Delamination in Wood, Wood Products and Wood-Based Composites

Delamination in Wood, Wood Products and Wood-Based Composites

Voichita Bucur Editor

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 [email protected]

ISBN 978-90-481-9549-7 e-ISBN 978-90-481-9550-3 DOI 10.1007/978-90-481-9550-3 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2010938326 © Springer Science+Business Media B.V. 2011 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. 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. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Foreword

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 properties of wood and wood-based materials. Her previous books Acoustics of Wood and Nondestructive Characterization and Imaging of Wood are outstanding reference 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. KazemiNajafi – 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 graciously 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 evaluation of delamination in wood and wood-based materials, and mechanical stress development in the woody cell wall in response to various stressors. A vibrationbased 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 ultrasonics. 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 formation, 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 technical 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 literature. Professor Bucur has completed another significant contribution to the wood science literature. Project Leader USDA Forest Products Laboratory October, 2009

Robert J. Ross, Ph.D.

Preface

Delamination occurs in all man made composite materials as well as in natural composites 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 information 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 nondestructive 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 acting 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 acoustical 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 colleague 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. Melbourne, Victoria October 2009

Voichita Bucur

Acknowledgements

Permission for the figures cited in this book have been granted by Copyright Clearance Center, (http://www.copyright.com), by different organisations and colleagues 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 enthusiastic help in collecting documents and books cited in this volume. Melbourne, Australia

Voichita Bucur

ix

Contents

Part I

General Aspects

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Voichita Bucur

3

2 Terms for Delamination in Wood Science and Technology . . . . . Voichita Bucur

17

3 Delamination Detection – A Vibration-Based Approach . . . . . . Voichita Bucur

33

4 Initiation and Growth of Delamination in Wood and Wood-Based Composites, a Fracture Mechanics Approach . . Voichita Bucur 5 A Theoretical Model of Collapse Recovery . . . . . . . . . . . . . . Philip Blakemore Part II

51 101

Methodology for Delamination Detection and Factors Inducing and Affecting Delamination

6 Delamination of Wood at the Microscopic Scale: Current Knowledge and Methods . . . . . . . . . . . . . . . . . . . . . . . . Lloyd Donaldson

123

7 Probing the Wood Coating Interface at High Resolution . . . . . . Adya P. Singh and Bernard S.W. Dawson

145

8 Delamination in Timber Induced by Microwave Energy . . . . . . Georgiana Daian

159

9 Delaminations Induced by Weathering in Wood and Wood-Based Composites Panels . . . . . . . . . . . . . . . . . Voichita Bucur 10

Delamination in Timber Induced by Drying . . . . . . . . . . . . . Nawshad Haque

173 197

xi

xii

Contents

Part III

Delamination in Different Products

11

Industry Prospective of Delamination in Wood and Wood Products Chih Lin Huang

215

12

Internal Checking During Eucalypt Processing . . . . . . . . . . . Philip Blakemore

237

13

Acoustic Tomography for Tension Wood Detection in Eucalypts . . Voichita Bucur

255

14

The Hygroscopic Warping of Cross-Laminated Timber . . . . . . . Thomas Gereke, Per Johan Gustafsson, Kent Persson, and Peter Niemz

269

15

Acoustic Emission Activity Induced by Delamination and Fracture of Wood Structure . . . . . . . . . . . . . . . . . . . Voichita Bucur

287

Delamination Detection in Wood – Based Composites Panel Products Using Ultrasonic Techniques . . . . . . . . . . . . . . . . Voichita Bucur and Saeed Kazemi-Najafi

307

Delamination Evaluation of in-Service Glulam Beams and other Structural Members Via Ultrasonics . . . . . . . . . . . Ferenc Divos

353

16

17

18

Moisture Induced Stresses and Deformations in Parquet Floors . . Samuel Blumer, Erick Serrano, Per Johan Gustafsson, and Peter Niemz

19

Glue Line Nondestructive Assessment in Timber Laminates with an Air-Coupled Ultrasonic Technique . . . . . . . . . . . . . . Sergio J. Sanabria, Christian Müller, Jürg Neuenschwander, Peter Niemz, and Urs Sennhauser

