temperature-dependent behaviour, or
moisture content (including moisture change or mechanosorptive strain) dependent
behaviour. All of which may be critical for accurately modelling the deformation
and stress distribution in the cell wall layers prior to steam reconditioning. Even if
alternate attempts to simplify the moisture related behaviours were pursued, but that
still accounted for the significant reduction in collapse recovery below 15% mois-
ture content, the lack of good temperature-dependent data in the different secondary
cell wall layers is currently a major impediment for developing the current model
further.
The other major improvement that could be made to the model developed here
would be to include multiple cells with more realistic geometries and arrangements.
Such an approach was attempted in the models by Astley et al. (1998), where real
cross-sections of Pinus radiata tracheids were scanned and skeletonised to form the
5 A Theoretical Model of Collapse Recovery 119
geometrical basis of a finite element model used to predict the macroscale elastic
properties. Unfortunately, while the skeletonisation process makes the realistic cell
geometries much simpler to implement in a finite element model, it is not possi-
ble here. Mostly this is because it is the two-dimensional spatial arrangement of the
different secondary cell wall layers that is considered critical to the collapse and col-
lapse recovery behaviour. Scanning in real cross-sections is still a possibility, but,
it would require much more complex image analysis programming to approximate
where the cell wall layer boundaries occurred for the relevant material properties to
be applied to the appropriate elements. The number of elements required in this type
of approach would also make the finite element model considerably more computa-
tionally intensive. Nevertheless, as computer processing continues to become faster
and cheaper, even in the near future this is unlikely to be a significant restraint for
a model with up to 100 cells. This approach would also largely avoid the need for
artificial constraints in the current model, such as the shear displacements required
to achieve a more realistic flattening of the cell lumen.
References
Astley RJ, Stol KA, Harrington JJ (1998) Modelling the elastic properties of softwood. Part II: The
cellular microstructure. Holz Roh Werkst 56:43–50
Blakemore P (2008) Optimisation of steam reconditioning for regrowth-ash and planta-
tion grown eucalypt species. PhD Thesis, The University of Sydney. 327 pp. http://hdl.
handle.net/2123/2343. Accessed 3 August 2010
Blakemore PA, Langrish TAG (2008) Effect of pre-drying schedule ramping on collapse recovery
and internal checking with Victorian Ash eucalypts. Wood Sci Technol 42(6):473–492
Chafe SC, Barnacle JE, Hunter AJ, Ilic J, Northway RL, Rozsa AN (1992) Collapse: an
introduction. CSIRO Division of Forest Products, Melbourne, 9 pp
Harrington JJ, Booker R, Astley RJ (1998) Modelling the elastic properties of softwood. Part 1:
The cell wall lamellae. Holz Roh Werkst. 56:37–41
Innes TC (1995) Stress model of a wood fibre in relation to collapse. Wood Sci Technol 29:363–376
Innes TC (1996) Improving seasoned hardwood timber quality with particular reference to
collapse. PhD Thesis, Faculty of Engineering, University of Tasmania, Tasmania, 207 pp
Kauman WG (1964) Cell collapse in wood. CSIRO Division of Forest Products, Melbourne, 59 pp
Oliver AR (1991) A model of the behaviour of wood as it dries (with special reference to Eucalypt
materials). Civil and Mechanical Engineering Department, University of Tasmania, Tasmania,
107 pp
Wardrop AB, Bland DE (1959) The process of lignification on woody plants. In: Proceedings of
the 4th international congress of biochemestry. Pergamon Press, New York, NY, pp 76–81
Williams M.L, Landel R.F, Ferry John D (1955) The temperature dependence of relaxation
mechanisms in amorphous polymers and other glass-forming liquids. J Am Chem Soc
77:3701–3707
Part II
Methodology for Delamination Detection
and Factors Inducing and Affecting
Delamination
Chapter 6
Delamination of Wood at the Microscopic Scale:
Current Knowledge and Methods
Lloyd Donaldson
Contents
L. Donaldson (B)
Bioproduct Development, Scion - Next Generation Biomaterials, 49 Sala St. Rotorua,
Private Bag 3020, Rotorua 3046, New Zealand
e-mail: lloyd.donaldson@scionresearch.com
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 123
DOI 10.1007/978-90-481-9550-3_6, C Springer Science+Business Media B.V. 2011
124 L. Donaldson
Fig. 6.1 Diagram of cell wall delamination for both radial and tangential planes, showing tran-
swall and intrawall fracture types. Interwall fracture is a special case of intrawall fracture directly
through the middle lamella or directly between individual tracheids or fibres
Fig. 6.3 Light micrographs of radiata pine showing: (a) Tangential delamination in thin
walled mild compression wood with examples of both transwall (t) and intrawall (i) fracture.
(b) Tangential delamination in thick walled tracheids showing exclusively intrawall fracture. (c)
Radial delamination associated with internal checking showing intrawall fracture at the middle
lamella (ml) region. Scale bar = 30 μm
126 L. Donaldson
bordered pits during the 270 years in service. Fracture in new wood (<3 years-old)
was dominated by intrawall delamination.
The ultrastructural characteristics of weathered pine wood (also from a Buddhist
temple) were examined by TEM and UV microscopy (Kim et al. 2008). Both middle
lamella and S3 layers showed dissolution and separation, accompanied by delami-
nation of the S1 layer in more degraded regions. The S2 layer showed the greatest
resistance to weathering but eventually disintegrated due to breakdown of the cell
wall matrix.
Internal checking (checking enclosed within the piece of timber), and intra-ring
checking (checking enclosed within a growth ring) are characterised by radial
delamination associated with the wood rays (Booker et al. 2000; Putoczki et al.
2007). Wood rays represent a natural point of weakness because of their perpendic-
ular orientation to the wood grain, their reduced lignification, and thin cell walls.
Every tracheid contacts with at least one ray, hence any radial delamination must
involve rays at some point. Whether checking initiates at the ray or merely fol-
lows the path of least resistance is not known, but at least in the case of intra-ring
checking, the cell wall failure occurs well above fibre saturation point (Booker et al.
2000). Both internal checks and intra-ring checks typically result from delamination
between adjacent tracheid, fibre or ray cell cell walls (Putoczki et al. 2007).
6.1.4 Shelling
This is very uncommon, but is known to occur particularly in Hoop pine (Araucaria
cunninghamii D. Don). A more extreme form of shelling associated with severe
drought on certain types of site, such as sand hills, has also been described in
radiata pine (Pinus radiata D.Don). This type of shelling results from reduced
lignification associated with tangential bands of affected tracheids, resulting in
intrawall delamination (Barnett 1976; Donaldson 2002) (Fig. 6.4).
Compression wood is the reaction wood formed on the lower side of leaning stems
and branches in softwoods as a response to gravitational stress. The S2 layer of
severe compression wood tracheids may often show radial delaminations known as
helical checks. These checks are associated with an absence of the S3 layer, and
with swelling caused by increased lignification of the outer S2 region (Boyd 1972;
Yamamoto 1998; Bamber 2001).
In larch (Larix decidua Mill.) compression wood under radial longitudinal shear,
shear strength was greater than comparable normal wood controls (Gindl and
Teischinger 2003). Compression wood shows intrawall fracture, exposing the S1
layer on the fracture surface, and this difference may occur as a result of changes
in microfibril angle and lignification, as well as the presence of intercellular spaces
(Gindl and Teischinger 2003). In compression wood, there is greater lignification
of the outer S2 region and less difference in microfibril angle between the S1 and
S2 layers, making fracture at the S1 /S2 transition more difficult. In addition, the
middle lamella region is less lignified in compression wood than in normal wood
(Donaldson 2001), and the presence of spaces at the cell corners may result in
notching effects (Gindl and Teischinger 2003). In contrast, normal wood showed
transwall and intrawall fracture between the S1 and S2 layers during fracturing. This
130 L. Donaldson
Huang (1995) used ultrasonic treatment to induce delamination of wood cell walls
in order to observe microfibril orientation. This was most effective on large diame-
ter tracheids with high microfibril angles. Treatment with Congo red was found to
enhance checking. This type of delamination appears to follow cellulose microfib-
rils in the S2 layer but the ultrastructure was not investigated in any detail. Similar
delaminations are induced by iodine precipitation within micro-porosities of the cell
wall (Donaldson and Frankland 2004).
The wood cell wall has a complex hierarchical organization, starting with cellu-
lose microfibrils in the form of 3 nm diameter strings of undefined length (Fengel
1978; Donaldson and Singh 1998). These microfibrils group together in clusters
to form macrofibrils 20–30 nm in diameter, embedded in an amorphous matrix
of lignin and hemicellulose (Donaldson 2007). The cell wall has 4 layers, an
outer amorphous layer called the compound middle lamella, incorporating the pri-
mary cell wall, which contains randomly oriented cellulose microfibrils and is
highly lignified. The secondary cell wall, which characterises wood cells, con-
tains three layers: an outer S1 layer with transversely oriented microfibrils and
relatively low levels of lignification, a wide S2 layer with variable but largely axial
6 Delamination of Wood at the Microscopic Scale 131
Wood shows anisotropic fracture properties depending on the method of stress appli-
cation (compression, tension or shear), and the direction of the applied force in
relation to the grain direction. Under compression, cell walls crumple, forming slip
planes or compression failures (Kucera and Bariska 1982; Wilkins 1986a, b), while
tension or shear results in fracture within (intrawall) or across (transwall) the cell
wall (Côté and Hanna 1983; Donaldson 1995, 1996) (Fig. 6.1). Intrawall fracture is
favored in thick cell walls with large microfibril angles, while transwall fracture is
favored in thin-walled cells with small microfibril angles (Wardrop 1951; Wardrop
and Addo-Ashong 1965; Côté and Hanna 1983; Ashby et al. 1985; Donaldson 1996;
Fruhmann et al. 2003b) (Fig. 6.3).
Analysis of the locations of intrawall delamination on tangential tracheid walls
provides information on the nature of preferred failure boundaries within the cell
wall (Donaldson 1995, 1996). Intrawall delamination mainly occurs at three differ-
ent locations: within the middle lamella, between the middle lamella and S1 layer,
and between the S1 layer and S2 layer (Côté and Hanna 1983). Failure probably also
occurs within the S1 layer itself, since examination of fracture surfaces by polarized
light microscopy (in transverse view) or by scanning electron microscopy (in lon-
gitudinal view) often shows exposure of the S1 layer (Koran 1967, 1968; Côté and
Hanna 1983; Zink et al. 1994; Donaldson 1995, 1996). This implies that the S1 layer
is a weak point that favors delamination. The ultrastructural features that contribute
to this have been investigated, suggesting that reduced lignification and the orien-
tation of cellulose microfibrils are important factors in determining the position of
fracture (Donaldson 1995, 1996).
Donaldson (1997) used TEM to study transwall fracture in radiata pine, showing
the characteristic stepping between cell wall layers that occurs in this type of fracture
(Fig. 6.5). Some evidence for plastic deformation within a small distance of the
fracture face was also described.
132 L. Donaldson
In softwoods, the compound middle lamella (including the primary cell wall)
lacks highly oriented cellulose, and is thus relatively isotropic, tending to show
very smooth intrawall fractures but somewhat irregular transwall fractures (Fig. 6.6)
(Donaldson 1997). There are significant changes in cell wall properties at the
S1 /middle lamella boundary from an isotropic material to a highly anisotropic mate-
rial, because of the highly oriented cellulose microfibrils, and this seems to favor
intrawall fracture. Putoczki et al. (2007) have used TEM to study delamination asso-
ciated with intra-ring checking. This delamination occurred mainly at the compound
middle lamella/S1 interface (80%), possibly in association with reduced lignification
in the S1 region (Donaldson 1995).
The interface between the S1 and S2 layers is a less abrupt change, with microfib-
rils gradually reorienting and forming a transition zone (Brändström et al. 2003;
Donaldson and Xu 2005). Nonetheless, fractures that expose the S2 layer tend to
be the most common (Koran 1967, 1968; Côté and Hanna 1983; Zink et al. 1994;
Donaldson 1995, 1996). Transwall fractures are less common and appear to involve
some degree of plastic deformation (Donaldson 1997), but are favored by thin tra-
cheid walls, with some indication of a genetic influence (Koran 1967, 1968; Côté
and Hanna 1983; Zink et al. 1994; Donaldson 1997). These same features may be
involved in internal checking (Putoczki et al. 2007) and other types of delamination.
Schmitt et al. (1996) used TEM to study brash (low level of shock resistance)
and tough (high level of shock resistance) fracture modes in Hickory (Carya spp).
Transwall fracture showed stepping up at the S1 / S2 transition or between the pri-
mary wall and S1 layer (Schmitt et al. 1996; Donaldson 1997). The middle lamella
6 Delamination of Wood at the Microscopic Scale 133
region was the most common point of failure in these samples. The relatively smooth
transwall fracture face of brash samples was associated with compression failures
in the cell wall, while tough specimens showed a highly irregular, mainly longitu-
dinal fracture face, associated with microcracks up to several millimeters from the
fracture, that were absent in brash samples. The microcracks may be involved in
dissipation of energy at some distance from the fracture plane in the tough samples.
The morphological differences explain variations in shock resistance between spec-
imens of similar density. Transwall fracture surfaces show a thin, densely staining
layer similar to that described by Donaldson (1997) suggesting a thin layer of elastic
deformation at the surface.
TEM has also been used to study surface structure, and delamination, in pulp
fibres resulting from beating and refining treatments (McIntosh 1967; Page and De
Grâce 1967; Kibblewhite 1972; Singh and McDonald 2000; Singh et al. 2003; Molin
and Daniel 2004; Brändström et al. 2005; Billosta et al. 2006). In the pulping pro-
cess, the site of fracture in softwoods has an important, species-dependent practical
ramification. Specifically, spruce has a lower energy requirement for mechanical
pulping compared to pine. This is because in spruce, the S1 layer is easily removed
by refining, whereas in pine, the S1 layer detaches from the S2 layer, but remains
134 L. Donaldson
Using thin wood foils, small angle X-ray diffraction, and a video extensometer,
Reiterer et al. (2001) studied the deformation and energy absorption of wood cell
walls under tensile loading. Microfibril angle was found to influence both longitu-
dinal extensibility, and deformation perpendicular to load. Maximum longitudinal
◦ ◦
strain increased from 0.5% to 11% as microfibril angle increased from 5 to 50 .
Most of the increased extensibility at higher microfibril angles was due to irre-
versible deformation of the cell wall. Reiterer et al. (2001) also found that tangential
◦
strain increased with microfibril angle, reaching a maximum at 27 . Tensile strength
◦
decreased with increasing microfibril angle, from 220 MPa at 5 to 35 MPa at
◦
50 . Energy absorption capacity was found to increase with microfibril angle, as
shown in the morphology of the fracture surface, which was smooth for low MFA,
indicating brittle behavior, and heavily torn and deformed in samples with high
MFA, indicating ductile behavior. The fraction of absorbed energy resulting from
elastic deformation was only about 10% in samples with high microfibril angles
(Stanzl-Tschegg 2006).
6 Delamination of Wood at the Microscopic Scale 135
2008). In the longitudinal orientation, beech and spruce had similar specific fracture
energy under torsional load.
the presence of cell wall delaminations, including rupturing and cracking within the
S2 layer in the abrasive planed sample (Murmanis et al. 1983).
Koran (1967) examined the effect of temperature on the structure of radial longitu-
dinal tensile fractures. The amount of transwall fracture was found to change with
◦
temperature, the maximum of 50% being at 0 C with a non-linear decline to less
◦
than 5% transwall fracture at 200 C. Transwall fracture also declined rapidly below
◦ ◦
0 C with less than 5% transwall fracture at –100 C. Temperature also influenced
the nature of surfaces produced by intrawall fracture, higher temperatures produc-
ing smoother surfaces with fewer interlayer transitions. The way in which bordered
pits fractured also changed, with separation at both the pit margin and pit border at
◦ ◦
25 , but only at the pit margin at 200 C. Fracture involving rays was predominantly
transwall at low temperatures, but was intrawall between ray cells and tracheids, at
higher temperatures. In a similar study on tangential walls, Koran (1968) found that
◦ ◦
at temperatures below 100 C, mainly S1 surfaces were exposed, while at 150 or
above, primary wall/middle lamella was exposed. This difference probably reflects
the effect of matrix softening at higher temperatures.
The effect of temperature on transverse tensile fracture was studied by Woodward
(1980). Thermal softening of hemicellulose in the cell wall matrix was likely to be
responsible for the observed effects, including more changes in fracture plane at
◦
higher temperatures (77 C) and indications of plastic strain. Changes in smoothness
of the fracture surface were also observed, confirming the earlier results of Koran
(1967, 1968).
Heat treatment of wood has been used to effectively improve dimensional sta-
bility and durability. Boonstra et al. (2006a, b) have studied the occurrence of
delamination in heat-treated wood using SEM. Heat-treated beech and birch showed
radial cracks adjacent to the rays. Heat treatment may therefore result in increased
brittleness with delamination between fibres.
the sample in resin should be considered. This will preserve the structural integrity
of the wood sample without inducing further damage, and also allow thin sec-
tioning for examination of the fracture surface (Donaldson 1995, 1996, Donaldson
et al. 2007).
Samples should be prepared as 1–3 mm long blocks with some means of iden-
tifying the fracture surface. For example, a radial or tangential fracture surface
can be removed from the wood sample by using a sledge microtome to prepare
a 1–200 μm-thick slice from the fracture surface. Once embedded and sectioned
in cross-section, it should be possible to differentiate the fracture surface from the
microtome-cut surface, where many lumens will be exposed by the cutting. Prepared
blocks should be dehydrated in acetone and embedded in Spurr resin. Cross-sections
can then be cut with a glass knife, keeping the block face as small as practical to
avoid any distortion, which can still occur even with embedded material. Sections
can be stained with toluidine blue and examined with combined brightfield and
polarised light microscopy, to determine the cell wall layers present at the fracture
surface (Donaldson 1995, 1996). Similar techniques can be applied, for example,
to glue-lines, although in some cases cutting unembedded material with a sledge
microtome may be a satisfactory approach, especially for harder materials (Singh
and Dawson 2006). Serial sectioning and 3D reconstruction have been used in com-
bination with light microscopy, to examine the complex delamination associated
with sawn surfaces (Donaldson et al. 2007).
Stanzl-Tschegg 2006; Putoczki et al. 2007). For SEM, samples are mounted directly
for observation of 3D surfaces, requiring only sputter-coating with metal to reduce
charging effects. If environmental SEM (ESEM) or low-voltage SEM (LVSEM) are
available, then even the sample coating can be eliminated, potentially allowing re-
examination of the same sample after successive treatments (Ando and Onda 1999;
Fruhmann et al. 2003a; Turkulin et al. 2005a, b; Donaldson 2007). Stereoscopic
imaging of rough fracture surfaces can also be very useful, although published
examples are rare (Donaldson et al. 2007) (Fig. 6.5).
The use of TEM to study delamination is less common than SEM because of its
more complicated specimen preparation. Its main advantage is the high resolution
imaging of fracture faces and in particular the identification of cell wall layers
present at the surface (Donaldson 1995, 1997; Putoczki et al. 2007) (Fig. 6.6).
TEM is however restricted to relatively small fields of view so is not well suited
to characterization of large fracture zones.
6.4 Summary
Delamination of wood exhibits a complex anisotropic behavior that is strongly
related to the microscopic structure of the wood, and to the presence and distribution
of microscopic defects. Microscopic analysis is therefore an important tool in under-
standing the delamination properties of wood and wood-based materials, as shown
by recent attempts to model this behavior, which have shown the need to incorpo-
rate 3D morphology into the mechanical models (Nairn 2006; Sedighi-Gilani et al.
2006; Sedighi-Gilani and Navi 2007).
References
Ando K, Hirashima Y, Sugihara M, Hirao S, Sasaki Y (2006) Microscopic processes of shear-
ing fracture of old wood, examined using the acoustic emission technique. J Wood Sci 52:
483–489
Ando K, Onda H (1999) Mechanism for deformation of wood as a honeycomb structure I: effect of
anatomy on the initial deformation process during radial compression. J Wood Sci 45:120–126
Ashby MF, Easterling KE, Harrysson R, Maiti SK (1985) The fracture toughness of woods. Proc
R Soc London Ser A: Math Phys Sci 398:261–280
Bamber RK (2001) A general theory for the origin of growth stresses in reaction wood: How trees
stay upright. IAWA J 22:205–212
Bamber RK, Summerville R (1981) Microscopic studies of the weathering of radiata pine sapwood.
J Inst Wood Sci 9:84–88
Barnett JR (1976) Rings of collapsed cells in Pinus radiata stemwood from lysimeter-grown trees
subjected to drought. NZ J For Sci 6:461–465
Bergander A, Salmén L (1997) Lamellar cracks, do they appear during radial compression of wood.
Nordic Pulp Paper Res J 12:216–219
Billosta V, Brändström J, Cochaux A, Joseleau JP, Ruel K (2006) Ultrastructural organization of the
wood cell wall can explain modifications caused in fibres during the pulping process. Cellulose
Chem Technol 40:223–229
140 L. Donaldson
Blanchette RA, Haight JE, Koestler RJ, Hatchfield PB, Arnold D (1994) Assessment of deteriora-
tion in archeological wood from ancient Egypt J Am Inst Conserv 33:55–70
Boatright SWJ, Garrett GG (1983) The effect of microstructure and stress state on the fracture
behaviour of wood. J Mat Sci 18:2181–2199
Booker RE, Haslett A, Sole J (2000) Acoustic emission study of within-ring internal checking
in radiata pine. Proceedings 12th international symposium on nondestructive testing of wood.
University of Western Hungary, Sopron, 13–15 September 2000
Boonstra MJ, Rijsdijk JF, Sander C, Kegel E, Tjeerdsma B, Militz H, J van Acker, Stevens M
(2006a) Microstructural and physical aspects of heat treated wood. Part 1. Softwoods. Maderas
Ciencia y Technologia 8:193–208
Boonstra MJ, Rijsdijk JF, Sander C, Kegel E, Tjeerdsma B, Militz H, J van Acker, Stevens
M (2006b). Microstructural and physical aspects of heat treated wood. Part 2. Hardwoods.
Maderas Ciencia y Technologia 8:209–217
Borgin K (1971a) The mechanism of the breakdown of the structure of wood due to environmental
factors. J Inst Wood Sci 28:26–30
Borgin K (1971b) The cohesive failure of wood studied with the scanning electron microscope. J
Microsc 94:1–11
Borgin K, Parameswaran N, Liese W (1975) The effect of aging on the ultrastructure of wood.
Wood Sci Technol 9:87–98
Boyd JD (1972) Tree growth stresses. V. Evidence of an origin in differentiation and lignification.
Wood Sci Technol 6:251–262
Brändström J, Bardage SL, Daniel G, Nilsson T (2003) The structural organisation of the S1 cell
wall layer of Norway spruce tracheids. IAWA J 24:27–40
Brändström J, Joseleau JP, Cochaux A, Giraud-Telme N, Ruel K (2005) Ultrastructure of
commercial recycled pulp fibres for the production of packaging paper. Holzforschung 59:
675–680
Chafe SC (1977) Radial dislocations in the fibre wall of Eucalyptus regnans trees of high growth
stress. Wood Sci Technol 11:69–77
Clair B, Gril J, Baba K, Thibaut B, Sugiyama J (2005) Precautions for the structural analysis of
the gelatinous layer in tension wood. IAWA J 26:189–195
Côté WA, Hanna RB (1983) Ultrastructural characteristics of wood fracture surfaces. Wood Fibre
Sci 15:135–163
Dadswell HE, Langlands I (1934) Brittleheart in Australian timbers: a preliminary study. J CSIR
(Aust.) 7:190–196
Daniel G, Volc J, Niku-Paavola ML (2004) Cryo-FESEM and TEM immuno-techniques
reveal new details for understanding white-rot decay of lignocellulose. CR Biologies 327:
861–871
Dill-Langer G, Lütze S, Aicher S (2002) Microfracture in wood monitored by confocal laser
scanning microscopy. Wood Sci Technol 36:487–499
Donaldson LA (1993) Distribution of inorganic elements in sub-fossil Totara wood. Holzforschung
47:451–57
Donaldson LA (1995) Cell wall fracture properties in relation to lignin distribution and cell
dimensions among three genetic groups of radiata pine. Wood Sci Technol 29:51–63
Donaldson LA (1996) Clonal variation in the fracture properties of radiata pine wood. In:
Donaldson LA, Singh AP, Butterfield BG, Whitehouse L (eds) Recent advances in wood
anatomy, NZFRI Ltd, Rotorua, pp 283–291
Donaldson LA (1997) Ultrastructure of transwall fracture surfaces in radiata pine wood using
transmission electron microscopy and digital image processing. Holzforschung 51:303–308
Donaldson LA (2001) Lignification and lignin topochemistry – an ultrastructural view.
Phytochemistry 57:859–873
Donaldson LA (2002) Abnormal lignin distribution in wood from severely drought stressed Pinus
radiata trees. IAWA J 23:161–178
Donaldson LA (2007) Cellulose microfibril aggregates and their size variation with cell wall type.
Wood Sci Technol 41:443–460
6 Delamination of Wood at the Microscopic Scale 141
Koran Z (1967) Electron microscopy of radial tracheid surfaces of black spruce separated by tensile
failure at various temperatures. Tappi 50(2):60–67
Koran Z (1968) Electron microscopy of tangential tracheid surfaces of black spruce separated by
tensile failure at various temperatures. Svensk Papperstidning 17:568–576
Kucera LA, Bariska M (1982) On the fracture morphology in wood. Wood Sci Technol 16:241–259
Lee KH, Singh AP, Kim YS (2007) Cellular characteristics of a traumatic frost ring in the
secondary xylem of Pinus radiata. Trees 21:403–410
Loidl D, Tschegg EK, Stanzl-Tschegg SE (2008) Fracturing of wood under superimposed tension
and torsion. Wood Sci Technol 42:61–74
Maurer A, Fengel D (1991) Electron microscopic representation of structural details in softwood
cell walls by very thin ultramicrotome sections. Holz Roh-Werstoff 49:53–56
McIntosh DC (1967) The effect of refining on the structure of the fibre wall. Tappi J 50:482–488
Molin U, Daniel G (2004) Effects of refining on the fibre structure of kraft pulps as revealed by
FESEM and TEM: influence of alkaline degradation. Holzforschung 58:226–232
Murmanis L, River BH, Stewart H (1983) Microscopy of abrasive-planed and knife-planed surfaces
in wood-adhesive bonds. Wood Fibre Sci 15:102–115
Nairn JA (2006) Material point method simulations of transverse fracture in wood with realistic
morphologies. Holzforschung 61:375–381
Nilsson T, Daniel G (1990) Structure and the aging process of dry archaeological wood. In: Rowell
RM, Barbour RJ (eds) Archaeological wood: properties, chemistry, and preservation, Advances
in chemistry series 225. American Chemical Society, Washington, DC, pp 67–86
Page D, De Grâce JH (1967) The delamination of fibre walls by beating and refining. Tappi J
50:489–495
Petterson RP, Bodig J (1983) Prediction of fracture toughness of conifers. Wood Fibre Sci 15:
302–316
Putoczki TL, Nair H, Butterfield B, Jackson SL (2007) Intra-ring checking in Pinus radiata D.Don:
The occurrence of cell wall fracture, cell collapse and lignin distribution. Trees 21:221–229
Reiterer A, Lichteneger H, Fratzl P, Stanzl-Tschegg SE (2001) Deformation and energy absorption
of wood cell walls with different nanostructure under tensile loading. J Mat Sci 36:4681–4686
Reiterer A, Tschegg S (2002) The influence of moisture content on the mode I fracture behavior of
spruce wood. J Mat Sci 37:4487–4491
Sandberg D (1999) Weathering of radial and tangential wood surfaces of pine and spruce.
Holzforschung 53:355–364
Sandberg D, Söderström O (2006) Crack formation due to weathering of radial and tangential
sections of pine and spruce. Wood Mat Sci Eng 1:12–20
Schmitt U, Richter HG, Wittke B (1996) Fracture morphology of Hickory (Carya spp.,
Juglandaceae) under single-blow impact loading. IAWA J 17:151–160
Schniewind AP, Ohgama T, Aoki T, Yamada T (1982) Effect of specific gravity, moisture content
and temperature on fracture toughness of wood. Wood Sci 5:101–109
Sedighi-Gilani M, Job L, Navi P (2006) Three-dimensional modeling of wood fracture in mode I,
perpendicular to the grain direction at fibre level. Wood Mat Sci Eng 1:52–58
Sedighi-Gilani M, Navi P (2007) Micromechanical approach to wood fracture by three-
dimensional mixed lattice-continuum model at fibre level. Wood Sci Technol 41:619–634
Sell J, Leukens U (1971) Investigations of weathered wood surfaces Part II: Weathering phenomena
of unprotected wood species. Holz-Roh Werkstoff 29:23–31
Sell J, Zimmermann T (1993) Radial fibril agglomerations of the S2 on transverse-fracture surfaces
of tracheids of tension-loaded spruce and white fir. Holz Roh-Werkst 51:384
Singh AP, Anderson CR, Warnes JM, Matsumura J (2002) The effect of planning on the micro-
scopic structure of Pinus radiata wood cells in relation to penetration of PVA glue. Holz als
Roh und Werkstoff 60:333–341
6 Delamination of Wood at the Microscopic Scale 143
Singh AP, Dawson BSW (2003) The mechanism of failure of clear coated wooden boards as
revealed by microscopy. IAWA J 24:1–11
Singh AP, Dawson BSW (2004) Confocal microscope – A valuable tool for examining wood-
coating interface. JCT Res 1:235–237
Singh AP, Dawson BSW (2006) Microscopic assessment of the effect of saw-textured Pinus
radiata plywood surface on the distribution of a film-forming acrylic stain. JCT Res 3:193–201
Singh AP, Dunningham EA, Plackett DV (1995) Assessing the performance of a commercial wood
stain by transmission electron microscopy. Holzforschung 49:255–258
Singh AP, Kim YS, Chung GC, Park BD, Wong AHH (2003) TEM examination of surface
characteristics of rubberwood (Hevea brasiliensis) HTMP fibres. Holzforschung 57:579–584
Singh AP, McDonald AG (2000) Comparison of radiata pine and rubberwood HTMP fibres by
microscopy and MDF panel properties. In Kim YS (ed) Proceedings of the 4th pacific regional
wood anatomy conference, Chonnam National University Press, Kwangju, Korea, pp 334–339
Singh AP, Ratz A, Dawson BSW (2007) A novel method for high-resolution imaging of coating
distribution within a rough-textured plywood surface. JCT Res 4:207–210
Singh AP, Sell J, Schmitt U, Zimmermann T, Dawson B (1998) Radial striations of the S2 layer in
mild compression wood tracheids in Pinus radiata. Holzforschung 52:563–566
Smith I, Vasic S (2003) Fracture behavior of softwood. Mech Mater 35:803–815
Somerville A (1980) Resin pockets and related defects of Pinus radiata grown in New Zealand.
NZ J For Sci 10(2):439–444
Stanzl-Tschegg S (2006) Microstructure and fracture mechanical response of wood. Int J Fracture
139:495–508
Temnerud E, Valinger E, Sundberg B (1999) Induction of resin pockets in seedlings of Pinus
sylvestris L. by mechanical bending stress during growth. Holzforschung 53:386–390
Thuvander F, Berglund LA (2000) In situ observations of fracture mechanisms for radial cracks in
wood. J Mat Sci 35:6277–6283
Thuvander F, Jernkvist LO, Gunnars J (2000) Influence of repetitive stiffness variation on crack
growth behavior in wood. J Mat Sci 35:6259–6266
Turkulin H, Arnold M, Derbyshire H, Sell J (2001) Structural and fractographic SEM analysis of
exterior coated wood. Surf Coat Int Part B. Coat Trans 34:67–75
Turkulin H, Holzer L, Richter K, Sell J (2005a) Application of the ESEM technique in wood
research Part II. Comparison of operational modes. Wood Fibre Sci 37:565–573
Turkulin H, Richter K, Sell. J (2005b) Application of the ESEM technique in wood research. Part
I: Optimization of imaging parameters and working conditions. Wood Fibre Sci 37:552–564
Turkulin H, Sell J (1997) Structural and fractographic study on weathered wood. Forschungs- und
Arbeitsbericht 115/36:1–43
Vasic S, Smith I, Landis E (2002) Fracture zone characterization micro-mechanical study. Wood
Fibre Sci 34:42–56
Vasic S, Stanzl-Tschegg S (2007) Experimental and numerical investigation of wood fracture
mechanisms at different humidity levels. Holzforschung 61:367–374
Voulgaridis EV, Banks WB (1981) Degradation of wood during weathering in relation to water
repellant long-term effectiveness. J Inst Wood Sci 9:72–83
Wardrop AB (1951) Cell wall organization and the properties of the xylem I. Cell wall organization
and the variation of breaking load in tension of the xylem in conifer stems. Aust J Sci Res Ser
B Biol Sci 4:391–414
Wardrop AB, Addo-Ashong FW (1965) The anatomy and fine structure of wood in relation to
its mechanical fracture. Proceedings of the 1st Tewkesbury Symposium, Melbourne, 26–30,
August 1965, pp 169–199
Wilkins AP (1986a) The nomenclature of cell wall deformations. Wood Sci Technol 20:97–109
Wilkins AP (1986b) Factors affecting the occurrence of broken fibres in macerated wood – A
research note. Wood Fibre Sci 18:208–210
Woodward C (1980) Fractured surfaces as indicators of cell wall behavior at elevated temperatures.
Wood Sci 13:83–86
144 L. Donaldson
Yamamoto H. (1998) Generation mechanism of growth stresses in wood cell walls: Roles of lignin
deposition and cellulose microfibril during cell wall maturation. Wood Sci Technol 32:171–182
Zimmermann T, Sell J (1997) The fine structure of the cell wall on transverse-fracture surfaces of
longitudinally tension-loaded hardwoods. Forschungs- und Arbeitsbericht 115/35:1–37
Zink AG, Pelikane PJ, Shuler CE (1994) Ultrastructural analysis of softwood fracture surfaces.
Wood Sci Technol 28:329–338
Chapter 7
Probing the Wood Coating Interface at High
Resolution
Contents
7.1 Introduction
Wood is a versatile biomaterial used in a wide range of applications. Although wood
is strong and durable the products made from it when placed in service, particularly
in outdoor environments, can deteriorate within a relatively short time because of
exposure to weathering factors, such as solar radiation, rain and decay microorgan-
isms. Application of coatings to the exposed surfaces of wood products, that can
prevent solar radiation and water from reaching wood tissues, can provide protec-
tion from wood deteriorating factors. Coating adhesion, which is among the factors
that play an important role in determining the performance of an applied coating, is
related to chemical and physical interactions with wood, the latter involving coating
attachment to wood via penetration into surface tissues, where cell lumens and cell
wall delaminations have an important role.
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 145
DOI 10.1007/978-90-481-9550-3_7, C Springer Science+Business Media B.V. 2011
146 A.P. Singh and B.S.W. Dawson
Prior to coating application wood panels are often surfaced, and the method of
surfacing (planing, sanding, band-sawing) variously influences wood-coating inter-
action and coating performance (Richter et al. 1995) because of the differences in
the introduced morphological and structural changes, which can occur at both cel-
lular and tissue levels, and can involve up to several cell layers from the surface.
For example, whereas planing with sharp knives results in only minor damages to
cells, generally as fine order cracks and delaminations in cell walls, being confined
to the outermost cell layers (Singh et al. 2002; Singh and Dawson 2004), abra-
sive planing or planing with dull knives cause severe damages, often resulting in
cell compression in addition to cell wall cracking and fracturing, and consequent
increase in the surface roughness of panels (Stewart and Crist 1982; Murmanis et al.
1986; Singh et al. 2002; Hernandez and Rojas 2002). The effect of sanding, which
is practiced to reduce the irregularities in the roughness of the surface, rendering
it more uniform for coating application, is quite different from that of planing.
Sanding causes cell compression, and often the fine particulate materials arising
from the abrasive action on cell walls in the surface layers become deposited within
the exposed cell lumens. However, the surface texture varies with the grain size,
finer grains producing smoother surfaces than coarser grains (Sinn et al. 2004; de
Moura and Hernández 2006). For some applications panels are saw-textured (band-
sawn) to produce highly rough textured surfaces. Such surfaces are aesthetically
pleasing, particularly when finished with coatings that allow the textured patterns
to be visible. The high surface irregularities are caused by the digging and ripping
actions of the saws used, resulting in large surface tissue masses to become torn
and twisted as well as cracks of varying sizes to form in cell walls in the affected
regions (Singh and Dawson 2006; Donaldson et al. 2007). A thorough understanding
of how applied coatings physically/mechanically interact with planed, sanded and
saw-textured wood surfaces is needed to effect developments leading to designer
products for high performance and specific applications, and the application of
high resolution imaging techniques in the last few years have yielded valuable
information.
This article will consider recent advances made in understanding the physi-
cal interactions between wood and applied coatings, particularly through studies
involving high-resolution probing of wood-coating interface, where mechanical
interlocking between the wood and coatings involves coating penetration into wood
pores of varying dimensions. While earlier studies, based on light, confocal and
scanning electron microscopy, provided valuable information on the pathways of
coating movement within wood and depths to which coatings can penetrate into
larger wood pores (rays, cell lumens) for a range of coating and wood types (Côté
and Robinson 1968; de Meijer et al. 1998, 2001; Nussbaum et al. 1998; Rijckaert
et al. 2001; Singh and Dawson 2003; Van den Bulcke et al. 2003), the microscopy
methods employed were inadequate in revealing whether coatings could also pene-
trate wood micro-pores (cracks and delaminations in cell walls). High resolution
imaging techniques to probe wood-coating interface have been employed only
within the last few years. In particular, the use of confocal laser scanning microscope
and field-emission scanning electron microscope has revealed the aspects of coating
penetration and distribution within surface wood tissues that have proved vital in
7 Probing the Wood Coating Interface at High Resolution 147
The work of Singh and Dawson (2004) involved CLSM examination of the wood-
coating interface in sliding microtome-cut sections from radiata pine (Pinus radiata)
panels that had been planed and subsequently finished with a polyurethane coating.
Prior to examination the sections were stained with a combination of two different
histochemical stains, toluidine blue and Sudan IV, which reacted with the wood and
the coating respectively. This proved to be an ideal stain combination for achiev-
ing the desired colour differentiation between the wood and the coating for imaging
with CLSM, and also with LM for a comparison. The sections were sequentially
stained with toluidine blue and Sudan IV and examined with LM, after mounting
in glycerol on a glass slide. The same sections were subsequently examined with
CLSM. The correlative microscopy approach undertaken to examine the same sec-
tions sequentially with LM and CLSM enabled a direct comparison to be made
with regard to the resolving capabilities of these microscopes. However, while the
main aim of the work undertaken by Singh and Dawson (2004) was to exploit the
high resolution capabilities of CLSM to advance our understanding of the physical
interactions occurring between the wood and the applied coating, particularly at the
wood-coating interface, the combined use of LM and CLSM also provided a more
complete knowledge of the pattern of coating distribution within the penetrated
tissue regions.
148 A.P. Singh and B.S.W. Dawson
Fig. 7.1 LM view of a section taken through machined wood-polyurethane interface. Penetration
of the coating (red colour) into cell lumens is resolvable, but it is not apparent whether the coating
has also penetrated into cell walls (light blue colour). Bar=20 μm
had formed in the surface tissue layers of the panel during machining. This infor-
mation constituted a significant advance in understanding the physical nature of
wood-coating interaction, as it became evident that mechanical interlocking of a
coating into the wood surface can involve coating filling of numerous micro-voids
within cell walls in addition to cell lumens, which can lead to a more robust coat-
ing adhesion. This knowledge advance was made possible by the use of CLSM and
its unique capability in high resolution imaging, and in obtaining sequential optical
sections through the depth of an object, and in producing a composite image based
on the sequential sections acquired.
In another recent study (Singh and Dawson 2006) CLSM was used in combination
with LM and SEM for high resolution imaging of wood-coating interface in radiata
pine plywood panels that had been saw-textured and subsequently finished with a
film-forming acrylic stain. Interest in the use of plywood products, particularly in
outdoor applications, such as sidings, is increasing because of their high strength
and stability, and also because they perform excellently in service outdoors, par-
ticularly when finished with stain coatings after saw-texturing (Williams and Feist
1994). In addition this product provides an aesthetically pleasing surface. It has also
been recognized that superior performance of such products outdoors is because
of the greater absorption of stains in rough surfaces produced from saw-texturing
(Williams and Feist 1994). However, a knowledge of how stain coatings interact
with wood tissues within highly rough surfaces, required a close examination of the
wood-coating interface by novel microscopy techniques. The extensive microscopic
studies undertaken by Singh and Dawson (2006), which also included correlative
LM, CLSM and SEM examination of the same sections, provided detailed infor-
mation on the micromorphology of surface wood tissues in relation to coating
distribution. This new knowledge formed the basis for understanding why stain coat-
ings applied to saw-textured plywood surfaces perform excellently under conditions
of outdoor exposure (Williams and Feist 1994).
The correlative microscopy study undertaken by Singh and Dawson (2006) was
designed in a way to enable the same sections taken from the wood-coating interface
to be examined sequentially by LM, CLSM and SEM, and thus a true comparative
assessment to be made of the capabilities of the three different types of microscopy.
Also, the combined microscopy provided more complete information on wood-
coating interface, based on the unique capabilities of each microscopy type as well
as the complementary information obtained.
The microscopy work was performed on transverse sections cut from radiata pine
plywood panels that had been saw-textured to produce a highly rough surface, to
150 A.P. Singh and B.S.W. Dawson
Fig. 7.3 LM views of sections taken through the band-sawn plywood-stain coating interface. The
plywood surface is highly irregular, with masses of distorted and ripped (arrow) tissues (blue
colour). The thickness of coating film (brown colour) is variable. Bar=200 μm
which a film-forming acrylic stain was applied. The sections cut through the wood-
coating interface with a sliding microtome were stained with toluidine blue. The
sections were then mounted in glycerol on glass slides and examined to first view
the micromorphology of the surface tissue layers (Fig. 7.3). The staining method
employed permitted excellent contrast differentiation between the coating (natural
brown colour) and wood cell walls (blue to bluish green colour after toluidine blue
staining). For subsequent correlative microscopy, selected areas from the sections
that had been examined by LM were then also viewed with CLSM operating in the
fluorescence mode at excitation and emission wavelengths suitable for achieving a
sharp colour-contrast differentiation between the coating and wood cell walls (Singh
and Dawson 2006), and then with SEM after further processing (Fig. 7.4a–c). For
SEM work, the sections were removed from the glass slides by immersing and then
floating them off in water in a Petri dish. Subsequently, the sections were air-dried
by placing them between two glass slides and then clamping to prevent from curl-
ing. The sections were then mounted on stubs, gold-coated in a sputter coater and
examined with a SEM (Fig. 7.4c).
The key advantage of the combined microscopy approach employed in Singh
and Dawson’s work was that the sections were first examined with LM, which
enabled low magnification images of the highly distorted surface tissues in the
7 Probing the Wood Coating Interface at High Resolution 151
a b
Fig. 7.4 LM, CLSM and SEM views of the same section that was cut through the band-sawn
plywood-stain coating interface. A group of highly distorted tissues, which is not well defined in
the LM view and appears fuzzy (arrow in a) is clearly resolvable in CLSM and SEM views (b and
c), revealing an intricate pattern of coating distribution within the distorted surface tissues (arrow,
arrowhead in b). (a and c) bar=200 μm; (b) bar=100 μm
saw-textured panels to be captured within a relatively short time (Fig. 7.3). The same
sections could then be examined with CLSM and SEM to view specific regions of
the sections in greater detail and with vastly improved definition, because of the
higher resolution capabilities of these instruments (compare Fig. 7.4a with b and c).
LM afforded a rapid assessment of the nature and extent of tissue deformations
and dislocations corresponding to the highly rough surface caused by band-sawing
(Fig. 7.3). The digging and ripping actions of the band-saw caused masses of sur-
face tissues to become dislocated, often being raised well above their original plane.
The dislocated tissues were also bent, changing the grain direction (Fig. 7.3). LM
also clearly revealed the pattern of coating distribution within the damaged surface
tissues, providing evidence of coating penetration into the cell lumens and other
larger voids, such as the deep cracks formed within the surface from the digging
and ripping actions of the band-saw (Fig. 7.3). Not surprisingly, the coating film on
the surface was highly irregular in thickness, being thickest at the bases of cracked,
dislocated tissues and thinnest overlying the apical parts of the raised tissues, appar-
ently resulting from uneven coating distribution across the surface irregularities
(Fig. 7.3). However, despite its usefulness in a rapid assessment of the above features
related to rough-textured surfaces, LM had severe limitations in resolving surface
152 A.P. Singh and B.S.W. Dawson
tissue masses that were twisted with respect to the orientation of the underlying
unaffected tissues, and the damages present at the cell wall level, such as cracks and
delaminations in cell walls. Correlative microscopy showed that such features were
clearly resolved by both CLSM and SEM (compare Fig. 7.4a with b and c).
CLSM, which offers greater resolution than achievable by LM with the added
advantage of optical sectioning, makes it possible to obtain information based on
a composite of images from sequential sections taken through the thickness of a
sample. This combination of unique capabilities of CLSM made it possible in the
work of Singh and Dawson (2006) to clearly resolve even the most severely twisted
surface tissues, which appeared fuzzy under LM, allowing a close examination of
the rather intricate pattern of coating distribution within the highly distorted surface
tissues. The CLSM observations revealed that the coating had penetrated cell wall
cracks of even very small dimensions, in addition to filling the larger voids, such
as the lumens of rays and tracheids and the voids formed from the cracking and
dislocation of surface tissues (Fig. 7.4b).
Although SEM is a surface imaging instrument, its main advantage over LM
and CLSM lies in achieving greater depth of focus and resolution, which made
it possible in the correlative microscopy work undertaken by Singh and Dawson
(2006) to also clearly resolve the twisted surface tissues and also the penetration
pathways of the coating (Fig. 7.4c), which supported the observations made using
CLSM.
A recent study (Singh et al. 2007) describes a technical breakthrough in the imaging
of wood-coating interface with FE-SEM, a high resolution SEM, to visualise coating
penetration into cell wall cracks and delaminations of extremely small dimensions,
and the information presented strengthens the concept that an effective mechani-
cal interlocking between the wood surface and an applied coating greatly enhances
coating adhesion and thus coating performance.
In their study Singh et al. 2007 used a penetrating stain and a highly rough tex-
tured wood surface as a substrate to examine the physical nature of their interaction
to better understand the influence of wood texture on coating penetration. The coat-
ing was an oil-borne stain and the wood substrate was radiata pine plywood that had
been saw-textured to produce a highly rough surface, as in the work described in
the section dealing with combined microscopy. A novel technique was developed
to examine the wood-coating interface in microtome-cut sections with FE-SEM
after treating the sections with a high atomic number reagent. The sections (90 μm
thick) cut transversely through the wood-coating interface with a sliding microtome
were treated with 1% aqueous osmium tetroxide (OsO4 ) for 3 h at room tempera-
ture. After washing in water, the sections were clamped between two glass slides
7 Probing the Wood Coating Interface at High Resolution 153
to air-dry before mounting on stubs applied with sticky carbon discs. The sections
were then coated with chromium in a sputter coater and examined with FE-SEM
operating in the BEI mode at the operational conditions described in detail earlier
(Singh et al. 2007). For comparisons, the same sections were also examined in the
secondary electron imaging (SEI) mode of the FE-SEM.
In the work of Singh et al. (2007), a combination of imaging in the BEI mode
based on the novel technique developed for pre-treating sections with OsO4 and the
high resolution capability of FE-SEM made it possible to clearly resolve and dif-
ferentiate the coating material from wood cell walls, enabling detailed observations
to be made on the intricate pattern of coating distribution within the saw-textured
surface tissues exhibiting a range of distortions and damages from tissue level down
to the nanometer-scale cell wall cracks present. Most SEM studies undertaken of
wood-coating interface prior to this work were based on imaging in the SEI mode
of conventional SEM (and not high resolution FE-SEM), and also the focus had
been to obtain information on the depth of coating penetration into larger voids in
the wood surfaces, such as the lumens of rays, tracheids and vessels.
As mentioned earlier, band-sawn wood surfaces have a rough texture, with highly
irregular contours resulting from the digging and ripping actions of the band-saw,
the feature being particularly striking for plywood. It had been previously shown
that saw-textured plywood performs well outdoors when finished with penetrating
coatings, such as stains (Williams and Feist 1994). The greater coating absorption
because of the high roughness of the surface was considered to be the main contrib-
utory factor, but an understanding of wood-coating interaction at the cellular level
remained a knowledge gap. In the work of Singh et al. (2007) the development and
application of the high resolution BEI technique in conjunction with FE-SEM to
examine the interface of a saw-textured plywood surface provided vital information
for understanding the basis for the excellent performance noted earlier of the stained
saw-textured plywood sidings under outdoor conditions.