20

365

379

From Present Researches to Future Developments . . . . . . . . . Voichita Bucur

389

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

393

Contributors

Philip Blakemore Division of Materials Science and Engineering, CSIRO, Clayton laboratories, Clayton South, VIC 3169, Australia, [email protected] Samuel Blumer b-h-e GmbH, Holzinnovationszentrum 1a, 8740 Zeltweg, Austria, [email protected] Voichita Bucur CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia, [email protected] Georgiana Daian The University of Melbourne, Department of Forest and Ecosystem Science, Melbourne, VIC 3010, Australia, [email protected] Bernard S.W. Dawson Wood and Biofibre Technologies, Scion Rotorua Te Papa Tipu Innovation Park, 49 Sala Street Whakarewarewa, 3010, Bay Of Plenty, New Zealand, [email protected] Ferenz Divos Faculty of Wood Science, University of West Hungary, Sopron, Hungary, [email protected]; [email protected] Lloyd Donaldson Bioproduct Development, Scion - Next Generation Biomaterials, 49 Sala St. Rotorua, Private Bag 3020, Rotorua 3046, New Zealand, [email protected] Thomas Gereke 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, [email protected] Per Johan Gustafsson Division of Structural Mechanics, Lund University, P.O. Box 118, SE-221 00 Lund, Sweden, [email protected] Nawshad Haque Division of Minerals, CSIRO Clayton, Bag 312, Clayton South, VIC 3169, Australia, [email protected] Chih Lin Huang Weyerhaeuser Technology Center, 32901 Weyerhaeuser Way S, Federal Way, WA 98001, USA, [email protected]

xiii

xiv

Contributors

Saeed Kazemi–Najafi Wood & Paper Science & Technology Department, Tarbiat Modares University, Noor, Iran, [email protected] Christian Müller Institute for Building Materials, Wood Physics, ETH Zürich, Schafmattstrasse 6, CH-8093, Zürich, Switzerland, [email protected] Jürg Neuenschwander Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland, [email protected] Peter Niemz Institute for Building Materials, Wood Physics, ETH Zürich, Schafmattstrasse 6, CH-8093, Zürich, Switzerland, [email protected] Kent Persson Division of Structural Mechanics, Lund University, P.O. Box 118, SE-221 00 Lund, Sweden, [email protected] Robert J. Ross Forest Products Research Laboratory One Gifford Pinchot Drive Madison, Madison, WI 53726, USA, [email protected] Sergio J. Sanabria Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland, [email protected] Urs Sennhauser Department of Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland, [email protected] Erik Serrano University of Vaxjo, Lucklings plats 1 SE 35195 Vaxjo, Sweden, [email protected] Adya P. Singh Wood and Biofibre Technologies, Scion Te Papa Tipu Innovation Park, Rotorua 3010, New Zealand, [email protected]

Part I

General Aspects

Chapter 1

Introduction Voichita Bucur

Contents 1.1 Background . . . . . . 1.2 Solid Wood . . . . . . 1.3 Wood-Based Composites 1.4 Summary . . . . . . . References . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

3 5 9 12 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 composite 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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_1, 

3

4

V. Bucur

Fig. 1.1 Delamination observed on cross sections of Douglas-fir laminated lumber. Note delamination 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)

between plies, debonding of wood–adhesive layers, or wood fibre fracture. Delamination, which is a debonding of two adjoining layers in the laminated wood-based composite, is probably the most frequently observed damage. Delamination can occur at several scales: Fig. 1.1 shows the cross section of Douglas – fir lumber laminates with macroscopic delamination, while at a submicroscopic scale, delamination can be observed between the S1 and S2 layers in spruce latewood tracheids, as can be seen in Fig. 1.2. Delamination may result from manufacturing errors, by imperfect bonding, by separation of adjoining piles, etc., or, during in service loading such as by accidentally excessive loading produced for example by snow or, by fatigue in cyclical environmental conditions of temperature and humidity.

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 composites 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 mechanisms, 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 prerequisite 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”.

1.2 Solid Wood Wood is a natural fibrous, layered composite which exhibits a remarkable combination of properties related to strength, stiffness and toughness (Vincent and Currey 1980; Schniewind 1989). As noted by Schniewind (1981) “wood is composed from

6

V. Bucur

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 principal 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 temperate 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