BEI imaging in combination with SEM has been used to examine a range of com-
posites consisting of two or more components of differing atomic numbers (Carter
1979; Harris et al. 1999; Richards et al. 1999; Herzog et al. 2004). In the BEI
mode high atomic number components appear brighter. Thus brightness intensity
serves as a useful basis for obtaining information on the location and distribution
of high atomic number components in a sample. Methods have also been developed
to boost the brightness of low atomic number substances by reacting them with
high atomic number additives as tracers or stains, which could enhance the yield
of backscattered electron signals (DeNee and Carpenter 1979; Schraufnagel and
Ganesan 1998). Initial attempts by Singh et al. (2007) to image the wood-coating
interface by FE-SEM operating in the SEI mode were unsuccessful as the contrast
differentiation between the coating and wood cell walls was poor (Figs. 7.5a and
7.6a), and a method had to be developed to obtain suitable contrast differentia-
tion. FE-SEM imaging of wood-coating interface in the BEI mode after treatment
of sections with OsO4 , a high atomic number substance which selectively reacted
with the coating material (and not wood cell walls) in the wood-coating sections,
proved to be the method of choice in the work undertaken by Singh et al. (2007).
154 A.P. Singh and B.S.W. Dawson
Fig. 7.5 FE-SEM-SEI (a) and FE-SEM-BEI (b) views of the same section that was cut through
the band-sawn plywood-stain coating interface. In the SEI view the coating is poorly differentiated
from wood cells, whereas in the BEI view the coating material is well highlighted because of its
superior brightness, revealing a complex pattern of coating distribution within the distorted surface
tissues. Bar=100 μm
Fig. 7.6 High magnification FE-SEM-SEI (a) and FE-SEM-BEI (b) of the same region of band-
sawn plywood-stain coating interface. Whereas in the SEI view the contrast of coating is similar to
that of wood cell walls, in the BEI view the coating appears much brighter and is thus readily dis-
tinguishable from the cell walls, revealing the pathways of coating penetration, including coating
penetration into tiny cell wall cracks (arrows). Bar=10 μm
7 Probing the Wood Coating Interface at High Resolution 155
7.4 Summary
The information presented in this article highlights the advances made in under-
standing the physical aspects of wood-coating interaction. In particular, the focus
here has been to review the work presented in three papers published within the last
few years (Singh and Dawson 2004, 2006; Singh et al. 2007), where developments
in microscopy techniques have lead to high resolution imaging of wood-coating
interface, employing CLSM alone as well as in combination with LM and SEM
as part of correlative microscopy performed on the same sections, and FE-SEM in
combination with backscattered imaging. The information published in these papers
provided an opportunity to examine the pattern of coating penetration and distribu-
tion for two contrasting coating types and wood surface textures. Particularly useful
has been the system that involved stain coatings and a rough-textured plywood, a
combination proving an ideal model system to study the distribution pattern of a
penetrating coating in a wood surface rendered highly porous from band-sawing,
that caused massive tissue distortions and created voids of wide ranging dimen-
sions, including nanometer-scale cell wall cracks, in the panel surface. The high
resolution microscopy techniques employed in these studies greatly enhanced our
156 A.P. Singh and B.S.W. Dawson
References
Carter HW (1979) Backscattered electron imaging of biological specimens. In: Proceedings of the
14th annual conference of the microbeam analysis society, San Antonio, pp 1–2
Côté WA, Robinson RG (1968) A comparative study of wood: wood coating interaction using
incident fluorescence and transmitted fluorescence microscopy. J Paint Technol 40:427–432
de Meijer M, Thurich K, Militz H (1998) Comparative study on penetration characteristics of
modern wood coatings. Wood Sci Technol 32:347–365
de Meijer M, Thurich K, Militz H (2001) Quantitative measurements of capillary coating
penetration in relation to wood and coating properties. Holz Roh Werkst 59:35–45
de Moura LF, Hernandez RE (2006) Effects of abrasive mineral, grit size and feed speed on the
quality of sanded surfaces of sugar maple wood. Wood Sci Technol 40:517–530
DeNee PB, Carpenter RL. (1979) Application of heavy metal staining (OsO4 )/backscattered elec-
tron imaging technique in the study of organic aerosols. In: Proceedings of the 14th annual
conference of the microbeam analysis society, San Antonio, pp 8–10
Donaldson L, Bardage S, Daniel G (2007) Three-dimensional imaging of a sawn surface: a com-
parison of confocal microscopy, scanning electron microscopy, and light microscopy combined
with serial sectioning. Wood Sci Technol 41:551–564
Harris LG, Gwynn I, Richards RG (1999). Contrast optimization for backscattered electron
imaging of resin embedded cells. Scan Micros 13:71–81
Hernandez RE, Rojas G (2002) Effects of knife jointing and wear on the planed surface quality of
sugar maple wood. Wood Fiber Sci 34:293–305
Herzog B, Goodell B, Lopez-Anido R (2004) Electron microprobe imaging for the characterization
of polymer matrix composites. Compos Part A 35:1075–1080
Murmanis L, River BH, Stewart H (1986) Surface and subsurface characteristics related to
abrasive-planing conditions. Wood Fiber Sci 18(1):107–117
Nussbaum RM, Sutcliffe EJ, Hellgren AC (1998) Microautoradiographic studies of the penetration
of alkyd, alkyd emulsion and linseed oil coatings into wood. J Coat Technol 70:49–57
Richards RG, Owen GRH, Gwynn I (1999) Low voltage backscattered electron imaging (<5 kV)
using field emission scanning electron microscopy. Scan Micros 13:55–60
Richter K, Feist WC, Knaebe MTC (1995) The effect of surface roughness on the performance of
finishes. Part 1. Roughness characterization and stain performance. For Prod J 45(7/8):91–97
Rijckaert V, Stevens M, Van Acker J, De Meijer M, Militz H (2001) Quantitative assessment of the
penetration of water-borne and solvent-borne wood coatings in Scots pine sapwood. Holz Roh
Werkst 59:278–287
7 Probing the Wood Coating Interface at High Resolution 157
Schraufnagel DE, Ganesan DP (1998) Tracers in vascular casting resins enhance backscattering
brightness. Scan Micros 12:631–639
Singh AP, Anderson CR, Warnes JM, Matsumura J (2002) The effect of planing on the microscopic
structure of Pinus radiata wood cells in relation to penetration of PVA glue. Holz Roh Werkst
60:333–341
Singh AP, Dawson BSW (2003) The mechanism of failure of clear coated wooden boards as
revealed by microscopy. IAWA J 24(1):1–11
Singh AP, Dawson BSW (2004) Confocal microscope – a valuable tool for examining wood-
coating interface. J Coat Technol Res 1(3):235–237
Singh AP, Dawson BSW (2006) Microscopic assessment of the effect of saw-textured Pinus radi-
ata plywood surface on the distribution of a film-forming acrylic stain. J Coat Technol Res
3(3):193–201
Singh AP, Ratz A, Dawson BSW (2007) A novel method for high-resolution imaging of coating
distribution within a rough-textured plywood surface. J Coat Technol Res 4(2):207–210
Sinn G, Gindl M, Reiterer A, Stanzl-Tschegg S (2004) Changes in the surface properties of wood
due to sanding. Holzforschung 58:246–251
Stewart HA, Crist JB (1982) SEM examination of subsurface damage of wood after abrasive and
knife planing. Wood Sci 14:106–109
Van den Bulcke J, Rijckaert V, Van Acker J, Stevens M (2003) Quantitative measurement of
the penetration of water-borne coatings in wood with confocal laser microscopy and image
analysis. Holz Roh Werkst 61:304–310
Williams RS, Feist WC (1994) Effect of preweathering, surface roughness, and wood species on
the performance of paint and stains. J Coat Technol 66:109–121
Chapter 8
Delamination in Timber Induced
by Microwave Energy
Georgiana Daian
Contents
8.1 Introduction
High intensity microwave treatment creates changes in wood structure. The appli-
cation of intensive microwaves generates the appearance of micro-voids and checks
throughout the cross section of wood in the radial/longitudinal grain direction, due
to the effect of microwave induced steam pressure on wood micro and macro-
elements (Torgovnikov and Vinden 2003). Under microwaves, the permeability
of some wood species in the radial direction can be increased by a factor of
170–1,200 times (Vinden and Torgovnikov 2000). For microwave-wood applica-
tions, such as wood impregnation, the occurrence of uniformly distributed internal
checks is intended. For other microwave treatments designed to create growth stress
relief in fast grown plantation hardwoods or to assist and accelerate wood drying, as
well as wood bending, the treatment defects (checks, cracks, collapse, splits) have
to be controlled and minimized.
G. Daian (B)
The University of Melbourne, Department of Forest and Ecosystem Science, Melbourne,
VIC 3010, Australia
e-mail: gdaian@gmail.com
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 159
DOI 10.1007/978-90-481-9550-3_8, C Springer Science+Business Media B.V. 2011
160 G. Daian
– Heating phase – the energy is transferred directly from the microwave field to the
wood and very little moisture loss is incurred.
– Streaming phase – the temperature increases beyond the boiling point of water and
the resultant internal vapor pressure drives out the liquid from the wood under the
action of pumping phenomenon. The moisture content promptly decreases.
– Enthalpic phase – vapor transport becomes the dominant migration mechanism
because of the elevated internal temperatures which sustain vaporization within
the wood.
8 Delamination in Timber Induced by Microwave Energy 161
– Thermal runaway phase – commences when the wood becomes dry in cer-
tain locations and the temperature increases rapidly. The hot spots can severely
damage the wood.
The manner in which the internal heat generated by microwaves power of high
intensity affects the solid wood structure has been studied by Torgovnikov and
Vinden (2003). The researchers found that if wood is supplied with microwave
power of high intensity, the moisture in the cells starts boiling and high steam pres-
sure is generated in the wood cells. The steam pressure leads to the rupture of the
weakest elements of wood structure, such as the pit membranes and ray cells. The
destroyed rays form voids whose distribution in the microwave modified wood is
influenced by the wood structure: the larger the number of rays, the greater the
number of cavities.
Torgovnikov and Vinden (2003) also noticed that by increasing the intensity of
the supplied microwave energy, the steam pressure increases. As a result, cellular
components such as traheids, libriform fibers, or vessel walls, rupture gradually in
proximity of the ray tissue by generating narrow checks. The checks tend to extend
in the radial-longitudinal planes, forming cavities. A further breakdown occurs at the
resin canals level in softwoods, facilitating the resin replacement. In hardwoods, the
vessels membranes (tyloses) break down, making the vessels permeable to liquids
and vapors.
Three degrees of structural modification induced by microwave power of high
intensity were defined (Torgovnikov and Vinden 2003):
– low degree: including rupturing wooden cells pit membranes, resin melting and
replacement in canals, partly rupturing ray cells;
– moderate degree: including rupturing wooden cells pit membranes, resin boil-
ing and replacement, destroying tyloses in the vessels (hardwood species) and
rupturing ray cells;
162 G. Daian
– high degree: including rupturing wooden cells pit membranes, destroying tyloses
in the vessels (hardwood species), rupturing ray cells, rupturing main cells (tra-
cheids, libriform) walls and vessels, and the formation of cavities being primarily
in the radial-longitudinal direction.
Cooperative Research Centre (CRC) for Wood Innovations (Australia) has inves-
tigated microwave processes and technologies for achieving various forms of
microwave modification of wood, ultimately proven to have industrial applicabil-
ity (Australian Forests and Timber News 2007; 2008). Three license agreements
were completed with industry: with TimTech Chemicals – an advanced preserva-
tion process; with Bostik Australia – adhesive technology; and Carter Holt Harvey –
softwood rail sleepers’ preservation.
Other similar studies have been undertaken worldwide as seen in the next
sections.
The response of wood at the microwave field oscillations depends on the wood abil-
ity to absorb and store electrical potential energy. This physical reaction is measured
by the wood permittivity, otherwise known as the dielectric parameters of wood.
Dielectric properties of wood play an important role in the development of
microwave processes which aim certain structural modification of the material; they
enable electromagnetic simulations for optimal system design. Knowing the dielec-
tric properties of wood and their variations during various stages of microwave
processing, the estimation of the absorbed power and stored energy is also possible
(Daian 2005).
The literature (Tsutsumi and Watanabe 1965; Norimoto 1976; Torgovnikov
1993; Kabir et al. 1997 and 2001; Olmi et al. 2000; Afzal et al. 2003; Daian 2005;
Daian et al. 2005 and 2006; Koubaa et al. 2008) comprises data on dielectric proper-
ties of some wood species, varying with moisture contents, densities, temperatures,
electromagnetic frequencies and electric field (i.e. acting in the three structural
directions of wood).
Torgovnikov (1994) indicates that under microwave power, the wood tempera-
tures may reach 100–170◦ C, giving rise to internal pressures of up to 0.7 MPa. The
only available method for measuring the dielectric permittivity of thick pieces of
wood at 2.45 GHz and suitable for temperatures as high as 150–170◦ C and high
pressures, was developed by Daian et al. (2005).
Daian et al. method (2005) is based on the short-circuit line measurement tech-
nique, measuring the impedance change caused by the presence of the wood within
a rectangular waveguide. The method was complemented by a numerical procedure
which extracts the permittivity from the measured impedance via the solution of a
transcendental equation in the complex plane. Unlike von Hippel algorithm (von
Hippel 1954), the procedure presented by Daian et al. (2005) considers the device
8 Delamination in Timber Induced by Microwave Energy 163
losses created by the presence of the pressure windows and pumping air holes in the
waveguide.
The analytical expression of the dielectric parameters of wood has been sug-
gested for many years by the algebraic mixture laws (Torgonikov 1993). Since the
mixture laws do not account for the dependence of permittivity on material’s micro
and macrostructure, they fail to accurately predict the complex permittivity of wood.
Daian et al. (2006) developed a 3D computer-based model to overcome the mixture
laws’ downside.
The model considers wood as a four-phase material (cell wall substance, bound
water, free water and air) and uses the typical arrangement of the wood cells and the
presence of bound and free water, in conjunction with their fractional volumes at any
particular density and moisture content. At macro-level, the wood was pictured as
an arrangement of intercrossing parallel layers indicating the cell cavities which are
partially or completely filled with moisture layers. Since the wood geometry is more
complex than just representing a material consisting of conductive layers separated
by insulating layers, transversal structures described as a group of parallel and series
connected capacitors representing the rays were inserted in the entire wood picture.
At micro-level, the structure of moist wood substance was considered as a random
mixture of wood substance and bound water with their correct fractional volumes at
various moisture contents.
To calculate the dielectric properties of the modeled wood piece, the 3D structure
was imaginarily divided into a mesh of small cells, each having specific dielectric
properties (i.e. complex permittivity of wood substance, air, free water or bound
water). Then, the main idea was to create a potential difference on two oppo-
site boundaries of the structure and solve for the potentials at all inner vertices of
the mesh.
Complex permittivities of structural and compositional components of wood
(wood substance/moist wood substance, free water, bound water, air and rays) for
a range of microwave frequencies (0.915 GHz and 2.45 GHz) and temperatures
(20–150◦ C) represent known variables for the computational algorithm and there-
fore, estimations were determined based on available data from literature, extrapo-
lations, Cole and Cole equations and existent evaluation methods (Daian 2005).
The 3D computer-based model reproduced the trends of the dielectric constant
reported by the in-house measurements and the literature (at various directions of
the electric field and moisture content), and the values of the complex permittivities
obtained via measurements (10% and 5% average deviations for dielectric constant
and loss factor, respectively) (Daian et al. 2006; Daian 2005).
8.4.1 Impregnation
Delamination effect of microwave treatment on wood structure has been seen by
researchers as a feasible solution for impermeable wood species in wood preserva-
tive treatment industry (Vinden et al. 2004; Sugiyanto et al. 2008) or other similar
applications: e.g. resins (Torgovnikov and Vinden 2006b; Przewloka et al. 2007) or
polymer impregnation (GeumHyun et al. 2003).
Vinden at al (2004) and Torgovnikov and Vinden (2006b) investigated the uni-
form formation of high degree modification in softwood and hardwood timber
(Fig. 8.2), created by microwave for applications such as preservative treatment and
resins impregnation (Fig. 8.3). The microwave modification effect, visibly achieved
in the radial-longitudinal direction with respect to the grain of Douglas-fir and Pinus
radiata, was microscopically described as the result of: the rupturing of the ray cells
tissue, resin softening and mobilization and formation of a large numbers of cavities.
In hardwoods (e.g. Quercus robur, Eucalyptus muelleriana, Eucalyptus regnans,
Eucalyptus obliqua), the ray cell rupturing is accompanied by controlled formation
of micro checks, at the interface of ray tissue and longitudinal fiber, and micro-voids
in the radial-longitudinal direction.
Sugiyanto et al. (2008) studied the anatomical changes in timber which was
treated by using a microwave system designed for conditioning wood surface for
preservative treatment. The microscope analysis indicated the occurrence of induced
micro- and macro-voids in radial-longitudinal planes of top area and at a limited and
controlled depth in Pinus radiata timber. The voids appeared as narrow checks with
widths predominantly greater than 0.05 mm in some areas and lower than 0.05 mm
in other areas.
GeumHyun et al. (2003) observed that a specific microwave treatment of wood
results in many checks in the tracheid’s cell walls when preceded by ultrasounds,
thus, improving the permeability of Pinus radiata and Pinus koraiensis by 2 and up
to 7–8 times, respectively, for enhanced polymer impregnation treatment.
8.4.2 Drying
Generally, drying defects, such as surface checking, collapse of the cross section and
internal collapse checks, appear even under the best controlled conditions, degrading
a significant proportion of timber (Harris et al. 2008).
Microwave drying of wood has been widely studied (e.g. Antti 1992 and 1995;
Zielonka and Dolowy 1998; Zielonka and Gierlik 1999; Antti and Perre 1999;
Perre and Turner 1999; Awoyemi 2004a; Liu and Zhang 2005; Liu et al. 2006;
Torgovnikov and Vinden 2006a; Piao et al. 2006; Jia and Afzal 2007) and con-
sidered as an alternative drying technology which appears to minimize the effects
of checking and collapse, when used as a preconditioning treatment (Torgovnikov
et al. 2003; Harris et al. 2008).
Despite the extensive work on microwave drying of wood, there is little infor-
mation regarding the incidence of collapse in the microwave drying applications.
Governed by the principle that by increasing the permeability of wood prior to
drying the moisture is able to move more freely and the evaporative front created
during drying is maintained for a longer period at the wood surface (reducing ten-
sion stress and in turn surface checking), Harris et al. (2008) investigated the effect
of low-intensity microwave pretreatment on drying degradation (i.e. checking or
splitting) of Eucalyptus Obliqua. The study revealed a significant reduction in the
formation of internal/honeycombing and surface checking in the microwave samples
pretreated with various microwave schedules, when compared with the controls.
8.4.3 Bending
The ability of wood to bend is subject to a large number of factors including: the
softening method, the type, direction and duration of loading, moisture content,
structural characteristics of wood, such as variation in size of the cells walls and
166 G. Daian
pores, and other natural characteristics which relate to the quality of wood (i.e.
knots, slop of grain, gum and resin pockets, etc). These factors can influence the
bending quality by producing irreversible damage to the wood.
Microwave softening of wood has been extensively studied and considered a
competitive softening method in wood bending (Norimoto 1979; Norimoto and
Gril 1989; Studhalter 2005; Ozarska and Juniper 2006; Studhalter et al. 2008;
Juniper 2008). Considering the microwave effects on the wood structure (presented
in a previous section), naturally and unlike other microwave wood applications
(e.g. impregnation), low to moderate microwave wood modification is intended in
bending to avoid internal checking of bent-wood components.
When a piece of wood is bent, the wood fibers on the outer side of the bend are
put in compression and those on the opposite side, in tension. The compression is
accompanied by the shortening of wood fibers and the tension by the lengthening
or stretching (Wilson 1029; Taylor 2001; Schleining 2002). These phenomena may
generate defects such as crosswise folds or wrinkles and lateral buckling on com-
pression side, and tensile breaking strains (fiber rupture or split) on tension side
(Fig. 8.4) (Peck 1957; Norimoto and Gril 1989).
The break of wood fiber during bending varies widely not only among species
but also within the same wood species regardless of the softening treatment used
(Daian and Ozarska 2008).
So (1997) found that improved outcomes for microwave softening are achieved
at high moisture contents (50%). Steam softening generally focuses on moisture
contents around fibre saturation point (Stevens and Turner 1970). The quality of
microwave bending at higher moisture content is enhanced when comparing with
steam softening. This can be explained by the fact that the water is the compo-
nent which dominates the microwave heating properties of wood (Norimoto 1979;
Torgonokov 1993).
Steam softening causes changes in wood mechanical properties by improving its
bending quality (Stevens and Turner 1970). In 1989; Norimoto and Gril assumed
that microwave softening can have a much better effect on the mechanical proper-
ties of wood, saying that microwave allow “new wood species and specimens with
lower quality to endure larger deformations with a reduced loss” when compared
with traditional steaming procedure. Later, a study (Awoyemi 2004b) confirmed
Norimoto and Gril’s assumption by revealing that the microwave modification of
Pinus radiata resulted in: 4.8–59.3% and 2.8–27.4% reductions in the tangential
a b c
Fig. 8.4 Examples of crosswise folds and tensile breaking strains in bending microwave-softened
wood (Daian 2009, unpublished personal photos)
8 Delamination in Timber Induced by Microwave Energy 167
and radial MOE, respectively; 5.7–65.5% and 0.9–43.5% reductions in the tan-
gential and radial MOR, respectively. An extensive study on mechanical behavior
of Eucalyptus regnans during microwave bending has been performed by Juniper
(2008), revealing as well the positive effects of microwave on the quality of bending.
Daian et al. (2009) investigated the effect of applicator design, power intensity
and exposure time on the microwave energy distribution and the shape of the treat-
ment zone within green wood boards of about 105 × 30 mm. The study identified
the optimal design solution for a microwave pre-drying process, able to uniformly
modify the timber structure in the cross section with minimum and small checks,
and without structural defects.
Several pre-treatment applicator systems with different dimensions and config-
urations were analyzed by Daian et al. (2009). For a particular configuration (an
oversized tunnel applicator and four tapered microwave waveguides placed around
the tunnel walls), the electromagnetic modeling and simulation indicated that the
size of the delamination zone significantly depends on tunnel width: the bigger the
tunnel width, the more uniform distribution along the dielectric sample. In addi-
tion, the intensity of the absorbed energy was found to be subject to the tunnel
applicator size: maximum absorption of the energy occurred as the tunnel applica-
tor narrowed. The two findings target the microwave pre-drying process objectives
but contradict in terms of the applicator size. As low energy absorption translates to
inefficient utilisation of the energy and the necessity to apply higher exposure time
to the microwave radiation (which could lead to extensive checking), the small size
of the applicator was considered optimum for the experimental trials.
For the modeled configuration, Daian et al. (2009) showed experimentally that a
uniform delamination area was achieved along the wood sample cross section for a
microwave energy supply of 20+20 kW (i.e. two generators were in use) and 40 s
exposure time. The delamination area was characterised by a few very small splits
in comparison with very large splits or ruptures at different locations inside wood
when lower microwave energy and higher exposure time were applied (Fig. 8.5). In
addition, it was found that much better results could be obtained when the timber
samples were quarter-sawn.
Fig. 8.5 Microwave delamination zone inside Blue Gum (Eucalyptus globulus) timber samples.
(a) 5+5 kW applied energy and high exposure time 110 s.; (b) 10+10 kW applied energy and lower
exposure time 75 s; (c) 20+20 kW applied energy and half exposure 40 s time in respect with
case b. (Daian et al. 2009)
8 Delamination in Timber Induced by Microwave Energy 169
8.6 Conclusions
The use of microwave processing techniques in wood production is a revolution-
ary concept which has been extensively studied for the last decade. Both high and
low intensity microwave treatment creates delaminations in wood structure. The
microwave-induced delaminations are described by ruptures at micro and macro
level of wood structure. The level to which wood delamination occurs is critical
in any particular microwave-wood application. The size of the modification area,
its distribution and the frequency of cells rupturing can be controlled and opti-
mized. Delamination is generally optimised through electromagnetic modeling and
simulations particularly targeting the microwave equipment design.
The dielectric properties of wood is a must-to-know parameter in the microwave
simulations processes. A 3D computer-based model was created to predict the
dielectric properties of wood. The model is able to accurately reproduce the experi-
mental parameters over a wide range of fractional volumes of wood constituents (i.e.
wood substance, air, free water and bound water). Thus, it represents a powerful tool
for efficiently minimising the experimental work.
The literature on solid wood delaminations in various microwave applications
reports that:
8.7 Summary
The use of microwaves has been long investigated in materials processing and a
broad range of plasma processes, chemical synthesis and processing, as well as in
waste remediation. During the last decade, the microwave-wood alliance captured
scientists’ attention considering that it can contribute to revolutionary developments
for the timber industry needs.
This chapter reviews the principles and physics of microwave delamination of
wood, fundamental to the study of the microwave-wood applications. The factors
which affect the degree of delamination in the microwave irradiated wood have been
extensively investigated by scientists and the findings are underlined in this chapter.
Studies concerning the control of microwave delamination process are reviewed as
well.
170 G. Daian
References
Afzal MT, Colpitts B, Galik K (2003) Dielectric properties of softwood species measured with
an open-ended coaxial probe. Proceedings 8th International IUFRO Wood Drying Conference,
Brasov, Romania, August 24–29, pp 110–115
Antti AL, Perré P (1999) A microwave applicator for on line wood drying: temperature and
moisture distribution in wood. Wood Sci Technol 33(2):123–138
Antti AL (1992) Microwave drying of hardwood: simultaneous measurements of pressure,
temperature, and weight reduction. For Prod J 42(6):49–54
Antti AL (1995) Microwave drying of pine and spruce. Holz als Roh- und Werkst 53(5):333–338
Australian Forests & Timber News (2007) Technological edge in timber treatment. Aust For
Timber News, Hartley Higgins, Ryan Publications Pty Ltd., 16(8):21
Awoyemi L (2004a) Effects of microwave modification on the kiln drying time of Eucalyptus
obliqua wood. J Indian Acad Wood Sci 1:83–88
Awoyemi L (2004b) Effects of microwave modification on the mechanical properties of Pinus
radiata heartwood. Indian Forester 130:749–756
Cooperative Research Centre (CRC) Wood Innovations (2004) Microwaves and wood processing-
Fact Sheet
Daian G (2005) Establish the dielectric properties of a number of wood species for a range of
frequencies, temperatures and pressures. PhD Thesis, Swinburne University of Technology,
Australia
Daian G, Taube A, Birnboim A, Shramkov Y, Daian M (2005) Measuring the dielectric properties
of wood at microwave frequencies. Wood Sci Technol 39(3):215–223
Daian G, Taube A, Birnboim A, Daian M, Shramkov Y (2006) Modeling the dielectric properties
of wood. Wood Sci Technol 40(3):237–246
Daian G, Ozarska B (2008) Determination of minimum radius of bending curvature of heartwood
and sapwood from young regrowth and quickly grown hardwood species using microwave
wood bending technology. Report for the FWPA Project No. PN07.2037: impact of sapwood
on the properties and market utilisation of plantation and young hardwoods
Daian M, Taube A, Torgovnikov G, Daian G, Shramkov Y (2009) Computer modelling
of the energy distribution within wood throughout microwave processing. Comput Mater
Continua 8:3
GeumHyun D, InAeh K, ByeungSu P. (2003) Effects on the polymer impregnation of microwave
treated wood. KFRI J. Forest Science (Seoul) 66:11–17
Harris GA, Torgovnikov G, Vinden P, Brodie GI, Shaginov A (2008) Microwave pretreatment of
backsawn messmate boards to improve drying quality: Part 1. Drying Technol 26:579–584
Jia D, Afzal M T (2007) Modeling of moisture diffusion in microwave drying of hardwood. Drying
Technol 25:449–454
Juniper LF (2008). Investigation into the mechanical behaviour of Eucalyptus regnans during
microwave bending. PhD Thesis. The University of Melbourne
Kabir MF, Daud WM, Khalid KB, Sidek HAA (1997) Dielectric properties of rubber wood at
microwave frequencies measured with an open-ended coaxial line. Wood Fiber Sci 29(4):
319–324
Kabir MF, Daud WM, Khalid KB, Sidek HAA (2001) Temperature dependence of the dielectric
properties of Rubber wood. Wood Fiber Sci 33(2):233–238
8 Delamination in Timber Induced by Microwave Energy 171
Koubaa A, Perré P, Hutcheon RM, Lessard J (2008) Complex dielectric properties of the sapwood
of aspen, white birch, yellow birch, and sugar maple. Drying Technol 26:568–578
Liu Z, Zhang B (2005) Application of microwave heating in wood drying. World For Res 18:
54–58
Liu Z, Li Y, Zhang B, Liu Z (2006) Microwave drying properties of Pinus massoniana lumber. J
Zhejiang For Coll 23:482–485
Metaxas AC, Meredith RJ (1993) Industrial microwave heating. Peter Peregrinus Ltd. On behalf
of the IEE, London
Norimoto M (1976) Dielectric properties of wood. Wood Res 59(60):106–152
Norimoto M (1979) Wood bending by microwaves. Wood Res Rev 14:13–26
Norimoto M, Gril J (1989) Wood bending using microwave heating. J Microwave Power
Electromagn Energy 24(4):203–212
Olmi R, Bini M, Ignesti A, Riminesi C (2000) Dielectric properties of wood from 2 to 3 GHz. J.
Microwave Power Electromagn Energy 35(3):135–143
Ozarska B, Juniper L (2006). Design brief: design and construction of prototype wood bending
facilities for “Wood Shapes Pty Ltd”. CRC Wood Innovations Report.
Peck EC (1957) Bending solid wood to form. U.S. Goverment Printing Office, Washington, DC
Perré P, Turner IW (1999) The use of numerical simulation as a cognitive tool for studying the
microwave drying of softwood in an over-sized waveguide. Wood Sci Technol 33:445–464
Piao J, Fujimoto N, Oohashi K, Tanikawa M, Kitada M, Sonobe H, Ueda Y (2006) Drying
properties of Sugi round timber with microwave heating. J Fac Agric Kyushu Univ 51:345–349
Przewloka SR, Hann JA, Vinden P (2007) Assessment of commercial low viscosity resins as
binders in the wood composite material Vintorg. Holz Roh Werkst 65:209–214
Schleining L (2002) The complete manual of wood bending: milled, laminated, and steam-bent
work. Linden Publishing , Fresno, CA
So W (1997) Study on the application of microwave-heating system for making bent-wood
furniture (II). Mokchae Konghak 25(2):52–60
Studhalter, B. (2005) Temperature and moisture content behaviour in microwave heated wood
prior to bending – Mountain Ash (Eucalyptus regnans). Diploma Thesis. The University of
Melbourne and The University of Applied Sciences, Biel, Switzerland.
Studhalter B, Ozarska B, Siemon GR (2008) Temperature and moisture content behaviour in
microwave heated wood prior to bending – Mountain Ash (Eucalyptus regnans). Published
“Online First” 23 December 2008. Holz als Roh- Werkstoff.
Stevens WC, Turner N (1970) Wood bending handbook. Fox Chapel Publishing, East-
Petersburg, PA
Sugiyanto K, Torgovnikov G, Vinden P (2008) Microwave wood modification of timber surfaces
for preservative treatment. Proceedings of global congress on microwave energy applications,
August 4–8, Otsu, Japan, pp 229–232
Taylor Z (2001) Wood bender’s handbook. Sterling Publishing, New York, NY
Tinga WR (1993) Microwaves material interactions and process design modeling. Ceram Trans
36:29–43
Torgovnikov GI (1993) Dielectric properties of wood and wood-based materials. Springer,
Heidelberg, ISBN 3-540-55394-0
Torgovnikov G I (1994) Some aspects of microwave power application in wood technology. Wood
Sci Dig 52:9–12
Torgovnikov GI, Vinden P (2003) Effect of intensive microwave radiation on wood structure.
Proceedings of the 9th international conference on microwave and high frequency heating,
Loughborough University, UK, pp 501–504
Torgovnikov GI, Vinden P, Mekhtiev M (2003) Microwave modification of hardwoods. Forest
wood product research and development project no. PN00.1301
Torgovnikov GI, Vinden P (2006a) Method of microwave treatment of wood. United States Patent
No 7089685
Torgovnikov GI, Vinden P (2006b) Microwave method for increasing the permeability of wood and
its applications. Advances in microwave and radio frequency processing, report from the 8th
172 G. Daian
international conference on microwave and high frequency heating held in Bayreuth, Germany,
September 3–7, 2001
Torgovnikov GI (2008) Microwave laboratory and expertise could be lost to Australia. Aust For
Timber News 17(8):8
Tsutsumi J, Watanabe H (1965) Studies on dielectric behaviour of wood (II): effect on moisture
content on dielectric constant ε and dielectric loss factor ε . Mokuzai Gakkaishi, pp 115–118
Vinden P, Torgovnikov GI (2000) The physical manipulation of wood properties using microwave.
International Conference of IUFRO, Tasmania, Australia, March 19–24, 2000
Vinden P, Romero FJ, Torgovnikov GI (2004) Method for increasing the permeability of wood. US
Patent No. 6742278
von Hippel AR (1954) Dielectric materials and applications. The Technology Press. New York, NY
Wilson TRC (1929) Wood bending (with appendix on apparatus for bending boat ribs).
Madison Wisconsin, United Stated Department of Agriculture–Forest Sevice – Forest Product
Laboratory
Zielonka P, Dolowy K (1998) Microwave drying of spruce: moisture content, temperature, and heat
energy distribution. For Prod J 48:77–80
Zielonka P, Gierlik E (1999) Temperature distribution during conventional and microwave wood
heating. Holz als Roh- Werkst 57:247–249
Chapter 9
Delaminations Induced by Weathering in Wood
and Wood-Based Composites Panels
Voichita Bucur
Contents
9.1 Background
The term “weathering” (Feist 1982) defines any of the physical, mechanical or
chemical process by which wood or wood based products undergo slow degrada-
tion induced by the weather (sunlight, wind, precipitations, diurnal and seasonal
changes in relative humidity, atmospheric pollution, etc). Knowledge about weath-
ering durability comes from practical experiences of end-users, from field tests and
from standardized laboratory tests. The weathering process affects only the surface
of wood or wood products. It was primarily accepted that the sunlight- ultravio-
let radiation, visible and infrared radiation- initiate the wood weathering. (Williams
2005). The UV radiation has sufficient energy to degrade lignin and carbohydrates,
while the visible light degrades wood extractives. Wood photo – degradation starts
after exposure to the sunlight (Bentum and Addo-Ashong 1977; Derbyshire and
Miller 1981, 1995; Groves and Banana 1986; Onishi et al. 1989). This process is
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 173
DOI 10.1007/978-90-481-9550-3_9, C Springer Science+Business Media B.V. 2011
174 V. Bucur
Table 9.1 Some values of the erosion of earlywood and latewood of different species exposed to
the North American climate for a period ranging from 4 to 16 years (Williams 2005)
very slow and is of about 5 mm thickness decreasing of a board during 100 years
(Feist and Mraz 1978). In Table 9.1 are given some values of the erosion of early-
wood and latewood of different species exposed to the North American climate for
a period ranging from 4 to 16 years. Erosion values for plywood made from differ-
ent species are given in Table 9.2. The erosion of earlywood is always greater than
that of latewood. The ratio between the erosion of latewood and earlywood in solid
wood, after 4 years of exposure varies between 2.37 and 2.5. The same ratio is dif-
ferent after 16 years of outdoor exposure and varies between 1.52 and 2.9 depending
on species. For the plywood, after 4 years of exposure this ratio varies between 2.45
for Douglas fir and 3.2 . . . 3.4 for Western red cedar and redwood plywood. After
16 years of outdoor exposure the variation of this ratio is very small and is between
1.75 for Douglas fir plywood and 1.46 for Western red cedar plywood and redwood
plywood.
During weathering wood colour changes, surface fibres are erode, checks are
developed, and mildew colonizes the surface. The boards warp and cup. Moisture
variations induced by dew, rain, snow, temperature variation and oxygen are causes
of wood degradation. The checking, splitting and warping degrade the surface more
Table 9.2 Erosion values for some wood products exposed to the North American climate for a
period ranging from 4 to 16 years (Williams 2005)
than the photo-oxidation. The wood – based composites can fail catastrophically
within several years if no protection against weathering. Wood decay affects in the
same time the wood exposed to weathering. This aspect is not treated in this chapter.
Literature reviews on wood degradation induced by weathering, were published
periodically and have dealt with colour change, erosion, free radicals, surface wet-
ting characteristics, anatomical changes, mechanical properties and conservation of
fine art objects (Sell and Wälchi 1969; Sell and Leukens 1971; Feist 1982; Feist and
Hon 1984; Feist 1990: Uzzielli 1998, 2006; Deglise and Dirol 2000; Unger et al.
2001; Williams 2005). These studies were motivated primarily by the behaviour of
different wood coatings exposed to weathering, for which the reference has been
always the non coated wood.
The understanding of wood weathering mechanism, the effects on physical and
chemical properties allow the development of methods for inhibiting or retard-
ing the degradation, which might maximize the service life of wood, wood based
products and wood-based composites in any type of climate. Wood exposed
outdoor to weathering develops surface checks when tensile stress due to the
anisotropy of shrinkage and swelling exceed the elastic limits, inducing fracture
(Schniewind 1963). In severe cases, checks affect the mechanical properties of the
structural elements and destroy the appearance of the historical monuments built
in wood.
The aim of this chapter is to review the studies focused on the generation of fis-
sures or cracks and the development of delaminations induced by weathering on
non coated wood and wood based composites. The aspects related to photochem-
ical and biological processes of natural or treated wood are not discussed in this
chapter.
Numerous wood species are used for outdoor purposes because of their durability
and relatively high mechanical properties. In outdoor utilization wood is subjected
to variable environmental conditions which determine changes in surface properties
such as color change, surface roughening, crack development, etc., therefore the
changes in surface properties of wood species during weathering are of significant
practical concern. The fundamental aspects of solid wood weathering have been
discussed firstly by Coupe and Watson (1967) and more recently by Hon (1981) or
Williams (2005). The literature is scarce on knowledge related to wood anatomy and
the mechanisms of crack formation, propagation and development of delaminations
in wood induced by weathering (Donaldson 2010). Wood durability testing under
natural outdoor exposure conditions requires long time experiments. To simulate
this long time of natural weathering, accelerated tests are proposed by different stan-
dards (ASTM D 2898; ASTM D 2017-05; ASTM D 2481-05) or research reports
(Roux and Podgorski 2000). Accelerated ageing tests have the merit of relative short
duration.
Note that accelerated ageing tests are well suited for ranking the durability
performance of difference species and evaluating wood preservatives.
176 V. Bucur
Fig. 9.1 Identification of checks at wood surface Christy et al. (2005, Figure 1, Figure 2, Figure 3).
(a) the selected area of a southern pine specimen (b) the brightness profile along a line across the
centre of the specimen (c) the checks detection on the brightness profile
Table 9.3 Possible interactions between the weathering and the development of delaminations in
non coated wood
wind, etc. The meteorological conditions of each site are typical. The severity of cli-
mate in different areas is extremely variable. Transferring weathering results from
one location to another seems to be really difficult. Very few attempts have been
made to develop a classification system for climatic stresses of building structures
in general and of wooden building in particular (Creemers et al. 2002). The major
aim was to categorise the intensity of decay hazard for wooden constructions based
on empirical equations including air temperature, number of days with minimum or
maximum precipitations, etc. The equation 9.1 gives the climatic index developed
for European climatic conditions:
Iglobal nr .Rsum
C.I = + (9.1)
20 500
where:
C.I = climatic index
I global = global irradiation on planes tilted 45◦ facing South (kWh/m2 )
nr = number of days with precipitations more than 0.1 mm
Rsum = annual sum of precipitation (cm)
The numbers 20 and 500 are introduced to weigh the relative importance of radi-
ation and rainfall and to keep the resulting climate index for the European sites
around 100.
Creemers et al. (2002) noted that the climatic index CI showed reasonable corre-
lation only to global irradiation and that including the data on precipitations “hardly
improved the model”. In our opinion this means that to validate the tests, for each
weathering experiment the specific climatic conditions must be declared in the final
written report.
– on the radial surface, the cracks are localised and appeared at the annual ring
border and in the early wood. Small cracks in the annual ring border are
predominant.
– on the tangential surface, the cracks appeared in the latewood and across the whole
exposed surface.
– macroscopic cracks propagated along the whole length of the specimen, having a
large depth and being wider than 0.25 mm. Small short cracks had a depth of one
or few annual rings.
Table 9.4 The average of total crack length per unit are (m/m2 ) a after 33 months outdoor
exposure and subsequent conditioning at 12% moisture content (data from Sandberg 1999)
Sweden and showed a difference in durability between the heartwood and sapwood,
heartwood being more durable then the sapwood. As astonishing as this could be,
in this study it was demonstrated that the origin, the time of felling, the annual ring
width and the distance from the pith had no influence on the durability of Scots pine.
eroded by rain and wind. Figure 9.4 shows a zone of pine earlywood eroded by
wind and weather. Because of the tensile stress developed by the moisture dynam-
ics, fracture is initiated in 3 . . . 10 cell rows (Fig. 9.5). Crack propagates through the
first earlywood cell rows and seldom more than two rows from the annual ring bor-
der. In coniferous the microdensitometric components are very different in latewood
and earlywood. This explains why the cracks initiate at the limit of the annual ring,
Fig. 9.4 Eroded zone of earlywood in pine after 33 months of outdoor exposure (Sandberg
1999, Figure 9)
9 Delaminations Induced by Weathering in Wood 183
Fig. 9.5 Crack initiation and propagation in spruce, at the border of the annual ring
(Sandberg 1999, Figure 7)
between the latewwod and earlywood. The spruce and pine have numerous large
bordered pits in radial cell walls. Their pits are degraded in the early stage of weath-
ering (Sandberg 1999; Miniutti 1967; Borgin 1970, 1971). The cracks propagate
radially into the pits torus, which destroy the equilibrium of the cell wall and induce
new stress in the cell wall. The next stage involves the degradation of the cell wall.
The cracks develop parallel to the microfibrils in S2 ; in early wood and latewood.
The microfibrils contain an important amount of cellulose. The microfibrils are the
most stable structural element of tracheids. The degradation advancement destroys
the cohesion between microfibrils and different cell layers.
The degradation pattern of the tangential plane is different from the radial one,
because the fine structure of these two walls is very different. The tangential cell wall
has no sculpturing to initiate the crack. In this case the weaker zone is the middle
lamella. The delamination in tangential cell wall initiates in the middle lamella and
propagates very sharply up to ten cell rows (Fig. 9.6). Delamination in the middle
lamellae is very evident in the latewood on the tangential exposed surfaces.
The initiation, the development and the propagation of cracks due to weathering
can be explained by the development of stress field induced by the gradient of mois-
ture content. The shrinkage and swelling in tangential direction are about the twice
of that in radial direction. The exposure to rapid alternation of rain and strong sun-
light induces high moisture gradients between the surface areas and the underlying
wood, which determines more formations of cracks on tangential surfaces than on
radial surfaces. The latewood has higher density and shrinks more than the early-
wood. This also explains why the crack initiates at the annual ring limit between the
184 V. Bucur
Fig. 9.6 Delamination in the middle lamella of the exposed tangential surface of spruce latewood
(Sandberg 1999, Figure 13)
latewood and earlywood and then propagates through the earlywood. The orienta-
tion -radial or tangential- plays an important role in crack initiation and propagation
because of the strong anisotropy of wood. Tangential surfaces have more and deeper
cracks than radial surfaces. Radial surfaces should be selected for a better behaviour
to weathering.
Sandberg and Söderström (2006) studied at microscopic level the crack forma-
tion due to weathering of radial and tangential sections of pine and spruce after 61
months. The following factors have a notorious influence on the development of
delamination:
– in spruce, the annual ring orientation. The tangential section had 1.7 . . . 2.2 times
greater mean total crack length per unit area than the radial sections.
– in pine, the annual ring orientation. The tangential section had 2.2. . . 2.6 times
greater mean total crack length per unit area than the radial sections
– in both species, tangential sections have more and deeper cracks than radial
surfaces, the cracks on the tangential sections occur in both earlywood and
latewood
– in both sections, in radial sections cracks occur primarily at the annual ring border,
sometime they extend also in the earlywood.
– in both species degradation of cell wall cracks tend to follow the microfibril ori-
entation in S2 . In pine, because of numerous piths, the radial cell wall of the
earlywood degraded at an earlier stage of weathering
9 Delaminations Induced by Weathering in Wood 185
a b
Fig. 9.8 Transversal section of yellow-poplar before and after outdoor exposure for 30 day. (Hon
and Feist 1986, Figure 4) (a) transversal section before exposure (b) transversal section after expo-
sure. The middle lamella is eroded, the surface is roughening (c) tangential section after 30 days
of exposure, slight separation of the procumbent cells in rays and roughening of cell walls
average of the largest crack length in aspen was 55 mm ±34 mm and in spruce was
123 mm ± 97 mm. The boards either of spruce or aspen, positioned near the periph-
ery of the tree had more longer (>120 mm) cracks than the internal boards, because
in those boards there is a predominant influence of the tangential anisotropic direc-
tion. Aspen got a high number of relatively short cracks rather spruce got fewer, but
longer and injurious. For both species, the boards near the bark have more injuries
than those near the pith. The boards near the pith contain more juvenile wood than
those near the bark. On the other hand, the juvenile wood has an important number
of medulary rays which prevent the generation and development of cracks.
a b
Fig. 9.9 Delamination observed on the transversal section of Southern pine cross section after
exposure to UV (λ > 200 nm) during 1000 h. (Williams 2005, Figure 7.26, Figure 7.27) (a) Cross
section before the exposure (b) cross section after 1000 h exposure to UV (λ > 200 nm). The
middle lamella was completely eroded producing delamination between the tracheids
188 V. Bucur
Fig. 9.10 Delamination observed on the radial section of Southern pine section after exposure to
UV (λ > 200 nm) during 1000 h. Deterioration of pits generated cracks and delamination between
tracheids (Williams 2005, Figure 7.28)
Fig. 9.11 Delamination observed on the transversal section of yellow poplar after exposure to UV
(Hon and Feist 1986, Figure 7, Figure 8, Figure 9) (a) cross section after 500 h irradiation with
λ > 200 nm; (b) cross section after 1000 h irradiation with λ > 220 nm (c) cross section after
2000 h irradiation with λ > 220 nm
9 Delaminations Induced by Weathering in Wood 189
Fig. 9.12 Delamination observed on the tangential section of yellow poplar after exposure to UV
(Hon and Feist 1986, Figure 10, Figure 11) (a) roughening of the tangential section after irradiation
with λ > 220 nm during 2000 h (b) delamination into the cell wall induced by the deterioration of
pits observed on tangential section after irradiation with λ > 220 nm during 2000 h after 1000 h
irradiation
Fig. 9.13 Delamination observed on stereoscopic micrographs of three species (auri – Acacia
auriculiformis, robusta – Eucalyptus robusta and ipe – Tabebula spp) before and after 600 h
artificial weathering (Kishino and Nakano 2004, Figure 7)
9 Delaminations Induced by Weathering in Wood 191
Fig. 9.14 Contact angle versus the parameters of cracks (area, average length and width) (Kishino
and Nakano 2004, Figure 6)
in wettability between species was due to the structural modifications (i.e. cracks)
induced by the artificial weathering treatment whereas the increase in weattability
was due to the chemical changes of wood surface.
The aim of section is to review the studies focused on the development delam-
inations induced by weathering in wood-based composites panels produced with
natural, untreated wood.
To our knowledge, the literature is very scarce in articles referring only to unfinished
wood products or wood based composites panels exposed to weathering (Shaler
et al. 1988; Biblis 2000). Related information on delamnation and surface quality
in non coated products can be extracted from the literature on panels’ durability, in
which the non coated products are used as reference to evaluate the performance
of different finishes. The corresponding literature is very abundant. Plywood was
probably the first product on which systematic studies on weathering performance
during outdoor exposure were published (Selbo 1969; Koch 1967, 1970; Black et al.
1976; Hunt and Matteson 1976; Hayashi et al. 2002, 2005).
Biblis (2000) studied the effect of weathering on surface quality and structural
properties of six species (redwood, western redcedar, Douglas fir, lauan, baldcypress
and southern yellow pine) of untreated commercial plywood siding after 6 years
of exposure in Alabama, USA. Qualitative information only was reported – “the
surfaces of all species developed splits and cracks”.
Williams et al. (2005) reported the resistance to checking and warping of western
juniper lumber and particleboard for siding and decking uses. The specimens have
192 V. Bucur
been exposed outdoor during 6 years. Rating of checking decreased severely during
the first year of exposure, after which the rating becomes relatively constant.
Williams and Feist (2007) reported cracking development on surface of yellow-
poplar and sweetgum plywood panels (16 mm) used for siding and outdoor exposed
for 16 years near Madison, Wisconsin, USA. Plywood surfaces were flat grain and
were made from lathe peeled veneers. The rating of cracking dramatically developed
during the first two years of exposure, after which was quite constant.
The durability of structural softwood after 30 years indoor exposure was reported
by Raknes (1997), and used to compare the performance of eight urea-formaldehyde
glues, casein and resorcinol for structural softwood bonding. ASTM D 1101 – 59
(today not in use) was used to define and measure visually the delamination of
structural elements. For the outdoor exposure, the effect of the weathering was evi-
dent after 10 years of exposure and was combined with fungi infestation, shrinkage
inducing stress in wood and glue.
The delaminations in finger-jointed and laminated southern pine posts has been
reported by Shaler et al. (1988) in relationship with the strength and durability of
phenol-resorcinol joints of treated and untreated southern pine posts. The specimens
used were bocks cut from the test billets submitted to cyclic exposure as recom-
mended by the American national standard for structural laminated timber AITC T
-110 ( exposure to vacuum, 25 in. Hg, for 30 min, followed by 75 psi air pressure,
and 24 h exposure to drying in an air forced-air oven at 160 F). The delamination
was measured visually and expressed as percent of glueline on the cross-section al
faces.
9.4 Summary
On our planet, there are probably more buildings such as residential, commercial
industrial or farming constructed with wood and wood-based composites than any
9 Delaminations Induced by Weathering in Wood 193
other structural materials. The widespread use of wood and wood-based composites
in the construction of buildings has both an economic and aesthetic basis. The
natural beauty of these products is difficult to match with other materials.
The weathering durability of wood, wood products and wood – based composites
is an important quality factor, essential for environmental friendly exterior uses in
buildings. Weathering produces unfavorable quality changes such as discoloration,
roughening and checking of surface or modification of physical and mechanical
properties. Weathering of wood depends on many environmental factors and among
them the ultraviolet component of sunlight is responsible for cell wall lignin de-
polymerization. This effect is amplified by reining water which erodes the surfaces
exposed to weathering. The erosion process id very slow and was estimated to be
of about 5 mm thickness decreasing of a board during 100 years of exposure in
Northern Hemisphere. The erosion of the earlywood is always greater than that of
latewood.
The checks and delaminations at wood surface can be detected visually or auto-
matically. Visual detection refers to the check length, width, depth or number.
Modern methods use reflection of laser and fluorescent light to detect cracks at
wood surface. The width of cracks has been determined with stylus profilometry.
Automatic detection of checks at wood surface was performed using waveform
analysis of the reflected digital image. Variation in the brightness profile of a board
is also a good indicator of checks presence. At macroscopic scale, the factor of
influence on the delaminations induced by long term outdoor exposure are: wood
anisotropy, wood texture, annual ring structure, the presence of specific wooden
tissues such as juvenile wood, reaction wood, heartwood and sapwood. It is com-
monly accepted that wood anisotropy plays a determinant role in crack formation.
The degradation pattern of RL plane is different from that of TL plane. At micro-
scopic scale, in softwoods, the delamination was produced by wind and rain or
snow erosion of tracheids. Cracks propagated through the first earlywood cell
rows near the annual ring border. The pits were degraded from the early stage of
weathering. The cracks developed parallel to the microfibrils in S2 , their cohesion
is affected. Artificial weathering exposure induced similar type of degradation as
outdoor exposure in a shorter time.
References
Achenbach JD, Gautesen AK, Mendelsohn DA (1980) Ray analysis of surface waves interaction
with an edge crack. IEEE Trans Son Ultrason SU 27:125–129
Achenbach JD, Komsky IN, Lee YC, Angel YC (1992) Self-calibrating ultrasonic technique for
crack depth measurement. J Nondestruct Eval 11(2):103–108
Angel YC, Achenbach JD (1984) Reflection and transmission of obliquely incident Rayleigh waves
bu a surface – breaking crack. J Acoust Soc Am 75:313–319
Arnold M, Lemaster RL, Dost WA (1992) Surface characterization of weathered wood using laser
scanning system. Wood Fiber Sci 24(3):287–293
ASTM D 1037-06a Standard Test Methods for Evaluating Properties of Wood-Base Fiber and
Particle Panel Materials, hardboard, medium density fiberboard, www.astm.org/Standards/
D1037.htm. Accessed 15 December 2008
194 V. Bucur
ASTM D 2017-05 Standard Test Method of Accelerated Laboratory Test of Natural Decay
Resistance of Woods, decay, evaluation, laboratory, natural, resistance. www.astm.org/
Standards/D2017.htm. Accessed 15 December 2008
ASTM D 2481-05 Standard Test Method for Accelerated Evaluation of Wood Preservatives
for Marine Services by Means of Small Size Specimens. www.fpl.fs.fed.us/documnts/. . ./
fpl_2008_schirp001.pdf. Accessed 15 December 2008
ASTM D 2898 – 08e1 (2008) Standard practice for accelerated weathering of fire-retardant-treated
wood for fire testing
Bentum ALK, Addo-Ashong FW (1977) Weathering performance of some Ghanaian timbers.
Technical Note 26, Forest Products Research Institute, Ghana
Biblis EJ (2000) Effect of weathering on surface quality and structural properties of six species
of untreated commercial plywood siding after 6 years of exposure in Alabama. For Prod J
50(5):47–50
Black JM, Mraz EA, Lutz JF (1976) Performance of softwood plywood during 10 years exposure
to weather. For Prod J 26(4):24–27
Borgin K (1970) The use of the scanning electron microscope for the study of weathered wood. J
Microsc 92(1):47–55
Borgin K (1971) The mechanism of the breakdown of the structure of wood due to environmental
factors. J Inst Wood Sci 5(4):26–30
Borgin K, Parameswaran N, Liese W (1975) The effect of aging on the ultrastructure of wood.
Wood Sci Technol 9:87–98
Christy AG, Senden TJ, Evans PD (2005) Automated measurement of checks at wood surface.
Measurement 37:109–118
Coupe C, Watson RW (1967) Fundamental aspects of weathering. Proc Ann Conv Br Wood Preserv
Ass 2:37–49
Creemers J, Meijer M de, Zimmermann T, Sell J (2002) Influence of climatic factors on the
weathering of coated wood. Holz Roh- Werkst 60:411–420
Deglise X, Dirol D (2000) Durabilité des bois et problèmes associés. Hermes Science Publishing,
Paris
Derbyshire H, Miller ER (1981) The photodegradation of wood during solar degradation. Part 1:
Effects of structural integrity of thin wood strips. Holz Roh- Werkst 39:341–350
Derbyshire H, Miller ER, Turkulin H (1995) Investigation into the photodegradation of wood
using microtensile testing. Part I: the application of microtensile testing to the measurement
of photodegradation rates. Holz Roh- Werkst 53:330–345
Donaldson L (2010) Delamination of wood at microscopic scale: current knowledge and methods.
(Chapter 6, this volume)
Dong R, Adler L (1984) Measurements of reflection and transmission coefficients of Rayleigh
waves from cracks. J Acous Soc Am 76(6):1761–1768
Evans PD, Donnelly C, Cunningham RB (2003) Checking of CCA –treated radiata pine decking
timber exposed to natural weathering. For Prod, J 53(4):66–71
Evans PD, Wingate – Hill R, Cunningham RB (1997) The ability of physical treatments to reduce
checking in preservative treated slash pine posts. For Prod J 47(5):51–55
Flaete PO, Hoibo OA, Fjaertoft F, Nilsen TN (2000) Crack formation in unfinished siding of
aspen (Populus tremula L.) and Norway spruce (Picea abies (L) Karst.) during accelerated
weathering. Holz als Roh- Werkst 58(3):135–139
Feist WC (1982) The structural use of wood in adverse environments. Van Nostrand Reinhold Co,
New York, NY, pp 156–178
Feist WC, Hon DNS (1984) Chemistry and weathering and protection. In Rowell RM (ed) The
chemistry of solid wood, Advances in Chemistry Series 2007 American Chemical Society,
Washington DC
Feist WC (1990) Outdoor wood weathering and protection. In Rowell RM, Barbour JR (eds)
Archeological wood: properties, chemistry and preservation, Proceedings of the 196 meeting
American Chemical Society, September 25–28 Los Angeles
9 Delaminations Induced by Weathering in Wood 195
Feist WC, Mraz EA (1978) Comparison of outdoor and accelerated weathering of unprotected
softwoods. For Prod J 28(3):38–43
Groves KW, Banana AY (1986) Weathering characteristics of Australian grown radiate pine. J Inst
Wood Sci 10(5):210–213
Hayashi T, Miyatake A, Harada M (2002) Outdoor exposure tests of structural laminated veneer
lumber. I Evaluation of physical properties after six years. J Wood Sci 48:69–74
Hayashi T, Miyatake A, Fu F, Kato H, Karube M, Harada M (2005) Outdoor exposure tests of
structural laminated veneer lumber. II Evaluation of the strength properties after nine years. J
Wood Sci 51:486–491
Hon DNS (1981) Weathering of wood in structural use. In “Environmental degradation of engi-
neering materials in aggressive environments. Proceedings of 2nd international conference on
environmental degradation of engineering materials. September 21–23 Blacksburg VA
Hon DNS, Feist WC (1986) Weathering characteristics of hardwood surfaces. Wood Sci Technol
20:169–183
Hunt MO, Matteson DA Jr (1976) Structural characteristics of weathered plywood. Mater Res
Stand 6:508 –516
Imamura Y (1993) Morphological changes in acetylated wood exposed to weathering. Bull Wood
Res Inst Kyoto Univ no 79:54–61
Kamden DP, Zhang J (2000) Characterization of checks and cracks on the surface of weathered
wood. Int Res Group on Wood Preservation Doc 2000, IRG/WP 00 – 40153
Kishino M, Nakano T (2004) Artificial weathering of tropical woods. Part 1: Changes in
wettability. Holzforschung 58:552–557
Koch P (1970) Delamination of southern pine plywood during three years of exterior exposure. For
Prod J 20(11):28–31
Koch P (1967) Minimizing and predicting delamination of southern pine plywood in exterior
exposure. For Prod J 17(2):41–47
Kollman FP, Coté WA (1968) Principles of wood science and technology. Springer, Berlin
Kucera LJ, Sell J (1987) Weathering behaviour of beech wood in the ray tissue region. Holz als
Roh – Werkst 45(3):89–93
Kuo M, Hu N (1991) Ultrastructural changes of photodegradation of wood surfaces exposed to
UV. Holzforschung 45(5):347–353
Li ZD, Achenbach JD, Komsky IN, Lee YC (1992) Reflection and transmission of obliquely
incident surface waves by an edge of a quarter space: theory and experiment. J Appl Mech
59:349–355
Lidington BH, Silk MG (1975) Crack depth measurements using a single surface wave probe. Brit.
J. NDT 17:165–167
Lopez C, Doval FF, Dorrio BV, Blanco-Garcia J, Bugarin J, Alen JM, Fernandez A,
Fernandez JL, Perez – Amor M, Tejedor BG, (1998) Fibreoptic reflectometric technique
for the automatic detection and measurement of surface cracks. Meas Sci Technol 9(9):
1431–1431
Luekens von U, Sell J (1972) Bewitterungsversuch der EMPA mit Holz und Anstrichen für
Holzfassaden. Separatabdruck aus der Schweizerischen Maler – und Gipsermeister - Zeitung
Nr 14
Miniutti VP (1967) Microscopic observations of ultraviolet irradiated and weathered softwood
surfaces and clear coatings. US Forest Service Res. Paper FPL 74
Onishi M, Tsujimoto Y, Sakuno T (1989) Degradation behaviour of exterior wood after outdoor
exposure for 2 years. Res Bull Tottori University Forests no 26:93–99
Raknes E (1997) Durability of structural wood adhesives after 30 years aging Holz Roh- Werkst
55:83–90
Raczkowski J (1980) Seasonal effects on the atmospheric corrosion of spruce micro-sections. Holz
als Roh- Werkst 38:231–234
Rietz RC (1961) Accelerated weathering of red oak treated with various preservatives used to treat
crossties. For Prod J 11(12):567–575
196 V. Bucur
Roux ML, Podgorski L (2000) The advantages of having in the future a European accelerated
weathering test for wood finishes. Surf Coat Int 83,8:399–403
Rydell A, Bergström M, Elowson T (2005) Mass loss and moisture dynamics of Scots pine (Pinus
sylvestris L) exposed outdoors above ground in Sweden. Holzforschung 59:183–189
Sandberg D (1999) Weathering of radial and tangential wood surfaces of pine and spruce.
Holzforschung 53(4):355–364
Sandberg D (1997) The influence of annual ring orientation on crack formation and deformation
in water soaked pine (Pinus sylvestris L) and spruce (Picea abies (L) Karst.) timber, Holz als
Roh- Werkst 55(3):175–182
Sandberg D (1996) The influence of pith and juvenile wood on proportion of cracks in sawn timber
when kiln dried and exposed to wetting cycles. Holz als Roh- Werkst 54(3):152
Sandberg D, Söderström O (2006) Crack formation due to weathering of radial and tangential
sections of pine and spruce. Wood Mater Sci Eng 1(1):12–20
Schniewind AP (1963) Mechanism of crack formation. For Prod J 13(11):475–480
Selbo ML (1969) performance of southern pine plywood during five years exposure to weather.
For Prod J 19(8):56–60
Sell J, Wälchi O (1969) Changes in the surface texture of weather exposed wood. Material und
Organismen 4:81–87
Sell J, Leukens U (1971) Investigations on weathered wood surface – Part II: weathering
phenomena on unprotected wood species. Holz als Roh- Werkst 29:23–31
Shaler SM, Wright B, Manbeck HB (1988) Strength and durability of phenol resorcinol joints of
CCA –treated and untreated southern pine. For Prod J 38(10):59–63
Sudiyani Y, Takhashi M, Imamura Y, Minato K (1999) Physical and biological properties of
chemically modified wood before and after weathering. Wood Res Bull Kyoto Univ no 86:1–6
Szymany R, McDonald KA (2004) Defect detection in lumber: state of the art. For Prod J
31(11):34–44
Unger A, Schniewind AP, Unger W (2001) Conservation of wood artifacts: a handbook. Springer,
New York, NY ISBN-10: 3540415807
US Forest Products Lab (1974) Wood Handbook: Wood as an engineering material. USDA
Agriculture Handbook 72, Madison, WI
Uzzielli L (1998) Historical overview of panel making techniques in central Italy. In Dardes K,
Roth A (eds) The structural conservation of panel paintings. Proceedings of the symposium
journal Paul Getty museum, pp 24–28 April, Los Angeles, pp 110–135
Uzzielli L (2006) The approach of wood science towards conservation of wooden objects belonging
to the cultural heritage. Proceedings of the cultural heritage and science Gent Belgium 5–7
December
Wahl P, Hanhijarvi A, Silvennoinen R (2001) Investigation of micro cracks in wood with laser
speckle intensity Opt Eng 40(5):788–792
Williams RS (2005) Weathering of wood. In Rowell RM (ed) Handbook of wood chemistry and
wood composites. CRC, Boca Raton, pp 139–185
Williams RS, Feist WC (2007) Durability of yellow-poplar and sweetgum and service life of
finished after long-term exposure. For Prod J 54(7/6):96–101
Williams RS, Sawan L, Sotos P, Knaebe M, Feist WC (2005) Performance of finished western
juniper lumber and particleboard during outdoor exposure. For Prod J 55(11):65–72
Chapter 10
Delamination in Timber Induced by Drying
Nawshad Haque
Contents
10.1 Introduction
In the context of wood drying, delamination can be regarded as a defect due to
poor drying where separation between wood fiber cells occurs on surface (surface
checks) or inside of wooden boards (internal checks) and in the form of splits (i.e.
end split) at the end of the timber board. Solid wooden board can be considered as
naturally laminated product. Before discussing delamination aspect of wood fibers
during drying in further detail, it is helpful to discuss about basic principles and
purpose of wood drying.
N. Haque (B)
Division of Minerals, CSIRO Clayton, Bag 312, Clayton South, VIC 3169, Australia
e-mail: Nawshad.Haque@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 197
DOI 10.1007/978-90-481-9550-3_10, C Springer Science+Business Media B.V. 2011
198 N. Haque
Wood drying (also called seasoning in the wood literature) is the removal of water
from the timber as economically and with as little damage as possible. A recent
textbook by Keey et al. (2000) covers many aspects of timber drying, including the
fundamental basis of this technology. Wood drying is an inevitable part of wood pro-
cessing. Typical times for drying softwood such as radiata pine (Pinus radiata) are
in the order of days but for hardwoods such as eucalypts are in the order of months
and year. Although high-temperature drying technology developed in Australasia
reduced drying times for softwoods from months to days, optimized process still
is a challenge for the industry to avoid product degrade. Reducing delamination
of board either internally or externally during wood drying is also part of such
challenges.
An important objective of seasoning timber is to dry it to the equilibrium mois-
ture content (EMC) before use. EMC is the moisture content (MC) of wood when
it is in equilibrium with the temperature and humidity of the surrounding air. Thus
the gross dimensional changes through shrinkage are carried out during drying and
before final use. Timber is dried to conform to the average of the maximum and
minimum EMC that will be attained by the wood in service under fluctuations of
different climatic conditions. The movement in the components of the finished prod-
uct, relative to the dimensions at the times of fabrication, is also kept to a minimum
if dry timber is used. Thus drying is the first step towards realizing the maximum
attainable dimensional stability from any timber during use. To eliminate move-
ment completely in wood, chemical modification of wood is a possible technology.
This is the treatment of wood with chemicals to replace the hydroxyl groups with
other hydrophobic functional groups of modifying agents (Stamm 1964). Among all
the existing processes, wood modification with acetic anhydride has considerable
promise due to the high anti-shrink or anti-swell efficiency (ASE) attainable with-
out damaging the wood properties. However, acetylation of wood has been slow in
commercialisation due to the cost, corrosion and the entrapment of the acetic acid
in wood. There is extensive literature relating to the chemical modification of wood
(Rowell 1983, 1991; Kumar 1994; Haque 1997).
Drying, if carried out promptly after the felling of trees, also protects tim-
ber against primary decay, fungal stain and attack by certain kinds of insects.
Organisms, which cause decay and stain, generally cannot thrive in timber with
a MC below 20%. In this context, wood MC is the total amount of water contained
in a piece of wood. In timber technology, moisture content is expressed as a percent-
age of the oven dry weight. MC determinations with the oven drying method give
an average MC for the piece. Actual MC at different locations within this piece may
vary quite considerably depending on the moisture gradients and drying characteris-
tics of the particular timber species. The average MC at the beginning of the drying
process for test pieces and kiln samples is termed the initial MC, and the average
MC of the stack and at the end of drying is termed the final MC. Several, though
not all, insect pests can live only in green timber. Dried wood is less susceptible to
decay than green wood (above 20% MC).
Apart from the above important advantages of drying timber, the following points
are also significant (Walker et al. 1993; Desch and Dinwoodie 1996):
10 Delamination in Timber Induced by Drying 199
• Dried timber is lighter, and hence the transportation and handling costs are
reduced.
• Dried timber is stronger than green timber in terms of most strength properties.
• Timbers for impregnation with preservatives have to be properly dried if
proper penetration is to be accomplished, particularly in the case of oil-type
preservatives.
• In the field of chemical modification of wood and wood products, the material
should be dried to a certain MC for the appropriate reactions to occur.
• Dry wood works, machines, finishes and glues better than green timber. Paints
and finishes last longer on dry timber.
• The electrical and thermal insulation properties of wood are improved by drying.
The external drying conditions (e.g. temperature, relative humidity and air velocity)
control the external boundary conditions for drying, and hence the drying rate, as
well as affecting the rate of internal moisture movement. The drying rate is affected
by external drying conditions (Walker et al. 1993; Keey et al. 2000), as will now be
described.
10.2.1 Temperature
If the relative humidity is kept constant, the higher the temperature, the higher the
drying rate. Temperature influences the drying rate by increasing the moisture hold-
ing capacity of the air, as well as by accelerating the diffusion rate of moisture
through the wood.
The actual temperature in a drying kiln is the dry-bulb temperature, which is the
temperature of a vapor-gas mixture determined by inserting a thermometer with a
dry-bulb. On the other hand, the wet-bulb temperature is defined as the temperature
reached by a small amount of liquid evaporating in a large amount of an unsat-
urated air-vapor mixture. The temperature sensing element of this thermometer is
kept moist with a porous fabric sleeve (cloth) usually put in a reservoir of clean
water. A minimum air flow of 2 m s−1 is needed to prevent a zone of stagnant
damp air formation around the sleeve (Walker et al. 1993). Since air passes over the
wet sleeve, water is evaporated and cools the wet-bulb thermometer. The difference
200 N. Haque
between the dry-bulb and wet-bulb temperatures, the wet-bulb depression, is used
to determine the relative humidity from a standard hygrometric chart (Walker et al.
1993). A higher difference between the dry-bulb and wet-bulb temperatures indi-
cates a lower relative humidity. For example, if the dry-bulb temperature is 100◦ C
and wet-bulb temperature 60◦ C, then the relative humidity is read as 17% from a
hygrometric chart.
The relative humidity of air is defined as the partial pressure of water vapor divided
by the saturated vapor pressure at the same temperature and total pressure (Siau
1984). If the temperature is kept constant, lower relative humidities result in higher
drying rates due to the increased moisture gradient in wood, resulting from the
reduction of the MC in the surface layers when the relative humidity of air is
reduced. The relative humidity is usually expressed on a percentage basis. For
drying, the other essential parameter related to relative humidity is the absolute
humidity, which is the mass of water vapor per unit mass of dry air (i.e. kg of water
per kg of dry air).
Broadly, there are two distinct methods by which timber can be dried:
• natural drying; and
• artificial drying.
Air drying is a natural drying method, while artificial drying includes kiln dry-
ing (mainly), vapor drying, solvent drying, infra-red drying, high frequency drying,
microwave drying, superheated steam drying, and chemical seasoning using salts.
10 Delamination in Timber Induced by Drying 201
Although most of these drying can take place at atmospheric pressure, drying in
vacuum is undertaken to gain the advantage of low boiling point of water for certain
hardwoods (e.g. radio-frequency or superheated steam vacuum drying). Solar dry-
ing utilizes solar energy in such a way that it makes the process relatively simple and
less expensive compared with kiln drying (Desch and Dinwoodie 1996), although
the analysis of solar kiln performance is relatively recent compared with the use of
solar kilns.
Air drying is the drying of timber by exposing it to the sun (Fig. 10.1). It depends on
the natural conditions of wind, sunshine and rain. The technique of air drying con-
sists mainly of making a stack of sawn timber (with the layers of boards separated
by stickers) on raised foundations, in a clean and dry place, under shade if avail-
able. Atmospheric air is the drying agent, and the rate of drying largely depends on
climatic conditions. The air enters the stack of timber at the top, particularly at the
edges of the stack, picks up moisture, is cooled and then drops to the bottom. Some
air flows horizontally through the stack, driven by the wind. For successful air dry-
ing, positive, continuous and uniform flow of air throughout the pile of the timber
needs to be considered, including the prevailing wind direction and the layout of the
air drying yard (Desch and Dinwoodie 1996).
Fig. 10.2 High temperature kilns for drying softwood (courtesy of Windsor Kilns Pty. Ltd)
the process, deliberate control of temperature, relative humidity and air circulation
is provided to give conditions at various stages (MC or times) of drying the tim-
ber to achieve effective drying. For this purpose, the timber is stacked in chambers,
called wood drying kilns (Fig. 10.2), which are fitted with equipment for manipula-
tion and control of the temperature and the relative humidity of the drying air and its
circulation rate through the timber stack (Walker et al. 1993; Desch and Dinwoodie
1996).
Kiln drying provides a means of overcoming the limitations imposed by erratic
weather conditions. In terms of the fundamental drying process, the process of kiln
drying does not differ from air seasoning. In both cases, unsaturated air is used as
the drying medium, and the principle of drying is the same, i.e. removal of moisture
from the interior to the surface of the timber. Almost all commercial timbers of the
world are dried in industrial kilns. A comparison of air drying, conventional kiln
and solar drying is given below:
• Timber can be dried to any desired low moisture content by conventional or solar
kiln drying, but in air drying, moisture contents of less than 18% are difficult to
attain for most locations.
• The drying times are considerably less in conventional kiln drying than in solar
kiln drying, followed by air drying.
• In air drying, a large amount of capital investment is needed for stacking a
large amount of timber stock over a longer period than in conventional or solar
kilns, although the installation for these kilns, as well as their maintenance and
operation, is expensive (in terms of capital items).
• Air drying needs a large land area, so the land rental is significant.
• In air drying, there is little control over the drying elements, so drying degrade
cannot be controlled.
• The temperatures employed in kiln drying typically kill all the fungi and insects
in the wood if a maximum dry-bulb temperature of above 60◦ C is used for the
10 Delamination in Timber Induced by Drying 203
drying schedule. However, all the fungi and insects may not be killed by air
drying temperatures and may subsequently attack the timber.
• In air drying, the rate of drying may be very rapid in the dry summer months,
making timber boards liable to crack and split (i.e. delamination), and too slow
during the cold winter months.
• The significant advantages of conventional kiln drying include higher throughput,
and precision (better control of the final moisture content). Conventional kiln
and solar drying both enable wood to be dried to any MC regardless of weather
conditions. This makes both solar and conventional kiln drying more appropriate
for most large-scale drying operations than air drying.
• defects that arise due to the shrinkage anisotropy, related to the warping of timber
boards; and,
• defects that arise due to uneven drying, associated with the rupture of the wood
tissue and delamination.
Defects related to warping include cupping, bowing, twisting, spring and dia-
monding. Defects related to the rupture of tissues include checks (surface, end
and internal), end splits due to delamination, honey-combing and case-hardening.
Some defects due to shrinkage anisotropy and uneven drying are shown in Fig.10.3.
Collapse is another form of defect that usually occurs above the fiber saturation
point (FSP) and is not related to shrinkage anisotropy. The term FSP, described first
by Tiemann in 1906, is the MC of wood at which wood cell walls are completely
saturated with water and the cell cavities are empty (Stamm 1964). The FSP for
most wood is around 30%. Collapse occurs as a result of the physical flattening of
water filled fiber cells due to the action of internal tension. Collapse is often seen as
a corrugation, or “washboarding” of the board surface (Innes 1996). Collapse gen-
erally occurs at high moisture content above FSP (Booker and Koga 2003). In the
context of wood drying, delamination is also considered a form of degrade due to
poor drying.
Australian and New Zealand Standard Organizations (AS/NZS 4787 2001) have
developed a standard for timber quality and set five criteria for measuring drying
quality. These are the MC gradient; the presence of residual drying stress (i.e.,
related to case-hardening); surface, internal and end checks (i.e. delamination);
collapse; distortions; and discolouration caused by drying. This standard has also
described the drying quality classification, how to assess each of these drying qual-
ity criteria, and the limits for each criterion to be acceptable within a quality class.
Check is a drying defect which occurs when tensile stresses cause the wood fibers
to separate and form cracks. Checks are cracks that are visible at the surface but
do not extend across a piece of board. Collapse is a drying defect that occurs when
tensile stress in the core results in the formation of internal cavities. Permanent set
is a change in the properties of wood that can occur during drying when the stresses
exceed the elastic limit. Permanent set prevents normal shrinkage of the timber and
can lead to more obvious defects such as casehardening and honeycombing. Most of
the drying defects develop due to uneven drying stress. Drying stress is the force per
unit area that occurs in some zones of drying wood as a result of uneven shrinkage
in response to the MC gradients that develop in wood. MC gradient is a progressive
difference in MC between the core and the surface of a piece of wooden board.
Timber has significantly anisotropic properties. Its strength is much greater along
the grain (longitudinal direction) than across the grain (radial and tangential
directions), typically by an order of magnitude. For example, with Douglas-fir
(Pseudotsuga menziesii) at 12% MC (density 545 kg m−3 ), the compressive strength
parallel to the grain is ten times higher than that in the perpendicular direction,
whereas the tensile strength parallel to the grain is fifty times higher than that
perpendicular to the grain (Desch and Dinwoodie 1996).
The strength properties of timber vary with the wood species, density, MC, grain
direction and with the direction of load applied on the material. During the drying
of timber, the steepest gradients in MC occur across the grain. Hence timber boards
usually fail mechanically and delamination occurs across the grain during drying.
The cross-grain properties are the most important ones when modeling drying-
induced stresses. By contrast, timber in structural applications is generally loaded
along the grain, in order to take the advantage of the higher strength in the lon-
gitudinal direction. Therefore, mechanical properties for structural applications are
usually measured along the grain. As a consequence, properties for timber loaded
across the grain are scarcer in literature than those for along the grain direction. Both
the mechanical and drying behavior of timber are also species-dependent, which
means that the relevant properties need to be determined for each species of timber.
A major challenge for researchers and industries in timber drying is to reduce
both drying time, and the loss of product value due to drying defects. An optimized
drying schedule trades off between these two opposing objectives (i.e. quality and
206 N. Haque
productivity). The aim of timber drying research is to achieve both faster drying and
better product quality. The optimized schedule dries timber in the fastest possible
time, which is limited by a maximum stress/strain level at the surface and throughout
the board thickness that is unlikely to cause the timber to crack (i.e. delamination).
Thus the delamination in the forms of internal and surface checks is minimized. The
absence of checks may be an indicator of lower stress and possibly the reduction of
other defects and distortion.
In modeling and simulation approach, essentially a drying model is used to pre-
dict drying rate, MC and temperature of slices of boards throughout the thickness
direction, and a stress model is used to predict and estimate developed stresses and
strains (Haque 2002). Based on the predicted stress/strain level, the drying rate is
optimized for a prescribed time horizon. This approach is an effective alternative
to traditional methods and significantly reduces the number of trials, thus saving
time and money. A combined drying and stress model can be used applying the
nonlinear model predictive control technique. This approach is implemented within
the MATLAB and Fortran programming environment to optimize drying schedules,
i.e. a set of temperatures and relative humidity (for time-based or MC based sched-
ule). These schedules have potential application in the timber processing industry to
increase throughput and for better quality product.
Wood is also hygroscopic, so it will adsorb moisture from the atmosphere if it
is dry and correspondingly yield moisture to the atmosphere when wet. The results
are swelling and shrinkage of wood, respectively. Shrinkage and swelling of wood
only occur when the wood MC falls below the FSP (Desch and Dinwoodie 1996).
Shrinkage in wood caused by changing MC is the fundamental cause of many drying
defects because the MC is usually non-uniform in the wood, causing differences in
shrinkage strain through the board thickness, leading to (instantaneous) stresses and
strains that can cause the timber to crack and delaminate. Other processes, such as
the development of viscoelastic and mechanosorptive strains, also have important
effects on the development of stresses and strains in timber. Not only is a knowledge
of shrinkage behavior in timber necessary to optimize drying schedules to produce
uncracked timber, but shrinkage properties are also required to assess the fractional
oversize cutting of green timber boards required at the saw-bench before drying.
The amounts of shrinkage and swelling are different in the three main planes of
the wood, i.e. longitudinal, tangential and radial directions. Wood is thus said to be
anisotropic with respect to shrinkage and swelling. The longitudinal shrinkage along
the grain is often small, volumetrically less than 0.1%, whereas the radial (from the
pith outwards) shrinkage is 1.5–4.5% and the tangential shrinkage (direction tan-
gential to the growth ring), often being the highest, is in the range of 2.0–9.5%
(Kollmann and Cote 1968). Wood shrinks, on average, about fifty times as much in
the transverse directions as in the fiber direction (Stamm 1964). The defects aris-
ing from shrinkage anisotropy are warping (various forms e.g. twisting, cupping,
diamonding, bowing, and spring or crook).
As soon as a tree is felled, the timber begins to lose moisture. Eventually, if the
log has no contact with moist surfaces, the wood will reach the EMC. Above FSP,
the timber loses water from cell lumens, producing few significant effects on the
10 Delamination in Timber Induced by Drying 207
dimensions of the cell walls (Stamm 1964; Walker et al. 1993). However, below
FSP, the timber loses water from the cell walls, causing them to shrink due to the
volume of moisture lost. Differences in structure between different species of tim-
ber, including different cell wall thicknesses, mean that there is a range of shrinkages
for various timbers. If all other timber properties (mechanical, physical, and trans-
port ones) are equal, a piece of timber showing higher shrinkage will be more likely
to crack than one showing lower shrinkage, because the same moisture-content gra-
dient will lead to greater differential shrinkage strain between the centre and surface
of the board showing higher shrinkage.
10.6.1 Hardwood
A combined hardwood drying and stress model was used for this example to predict
drying stress and optimize drying schedule that can reduce or avoid delamination.
A modified approach should be used to develop a practically applicable opti-
mized drying schedule as shown in Fig. 10.4. Since the early of part of drying is
most critical in terms of product quality, drying from green to this stage where the
strain is maximum has to be carried out using the fixed drying schedule tested in the
laboratory. It is possible that, after this critical stage, drying can be accelerated sig-
nificantly using the optimization procedure. Once the strain reaches its maximum, it
can be kept constant for faster drying as long as the dry-bulb temperature does not
exceed the upper limit.
For the schedule shown in Fig. 10.4, the drying started from an initial MC of
70%. After around 10 days of drying, stress and resulting strain started to develop
in timber board. Theoretically strain should develop once the MC reduces below
FSP. However, the strain (across grain direction) started to be predicted while the
80 5
Moisture
Drybulb Strain
70
Wetbulb
4
Instantaneous strain (%)
60
50 3
40
30 2
20
1
10
0 0
0 10 20 30 40 50 60 70 80 90
Time (days)
Fig. 10.4 Predicted strain and MC for a typical optimized drying schedule for hardwood
208 N. Haque
MC is around 55%. This is because although the average MC was around 55% but
the surface MC reduced below FSP. The strain reached at its maximum to about 4%
(or 0.04 mm mm−1 ) at around 50 days. If the strain develops above 4% then it is
likely that the timber will delaminate and internal or surface checks can occur. Once
the maximum stress level is passed then the drying can be accelerated by increase
temperature and wet bulb depression. This strategy can be implemented for this kind
of drying schedule.
The use of higher dry-bulb temperature (60◦ C) at the early stage of drying
of hardwood boards such as blackbutt (Euclayptus pilularis) produced excessive
degrade observed in an experimental trial in a laboratory kiln, i.e. numerous surface
checks developed on the boards and end splits were present. The reason may be that
the surface of the board was drying too fast while the centre of the board was still
too wet, or that significant collapse occurred. This higher MC gradient would cause
delamination. Also from industrial experience and literature (Vermass 1995; Innes
1996), generally a low dry-bulb temperature is used for drying eucalyptus in the
early part of drying to avoid cracks (i.e. delamination) and end splits. This limiting
(low dry-bulb) temperature for drying optimization also comes from the consider-
ation that collapse occurs above certain temperatures for some hardwoods. Innes
(1996) introduced a concept called the “collapse threshold temperature” from the
predictions of a single fibre model and the observations of drying trials. He found
that if a timber board with a MC above the FSP is dried at temperatures above the
collapse threshold temperature, then it will collapse. It was estimated that the col-
lapse threshold temperature is below 40◦ C for ash type eucalypts. He found that the
collapse threshold temperatures was 24◦ C and 30◦ C for Tasmanian and Victorian
Eucalyptus regnans, respectively, from both slice and board test methods. However,
blackbutt is not an ash-type eucalyptus. Innes (1996) also stated that the collapse
threshold temperatures are different for the earlywood and latewood of a wood sam-
ple, but he did not report any test results. The ambient temperatures in the air-drying
yard at Boral Timber’s Herons Creek site are frequently above 35◦ C during summer
months, but there is no significant collapse found with the air-dried and finally kiln
dried timber (mainly blackbutt) at this sawmill site. Thus it is reasonable to assume
that the collapse threshold temperature may be higher for Eucalyptus pilularis.
10.6.2 Softwood
A combined softwood drying and stress model was used for this example to predict
drying stress developed for a 40 mm thick board of radiata pine (Pinus radiata).
The basis of these models has been described in earlier studies (Haque 2007; Pang
and Haslett 1995). Typical values found in literature for radiata pine wood proper-
ties such as density (i.e. 450 kg m-3 ), initial MC (i.e. 150% dry-basis) and wood
permeability were used. The air flow for drying was assumed to be 7 m s−1 .
Generally dry-bulb temperature below 60◦ C is considered low temperature dry-
ing, and at 90◦ C is termed accelerated conventional temperature (ACT) drying. The
predicted drying time is shown in Fig. 10.5 which indicates that the drying time
10 Delamination in Timber Induced by Drying 209
160
50/40
140 60/40
Moisture content (%) 60/30
120
70/40
100 70/30
90/60
80
120/70
60 140/90
40
20
0
0 20 40 60 80 100 120 140
Time (hours)
Fig. 10.5 Predicted drying time for a 40 mm board for a range of schedules (50/40 means 50◦ C
dry-bulb temperature and 40◦ C wet-bulb temperature)
can be from 9 h for 140/90◦ C schedule to up to 130 h for a 50/40◦ C schedule. The
predicted drying time decreases with increasing dry-bulb temperature and wet-bulb
depression, which is expected.
1
Surface stress (MPa)
0 90/60
0 20 40 60 80 100 120 70/40
70/30
–1
60/30
60/40
–2 50/40
–3
–4
Time (hours)
1.2
1
0.8
Stress at centre (MPa)
0.6 90/60
0.4 70/40
70/30
0.2
60/40
0 60/30
0 20 40 60 80 100 120 50/40
–0.2
–0.4
–0.6
–0.8
Time (hours)
Fig. 10.7 Predicted stress at the centre of a 40 mm board for a range of schedules
It is evident from the figures that in the beginning of drying, there is little stress
both at the surface and centre of the board since the board is still wet. There is some
condensation at the surface, and the surface moisture content increases above the
initial value that is usually less than the average. The corresponding stress profile
shows that there is no significant stress at that time, since the moisture contents are
above the fiber saturation point throughout the timber board. After about 5 h a small
amount of stress (about 0.5 MPa) is predicted near the surface. Gradually stress
develops and increases up to a maximum of 1 MPa. The predicted maximum stress
is higher for the schedules with higher wet-bulb depressions. Once the MC gradi-
ent is well developed from the surface to the centre, the associated stress is larger
through the board, with tension stress at the surface and compression at greater
depths at the centre. In this case, positive stress is considered as tensile and negative
stress is compressive in nature. The surface and centre shows opposing behavior,
as theoretically expected. The effect of temperature that causes faster drying rates
is also evident. As drying progressed, the MC gradient is significant (surface to
centre), and this causes a very large amount of tensile stress (on surface) and com-
pression stress (towards the centre). Towards the end of drying, the MC gradient
becomes shallower or milder than that for the early drying stages, resulting in lower
stress. At the end of drying, since the MC becomes flattened, and the stresses are
also reduced. If the developed stress reaches above the strength of wood, delami-
nation will occur either internally or on the surface resulting in internal or surface
checks on the radiata pine timber board.
2
120/70 Surface
1.5 140/90 Surface
120/70 Centre
1
Surface stress (MPa)
140/90 Centre
0.5
0
0 2 4 6 8 10 12 14
–0.5
–1
–1.5
–2
Time (hours)
Fig. 10.8 Predicted stress at the centre and surface of a 40 mm board for two high temperature
schedules
drying although the maximum stress was above 1.5 MPa compared with 1 MPa for
ACT drying.
The consequence of this drying for delamination to occur is that the predicted
stress is below the threshold failure stress of radiata pine. The failure stress across
the grain of wood is around 4.5 MPa at 150◦ C. Thus it is unlikely that delamination
will occur under these drying conditions since developed drying stress is below the
failure stress limit.
10.7 Summary
Delamination is the separation of wood fibre cells in the context of solid wood prod-
uct such as timber board. Wood is a natural composite material with anisotropic
properties. In the context of wood drying, delamination can be regarded a form of
degrade due to poor drying condition or due to wood properties such as wood with
low strength. In wood drying, these defects are commonly known as surface checks,
internal checks and end splits on timber boards. Some basic concepts of drying have
been elaborated in this chapter, following the prediction of drying time, developed
drying stress, strain on the board surface and centre for a range of drying schedules
using mathematical models both for a hardwood and a softwood. The combined
drying and stress model can be helpful to optimize drying process that can reduce
delamination in solid wood during drying. The higher the drying temperature, the
shorter will be the drying time. However, the likely developed stress during drying is
predicted to be higher due to aggressive drying conditions. The optimum drying con-
dition seeks the trade-offs between these two opposing behaviors thus developing
drying schedule that dries timber in least time and without significant degrade such
as delamination. This is the main goal of many timber drying research programs.
212 N. Haque
Acknowledgements Some of the contents of this chapter were taken from the authors’ Ph.D.
thesis supervised by Associate Professor Tim Langrish at the University of Sydney. Parts of this
chapter also were developed from the authors time of earlier works at New Zealand Forest Research
Institute and its’ Ensis Joint Venture with CSIRO. The author thanks his current employer CSIRO
Minerals for providing time to write this chapter.
References
Booker R, Koga S (2003) How collapse and internal checking happen in boards during drying.
Wood processing newsletter 33, July 2003. New Zealand Forest Research Institute, Rotorua
Desch HE, Dinwoodie JM (1996) Timber: Structure, properties, conversion and use. 7th edn.
Macmillan Press, London, 306p
Haque MN (1997) The chemical modification of wood with acetic anhydride. MSc Dissertation.
The University of Wales, Bangor, 99p
Haque MN (2002) Modelling of solar kilns and the development of an optimised schedule for
drying hardwood timber. The University of Sydney, Australia, 354p
Haque MN (2007) Simulation of temperature and moisture content profiles in a Pinus radiata board
during high-temperature drying. Drying Technol 25(4):547–555
Innes T (1996) Improving seasoned hardwood timber quality with particular reference to collapse.
PhD Thesis, University of Tasmania, Australia, 172p
Keey RB, Langrish TAG, Walker,JCF (2000) Kiln-Drying of lumber. Springer, Berlin, 326p
Kollmann, FFP, Cote WA Jr. (1968) Principles of wood science and technology. I. Solid Wood.
Springer, NewYork, NY, 592p
Kumar S (1994) Chemical modification of wood. Wood and Fiber Sci 26(2):270–280
Pang S, Haslett AN (1995) The application of mathematical models to the commercial high-
temperature drying of softwood lumber. Drying Technol 13(8–9), 1635–1674
Pordage LJ, Langrish, TAG (1999) Simulation of the effect of air velocity in the drying of
hardwood timber. Drying Technol 17(1–2):237–256
Rowell RM (1983) Chemical modification of wood. For Prod Abstr 6(12):363–382
Rowell RM (1991) Chemical modification of wood. In: Hon D.N.-S, Shiraishi N (eds) Wood and
cellulosic chemistry, pp 703–756, Marcel Dekker, New York, NY
Siau, JF (1984) Transport processes in wood. Springer, NewYork, NY 245p
Stamm AJ (1964) Wood and cellulose science. Ronald Press, New York, NY, 509p
Standard Australia (2001). Timber – assessment of drying quality. Australian/New Zealand
Standard (AS/NZS) 4787, Sydney, 24p
Vermass HF (1995) Drying eucalypts for quality: material characterstics, pre-drying treatments,
drying methods, schedules and optimisation of drying quality. S Afr For J, 174:41–49
Walker JCF, Butterfield BG, Langrish TAG, Harris JM, Uprichard JM (1993) Primary wood
processing. Chapman and Hall, London, 595p
Part III
Delamination in Different Products
Chapter 11
Industry Prospective of Delamination in Wood
and Wood Products
Contents
11.1 Introduction
The highly-competitive forest industry is facing mounting pressures from both sub-
stitutions and imports. The North America industry is not only adjusting to the
changing wood quality and timberland ownership but is also coping with the cycli-
cal downturn of the market. It is very challenging to balance high-mill throughput
and lumber quality to meet customer needs while remaining profitable in a rapidly
changing industry. Logistics costs and governmental regulations are just a cou-
ple of obstacles to overcome. These hurdles are unappreciated by those with little
experience in the operations. Of the many great ideas presented, only a few are oper-
ationally feasible. For example, while it is possible to sort raw materials at various
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 215
DOI 10.1007/978-90-481-9550-3_11, C Springer Science+Business Media B.V. 2011
216 C.L. Huang
stages throughout the value chain to reduce lumber variability, such efforts may not
be cost-effective for the following reasons:
The reduced availability of public forest lands, due to the Endangered Species
Act in the Pacific Northwest region in the late 1980s, resulted in a spike in log
prices and, consequently, resulted in mill closures (Parrish 1991). Other examples
of regulatory impacts (Howard and Westby 2007) include the:
The industry is adapting to these paradigm shifts and regulatory changes while
minimizing the risks, and at the same time, maintaining its core competency:
growing trees to make wood and paper products.
Many products come from the forests: solid wood products (SWP) such as lum-
ber, wood panels including plywood, oriented strand board (OSB), and engineered
wood products (EWP) mainly used for residential housing construction. In addition
to these building products, pulp is used to make paper and corrugated packaging,
fluff pulp for personal care products, as well as a myriad of other commercial prod-
ucts. Although specialty and proprietary products have their own niche markets,
forest products are mainly commodities, vulnerable to fluctuating prices in a highly
competitive market. Product manufacturers must meet governmental regulatory and
code requirements while satisfying the customer’s demand for lower prices. This
is a challenge. Manufacturers must make sufficient revenue to cover the cost of
manufacturing which includes raw materials, capital depreciation, skilled labor, and
escalating energy costs.
In many cases, making lower-grade products will not cover the cost of manu-
facturing. For example, normally sawmills in the south-eastern United States lose
money for every piece of southern yellow pine (SYP) lumber that is below #2 grade.
Even with the declining wood quality, manufacturers have found that by applying
technologies to create value-added products, they can increase the perceived value
of the wood products instead of trading them as a commodity. The Framer Series
Lumber (FSL) is an example of this kind of technology-oriented strategy to fulfill
customer needs.
The durability against fluctuating environmental conditions and biodegradations
makes wood preservation an important part of the industry. Wood products not only
compete with substitutions but also with other wood products. Lumber cut from
plantations has a higher percentage of juvenile wood and has a higher percentage
11 Industry Prospective of Delamination in Wood and Wood Products 217
of large knots than that from a natural forest stand. Since the raw material sup-
ply for high-strength lumber is diminishing (mostly as a result of shorter rotation
periods), products like EWP will continue to gain market shares in structural
applications.
EWP uses many components, in addition to wood, to make its products:
These failures can originate from the growing tree itself or compounded with
harvesting and manufacturing practices. Some failures can also be induced while the
product is in service. Checks or splits in wood may not be considered as a “defect or
failure” in some processes. For instance, in lumber treating, checks and incisions aid
in the penetration of the chemicals to the wood. Easy to split is desirable in firewood
production. In the case of pulping, the delamination in the chips is an advantage for
a thorough distribution of pulping solution.
The type, direction, frequency, duration, and intensity of stresses or impacts on
wood product’s weak areas, such as the thin-wall earlywood and the swirling grain
around a knot, can initiate splits, fractures, and breaks in the product. Wood-cell
structures like the angle-ply-laminated cell wall and the radial and longitudinal
alignment of specific types of cells are critical factors related to the propagation
of the crack. For EWP, delamination is mainly caused by adhesive failure and the
strength of its interaction to the wood.
Numerous articles on splits, shakes, and cracks of SWP and EWP delaminations
have been published in scientific journals and trade magazines. These topics have
also been covered in previous chapters as well. This chapter will first review the
unique structure of wood and then discuss splits, shakes, checks, cracks, and delam-
inations that originate in the unique wood structure of the living tree, as well as,
through manufacturing, and finally, the products in-service.
In the line above from Joyce Kilmer’s poem “Trees”, the word “poems” can be
substituted with “wood products”. Wood is a unique composite material with great
variability because of its natural origin. Trees and their roots are exposed to the envi-
ronment, and in the course of growing, withstand harsh winters and summers. As
with all living organisms, trees use various life-history strategies to adapt to adverse
growing conditions. Even for wood cut from a genetically cloned forest, there are
intrinsic within-tree variations as well as environmental heterogeneity variability
between trees.
The earliest utilization of wood by humans may have been for heating and for
cooking fuel. It is known that species with straight grains are easier to split while
species with twist and interlocking grains are difficult to split. Some aboriginals
were known to beat the stems repeatedly to separate splints for basket weaving and
for other utilitarian purposes.
The tendency of wood to split along the longitudinal-radial (LR) planes is related
to its longitudinal and radial alignment of cells, as well as, the orientation of the
ray cells in the radial direction. During the vigorous growing season when plenty
of water is available, for species with distinct ring structure, the tree’s cells in the
cambium divide and expand their size by turgor pressure, which makes the tree
grow taller and bigger in diameter. When a tree senses a shortage in water supply,
instead of pushing for more dimensional growth, it gradually reduces the rate of cell
division and also deposits more photosynthetate on the cell walls of the small lumen
(latewood or summerwood).
When the temperature is too cold to “fix carbon”, the tree goes dormant; when
the growing season returns, the tree starts producing large, thin-wall cells (early-
wood or springwood) again. For softwood, the density of latewood is two to three
times that of the earlywood, and the first formed earlywood has the lowest den-
sity. Besides easy splitting characteristics along the LR planes, ring curvature and
the discontinuity of density from the first formed earlywood are additional macro-
characteristics of wood related to the initiation and propagation of delaminating or
splitting failures.
Cellulose, lignin, and hemicellulose are the major cell wall components.
Cellulose is the world’s most abundant organic material. Linear polymers of glu-
cose or fibrils are tightly bundled or braided into microfibrils which are priminarily
crystalline. The microfibril bundles are embedded in a matrix of amorphous hemi-
cellulose and lignin. The crystalline regions of microfibrils are inaccessible to water.
Lignin is an amorphous, hydrophobic phenolic polymer, which softens in raised
temperatures. Hemicellulose consists of the amorphous cellulose chains that bridge
the lignin and the microfibril.
After the cellulosic framework is laid down, the lignin starts to bond to the hemi-
cellulose sites near the cell corner, then fills in the pectinaceous middle lamella, and
finally, influxes the cellulosic framework from the cell corner towards the lumen.