A more complex model was proposed by Mark (1981), and consists of a matrix and framework. The corresponding mechanical properties of the cell wall material can be derived from the natural polymer constituent (cellulose, hemicellulose and lignin) properties by the rule of mixture. The stiffness and strength of cellulose itself are considerable, the theoretical value for Young’s modulus and tensile strength being in the order of 250 GPa and 25 GPa (Mark 1967) respectively. Wood mechanical properties are considerably inferior to those of pure cellulose. Figure 1.4 shows the degradation of Young’s modulus from cellulose to wood. To E was analysed for several situations and the illustrate this aspect, the ratio σrupture smaller the ratio, the better the material will be in resisting crack propagation. In an ideal solid this ratio is in the order of 10, however this ratio is about 100 in longitudinal anisotropic direction of wood. The reduction of the Young’s modulus E from cellulose to wood is due to largely to the very complex structural arrangement of this material in which the microfibrillar angle plays a very important role. The development of computation techniques in the last 25 years, and the progress achieved in mechanical characterisation of solids in general and of composite materials in particular, affected positively the development of modelling of the wood structure. Gibson and Ashby (1988) proposed a cellular structure model with hexagonal cell shape and used for calculation the principles of cellular solid mechanics. Some improvements of this approach were given by Kahle and Woodhouse (1994) and Watanabe et al. (2000, 2002), which considered the cell wall material as transversely isotropic. Significant progress in Wood Science has been achieved using multiscale models which were elaborated by using three-dimensional finite element simulation of representative softwood related cellular models. In addition data related to the microstructural characteristics such as the micrifibril angle and the chemical composition of the cell wall such as lignin, hemicelluloses, water and crystalline cellulose were also integrated into their models (Harrington et al. 1998; Astley et al. 1998; Yamamoto 1999; Persson 2000; Watanabe and Norimoto 2000; Yamamoto et al. 2005; Hofstetter et al. 2005, 2006; Fritsch and Hellmich 2007). Using the experimental observations of wood behaviour at different scales, Hofstetter et al. (2007) proposed a very original approach considering simultaneously the continuum mechanics for the solid-type behaviour of the cell wall and on the other hand, the unit cell method, for the plate-type behaviour of the softwood microstructure. It was stated that the activation of different load-carrying mechanisms of cellular structure depends on the loading state of wood, such as for example: – the plate-type bending and shear deformations of the cell walls which are dominant 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 random/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 elastic 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

predictive capacity of this model could be substantially improved by using more accurate values of the elastic constants at a microscopic scale, which can be obtained with the development of specific acoustic microscopic technique as suggested by Bucur (2003). All these studies related to the modelling of wood structure clearly suggest that delamination can occur between different layers at submicroscopic, microscopic and macroscopic structural levels.

1.3 Wood-Based Composites With regards to the wood-based composites, the mechanical behaviour of two groups of products must be analysed: the laminated wood products such as glulam, plywood, laminated veneer lumber (LVL), parallel-strand timber (PSL), structural particleboard, oriented strandboard (OSB), the fibre-based products such as fibreboard particleboard, paper and fiber reinforced composite such as fibre-cement boards, carbon fibre-reinforced plywood, and wood and glass-fibre composites, paper, etc. Performance criteria for wood-based composites relate directly to product end use. Laminated products are frequently used for structural purposes. This requires consideration of engineering strength needs, safety and short and long term response of the material to the service environment. Structural, exterior-grade products have the most demanding bond-quality requirements, since glue line failure could be catastrophic to these structures. In these situations glue line strength, durability and reliability must be assured, by computational analysis and bond quality testing programs. Computational models to simulate mechanical behaviour of new woodbased composites are critically needed because of cost-effectiveness. The effects of varying raw material characteristics on the mechanical properties of prospective new products can be thoroughly analysed. The intensive and expensive bond quality testing programs also can be improved by modeling. The factors affecting the quality of adhesion in wood-based composites are related to the heterogeneous and anisotropic character of wood reflected in the anatomical characteristics, permeability, density and moisture content, fibre bonding sites, and on the other hand in the nature of adhesives (thermosetting or thermoplastic). As noted by Schniewind (1981) “bond formation depends upon the development of physical and chemical interactions both within the bulk adhesive polymer and at the interface between adhesive and wood. Interactions within the adhesive accumulate to give cohesive strength while the forces between adhesive and wood provide adhesive strength. Both should exceed the strength of the wood allowing substantial wood failure during destructive testing of high-quality bond”. Optimum bond formation requires intimate contact between adhesive and wood substrates to ensure macromolecular interaction over a large area. Different techniques (X-ray, NMR, microindentation, etc.) were developed for the mechanical characterisation of the wood-adhesive interface. Figure 1.5 shows the light microscopy image of a spruce parallel-strand lumber specimen which contains fractured

10

V. Bucur

Fig. 1.5 Transverse section of spruce parallel-strand lumber which contains fractured and delaminated zones. The arrows indicate the zones tested via nanoindentation (Konnerth and Gindl 2006, Figure 1)

and delaminated zones tested with the nanoindentation technique developed by Konnerth and Gindl (2006). Modelling of the mechanical behaviour of laminated wood composites for predicting elasticity and strength has been reported for more than 30 years in numerous articles. A small snapshot of these include: Hunt and Suddarth (1974) who predicted the Young’s modulus and the shear modulus of medium-density flakeboard, Okuma (1976) studied the plywood properties influenced by the glue line, Gerrard (1987) proposed an equivalent orthotropic elastic model for the properties of plywood, Shaler and Blakenhorn (1990), Wang and Lam (1998) or Lee and Wu (2003) predicted the mechanical properties of oriented flakeboard. The mechanical behaviour of laminated veneer lumber, LVL, has been studied by Bejo and Lang (2004), Castro and Paganini (2003), Hata et al. (2001), Kamala et al. (1999), Lang et al. (2003), Park and Fushitani (2006). Finite element modelling of laminated wood composites as a multilayer system was proposed by several authors (Triche and Hunt 1993; Suo and Bowyer 1995; Clouston et al. 1998; Morlier and Valentin 1999; Nafa and Araar 2003; Wu et al. 2004) for predicting tensile, compression or bending strength and stiffness using failure criteria. Clouston and Lam (2001, 2002) and Clouston (2007) proposed an advanced methodology for analysing the multiaxial stress states in small specimens of parallel wood-strand composites, using a 3D non-linear stochastic finite element model and Monte Carlo simulations. The Tsai-Wu strength theory to predict the ultimate load carrying capacity of a centre point off-axis bending member made from Douglas fir laminated veneer, incorporating the size effect was reported by Clouston et al. (1998).

1

Introduction

11

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 weatherability 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 different 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 fiberreinforced 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 – hardwood 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 products, forest waste and construction waste materials has given rise to the development of new or improved technological processes such as the water vapour explosion process. 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 testing 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)

low cost processes. Durability of these products is related to matrix formulations, processing methods and curing regimes. If natural fibre reinforced cement products are to be readily available for low cost housing much research still remains to be conducted for improving the durability of the products.

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 structural 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 development 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 woodstrand 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 conference 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 performance 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