The layered structure of the cell wall includes a thin intercellular layer, the middle
lamella, a thin primary cell wall, and a thick secondary cell wall. The combination
of the thin middle lamella and the primary cell wall layers is termed the compound
middle lamella (CML).
11 Industry Prospective of Delamination in Wood and Wood Products 219
Fig. 11.1 Scanning Electron Microscope (SEM) Photo Showing Microfibril in S2 Layer of
Loblolly Pine Earlywood Tracheids. (Note the S1 layer at the lower left corner of the photo)
The primary cell wall has randomly distributed microfibrils; however, the sec-
ondary cell wall has three layers with the microfibril angle (MFA) winding in
different directions within each layer. The first formed thin S1 layer has microfibril
winding transversely, the thick S2 layer has microfibril twisting in a longitudinal
helix (Fig. 11.1), and finally, the thin S3 layer facing the lumen has a microfibril
angle close to the horizontal direction. The definition of MFA is the winding angle
of the S2 microfibril in respect to the longitudinal axis of the cell.
The micro-scale discontinuity across the double cell wall is between the CML
and the S1 layers because of differences between random versus organized microfib-
rils. Formation of the crystalline region creates contraction (the tension stress), while
the influx of lignin into the cellulosic framework expands the cell wall dimension
creating compression stress. Hemicellulose facilitates the lignification processes.
The first formed earlywood expands between the rigid xylem and the flexible cam-
bium, so the theoretical growth stress profile is compression on the xylem and
tension on the cambium side. Such growth stress pattern combined with the MFA
may tilt the orientation of a dividing and growing cell and result in spiral grain
patterns within the tree (Schulgasser and Witztum 2007).
Wood shrinks as the equilibrium moisture content (EMC) falls below the fiber
saturation point (FSP), which is around 30% moisture content. The swelling or
shrinking of wood generates a tremendous force. Such forces were used by ancient
Egyptians to quarry stones from rocks.
Wood from plantation loblolly shrinks in the longitudinal, radial, and tangen-
tial directions for about 0.2, 4, and 7.5% respectively. The resistant force against
shrinkage is closely related to the MFA and the amount of ray tissue in the wood.
If the microfibril is coiled like a spring with its MFA greater than 35 degrees, the
longitudinal shrinkage of the cell increases exponentially with the increasing MFA.
220 C.L. Huang
If the cell has a thick cell wall that can hold a high volume of water at FSP, then the
transverse shrinkage should be high. Ray cells provide more nucleuses to initiate
radial splitting and somewhat release the restraint on tangential shrinkage.
One of the driving forces of wood drying shrinkage is the amount of hemicellu-
lose, especially the galactan, which is found to be an important variable related to
the longitudinal shrinkage of wood (Floyd 2005). There are two types of reaction
wood: compression wood (CW) on the under side of a leaning softwood and ten-
sion wood (TW) on upper side of a leaning hardwood tree. The main function of
CW and TW of a tree is to regain its upright position which is critical for the tree’s
survival. Severe CW tracheids lack an S3 layer and their S2 layer contain checks,
making the tracheids appear like coiled springs enclosed by S1 and CML. TW has
high crystallinity which may generate contraction forces to pull the leaning stem
straight (Yamamoto 1998). The pushing force of severe CW may be hypothesized
as the extending length of a coiled spring upon the lateral force exerted by gravity.
Both CW and TW have high galactan content. CW also has high MFA, so the lon-
gitudinal shrinkage of CW can be 10–20 times higher than normal wood. The MFA
of hardwood is generally smaller than 30 degrees, so the impact of MFA on the
longitudinal shrinkage of hardwood is negligible. S3 is missing in TW’s gelatinous
fiber which is loosely bonded with S1, so the resistant force to longitudinal shrink-
age is influenced mostly by the large microfibril winding angle of the S1 layer. The
missing S3 layer, high galactan content, low lignin content, and the influence of
large angles of S1 microfibrils also make the longitudinal shrinkage of TW (two to
eight times) higher than normal wood. The S2 layer of the CW tracheids has high
MFA and contains cell wall checks. These combined with high galactan content
make the rate of longitudinal shrinkage of CW rises sharply when EMC falls below
10%, which may cause transverse cracks on the wood product. The brittle reaction
wood also makes the wood product prone to brash failure.
Various types of fiber separations and breakages may happen in a tree’s life. For
example, cracks and splits can be caused by:
The fiber separations in the tree may be visible, healed over, or hidden.
Mechanical damage during harvesting, handling, and mill processing can create
new splits or cracks or may exacerbate existing ones. During harvesting, end-splits
are the result of stress relief that occurs during cross-cutting young eucalyptus
stems. Such splitting is well-known. Remedies used to alleviate the internal stresses
11 Industry Prospective of Delamination in Wood and Wood Products 221
are: making circumferential grooves on the trunk; banding the log ends to restrain
the propagation of end splits; coating the log ends to delay drying stresses; and
processing logs fresh from harvesting. These examples of known practices are used
to reduce the impacts of end splitting on eucalyptus saw logs (Malan 1995).
Damage to the tree can occur when the tree is dropped during harvesting or
through rough handling during transportation to the mill. Such rough handling can
cause significant amounts of fractures and breakages especially around large knots
in the upper portion of the logs. These knots develop from intensively-managed
plantations. An experienced forester will consider the potential damage to neigh-
boring trees from harvesting during the pre-harvest tagging of high value trees, such
as those for pole-timber and for face-veneer. Using appropriate harvesting methods,
good handling techniques, and providing on-going training of skilled operators are
ways to minimize mechanical damage and ensure maximum value is recovered from
the plantation stand.
Volume deductions due to the log defects are taken when grading for different
log markets. The log is downgraded based on the amount of splitting there is in the
log from ring shake, felling shake, end blooming, breakage, splits, or seams. These
types of defects also impact the clear wood recovery for appearance grade products.
However, in North America, the main impact of end splits or log breakage on the
value of the saw log is the trimming loss in the two-foot increments for structural
lumber products.
Within the log, earlywood intra-ring checking, and radial splitting of radiata pine,
eucalyptus, and other species are well documented defects often observed in the low
density sapwood areas of the butt logs. The hairline delaminations exacerbate after
drying (Fig. 11.2).
Wetwood (wood in a standing tree that has become internally infused with water),
encourages the growth of anaerobic micro-flora. Wetwood is often found in hem-
lock, white fir, cottonwood, elm, as well as other species, and is associated with
heart or ring shake (the onion elm). Ring shake is the separation of fibers in the
growth rings. The separation occurs from environmental stresses on the tree. These
stresses can be from storms, winds, droughts or poor drainages that waterlog the
roots. The separation occurs at the low-density or micro-organism weakened cells.
Anaerobic bacteria, which grow in the damaged areas of the tree, may release
an odor and contribute to localize staining in lumber. Not only is wetwood difficult
to dry, the release of chemicals in the wetwood from drying may corrode the kiln
(Pong and Ward 1979). Ring-shake can also cause serious delamination failure in
wood products as shown in the glulam photo (Fig. 11.3).
Severe bending of the tree trunk due to gale-force wind or ice-damage buckles
the cell wall and the resulting heal-over may create transverse wrinkles on the trunk
of certain tree species. Although the strength of wood products with fiber buckling is
only slightly reduced, the toughness of the wood can be significantly decreased. For
instance, scaffolding boards containing compression creases or compression wood
boards may cause brash failure leading to serious accidents, injuries, or deaths.
The tremendous compressive stress in large trees of low-density hardwood may
develop compressive creases near the pith area or the brittleheart; such low-density
wood is difficult to use. Disease and insect infestations can result in canker and tip
dieback, which results in multiple leaders and split forks, further compounding the
tree’s vulnerability to disease and breakage.
Through genetic improvements, improved silvicultural practices, and proper har-
vesting and handling techniques future saw log values can be improved, especially
for species susceptible to split, crack, shake, and check. Properly controlled tim-
ing, intensity, and frequency of silvicultural activities can minimize abrupt changes
in the tree’s growth and reduce the chances of physical and biological damages to
the tree.
While these practices are desirable, it is a challenge for the industry to balance
improvement costs against profits. The production rate, various imposed EPA and
governmental regulations, as well as the ultimate wood quality, all have a strong
11 Industry Prospective of Delamination in Wood and Wood Products 223
impact on the profits. These variables are difficult to study because of the industry’s
uncertain future markets and changing regulations. Scientists and engineers offer
many solutions to the industry, but the solutions are often narrow in focus without
considering constraints in operation, changes in the market, or impacts of future
regulations.
Unlike the nondestructive evaluation (NDE) instrument used to test the sound-
ness of trees in public areas, tree and log are mostly graded by visual characteristics
in forest industry. More sensors are being used in the log breakdown and manufac-
turing processes where “the wood really meets the steel”.
Lumber grade recovery (#2 and better) and mill throughput (MBF/hour) are essen-
tial for a mill’s profitability in the southeastern United States. Log quality, sawing,
and kiln performances are among the important factors in determining the success of
the sawmills. Natural and manufacturing imperfections downgrade the SWP. Knots,
burl, decay, holes, intra-ring checking, and compression wood (CW) are natural
defects that originate in the tree. Further damage from poor manufacturing processes
can deform, distort, fracture, fragment, or crush the wood.
Fiber-separation and fiber-breakage such as shelling, slivers, tear outs, splits,
shakes, checks, cracks, loosened, chipped, torn grains (fragments still connected),
are defects created during the manufacturing processes. Following are examples of
causes and where the fiber damage can occur in the manufacturing process:
• species
• grain orientation during cutting,
• sharpness of cutting edges,
• setting of the gang, twin, trimmer, edger, planer, and the feeding and conveying
systems.
Defects can occur in the lumber when it is improperly processed through mis-
aligned systems (conveyors, hold-downs, saws and knives), poorly maintained
equipment (dull cutting edges) or improper drying schedule (drying stresses). In
cutting, the friction buildup between cutting tip and the wood will be exacerbated
by wood with high-silica content (e.g. certain high-density hardwood species). The
saw tip itself limits cutting speed (e.g., steel versus carbide). Localized high tem-
perature from the saw’s friction may degrade or soften the surface of the cutting tip.
Proper, routine saw filing is critical in a sawmill to ensure quality cuts. Saw-tooth
design and other bench techniques a saw filer uses can improve the cutting profi-
ciency, while saw blade coatings or cooling treatments can extend the life of the
cutting blades.
224 C.L. Huang
Similar to the log breakdown process, damage from mechanical stresses can be
caused from the lumber drop sorter, press rolls, hold downs, and the occasional
jackstraw or lumber spill that occurs in conveying and stacking the lumber. Rough
handling can create new or exacerbate existing splits and cracks in a piece of
lumber.
Further damage to the lumber can occur down stream in the mill process. Drying
stresses can cause surface and internal checks as well as lumber warp, which create
problems in the planer mill. To reduce drying stresses, in-kiln moisture and temper-
ature controls are used to adjust the drying conditions. In addition to the moisture
and temperature meters, scanners and sensors are used by modern, high-throughput
sawmills to minimize the impacts of natural defects and maximize the consistency
and reliability during the manufacturing process.
Log scanners are used at the merchandiser to optimize the bucking solution, and
at the primary breakdown to determine the best cutting pattern. Range finders are
positioned throughout the mill stream to monitor deflection, snaking, or other types
of misalignments and detect any imbalances at various cutting centers. Transducers
are mounted to monitor the vibrations of motor and hydraulic systems.
These monitoring sensors give important feedback data to the mill person-
nel. With this feedback, mill personnel can maximize the lumber throughput
while minimizing wastes and defects. The data from the sensors are used for
quality control (QC), alarms, and fine tuning the equipment during scheduled
maintenances.
The feed-forward and feedback of the sensor information can optimize the pro-
cess flow and prevent upset conditions. Frequently, however, the root cause of an
upset condition can be traced back to the raw material source. In this era of fast
changing technologies, redundant functions are often bundled into the data col-
lection packages, when only basic data are needed. One of the concerns from the
abundance of sensor data is that the mill personnel become overwhelmed with data
which are rarely analyzed. The human brain is a super computer with millions of
processors integrated together, but most computers only have one central processing
unit (CPU). A veteran sawyer can send a log along a conveyor to the cutting cen-
ter while watching the best cutting solution being displayed on computer monitor.
The computer cannot dynamically process the same observations and subsequent
decisions the sawyer makes on the variety of log mixes that may set off problems
downsteam in the mill process such as jackstraws, over- and under-drying, or planer
skips. Especially in new mill construction or modifications to existing mills, besides
reporting and archiving the mean, deviation, and the trends, involvement from expe-
rienced operators is essential for the engineers to extract critical information from
the data to improve the mill processes.
Consequences of lumber defects are manifested in terms of poor return to the
mill, and far more seriously, failure of the wood products in service. Defects of
SWP are defined in various grading books. Delamination defects can be minimized
through proper lumber handling and mill processing. Below are some examples of
the defects:
11 Industry Prospective of Delamination in Wood and Wood Products 225
• Skips and torn grains are mainly caused by misalignments and dull cutting edges;
• Splits are the separation along the fiber due to the physical and mechanical
stresses;
• Shakes are the separation of wood along the growth ring or a macro-delamination
along the weakened areas;
• Checks are mostly caused by seasoning or drying; and
• Cracks are the breakages across the fiber direction.
With the exception of machine stress rated lumber (MSR), machine evaluated
lumber (MEL), and some proprietary products such as FSL, the majority of SWP
in North America is still graded only on a visual basis. Although experienced
graders can consistently grade lumber with a high degree of accuracy, the mod-
ern sawmill is using scanning technology for grading lumber. This prevents grading
differences among graders and overcomes the grader fatigue factor from long shifts.
However, there are some disadvantages the grade scanners have over the human
graders.
The human brain has amazing capabilities for resolving visual defects. For exam-
ple, various types of hairline cracks, blonde knots, and worm holes can be detected
by the grader and the grading inspector, but are often missed by a high-speed grad-
ing machine. A missed hairline crack may cost the mill dearly in terms of additional
handling on the below-grade products (the products failed the inspection), as well
as potential customer claims from failed in-service products.
The impact of defects on SWP appearance grade applications is more straight-
forward, because the defects usually mean only a downgrade in value. However,
defects from fiber separation or breakages in a structural member are more severe
in consequence.
The influences of the defect on strength in tension, compression, or shear are
closely related to the orthotropic mechanical properties of clear wood. The ultimate
compression strength of metal, concrete, and natural stones is higher in compression
than in tension. By contrast, however, the ultimate strength of wood in tension is two
to three times that of compression due to its aligned microfibrils along the cell wall
of the hollow tubular structures.
Stiffness and strength of clear wood samples are affected by the combined effects
of the slope of grain (SOG), which is indicated by the checks on lumber surface
(Fig. 11.4), and unavoidable natural defects such as knots, decay, and compression
wood (CW). Compression wood (CW) may be hidden, but brash failure in products
made from CW or decay can result in serious injuries.
Detecting these defects before they manifest themselves in product failures can
be difficult. Interactions from the type, size, and location of the defects as well
as various structural designs and applications can be complex. For example, the
impact of grain deviation caused by knots is significant in tension, less important
in compression, and may be insignificant in shear strength. However, knots near the
edge, especially the side in tension, create high-stress concentrations which initiate
failures in structural lumber.
226 C.L. Huang
Fig. 11.4 Checks along the grain on loblolly pine lumber surface
Glulam is one of the oldest EWP. With two or more layers of finger-jointed lum-
ber glued together, (edge-to-edge and face-to-face), straight or intentionally curved
dimensional products can be made, filling a variety of market needs. For products
requiring sharp curves, thin lamstocks are used in the set up to reduce bending stress.
To make glulam products, the ends of graded lumber are inspected, prepared, and
finger jointed into a lamination. These laminations are planed, coated with adhe-
sive, and assembled into a specific pattern. The assembly is then clamped and cured
before being finished and fabricated into different classes of products.
A finger joint may cost more in the process, but it trims off less waste than a
scarf joint does. The optimum length of a structural joint is three to four times the
pitch. However, too many tips may create more stress concentration points or crack
initials. To prevent splitting at the valley when pressed, the length of the finger is
slightly shorter than the depth of the valley. The tip of the finger also needs to be
thin to minimize the size of voids.
Fiber reinforced polymer (FRP) strips can be used on the tension side of a glu-
lam to enhance bending strength. The FRP can also be pre-stressed to reduce peak
stresses. Primer or modified adhesive formulations may be necessary to laminate the
FRP-wood hybrid composite. Visual inspection and nondestructive testing are used
to cull defected products as a final step in manufacturing.
A drawback of glulam compared to other EWPs is the frequency of knots or
finger joints which reduce the strength of glulam products in service. Thinner
11 Industry Prospective of Delamination in Wood and Wood Products 227
laminations cost more in handling; however, they have a higher degree of random-
ization of the defects in the products.
Veneers have long been used in appearance applications where surface quality is
important such as in the furniture industry. EWP products made with veneers are:
• Plywood,
• Laminated veneer lumber (LVL), and
• Parallel strand lumber (PSL).
Lathe checks are unavoidable in rotary peeling of logs for veneer. These lathe
checks are caused by impact forces, knife rake angle as well as the flattening of a
curved veneer sheet when it leaves the lathe. Lathe checks are initiated on the loose
side of the veneer (the side facing the pith). The smaller the log or the thicker the
veneer, usually the more serious is the lathe-check. Veneers with rough surfaces and
deep lathe checks require a higher than normal glue spread, or extra filler to prevent
dry-out and over-penetration of the adhesive. Insufficient adhesive spread combined
with a rough surface will result in a starved glueline.
Proper preconditioning of the log, controlling the peeling speed of the lathe,
increasing nose bar pressure, heating the knife, and incising the tight side of the
veneer may reduce the adverse impacts of lathe checks. At the molecular level,
extractives in the wood may reduce the wettability of the veneer surface. Wood
extractives can also change the pH of the adhesive causing improper bonding or
curing.
Besides these surface properties, delamination of plywood and LVL is often
traced back to two areas: the dryer and the hot press. In the dryer, the under-
dried high moisture content (MC) veneers make the adhesive application and the
adhesive-wood reaction inconsistent. After peeling the log (or block), green veneers
are sorted into MC groups for a uniform drying. If necessary, radio frequency or
microwave dryers are then used to re-dry the wets to a target MC. After applying
the adhesive to the surface of the dried veneer, the coated veneers are laid up to a
target layer and thickness and then cured in a hot press. In the press, poor curing due
to improper MC-temperature profile or inadequate press-time or both often results
in delamination of plywood and LVL products.
For PSL, the process is different. The dried veneers for PSL production are
clipped into narrow strips, coated with adhesives, dropped into a forming trough,
and cured in a continuous press. The press is drawn by a steel belt into the throat of
a pair of platens with side dams. At this point, microwave energy is directed to the
strips through windows on the side dam. Densification and uniform heating in PSL
manufacturing ensures that the veneer strips are evenly distributed and that the resin
has thoroughly interlocked with the strips. This results in strong bonds between the
veneer strips. The PSL manufacturing process is also able to utilize veneer fish tails
and trims, making the PSL technology not only efficient in raw materials usage, but
also cost-effective.
Unlike veneer products that require large peeler blocks to make veneer sheets for
layups, strand products such as LSL and OSB, use low-density and low-cost small
228 C.L. Huang
diameter logs to produce long, thin strands (3 inches to 6 inches for OSB and as long
as 12 inches for LSL). These strands make for efficient forming and densification in
the manufacturing process.
However, by flaking or stranding the logs, within-strand fiber separations are
created, which result in strand breakage during the drying and blending processes,
especially for the longer strands. By preconditioning the log to soften it during
flaking or stranding, strand breakage is reduced. However, better stranding tech-
nology is needed to minimize damage from cross-grain cutting. Making processing
improvements is challenging because of the inherent variability in log properties
coupled with the demand for high mill throughput; normally, there are limited
resources in the mill to conduct research.
The coated flakes are formed into a billet by a disc-former to align the longer
flakes along OSB panel’s length for both top and bottom layers. The core layer is
formed by dropping coated shorter flakes onto a fin-former to align the flakes to the
width of the panel. The billets are cured in either a multiple-opening or a continuous
press for OSB or in a steam-injection press for LSL.
The strands themselves become well-bonded with resin because of the high reac-
tivity between pMDI and wood. Unlike the veneer products, strand products like
OSB and LSL is less likely to delaminate. However, when the edge of OSB panel
gets wet in-service, the swelling of the strand may damage the adhesive bond.
Since OSB and LSL use small diameter logs for making strands, defects become
randomized making a lower cost and more uniform product than that of Glulam
or veneer products. Although the alignment is not as good as in glulam or veneer
products, the strength and stiffness of OSB and LSL are significantly enhanced by
densification. Product delamination is normally not a problem with OSB or LSL.
Instead, thickness or edge-swelling is the main disadvantage of the highly-densified
OSB and LSL Products.
It is challenging to model the curing and bonding process of strand products
because preheating, moisture content, press pressure, press temperature, and press
time influence the complicated adhesive-wood interactions. For example, using
a mild press condition not only reduces the emissions of volatile organic com-
pounds (VOC) but also minimizes delaminations. Delamations or blows occur
because of the eruption of trapped steam or vapor among the resin-coated strands.
Carbonization is also reduced because of the milder press conditions. Mild pressing
conditions; however, can extend the press time required to cure the resin.
There are trade-offs, however, between speed and quality in manufacturing. The
industry wants a low-cost, minimum press time, room temperature or ambient cur-
ing adhesive, with a strong bond. Creative adhesive formulations, innovative press
designs, and reliable monitoring technologies are sought after to achieve both high
production rates and consistently strong bonds between the wood and the adhesive.
Adhesive selection is determined by the application in which it will be used
(exterior versus interior), mechanical property requirements of the end product,
performance durability, manufacturing process throughput, and purchasing cost. In
most industrial cases, the adhesive cost and the process throughput are the dominant
11 Industry Prospective of Delamination in Wood and Wood Products 229
factors in choosing the adhesive. A number of resins on the market have advantages
and disadvantages:
A release agent is required on the face layers of the OSB press when using
the fast-curing pMDI to prevent adhesive buildup on the platens. To avoid hav-
ing to use the release agent, PF is normally used for OSB face layers and pMDI
in the cores. However, the PF-pMDI compatibility may cause delamination in the
transition zones.
Depending on the type and the formulation of the adhesives, the moisture con-
tent needs to match the allowable ranges of the adhesives and the press conditions
to ensure a strong bond. Adhesives can be diluted with water and solubilized by
steam which may results in excessive adhesive flow and causes poor bonding. On
the other hand, pMDI, which reacts with water in room temperature but it is not
water miscible, is used in the steam-injection press of LSL manufacturing.
Fillers like walnut shell flour and china clay are usually added to the mix to
reduce adhesive cost, fill-in voids, and increase the rigidity of the adhesive in veneer
products. Other types of additives can be used to improve the flow, curing, and
durability of the adhesives.
With regard to health and safety, using formaldehyde scavengers or other binders
is one solution to reduce the emissions. The human body is highly reactive to
pMDI atomized by spraying or combustion; therefore, proper personal protection
equipment (PPE) is required in the manufacturing processes.
Many factors contribute to the cause of delaminations during the manufactur-
ing process of EWP. Since wood is a poor heat conductor, inadequate curing of the
adhesive at the center glueline is a common cause of plywood and LVL delamina-
tions. Temperature sensors at the center glueline of LVL are used to monitor the
press temperature to ensure proper curing. Preheating veneers, steam-injection, or
microwave/RF energy is necessary to cure thick EWP products such as PSL and
LSL. Another source of delamination can be traced to substandard adhesives, which
are often discovered by QC personnel to be the cause of poor bonding. This makes
it critical to select a reliable supplier who delivers a consistent-quality resin. An on-
line, ultrasound blow detector not only serves as a QC tool, but the information also
230 C.L. Huang
can be used for fine tuning the manufacturing processes while minimizing waste
and rejects.
Factors that contribute to a strong bond are:
Some adhesives are stronger and more durable than wood, so the quality of the
bond is determined by the load at failure and the percent of wood failure. In gen-
eral, the penetration of the adhesives is usually deeper in the earlywood lumen
of softwood or in the vessel of hardwood. The penetration of adhesive into the
cell wall to create interlocking networks is dependent on the reactivity and the
molecular size of the adhesive. The reaction starts from the secondary cell wall
facing the lumen penetrating any exposed fissures and pores. Additives and pretreat-
ment are ways to facilitate a strong bond that will prevent the delamination of the
products.
Fig. 11.5 Lumber (after tested for bending strength, modulus of rupture, MOR) of the same log
(loblolly pine), note the splits and cracks around the knots
232 C.L. Huang
mechanism even more complicated, especially at the micro-scale level where identi-
fying the nucleation, detecting the initiation, and tracking the propagation of a crack
can be a challenge.
Uneven shrinkage of the adhesive and the wood develops curing stresses which
becomes yet another variable to consider. The properties of the bond itself change
with in-service temperature and moisture conditions. Stress distribution and the
manner in which stresses are transferred to the composing elements will affect the
performance of the EWP. Unlike SWP, the crack propagation of EWP may follow
an irregular path of least resistance.
Depending on the toughness around the tip of a crack, the delamination may
be arrested, stabilized or worsened, leading to accelerated crack propagation and
ultimate failure. Whether the existence of a delamination is a structural con-
cern or not, is governed by the coverage ratio of the delaminated area to the
sound area. For example, a partial shallow check in a glulam column or post will
not affect the structural integrity of a building. However, on the other hand, a
deep delamination will reduce the effective load-bearing properties of a beam or
header.
Redundancies are normally integrated into the building codes after evaluat-
ing the in-service product applications, the type and frequency of support loads.
Bracing and connectors are also important factors regarding the load or stress-
transfer in the structure. Nondestructive evaluation (NDE) methods are helpful in
detecting and evaluating the safety of a structural member. Self-healing adhesives
may also provide a mechanism to arrest crack initiations or reverse crack propaga-
tions. Advancements in sensor and nanotechnology will provide solutions to make
stronger and safer wood products in the future.
In general, neither the industry nor the regulatory agencies are the cause of unsafe
buildings. Frequently, it is improper installation, insufficient maintenance, or exces-
sive loads that are the culprits for the loss of lives or property damage. Although
there are reasonable redundancies required by code for a structure, unusual natural
forces, such as hurricane force winds, can overwhelm the load limits. Homeowners
too can be at fault. Backyard decks are a popular weekend project for homeown-
ers, but if the deck is poorly designed or poorly maintained, the deck could fail
because of excessive loads or from decay. These failures lead to the serious per-
sonal injuries and property damage. As expected, the durability of wood products is
more demanding for outdoor applications.
11.6 Summary
With the increasing complexity in building codes and sophisticated computer-aided
designs, structural wood products are being used at near their maximum allowable
design limits. Wood checks, splits, cracks, and delaminations have potential to cause
great losses for the consumer. The industry is responding by making improvements
in their products to satisfy the end-user requirements using QC and QA processes.
The industry is also improving the raw material base by selectively breeding high-
quality clones and reducing growth-stresses in the plantationwood by timely and
appropriate silvicultural prescriptions. The mills too are responding by using NDE
and automated technology to monitor manufacturing processes and to inspect the
final products in an effort to increase mill production rates and maintain profit
margins.
While there are many challenges to overcome, the industry is adapting to the
changes and developing new processes for value added products that best use the
renewable, forest resource. The discussions covered in this chapter are based on the
best, currently available information, as well as industry experiences. Technology
and knowledge continue to advance while the markets remain in a constant state of
change making the future industry concerns different than can be anticipated today.
References
Akhtar M, Attridge MC, Myers GC, Blanchette RA (1993) Biomechanical pulping of loblolly pine
chips with selected white-rot fungi. Holzforschung 47:36–40
Brändström J, Daniel G, NilssonT (2002) Use of soft rot cavities to determine microfibrilangles in
wood; advantages, disadvantages and possibilities. Holzforschung 56:468–472.
Floyd S (2005) Effect of hemicellulose on longitudinal shrinkage in wood. In Entwistle K, Walker
JCF (eds) The hemicellulose workshop 2005. Wood Technology Research Center, University
of Cantebury, Christchurch, pp 115–120
Howard JL, Westby R (2007) U.S. Forest products annual market review and prospects, 2004–
2008. Forest Products Laboratory research note FPL-RN-0305.
Malan FS (1995) Eucalyptus improvement for lumber production. Sem. intern. de util. da mad.
Para serraria. Sao Paulo. Brasil, pp 1–19.
Parrish RB (1991) The changing resources, new technology, and market opportunities for wood
products. In: Proceedings of engineered wood products, processing and design, Atlanta, GA,
pp. 1–8.
Pong WY, Ward J (1979) The wetwood problem in the forest products industry. Proceedings of the
30th western dry kiln clubs meeting. Corvallis, OR: 115–124
Schulgasser K, Witztum A (2007) The mechanism of spiral grain formation in trees. Wood Sci
Technol 41:133–156
Wang JZ, DeGroot R (1996). Treatability and durability of heartwood. In: Ritter MA, Duwadi SR,
Lee PDH (eds) National Conference on Wood Transportation Structures: 252–260
Yamamoto H (1998) Generation mechanism of growth stresses in wood cell walls: roles of lignin
and cellulose microfibril during cell wall maturation. Wood Sci Tech 32:171–182
Chapter 12
Internal Checking During Eucalypt Processing
Philip Blakemore
Contents
12.1 Introduction
Collapse is a severe type of shrinkage that occurs to some extent in most species
and timber types, but is particularly prevalent amongst certain species. The low
to medium density Eucalypt species are particularly prone (Chafe et al. 1992).
Collapse is “abnormal” in that it occurs in saturated timber above the Fibre
Saturation Point (FSP) when the cell lumen is still saturated with liquid water,
whereas normal shrinkage occurs below the FSP where moisture is lost from the cell
walls, and the cellulose microfibrils in the walls essentially move closer together. It
is collapse that causes much of the surface and internal checking problems when
drying timber from many eucalypt species.
P. Blakemore (B)
Department of Materials Science and Engineering, CSIRO, Clayton South, VIC, Australia
e-mail: philip.blackmore@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 237
DOI 10.1007/978-90-481-9550-3_12, C Springer Science+Business Media B.V. 2011
238 P. Blakemore
Where
P = total liquid tension (N m−2 or kg s−2 m−1 )
= surface tension (N m−1 or kg s−2 )
r1 , r2 = principal radii of the curved surface (m)
The second equation is Kelvin’s equation, which relates the total liquid tension
to the relative vapour pressure above the meniscus:
ρRT po
P= loge (12.2)
M p
Where
P = total liquid tension (N m−2 )
ρ = density (kg m−3 )
R = gas constant (J mol−1 K−1 )
T = absolute temperature (K)
M = molecular weight (kg mol−1 )
p = vapour pressure (Pa)
po = pressure of saturated vapour (Pa)
12 Internal Checking During Eucalypt Processing 239
Table 12.1 Calculated liquid-tension forces and associated vapour pressures for various capillary
radii at 20◦ C
15 9.70667 93.1
60 2.42667 98.2
150 0.97067 99.3
1500 0.09707 99.9
In Eq. (12.1), the principal radii (r1 and r2 ) are equivalent to the radii of the
largest interstitial openings or capillaries that are present in the pit membranes.
Thus, the smaller those capillaries are, the larger the tensions that develop in
the water in the saturated cell lumens. It also follows that in general the smaller
those capillaries are, the lower the permeability will be for the wood. Hence, low
permeability and inherently slow drying are intrinsic characteristics of collapse
prone wood.
The second equation relates the liquid tension to the temperature and relative
humidity of the air above the meniscus. Both equations together provide a means
of estimating the liquid tension present in a cell lumen for a given capillary size
and temperature. Examples of the magnitude of liquid-tension forces and relative
vapour pressures that develop for various capillary radii at 20◦ C are shown in
Table 12.1 Kauman (1964) suggested that in the cells that collapse, the interstitial
openings in the fibre to fibre or fibre to ray parenchyma pits would be in the range of
60–100 nm. Thus, the order of magnitude of the liquid-tension forces that are likely
to be present in collapsing cells is in the range of 1.456–2.427 MPa. However, Chafe
et al. (1992) suggests that capillaries with a maximum radius of 15 nm are known
to be present in saturated cell walls of some ash species, and therefore tensions of
as high as ∼9.7 MPa could be possible.
Apart from the size of the interstitial capillaries, the presence of extraneous mate-
rials, such as cytoplasmic debris and extractives, is also considered important for
permeability. Their presence is considered to reduce the size of the capillaries that
are available, therefore it can significantly increase the risk of collapse occurring.
Hillis (1984) notes that eucalypts in general have relatively high levels of extrac-
tives compared with other genera. The extractive levels could also be considerably
variable even within the one species. This is another contributing factor as to why
many eucalypts are both slow drying and collapse prone.
Kauman (1964) also clearly distinguishes the role that compressive drying
stresses play in collapse severity. Severe drying schedules that increase drying
stresses are likely to increase the amount of collapse that occurs. Given that col-
lapse itself contributes significantly to drying stresses, this in a sense provides a
type of positive feedback into the amount of collapse that occurs. The affect of
drying stresses though, is nevertheless still considered to be a secondary effect, in
that it is highly unlikely that compressive drying stresses can cause collapse in and
of themselves. Their role is only in extenuating the degree of collapse that occurs
because of the liquid tension forces that are present.
240 P. Blakemore
The liquid tension theory for collapse, as discussed above, has two main assump-
tions. The first is that collapse can only occur in cells that are fully saturated, or at
least that the radii of any bubbles that are present are less than the capillary radii
being considered in Eq. (12.1). The liquid tension forces will principally act on the
meniscus of largest radii, hence, if a bubble with a greater radii were present than
the radii of the largest capillary, the tension forces will expand it rapidly to relieve
the stresses without collapse occurring. The second assumption is that the sap has
sufficient cohesive strength to transmit the liquid tension. Again, if this were not the
case, cavitation would occur before cell collapse occurred.
Apart from the magnitude of the liquid tensions developed, the other main deter-
minate of whether or not collapse will occur is the strength of the cell walls.
Unfortunately, the mechanical properties of individual cell walls, or more impor-
tantly, the mechanical properties of the various secondary layers that make up the
cell wall, are almost impossible to measure directly. Hence, as a first approxima-
tion, the average compressive strength in the perpendicular cross-section of small
test samples is often used as an approximate value for cell wall strength. For exam-
ple, Chafe et al. (1992) quote a transverse compressive strength of 4.0 MPa for
Eucalyptus regnans (Bolza and Koot 1963) such that collapse was likely to occur if
the liquid tension force was greater than this value. With reference to Table 12.1
above, collapse would therefore occur in this species providing that the biggest
capillaries present had a radius of about 36.5 nm or less.
To more accurately investigate the strength properties of the cell wall for predict-
ing the onset of collapse, Innes (1995b) attempted a mathematical model that was
primarily intended to look at the effect of temperature on cell wall strength. The
cell wall in his model was broken down into the three secondary cell wall layers to
account for the fact that the alignment of the cellulose microfibrils within those lay-
ers strongly influences the anisotropic mechanical properties within in each layer.
Most importantly, the S1 and S3 layers, with large microfibril angles, essentially
provide a form of circumferential stiffening to resist collapse. Using a plain-strain
assumption, his model was essentially that of a three- layered thick-walled cylinder.
Innes (op.cit.) developed an analytical solution for this model based on the average
anatomical dimensions of the Tasmania oak group of ‘Ash’ species (E. delegatensis,
E. rengns, E. obliqua). Again though, given the difficulty of measuring cell wall
mechanical properties, all of the cell wall properties were essentially based on edu-
cated guesses, using measurements taken on small wooden test samples as a guide.
The Innes model predicted that the onset of collapse at 25◦ C occurred when the
liquid tension reached approximately 5.33 MPa.
Blakemore (2008) used finite element modelling software to generate numerical
solutions to a similar model to the Innes (1995b) one. This was done to start to over-
come some of limitations with the Innes (1995b) model. The first such limitation
was that in reality cells are not isolated, and it is likely that the behaviour of double
cell wall layers (three layers of secondary cell wall on either side of the compound
middle lamella) is different than that of a single wall layer. The Blakemore model
also allowed the use of a more rectangular cell shape, which is also possibly more
realistic. More details of the Blakemore’s model are provided in Blakemore (2008).
12 Internal Checking During Eucalypt Processing 241
The inherent geometric stability of this multi-cell model meant that a shear dis-
placement was also introduced to ensure that the deformed shape for the rectangular
shaped cells was closer to what is observed in real wood samples. The shear dis-
placement was partly justified on the bases that it may reflect the real effect of drying
stresses on the deformation that occurs during collapse. In the Blakemore model, an
internal liquid tension of about 4.9 MPa was required to initiate collapse at 25◦ C.
Unfortunately, many of the serious simplifications in the Innes (1995b) model still
remain in the Blakemore (2008) model; and hence the model’s utility is still very
limited. For example, it is likely that the predicted internal liquid tensions that lead
to collapse are still relatively high simply due to the inherent stability of the regular
geometric shapes used. It seems, likely that irregularities in real cell walls could
lead to weak points where collapse would occur at much lower tensions than those
predicted in the models.
One feature the Innes (1995b) and Blakemore (2008) models help to demonstrate
clearly is the importance of temperature as a factor in determining the severity of
collapse that occurs. The effect of temperature has been well established over a
period of time (Greenhill and Dadswell 1940; Ellwood 1952; Kauman 1960; Innes
1996a, b). Mostly this effect is due to a softening of the cell walls at higher tem-
peratures as strength is reduced and creep increases. Additionally, if the increased
temperatures also contribute to increased drying stresses, this may also increase the
collapse severity.
Innes (1996a, b) suggested that there may even be a temperature threshold
below which collapse does not occur. However, Ilic (1999), while acknowledging
the importance of temperature on collapse, doubted the validity of a threshold in
the temperature ranges suggested by Innes (1996a, b). Ilic (1999) noted that some
collapse had still been observed in samples stored in refrigerators at about 4◦ C.
they can be due to collapse shrinkage, which occurs in the early stages of drying
when the cell lumens are still saturated. Both normal and collapse shrinkage are
strongly anisotropic. The ratio of he tangential to radial directions is typically about
∼2 (normally within a range of 1.5–3.0) (Kauman 1964). There are many factors
that contribute to transverse anisotropy, and the contribution of each varies between
and within different species. The main contributing factors include: ray restraint,
earlywood/latewood interaction, and the differences in the radial and tangential cell
walls (stiffness, chemical components or layering, microfibril angle and pit arrange-
ments and microfibril aggregations). Major reviews of shrinkage anisotropy, and
of these contributing factors, have been undertaken by Pentoney (1953); Kelsey
(1963); Stamm (1964); Kollmann and Côté (1968); Boyd (1974); Skaar (1988) and
more recently Booker 2003).
The main source of spatial variation in shrinkage is due to differences between
the earlywood and latewood bands of the growth rings. In particular, collapse shrink-
age can vary very significantly between earlywood and latewood. The thinner,
weaker walls in the earlywood mean that collapse severity is often much greater
in earlywood. It is this variation in collapse intensity that leads to the commonly
observed ‘washboarding’ effect (Fig.12.1), which is seen on the wide faces of quar-
tersawn boards. If the latewood band above and below a band of collapse prone
earlywood are significantly denser and stiffer, this can lead to severe stress gra-
dients within the earlywood. Indeed, Ilic (1999) found that internal checking was
particularly likely in boards of E. regnans, where the earlywood air-dried density
was below about 450 kg m−3 and the latewood above 600 kg m−3 . It is the combi-
nation of the shrinkage anisotropy and the large differences in the properties of the
earlywood and latewood that means that most collapse related checks are initiated in
the earlywood, and are radially aligned and elliptical in shape. The presence of rays
is probably also contributed to the radial alignment as they can act as lines of weak-
ness for check initiation. As wood properties are often quite uniform within arcs
of a growth ring, it is not uncommon to see collapse related internal checks evenly
spaced within a single earlywood ring (Fig.12.2). Surface checking that is caused
by collapse, similarly tends to mostly occur in bands of earlywood that intercept the
surface. This is particularly noticeable in backsawn boards (Fig.12.3). In most euca-
lypts then, when only intra-ring checks are found in the earlywood, those checks are
almost always collapse induced checks.
Fig. 12.1 Example of collapse washbording and internal checking in a board of quartersawn
E. nitens (75 × 25 mm)
12 Internal Checking During Eucalypt Processing 243
Fig. 12.2 Examples of collapse related internal checking in earlywood rings in Victorian Ash
(E. regnans or E. delegatensis)
The temporal variation in shrinkage is due to the drying moisture content gradi-
ents that develop as drying progresses. There is generally good data in the literature
on the development of moisture content and drying stresses based mostly on normal
shrinkage (e.g. See Fig. 12.4: US Forest Products Laboratory Forest Service 1999).
Unfortunately, there is little good data in the literature on the temporal occurrence
of collapse and intra-ring internal checking as drying progresses. Plots of overall
dimensional shrinkage (Fig. 12.4 – this data is from material used by Blakemore
and Langrish 2008b) suggest that while not as linear as normal shrinkage, it progres-
sion is still reasonably constant. This suggests that collapse shrinkage may progress
from the surface to the core of the boards. However, the exact timing and location
of internal checking is not as clear. For instance, Blakemore and Langrish (2008b),
found that ramping of the pre-drying schedules below 50% moisture content had
minimal effects on the levels of internal checking present. This might suggest that
most of the internal checking had already occurred by the time the mean moisture
content of the board was below about 50%, and yet measurable collapse was still
being observed.
The exact interaction of collapse related drying stresses with normal shrinkage
related drying stresses is also not clear. It is normally assumed that most of con-
ventional drying stresses are due to normal shrinkage as the moisture content drops
Fig. 12.3 Grouping of collapse related surface checks in the earlywood zones of a 200 mm
wide backsawn board of plantation grown E. nitens. The edges are closer to quartersawn grain
orientation, and hence no surface checks are present on this face
244 P. Blakemore
20
16
14
12
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
MC (%)
Fig. 12.4 Area (radial × tangential) shrinkage against moisture content for a number of 100 ×
50 mm boards of Victorian Ash (E. regnans and E. delegatensis) with a collapse ranging from
minimal to severe
below the fibre saturation point. The conventional development of normal drying
stresses are shown to the left of Fig. 12.5. However, if collapse is occurring well
in advance of the movement of the fibre saturation point, it may to some extent be
reducing the severity of normal shrinkage related drying gradient. Conversely, col-
lapse that has occurred at the surface may increase the severity of the collapse that
is occurring later, below it deeper into the board.
The most obvious visual indication of severe drying stresses is when the intra-
ring collapse related internal checks are extended into inter-ring checks. In severe
cases, this is classified as honeycomb checking (Fig. 12.5). The interaction of drying
stresses and board geometry is responsible for the sunken faces that are often seen on
the wide face of boards of severely collapsed boards (e.g. Fig. 12.1). Kauman (1964)
provides a clear explanation for these phenomena. Essentially, as the outer surface
layers (case) dry out, a tension set is induced in these layers (Fig. 12.6). As drying
continues, after stress reversal has occurred, these dryer and much stiffer layers
provide a high resistance to the shrinkage that is now occurring in the centre of
the board. In the width direction, these surface layers or columns are close together
and hence the edges are kept straight. However, in the thickness direction, the edge
columns are much further apart and hence the surface columns on the wide faces
act like uniformly loaded beams supported only at the ends. These drying stresses
can also interact with board dimensions to affect the severity of internal checks.
For example, Chafe and Carr (1998a), found severe inter-ring internal checking in
boards that were 100 × 50 mm dried at 30◦ C and an initial relative humidity of 90%,
while in matching boards of either 50 × 50 mm or 100 × 25 mm, dried under the
12 Internal Checking During Eucalypt Processing 245
Fig. 12.5 Diagram of normal drying stress development and checking development in backsawn
eucalypts
same conditions, only intra-ring checking was found. However, no such inter-ring
checking was observed in 100 × 40 mm dimension boards dried by Blakemore and
Langrish (2007) also dried at 30◦ C and 95% RH. And no such inter-ring checks were
observed in trials with 105 × 50 mm specimens initially dried at 20◦ C (Blakemore
and Langrish 2008b). Ramped drying schedules were also tried in the Blakemore
and Langrish (2008b) trial, and again, no inter-ring checking was caused when,
below a mean moisture content of 50%, the temperatures were progressively raised
to 55◦ C. This is much higher than the maximum temperature of 30◦ C used by Chafe
and Carr (1998a) in pre-drying. So it is possible that the inter-ring checking may
have been caused by significantly wet cores that were still present when oven-drying
was undertaken when the samples were nominally at 12% MC.
To some extent, collapse can occur to a small extent in almost all timber species.
However, Kauman (1960) suggests that “severe collapse is largely confined to
medium density angiosperms” and further identifies the “Ash” group of eucalypt
species as amongst the “best known collapsing species”. In this grouping of species,
the most economically important are E. regnans, E. delegatensis and E. obliqua.
Currently, there are two cool to mild temperate species, E. globulus and E. nitens,
that are being grown extensively in plantations in southern Australia and around the
world, predominately as a source of pulp fibre. While both are collapse prone, the
severity of the problem is considerable worse in E. nitens.
Unfortunately, there is no commonly accepted method of measuring collapse
that would provide a quantitative basis for comparing eucalypt species. There are
too types of collapse measurements that are made for experimental purposes. The
first attempts to specifically quantify the amount of collapse that has occurred and
provides the more detailed measurements. When measuring the total shrinkage of a
wood sample, the measurement has a component of normal shrinkage and collapse
shrinkage in it. To distinguish between the two components of shrinkage, measure-
ments are also madded on thin sections (<1 mm along the grain) cut from either
end of the measurement blocks. These thin sections are used to provide a measure
of collapse free shrinkage. As the fibres in most eucalypts are longer than 1 mm,
there should be almost no intact fibres present in such a thin sample. If there are
no intact fibres, there can be no saturated lumens and therefore collapse can not
occur. Essentially then, such a sample should provide a measure of normal shrink-
age only. Normal shrinkage is what occurs as bound water is removed form the
cell walls. Hence, by subtracting this measure of normal shrinkage from the total
shrinkage that occurs in a matching longer sample, a measure of collapse shrinkage
is obtained.
The main confounding factor in this measure of collapse is that the drying
stresses will variably interact and restrain the amount of normal shrinkage that
occurs. As most of the drying that occurs in the thin slices is likely to be end dry-
ing, this means that almost no transverse drying stresses should be present, and
hence the expression of normal shrinkage is likely to be slightly less in the larger
section than in the matching thinner section. This can sometimes lead to spuri-
ous positive collapse recovery values in low collapse specimens (e.g. Blakemore
and Langrish 2007). Nevertheless, this general technique has been used repeat-
edly by many researchers (e.g. Greenhill 1938; Kauman 1960; Ilic and Hillis
1986; Chafe 1987; Chafe 1990; Chafe and Ilic 1992a, 1992b, 1992c; Chafe 1993;
12 Internal Checking During Eucalypt Processing 247
Chafe 1994; Innes 1995a; Chafe and Carr 1998a, 1998b; Ilic, 1999; Yang et al.
2002, 2003). The next most commonly used technique for measuring collapse is
to more simply just measure recoverable, or recovered collapse. This typically
involves measuring the total shrinkage before and after reconditioning (e.g. Chafe
1985, 1986) at a standardised moisture content. In this way, the total amount
of collapse is not measured, but assuming that the reconditioning is undertaken
appropriately, most of it should be recovered. While less accurate than the pre-
vious method, the technique is considerably easier to implement. Data of this
type is more generally available as it can be calculated from the data collected
by Kingston and Risdon (1961) for Australian timber species. The reason it is
less accurate, is that there are a number of variables that can affect collapse
recovery, the most important being moisture content (Blakemore and Langrish
2007, 2008b). In that research, Blakemore and Langrish (2007, 2008b) recon-
firmed that in two severely collapse prone eucalypts (Eucalyptus regnans and
E. delegatensis) the recovery of collapse decreased significantly below a moisture
content of about 15%. The fact that by methodology, all of the Kingston and Risdon
(1961) samples are reconditioned at 12% means that the recovery of collapse is
likely to be sub-optimal. Although, there is a chance that the standardised 2 h of
steaming with such small samples may have been enough to raise the moisture con-
tent back up above 15%. Figure 12.7 shows the recovered collapse values for all
of the Eucalypt species in Kingston and Risdon (1961). The species are arranged
in order of increasing basic density. This figure confirms the extent of the collapse
severity amongst the Ash group of eucalypts, such as E. regnans and E. delegatensis.