Hofstetter K, Hellmich C, Eberhardsteiner J (2007) Micromechanical modelling of solid-type and plate-type deformation patterns within softwood material. A review and an improved approach. Holzforschung 61:343–351 Hofstetter K, Hellmich C, Eberhardsteiner J (2006) The influence of the microfibril angle on wood stiffness: A continuum micromechanics approach. Comput Assisted Mech Eng Sci 13: 523–536 Hofstetter K, Hellmich C, Eberhardsteiner J (2005) Development and experimental validation of a continuum micromechanics model for wood. Eur J Mech Solid 24:1030–1053 Hunt MO, Suddarth SK (1974) Prediction of elastic constants of particleboard. Forest Prod J 24(5):52–57 Jeronimidis G (1980) Wood, one of nature’s challenging composite. In: Vincent JFV, Currey JD (eds) “The mechanical properties of biological materials”. Cambridge University Press, London, pp 169–182 Kahle E, Woodhouse J (1994) The influence of cell geometry on the elasticity of softwood. J Mater Sci 29:1250–1259 Kamala BS, Kumar P, Rao RV, Sharma SN (1999) performance test of laminated veneer lumber (LVL) from rubber wood for different physical and mechanical properties. Holz Roh- Werkst 57:114–116 Kelly A (ed) (1989) Concise encyclopedia of composite materials. Pergamon, Oxford Konnerth J, Gindl W (2006) Mechanical characterization of wood-adhesive interphase cell walls by nanoindentation. Holzforschung 60:420–433 Lang EM, Bejo L, Divos F, Kovacs Z, Anderson RB (2003) Orthotropic strength and elasticity of hardwoods in relation to composite manufacture. Part III. Orthotropic elasticity of structural veneers. Wood Fiber Sci 35:308–320 Lee AWC, Hong Z, Phillips DR, Hse CY (1987) Effect of cement /wood ratios and wood storage conditions on hydration temperature, hydration time and compressive strength of wood – cement mixtures. Wood Fiber Sci 19:262–268 Lee JN, Wu Q (2002) In – plane dimensional stability of three-layer oriented strandboard. Wood Fiber Sci 34:77–95 Lee JN, Wu Q (2003) Continuum modelling of engineering constants of oriented strandboard. Wood Fiber Sci 35:24–40 Lenth CA, Kamke FA (1996) Investigations of flakeboard mat consolidation. Part I. Characterizing the cellular structure. Wood Fiber Sci 28:153–167 Lopez-Anido R, Gardner DJ, Hensley JL(2000) Adhesive bonding of eastern hemlock glulam panels with E-glass / vinyl ester reinforcement. Forest Prod J 50, 11/12:43–47 Mark RE (1981) Molecular and cell wall structure of wood. In: Wangaaed FF (ed) Wood: Its structure and properties. Educational Modules for Material Science and Engineering Project. Pensilvania State University, University Park, Pensylvania, USA, pp 43–100 Mark RE (1967) Cell wall mechanics of wood tracheids. Yale University Press, New Haven, Connecticut Morlier P, Valentin G (Eds) (1999) Damage in wood. COST Action E8, Bordeaux Moslemi AA, Pfister S (1987) The influence of cement/wood ration and cement type on bending strength and dimensional stability of wood-cement composite panels. Wood Fiber Sci 19: 165–175 Moulin JM, Pluvinage G, Jodin P (1990) FGRG : Fiberglass reinforced gluelam – a new composite. Wood Sci Techn 24:289–294 Murata K, Masuda M (2006) Microscopic observation of transverse swelling of latewood tracheid: Effect of macroscopic/mesoscopic structure J Wood Sci 52:283–289 Nafa Z, Araar M (2003) Applied data for modelling the behaviour in cyclic torsion of beams in glued-laminated wood: Influence of amplitude. J Wood Sci 49:36–41 Ogawa H (2000) Architectural application of carbon fibers. Development of new carbon fiber reinforced glulam. Carbon 38:211–226 Okuma M (1976) Plywood properties influenced by the glue line. Wood Sci Techn 10:57–68

1

Introduction

15

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 composites. 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 characteristics 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 2.1 General Terms . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Terms for Delamination in Solid Wood . . . . . . . . . . . . . . 2.3 Terms for the Delamination in the Cell Wall . . . . . . . . . . . . 2.4 Terms for the Delamination in Laminated Wood Products . . . . . 2.5 Terms for the Delamination in Wood-Based Fibre and Particle Panels 2.6 General Classification of Delamination . . . . . . . . . . . . . . 2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

17 18 19 19 25 26 29 29

2.1 General Terms In Material Science, delamination is defined as a sub critical damage to the interfaces between the plies in a laminate composite that causes a reduction in the load carrying capacity of composite (Morris 1992). The terms which describe delamination in wood and wood- based composites are very numerous and often confusing due to a multitude of reasons (the use of terms which were considered inappropriate in recent days, new technologies related to microscopic observation of the structure, etc). A comprehensive understanding of these terms is essential for the uses of wood products under competitive conditions of modern technology. This chapter discusses the terms that refer to delamination in solid wood, in wood cell wall, in laminated products, and in fibrous and particle board wood-based composites.

V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_2, 

17

18

V. Bucur

2.2 Terms for Delamination in Solid Wood Terms which express delamination in solid wood have been defined in very well known reference textbooks (Kollmann and Cote 1968; Panshin and de Zeeuw 1980) and standards under the label of defects which develop in wood after it has been cut. In what follows we quoted the terms as referred in ASTM D9 – 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

of the pith, occasionally moving radially to an adjacent latewood ring. A synonym is “cup shake”. Meyer and Leney (1968) described ring shakes from standing conifer trees as compound middle lamella failures, usually in latewood, with loose fibres and deposites of extraneous material on their shake surface. ◦ Handsplit and resawn shakes – a shake having a split face and a sawn back ◦ Tapersplit shake – a shake having two split faces and a natural shingle like taper ◦ Straightsplit shake – a shake having two split faces and with no pronounced taper – Split a separation of the wood parallel to the fiber direction, due to the tearing apart of the wood cells.

2.3 Terms for the Delamination in the Cell Wall 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 numerous 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 phenomena 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 structure inspection”. One can speculate about the contribution of new technologies for higher resolution microscopy in relation to wood ultrastructure which influence its mechanical behaviour.