The weak to moderate general negative relationship between density and collapse is
1000 20
900 18
800 16
Recovered Collapse (%)
Basic Density (kg m–3)
700 14
600 12
500 10
400 8
300 6
200 4
100 2
0 0
E. regnans (35 y.o.)
E. obliqua
E. lindleyana
E. mannifera
E. haemastoma
E. papuana
E. campanulata
E. bridgesiana
E. cambageana
E. dalrympleana
E. rubida
E. rubida
E. radiata
E. fastigata
E. stuartiana
E. australiana
E. ovata
E. ovata
E. obliqua
E. globoidea
E. saligna
E. marginata
E. calophylla
E. cypellocarpa
E. wilkinsoniana
E. muelleriana
E. macrorrhyncha
E. consideniana
E. robusta
E. carnea
E. maculata
E. resinifera
E. cloeziana
E. paniculata
E. bosistoana
E. longifolia
E. gomphocephala
E. tetrodonta
E. molucanna
E. melliodora
E. spenceriana
E. leucoxylon
E. sideroxylon
E. delegatensis
E. polyanthemos
E. decorticans
E. regnans
E. pilularis
E. nitens
E. grandis
E. globulus
E. globulus
E. dives
E. globulus
E. viminalis
E. viminalis
E. grandis
E. dives
E. pilularis
E. botryoides
E. camaldulensis
E. astringens
E. tereticornis
E. microcorys
E. largiflorens
E. albens
E. tessellaris
E. wandoo
E. robertsonii
E. jacksoni
E. baxteri
E. dunnii
E. rossii
E. sieberi
E. stjohnii
E. maidenii
E. cladocalyx
E. goniocalyx
E. diversicolor
Species
Fig. 12.7 Recovered collapse as derived from Kingston and Risdon (1961) shrinkage data for a
range of Eucalypt species. Recovered collapse is calculated as the after reconditioning shrinkage
at 12% subtracted from the before reconditioning shrinkage value at 12%. The error bars are 95%
confidence intervals for the mean values. Where no error bars are shown this is because the sample
size was too small to calculate a standard error
248 P. Blakemore
in keeping with that which has been observed by a number of authors (Chafe 1985,
1986; Ilic and Hillis 1986; Chafe 1994; Ilic 1995; and Blakemore and Langrish
2008a, b).
As early as 1917, James Grant and George Grant (father and son) cited by (Greenhill
1938), independently discovered that steam could be used to recover collapse in
Australian eucalypts. Essentially, steam reconditioning involves placing the timber
in a well insulated chamber and filling the chamber with saturated steam for a num-
ber of hours, depending on the species and thickness of the boards. Greenhill (1938,
1940) carried out a range of reconditioning experiments on samples of Eucalyptus
regnans and E. delegatensis. He clearly demonstrated the importance of heat in col-
lapse recovery and the need to get the temperature in the steaming chamber as close
as possible to 100◦ C (Fig. 12.8).
A number of industrial rules of thumb have been in use in Australia for many
decades, based mostly on this research by Greenhill (1938 and 1940) and some
work by Mackay (1972). Unfortunately, the small size of the specimens used in
this research, and the fact that they all the specimens came from one or two
boards, means that robustness of all of these rules has been somewhat question-
able. Blakemore and Langrish (2007, 2008b), conducted a range of reconditioning
experiments on larger dimension boards (100 × 40 and 100 × 50 mm) from a wider
range of trees. The results of this confirmed that below a mean moisture content of
Fig. 12.8 The effect of reconditioning temperature and drying conditions on collapse recovery
(Greenhill, 1938)
12 Internal Checking During Eucalypt Processing 249
BEFORE RECONDITIONING
20% 15%
32.8% 16.9%
Fig. 12.9 Cross sectional scans of internal checking for two end-matched boards dried with the
same ramped pre-drying schedule. Reconditioning was undertaken at the nominal mean moisture
content of 20%, (board on left) and 15% (board on right). Percentages shown above each board are
measured core moisture contents
250 P. Blakemore
110 1040
Surface
9 mm top
100 1030
middle
90 9 mm bottom 1020
Temperature (°C) Width Swelling (× 0.1 = %)
Chamber
80 Side 1010
Loadcell
70 Datalogger Temp 1000
Width
Mass (g)
60 990
Mass
50 980
40 970
30 960
20 950
10 940
0 930
0 1 2 3 4 5 6 7 8
Time (hrs)
Fig. 12.10 Measurements of mass, temperature (Surface, 9 mm in from top and bottom surface
and core) and width changes during steam reconditioning for sample board with a basic density ≈
490 kg m–3 (From Blakemore and Langrish 2008b)
One of the simplest methods for modelling the time to heat up the boards is based
on dimensional analysis, for which a relevant dimensionless group is the Fourier
number (Fo):
ατ kτ
Fo = 2 = (12.2)
l ρcl 2
where:
α = thermal diffusivity — k/ρc(m2 s−1 )
k = thermal conductivity (W m−1 K−1 )
τ = time (s)
c = heat capacity (J kg−1 K−1 )
l = a characteristic dimension—in this case, half the board thickness (m)
A value for heat capacity, c, was obtained from the US Forest Service Forest
Products Laboratory (1999) with the following equations.
Co = 0.1031 + 0.003867 × θ (12.3)
12 Internal Checking During Eucalypt Processing 251
where:
(Co + 0.01 · MC · Cw )
c=
(1 + 0.01 · MC)
(12.4)
+MC −0.06191 + 2.36 × 10−4 · θ − 1.3310−4 · MC
where:
Where:
ρ = density (kg m−3 )
Equation (12.4) has been used to estimate the heat-up time by assuming that a
rectangular parallelepiped with uniform material properties will reach an approxi-
mately uniform temperature distribution when Fo = 1. Figure 12.11 (page 101) from
Carslaw and Jaeger (1997) shows that for an infinitely long and wide solid with par-
allel sides, Fo = 1 accounts for about 90% of the temperature change from the initial
to final temperatures in the core of the board. Given the narrower rectangular nature
of the sub-samples in this case, with additional heat conduction from the edges and
ends, it would be expected that the percentage of the temperature change in the
core of the board would be greater than 95%. For this analysis, the assumed mate-
rial property values were: ρ = 490 kg m−3 , moisture content (MC) = 20%, θ =
333 K (60◦ C — halfway between 20 and 100◦ C), Cw = 4.19 kJ kg−1 K−1 and
l = 0.022 m.
Using these values, the characteristic heat-up time (τ ) was estimated to be
approximately 80 min. Figure 12.10 shows that the observed heat-up time for the
core of a 100 × 50 mm sample board, with mean moisture content nominally of
20%, was approximately 70 min. The simplicity of this analysis makes it easy to
estimate heat-up times, and the estimated error of ±20% is tolerable for this pur-
pose, especially given the conservativeness of current industrial practice where it is
common to steam for up to 6–8 h depending on the thickness of boards. The main
reason for the long steaming is to allow for moisture uptake by boards that may be
over-dry, but the effectiveness of this approach has never been validated.
252 P. Blakemore
12.8 Summary
All forms of checking in timber, both surface and internal, are due to differential
shrinkages that result in internal drying stresses. In the case of mid to low den-
sity eucalypts, the occurrence of collapse shrinkage severely increases these drying
stresses and the likelihood of checking. This is true at both the board scale and the
growth ring scale where differences in collapse shrinkage and stiffness between the
zones of earlywood and latewood in a growth ring, can lead to regular intra-ring
checking in the earlywood. This form of checking can be one of the more obvious
indications of collapse related checking.
Reconditioning the timber in saturated steam is a long established method for
recovering collapse shrinkage. It is important that this be undertaken with the board
at the correct moisture content. A mean moisture content of about 18% is ideal,
although there is little difference in recovery in the range of 18–25%. Certainly
below about 15% moisture content collapse recovery starts to decline severely.
Perhaps, more important is the core moisture content, as it is the hardest to con-
trol. For recovery in the core of the boards, it is important that it be below about
25% moisture content. At higher moisture contents collapse recovery is incomplete,
at much higher moisture content, intra-ring checking may be caused and a darkening
of the wood color can also occur.
The reconditioning affect mostly appears to be a temperature related effect. As
such the steaming treatment only needs to be undertaken long enough to get the
core of the boards as close to the steam temperature or 100◦ C as possible. In boards
where the core of the board has been over-dried below 15% moisture content, recov-
ery can be achieved if the moisture content is raised back above this level. But
depending on the thickness of the boards the length of steaming required can be
prohibitively long (>24 h) and expensive.
Provided that the drying schedules have minimised the amount of collapse that
has occurred, and only intra-ring checking has occurred. Gradual, ramping up of
the severity of the pre-drying schedules, once the mean moisture contents are below
about 50%, has appeared to have minimal effect on the internal checking levels
12 Internal Checking During Eucalypt Processing 253
present. Even with current best pre-drying practices, much of the intra-ring checking
appears to be due to inherent properties rather than the drying conditions.
Good steaming practices will result in most of the intra-ring checks being closed.
The significance of those closed checks is unclear. An industry perception is that
upon ripping or deep moulding, such close check will always result in feathering
effect on the surface of such products. In such cases whether the original checks
were fully closed, or partially closed is unclear.
References
Blakemore PA (2008) Optimisation of steam reconditioning for regrowth-ash and plantation grown
eucalypt species. PhD Thesis, The University of Sydney. http://hdl.handle.net/2123/2343.
Accessed 3 August 2010
Blakemore PA, Langrish TAG (2007) Effect of mean moisture content on the steam reconditioning
of collapsed Eucalyptus regnans. Wood Sci Tech 41:87–98
Blakemore PA, Langrish TAG (2008a) Effect of collapse on fitted diffusion coefficients for
Victorian ash eucalypts. Wood Sci Tech 42:535–549
Blakemore PA, Langrish TAG (2008b). Effect of pre-drying schedule ramping on collapse recovery
and internal checking with Victorian Ash eucalypts. Wood Sci Tech 42:473–492
Bolza E, Kloot NH (1963) The Mechanical properties of 174 Australian timbers. CSIRO Division
of Forest Products, Melbourne, p 112
Booker R (2003) Shrinkage and theories of differential shrinkage. In: Wood research, knowledge
and concepts for future demands. EMPA Wood Lab Res Work Rep 115/50:29–46
Boyd JD (1974) Anisotropic shrinkage of wood: Identification of the dominant determinants.
Mokuzai Gakk. 20:473–482
Campbell GS, Hartley J (1984) Drying and dried wood. In: Hillis WE, Brown AG (eds) Eucalypts
for wood production. CSIRO/Academic, Sydney, pp 328–336
Carslaw HS, Jaeger JC (1997) Conduction of heat in solids. Clarendon Press, Melbourne, 510 pp
Chafe SC (1985) The distribution and interrelationship of collapse, volumetric shrinkage, moisture
content and density in trees of Eucalyptus regnans F. Muell. Wood Sci Tech 19:329–345
Chafe SC (1986) Radial variation of collapse, volumetric shrinkage, moisture content and density
in Eucalyptus regnans F. Muell. Wood Sci Tech 20:253–262
Chafe SC (1987) Collapse, volumetric shrinkage, specific gravity and extractives in Eucalyptus
and other species. Part 2: The influence of wood extractives. Wood Sci Tech 21:27–41
Chafe SC (1990) Changes in shrinkage and collapse in the wood of Eucalyptus regnans F. Muell
following extraction. Holzforschung 44(4):235–244
Chafe SC (1993) The effect of boiling on shrinkage, collapse and other wood-water properties in
core segments of Eucalyptus regnans F. Muell. Wood Sci Tech 27:205–217
Chafe SC (1994) Preheating green boards of mountain ash (Eucalyptus regnans F. Muell). I. Effects
on external shrinkage, internal checking and surface checking. Holzforschung 48:61–68
Chafe SC, Barnacle JE, Hunter AJ, Ilic J, Northway RL, Rozsa AN (1992) Collapse: an
introduction. CSIRO Division of Forest Products, Melbourne, 9 pp.
Chafe SC, Carr JM (1998a). Effect of board dimensions and grain orientation on internal checking
in Eucalyptus regnans. Holzforschung 52:430–440
Chafe SC, Carr JM (1998b) Effect of preheating on internal checking in boards of different
dimension and grain orientation in Eucalyptus regnans. Holz Roh- Werkst. 56:15–23
Chafe SC (1994) Preheating green boards of Mountain Ash (Eucalyptus regnans F. Muell) II.
Relationships amongst properties. Holzforschung 48:163–167
Chafe SC, Ilic J (1992a) Shrinkage and collapse of thin sections and blocks of Tasmanian mountain
ash regrowth. Part 1: Shrinkage, specific gravity and the fibre saturation point. Wood Sci Tech
26:115–129.
254 P. Blakemore
Chafe SC, Ilic J (1992b) Shrinkage and collapse of thin sections and blocks of Tasmanian mountain
ash regrowth. Part2: The R-ratio and changes in cell lumen volume. Wood Sci Tech 26:181–187
Chafe SC, Ilic J (1992c) Shrinkage and collapse of thin sections and blocks of Tasmanian mountain
ash regrowth. Part3: Collapse. Wood Sci Tech 26:343–351
Ellwood EL (1952) The seasoning of rotary peeled veneer from Eucalyptus regnans F.v.M. Aus J
App Sci 1:53–70
Greenhill WL (1938) Collapse and its removal: Some recent investigations with Eucalyptus
regnans. CSIRO Division of Forest Products, Melbourne, 32 pp.
Greenhill WL (1940) Collapse and its removal. Aus Timb J 6:160–161, 171, 228–229, 239, 241
Greenhill WL, Dadswell HE (1940) The density of Australian Timbers. 2.-Air-dry and basic
density data for 172 timbers. CSIRO Division of Forest Products, Melbourne, 75 pp
Hillis WE (1984) Wood quality and utilization. In: Hillis WE., Brown AG (eds) Eucalypts for wood
production. CSIRO, Melbourne, 259–289
Ilic J (1995) Advantages of prefreezing for reducing shrinkage-related degrade in Eucalypts –
General considerations and review of the literature. Wood Sci Tech 29:277–285
Ilic J (1999) Shrinkage-related degrade and its association with some physical properties in
Eucalyptus regnans F. Muell. Wood Sci Tech 33:425–437
Ilic J, Hillis WE (1986) Prediction of collapse in dried eucalypt wood. Holzforschung 40:109–112
Innes TC (1995a) Collapse free pre-drying of Eucalyptus regnans F. Muell. Holz Roh Werkst
53:403–406
Innes TC (1995b) Stress model of a wood fibre in relation to collapse. Wood Sci Tech 29:363–376
Innes TC (1996a) Collapse and internal checking in the latewood of Eucalyptus regnans F.Muell.
Wood Sci Tech 30:373–383
Innes TC (1996b) Pre-drying of collapse prone wood free of surface and internal checking. Holz
Roh Werkst. 54:195–199
Kauman WG (1960) Contribution to the theory of cell collapse in wood: Investigations with
Eucalyptus regnans. Aus J App Sci 11(1):122–145
Kauman WG (1964) Cell collapse in wood. CSIRO Division of Forest Products, Melbourne
Kelsey KE (1963) A critical review of the relationship between the shrinkage and structure of
wood. CSIRO Division of Forest Products, Melbourne
Kingston RST, Risdon CJE (1961) Shrinkage and density of Australian and other South-west
Pacific woods. Division of Forest Products, Melbourne
Kollmann FFP, Côté WA Jr (1968) Principles of wood science and technology I. Solid Wood.
Springer, Berlin
Mackay JFG (1972) Recovery and collapse in E. delegatensis by use of anhydrous ammonia and
steam. Wood Fiber 4(3):126–129
Oliver AR (1991) A model of the behaviour of wood as it dries (with special reference to
Eucalypt materials). Research Report CM91-1, Civil and Mechanical Engineering Department,
University of Tasmania
Pentoney RE (1953) Mechanisms affecting tangential vs. radial shrinkage. J For Prod Res Soc
3(2):27–32
Skaar C (1988) Wood-water relations. Springer, Berlin
Stamm AJ (1964) Wood and cellulose science. The Ronald Press Comp., New York, NY
Tiemann HD (1915) The effect of different methods of drying on the strength of wood. Lumber
World rev 28(7):19–20
US Forest Service Forest Products Laboratory (1999) Wood handbook: Wood as an engineer-
ing material. US Department of Agriculture, www.fpl.fs.fed.us/documnts/fplgtr/fplgtr113/
fplgtr113.htm. Accessed 3 August 2010
Yang JL, Ilic J, Evans R, Fife D (2003) Interrelationships between shrinkage properties, microfib-
ril angle, and cellulose crystallite width in 10-year-old Eucalyptus globulus. NZ J For Sci
33(1):47–61
Yang JL, Fife D, Ilic J, Blackwell P (2002) Between-site and between-provenance differences in
shrinkage properties of 10-year-old Eucalyptus globulus Labill. Aus For 65:220–226
Chapter 13
Acoustic Tomography for Tension Wood
Detection in Eucalypts
Voichita Bucur
Contents
13.1 Introduction
The last two decades of the 20th century have been characterized by an extraor-
dinary development of different methodologies for nondestructive evaluation of
wood products, which has succeeded in the imaging of the internal structure of
wood at both macroscopic and microscopic levels. The efforts devoted towards
the development of nondestructive technologies for the evaluation of wood prop-
erties of trees, stems, logs and lumber are summarized in reference books (Pellerin
and Ross 2002: Bucur 2003) and in the proceedings of the international symposia
on non-destructive testing of wood. At the present time the predominant attention
of the forest industry is oriented through imaging techniques development of the
internal structure of wood or wood products, based on the propagation of different
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 255
DOI 10.1007/978-90-481-9550-3_13, C Springer Science+Business Media B.V. 2011
256 V. Bucur
Fig. 13.1 Radial delamination in S2 layer which do not extend to the cell lumen, in Eucalyptus
spp.. Transverse section magnification 10,500 x (Chafe 1977, Figure 4). Top left insert –
Magnification 29,000 x – lumen boundary: S3 absent, steeply inclined microfibrils. Bottom insert –
Magnification 29,000 x – microfibrils misalignment and radial dislocations L = lumen, G =
gelatinous layer, Arrows = innermost gelatinous layer
13.2.1 Materials
Six disks with varying degree of pith eccentricity has been selected, namely five
disks were cut from Eucaluptus delegatensis logs and one disc from a plantation
grown Eucaluptus nitens log. The wider side of the eccentric discs were anticipated
to contain various portions of tension wood. The discs were in green conditions.
Note that the discs are representative of the transversal section of trees and logs.
A combined method using ultrasonic waves was developed to detect and locate
the tension wood. The direct transmission technique was developed to detect the
tension wood along the axis L. The stress wave technique was used to locate the
258 V. Bucur
10
10B
9 11
8 11B
7 12
12B
1
6
5 2
Fig. 13.2 Time of flight measurements on disc, in longitudinal anisotropic direction of wood, on
different rays and locations (Yang et al. 2007b, Figure 1)
corresponding tension wood zones, with a 2D imaging technique for the inspection
of the transversal section of the disc (plane RT).
Ultrasonic velocity method allows the measurement of the time of flight in longi-
tudinal anisotropic direction of wood (L), on different rays and locations, as shown
in Fig. 13.2. The corresponding velocity was noted (VLL ). Broad band 1 MHz trans-
ducers were used for these measurements, in direct transmission technique, using
conventional Panametrics equipment (Fig. 13.3).
Acoustic tomography was performed by measuring the stress wave velocity in
radial and tangential anisotropic directions, with the FAKOPP multi-channel timer
using a linear filtered back projection technique for image reconstruction (Divos and
13 Acoustic Tomography for Tension Wood Detection in Eucalypts 259
Fig. 13.3 Conventional ultrasonic equipment for measurements on discs (Yang et al. 2007b,
Figure 1)
Fig. 13.4 Stress wave signal on larch in time domain (Divos and Divos 2005, Figure 1). Axis X =
time scale. The grid size on the time scale is 50 μs. Axis Y = amplitude scale in relative units.
Note the 40 kHz signal superimposed on the 2 kHz signal
Divos 2005). A hammer blow was used to produce stress waves in a wide frequency
range. As can be seen from the Fig. 13.4 a 40 kHz signal is superimposed on the
2 kHz signal. The device allows a strong amplification of the signal and the fre-
quency of 40 kHz was used for the wavelength calculation, which is 25 mm. This
means that the minimum detectable defect size is 25 mm, which is the effective
resolution of this device. The resolution of the stress wave acoustic tomography is
influenced by the frequency, the number of probes and the inversion technique for
image reconstruction.
Acoustic waves are emitted sequentially from the source probe position and
recorded at receiver source position. For radial inspection with bulk waves, the trans-
ducers were inserted in the radial direction to the disc periphery (Fig. 13.5a), while
260 V. Bucur
Fig. 13.5 Transducers positioning for acoustic tomography. (a) inspection in R direction and
(b) inspection in T direction (Yang et al. 2007b, Figure 2)
for tangential inspection, with surface waves, the transducers were inserted into the
disc in T direction (Fig. 13.5b).
Acoustical imaging with the stress wave technique was performed first on discs
having a big eccentricity, corresponding to an important proportion of tension wood
and secondly on concentric discs with normal wood. The tomographs were obtained
with 8 and 16 probes. The distance for the tomographic maps was calculated assum-
ing a straight path between the two transducers. The tomographic resolution is
related to the pixel size. The number of cells cannot exceed the number of mea-
surements (for 16 transducers, 120 independent acquisitions). The time uncertainty
is 1 μm and the calculated limit of the spatial resolution is 25 mm.
5000
Ray 1 Ray 4
4000
3500
3000
2500
0 50 100 150 200
Distance from cambium (mm)
Fig. 13.6 Variation of ultrasonic velocity in L direction (bulk waves) on a disc following four rays.
(a) the rays on the disc, (b) velocity VLL on four rays (Yang et al. 2007b, Figure 3)
Fig. 13.7 The velocity map with radial stress wave velocity on the eccentric and concentric discs
with 8 probes (Yang et al. 2007a, Figure 2)
For eccentric discs, the tangential stress wave velocity maps (Fig. 13.8) were
obtained at different distances from the discs periphery (ex.: 20 mm, 140 mm). The
range of surface tangential velocities was between 1450 m/s (green) and 395 m/s
(red). Surface tangential stress wave velocities were lower in tension wood zones
than in the opposite wood, the lateral wood and the juvenile wood. As have seen
in radial maps, low velocity values were observed on the wider side of the eccen-
tric disc (tension wood zones). Smaller inspected surfaces gave better resolution of
wood structure, as for example the maps obtained for 140 mm from the periphery
compared to those at 20 mm from the periphery. The tangential stress wave velocity
13 Acoustic Tomography for Tension Wood Detection in Eucalypts 263
Fig. 13.8 The velocity map with surface tangential stress wave velocity on an eccentric disc at
difference distances from the discs periphery (ex.: 20 mm, 140 mm) (Yang et al. 2007b, Figure 4)
maps for the concentric discs (Fig. 13.9) are overall uniformly colored, except for a
path of much lower velocity that corresponds to the pith area and juvenile wood.
It is obvious that acoustic tomography allows the location of pith’s area on logs
and trees. Improvement of velocity map resolution could be achieved by using more
transducers and/or increasing frequency, so that the variation in wood properties can
be more precisely estimated.
264 V. Bucur
Fig. 13.9 The velocity map with surface tangential stress wave velocity on a concentric disc (Yang
et al. 2007b, Figure 5)
13 Acoustic Tomography for Tension Wood Detection in Eucalypts 265
1 2
Fig. 13.10 Comparison between the velocity map resolution with radial stress waves, with 8
transducers (left) and with 16 transducers (right) on two disks (Yang et al. 2007b, Figure 5)
Fig. 13.11 Image of the transversal section of the eccentric disk with tangential stress waves and
with 8 (left) or 16 (right) transducers located at different distances from the periphery (Yang et al.
2007b, Figure 6)
266 V. Bucur
13.5 Summary
The detection and the location of reaction wood (compression or tension wood) in
trees and logs is a major interest for wood industry. In tension wood which has an
important thick gelatinous layer (or G – layer) in the cell wall, radial delaminations
have been observed in Eucalyptus spp. by Chafe (1977) as local disruptions to the
microfibril orientation. On the other hand, the tension wood has lower properties
than normal wood because of morphological difference between the lignified sec-
ondary wall S2 and the unlignified G –layer (Donaldson 2001). These structural
particularities as well as the disruption of microfibril orientation in tension wood
are evident with ultrasonic techniques such as acoustic tomography and ultrasonic
velocity method. The detection of tension wood in L direction was performed using
a direct transmission ultrasonic technique and measuring the time of flight with
1 MHz probes and calculating the corresponding velocity noted VLL . The corre-
sponding values are higher in tension wood (average 3847 m/s), than in opposite
wood (average 3187 m/s), or lateral wood considered as normal wood (average
3544 m/s). The location of tension wood was possible using a stress wave method
(frequency 40 kHz) and a linear filtered back projection technique for image recon-
struction. A hammer blow was used to produce stress waves and the corresponding
signal was strongly amplified. Acoustic waves are emitted sequentially from the
source probe position and recorded at receiver source position. Maps were obtained
with longitudinal bulk waves when the transducers were inserted in the radial direc-
tion to the disc periphery and with surface waves when the transducers were inserted
in T direction. The tangential stress wave velocity maps were obtained at differ-
ent distances from the discs periphery (ex.: 20 mm, 140 mm).For radial inspection
the range of ultrasonic values was between 1630 m/s (green) for normal wood and
1542 m/s (red) for tension wood, in radial stressed measurements. For tangential
inspection the range of ultrasonic velocity was between 1450 m/s (green) for normal
wood and 395 m/s (red) for tension wood. Surface tangential stress wave veloci-
ties were lower in tension wood zones than in the opposite wood, the lateral wood
and the juvenile wood. The resolution of the velocity maps was 25. 10−3 m. The
resolution of stress wave based acoustic tomography is influenced by the applied
frequency, the number of sensors (Fig. 13.10 and Fig. 13.11) and the inversion tech-
nique for image reconstruction. The acoustic procedures described here can assist
in managing wood quality, assessing forest value and improving timber quality of
future plantations.
References
Andrews M (2003) Which acoustic speed ?. In: Proceedings of the 13th international symposium
nondestructive testing of wood. August 19–21 University of California, Berkley, pp 156–165
Attia al Hagrey S (2007) Geophysical imaging of root zone, trunk and moisture heterogeneity. J
Exp Botany (Special issue paper doi: 10;1093:/jxb/erl237) 1–16
Archer R (1986) Growth stresses and strains in trees. Springer, Berlin
Bamber RK (2001) A general theory for the origin of growth stresses in reaction wood. How trees
stay upright. IAWA J 22:205–212
13 Acoustic Tomography for Tension Wood Detection in Eucalypts 267
Berryman JG (2000) Analysis of approximate inverses in tomography. Part I and II. Optimisation
Eng 1:87–115 ,437–473
Biagi E, Gatteschi G, Masotti L, Zanini A (1994) Tomografia ad ultrasuoni per la caratterizzazione
difettologica del legno. Alta frequenza – Rivista di elettronica 6(2):48–57
Blackmore P (2008) Optimisation of steam reconditioning for regrowth – ash and plantation –
grown eucalypt species. Ph D thesis, University of Sydney, Australia
Boyd JD (1950) Tree growth stresses. Part II. The development of shakes and other visual failures
in timber. Aust J Appl Sci 1:296–312
Boyd JD (1972) Tree growth stresses. V. Evidence of an origin in differentiation and signification.
Wood Sci Technol 6:251–262
Bucur V (2003) Nondestructive characterization and imaging of wood. Springer Series in Wood
Science, Berlin, 354 pp
Bucur V (2005) Acoustics of wood, 2nd edn. Springer Series in Wood Science, Berlin
Bucur V (2005) Ultrasonic techniques for nondestructive testing of standing trees. Ultrasonics
43(4):237–239
Bucur V, Janin G, Herbe C, Ory JM (1991) Ultrasonic detection of reaction wood in European
species. Proceedings of the 10th congres Forestier Mondial, 17–26 September, Paris
Chafe SC (1977) Radial dislocations in the fibre wall of Eucalyptus regnans trees of high growth
stresss. Wood Sci Technol 11:69–77
Coutand C, Jeronimis G, Chanson B, Loup C (2004) Comparison of mechanical properties of
tension wood and opposite wood in Populus spp. Wood Sci Tech 38:11–24
Divos F, Divos P (2005) Resolution of stress wave based on acoustic tomography. Proceedings
of the 14th international symposium NDT on wood, http://www.ndt.net/article/v11n04/divos/
divos.htm. Accessed 5 April 2006
Donaldson LA (2001) Lignification and lignin topochemistry – an ultrastructural view.
Phytochemistry 57:859–873
Fengel D, Wegener G (1984). Wood chemistry, ultrastructure, reactions. De Gruyter, Berlin
Ferrand JC (1981) Recherches des solutions pratiques a apporter aux problèmes poses par les con-
traintes de croissance des arbres forestiers. Ph D Thesis Institut Nat. Polytechnique de Lorraine,
France
Gun TH, Hutchins DA, Green RJ, Andrews MK, Harris PD (2005) Noncontact, high resolu-
tion ultrasonic imaging of wood samples using coded chirp waveforms. IEEE Trans UFFQ
52(2):280–288
Jacobs MR (1938) The fiber tension of woody stems, with special reference to genus Eucalyptus.
Commonwealth Forestry Bureau, Canberra, Bull no 22.
Kubler H (1987) Growth stresses in trees and related wood properties. For Abstr 48:131–189
Lawday G, Hodges PA (2000) The analytical use of stress waves for the detection of decay in
standing trees. Forestry 75(2):447–456
Lee SJ, Oh JK, Yeo H, Lee JJ, Kim KB, Kim KM (2007) ield applications of nondestructive
testing for detecting deterioration in Korean historic wood buildings. Proceedings of the 15th
international symposium NDT on wood, Duluth, pp 227–232
Lin CJ, Kao YC, Lin TT, Tsai MJ, Wang SY, Lin LD, Wang YN, Chan MH (2008) Application
of an ultrasonic tomographic technique for detecting defects in standing trees. Int Biodeterior
Biodegradation 62:434–441
Martinis R (2002) Nondestructive techniques for decay diagnosis on standing trees. Ph D thesis,
University of Florence, Italy
Martinis R, Socco V, Sambuelli L, Schmitt O, Nicolotti G, Bucur V (2004) Tomographie
ultrasonore pour les arbres sur pied. Ann Forest Sci 61(2):157–162
Maurer H, Schubert S, Bachle F, Clauss S, Gsell D, Dual J, Niemz P (2006) A simple anisotropy
correction procedure for acoustic wood tomography. Holzforschung, 60:567–573
Müller M, Burghammer M, Sugiyama J (2006) Direct investigation of the structural properties of
tension wood cellulose microfibrils using microbeam X-ray fiber diffraction. Holzforschung
60:474–479
268 V. Bucur
Pellerin R, Ross RJ (2002) Nondestructive evaluation of wood. Forest Products Society, Madison,
Wisconsin, 210 pp
Polge H, Thiercelin F (1979) Growthstress appraisal through increment core measurements. Wood
Sci 12:86–92
Sandoz JL, Benoit Y (2007) Acousto – ultrasonic nondestructive evaluation of historic wood
structures. Proceedings of the 15th international symposium NDT on Wood, Duluth, p 245
Schubert S, Gsell D, Dual J, Motavalli M, Niemz P (2009) Acoustic wood tomography on trees
and the challenge of wood heterogeneity. Holzforschung 63:107–112
Socco V, Martinis R, Sambuelli L, Comino E, Nicolotti G (2000) Open problems concerning
ultrasonic tomography for wood decay diagnosis. Proceedings of the 12th symposium NDT
of Wood, University of Western Hungary, Sopron, p 468
Socco V, Sambuelli L, Nicolotti G (2004a) Ultraspnic tomography for nondestructive testing of
living trees. GNGTS – Atti del 19 Convegno Nationale /03.10
Socco V, Sambuelli L, Martinis R, Comino E, Nicolotti G (2004b) Feasability of ultrasonic
tomography for nondestructive testing of decay in living trees. Res Nondest Eval 15:31–54
Timell TE (1986) Compression wood in gymnosperms. Springer, Berlin
Tomikawa Y, Iwase Y, Arita K, Yamada H (1990) Nondestructive inspection of wooden poles using
ultrasonic computed tomography. IEEE Trans UFFC 33(4):354–358
Wang X, Carterr P, Ross RJ, Brashaw BK (2007) Acoustic assessment of wood quality of raw
forest materials – a path to increased profitability. For Prod J 57(5):6–14
Yang JL, Bucur V, Ng D (2007a) Acoustic detection of tension wood in eucalypts. Paper at 14th
international congress on sound and vibration. 9–12 July, Cairns, Australia
Yang JL, Bucur V, Ng D, Edbon N (2007b) Detection of tension wood in eucalypts discs
using ultrasonic and stress wave techniques. Proceeding of the 15th international symposium
nondestructive testing of wood. Forest Products Society, Madison, pp 143–148
Zhang YY, Lu ZQ (1996) Acoustical tomography based on second order Born transformation
perturbation approximation. IEEE Trans Ultrasonics 43(2):296–302
Zobel BJ, van Buijtenen JP (1989) Wood variation. Its causes and control. Springer, Berlin
Chapter 14
The Hygroscopic Warping of Cross-Laminated
Timber
Thomas Gereke, Per Johan Gustafsson, Kent Persson, and Peter Niemz
Contents
14.1 Introduction
This chapter focuses on moisture-induced deformations in three-layered cross-
laminated timber with a symmetrical build-up, where the fibre direction of the
middle layer is oriented perpendicular to that of the outer layers. Dimensional sta-
bility, i.e. the ability to resist warping, is of main interest for the application of such
wood panels. The cross lamination of the layers is advantageous to warping. The
moisture-induced expansion/contraction of each single layer is partly restrained by
the adjacent layers. The free swelling and shrinkage of adjacent layers differ approx-
imately by a factor of 10 (radial/longitudinal) to 20 (tangential/longitudinal). As a
T. Gereke (B)
Composites Group, Department of Civil Engineering & Department of Materials Engineering,
The University of British Columbia, 6250 Applied Science Lane, Vancouver, B.C.,
Canada V6T 1Z4
e-mail: thomas.gereke@composites.ubc.ca
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 269
DOI 10.1007/978-90-481-9550-3_14, C Springer Science+Business Media B.V. 2011
270 T. Gereke et al.
consequence of this difference, stresses and even cracks may occur. In large-scale
panels warping was observed. This reduces the serviceability in the practice. Due
to a climate gradient considerable distortions (warp) in the form of cup and twist
may occur. The types of warp are shown in Fig. 14.1. The cup deformation may
occur in two directions, the xz- and the yz-plane. This is often termed cup and bow
deformation (Ormarsson 1999).
To model the hygroscopic distortions, e.g., by means of a finite element simu-
lation, the moisture distribution in the panel is needed. Moisture transport in wood
below the fibre saturation point is a process governed by diffusion, which may be
simulated by Fick’s law (Siau 1995). The material parameter needed is the diffusion
coefficient D [m s−2 ]. It is well documented in the literature for Norway spruce
(Vanek and Teischinger 1989; Siau 1995; Hukka 1999). The information in liter-
ature about the moisture diffusion coefficient of adhesives is, on the other hand,
very limited. In this study, the diffusion coefficient of the adhesive was determined
by a combination of experimental measurements and numerical simulations of the
moisture distribution in the panels.
Water diffuses into wood through a boundary layer that provides resistance to the
diffusion if airflow at the wood surface is slow. The flux vector J perpendicular to the
surface with the normal vector n is driven by the difference in concentration of the
wood surface cs and the concentration ca that corresponds to the relative humidity
(RH) of the ambient air:
nJ = h (ca − cs ) (14.1)
The dot denotes derivative with respect to time. The three-dimensional model
was previously validated to distortions of sawn timber by Ormarsson (1999). The
elastic compliance matrix S contains the moduli of elasticity E, shear moduli G
14 The Hygroscopic Warping of Cross-Laminated Timber 271
and Poisson’s ratios ν (Bodig and Jayne 1982). The compliance matrix of wood is
orthotropic and defined as
⎡ ⎤
EL−1 −νLR ER−1 −νLT ET−1 0 0 0
⎢ ⎥
⎢ −ν E−1 E−1 −ν E−1 0 ⎥
⎢ RL L R RT T 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ −νTL EL−1 −νTR ER−1 ET−1 0 0 0 ⎥
S=⎢
⎢
⎥
⎥ (14.3)
−1
⎢ 0 0 0 GLR 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢
⎣ 0 0 0 0 G−1
LT 0 ⎥
⎦
0 0 0 0 0 G−1
RT
Due to moisture dependency of the moduli, the rate of the compliance matrix
Ṡ has to be considered. The vector α contains either the shrinkage or the swelling
coefficients in the orthotropic directions:
αT = αL αR αT 0 0 0 . (14.4)
The test set-up was made according to the standard DIN EN ISO 12572 (2001).
Cylindrical specimens with vertical oriented annual rings (i.e. tangential diffusion
direction) were used. Three layers, each 10 mm thick, were assembled by applying
one-component polyurethane adhesive (Purbond HB 180). After turning to the final
dimenaions, these specimens measured 30 mm in thickness and 140 mm in diameter.
They were conditioned before and after gluing at 65% RH until a balance of weight
was achieved. Then, the samples were fastened with rubber sleeves on cups, which
272 T. Gereke et al.
were filled with distilled water. The sleeves prevented the specimens from moisture
sorption at the edges. A climate gradient of 65/100% RH was applied by placing
the cups in a climate room. After 14, 21, 28 and 170 days, two rectangular samples
with a dimension of 40 mm × 80 mm were cut from the centre of each specimen.
Subsequently, they were split into layers of 5 mm thickness, whose moisture content
was determined by drying. According to Teischinger and Vanek (1987) two methods
are suited to produce layers of that thickness: cutting with a band saw and splitting.
With a thin saw blade, the cutting method provides more exact layer thicknesses.
However, it may affect the moisture content by heating due to friction. Thus, the
splitting method was chosen here.
In Table 14.1, the thicknesses and growth ring orientations of the panels and the
individual layers are shown. The panels were made of Norway spruce (Picea abies
[L.] Karst.) and glued either in the laboratory with one-component polyurethane
(adhesive application 200 g m−2 , one-sided, forming pressure 0.8 MPa, pressing
time 3 h) or glued by an industrial manufacturer with urea resin. Both the edges of
the boards and the layer-to-layer surfaces were glued. The material was before and
after the gluing conditioned at 20◦ C and 65% RH until equilibrium was reached.
The group of specimens indicated by AR refers to experiments where the influ-
ence of the annual ring orientation θ was studied. In global panel coordinates
(x, y, z), θ is defined as the angle between the tangential axis and the horizontal
coordinate of the geometric coordinate system as shown in Fig. 14.2. Three orien-
tations were tested: 0◦ , 45◦ and 90◦ , where θ =0◦ describes horizontally oriented
annual rings and θ =90◦ describes vertically oriented annual rings. The group of
specimens indicated by LR relates to testing where the influence of the layer ratio
LR on cup deformation was studied. The layer ratio is defined as
2aOL
LR = , (14.6)
atot
Group ID Production a αBL = aTL (mm)b aML (mm)b atot (mm)b θ (◦ ) LR (–)
where 2aOL is the thickness of the two outer layers and atot is the panel thickness.
LR was varied from 0.37 to 0.67.
The hygroscopic warping of three-layered cross-laminated wood panels was
determined on panels measuring 300 mm × 300 mm, which were initially con-
ditioned at 65% RH. The small faces of the samples were sealed with lacquer to
enforce the moisture flow within the panels in the thickness direction. The speci-
mens were then placed in a box on three supports as shown in Fig. 14.3a, b. The
contact area between the panel edges and the box top cover was insulated by a rub-
ber joint Fig. 14.3a. The climate difference of 65% RH and 100% RH between the
upper and lower surfaces was induced. These two levels of relative humidity were
obtained by means of water in the bottom of the box and a constantly conditioned
Fig. 14.4 Dimensions in mm, location of measuring points and coordinate system (x: fiber
direction of the outer layers, y: fiber direction of the middle layer)
climate room, in which the test set-up was stored. Relative displacements in the
z-direction of the measuring points, indicated in Fig. 14.4, were recorded by means
of dial gauges placed in a steel plate (Fig. 14.3d). Stop positions guaranteed identical
placement of the steel plate in every measurement. The displacement measurements
were carried out on days 1, 2, 4, 7, 10, 14, 21 and 31.
The cup deformation in the xz-plane (see Fig. 14.1) was determined as
1 A,D,G
cupxz = ūz + ūC,F,J
z − ūB,E,H
z (14.7)
2
and the cup deformation in the yz-plane as
1 A,B,C
cupyz = ūz + ūG,H,J
z − ūD,E,F
z . (14.8)
2
i,j,k
The notation ūz indicates the mean value of the displacements in the three
measuring points i, j and k as shown in Fig. 14.4. Twist deformation is omitted from
the analysis, since it is small compared to cup.
m(t) − m0
ĉ = , (14.11)
V0
where m0 [kg] and V0 [m3 ] are the weight and the volume of the sample at t=0
(corresponds to the beginning of the experimental test, i.e. the beginning of the
moisture gradient). The condition for steady-state diffusion is
∂ ĉ
= 0. (14.12)
∂t
The barrier time was determined to t∗ = 1000 h, since concentration diversifies
insignificantly afterwards. The reduction factor in Eq. (14.8) is κ = 0.45 [–].
It is assumed that moisture diffuses through the adhesive as water vapour.
According to Siau (1995) water vapour diffusion in wood slows down at higher
MC. Thus, the following formulation was evaluated from the experimental tests:
c (ϕ) = f1 ϕ 5 + f2 ϕ 4 + f3 ϕ 3 + f4 ϕ 2 + f5 ϕ + f6 (14.14)
has been found for spruce wood with ρ0,wood = 450 kg m−3 . The shape factors in
Eq. (14.14) are f1 = 170.81, f2 = −406.49, f3 = 366.60, f4 = −150.82, f5 = 32.91
276 T. Gereke et al.
Eight-node volume elements were used for the modelling. The fineness of the ele-
ment mesh was chosen in order to obtain accurate results. The mesh density (number
of elements)
NOE = Nx · Ny · (3Nz + 6) (14.16)
was chosen to Nx = 60, Ny = 120 and Nz = 5, which are the number of elements
in the global panel directions and the thickness mesh density of one wood layer
(Nz ). The adhesive layers were modelled by three volume elements in the thickness.
The annual ring orientation was taken into account by rectangular local coordinate
systems. Thus, a possible curvature of the annual rings and different lamellas within
a layer were not considered.
The elastic parameters describing the compliance matrix S (Eq. (14.3)) depend
on the moisture content. From the findings of Neuhaus (1981) the following relation
between the elastic parameters C (may be substituted by E, G or ν) and the MC could
be found:
C = a0 + a1 ω + a2 ω2 + a3 ω3 (14.17)
The shape factors ak (k = 0 . . . 3) are presented in Table 14.2. The data shows that
the moduli decrease with MC. ET and ER decrease within the hygroscopic range
to about 50% of their value at the ovendry state, while EL only decreases to 85%.
Table 14.2 Parameters ak of the relation between the moduli of elasticity E, the shear mod-
uli G, the Poisson’s ratios ν and moisture content (Eq. (14.13)) according to Neuhaus (1981),
0% ≤ ω ≤ 28%
Parameter a0 a1 a2 a3
The shear moduli also decrease with moisture content: GLR and GLT decrease to
about 50% and GRT decreases to 30% from the ovendry state to the fibre saturation
point. The Poisson’s ratio show a differing behaviour: νLR remains nearly constant
within the hygroscopic range while νLT and νRT increase when the moisture content
is increased.
The swelling coefficients were chosen according to Neuhaus (1981) and Sell
(1997) to αR = 1.7 · 10−3 and αT = 3.3 · 10−3 and according to Dahlblom et al.
(1999) (see also Ormarsson 1999) to αL = 5.0 · 10−5 (all non-dimensional).
The mechano-sorptive material parameters are based on results obtained by
Santaoja et al. (1991) and Mårtensson (1992) and are assumed to be independent
of the moisture content. The mechano-sorption coefficients in the orthotropic direc-
tions and planes that describe the mechano-sorption material matrix m in Eq. (14.4)
are mL = 1.0 · 10−4 , mR = 0.15, mT = 0.2, mLR = 8.0 · 10−3 , mLT = 8.0 · 10−3 and
mRT = 0.8 (all MPa−1 ). The coupling coefficients between the different directions
are chosen to μLR = 0, μLT = 0 and μRT = 1.
The adhesive, polyurethane, is assumed to act as a linear elastic isotropic material
(Konnerth et al. 2007). The adhesive layers were assumed to be 0.1 mm thick. The
elastic properties were chosen according to the tests of Konnerth et al. (2007) to
Eadh = 470 MPa and νadh = 0.3.
14.4.1 Moisture
The calculated moisture profiles and the measured MC are shown in Fig. 14.5. The
climate difference resulted in distinctive moisture profiles. The influence of the glue
lines is obvious and very significant. The largest gradient in MC could be detected
between the bottom and the middle layer. These two layers showed a rapid increase
in moisture during the first 14 days and also a significant increase in moisture from
day 14 to 170 (Fig. 14.6). The MC in the top layer did not change very much from
day 14 to day 170.
The results of the experimental tests after 31 days are shown in Table 14.3. The
major cupping was found in the yz-plane since the perpendicular to fibre direc-
tion of the outer layers is in the y-direction. This results in a dominant swelling
of the bottom layer in this direction. Due to the hindering by the middle layer, the
panel warps in the yz-plane. The simulated warping of panel AR90 is displayed in
Fig. 14.7. The figure shows the fact that the edges moved upwards while the centre
moved downwards.
The influence of the annual growth ring orientation on the cup deformation is
conspicuous (Table 14.3) and governed by the difference in swelling, stiffness prop-
erties and material orientation. The largest cup was found for panels with horizontal
Table 14.3 Measured warping after 31 days: mean and (standard deviation), ID according to
Table 14.1, n – number of samples, ρ0 – ovendry density (dry mass/dry volume)
Fig. 14.7 Simulated warp in a panel of series AR90 (one half of the panel is displayed,
displacements × 10, t = 744 h)
annual rings (AR0) reflecting the large swelling of the bottom layer. Tangential
swelling is about two times higher compared to radial swelling as it occurs in series
AR90. The smallest cupping was recorded for the panels with θ =45◦ (AR45), which
is due to the very small stiffness of softwood in that direction. It is about one half of
the tangential stiffness and about one quarter of the radial stiffness as a result of the
small GRT (Keunecke et al. 2008).
The tests of different layer ratios show increased cupping cupyz with increased
LR (Table 14.3). The simulated results of a variation of 0.5 ≤ LR ≤ 0.8 are plotted
in Fig. 14.8 for two panel thicknesses, 20 mm and 30 mm. An increased LR gives
an increased ratio of outer layers, i.e. a larger ratio of transverse oriented fibres in
the y-direction. This effect leads to a larger cupyz at higher layer ratios. The results
show an exponential increase in cup deformation.
The experimentally obtained time variation of cupyz is shown in Fig. 14.9.
Almost all panels attain their maximum cup between day 3 and day 12. Series AR45
and LR37 showed no significant maximum. For AR45 cupyz became constant from
day 5 whereas for LR37 a slightly increased cupyz was recorded during the whole
testing period. The MC of the bottom layer was found to increase rapidly (Fig. 14.6).
Hence, the expansion of the bottom layer, which is mainly in the y-direction, leads
to a strong increase in cupyz . The decrease in cupyz after the maximum is reached
is caused by an increase in MC of the middle layer and later of the top layer. The
moisture increase and, therefore, swelling of these two layers act in opposition to the
deformation of the bottom layer. Thus, cupyz decreases and cupxz increases. The sim-
ulated history plot of cupping in AR90 shows a good agreement to the experimental
test results (Fig. 14.9).
Compared to beech panels, as they were investigated by Gereke et al. (2009a),
spruce panels showed a good dimensional stability (Gereke et al. 2009b). The max-
imum cupyz in three-layered cross-laminated beech panels with layer thicknesses
280 T. Gereke et al.
Fig. 14.8 Simulated influence of the layer ratio LR on cupyz for two panel thicknesses, 20 mm and
30 mm, t = 744 h
of 10 mm and LR = 0.67 was detected to 1.14 mm. The cup in spruce panels was
measured to be 70% smaller (AR90).
A sensitivity study of the mechanical material parameters was performed on a
panel of type AR90. Nearly no influence could be detected for varying the shear
moduli and the Poisson’s ratios. Panel type AR90 also acts little sensitive on a modi-
fication of ET . Fig. 14.10 shows the influence of a 30% increase and a 30% decrease
Fig. 14.9 Nonlinear regression of the time variation of the cup deformation in the yz-plane, cupyz
14 The Hygroscopic Warping of Cross-Laminated Timber 281
50 cupxz
cupyz
40
Difference to reference (%)
30
20
10
all OL ML all OL ML
0
all OL ML all OL ML
–10
–20
–30
EL ER αL αR m L mR ER* αR*
50 cupxz
cupyz
40
Difference to reference (%)
30
20
10
all OL ML all OL ML
0
all OL ML all OL ML
–10
–20
–30
EL ER αL αR m L mR ER* αR*
Fig. 14.10 Relative change of the cup of panel type AR90 caused by 30% increase (upper) and
30% decrease (lower) of material parameter values for the complete panel (all) and the individual
layers (OL – outer layers, ML – middle layer). ∗ The mechano-sorptive fraction is omitted both in
the parameter study and in the reference calculation
282 T. Gereke et al.
Fig. 14.11 Simulated stress distribution σy along the y-axis in the bottom layer in panel AR90 after
744 h of climate difference 65/100% RH, x=150 mm, z=0 mm (upper) and z=10 mm (lower),
30% decrease of the material parameter values, ref – reference panel
The mechano-sorptive strain was found to be very important for the cup defor-
mation and for the results obtained in the parameter study. The mechano-sorptive
coefficient mR affects cupyz with about 54% when decreased and -34% when
increased. If omitting the mechano-sorptive effect, then the simulated cupping
would become about 10 times higher and deviate very much from the cupping
recorded in the experimental tests.
The mechano-sorptive strain is important also for the influence on the cup-
ping of ER and EL . Increase of these two parameters, all other parameters
kept constant, give decreased cupping if mechano-sorption is considered in the
analysis and increased cupping if not considered. With an increase in the coeffi-
cients of hygroexpansion the deformations increases and vice versa. A variation
of α L in the outer layers of 30% influences cupxz with about 30%. In con-
trast, the variation of EL , ER and α R has almost no influence on cupxz if the
mechano-sorptive strain increment is considered. The modification of the coeffi-
cient of hygroexpansion causes a larger expansion of the layers in the appropriate
direction. The longitudinal hygroexpansion influences the major direction. This
results in a larger cupxz but only slightly changes cupyz . The panel reveals the
opposite behaviour when α R is modified. Here, the minor direction is mainly
affected.
The stress distribution σy (radial direction) along the y-axis in the middle of
the panel (x=150 mm) is displayed for the positions z=0 mm (bottom face) and
z=10 mm (in the bottom layer at the glue line) in Fig. 14.11. Stresses are signifi-
cantly influenced by a variation of ER and α R , which lead to increased stresses. The
variation of the mechano-sorptive coefficient mR resulted in decreased stresses at the
glue line.
The stiffness of the glue line influences the warping only in a slight manner. In
a range up to Eadh =10 GPa, which conforms with results obtained by Konnerth
et al. (2006, 2007), we observed a linear decrease in cupyz of 5% while cupxz
is not influenced. Thus, the influence on warping of potentially different mod-
uli of elasticity of the two adhesives applied in the experimental tests can be
neglected.
14.5 Summary
The hygroscopic warping of three-layered cross-laminated timber is mainly affected
by the stiffness, the moisture-induced swelling and the mechano-sorptive coefficient
in the minor direction. These parameters can be altered by changing the layer ratio
and the orientation of the annual growth rings. According to the results of this study
the orientation of the annual growth rings should be chosen to θ =45◦ . The low stiff-
ness perpendicular to the grain leads to small cupping. Good results were obtained
for θ =90◦ as well. The small moisture-induced swelling in the radial direction leads
to small cupping. θ =0◦ should be avoided in practical applications. Furthermore,
the layer ratio LR should be chosen as small as possible but without losing the barrier
284 T. Gereke et al.
effect due to the cross-lamination. Depending on the panel thickness, the outer lay-
ers can become thin. Small thickness results in increased risk of cracking in the outer
layers when the panel is exposed to drying. Therefore, moisture-induced stresses and
cracking was studied by Gereke (2009). The possibilities and effects of alterations
of the panel structure, e.g., by using some alternative middle layer material, has also
been studied by Gereke (2009).
Acknowledgements Financial support from the European Cooperation in Science and
Technology (COST, Action E49) is gratefully acknowledged.
References
DIN EN ISO 12572 (2001) Hygrothermal performance of building materials and products –
Determination of water vapour transmission properties
Bodig J, Jayne A (1982) Mechanics of wood and wood composites. Van Nostrand Reinhold
Company, New York, NY
Dahlblom O, Persson K, Petersson H, Ormarsson S (1999) Investigation of variation of engineering
properties of spruce. 6th international IUFRO wood drying conference: Wood Drying Research
& Technology for Sustainable Forestry Beyond 2000. University of Stellenbosch, South Africa
Frandsen HL, Damkilde L, Svensson S (2007) A revised multi-Fickian moisture transport model
to describe non-Fickian effects in wood. Holzforschung 61:563–572
Gereke T (2009) Moisture-induced stresses in cross-laminated wood panels. PhD thesis, ETH
Zurich
Gereke T, Schnider T, Hurst A, Niemz P (2009a) Identification of moisture-induced stresses in
cross-laminated wood panels from beech wood (Fagus sylvatica L.) Wood Sci. Techn 43(3-4):
301–315. Published on line: September 23, 2008
Gereke T, Gustafsson PJ, Persson K, Niemz P (2009b) Experimental and numerical determina-
tion of the hygroscopic warping of cross-laminated solid wood panels Holzforschung 63(3):
340–347
Hanhijärvi A (1995) Modelling of creep deformation mechanisms in wood. Technical Research
Centre of Finland, Espoo
Hukka A (1999) The effective diffusion coefficient and mass transfer coefficient of nordic softwood
as calculated from direct drying experiments. Holzforschung 53:534–540
Keunecke D, Hering S, Niemz P (2008) Three-dimensional elastic behaviour of common yew and
Norway spruce. Wood Sci Technol 42:633–647
Konnerth J, Gindl W, Müller U (2007) Elastic properties of adhesive polymers. Part I:
Polymer films by means of electronic speckle pattern interferometry. J Appl Polym Sci 103:
3936–3939
Konnerth J, Jäger A, Eberhardsteiner J, Müller U, Gindl W (2006) Elastic properties of adhesive
polymers. Part II. Polymer films and bond lines by means of nanoindentation. J Appl Polym
Sci 102:1234–1239
Leicester RM (1971) A rheological model for mechano-sorptive deflections of beams. Wood Sci
Technol 5:211–220
Mårtensson A (1992) Mechanical behaviour of wood exposed to humidity variations. Report
TVBK-1006, Lund University, Lund, Sweden
Neuhaus F.-H (1981) Elastizitätszahlen von Fichtenholz in Abhängigkeit von der Holzfeuchtigkeit.
Ruhr-University Bochum
Ormarsson S (1999) Numerical analysis of moisture-related distortions in sawn timber. Phd thesis,
Chalmers University of Technology, Göteborg, Sweden,
Ranta-Maunus A (1990) Impact of mechano-sorptive creep to the long-term strength of timber.
Holz als Roh- und Werkstoff 48:67–71
14 The Hygroscopic Warping of Cross-Laminated Timber 285
Voichita Bucur
Contents
15.1 Introduction
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 287
DOI 10.1007/978-90-481-9550-3_15, C Springer Science+Business Media B.V. 2011
288 V. Bucur
evaluation of the topography of the fracture plane and finally, studies for different
failure modes. (Grosse and Finck 2006; Grosse at al. 2003).
Acoustic emission phenomena occur in wood or wood-based composites, like
in other solids, when subjected to a mechanical loading or other stress induced
conditions. The materials generate transient elastic waves caused by the release
of localized stress energy. Acoustic emissions can be detected in a wide range
of frequencies (100 kHz . . . 1 MHz), the most appropriated for wood being with
piezoelectric transducers between 100 and 200 kHz. (Drouillard 1990; Ono 1997;
Beall 2002; Kawamoto and Williams 2002; Bucur 2005). Resonant and wideband
transducers can be used in contact or not with the specimen. The rapid irreversible
stress-releasing events (delamination, dislocations, cracks, fibre breakage, etc) gen-
erate a spectrum of stress waves starting at 0 Hz and typically falling off at several
MHz. The monitoring of the level of acoustic emission activity forms one of the
basis of non-destructive testing in-situ of fracture phenomena, which are local
damages in wood and wood-based composites.
Fig. 15.1 A typical acoustic event and its main parameters: event duration, peak amplitude,
threshold voltage (ISO 12716)
energy which is the AE waveform squared and integrated over time. Power spectrum
was also used. When attempting to extract more information from the waveform and
emission patterns, a range of factors must be take into account such as source emis-
sion characteristics, propagation effects and transducers excitation effects. The use
of acoustic emission and Weibull statistics to characterize the specimen is an effec-
tive way to perceive subtle changes in material fracture characteristics (Okoroafor
and Hill 1995; Petri 1996; Hill et al. 1998).
The acoustic emission signal emanating from the sample is detected via a trans-
ducer in contact or not with the specimens (Murphy et al. 1990; Green 2004).
Furthermore the signal is amplified and processed in real time with appropri-
ate computer software. Fundamentals of acoustic emission signal processing and
instrumentation are specified in ASTM E750-98 (transducers, preamplifiers, filters,
amplifiers, cables threshold and counting instrumentation).
Acoustic emission activity is strongly dependent on irreversible, non elastic
deformations in wood. Due to the nature of signal source, the acoustic emission
tests are not perfectly reproducible. The wavelength of the signals is an important
parameter in planning the experiments. Signal localization is the basis of all anal-
ysis techniques. Problems related to the influence of ambient noise, the attenuation
of signals resulting in low-signal to noise ratio require sophisticated data processing
techniques. Sensor polarity, sensor orientation, source coordinates, P and S waves
displacement amplitudes recorded at each sensor, spectral amplitudes, the polarity
of the wave phases must to be under control. Knowing how the signals are recorded
is essential in understanding the acoustic emission technique.The interpretation of
the acoustic emission signals is very much dependent on the experimental condi-
tions and of the skill and art of the operator armed with wide theoretical knowledge
for the complex understanding of wave propagation and fracture phenomena.
Advance in acoustic emission technique for composites was achieved with the
development of artificial neural network back propagation algorithm (Fausett 1994;
Sasikumar et al. 2008). At the beginning of 1990s, new techniques for acoustic
emission signal processing emerged for the detection and characterization of failure
290 V. Bucur
of materials such as: wavelet transform (Chui 1992; Serrano and Fabjo 1996) or
simulation techniques (Grabec and Sachse 1997). Nonlinear acoustic modulation
technique was also developed (Solodov 1998; Ballad et al. 2004; Stoessel et al.
2003) for crack detection. Nonlinear vibrations of cracks produce nonlinear acoustic
emission signals which are high order harmonic radiations in surrounding air that
enables to locate and image cracks remotely.
Wood is a “dynamic acoustic” material from which the acoustic emission signals
emerge rapidly and randomly. An excellent review of the state of the art, techno-
logical challenges and future developments of parameter based acoustic emission
techniques for wood and wood based composites has been published by Kawamoto
and Williams (2002). In this section will be described the effect of species, grain
angle, periodic structure of the annual rings, tension wood and reused, old wood
on the acoustic emission activity of wood and the corresponding delamination and
fracture phenomena, studied with parameter- based techniques.
Fig. 15.2 Variation of the number of AE counts versus the tensile strength for four species (spruce
and pine and alder, oak and ash). Reiter et al. (2000, Figure 8)
15 Acoustic Emission Activity Induced by Delamination and Fracture 291
a b
Fig. 15.3 The onset of microcracks and the acoustic emission amplitude is shown in Reiter et al.
(2000, Figures 9) (Figure 9b, 10b)
the tensile strength for four species (spruce and pine and alder, oak and ash). The
softwoods exhibited higher AE counts than hardwoods. The onset of microcracks
and the acoustic emission amplitude is shown in Fig. 15.3. The microcrack forma-
tion around the crack tip and fiber bridging produce AE activity. High number of
AE events for softwoods are related to stable crack propagation phase. In hardwood
numerous crack arrests can be seen induced by unstable crack propagation, fewer
microcracks are formed, fiber bridging is not effective and AE counts are lower. For
pine the microcrack formation started at 86% of the maximum horizontal force. The
AE amplitudes are numerous and greatest at the beginning (below 400 s) and then,
after the macrocracks started to propagate, the amplitudes decreased continuously
until the end of the experiment. For the alder, the beginning of microcracks forma-
tion is at about 93% FH Max . The growth and arrest of crack is accompanied by very
high variation of the amplitudes. The signal amplitude decreased as soon as crack
arresting took place.
Reiter et al. (2002) compared wood fracture behavior in RL systems for the
spruce, alder, oak and ash. For this purpose the parameters selected were AE counts,
brittlennes, B, KIc and Gf (Table 15.1). Spruce exhibited the highest AE number of
Table 15.1 Fracture parameters and acoustic emission counts number for RL crack propagation
system for different species (data from Reiter et al. 2002)
Note: L is the ligament length, k init is the slop is proportional with the effective modulus of elas-
ticity, KIc is the critical stress intensity factor, Gf is the specific fracture energy, which characterize
the whole fracture process including crack initiation and propagation until complete separation of
the specimen into two pieces
292 V. Bucur
counts, the smallest brittleness and the smallest Gf . Lower B values indicate a more
brittle behavior. The presence of rays in hardwoods explained the significant dif-
ference between the behavior of spruce and other hardwood species. The speed of
loading can have an important effect on crack propagation and acoustic emission
activity (Ogawa and Sobue 1999).
Aicher et al. (2001) examined the delamination in softwoods loaded in tension
in plane normal to the tangential direction. Acoustic emission source location and
damage localization was possible within 100 mm distance. The characteristic rapid
increase in AE events rate was not accompanied by a visible change in specimen
stiffness. Ringger et al. (2003) noted that the location of acoustic emission source
is possible if the acoustic anisotropy of species (ex: spruce and beech) is taken into
account.
It is generally accepted that all physical and mechanical properties of wood are
affected by the slop of grain. The effect of grain angle on fracture toughness and
acoustic emission parameters has been studied by Ando et al. (1992) on Picea
jezoensis Carr., on single edge-notched specimens at 0◦ , 15◦ , 30◦ , 45◦ , 60◦ , 75◦
and 90◦ . For all specimens the notches were introduced in tangential direction.
The relationships between the grain angle and the critical crack opening dis-
placement (COD) and KIc are given in Fig. 15.4. The increasing of grain angle
induces the continuous decreasing of KIc , while the acoustic emission activity is
very different from small angle (0◦ . . . 15◦ ) compared with large angles (> 30◦ ).
The m coefficient decreases quite linearly with increasing grain angle in Picea
jezoensis.
Figure 15.5 depicts the variation of the acoustic events number, the signal ampli-
tude, at various levels of loading up to the proportional limits. Two groups of
different signals have been observed, at small angle (0◦ and 15◦ ) only signals
of low amplitude 40–50 dB were generated in the early stage. The increasing of
load determined the gradual increasing of higher amplitudes. For large angles,
signals of higher amplitudes (65–75 dB) were generated from the early stage of
fracture.
The ex-situ inspection of fractured specimens showed that at small angles the
crack propagated along the grain, with transwall typical fractures in earlywood. At
large angles the cracks propagate along the grain, as interwalls fractures. The acous-
tic emission signals generated at small angle, before crack propagations correspond
probably to the microcracks at the tip zone induced by delaminations between cell
wall layers.
Cyra and Tanaka (2000) studied fracture phenomena and acoustic emission
in relation to routing cutting process. The acoustics events were related to the
grain angle, the state of cutting and the surface roughness. The acoustic emission
technique is promising for routing monitoring.
15 Acoustic Emission Activity Induced by Delamination and Fracture 293
a b
c d
Fig. 15.4 Relationship between the grain angle and fracture parameters and acoustic emission
parameters (Ando et al. 1992, Figures 3, 4, 6, and 8), (a) COD and grain angle, (b) KIc and grain
angle, (c) AE cumulative events (counts) and grain angle up to proportional limit of Load –COD
curves, (d) m and grain angle Mokuzai Gakkaishi (1992, Figures 4, 5, 6, and 8)
Table 15.2 Fracture modes at different grain angle orientation of specimens under torsional
loading (data from Chen et al. 2006)
Red lauan II RL, III RT I RL, I RT II RL, III RT and III TR; II TL, III TR and III RT
Sitka Spruce II TL, III TR I TL, I TR II TL, III TR and III RT; II RL, III RT and III TR
Ansell (1982) was probably the first to demonstrate the influence of the earlywood-
latewood ratios under tensile loading on acoustic emission activity, expressed by the
shape of AE strain curve. Dill – Langer and Aicher (2000) observed micro fracture
nucleation of spruce under tensile loading; it was an on-set of AE prior to the first
visible crack growth step.
Ando et al. (1991) studied the effect of the location of the crack tip in an annual
ring of sugi (Cryptomeria japonica) in single edge notched specimens of the TR
crack propagation system. The critical stress intensity factor KIc varied according
to the location of the crack tip, from the pith to the bark or from the bark to the
pith. The crack tip was located in earlywood or in latewood. When located in the
15 Acoustic Emission Activity Induced by Delamination and Fracture 295
earlywood, KIc increased with decreasing distance between the crack tip and the
latewood located forward of the tip. The patterns of the AE amplitude distribution
at various levels of loading up to the proportional limit in load-COD diagrams were
studied. Two populations of peaks amplitudes were observed, at 45 dB and 70 dB,
corresponding with two microcracks at different energy levels and recognized as
transwall failure and intrawall failures (Ando and Ohta 1995). It was supposed that
the variability of KIc by the crack tip position in the annual ring is due either to
the difference in cell shape or cell wall thickness around the crack tip, and to the
difference in stress concentration induced by crack location and the direction of
crack propagation. Ando and Ohta (1999) extended the previous studies to Sitka
spruce (Picea sitchensis), sugi (Cryptomeria japonica) and akamatsu (Pinus densi-
flora) taking into account the anisotropy and the heterogeneity of different zones in
the annual ring with FEA. Figure 15.6 shows the variation of KIc as a function of the
location of the crack tip in the annual ring for sugi (Cryptomeria japonica), spruce
(Picea sitchensis) and akamatsu (Pinus densiflora), for pith side notched and bark
side notched specimens.
The stress at the tip is expressed by
σtip = α σ and α = σtip /σ
where:
σ tip is the stress at the crack tip, in the tangential direction obtained with FEA
α is the stress concentration factor
σ is the nominal stress
Figure 15.7 shows the variation of stress concentration factor α as a function
of the location of the crack tip in the annual ring for sugi (Cryptomeria japon-
ica), spruce (Picea sitchensis) and akamatsu (Pinus densiflora) for pith side notched
(filled circles) and bark side notched (open circles) specimens. When the degree of
stress concentration was small an acoustic emission signal was generated and an
intrawall failure was observed, before the crack initiation. When the degree of stress
concentration was large, a signal of large amplitude was generated and transwall
failure was observed.
The annual ring scale was also studied by Dill –Langer and Aicher (2000). They
monitored simultaneously the crack propagation and the acoustic emission activ-
ity of notched spruce specimens in tension load, in RT and TR systems. A third
configuration was studied for specimens at 45◦ between radial growth direction
and load axis, with notch at 45◦ versus R. The damage mechanism was stud-
ied at micro (tracheids diameter 50 μm) and mezzo scales (annual ring width
3. . .5 mm). Confocal laser scanning microscopy was used for in-situ observation
of crack growth. Two characteristic damage phenomena have been observed. When
crack propagation is in TR system, or T propagation, the rupture of earlywood cell
walls was observed (intrawall failure). When crack propagation in RT system, or
R propagation, the debonding of the interface middle lamella between two adja-
cent tracheids was observed (interwall failure). The crack path was in zigzag for
the third configuration with the specimen oriented at 45◦ . “The more the crack
296 V. Bucur
Fig. 15.6 Variation of KIc in the annual ring as a function of the relative position of the crack for
sugi, spruce and akamatsu Permission J Wood Sci 45:275–283 Figure 9
approaches the earlywood/latewood transition the more it deviates from the initial
direction until propagation coincides with 45◦ At this stage the crack surface con-
sists predominantly of ruptured cell walls comparable to the rupture in RT system.
Having reached half the specimen with the crack turns 90◦ anticlockwise propa-
gating through late and transition wood. Thereby the crack surface is smooth as
15 Acoustic Emission Activity Induced by Delamination and Fracture 297
Fig. 15.7 Variation of stress concentration factor versus the relative position of the crack for sugi,
spruce and akamatsu Ando and Ohta (1999, Figure 10)
Acoustic emission technique from the early 1980 was related to checking detection
and wood drying (Kawamoto and Williams 2002). Acoustic emission signals dur-
ing drying are related to events produced by checking and water movement and it
is difficult to distinguish the source of emission. Wave pattern recognition using
cluster analysis was limited in applications for monitoring and controlling kilns
(Schniewind et al. 1996; Lee et al. 1996). The major AE sources are the surface
tensile stress induced by the water movement below the FSP and the thermal stress
related to temperature variation. Wood microscopic structure shows slips in the crys-
talline segments of cellulose (Booker 1994). The checking occurs when the rate of
298 V. Bucur
slips exceeded a critical value. The slip lines generate slow AE events. Brittle micro-
cracks in the cell wall and delaminations generate rapid AE events (Sato et al. 1984;
Schniewind 1989).
Acoustic emission energy has been used to identify the damage during drying
by Kowalski et al. (2004). Roughly three groups of AE signals were identified. The
first group, observed at the beginning of drying process – with small amount of
energy and an important number of events, inducing microcracks. The second group,
identified during drying, with increasing acoustic emission energy and diminishing
the number of events, when the surface of the specimen shrinks and the moisture
content is around the fiber saturation point, inducing macrocraks. The final detected
stage of drying corresponds to specimen core drying, and has relatively low energy
AE signals. For better understanding of fracture phenomena during drying and the
related acoustic emission activity, problems related to the attenuation of signals and
transducers sensitivity must be solved.
15 Acoustic Emission Activity Induced by Delamination and Fracture 299
Fig. 15.9 Acoustic emission signals during oak drying (Kim et al. 2005, Figures 2 and 3).
(a) cumulative counts versus drying time (b) AE waveform during surface check (c) AE waveform
during water movement
Recycling old wood salvaged from old structures, is a real problem for the modern
and green sustainable society (Ando et al. 2006, 2007). Acoustic emission technique
300 V. Bucur
Table 15.3 Recognition rate with ANN classifiers using AE parameters and principal components
inputs (data from Kim et al. 2005)
is appropriated for examining the differences between new and old wood in shear-
ing fracture (Japanese standard JIS Z 2101-1994). Specimens of Japanese red pine
(Pinus densiflora) from 270 old structural members were compared with specimens
of new wood, lumbered within 3 years before testing. The cumulative AE event
counts versus shearing stress is presented in Fig. 15.10. In the initial stage of loading
a c
b
Fig. 15.10 Behavior of old and new wood (Ando et al. 2006, Figure 3, 5) (a) cumulative AE
events versus shearing stress in new wood; (b) cumulative AE events versus shearing stress in old
wood; (c) m value versus relative stress (ratio of stress to maximum stress)
15 Acoustic Emission Activity Induced by Delamination and Fracture 301
the behaviour of new wood is very different from that of old wood. The difference
in behaviour of old and new wood is well put in evidence in Fig. 15.10c with the
variation of m value versus the relative stress. The increase in m value expresses
the acoustic emission events of small amplitudes with stable crack propagation.
The decrease in m value signifies the frequent occurrence of high-amplitudes AE
events, with predominantly unstable fractures. The fractographic analysis revelled
in new wood a smooth flat surface with intrawall failure. The fracture surface of old
wood was rough and irregular of trans-wall type, initiated from the bordered pits.
Under shearing test the old wood underwent stable crack propagation before the
final fracture.
The acoustic emission energy, monitored with a high speed waveform acquisi-
tion system was reported by Landis and Whittaker (2001). The progressive crack
growths along the grain in direction in notched eastern hemlock (Tsuga Canadensis)
specimens has been studied in Mode I (Fig. 15.11). Crack length data and load-
CMOD data were used to calculate the fracture energy. The release of the acoustic
emission energy was calculated by integrating the instantaneous power of an elastic
wave over all frequencies, and by multiply the result by the length of the AE wave-
form. From data plotted in Fig. 15.12 (which gives the variation of the measured
load and cumulative acoustic emission energy release during fracture) it can be
noted that:
Figure 15.13 shows the linear relationship between the acoustic emission energy
rate and the fracture energy release.
A more refined approach to examine the mechanisms of energy dissipation dur-
ing fracture perpendicular to grain and crack propagation in radial direction, in
spruce was proposed by Watanabe and Landis (2007). It was hypothesized that the
total dissipation energy during fracture is composed from two main components,
the first one, corresponding to the dissipation of energy induced by short bursts
and unstable crack growth, reflected by the strong acoustic emission activity and
the second one, corresponding to additional energy, dissipated in the form of more
gradual processes that include creep deformation, and slow crack growth, as can
be seen from Fig. 15.14. Towards the end of the experiment the acoustic energy
rises again, but before that, the fracture was slowed probably because of bridging
effect combined with the action of other cohesive forces. More research is needed
to understand the rapid rise of energy at the end of the test. It was advanced that this
behavior is related to cohesive forces at the crack tip.
15.5 Summary
Fig. 15.14 The variation of AE energy, the consumed energy and load versus time, for a specimen
of Canadian Eastern spruce tested in mode I (Watanabe and Landis 2007, Figure 5)
tri-dimensional disordered lattice model of dynamic fracture, which can relate the
acoustic response to the internal damage of the sample.
References
Aicher S, Höfflin L, Dill-Langer G (2001) Damage evolution and acoustic emission of wood at
tension perpendicular to fiber. Holz als Roh Werkst 59:104–116
Ando K, Takita A, Hirashima Y, Sasaki Y (2007) Fractography of old wood. Nagoya Univ Forest
Sci 26:1–7
Ando K, Hirashima Y, Sugihara M, Hirao S, Sasaki Y (2006) Microscopic processes of shearing
fracture of old wood, with acoustic emission technique. J Wood Sci 52, 6:483–489
Ando K, Ohta M (1999) Variability of fracture toughness by the crack tip position in an annual
ring of coniferous wood. J Wood Sci 45:275–283
Ando K, Ohta M (1995) Relationship between the morphology of micro-fractures of wood and the
acoustic emission characteristics. Mokuzai Gakk 41:640–646
Ando K (1993) Direct observation of micro-fracture process of wood by SEM and its acoustic
emission characteristics under tension test. Proceedings 9th conference on acoustic emission,
Osaka, Japan, pp 85–90
Ando K, Sato K, Fushitani M (1991) Fracture toughness and acoustic emission characteristics of
wood. I. Effect of the location of a crack tip in an annual ring. Mokuzai Gakk 37:1129–1134
Ando K, Sato K, Fushitani M (1992) Fracture toughness and acoustic emission characteristics of
wood II: effects of grain angle. Mokuzai Gakkaishi, 38(4):342–349
Ansell MP (1982) Acoustic emission from softwoods in tension. Wood Sci Technol 16:35–38
American Society for Nondestructive Testing – ASNT (2005) Acoustic emission testing. In
Nondestructive Testing Handbook, 3rd edition, vol 6, Published by ASNT, Columbus OH
ASTM E750-98 Standard practice for characterizing acoustic emission instrumentation
Ballad EM, Vezirov SY, Pfleider K, Solodov IY, Busse G (2004) Nonlinear modulation technique
for NDE with air-coupled ultrasound. Ultrasonics 42:1031–1036
Beall FC (2002) Overview of the use of ultrasonic technologies in research on wood properties.
Wood Sci Technol 36(3):197–212
Booker JD (1994) Acoustic emission and surface checking in Eucalyptus Regnans boards during
drying. Holz als Roh Werkst 52:383–388
Bucur V (2005) Acoustics of wood. Springer, Heidelberg
Chen Z, Gabbitas B, Hunt D (2006) Monitoring of fracture of wood in torsion using acoustic
emission. J Mater Sci 41(12):3645–3655
Chui CK (1992) Introduction to wavelets. San Diego, Academic
Cunderlik I, Molinski W, Raczkowski J (1996) The monitoring of drying cracks in the tension and
opposite wood by acoustic emission and SEM. Holzforschung 50:258–262
Cyra G, Tanaka C (2000) The effects of wood-fiber directions on acoustic emission in routing.
Wood Sci Technol 34(3):237–252
Dill –Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic emis-
sion. Proceedings internation conference Wood and wood fiber composites, Stuttgart, pp
93–104
Drouillard TF (1990) Anecdotal history of acoustic emission from wood. J Acoust Emission
9(3):155–176, 1990.
Fausett LV (1994) Fundamentals of neural networks: architecture, algorithms and applications.
Prentice Hall, Englewood Cliffs
Grabec I, Sachse W (1997) Synergetics of measurements, prediction and control. Springer, Berlin
Green RE Jr (2004) Non-contact ultrasonic techniques. Ultrasonics 42:9–16
Grosse C, Ohtsu M (2008) Acoustic emission testing basics for research – applications in civil
engineering. Springer, Heidelberg
15 Acoustic Emission Activity Induced by Delamination and Fracture 305
Grosse CU, Finck F (2006) Quantitative evaluation of fracture processes in concrete using signal-
based acoustic emission techniques. Cem Concr Compos 28:330–336
Grosse CU, Reinhardt HW, Finck F (2003) Signal-based acoustic emission techniques in civil
engineering. J Mat Civ Eng. 15(3):274–279
Hill R, Brooks R, Kaloedes D (1998) Characterization of transverse failure in composites using
acoustic emission. Ultrasonics 36:517–523
ISO 12716 – 2001 Non-destructive testing – Acoustic emission inspection
JIS Z 2101 – 1994 Methods of test for woods
Kawamoto S, Williams RS (2002) Acoustic emission and acousto-ultrasonic techniques for wood
and wood-based composites: a review – General Technical Report FPL-GTR-134. Madison, WI
Kim KB, Kang HY, Yoon DJ, Choi MY (2005) Pattern classification of acoustic emission signals
during wood drying by principal component analysis and artificial neural network. Key Eng
Materials 297– 300:1962–1967
Kowalski SJ, MolinskiW, Musielak G (2004) The identification of fracture in dried wood based on
theoretical modelling and acoustic emission; Wood Sci Technol 38(1):35–52
Landis E N (2008) Acoustic emission in wood. In: Grosse C, Ohtsu M (eds) Acoustic emission
testing basics for research – applications in civil engineering. Springer, Heidelberg, pp 311–322
Landis E N, Whittaker DB (2001) Acoustic emission as a measure of fracture energy. Proceedings
of the 1st conference of the European society for wood mechanics, Vila Real, Portugal
Lee SH, Quales SL, Schniewind AP (1996) Wood fracture, acoustic emission and the drying
process. Part II. Acoustic emission pattern recognition analysis. Wood Sci Technol 30:283–292
Minozzi M, Caldarelli G, Pietronero L, Zapperi S (2003) Dynamic fracture model for acoustic
emission. Eur Phys J B 36:203–207
Muravin B (2009) Acoustic emission, science and technology. J of Building and Infrastructure
Engineering of the Israeli Assoc of Engineers and Architects (in press). www.muravin.com.
Accessed 22 July 2010
Murphy JC, Majerowicz S, Green RE Jr, Glass JT (1990) Laser interferometric probe for detection
of acoustic emission. Mater Eval 48:714–720
Ogawa M, Sobue N (1999) Effect of loading speed on fracture of timber with a crack. Mokuzai
Gakk 45(6):461–470
Okoroafor EU, Hill R (1995) Investigation of complex failure modes in fibre bundles dur-
ing dynamic mechanical testing using acoustic emission and Weibull statistics. J Mater Sci
30:4233–4243
Ono K (1997) Acoustic emission. In: Crocker MJ (ed) Encyclopedia of acoustics, Wiley, New
York, NY Chapter 68:797–809
Persson K (1997) Modeling of wood properties by a micro-mechanical approach. Ph D Thesis,
Lund University Report TV SM – 3020
Petri A (1996) Acoustic emission and microcrack correlation. Phil Mag B 77(2):491–498
Reiter A, Stanzl-Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of
softwood and hardwood. Wood Sci Technol 34(5):417–430
Reiter A, Stanzl-Tschegg SE, Tschegg EK (2002) Fracture characteristics of different wood species
under Mode I loading perpendicular to the grain. Mater Sci Eng A 332:29–36
Ringger T, Höfflin L, Dill-Langer G, Aicher S (2003) Measurement of the acoustic anisotropy of
soft and hardwood; effect of source location. Otto-Graff J 14:231–253
Sachse W, Kim KY (1987) Quantitative acoustic emission and failure mechanics of composite
materials. Ultrasonics 25:195–203
Sasikumar T, Rajendraboopathy S, Usha KM, Vasudev ES (2008) Artificial Neural Network
Prediction of Ultimate Strength of Unidirectional T-300/914 Tensile Specimens Using Acoustic
Emission Response J. Nondestructive Eval. 27(4):127–133
Sato KN, Kamei M, Fushitani M, Noguchi M (1984) Acoustic emission generated upon mechano-
sorptive creep of wood. Mokuzai Gakk 30(8):653–659
Schniewind AP (1989) Concise encyclopedia of wood and wood-based material. Pergamon Press,
Oxford
306 V. Bucur
Schniewind AP, Quales SL, Lee SH (1996) Wood fracture, acoustic emission and the drying
process. Part I Acoustic emission associated with fracture. Wood Sci Technol 30:273–282
Scott IG (1991) Basic Acoustic Emission. Gordon and Breach Science Publishers, New York
Serrano EP, Fabjo M (1996) Application of wavelet transform to acoustic emission signal
processing. IEEE Trans Signal Proc 44(5):1270–1275
Solodov I Y (1998) Ultrasonics of non-linear contacts: propagation, reflection and NDE-
applications. Ultrasonics 36:383–390
Stephens RWB, Leventhal HG (1974) Acoustic and vibration. Chapman and Hall, London
Stoessel R, Predak S, Solodov I, Busse G (2003) In:Green RE Jr, Djordjevic BB, Hentschel MP
(eds) Nondestructive materials characterization, Springer, Berlin, XI:117
Watanabe K, Landis EN (2007) An acoustic emission based study of energy dissipation in radially
loaded spruce. In: Navi P, Guidon A (eds) Proceedings of the 3rd international symposium on
wood machining, Lausanne, Switzerland, pp 179–182
Chapter 16
Delamination Detection in Wood – Based
Composites Panel Products Using Ultrasonic
Techniques
Contents
16.1 Introduction
Wood – based composite panel products (WBCP) are manufactured from veneer,
wood particles, strands or fibres bind together with different types of adhesives such
as urea-formaldehyde resin, phenol-formaldehyde resin, melamine formaldehyde
resin, methylene diphenyl diisocyanate or polyurethane resins. The nature and the
quality of the raw material and of the adhesives determine the characteristics of
the products (mechanical properties, water resistance, dimensional stability, surface
quality and machinability).
The products existing on the market can be classified such as:
– glued laminated timber – glulam, crosslam, glulam slabs,
– veneer based panels – plywood, laminated veneer lumber,
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 307
DOI 10.1007/978-90-481-9550-3_16, C Springer Science+Business Media B.V. 2011
308 V. Bucur and S. Kazemi-Najafi
Bulk waves or surface waves can be used for the characterization of the mechani-
cal behaviour of wood-based composites panel products. The waves characteristics
related to the propagation in an infinite solid of bulk waves – longitudinal and shear
waves, (named also P and S waves) are shown in Fig. 16.1. The longitudinal waves
are characterized by the fact that the direction of wave propagation is parallel to the
direction of particle motion (polarization). In the case of shear waves the particle
motion is perpendicular to the direction of wave propagation.
For the inspection of plate type specimens, as wood based composites panels,
it is also interesting to use Lamb waves. The Lamb waves are elastic wave modes
Fig. 16.1 Schematic representation of propagation of bulk longitudinal and shear waves in
solids (Olympus, Panametrics, Figure 3, page 41)). http://www.olympus-ims.com/data/File/
panametrics-UT.en.pdf (visited 16 june 2009)
310 V. Bucur and S. Kazemi-Najafi
– extensional mode which is symmetric versus the axis of the plate; compression
and tension effects can be observed; the displacement is due to Poisson’s effect.
Each symmetric mode has infinite number modes (s0 , s1 , s2 , . . . , sn ). The particle
motion is parallel with the direction of wave propagation. This mode is relatively
nondispersive (not dependent on frequency f ) if hf < 0.5 and where h is the plate
thickness.
– flexural mode which is antisymmetric versus the axis of the plate. Each anti-
symmetric mode have an infinite number modes (a0 , a1 , a2 , . . . an ) The particle
motion is perpendicular to the direction of wave propagation. This mode is highly
dispersive
The lowest order Lamb wave modes (s0 and a0 ) which are the two fundamentals
are commonly termed “plate waves” and are mostly used for non-destructive testing
of wood – based composite panels.
Lamb wave propagation occurs when the wavelength λ is as 0.1h < λ < 10h,
where h is the plate thickness. For experimental reason λ > 10h was recommended
by Bray and Stanley (1997) while λ > 5h and λ > 3h were proposed by Huang
16 Delamination Detection in Wood – Based Composites Panel Products 311
(1999) for experiments with composites and with extensional and flexural waves,
respectively. It is understood that for a type plate specimen, the length and width
of the plate must be much greater than the wavelength used for the inspection. The
layered solid in which plate waves propagate can be understand as a homogeneous
solid because of the long wavelength.
The techniques based on Lamb wave propagation are able to measure the flexural
and transverse shear rigidity of laminated fibre composites.
Defect detection using Lamb waves can be performed by measuring wave veloc-
ity or attenuation. Numerous reference books and articles (Bland 1988; Birks and
Green 1991; Nayeh 1995; Rose et al. 1987, 1992; Schmerr 1998; Rose 1999; Bar-
Cohen and Chimenti 1986; Tang 1988; Tang and Henneke 1989a; Huang et al. 1998)
reported the utilization of velocity and attenuation measurements for the evalua-
tion of elastic constants of fibre composites and for the detection of defects such
as delaminations and disbondings. The measurements of attenuation of ultrasonic
waves are more complicated than of velocity. The coupling medium could be a high
source of error for attenuation measurements, but is not as critical in velocity mea-
surements. Non contact transducers are recommended to avoid all these problems.
Tucker (2001), Tucker and Bender (2003), and Tucker et al. (2003) successfully
utilised the Lamb waves for continuous nondestructive inspection of wood-based
composite panels and of wood-plastic composites.
The linear inspection techniques have been developed in the hypothesis that the
acoustic wave amplitude is infinitesimally small and the response of the material is
assumed to be linear to the excitation signal, obeying Hook’s law. Under the label–
linear ultrasonic techniques- three main groups of techniques are recognized:
– reflection technique, or pulse – echo technique, for which only one transducer
is used. The ultrasonic wave is directed into the specimen and after propagating
twice through its thickness is recorded by the same transducer. This technique
works with continuous waves or with pulses
– through –transmission technique for which two transducers are used, the trans-
mitter and the receiver. The energy injected by the transmitter travel through the
specimen and is recorded by the receiver. This technique works with continuous
waves or with pulses
– emission technique, known also as acoustic – emission technique, uses only one
transducer which is a receiver and which collects waves emitted by the specimen
under mechanical stress. This technique is not described in this chapter.
For all these techniques the contact transducers are coupled to the specimen with
a coupling medium or with a delay line. Non contact transducers called also air
coupled transducers can be used for the generation of bulk waves or surface waves
without contact between the specimen and the transducers. In this case the coupling
medium is the air (or an other gas).
312 V. Bucur and S. Kazemi-Najafi
as a receiver. Driven by the pulser, the transducer generates energy converted into
mechanical vibration which travels twice through the material to be tested. When the
signal travels through a structural discontinuity of the material, a flaw (i.e. delami-
nation), part of the energy will be reflected back from the flaw surface. The reflected
signal is captured by the receiver and is displayed on the screen of the oscilloscope.
The ratio between the distance and the travelling time of the signal gives the veloc-
ity of propagation of the ultrasonic wave through the material under test. A more
sophisticated signal treatment gives information about the location, the size, the
orientation and other features of the defect. In pulse echo methods, a single-sided
access is needed and this is an advantage for testing panels. The transducers used
for this purpose must generate pulses of broad band frequency and short rising time.
For high attenuator materials such as wood – based composites this technique has
some limitations, especially in high frequency range.
The echo technique allows the direct localization of a reflector as for example
the back wall of the specimen or flaw. A clear echo from the back of a specimen
means that the specimen is free of defects. The presence of a defect in a specimen
can be identified when echoes with short time of flight and echoes with long time
of flight are measured (Fig. 16.3b). If the velocity in the sound material is known
as well as the size of the specimen, the defect can be located. Figure 16.4 shows a
structural element with important cracks detected with pulse echo technique. The
measurements were performed with longitudinal waves and 100 kHz. A very strong
echo was observed at 6.2 cm, which corresponds to the big crack observed on the
photograph. The echo from the back wall is at 12.5 cm for which the velocity was
1630 m/s.
Fig. 16.4 A structural element inspected with pulse echo technique (Hasenstab et al. 2006).
http://www.ndt.net/article/ecndt2006/doc/Th.2.4.1.pdf (a) view of the structural element in pine
(15.5 × 12.5 × 75 cm) (original Figure 3, page 4); (b) pulse echo measurements along the sound
zone in A-scan mode and in B-scan mode, with P wave transducers of 100 kHz (original Figure 2,
page 4); (c) ultrasonic image of the specimen scanned in B-mode, the crack is clearly identified at
6.2 cm (original Figure 5, page 5)
the matrix, fractured fibres, etc) are detectable with transducers oriented at an angle
to the surface of the specimen (Moran et al. 1985; Wooh and Daniel 1990; Gorman
1991; Steiner et al. 1995). It was demonstrated that a combination of normal and
oblique incidence with pulse-echo ultrasonic techniques can be used to produce a
highly detailed volumetric image of complex damage states dominated by transverse
matrix cracks and delamination.
A comparison between the pulse echo method and the through transmission
method in B-scan mode is shown in Fig. 16.6. The echo from the defect can be
seen between the echoes from the front and back wall of the specimen. In through
transmission technique the defect is represented by a lower amplitude signal.
16 Delamination Detection in Wood – Based Composites Panel Products 315
Fig. 16.6 Comparison between the pulse echo method and through transmission method in
B – scan mode (Stoessel 2004, Figure 21a, b). (a) reflection technique in B scan mode;
(b) through transmission technique in B scan mode. http://elib.uni-stuttgart.de/opus/volltexte/
2004/1622/pdf/Dis_Stoessel.pdf
316 V. Bucur and S. Kazemi-Najafi
Fig. 16.8 Focused slanted transmission of air coupled ultrasound (Solodov et al. 2004b, page
506). (a) experimental configuration θ angle of incidence, θ 0 “resonance” angle α azimuthal
angle; (b) plate wave excitation with air coupled ultrasound, the ultrasonic wave in air and the
plate wave in the sample move in phase along the surface (Döring et al. 2006, Figure 2, page 2).
http://www.ndt.net/article/ecndt2006/doc/P123.pdf
Vair
sin θ = (16.1)
Vplate
where Vplatei is the velocity in the plate, ω is the frequency, t is the thickness of
the plate, Ei is the Young’s modulus, ρ the density and ν the Poisson ratios. Their
product is <1.
The plate wave velocity can be used to calculate the Young’s moduli.
318 V. Bucur and S. Kazemi-Najafi
a b
Fig. 16.9 Wood anisotropy with flexural waves. (a) in a polar plot of flexural wave velocity in
beech veneer as a function of orientation angle (Döring et al. 2006). (Note: V = 520 m/s at 80◦
and V = 1050 m/s at 160◦ ) http://www.ndt.net/article/ecndt2006/doc/P123.pdf; (b) in a spruce
veneer lamina 1.2 mm thickness, in earlywood and latewood with focused slanted transmission
of air coupled ultrasound. Note phonon focusing due to wave guide effect in annual ring of wood
(Solodov et al. 2006, Figure 6, page 7). http://www.ndt.net/article/ecndt2006/doc/Th.4.7.1.pdf
Fig. 16.10 Experimental configuration for focused slanted transmission mode (FSTM) and for
focused slanted reflection mode (FSRM) for the detection of plate acoustic waves (PAW ) and
surface acoustic waves (SAW) when the plate thickness is grater than a few wave length (Solodov
and Busse 2006, Figure 4a, b) http://www.ndt.net/article/ecndt2006/doc/We.2.4.2.pdf. (a) FSTM
and FSRM configuration of ACU (air coupled ultrasound wave) conversion; (b) imaging set-up,
SAW = surface acoustic wave, PAW plate acoustic
16 Delamination Detection in Wood – Based Composites Panel Products 319
substantial increase of the amplitude of the ultrasonic signal was observed in var-
ious solids under phase matching conditions for plate and surface acoustic waves.
At the phase matching between the incident wave and the plate wave an efficient
excitation of the plate wave is obtained. The obliquely incident air coupled ultra-
sound re-radiate acoustic energy in air from the rear side of the specimen. Only the
zero-order modes, symmetrical and anti symmetrical are excited. “The higher order
modes are not excited because their acoustic displacements are localised in the inte-
rior of the sample and the field is ‘locked’ inside the sample thereby preventing
access from/to the surface”. The acusto-optical laser scanning system detects the
fields of plate wave and of the surface wave. The wave field is scanned with a laser
vibrometer and an animated picture of wave propagation is obtained. Experiments
Fig. 16.11 Anisotropy of the beech veneer expressed with plate waves propagating in LT plane
(Solodov and Busse 2006, Figure 3, page 507). (a) wave front imaging mode used to put in evi-
dence the anisotropy of beech veneer with plate acoustic waves (PAW); (b) anisotropy in single-ply
and cross-ply beech with flexural waves
320 V. Bucur and S. Kazemi-Najafi
Table 16.1 Velocities of surface acoustic waves (Rayleigh waves) in wood (Döring et al. 2006)
(m/s)
with 450 kHz piezo-ceramic transducers with a focus spot of 3–4 mm show, in
wavefront imaging mode, the anisotropy of beech veneer inspected in plate wave
field (Fig. 16.11). The elongation of the wavefront in the 45◦ direction indicates the
fibres direction. It was suggested the utilisation of this method for a rapid interro-
gation of the elastic anisotropy of the material over large areas. The values of the
surface acoustic waves velocities at 400 and 200 kHz for fir, beech, balsa and oak
are given in Table 16.1. For the wood specimens, the gain in amplitude with focused
slanted transmission mode, compared with normal transmission mode is about 10
times. It is to note that “unlike the case of bulk acoustic waves which are confined
into the interior of the solid, a surface access of plate acoustic waves/surface acoustic
waves field makes them particularly attractive for direct visualisation of defects. The
wave-defect interaction causes both amplitude and phase distortions of the guided
wave field and therefore the wavefront imaging mode permits remote discerning of
a wide range of flaws”.
The applications of this new methodology proposed by Solodov and Busse
(2006) include remote imaging of defects, mapping of in-plane elastic anisotropy
and thickness measurement of paint coatings during drying.
Numerous authors have developed the theoretical background for the contact or non
contact ultrasonic transducers. A selection of references is cited here (Lynworth
1989; Hutchins and Schindel 1994; Papadakis 1999; Bharadwaj 2002; Buckley and
Loertscher 1999; Buckley 1999, 2000; Green 2004). Figure 16.12 synthesizes the
types of ultrasonic probes which can be used for the inspection of solids (water jet,
gas coupled ultrasonics, water bubbler, electromagnetic acoustic transducers and
capacitive transducers, Laser acoustic stress wave receiver)
16 Delamination Detection in Wood – Based Composites Panel Products 321
Fig. 16.12 Ultrasonic probes used for testing of solids (Green 2004, Figure 3, page 10)
For wood – based composites testing, the contact transducers are piezoelec-
tric transducers, the airborne transducers can be either piezoelectric or capacitive
transducers.
b c
Fig. 16.13 Piezoelectric ultrasonic transducer (Olympus, Panametrics 2009, Figure 11, pages
15, 44). http://www.olympus-ims.com/data/File/panametrics-UT.en.pdf (visited 16 June 2009). (a)
transversal section; (b) propagation of longitudinal waves Figure from page 44; (c) propagation of
shear
tested. The specific case of wood material having low density, requires probes with
high intensity and low frequency mostly between 20 and 200 kHz, for field experi-
ments. Higher frequency (1–3 MHz) can be used for laboratory measurements and
experiments for mechanical characterization. The low frequency signal has small
attenuation. However the wavelength of this signal limits the flaw size detectability.
To inspect in situ solid wood or wood-based composites powerful low frequency
transducers are necessary.
Fig. 16.14 Air-coupled transducers. (a) Several types of air-couples transducers (Loertscher et al.
1996). http://www.ndt.net/artcle/qmi/qmi.htm. (1) Flat air-coupled transducer, backed by air and
with a thin front layer with low impedance; (2) Focusing air-coupled transducer with spherically
shaped ceramic disk; (3) Focusing air-coupled transducer with composite ceramic element; (4)
Focusing air-coupled transducer. Flat disk with refractive lens added; (5) Focusing air-coupled
transducer. Flat disk with reflecting optics added. (b) non contact transducers for lumber inspection
in through-transmission technique (Vun et al. 2008)
the computer controls. The automatic scanning is performed by means of two tech-
niques: the object is fixed and transducers are moving or the transducers are fixed
and object is moving. Scanning can be performed with one, two or an array of trans-
ducers. Using scanning procedure the ultrasonic parameters such as time of flight,
amplitude, centre frequency, phase angle, etc. are measured for every point of the
specimen situated in the scanning plan. The data set is plotted using colours or
shades of grey to produce detailed images of the specimen in different scan modes.
The types of images which can be obtained with different scanning procedures are
given in previous table. In that follows several aspects will be discussed in more
detail.
16 Delamination Detection in Wood – Based Composites Panel Products 325
Table 16.2 Scanning modes for ultrasonic imaging (data from Stossel 2004)
Fig. 16.15 Imaging in reflection and transmission with B-scan mode (Stoessel 2004, Figure 21,
page 7). (a) reflection (b) transmission. Permission from http://elib.uni-stuttgart.de/opus/volltexte/
2004/1622/pdf/Dis_Stoessel.pdf
16 Delamination Detection in Wood – Based Composites Panel Products 327
Fig. 16.16 Imaging of an inclusion in a plate with B-scan and C-scan modes (Stoessel 2004,
Figure 22, page 78) (a) experimental (the sample, the position of the defect and of the
probes) (b) B-scan for position detection (t) (c) C-scan image. http://elib.uni-stuttgart.de/opus/
volltexte/2004/1622/pdf/Dis_Stoessel.pdf
parameters. Broadband transducers provide short pulses for high axial resolution
and transmit a wide frequency range. With receiver filters it is possible to select
frequencies which provide best defect indications. For C-and B-scans different fre-
quency spectra from one transducer can be selected: one for best lateral and another
for high axial resolution”. Figure 16.17 illustrates this statement.
Ultrasonic imaging of solid wood or wood based composites using B and C-
scan modes were reported by Niemz et al. (1996 and 1999), Schmoldt et al. (1996),
Neuenschwander et al. (1997), Berndt et al. (1999), Kabir et al. (2001), and Kabir
and Araman (2002).
Fig. 16.17 Pulse response and frequency spectra of a glassplate reflector (Hillger 1997)
http://www.ultrasonic.de/article/wsho0597/hillger/hillger.htm and http://www.ultrasonic.de/
wshop/wshop_ap/wshop_ap.phtml. (a) Pulse response of a 50 MHz PVDF-foil transducer, scale:
0.1 V/div. and 50 ns/div. (b) Frequency spectra of a glassplate reflector, with 1: Broadband
transducer, 2: with B-scan filter, 3: with C-scan filter, scale: 1 MHz/div.; 4 dB/div. The scale of the
Figure is axis x =1 MHz /div; axis y = 4 dB/div
and anisotropic solids the non-linear effect is enhanced by the presence of hetero-
geneities and defects (like delamination of layers, cracks, etc) which are strong
local sources for non linear phenomena because of the non-linear motion of their
boundaries (Solodov 1998). The motion occurs in mixed modes, in normal and
tangential directions. Because of the local disbonding compressive stress can be
transferred better than tensile stress. At the edge of the cracks for example, the
efficiency of higher harmonics generation is of many order of magnitude higher
than in “classical” nonlinear behaviour. This effect was successfully used for the
detection and location of existing and of potential defects (Solodov 2001; Strössel
et al. 2001; Solodov et al. 2004b; Solodov and Busse 2006). The group from
Stuttgart University, Germany – Professor Busse, is very active in promoting ultra-
sonic nondestructive methods based on non linear effects in different materials
including wood. Solodov et al. (2004b) reported the detection of a delamination
in a wooden rod of 21 cm long, by imaging the second harmonic with B scan
(Fig. 16.18). At about 9 cm long an open crack below the surface was detected,
with the increasing of the amplitude with 20 dB. The maximum of the second
harmonic amplitude corresponds to the crack tip. In this zone the “claping” takes
place.
In the latest four decades big efforts have been made to introduce and develop
new ultrasonic nondestructive techniques and adequate instrumentation for wood –
based composites panels inspection and characterization. A selection of refer-
ences is cited here (Burmester 1968; Beall 1987a, b, 1989; Ty 1989; Bucur
1992; Bucur 1995; Niemz et al. 1996; Bucur et al. 1998; Bekhta et al. 2000;
Vun et al. 2000; Tucker 2001; Aicher et al. 2002; Kazemi-Najafi and Bucur
2002; Vun et al. 2003; Kleinschmidt 2003; He et al. 2004; Kazemi Najafi et al.
2005; Dill-Langer et al. 2005; Hasenstab and Krause 2005; Kazemi Najafi et al.
2007). The topics of these researches have been focused mainly on physical and
mechanical characterization of panel products, using different ultrasonic tech-
niques. Special emphasis was put on the study of the influence of experimental
conditions (effect of frequency, testing configuration with contact and non con-
tact transducers, specimen geometry, moisture content, etc) on the capability of
these methods to predict the mechanical properties of panels and to detect the
presence of delamination and adhesion deficient zones. In that follows two tech-
niques will be discussed: the through transmission technique and the plate wave
technique.
a b
Fig. 16.19 Delamination detection with through transmission technique (Dill-Langer et al. 2005).
(a) Specimen and testing configuration – bondline of two glulam lamellae; (b) Parameters of the
ultrasonic signal for delamination detection; (c) delamination detection with B scan with the time
of flight and normalised amplitude http://www.ndt.net/article/v11n04/dill-langer1/dill-langer1.htm
16 Delamination Detection in Wood – Based Composites Panel Products 331
the adhesive deficient zone, it was observed a considerable increased of the time-
of-flight, which means a reduction of the velocity and a considerable reduction
of the normalized values of the amplitude attenuation of both, initial and global
amplitudes.
Furthermore the same authors performed a more complex experiment with spec-
imens of structural dimensions containing numerous glue-lines, as can be seen
Fig. 16.20a. For this purposes, two glulam blocks cut from commercially produced
beams were glued together. Each block of 900 mm width was composed from 22
lamellae of 40 mm thickness. A central circular area of 300 mm simulated a defec-
tive zone. As can be seen from Fig. 16.20b in the defective zone, the time of flight
is increasing and the normalized amplitudes are decreasing. In order to improve the
experimental condition and to minimize the measurements dispersion and to reduce
the signal to noise ratio, Dill-Langer et al. (2005) performed some comparative tests
using three coupling media such as:
• dry coupling (no coupling medium, direct contact between transducer and timber
surface)
• coupling by means of an elastomer film (transducer coupled to the elastomer film
of 0.9 mm thickness by means of silicon paste, elastomer film coupled directly to
the timber surface)
• coupling by means of silicon paste
a b
Fig. 16.20 Delamination detection on a structural specimen (Dill-Langer et al. 2005). (a) Glulam
specimen of structural dimensions (900 × 900 mm) and testing configuration; (b) Zone of
delamination between 30 and 60 cm detected with normalised time of flight and normalized
amplitude
332 V. Bucur and S. Kazemi-Najafi
For all three conditions transducers with flat ends and a diameter of 64 mm were
equally pressed to the timber by means of a frame with a compression screw. The
test results revealed that the use of coupling media lead to a very high increase of
the signal-to-noise ratio as compared to the dry coupling. This factor is 36 in case
of elastomer film and 82 in case of silicon paste. The time-of-flight showed only a
minor dependency on coupling conditions. However, it was interesting to note that
the mean time-of-flight values for dry coupling were slightly higher (8–15%) as
compared to those obtained with the more efficient coupling by film or paste. In the
case of dry contact, this fact may be explained by the less efficient contact between
the probe and the specimen which induces difficulties in proper identification of
signal parameters (because of high level of noise). The variability of time-of-flight
readings decreased considerable from dry coupling (1.5%) to much lower values in
case of film (0.75%) and paste (0.8%) coupling conditions. It was concluded that
the use of an elastomer film as coupling medium is the best compromise for optimal
coupling performance and best practical handling conditions. It is evident, that the
control of the compression force of the transducers to the specimen surface is a
prerequisite condition for measurements reproducibility.
Methodology for delamination detection in time and frequency domains was
proposed by Dimanche et al. (1994), to detect the minimum size and locate the
defect. Two models were proposed, one in time domain based on the time of flight
measurements and the other in frequency domain, using ultrasonic spectroscopy.
The minimum size of a defect was defined having a diameter Dmin , which is
defined as Dmin = Vf where V is the ultrasonic velocity in wood and f is the fre-
quency of the ultrasonic pulse. For the inspection of glulam it can be supposed that
the velocity is V = 1800 m/s and the frequency is 100 kHz. In this case the minimum
detectable defect is about 1.8 cm.
The principle of the methodology developed by Dimanche et al. (1994) is based
on the assumption that the diffraction effect takes place at the defect edge as in
Fig. 16.21 and the acoustic energy is dispersed. Depending on the relative position
of the receiver, a mask effect produced by the defect can be observed as an acoustic
shadow. The acoustic shadow is defined by the following parameters: the distance d
between the defect and the receiver and the diffraction angle μ which is a function
of velocity, frequency and probe diameter D. The “masking effect” increases with
the distance d and with the diffraction angle μ which is defined with the equation
V/f
μ = arcsin 1.22 (16.3)
D
If the defect is wider than the wavefront the signal will be not received by
the receiver. Travelling through a sample of thickness e, the size of the minimum
detectable defect can be calculated such as:
16 Delamination Detection in Wood – Based Composites Panel Products 333
Fig. 16.21 The geometry of the defect and the diffraction angle (Dimanche et al. 1994).
μ-diffraction angle, d-distance defect – receiver, D-probe diameter. (a) path wave in a sound zone;
(b) path wave in a defective zone. The diffraction takes place at the edge of the defect
V/f
Dmin ≥ 2.(e − d).tg arcsin 1.22 (16.5)
D
By combining the Eqs. (16.3 and 16.4), the minimum size of the detectable defect
can be calculated as shown in Fig. 16.22 which have the axis X expressed as a ratio
between the distance d of the defect to the receiver and the thickness of the sample e
and the axis Y as the ratio between Dmin and e. For each frequency two symmetrical
lines exist. The diagram has three main zones
– zone in which the detection is impossible – the areas under the straight line are
superimposed and neither Eq. (16.6) nor Eq. (16.4) are not satisfied
– zone of possible defect detection, but difficult, because of diffraction effect. In this
case, Eq. (16.3) is satisfied but Eq. (16.4) is not
– zone of easy defect detection, where the Eq. (16.4) is satisfied but Eq (16.3) is
not. The signal collapse will make the defect detection obvious, by observing the
attenuation of ultrasonic pulse.
The detection can be improved by increasing the frequency which means the
modification of the wavelength and by decreasing the diffraction angle.
The limitations of this methodology are:
– increasing the frequency would reduce the penetration depth of the ultrasonic
pulse
– modification of the diffraction angle requires focused probes – difficult in the low
frequency domain. The experimental inspection would be limited at the focused
region, while an increased diffraction would take place outside it.
334 V. Bucur and S. Kazemi-Najafi
Fig. 16.22 Theoretical prediction of the minimum defect size by analysing the amplitude of the
signal in time domain of beams A and B (Dimanche et al. 1994). (a) the geometry of beams A
and B; (b) minimum size defect prediction on beams A and B at 71 kHz frequency probe; Dmin
= minimium defect size, e – thickness of the sample, μ-diffraction angle, d -distance defect –
receiver, D – probes diameter
This methodology was experimented with two spruce glulam beams composed
of 10 lamellae (200 mm) with defective adhesion zones as shown in Fig. 16.22a.
The frequency was 71 kHz. The defects were located and identified for both beams
as seen in Fig. 16.23. The defective areas correspond to higher attenuations of the
received pulse (see the lower part of the graphs). This very elegant and attractive
methodology
seems
to have only one limitation introduced by the relationship:
V/f
arcsin 1.22 D < 1.
Fig. 16.23 Variation of the time of flight (μs) and attenuation (dB) as a function of beam length
for beam A and beam B on spruce. The size of the lamellae 2 mm thick and 150 mm wide, the
beam length 2 m, beam height 200 mm with 10 lamellae (Dimanche et al. 1994)
336 V. Bucur and S. Kazemi-Najafi
Fig. 16.24 Experimental device for OSB testing Vun (2003, Figure 2.1, page 16).
http://etd.lsu.edu/docs/available/etd-0708103-163628/unrestricted/Vun_dis.pdf. (a) through trans-
mission technique – transducers in contact (b) non- contact transducers
16 Delamination Detection in Wood – Based Composites Panel Products 337
net power of the ultrasound energy transmitted through the material, and the trans-
mission coefficient defined as a function of impedances in air and in the material.
These last variables provide more information than velocity on the microstructure,
texture and degree of debonding. A comparison between the direct contact tech-
nique and non contact technique is shown in Fig. 16.25 in which is plotted the
variation of velocity and attenuation as a function of density and mechanical param-
eters. The attenuation measured with non contact technique decreased as the density
increased. A minimum was reached at 900 kg/m3 . After this point an increase of
attenuation was observed. For the direct contact technique, after 900 kg/m3 , the
variation of attenuation is quite linear. The variation of attenuation versus density
is due to the modification of the internal structure of the specimens (ex: high, low
and random alignment). In the first half of the curves, probably the low density is
related to interspatial voids, rather in the second half of the curve, the increasing
plastic-strain is more related to the hardening modifications. The ultrasonic vari-
ables were correlated with the standard parameters – internal bonding (ASTM –
D1037), Young’s modulus of elasticity E1 and modulus of rupture in static bending.
It was concluded that the non contact system “provides a suitably remote measure-
ment convenience” and that this methodology is recommended for on-line quality
monitoring.
Grimberg et al. (2005) presented a laboratory methodology for the detection of
the delamination in a 3 layers composite specimen using non-contact ultrasonic
transducers. The geometry of the specimen was: section 49 × 49 cm2 , thickness
24 mm. The specimen was composed from a 20 mm thick poplar particleboard core
and a 1 mm beech veneer layer on each side of the core. The delamination was
simulated by inserting a regular Teflon frame of 80 μm thickness. The geometry
of the delaminated zones is shown in Fig. 16.26. The main characteristics of the
transducers were: 100 kHz central frequency, 25 mm the diameter of the piezo-
composite plate and 10 mm the air column. The zones with artificially induced
delamination had a section of 6 × 6 cm2 and were distributed as shown in pre-
vious Fig. 16.26a. The specimen was scanned in A- mode, with a step of 1 mm. The
reconstructed image is shown in Fig. 16.26b, on which the zones with delamina-
tion are clearly visible. It was concluded that the delamination in a multilayered
wood – based composite material can be detected using ultrasonic non contact
transducers.
Gan et al. (2005) reported a non-contact method for the detection of internal
checks in pinus radiata. Using two broad bandwidth non-contact capacitive trans-
ducers, combined with an advance technique for signal treatment. A coded chirp
signal was used to provide a specific waveform that could be post processed to
provide sufficient sensitivity for transmission across a wood sample of 34 mm thick-
ness. The signal to noise was greatly improved using two signal recovery techniques,
namely pulse compression and swept frequency multiplication. The presence of
the microcracks as well as details related to the annual ring structure can be well
observed with this imaging technique.
338 V. Bucur and S. Kazemi-Najafi
Fig. 16.25 Comparison between direct contact and non contact technique on OSB of different
alignment levels. Vun (2003, Figure 2.3, page 28). http://etd.lsu.edu/docs/available/etd-0708103-
163628/unrestricted/Vun_dis.pdf. (a) velocity and (b) attenuation versus the density of OSB. NC
and DC = non contact and direct contact methods; 1 IL = one layer board; 4% RC = 4% resin
content; HAL = high alignment level – 80%; LAL = low alignment level – 58%; RAL = random
alignment level – 26%
16 Delamination Detection in Wood – Based Composites Panel Products 339
Fig. 16.26 Debonding zones in a specimen and corresponding ultrasonic image (Greenberg et al.
2005, Figures 3 and 6a, b). http://www.ndt.net/article/wcndt00/papers/idn522/idn522.htm. Legend
specimen with debonding zones (plywood of two layers of 1 mm beech veneer, glued on the two
faces of 2 mm poplar particleboard, debonding zones simulated by inserting thin rayon foils 80 μm)
(a) imaging with time of flight with air coupled transducers 60 kHz; (b) imaging with 2D digital
Gaussian Kernel filter
The literature is very abundant on studies related to the propagation of Lamb waves
in anisotropic materials and to their interaction with defects (Alleyene and Cawley
1992; Rose et al. 1992; Potel and de Belleval 1993; Guo and Cawley 1993; Potel
et al. 1996, 2008; Chimenti 1997; Wang et al. 2001; Kundu et al. 2001; Declercq
et al. 2005, 2006). Aerospace industry developed techniques based on Lamb waves
to inspect fibre composites plates. Fibre composites and wood-based composites
have many structural similarities which can be expressed by several parameters
such as: heterogeneity, anisotropy, viscoelasticity, relatively low transverse shear
moduli, and relatively thin, layered, plate-like geometry. In addition, both materials
exhibit large attenuation coefficients in the high- frequency ultrasonic range. This
aspect forces the ultrasonic nondestructive techniques into the low frequency range.
340 V. Bucur and S. Kazemi-Najafi
a b
Fig. 16.27 Plate wave test setup (Tucker 2001, Figure 6, page 22; Figure 10, page 30;
Figure 43, page 100) https://research.wsulibs.wsu.edu:8443/dspace/bitstream/2376/53/1/b-tucker-
052101.pdf. (a) experimental plate wave setup; (b) defect detection; (c) delamination detection –
6 cm 2 area located in the centre of the MDF board, with 25 kHz transducers
thermography. Good agreement was observed between the images obtained with
both techniques.
Solodov et al. (2004b) reported a periodic delamination in a wood – based com-
posite plate constituted from oak veneer and particle board as shown in Fig. 16.29.
The image has been obtained with focused slanted transmission of air coupled ultra-
sound – forth harmonic. The delamination pattern was formed by 1.5 cm wide
periodic strips of unglued veneer. The nonlinear response of the delaminated zone
is clearly visible by a sharp local increase in the higher harmonic amplitude due to
the “clapping” mechanism.
Grimberg et al. (2000) reported a method for the delamination detection in
a multi-layer wood- based composite, using a Lamb wave technique with non
contact transducers. The geometry of the specimen was: section 49 × 49 cm2 ,
thickness 24 mm. The specimen was composed from a 20 mm thick poplar par-
ticleboard core and a 1 mm beech veneer layer on each side of the core. The
342 V. Bucur and S. Kazemi-Najafi
Fig. 16.29 Forth harmonic image with focused slanted transmission of air coupled ultrasound of
periodic delaminations between oak veneer lamina and particle board plate (Solodov et al. 2004b,
Figure 11, Page 510
16 Delamination Detection in Wood – Based Composites Panel Products 343
a b
Fig. 16.30 Industrial equipment for delamination detection on LVL (courtesy of Airstar)
(https://airstar1.com/BLOW%20DETECTOR/.htm)(a) transversal view of the equipment for test-
ing of LVL and air coupled ultrasonic transducers; (b) frontal view of the same equipment with
Power Sonic Resonance Technology (https://ews-usa.com/images/ewsproducys/eval.ofwood.pdf)
16.4 Summary
Wood – based composite panel products are manufactured from veneer, wood par-
ticle, strands or fibres bind together with different type of adhesives. Delamination
is one of the most important defect which occurs in these products. The aim of
this chapter is to review the ultrasonic techniques used for delamination detection.
Delamination induces modifications of the mechanical and elastical properties of
the materials, which can be observed with ultrasonic techniques. Ultrasonic inspec-
tion involves the utilisation of stress waves having a frequency higher then 20 kHz.
Linear and non linear ultrasonic techniques have been developed for ultrasonic
inspection of wood based composites. Linear techniques have been developed in the
hypothesis that the acoustic wave amplitude is infinitesimally small and the response
of the material is assumed to be linear tp the excitation signal. Hook law is valid.
Under the label of –linear ultrasonic techniques- three main groups of techniques
are recognized: the reflexion technique or the pulse echo technique, the transmis-
sion technique, and the emission technique. This last technique is not described
in this chapter. For ultrasonic signal transmission to the specimens, contact and non
contact transducers called also air coupled transducers, can be used. Bulk waves and
Lamb waves (plate waves) are used for the mechanical and elastical characteriza-
tion. Technological advancements with non contact transducers have made possible
the development of studies related to the non linear behaviour of materials. For
wood-based composites testing the contact transducers are piezoelectric, rather the
air coupled transducers can be either piezoelectric or capacitive transducers. Images
of panels’ internal structure can be obtained with different scanning procedure. The
most common modes are: A-scan, B-scan and C scan. The non linear behaviour
of anisotropic materials can be observed by the increasing number of higher har-
monics having increasing amplitude with the distance. In these materials the non
linear effect is enhanced by the presence of defects, because of non linear motion
of their boundary. Delamination detection in wood-based composites was studied
16 Delamination Detection in Wood – Based Composites Panel Products 345
with through transmission technique and with plate wave technique. In both cases
the transducers can be in direct contact with the specimen or can be air coupled.
Laboratory experiments have been reported for the delamination detection in clear
specimens or in specimens of structural size. An industrial non contact technique
was patented for the detection of delamination in MDF OSB OSL LVL LVP, etc.
The principle of this technique is based on the fact that a resonance occurs at the
frequency at which an integral number of half-waveslengths fit into some dimen-
sions of the workpiece, which vibrates at its natural frequencies. The presence of a
delamination modifies the frequency. A real time picture is displayed on each work-
piece and is used for further operations. The main advantage of this technique is its
capability to sample almost the entire workpiece. The air coupled transducers are
encapsulated and protected against dust, noise, etc. the non contact technology for
delamination detection is very robust, inherent safety and cost effective.
References
Aicher S, Dill-Langer G, Ringger T (2002) Non-destructive detection of longitudinal cracks in
glulam beams. Otto-Graf-J 13:165–182
Alleyene DN, Cawley P (1992) The interaction of Lamb waves with defects. IEEE Trans
Ultrasonics Feroelectric Freq Control 39:381–397
Auld BA (1973) Acoustic fields and waves in solids, vol I and II. Wiley, New York
Aymerich F, Meili S (2000) Ultrasonic evaluation of matrix damage in impacted composite
laminates. Composites: Part B 31:1–6
Bar-Cohen Y, Chimenti DE (1986) Nondestructive evaluation of composite laminates by leaky
Lamb waves. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative NDE,
vol l. 5B. Plenum Press, New York, pp 1199–206
Beall FC (1987a) Fundamentals of acoustic emission and acousto-ultrasonics. In: Proceedings of
6th International Symposium Nondestructive Testing of Wood, Washington State University,
U. S Pullman, WA. 1987 September 14–16, pp 3–22
Beall FC (1987b) Acousto – ultrasonic monitoring of glueline curing. Wood Fiber Sci 19:204–214
Bekhta PA, Niemz P, Kucera, LJ (2000) The study of sound propagation in the wood-based com-
posite material. In: Proceeding of 12th International Symposium on Nondestructive Testing of
Wood, Sopron, Hungary, Sopron, 13–15 September 2000, pp 33–41
Berndt H, Schniewind AP, Johnson GC (1999) High resolution ultrasonic imaging of wood. Wood
Sci Techn 33:185–189
Bharadwaj MC, Neeson I, Stead G (2000) Introduction to contact free ultrasonic characteri-
zation and analysis of consolidated materials. In: Proceedings of Application of NDE in
powder metals. Invited paper Iowa State University, NDT net-june 2000, Vol 5, No. 06,
http://www.ndt.net/article/v05n06/bhardwaj/bhardwaj.htm
Bharadwaj MC (2002) Nondestructive evaluation: introduction of non-contact ultrasound. In:
Schwartz M (ed) Encyclopedia of smart materials. Wiley, New York, pp 690–714
Birks AS, Green RE Jr.(1991). Ultrasonic testing. vol 7, 2nd edn. In: McIntire P
(ed) Nondestructive testing handbook. American Society for Nondestructive Testing,
Columbus, OH
Bjorno L (2001). Forthy years of non-linear ultrasound. Ultrasonics 40:11–17
Bland DR (1988) Wave theory and applications Oxford University Press, New York
Blomme E, Bulcaen D, Declerc NF (2002) Recent observations with air coupled NDE in the
frequency range of 650 kHz to 1.2 MHz. Ultrasonics 40:153–157
Bray D E, Stanley RK (1997) Nondestructive evaluation: a tool in design, manufacturing and
service. CRC Press, New York, NY
346 V. Bucur and S. Kazemi-Najafi
Buckley J (1999) Principles and applications of air-coupled ultrasonics. INSIGHT 40, 11:755–759
Buckley J (2000) Air-coupled Ultrasound – A millennial review. In: Proceedings of the 15th
World Conference of Non-Destructive Testing, Roma, Italy, 15–21 October 2000, Available
on http://www.ndt.net/article/wcndt00/papers/idn507/idn507.htm
Buckley J, Loertscher H (1999) Frequency considerations in air-coupled ultrasonic inspection. In:
British Institute of Non-Destructive Testing (BINT), September 1999, UK Cheshire, Plenum
Press, 41(11):696–699, Available on: http://joe.buckley.net/papers/BINDT99%20Airscan.PDF
Bucur V (1992) Anisotropy characterization of structural flackboard with ultrasonic methods,
Wood Fiber Sci 24(3):337–346
Bucur V (1995) Acoustics of wood. CRC Press, Boca Raton, FL, pp 266
Bucur V (2003a) Nondestructive characterization and imaging of wood, Springer, Heidelberg,
p 354
Bucur V (2003b) Techniques for high resolution imaging of wood structure: a review. Meas Sci
Technol 14:R 91–R98
Bucur V, Ansell M P, Barlow C Y, Prithard J, Garros S, Deglise X (1998) Physical methods for
characterization wood composites panel products, Holzfurchung 52(5):553–561
Burmester A (1968) On the correlation between sound velocity and density, tensile strength
perpendicular to the surface and bending strength of particleboard. Holz als Roh werkstoff
113–117
Castaings M, Cawley P (1996) The generation, propagation, and detection of Lamb waves in plates
using air-coupled ultrasonic transducers. J Acoust Soc Am 100(5):3070–3077
Castaings M, Cawley P, Fralow R, Hayward G (1998) Single side inspection of composite materials
using air-coupled ultrasound. J Nondestructive Eval 17:37–45
Chimenti DE (1997) Guided waves in plates and their use in materials characterization. Appl Mech
Rev 50(5):247–284
Dang C, Schmerr LW, Sedov A (2002) Ultrasonic transducer sensitivity and model-based
transducer characterization. Research. Nondestructive Eval 14(4):203–228
Declercq NF, Degrieck J, Leroy O (2005) Bounded beam interaction with plate-edge at Lamb
angle. Acta Acustica united with Acustica 91(2):326–332
Declercq NF, Degrieck J, Leroy O (2006) Ultrasonic polar scans: numerical simulation on
generally anisotropic media. Ultrasonics 45(1–4):32–39
Delsanto PP (2007) Universality of nonclassical nonlinearity. Applications to non-destructive
evaluations and ultrasonics. Springer, New York
Dill-Langer G, Bernauer W, Aicher S (2005) Inspection of glue-lines of glued-laminated timber
by means of ultrasonic testing. In: 14th International Symposium on Nondestructive Testing on
Wood, Hanover, Germany, pp 49–60
Dill-Langer G, Bernauer W, Aicher A (2006) Inspection of glue-lines of glued laminated timber by
means of ultrasonic testing. NDT.net 11(4) http://www.ndt.net/article/v11n04/dill-langer1/dill-
langer1.htm
Dimanche M, Capretti S, Del Senno M, Facaoaru I (1994) Validation of a theoretical approach
for the detection of delamination in glued laminated beams. In: proceedings of 1st European
symposium on nondestructive evaluation of wood, Sopron, pp 250–260
DIN EN 1330 – 4 (2000) Non-destructive testing – terminology Part 4: terms used in ultrasonic
testing. Trilingual version DIN – Deutsches Institut für Normung Beuth Verlag GmbH Berlin,
Berlin
Döring D, Pfleiderer K, Solodov I, Busse G (2006) Automated NDT of stiffness anisotropy with
air-coupled ultrasound. ECNDT 2006, p 8. http://www.ndt.net/article/ecndt2006/doc/P123.pdf.
Accessed 7 November 2009
Electronic Wood Systems North America (2006) Evaluation of wood-based panel proper-
ties and delamination detection by means of ultrasound, based on current technology on
online sonic measurements. See www.ews-usa.com and Electronic Wood Systems GmbH,
Germany www.electronic-wood-systems.de and http://ews-usa.com/index.php?option=com_
content&task=view&id. Accessed 15 March 2006
16 Delamination Detection in Wood – Based Composites Panel Products 347
Gan TH, Hutchins DA, Green RJ, Andrews MK, Harris PD (2005) Noncontact, high-resolution
ultrasonic imaging of wood samples using coded chirp waveforms. IEEE Trans Ultrason
Ferroelect Freq Control 52(2):280–288
Gorman MR (1991) Ultrasonic polar backscatter imaging of transverse matrix cracks, J Compos
Mater 25:1499–514
Graff K (1975) Wave motion in elastic solids. Ohio State University Press, Columbus, OH
Grandia WA, Fortunko CM (1995) NDE applications of air-coupled ultrasonic transduc-
ers. In: Proceedings of IEEE International Ultrasonics Symposium, Seattle, WA, USA,
pp 697–709
Green RE Jr. (1973) Ultrasonic investigation of mechanical properties, vol 3 of treatise on materials
science and technology. Academic, New York
Green RE Jr. (1999) Application of non-contact ultrasonics to nondestructive characterization of
materials. In: Green RE Jr (ed) Nondestructive characterization of materials, vol 10, Kluver
Academic, Plenum Press, New York, pp 419–420
Green RE Jr. (2004) Non-contact ultrasonic techniques. Ultrasonics 42:9–16
Grimberg R, Savin A, Lupu A, Iancu L, Rotundu C (2000) A method to determine the debond-
ing zones in multilayer wood materials. In: Proceedings of the 15th world conference of
non-destructive testing, Roma, Italy, 2000, IDN 522
Grimberg R, Savin A, Steigmann R, Bruma A (2005) Ultrasound and visual examination of wood
based products, In: The 8th international conference of the slovenian society for non-destructive
testing -application of contemporary non-destructive testing in engineering, September 1–3,
Portorož, Slovenia, pp 109–115
Guo N, Cawley P (1993) The interaction of Lamb waves with delaminations in composite
laminates. J Acoust Soc Am 94:2240–2246
Haberger CC, Mann RW, Baum GA (1979) Ultrasonic plate waves in paper. Ultrasonics 17:
57–62
Han Jun-Bo, Cheng JC, Wang TH, Berthelot Y (1999) Mode analyses of laser- generated tran-
sient ultrasonic Lamb waveforms in a composite plate by wavelet transform. Mater Eval 57(8):
837–840
Hasenstab A, Krause M (2005) Defect localisation in wood with low frequency ultrasound echo
technique. In: proceedings of 14th international symposium on nondestructive testing of wood,
Hannover, pp 329–336
Hasenstab A, Krause M, Osterloh K (2006) Testing of wooden construction elements with
ultrasonic echo technique and X –ray. ECNDT 2006 –Th.2.4.1:1–8
http://www.ultrasonic.de/article/ecndt2006/doc/Th.2.4.1.pdf
Hayward G, Gachagan A (1996) An evaluation of 1–3 connectivity composite transducers for
air-coupled ultrasonic applications. JASA 99(4):2148–2157
Hayward G (1997) Air-coupled NDE: constraints and solutions for industrial implementations. In:
Proceeedings of IEEE Ultrasonics Symposium, Toronto, Ontario, Canada, pp 665–673
He Y, Manful D, Bardossy A, Dill-Langer G, Ringger T, Aicher S (2004) Fuzzy logic based de-
noising of ultarsound signals from non-destructive testing, Otto-Graf-J 15:103–120
Hillger W (1992) Ultrasonic imaging of damages in CFRP-laminates. Acoust Imaging 19:575–579
Hillger W (1997) Ultrasonic imaging of internal defects in composites. NDTnet 2(5).
Paper at UTonline application workshop May ’97. http://www.ultrasonic.de/article/
wsho0597/hillger/hillger.htm and http://www.ultrasonic.de/wshop/wshop_ap/wshop_ap.phtml.
Accessed 12 June 2007
Hosten B (1991) Reflection and transmission of acoustic plane waves on an immersed orthotropic
and visoelastic solid layer. JASA 89:2745–2752
Hosur MV, Murthy C RL, Ramamurthy TS, Shet A (1998) Estimation of impact-induced damage
in CFRP laminates through ultrasonic imaging. NDT E Int 31(5):359–374
Huang W (1998) Application of Midlin plate theory to analysis of acoustic emission waveform in
finite plates. In: Thompson DO, Chimenti DE (eds) Proceedings 24th Annual Symposium on
Quantitative NDE, July 27–August 1, 1997, San Diego, CA, Plenum Press
348 V. Bucur and S. Kazemi-Najafi
Huang W, Ziola SM, Dorighi JF, Gorman MR (1998) Stiffness measurement and defect detection in
laminated composites by dry-coupled plate waves. In: Bossi RH, Pepper DM (eds) Proceedings
of SPIE,. March 31–April 2, San Antonio, TX, USA
Hurley V (2008) A successful case study of R&D, commercialization and adoption of new technol-
ogy by a wood products company. In: Wood Innovation 2008 Proceedings 18–19 September,
Melbourne: 182–183 (www.fridayoffcuts.com )
Hutchins DA, Schindel DW (1994) Advances in non-contact and air – coupled transducers.
In: Proceedings IEEE Ultrasonic Symposium 1994, Electronic Identifier 10.1109/ULTSYM.
1994.40181,1–4 November 1994, pp 1245–1254
Kabir M F, Araman P A (2002) Nondestructive evaluation of defects in wood pallet parts by ultra-
sonic scanning. In: Proceedings of 13th international symposium on nondestructive testing of
wood, California, Berkley, USA
Kabir MF, Schomoldt DL Shafer ME (2001) Roller-transducer scanning of wooden pallet parts
for defect detection. In: Thompson DO, Chimenti DE (eds) Review of Progress in Quantitative
NDE, vol 20. pp 1218–1225
Karim MR, Mal AK, Bar-Cohen Y (1990) Determination of the elastic constants of composites
through the inversion of leaky Lamb wave data. In: Thompson DO, Chimenti DE (eds) Review
of Progress in Quantitative NDE, vol 9A, Plenum Press, New York, pp 109–116
Kazemi Najafi S, Bucur V, Ebrahimi G (2005) Elastic constants of particleboard with ultrasonic
technique. Mater Lett 59:2039–2042
Kazemi Najafi S, Abbasi Marasht A, Ebrahimi G (2007) Prediction of ultrasonic wave velocity in
particleboard and fiberboard. J Mat Sci 42:789–793
Kazemi Najafi S, Bucur V (2002) Nondestructive characterization of particleboard with acoustic
methods. In: Proceedings of 6th conference of French acoustical Soceity, April 2002, Lille,
France
Kleinschmidt H (2003) Ultra-Scan delamination detection with new power sonic resonance tech-
nology increases panel board production. In: Proceedings of 7th European panel products
symposium, North Wales Conference Center, Llandudno UK - October 2003, pp 111–115
Kolsky H (1963) Stress waves in solids. Dover, New York, NY, USA
Krautkrämer J, Krautkrämer H (1990). Ultrasonic testing of materials, 4th edn. Springer, Berlin
Kundu T, Potel C, de Belleval JF (2001) Importance of the near Lamb mode imaging of
multilayered composite plates. Ultrasonics 39:283–290
Lionetto F ; Tarzia A ; Maffezzoli A (2007) Air-coupled ultrasound :a novel technique for mon-
itoring the curing of thermosetting matrices. IEEE Trans Ultrason Ferroelect Freq Control
54:1437–1444
Loertscher H, Grandia B, Strycek J, Grandia W (1996) Airscan transducers, techniques and
applications. NDTnet 1(9):4. http://www.ndt.net/article/qmi/qmi.htm
Lynworth LC (1989) Ultrasonic measurement for process control. Theory, techniques, applications.
Academic, Boston, MA
Martin RW, Chimenti DE (1987) Signal processing of leaky Lamb wave data for defect imaging in
composite laminates. In: Thompson DO, Chimenti DE (eds) Review of Progress in Quantitative
NDE, vol 6A, Plenum Press, New York, pp 815–824
Moran TJ, Crane RL, Andrews RJ (1985) High-resolution imaging of microcracks in composites,
Mater Eval 43:536–540
Musgrave MJP (1970) Crystal acoustics. Holden-Day, San Francisco, CA
Nayeh AH (1995) Wave propagation in layered anisotropic media: with applications to composites.
Elsevier, New York
Niemz P, Kucera L J, Poblete H, Baradit E(1996) On the application sound velocity to the determi-
nation of bending strength of particleboard. In: Proceedings of 10th international symposium
on nondestructive testing of wood, Lausanne, Switzerland
Niemz P, Kucera L J, Schob M, Scheffler M (1999) Possibility of defect detection in wood with
ultrasound. (Experimentelle Untersuchungen zur Erkennung von Defekten in Holz mittels
Ultraschal). Holz als Roh und Werk 57:96–102
16 Delamination Detection in Wood – Based Composites Panel Products 349
Steiner KV, Eduljee RF, Huang X, Gillespie JW Jr (1995) Ultrasonic NDE techniques for the
evaluation of matrix cracking in composite laminates. Compos Sci Technol 53:193–198
Stössel R, Krohn N, Pfleider K, Busse G (2001) Air-coupled ultrasound inspection of various
materials. Ultrasonics 40:159–163
Stössel R, Krohn N, Busse G (2000) Measurements with air-coupled ultrasound. In: Proceedings
of 7th Internat. Congres on Sound and Vibration, July 4–7, Garmish – Partenkirchen vol II:
795–802 and Acoust Phys J (2002) 48(3):159–163
Stössel R (2004) Air-couples ultrasound inspection as a new non-destructive testing tool for quality
assurance. PhD Thesis, University of Stuttgart, Germany
Strycek JO, Loertscher H (1999) Ultrasonic air-coupled inspection in advanced material. NDT.net
4:46–50
Strycek JO, Grandia WA, Loertscher H (1997) Wave modes produced by air coupled ultrasound.
NDTnet May 1997, 2(5). http://www.ndt.net/article/wsho0597/qmi2/qmi2.htm. Accessed 10
September 2008
Tang B, Henneke EG II (1989a) Lamb wave monitoring of axial stiffness reduction of laminated
composite plates. Mater Eval 47(8):928–932
Tang B, Henneke EG II (1989b) Long wavelength approximation for Lamb wave characterization
of composites laminates. Res Nondestructive Eval 1:51–64
Tang B, Henneke EG II, Stiffler RC (1988) Low frequency flexural wave propagation in laminated
composite plates. In: Duke JC Jr (ed) Acousto – ultrasonics: theory and application. Plenum
Press, New York, NY, pp 45–65
Tucker BJ, Bender DA, Pollock DG, Wolcott MP (2003a) Ultrasonic plate waves evaluation of
natural fiber composite panels. Wood Fiber Sci 35:266–281
Tucker BJ, Bender DA (2003) Continuous ultrasonic inspection of extruded wood-plastic compos-
ites. Forest Products J 53(6):27–32
Tucker BJ (2001) Ultrasonic plate waves in wood-based composite panels. Ph. D Dissertation,
Department of Civil and Environmental Engineering, Washington State University,
p 112
Ty Ch (1989) Modulus of elasticity of particleboard determined by nondestructive testing methods.
J Agric For 38(2):151–164
Viktorov I A (1967) Rayleigh and Lamb waves: physical theory and applications. Plenum Press,
New York, NY
Vun RY (2003) Ultrasonic characterization of engineering performance of oriented strand-
board. Louisiana State University – etd (electronic thesis and dissertations), http://etd.lsu.edu/
docs/available/etd-0708103-163628/
Vun RY, Wu Q, Bhardwaj MC, Stead G (2003) Ultrasonic characterization of structural properties
of oriented strandboard: a comparison of direct-contact and non-contact methods. Wood Fiber
Sci 35(3):381–396
Vun RY, Wu Q, Monlezun CJ (2003) Ultrasonic characterization of horizontal density variations
in oriented strandboard. Wood Fiber Sci 35(3):482–498
Vun RY, Hoop C, Beall FC (2005) Monitoring critical defects of creep rupture in oriented
strandboard using acoustic emission: incorporation of EN300 standard. Wood Sci Techn
39(3):199–214
Vun RY, Hoover K, Janowiak J, Bhardwaj M (2008) Calibration of non-contact ultrasound as
an online sensor for wood characterization: effects of temperature, moisture, and scanning
direction. Appl Physics A Mat Sci 90(1):191–196
Vun YR, WuQ, Bhardwaj M, Stead G (2000) Through – thickness ultrasonic transmission proper-
ties of oriented strandboard. In: Proceedings of 12th international symposium on nondestructive
testing of wood, Sopron, pp 76–68
Wang CS, Wu F, Chang FK (2001) Structural health monitoring from fiber reinforced composites
to steel reinforced concrete. Smart Mat Struct 10:548–552
Wooh SC, Daniel IM (1990) Enhancement techniques for ultrasonic nondestructive evaluation of
composite materials. J Eng Mat Tech 112:175–182
16 Delamination Detection in Wood – Based Composites Panel Products 351
Wooh SC, Daniel IM (1994) Three-dimensional ultrasonic imaging of defects and damage in
composite materials. Mater Eval 52(10):1199–1206
Wooh SC, Wei C (1999) A high-fidelity ultrasonic pulse-echo scheme for detecting delaminations
in composite laminates. Composites: Part B 30:433–441
Žukauskas E, Cicėnas V, Kažys R (2005) Application of air–coupled ultrasonic technique for sizing
of delamination type defect in multilayered materials. Ultragarsas 54(1):7–11
Chapter 17
Delamination Evaluation of in-Service Glulam
Beams and other Structural Members
Via Ultrasonics
Ferenc Divos
Contents
17.1 Introduction
F. Divos (B)
Faculty of Wood Science, University of West Hungary, Sopron, Hungary
e-mail: divos@fmk.nyme.hu; divos@fakopp.com
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 353
DOI 10.1007/978-90-481-9550-3_17, C Springer Science+Business Media B.V. 2011
354 F. Divos
on a school holiday, died when the roof of a skating rink fell in the Bavarian town
of Bad Reichenhall, after heavy wet snow falls (Associated Press 2006).
Since the tests should be performed in the field, simple devices must be used
and the experimental processes modified to improve the workability and efficiency
of tests. A preliminary inspection phase must be followed by an advance phase
of inspection. In the preliminary phase, a comprehensive in situ assessment of the
structure in its current condition must be made via a visual inspection. This visual
inspection defines damage severity and extends of deterioration, having in mind that
the engineering challenge is to assess member integrity. The first symptoms that can
be detected are surface cracks. (The biological degradations are not discussed in this
chapter). In the advance phase, the methods and procedures to be used for inspection
are defined, and the experimental results analyzed. This phase must be ended with
practical recommendations.
One of the most frequent observed defects are the cracks. The primary charac-
teristic of a crack is the depth. The common test for crack depth determination is
the penetration depth of 0.1 mm thick feller gauge (Fig. 17.1). This simple tech-
nique provides useful information about the crack, but has its limitation, especially
when the crack path is not straight (Fig. 17.2). Due to the above limitation in testing,
In glulam structures, two main crack types are recognised, the crack between the
lamellae and the cracks inside the lamellae material.
In the past, glulam elements containing pith have often been used by some manu-
facturers, typically at the central zone of the beam on which the mechanical loading
corresponds to the neutral axis. However an extreme case of utilization of a lamellae
is shown in Fig. 17.5a. Where the outer laminates are pith – containing. Cracks are
356 F. Divos
Fig. 17.5 Cracks in lamellae with piths zones. (a) two lamellae with pith in transversal section;
(b) cracks due to the pith at beam surface
also clearly visible on the surface (Fig. 17.5b). Note that the crack propagates to
the pith.
The beam ends are particularly vulnerable to initiation and growth of delamina-
tion because the dowels or screws of the fastening can be a crack initiator. A crack
having such origin is shown in Figs. 17.5 and 17.6.
Cracks form a material discontinuity and as such are a vibration propagation barrier.
Ultrasonic or stress wave tools are applicable for the crack depth determination.
17 Delamination Evaluation of in-Service Glulam Beams 357
Placing a vibration source close to the crack line and a receiver sensor on the
opposite side of the crack, the transit time between the emitter and receiver give us
information about the crack depth. The recommended distance between sensors is
2–5 cm. To compute the crack depth it is also necessary to know the transit time
in the intact material (see: material without crack). The following equation gives us
the crack depth (L)
where:
Tc transit time across crack,
Ti transit time in case of intact material,
D Distance between sensors,
V P-wave velocity perpendicular to fibers in intact material.
Figure 17.7 shows two examples for the measurement of the time of flight of
the ultrasonic pulse. The first one uses ultrasonic surface waves. In this case a
pair of ultrasonic probes (1 MHz) is displayed at 45◦ and pressed to the surface
a) b)
Fig. 17.7 Crack depth determination via acoustic technology. (a) with ultrasonic transducers at
45◦ (b) with stress wave sensors
358 F. Divos
of the beam. The second example shows the probes used for stress wave technique
(40 kHz), which uses a hammer and an amplification system to excite vibration into
the tested element. It is easy to understand that with ultrasonic techniques, which are
more precise, the value corresponding to crack depth measurement is greater than
that determined using feeler gauge penetration. The ultrasonic techniques work very
well for normal cracks. Including all error sources the relative error of crack depth
determination is around 10%. The ultrasonic techniques are strongly limited if the
crack is filled by paint.
MOE = ρV 2 (17.2)
where:
ρ density
V P-wave velocity
In-situ determination of density is however rather difficult but possible using
gamma or X rays. In practice the density values cited in the literature, for given
species is suitable for computation of MOE. The velocities can be measured rela-
tively easily and it is recommended to limit the distance between the two probes at
20 times the lamella thickness. Figure 17.8 shows the P-wave velocity measurement
using an ultrasonic device. A plastic – PVC – bar connects the sensors, helping to
keep the distance constant between during measurements.
The shear strength between the lamellae provides critical information about the glu-
lam beam mechanical capacities. Predicting the shear strength of the glue layer via
nondestructive techniques is possible but quite difficult as can be seen in Fig. 17.10.
17 Delamination Evaluation of in-Service Glulam Beams 359
a) b)
Fig. 17.9 P-wave velocity determination perpendicular to the grain. (a) the device (b) defect
location on the time of flight map. Dark (red) spot shows the location of the defective region
Shear sensors equipped with a knife type wave guide are applied to the opposite
faces of adjacent elements The plane of the knife determines the shearing plane.
This plane is prallel to the glue layer. In this experimental situation there were
no glue in between the lamellae. It was suggested that (Fig. 17.11) the atten-
uation measurements via shear wave amplitude can be a potential tool for the
360 F. Divos
a) b)
Fig. 17.10 Shear transducers. (a) The position of shear transducers on the transversal section of
the laminar structure; (b) the probes
Hungary abounds in historical buildings, several of which have wooden roof and
ceiling structures. Most of these buildings are in a run-down state, needing reno-
vation. Wood experts evaluate the bio-degradation of wood, identifying fungi and
insect attack by visual inspection, by touching the material or using a simple screw
driver.
In this chapter the effectiveness of the ultrasonic technique, allowing the time of
flight measurements is demonstrated on structural elements and on the ceiling struc-
ture of an old baroque castle – Esterhazy Castle – in Pápa in Hungary (Fig. 17.12).
17 Delamination Evaluation of in-Service Glulam Beams 361
a) b)
Fig. 17.12 Esterhazy Casel in Papa, Hungary. (a) genral view of the restored casel, (b) plane view
of the first flor
The first floor of the two storied, U-shaped building is more than 500 years old, and
had been used as a fortress until 1752. From the original structure, the builders have
kept the walls of the ground floor. The length of the building is 165 m. The roof and
the top ceiling structure (a system of dowelled, closely aligned beams) is made on
wood (larch, oak, spruce and lime). The first floor and the walls of the second floor
are made on stone and bricks, and have been built probably during the 18th century.
The ceiling structure of the first flor is different from that of thesecond one, the first
being vaulted, and the second composed of closely fitted, doweled beams.
The inspection of the structure started with the visual examination of the sound
and decayed zones of all elements. The second phase of the inspection with ultra-
sonic device and micro drilling device are commented in this chapter and are related
to the physical and mechanical tests of the Sections 16 and 38. Figure 17.13 shows
some aspects during in situ measurements For the ultrasonic test, the piezoelectic
transducers of the device-FAKOPP timer, have been equipped with 60 mm long nails
to facilitate the inspections of wooden beams. Using a long cable of 11 m lenhght,
all beams have been measured without difficulty. The test for velocity measurements
is fast, two experienced person can carry out the test within 30 s.
To obtain more data about the physical state of the inner layers of the beams,
micro drilling technique was used. The screw withdrawal force measurements
(Fscrew ) were performed for 5 mm diameter drill to 120 mm depth. The consis-
tency and the odor of wood particles falling out from the hole were also analyzed
to state about the wood quality of the inspected beams. For the inspected beams,
the strength predictor parameters are stress wave velocity V and screw withdrawal
resistance. The predictor coefficient was calculated as Fscrew · V2 . It was demon-
strated previously (Divos and Tanaka 1997, Divos et al. 1998, 1999) that screw
withdrawal resistance is well correlated shear modulus, withdrawal force and with
density. This suggested to use the empirical relationship σ = FCS v2 (similar to
E = ρv2 ). Moreover, Kollmann (1965) noted a strong relationship between modu-
lus of elasticity and modulus of rupture in bending of full size beams. Using the
screw withdrawal force and the velocity of stress wave the following empirical
strength predictor equation applied for coniferous wood species was derived. The
362 F. Divos
a)
b)
applied units in the equations are: MORest [MPa], Fscrew [kN] and v, the velocity
[km/s]:
The correlation coefficients between the bending strength and MORest is 0.74.
Figure 17.14 shows the correlation between modulus of rupture in bending of fill
140
120
MOR [Mpa]
100
80
Fig. 17.14 Relationship
between the modulus of 60
rupture and the predictor
parameters. represents 40
coniferous, ∇ represents 40 60 80 100 120 140
hardwood specimens Predictor [MPa]
17 Delamination Evaluation of in-Service Glulam Beams 363
Fig. 17.15 Strength distribution over the 214 years old ceiling in larch – a system of doweled,
closely aligned beams. The numbers represent the residual stress bending strength in MPa
size specimens and the predictor parameters.. The residual bending strength of indi-
vidual wooden beams have been estimated with ± 9 MPa accuracy. Figure 17.15
shows the strength distribution in the larch ceiling.
17.8 Summary
the shear strength of the glue layer via nondestructive techniques is possible using
shear probes equipped with a knife type wave guide. The plane of the knife deter-
mines the shearing plane. This plane is parallel to the glue layer. It was suggested
that the attenuation measurements via shear wave amplitude can be a potential tool
for the evaluation of delamination between glued layers for in –situ beams.
References
Associated Press (2006) German ice rink collapse. AP Image January 3, 2006 by Diether Endlicher,
www.highbeam.com. Accessed 3 August 2010
Brashaw BK, Vatalaro RJ, Wacker JP, Ross RJ (2005) Condition assessment of timber bridges.
1. Evaluation of a micro-driling resitance tool. FPL – GTR 159 USDA Forest Service Forest
Products Laboratory. Madison WI, USA
Divos F, Divos P, Divos G (2007) Acoustic techniques: from seedling to wood structures.
Proceedings of the 15th international symposium on nondestructuctive testing of wood. Duluth,
MN, pp 3–12
Divos F, Nemeth L, Bejo L (1999) Evaluation of the wooden structure of a baroque place in Papa,
Hungary. Proceedings of the 11th international symposium on nondestructive testing of wood.
Lausanne, Suisse, pp 153–160
Divos F, Tanaka T (1997) Lumber strength estimation by multiple regression, Holzforshung
51:467–471
Divos F, Tanaka T, Nagao H, Kato H (1998) Determination of shear modulus on construction size
timber. Wood Sci Technol 32:393–402
Kollmann, F. 1965 Relationship between elasticity and bending strength of wood, Proceedings of
the 2nd symposium on nondestructive testing of wood. Spokane, WA
Lee SJ, Oh JK, Yeo H, Lee JJ, Kim KB, Kim KM (2007) Field application on nondestructive
testing for detecting deterioration in Korean historic wood buildings. Proceedings of the 15th
international symposium on nondestructuctive testing of wood. Duluth, MN, pp 227–232
Pellerin RF, Ross RJ (2002) Nondestructive evaluation of wood. Forest Products Society,
Madison, WI
Ross RJ, Pellerin RF (1994) Nondestructive testing of assessing wood members in structures,
USDA, Forest Products Laboratory, FPL-GTR-70, Madison, WI
Sandoz JL, Benoit Y (2007) Acousto-ultrasonic nondestructive evaluation of historic wood struc-
tures. Proceedings of the 15th international symposium nondestructuctive testing of wood.
Duluth, MN, p 245
Tanaka T, Divos F, Fazan, T (1998) Nondestructive evaluation of residual bending strength of
wood with artificial defects by stress wave, Proceedings of the 11th international symposium
on nondestructive testing of wood, Madison, WI
Wacker JP, Wang X, Ross RJ, Brashaw BK (2007) Condition assessment of historic vessels.
Proceedings of the 15th international symposium nondestructuctive testing of wood. Duluth,
MN, pp 223–226
Wang X, Wacker JP (2006) Condition assessment of main structural members of US Brig Niagara.
Final Report Project no 187–2419. Erie Maritime Museum, Erie, PA
Chapter 18
Moisture Induced Stresses and Deformations
in Parquet Floors
Samuel Blumer, Erick Serrano, Per Johan Gustafsson, and Peter Niemz
Contents
18.1 Introduction
During the last decade the use of wood flooring systems in Europe has increased
dramatically. In Sweden for example, the proportion of wood flooring systems rose
steadily from 30% in the seventies to its current proportion which is 80%. This
rapid growth has fostered the development of new products, enabling the industry
to maintain and increase its market share.
The main objective of this paper is to improve understanding of the behaviour
of parquet floors exposed to different climates by applying numerical analysis
techniques using the commercial finite element program ABAQUS. In addition,
S. Blumer (B)
b-h-e GmbH, Holzinnovationszentrum 1a, 8740 Zeltweg, Austria
e-mail: samuel.blumer@b-h-e.at
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 365
DOI 10.1007/978-90-481-9550-3_18, C Springer Science+Business Media B.V. 2011
366 S. Blumer et al.
Bernoulli’s beam theory applied in two dimensions to the square specimens gives
a simple and additional validation instrument for determining the deformational
behaviour of parquet flooring (Bodig and Jayne 1982). This approach is comple-
mented by a parametric study of the long time behaviour of parquet planks that
emphasises the influence of creeping on the aforesaid deformations and failure
modes.
Parquet floor product from Sweden was tested. The parquet specimen as seen in
Fig. 18.1, has three main layers: the surface layer – denoted SL – is 3.6 mm thick,
the core layer – denoted CL – is 8.6 mm thick and backing layer – denoted BL – is
2 mm thick. The layers are glued together crosswise with urea formaldehyde resin.
The geometry of the specimen, a parquet plank, is a plate of size 188 × 14.2 ×
2500 mm.
188 mm
3.6 mm
8.6 mm
Z
2 mm
14.2 mm
X
Y click joint Surface layer: Oak (Quercus robur L.)
Core layer: Pine (Pinus sylvestris L.)
local coordinate system Backing layer: Veneer pine
Laboratory tests on the materials which compose the parquet planks (referred
to as basic material) were performed to determine mechanical characterization
and for providing data for calibration and validation of the finite element method
calculations.
Wood species used in these experiments are: pine, oak, beech and ash. The adsorp-
tion behaviour, density, static modulus of elasticity in the longitudinal direction and
the hygroexpansion factors of pine (Pinus sylvestris L.) and oak (Querqus robur L.)
have been determined. Data for beech and ash were obtained from the literature.
The static modulus of elasticity in the longitudinal direction was determined
using a three point-bending test. The measured values were in the same range
and showed the same variation as noted in the literature (Kollmann 1982, Wood
Handbook 1999).
For determining the adsorption behaviour, 20 samples from each species were
conditioned in a climatic chamber at 20◦ C and 25% relative humidity (RH) until
18 Moisture Induced Stresses and Deformations in Parquet Floors 367
they reached equilibrium moisture content (EMC). Thereafter, the climate was
changed from 25% to over 50% and then up to 85% relative humidity, all at a
temperature of 20◦ C. The shrinkage and swelling coefficients in the longitudinal,
radial and tangential directions were measured by changing the climate from 60%
RH down to 25% RH (shrinking, 20 specimens) and then/from 60% RH up to 85%
RH (swelling, 20 specimens) respectively, all at a temperature of 20◦ C.
Position A
G H I
D E F measurement points for
horizontal deformation
2 (Z) A B C
3 (Y) 1 (X)
measurement points for
direction of x axis
horizontal deformation
Position B
Fig. 18.2 Measurement of the plate’s vertical deformation in two directions and evaluation with
the beam theory of Bernoulli (t: thickness change, κ: curvature in x and y direction respectively)
368 S. Blumer et al.
coordinate system, thereafter 90◦ rotation counter clockwise of the plate was done
for the second measurements. The plate was supported at the downside of point
B, G and I., B’, G’ and I’ respectively. Three variables (thickness of the plate, cur-
vature in the x and y directions respectively), describing the vertical deformation of
the specimen, have been evaluated from the eight data measurements using a least
square fit and Bernoulli beam theory.
Several finite element models were created to simulate the behaviour of the par-
quet planks under different climatic conditions, such as a calibration model, a model
for the study of distortional effects and a model for gap opening. These models are
described in following sections. The influence of different parameters such as the
properties of the material, material orientation, glue line properties and the geometry
of the product on the stress and deformations will be tested using these models.
• The diffusion coefficient of pine and oak wood in the radial and tangential
directions are equal;
• The estimated diffusion coefficient is assumed constant below 15% MC, (Jönsson
2005)
• The model does not explicitly consider any interfacial layer between the wood
layers and the glue layer. The interfacial layers have been reduced to a continua-
tive 0.l mm thick composite layer. The glue lines between the backing and core
and core and surface layer respectively, were modelled as a 0.1 mm thick layer
(UF resin) with material data taken from Hagstrand (1999).
The relationships between the diffusion coefficient, the specific heat, the density
and the thermal conductivity (Carslaw and Jaeger (1959) and Eriksson (2005) are
given in the Eqs. (18.1) and (18.2))
∂T ∂ λ(T) ∂T
= (18.1)
∂t ∂x cρ ∂x
∂u ∂ ∂u
= DW (u) (18.2)
∂t ∂x ∂x
– density ρ [kgm−3 ]
– thermal conductivity λ [Js−1 m−1 K −1 ]
– time t [s]
Thermally coupled and quadratic interpolated brick elements have been used for
the calculation. To simplify the transport model, the moisture content at the sur-
face was set in equilibrium with the moisture content corresponding to the relative
humidity of circulating air, μsurf = μair . This imposed boundary condition is called
the boundary condition of Drichlet (Koc and Houska 2002).
An effective coefficient of diffusion for the entire parquet plank was deter-
mined and compared to experiments on specimens with moisture-isolated edges
(test series 2).
The bending deformations of the plate in plane xz and yz have been calculated
and compared to experiments on test series 2. The static boundary conditions were
included consistent with the test set up of test series 2. The degrees of freedom u1 ,
u2 and u3 were restrained on the lower edge of point B and in the vertical direc-
tion u2 on the lower edge of point G and I respectively, see Fig. 18.4. The gaps
between the pine strips in the core layer were modelled with ABAQUS. The layer
consisted of three strips (left, middle and right) each with different direction of the
growth rings (longitudinal L, radial R and tangential T). The material orientation
of the surface layer has been varied from 0◦ (Tangential direction parallel to u1 or
x direction according to Fig. 18.3) to 90◦ (Tangential direction perpendicular u1 or
x direction). The angle of the growth rings has been set similar to the test speci-
mens of test series 2. Transforming the stiffness matrix of the oak layer’s different
strips simulated the influence of the growth ring’s direction. The transformation
was done for both the stiffness matrix and the hygroexpansion factors. A coordinate
150
425 65 425
SL: left strip SL: middle strip SL: right strip
R (variable)
2 (Z) A B C
3 (Y) 1 (X)
T
core layer
L
R R
backing layer T
L
Fig. 18.3 Geometry and material of the parquet plank. The angle of the growth rings in the surface
layer differs between the strips
370 S. Blumer et al.
transformation of the orthogonal coordinate system around the longitudinal axis (L)
has been performed.
A finite element model (named the analytical model B and shown in Fig. 18.4) was
applied to predict the distortional behaviour of the parquet plank’s central compo-
nent. Parametric study on the influence of geometry, material and creeping of the
surface layer were performed. The model corresponded to half the width of the strip
and a depth of 26 mm. This depth included 2 half width core sticks (2×12.5 mm)
and 1 mm spacing between the sticks. The depth of the model was relatively small
compared to the length of the parquet planks, which is 2500 mm. The parquet strip
was 10 mm wide, 14.4 mm thick in a three-layer structure (3.6 mm surface layer
(SL), 8.6 mm core layer (CL) and 2 mm backing layer (BL)) glued together with
two 0.1 mm thick UF resin layers. The vertical deformation of the parquet planks
has been calculated between point A and point B. These points were located on
nodes, point B on the boundary edge whereas point A was located 10 mm from
the boundary to minimize the local deformation shape of the unconstrained face in
plane yz at x = 0. The surface in plane yz at the value x = 94 mm was constraint in
u1 or x direction according to the coordinate system shown in Fig. 18.2. The surface
could not be blocked in u2 or z direction in order to allow free movement of the
surface layer in the vertical direction. The edge below was also constraint in u2 or
z direction for stability reasons. The surfaces in plane xz were constraint in u3 or y
direction. This boundary condition simulated an infinite depth of the parquet plank.
Coupled temperature-displacement and quadratic interpolated elements have been
chosen for the model.
Fig. 18.4 Analytical model A: Geometry, static system and boundary conditions used for the
modelling of the distortional effects
The geometry, the static system and the boundary conditions of the model C are
shown in Fig. 18.5. The proposed model was applied to predict the deformation
18 Moisture Induced Stresses and Deformations in Parquet Floors 371
2 (Z)
94.000 94.000
1 (X)
2 8.6 3.6
A B
A.3 B.3
uA.1 uB.1
behaviour of the click joint and the gaps of the surface layer. The model corre-
sponded to half the width of the parquet plank on the right and left side of the click
joint. Because stresses mainly occurred in xz, the model was reduced to two dimen-
sions. The side edges of the model were coupled to the reference points A and B in
u1 or x direction. The symmetrical behaviour has been introduced to the model by
constraint equations. The rotation of the edge at point A is the same as that of the
corresponding surface at point B (ϕA.3 = ϕB.3 ). The horizontal deformation in u1
or x direction had similar values but opposite signs (uA.1 = −u1.B ). The model was
based on an elastic layer with a very small E-modulus, to give the stability in u2 or
z direction. This was done in order to simplify the model, such that no algorithm for
modelling the contact with the foundation had to be used. The contact of the model
in the click joint was modelled by a contact algorithm triggering reaction forces
in the case where the elements of the tongue come in contact with the element of
the groove in the joint region. Seams have been introduced for simulating the gaps
in the surface layer. Coupled temperature-displacement and quadratic interpolated
triangular and quadratic elements were chosen for this model.
In the following the moisture induced stresses and deformation in parquet floors
determined with the models A, B and C will be discussed.
The estimation of the moisture transport using different effective diffusion coef-
ficients resulted in Deff = 1.8e−11 [m2 s−1 ]. The value obtained was about 1.5−2
times smaller than that given by Simpson (1993). This difference may be caused
by the influence of the two glue lines in the parquet element that acts as a moisture
barrier.
372 S. Blumer et al.
L L
Z
X
Y v L
w L
The specimens without edge isolation (denoted A11–A20 for non lacquered
specimens) reached equilibrium moisture content in a climate of 20◦ C/25% RH after
28 days conditioning. Thus, the numerical analyses were performed as steady state
calculations. The comparison of the vertical deformation v (Fig. 18.6) between the
test specimen and numerical analysis according are shown in Figs. 18.7 and18.8
for lacquered specimens A1 . . .. A10, and for non lacquered specimens were
denoted A11. . ..A20. The bending of the plank in xz plane was strongly dominating
(v >> w). Figure 18.9 gives the comparison between the experimental and the
numerical results with the Analytical model A, for lacquered and non lacquered
–0.1
–0.15
Fig. 18.7 Test results versus –0.2
numerical FEM calculations,
–0.25
moisture content change from
10.25 to 6.85%. Bending –0.3 Mean values
deformation of lacquered –0.35
specimens (A1–A10 are the
specifications of the samples A1 A2 A3 A4 A5 A6 A7 A8 A9 A10
of test series 2) Specification of lacquered specimens
–0.1
–0.15
Fig. 18.8 Test results versus –0.2
numerical FEM calculations,
–0.25
moisture content change from
10.25 to 6.85%. Bending –0.3
Mean values
deformation of non lacquered –0.35
specimens (A11–A12 are the
specifications of the samples A11 A12 A13 A14 A15 A16 A17 A18 A19 A20
of test series 2) Specification of non lacquered specimens
18 Moisture Induced Stresses and Deformations in Parquet Floors 373
–0.25
–0.3
–0.35
–0.4
–0.4 –0.35 –0.3 –0.25 –0.2 –0.15 –0.1 –0.05 0
Bending v, numerical analyse [mm]
specimens, for moisture content decreasing between 10.25 and 6.85%. Further
comments regarding this model are:
– the introduction of different angles of the growth rings for each of the tree strips
of the surface layer shown in Fig. 18.3 had an important influence on the plate
deformation behaviour;
– the surface treatment (lacquered or non lacquered) did only slightly influence the
bending behaviour of the square samples.
The cupping effect under different drying conditions using different materials, angle
of the growth rings and geometry has been calculated. At the start of the calcu-
lation, the boundary condition of the surface layer was determined for moisture
content decreasing from 7.5 down to 5% and for different angles of growth rings
(Fig. 18.10). The cupping minimum occurs under 45◦ in both strips of the surface
layer. The influences of the surface thickness on the cupping of the parquet are
0.8
Angle = 30 deg
0.7 Angle = 45 deg
Angle = 90 deg
0.6 Angle = 0 deg
Cupping v [mm]
0.5
0.4
Fig. 18.10 Influence of the 0.3
surface layer’s angle of Beech
Beech
Angle SL 3.6mm
==Beech
30deg Beech
Angle 30deg
growth rings after reduction 0.2 Angle
Angle
Oak
Oak=SL
45
=Oak
45deg
3.6mm
deg Oak
Angle
Angle=SL
90
=Ashdeg
03.6mm
deg
of the moisture content from 0.1 Ash
Ash
Angle
Angle==090
deg
deg
Ash
7.5 down to 5%. (Local
0
tangential direction parallel to 0 20 40 60 80 100 120
horizontal plane at α = 0◦ ) Time [days]
374 S. Blumer et al.
Cupping v [mm]
0.5
0.4
0.3
Surfacelayer
Surface 4.8 mm
layer2.6mm
2.6mm
0.2 Surfacelayer 3.6 mm
layer3.6mm
3.6mm
Surface
Surfacelayer
Surface 2.6 mm
layer4.8mm
4.8mm
0.1
0
0 20 40 60 80 100 120
Time [days]
0.5
0.4
0.3
Beech SL 3.6mm
Beech
0.2 Oak Oak
SL 3.6mm
Ash SL
Ash3.6mm
0.1
0
0 20 40 60 80 100 120
Time [days]
shown in Fig. 18.11, for moisture content decreasing from 7.5 to 5%. The influ-
ences of the surface layer’s material – beech, oak and ash – on the cupping of the
parquet are shown in Fig. 18.12, for the same moisture content decreasing range.
The significant influence of the geometry (layers thickness) and of the species can
be observed. Beech had the highest cupping , a maximum of 0.78 mm after 20 days
while ash had the lowest cupping , 0.44 mm after 20 days.
Figure 18.13 shows the model and the cuts for stress calculation for the model
C where S11 is the horizontal stress, S22 the vertical stress and S12 the shear
stress. The influence of the materials used for the surface and core on the verti-
cal stress as a function of the horizontal distance to the middle of the gag is shown
in Fig. 18.14.
Figure 18.15 shows the variation of the horizontal stress through the cut of the
parquet plank for different geometries of the surface layer and core. It is notable that
the absolute values of the stresses are extremely mesh-size dependent. Maximum
18 Moisture Induced Stresses and Deformations in Parquet Floors 375
Fig. 18.13 Model and cuts for stress calculation. S11 : Horizontal stress, S22 Vertical stress and
S12 Shear stress
values of the stresses cannot be evaluated in this model. For this reason the curves
have to be compared on the basis of their gradients. The main target of the para-
metric study was to minimize the gradient of the vertical, horizontal and shear
stresses. The vertical stress S22 and the shear stresses S12 in the glue line can lead
to delamination. A steeper curve close to the gap indicates an increased risk for
crack formation and propagation of delamination. Here, the highest gradient of ver-
tical stresses can be observed for the beech wood. The creeping is stress depending
as demonstrated by Jönsson (2005) and Hanhijärvi (1995). Thus, higher horizontal
stresses can lead to higher creeping effect in the wood. A stress gradient in the sur-
face layer may result in stronger creeping of the surface layer at the bottom as on the
top. The effect of a creeping gradient in the surface layer has been modelled and the
variation of the vertical stresses S22 after creeping gradient is shown in Fig. 18.16.
The minimum is observed at about 6 mm distance to the gap.
Fig. 18.14 Vertical stresses S22 in CUT A: Different materials in the surface and core layer (HDF)
respectively
376 S. Blumer et al.
15
SL 4.8 mm/CL 8.6 mm
10
Groove
Core layer (CL)
5
Fig. 18.15 Horizontal stresses S11 in CUT B Influence of the geometry of the surface layer
3
φR,T = 0.8 2.5
Vertical stresses S22 [MPa]
2
φR,T = 1.6 1.5
ΦR,T = 0
1
φR,T = 3 0.5
ΦR,T = 0.8..1.6..3 (gradient)
0
hoak
–0.5
–1
hpine
–1.5
–2
0 5 10 15 20
Horizontal distance to the gap [mm]
18.4 Conclusion
As concluding remarks it is noted that a model that includes the whole parquet sys-
tem helps find optimal solutions as a function of stresses in the glue line and gap
opening of the surface layer. In the previous discussion, it was demonstrated that
the finite element method brings several advantages compared to traditional testing
in laboratory conditions. The time needed for simulating changing climatic cycles
is much smaller compared to laboratory tests. In the future, the design process for
18 Moisture Induced Stresses and Deformations in Parquet Floors 377
wood flooring systems should include basic material testing, finite element analyses
and, finally, testing of the developed product. Other advantages with modelling tech-
niques include the possibility to optimize the geometry of the joint and the lay-up of
the planks in a rather straightforward manner. The material and the angle of growth
ring in the surface layer have a considerable influence on the deformation and stress
distribution of the parquet planks. An angle of 45◦ (between tangential direction and
horizontal plane) in the surface layer minimized the cupping deformation. From the
design perspective, results based on calculations with elastic properties of the glue
line without introducing creeping factors are conservative; bigger deformations than
experienced in practice are predicted. The material properties of the glue line and
lacquer are difficult to determine, although the finite element method can be used for
parameter estimation. The long time behaviour of the glue line did not significantly
influence the deformation and stress distribution. It seems to be a good approach in
terms of modelling to assign the UF resin layer properties making it less hygroscopic
than wood and acting as a linear elastic layer. A hygroscopic material model may
make more sense for the wood material than for the glue line. Periodic loading can
increases the creeping effect, delamination may also occur after several summer –
winter cycles.
18.5 Summary
The indoor climate in buildings has changed in the last decade due to more efficient
climatic systems, floor heating systems and larger open floor areas with more natural
light. All these factors have induced increasing ranges of relative humidity between
different seasons. Also with decreasing relative humidity (in the winter 30–50%
RH, in the summer 70–90% RH), floor-heating systems increase the temperature
in wooden parquet planks for example. Such variations can result in troublesome
deformations, delamination of the surface layer and development of cracks in the
parquet flooring boards. Sometimes there is only deterioration of the appearance
but the durability of the flooring system can also be reduced. Many laboratory tests
have to be done before reaching an optimal design of the parquet elements. Due
to the high costs and time constraints of experiments, other supplementary research
methods should be tested and evaluated. The articles’ main objective was to improve
understanding of the behaviour of parquet floors exposed to different climatic condi-
tions by using numerical calculation. The use of the finite element models provides
options for design purposes of wood flooring systems. Several finite element models
to aid adequate design have been created, tested and applied. After calibration and
validation of the calculation method, parametric studies on the influence of material
properties, geometry of the parquet floors and the long-term behaviour of the wood
and glue line were performed. The results show a strong relation between material
and geometry choice on the deformation, for example the gap opening and the stress
distribution in the glue line, which can induce delamination of the surface layer and
distortional effects of the parquet boards.
378 S. Blumer et al.
References
Bodig J, Jane A (1982) Mechanics of wood and wood composites. – Van Nostrand Reinhold,
New York, NY
Carslaw HS, Jaeger JC (1958) Conduction of heat in solids. Oxford University Press, London
Eriksson J (2005) Moisture transport and moisture induced distortions in timber. – Doctoral Thesis,
Department of Applied Mechanics, Chalmers University of Technology, Göteborg
Hagstrand PO (1999) Mechanical analysis of melamine-formaldehyde composites – Doctoral
Thesis, Department of Polymeric Materials, Chalmers University of Technology, Göteborg
Hanhijärvi A (1995) Modelling of creep deformation mechanisms in wood. – Dissertation,
Technical Research Centre of Finland, Espoo
Jönsson J (2005) Moisture induced stresses in timber structures. -Doctoral Thesis, Report TVBK-
1031, Division of Structural Engineering, Lund University
Koc P, Houska M (2002) Characterisation of the sorptive properties of spruce wood by the inverse
identification method. Holz als Roh und Werk 60:265 – 270
Kollmann F (1982) Technologie des Holzes und der Holzwerkstoffe 2. Auflage- Springer, Berlin,
Heidelberg, New York, NY
Simpson WT (1993) Determination and use of moisture diffusion coefficient to characterize drying
of northern red oak (Quercus rubra). Wood Sci Technol 27:409–420
Wood Handbook (1999) Wood as an engineering material. -USDA Forest Products Laboratory,
Forest Laboratory, Madison, WI
Chapter 19
Glue Line Nondestructive Assessment in Timber
Laminates with an Air-Coupled Ultrasonic
Technique
Contents
19.1 Introduction
Glued solid wood products have gained much importance during the last years, as
they allow an efficient and versatile use of the renewable timber material. Current
standardized methods for bonding quality assessment consist of tests of small spec-
imens cut from the structure during production or visual in-service inspection.
Ultrasonic diagnostics are traditionally based on discrete point measurements using
contact techniques. Transducers are generally pressed onto the timber surface with a
coupling gel, liquid, or membrane couplant. Large glue line defects in glued timber
constructions have been detected with this method (Dill-Langer et al. 2005). The
disadvantages are a low precision in signal level measurements, which are highly
dependent on the coupling pressure, and that the coupling agent may deteriorate the
object. Better repeatability and one-dimensional continuous scanning is achieved
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 379
DOI 10.1007/978-90-481-9550-3_19, C Springer Science+Business Media B.V. 2011
380 S.J. Sanabria et al.
with roller transducers, which have been applied to defect inspection in hardwood
lumber (Kabir et al. 2002). Non-linear effects have been used to assess delamina-
tions between veneer lamina and particle board (Solodov et al. 2004); a welding
piezoelectric stack couples high power ultrasound into the sample and its surface is
scanned with a laser vibrometer.
Air-coupled ultrasonics (ACU) provides much more flexibility than traditional
techniques since the transducer can be moved at a certain distance from the sur-
face of the object, so fine and reproducible scanning in any direction is possible. A
high-power low-frequency ACU system for split detection in wood composites is
currently used in production lines (Niemz and Sander 1990). Ultrasonic imaging
has been performed in solid wood using through-transmission mode for inspec-
tion of density, knots, microcracks and drilled holes (Gan et al. 2005; Hasenstab
2006). Delaminations in wood panel paintings between solid wood and a thin plaster
layer have been assessed with both through-transmission and single-sided inspection
(Siddiolo et al. 2007).
In this work we present preliminary results of the application of ACU to assess
disbonding in glued solid wood objects. A specific measurement set-up and data
evaluation based on voltage level measurements of recorded A-scans allows pre-
cise imaging of areas with and without adhesive. Advantages and limitations of this
method are discussed.
Vglued
Lglued/non glued = 20 · log10 = Twood→glue→wood − Twood→air→wood
Vnon glued
1 ·Z2
T1→2→1 = 20 · log10 (Z4·Z+Z )2 Zi = ρi · ci
1 2
(19.1)
Where: Lglued/non glued (dB) is the amplitude level ratio between ACU signals prop-
agating through glued and non-glued material; Vglued (V) and Vnon glued (V) are
corresponding amplitude measurements in the recorded A-scans. T1→2→1 (dB) is
19 Glue Line Nondestructive Assessment in Timber Laminates 381
the transmission coefficient for a single echo propagating through a layer of mate-
rial 2 between two semi-infinite media of material 1. Zi (Pa·s/m), ρi (kg/m3 ) and
ci (m/s) are the acoustic impedance, density and speed of sound in the propaga-
tion direction for medium i. From available data cspruce = 1300 m/s (measured in
T orthotropic direction from ACU data following a similar method to (Vun et al.
2003) and ρspruce = 409 kg/m3 (gravimetrical determination), therefore Zwood =
0.532·106 Pa·s/m . From literature data (Deutsch et al. 1997) Zglue = 2.2·106 Pa·s/m
and Zair = 0.000427 · 106 Pa · s/m (dry air T = 20◦ C). From Eq. (19.1) it follows
that Twood→glue→wood = −4.1 dB and Twood→air→wood = −49.9 dB.
Table 19.1 Geometry of glued timber samples manufactured for ACU measurement
Type Description
The measurement setup is shown in Fig. 19.2. Two ACU broadband planar transduc-
ers (model NCG100-D50, The Ultran Group) with a central frequency of 120 kHz
and 50 mm active diameter were aligned perpendicularly to the surfaces of the sam-
ple, one transmitting an ultrasonic signal and the other one receiving it. The distance
between the transmitter and the sample (210 mm) was chosen to minimize the diam-
eter of the sound field penetrating the latter. A three-axis system from ISELR
moves
the two transducers together as a fixed unit; scanning the surface of the samples with
steps of 1 mm in the fast axis and 4 mm in the slow axis. A sinusoidal pulse of 115
Vpp amplitude and 33 μs length windowed with a Gaussian function was applied
to the transmitter. Received waveforms were amplified with a gain of 52 dB and
digitized with a sampling frequency of 2.5 MHz and 14 bits resolution, the gener-
ated A-scans being stored for each scanned position. No averaging of A-scans was
performed. C-scans were generated from a peak or root mean square (RMS) voltage
measurement for each A-scan and a defined time gate [t1 , t2 ]:
t2
1
VPEAK = max V(t) VRMS = V(t)2 dt (19.2)
[t1, t2] t2 − t1
t1
The distance between receiver and sample (80 mm) allows separating in time
multiple reflections between their surfaces (3) from measured waves (1) and (2).
A-scans received through bonded material (1) present a signal-to-noise ratio of 55
dB, which allows for enough dynamic range to record waveforms from glued and
non-glued areas in a single scan. Waves diffracted at the edges of the sample (4)
are blocked by a frame built around the inspected object. The frame is made from
wood (Norway spruce) covered by several layers of paper with small gaps of air
19 Glue Line Nondestructive Assessment in Timber Laminates 383
Fig. 19.2 Experimental set-up. Top image: Inspection principal and main propagation paths. 1
and 2 are waves propagating through the sample for glued and non-glued areas respectively. 3 are
multiple reflections between the receiver and the surface of the sample. 4 are waves diffracted at the
edges of the sample, which are blocked by a frame built around the object. The noise level in the
A-scans is 1.3 mVRMS . Bottom image: Profile of voltage level along the fast axis normalized with
respect to the glued area. The gap thickness decreases linearly between fast axis 0 and 250 mm
from 1 mm down to the glue line thickness
in between. Waves propagating through the frame are highly attenuated due to the
accumulation of impedance mismatch losses.
Fig. 19.3 ACU imaging of glue presence for the sample in Fig. 19.1. Top image: Photograph of
open board; the transition between glued and non-glued area is highlighted. Bottom image: ACU
C-scan of the sample. Feature 1 of the transition between glue/no glue and the resin pocket (2) can
be visualized
and right sides of the surface of the image can be clearly distinguished; the ampli-
tude values being higher than the ones measured in non-glued areas. In both images,
details of the transition between glued and non-glued regions can be resolved. The
spatial resolution of the images is limited by the sound field diameter (about 35 mm);
features smaller than 20 mm cannot be resolved. Preliminary tests applying spatial
deconvolution algorithms to the ultrasonic images showed an improvement of the
resolution limit down to 10 mm.
Fig. 19.4 ACU imaging of glue presence and absence of a sample type D. Top image: Photograph
of open board; in this case there was a small drop of glue (feature 1) joining the two lamellas,
separated from the glued area on the right side of the sample. Bottom image: ACU C-scan. The
presence of the drop of glue can be clearly recognized; however, the non-glued area between feature
1 and the glued region on the right side is smeared by the finite diameter of the sound field
19 Glue Line Nondestructive Assessment in Timber Laminates 385
Figure 19.2 shows in logarithmic scale the voltage level variations between glued
and non-glued area for a typical fast axis amplitude profile of a sample type C, and
for specific air gap thicknesses between the two lamellas. The voltage level in the
non-glued area shows a minimum of –50 dB with respect to the glued area, in good
agreement with the estimation for Twood→air→wood . As the gap thickness between
boards decreases, this level rises due to multiple reflections of the ultrasonic wave
adding constructively in the gap. About 20 mm from the boundary to the glued
area propagation through bonded material becomes dominant, owing to the finite
diameter of the sound field. An amplitude rise is observed in the delaminated area
from fast axis 0 to 50 mm. It probably corresponds to residual ultrasonic energy
diffracting through air at the edges of the sample, which is not blocked by the frame.
Best measurement performance was achieved by limiting secondary ultrasonic
propagation paths by evaluating a reduced number of cycles at the beginning of
the received waveform. RMS voltage and peak voltage give similar results with
sufficiently short temporal gates (less than 40 μs). Repeated ACU measurements of
the same object showed variations of less than 0.1 dB (error < 1%).
A homogenous amplitude level was observed in the ACU images of samples type
A and B. The average voltage level measured for type B glued samples is –1 dB with
respect to the value for type A solid wood samples; a smaller difference than pre-
dicted by Twood→glue→wood , which further enhances the contrast of ACU images.
A probable reason is the constructive interference of multiple reflections of the
ultrasonic wave in the glue line.
19.5 Conclusion
We have demonstrated that air-coupled ultrasound is well-suited for glued timber
inspection; combining the high sensitivity to disbonded interfaces of traditional
ultrasonic methods with a phenomenal reproducibility in amplitude measurements,
386 S.J. Sanabria et al.
19.6 Summary
References
Brekhovskikh LM (1980) Waves in layered media. New York, NY, Academic
Deutsch V, Platte M, Vogt M, Verein Deutscher Ingenieure (1997) Ultraschallprüfung Grundlagen
und industrielle Anwendungen. Springer, Berlin
Dill-Langer G, Bernauer W, Aicher S (2005) Inspection of glue-lines of glued-laminated tim-
ber by means of ultrasonic testing. In: Proceedings of the 14th international symposium on
nondestructive testing of wood. Eberswalde, pp 49–60
19 Glue Line Nondestructive Assessment in Timber Laminates 387
Gan TH, Hutchins DA, Green RJ, Andrews MK, Harris PD (2005) Noncontact, high-resolution
ultrasonic imaging of wood samples using coded chirp waveforms. IEEE Trans Ultrason
Ferroelectr Freq Control 52(2):280–288
Hasenstab A (2006) Integritaetspruefung von Holz mit dem zerstoerungsfreien
Ultraschallechoverfahren.Technische Universitaet Berlin. PhD Thesis
Kabir MF, Schmoldt DL, Schafer ME (2002) Time domain ultrasonic signal characterization for
defects in thin unsurfaced hardwood lumber. Wood Fiber Sci 34:165–182
Niemz P, Sander D (1990) Prozessmesstechnik in der Holzindustrie. VEB Fachbuchverlag, Leipzig
Schmerr LW, Song SJ (2007) Ultrasonic nondestructive evaluation systems models and measure-
ments. Springer, New York, NY
Siddiolo AM, D’Acquisto L, Maeva AR, Maev RG (2007) Wooden panel paintings investiga-
tion: an air-coupled ultrasonic imaging approach. IEEE Trans Ultrason Ferroelectr Freq Control
54(4):836-846
Solodov I, Pfleiderer K, Busse G (2004) Nondestructive characterization of wood by monitoring
of local elastic anisotropy and dynamic nonlinearity. Holzforschung 58:504–510
Vun RY, Wu QL, Bhardwaj MC, Stead G (2003) Ultrasonic characterization of structural properties
of oriented strandboard: a comparison of direct-contact and non-contact methods. Wood Fiber
Sci 35(3):381–396
Chapter 20
From Present Researches to Future
Developments
Voichita Bucur
Content
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
V. Bucur (B)
CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton,
Victoria 3168, Australia
e-mail: voichita.bucur@csiro.au
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 389
DOI 10.1007/978-90-481-9550-3_20, C Springer Science+Business Media B.V. 2011
390 V. Bucur
Damage involves mechanisms at all scales, from the molecular level revealed by
sophisticated instruments to massive scale during physical inspections. The dam-
ages can be induced by mechanical or thermal loads, environmental influences,
or result from residual stress or combinations of them, such as thermo-mechanical
stress. The ranges of different mechanisms and scales in damage accumulation in
wood or wood-based composites raise permanent problems for a detailed experi-
mental characterization as well as for modelling of delaminations. Implementation
of more sophisticated approaches for mechanical behaviour of wood-based com-
posites will require standardization of all three pure fracture modes and mixed
fracture modes characterization test methods for delamination onset threshold
and growth.
There are new research areas to be suggested, such as: testing of delamination
growth under dynamic loading, testing of multidirectional laminates, adaptation
of tests to new wood-based composites with through- thickness reinforcements,
or determination of in-situ interlaminar shear strength as a controlling factor for
the initiation of delamination. Although the delamination induced by dynamic frac-
ture toughness is of fundamental importance for wood machining and for pulp and
paper industry, the dynamic delamination test is not easy to perform because it is
experimentally difficult to induce high speed delamination growth in a simple and
controlled manner (Ravi-Chandar 2004; Freund 1998).
It must be borne in mind that traditional current, as well as new wood or
wood-based composite structures can be highly vulnerable to damages, in partic-
ular delaminations that might have been introduced during manufacturing, tooling,
processing or in service. Delaminations are difficult to detect by visual inspection,
thus reliable and if possible inexpensive detection methods and technologies (active
or passive) must be developed to improve safety and reliability of new wood – based
laminated composites structures in service. It is of vital importance to identify the
delaminations in new wood-based composites structures at the early stage, so as to
prevent any potential failure.
For further development in structural health monitoring of wood-based compos-
ites, the key issues are the prediction of delamination in different products and the
improvement of the design of new advanced wood - based composites and struc-
tures, to prevent and minimize the products from delaminations. Without doubt,
these fascinating challenges will be solved in the future by scientists and engi-
neers with new perspectives involving in this highly interdisciplinary field, which
has enormous potential for practical applications.
References
Freund LB (1998) Dynamic fracture mechanics. Cambridge University Press, Cambridge
Lemaitre J, Desmorat R (2005) Engineering damage mechanics. Springer, Berlin, Heidelberg
Ravi-Chandar K (2004) Dynamic fracture. Elsevier, Amsterdam
Sridharan S (ed) (2008) Delamination behaviour of composites. Woodhead Publishing, Cambridge
Tay TE (2003) characterization and analysis of delamination fracture in composites: an overview
of development from 1991 to 2001. Appl Mech Rev 56:1–31
Index
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, 393
DOI 10.1007/978-90-481-9550-3, C Springer Science+Business Media B.V. 2011
394 Index
E G
Ecological relevance, 91 Gain, 34, 201, 217, 308, 320, 382, 386
Effective Gap opening, 368, 370–371, 374–376
crack length, 88 Geometry, 11, 54, 60, 62, 87, 105, 111, 113,
Elastic constants, 8–9, 55, 78, 113, 115, 309, 163, 167, 244, 329, 333–334, 337, 339,
311 341, 366–371, 373–374, 376–377, 381
Elasticity, 10, 91, 116, 270, 276, 283, 291, 337, Glue, 9–10, 24–25, 126, 134, 138, 192, 199,
339, 358, 361, 366 227, 277, 283, 331, 340, 358–359,
Elliptical, 242 367–368, 371, 374–377, 379–386
Energy Glued laminated timber – glulam, crosslam,
balance, 54 glulam slabs, 307
Energy release rate, 38, 54–55, 58–59, 72–73, Grading, 221, 224–225, 358, 363
88, 90, 390 Grain
Engineered wood products (EWP), 216–217, angle effect, 89
226–232, 235 slope, 166
Engineering materials, 26, 45 Griffith, 52
Errors, 4, 8 Griffith theory, 52
Event, 24, 45, 64, 288–289, 291–294, 297–301, Growth rings, 18, 78, 82, 128, 206, 222, 225,
308 242, 252, 272, 278, 283, 369, 373, 377
Experimental determination of Guided waves, 320
fracture toughness, 391
stress intensity factor, 391 H
Exposure Hardness, 138, 178, 252, 297
artificial, 178, 186–192 Hardwood, 6, 11, 126, 130, 161–162, 164, 176,
outdoor, 149, 174–175, 178–186, 191–192 185, 207–208, 211, 220, 222–223, 230,
235, 291–292, 362, 380
F Harmonic, 39, 290, 328–329, 341–342
Failure Health monitoring, 3, 12, 27, 29, 33–34,
modes, 91, 288, 366, 390 37–38, 40, 44–45, 64, 389, 391
stress, 211, 389 Heat transport, 160
Fatigue Hemicelluloses, 7
tests, 293 Hidden specimens, 220–221, 225, 299–300
Fiber High resolution, 24, 138–139, 145–155, 386
board, 229, 308 Hook’s Law, 311, 344
pullouts, 308 Humidity, 4, 12, 26, 80, 136, 173, 177–178,
Finite 189, 198–203, 206, 233, 239, 244, 270,
ABAQUS, 365 273, 275, 308, 343, 366–367, 369, 377
element method, 43, 116, 366, 376
finite element analysis (FEM), 27, 33, 38, I
45, 76, 78, 102–103, 112, 117, 372–373 Impact testing, 223, 308
Flaws, 34, 53, 308, 313, 320 Impregnation, 159, 163–166, 169, 199
Flexural, 39, 42, 310–311, 316–319 Infrared spectroscopy, 332
Flexural vibration, 39 Inhomogeneities, 385
Focused air-coupled transducers, 324 Initiation, 5, 20, 26, 29, 51–91, 135–136,
Fourier Transform Technique, 35, 37 183–185, 218, 232, 235, 242, 291, 293,
Four-point-bending test, 367 295, 303, 356, 389–391
Fractography, 80, 301 In-plane, 317, 320, 336
Fracture In-service damage, 308
process zone, 56–57, 65, 72, 74, 92 In-situ, 64, 69, 82–83, 134, 288, 295, 297, 358,
Frequency 391
natural, 343 Inspection
spectrum, 327 continuous, 45
Fundamental one side, 316
frequency, 327 two side, 316
396 Index
Integrity, 3, 25, 29, 34, 45, 137–138, 232, 308, 233, 256, 258, 261, 269, 282–283, 309,
354, 386 313, 316, 322, 366–367, 369–370, 380
Interface, 4, 9, 11–12, 24, 34–35, 42–43, Low frequency vibrations, 33, 38, 389
58–59, 74, 87, 127, 132–133, 145–156, Low temperature, 137, 208–210
165, 295, 318, 321 Lumber, 3–4, 9–10, 18–19, 25, 191, 215–217,
Interface wood-cement, 12 221–227, 231, 233–234, 255–256, 297,
Interfacial cracking, 59 300, 307, 324, 380
Interfacial layers, 368
Interlaminar fracture toughness, 391 M
Inverse Machining, 87, 149, 252, 391
Fourier transform, 35, 37 Macrocracks, 56, 64, 291
Irwin lamellar tearing, 55 Main, 19, 35, 45, 56, 64, 69, 87, 112, 116, 139,
Iterative, 110 147, 152–153, 162–163, 179, 192, 206,
220–221, 228, 240, 242, 246, 251, 256,
J 269, 289, 302, 308–309, 311–312, 322,
J-integral, 57–59 333, 337, 343, 355, 365–366, 375, 383,
Joints, 24–25, 72, 84, 87, 192, 226 386, 389–390
amplifier, 312
K Manufacturing
Knots, 25, 166, 204, 217, 221, 223, 225–226, defects, 223–224
231, 380–381 process, 28–29, 223–230
Mass production, 3
L Material properties, 77–78, 101, 103–105,
Lamb wave, 33–35, 37–38, 45, 309–311, 317, 111–112, 115–116, 251, 367, 377
339–341, 344 Maximum, 64, 66–67, 72, 134, 136–137, 168,
Laminated veneer lumber (LVL), 3, 9–10, 12, 178, 185, 198, 202, 206–208, 210–211,
19, 29, 227, 229, 307, 343–344 216, 221, 239, 246, 279, 291, 293, 300,
Laminates, 4, 19, 355, 379–391 325, 329, 353, 374
Laser scanning microscopy, 145, 147–152, 295 Mechanical
Laser vibromerty, 39, 319, 380 performance, 353
Lateral, 104, 166, 220, 260, 262, 266, 323, 327 properties, 3, 6–11, 34, 39, 78, 113, 116,
Leaky surface acoustic waves (SAW), 318–320 166, 175, 193, 205, 225, 233, 240, 292,
Levels, release rate, 38, 54–55, 58–59, 72–73, 307, 329
88, 90, 390 Mechanical properties of cell wall, 7, 113, 240
Life Medium density fiberboard (MDF), 26, 85,
prediction, 390 87–90, 229, 340–341, 343
Lignin, 6–7, 63, 91, 126, 130, 135, 173, 181, Microcracks, 20, 56, 59, 64, 69, 127, 133, 136,
185, 187, 189, 193, 218–220, 233, 235 185, 291–293, 295, 298, 337, 380
Linear elastic fracture mechanics (LEFM), 53, Microdefects, 35
56, 60, 74, 92 Microfibril angle, 6, 79, 82, 113, 129–131,
Load 134, 219, 240, 242, 260
displacement curves, 59, 64, 67, 74–75, 79, Microscopy
134, 297 confocal, 134, 138
modes, 64 electron, 81, 131, 136, 138–139, 146,
rate, 38, 60, 71–72 149–152
Local, 20, 23–24, 27, 29, 33, 37, 40, 54, 71, light, 9, 21–22, 24, 131, 137–138, 147,
130, 136, 185, 241, 256, 276, 288, 317, 149–152
328, 341, 366, 370, 373 Microstructural, 7, 56, 75, 91
residual stress, 3, 87–88, 112 Micro-voids, 149, 159, 165
Localization, 287, 289, 292, 303, 313 Microwaves, 159–170, 200, 227, 229, 235,
Longitudinal, 6–7, 18, 20, 41, 53, 77, 79, 113, 343, 390
115, 125, 129, 131–137, 159, 161–162, Mineral bonded particleboard and fiberboard,
164–165, 169, 186, 205–206, 217–220, 308
Index 397
Point source, 36, 201 Reconditioning, 101, 112, 118, 165, 227–228,
Poisson’s ratios, 8, 271, 276, 280 247–250, 252
Polymers, 103, 119, 217–218 Reference, 11, 18, 20, 34, 39, 51, 54, 63,
Poor cure, 227 74–75, 103–104, 175–176, 191, 240,
Porosity, 34 255, 261, 281–282, 309, 311, 320, 325,
Prediction of life, 390 329, 371
Pressure, 18, 28, 104–105, 108, 111–112, stress, 18, 20, 34, 54, 74–75, 92, 104, 175,
159–163, 167, 192, 200–201, 215, 218, 240, 261, 282, 309, 311, 320, 371
227–228, 230, 233, 235, 238–239, 272, Reflection, 34, 176, 193, 311, 315, 318, 326,
379, 380, 390 343
sound, 380 elastic waves, 34, 311
Principle of superposition, 38 Refraction, elastic waves, 311
Processing, 12, 40, 119, 150, 160, 162, 167, Relative humidity, 26, 80, 173, 189, 199–200,
169, 198, 206, 220–221, 224, 228, 202, 206, 239, 244–245, 270, 273, 275,
237–253, 256, 288–289, 299, 303, 308, 366–367, 369, 377
343, 391 Reproducibility, 323, 332, 385–386
internal checking, 220, 237–253, 343 Residual
Production process, 11, 169, 217, 227, 229, strength, 363
236, 308, 323, 343 stress, 3, 87–88, 112, 363, 391
Propagation, 5, 7, 20–22, 34–35, 53–54, 56–60, Resin pockets, 125, 128, 166, 381–382,
63, 64, 66, 70, 74, 82–83, 87–88, 90, 384–385
92, 124–126, 134–136, 175, 183–184, Resistance, 3, 11, 24, 54, 57–60, 66, 75, 82,
217–218, 221, 232, 235, 255–256, 289, 88, 127–128, 132–133, 135, 191, 232,
291–292, 294–297, 301–303, 309–313, 234, 244, 270, 275, 307, 361
317, 319, 322, 327, 336, 339, 356, 363, Resonance, 316–317, 343–345
375, 381–383, 385, 390 Restraint cracking, 220
Pulp, 11, 22, 92, 133, 135, 216–217, 233, 246, Review, 39, 45, 51, 66, 155, 160, 169, 175,
256, 391 191, 217, 242, 290, 308, 344, 389
P wave, 314 Rise time, 288, 299, 302
Rock, 219
Q Rods, 77, 323, 328
Q factor, 193
Root mean square (RMS) voltage, 288, 302,
Quality assessment, 3, 256, 316, 379
382
Quantitative, 3, 34, 75, 148, 192, 246, 288,
Roughness, 75, 83, 146, 153, 292
303, 390
Round wood, 234
R Rupture, 7, 25, 53, 56, 59, 72, 83, 160–161,
Radial, 6–7, 18–19, 23, 25, 53, 75, 77, 81–84, 166, 168–169, 204, 231, 293, 295–296,
103, 124–131, 134–138, 159, 161–162, 337, 361–362
164–165, 167, 169, 179–180, 183–185,
188, 205–206, 216–262, 217–221, S
233–234, 242, 244, 256–259, 265–266, Safe design, safety, safe performance, 5, 9, 44,
269–270, 275, 279, 282–283, 295, 302, 91–92, 229, 232, 235, 308, 343, 345,
367–369 391
Radiations, 126, 145, 160, 167–168, 173, Safety factors, 91–92
177–178, 185–189, 233, 290, 317, 321 Samples, 82, 112, 127, 133–139, 165, 168,
Radiographic, 303 198, 225, 240–241, 246–248, 251, 260,
Ratios, 8, 180, 271, 276, 279–280, 294, 317 271–273, 275, 278, 366–367, 372–373,
Raw material, 9, 11, 215–217, 220–224, 227, 381–382, 385–386, 390
236, 307 Sampling rate, 382
Rayleigh, 320 Scanning electron microscopy (SEM), 21–24,
R-curve, 57–59, 75, 89–90 62, 72, 131, 134, 136–139, 146,
Reaction wood, 129–130, 179, 193, 220, 256, 149–155, 219, 297
261, 266, 297 Scanning modes, 325
Index 399