2.4 Terms for the Delamination in Laminated Wood Products Structural laminated products include plywood, various composites of veneer and of wood based laminates such as laminated veneer lumber, glued laminated lumber, wood fibre-reinforced polymer composites, etc. Plywood as defined in ASTM D 1038 – as “usually a crossbanded assembly made of layers of veneer or veneer in combination with a lumber core or other woodbased panel material jointed with an adhesive. Plywood is generally constructed of an odd number of layers with grain of adjacent layers perpendicular to one another. Outer layers and all odd-numbered layers generally have the grain direction oriented parallel to the long dimension of the panel”.

20

V. Bucur Table 2.1 Nomenclature of cell wall deformation as referred in publications until 2007

Term

Author

Buckle

Robinson (1920)

Buckling

Scurfield et al. (1972) Kucera and Bariska (1982)

Buckling

Buckling of the cell wall

Wilkins (1986)

Cell wall crinkle

Green (1962)

Cell wall fold

Green (1962)

Occurrence

Thin walled cells and tissues

Observable at microscopic level

Common in trees and stressed wood

Compression crease Wilkins (1986) refereed as microscopic Compression crease, Wilkins (1986) refereed as macroscopic Compression wood Keith (1974) and microscopic compression failure

Observable at microscopic level Observable with naked eye

Corner crinkle

Green (1962)

Corrugation

Green (1962)

Crack arrested

Thuvander and Berglund (2000)

Crack growth and microcracks

Dill-Langer et al. (2002)

Crack propagation

Fruhmann et al. (2003) Robinson (1920)

Observable with light microscope Observable with light microscope Observable with light microscope Observable with confocal Laser Scanning Microscope Observable in ESEM Spruce wood

Crinkle

Description Buckling of cell, equivalent to Brush’s (1913) term “bending” Buckling of fibres precedes macroscopic failure Deformation characterized by coarse transverse folds and longitudinal cracks in the inner cell wall layers Reference of the level of observation must be made Structural deformation of cell wall frequently referred to as slip planes Cell wall distortion or discontinuity which is more pronounced than a cell wall crinkle. Produced by growth stress in trees or in wood by applied perpendicular stress. Horizontal rows of slip planes

Horizontal rows of slip planes

Deformation with a marked resemblance to Scurfield et al. (1972) defined as “wrinkling of the cell wall linings Minute crinkle originating in, or confined to, the thickenings at the corners of tracheids Cell deformations ranging from smooth undulations to sharp peaked folds crack propagation stopped in the latewood Progress of crack in early wood

Crack initiation and propagation in earlywood Local or horizontal bands of cell wall “crinkles”

2

Terms for Delamination in Wood Science and Technology

21

Table 2.1 (continued) Term

Author

Occurrence

Description

Crinkling

Bienfait (1926)

In earlywood during early gross failure

Permanent “telescoping deformation”

Fine microscopic crease, notes C1

Dinwodie (1968)

Fine microscopic crease, notes C2 Fold Fine and transwall fractures First step of crack propagation Fracture surface

Dinwoodie (1968)

Gross compression failures with light microscopy

Gross microscopic crease C3 Initial compression failures Interwall deformations Irregular crack profile Macroscopic buckling

Macroscopic compression crease Macroscopic compression failure

Macroscopic compression failure lines or creases

Robinson (1920) Donaldson (1995)

1 to 2 slip planes in depth, covering more than 2 cell wall in width 3 to 6 slip planes in depth Spruce earlywood Folding of cell Radiate pine Observations with SEM

Sippola and Pinus sylvestris Fruhmann (2002) Reiter et al. (2002) Different species Reiter and Sinn (2002) Bienfait (1926), Tissue with initial Dadswell and failure present, Langlands (1934) also occurs in both heart and truewood of Eucaliptus diversicolor Dinwoodie (1968) Bienfait (1926) Côté and Hanna (1983) Vasic and StanzlTschegg (2007◦ ) Côté and Hanna (1983)

Numerous in earlywood Observable with SEM observable in ESEM oak Observable with SEM

Dinwoodie (1966)

Keith (1974)

Observable with SEM

Kucera and Bariska (1982)

Observable with SEM

Observations with SEM Observations with SEM

Continuous deformation formed after initial failure development

More than 5 slip planes in depth The lining up of slip planes to form definite zones of failure Slip planes as described by Keith and Côté (1968) fracture through vessels Buckling of fibres which is preceded by cell wall deformation and related to slip plane formation Horizontal zone of dislocations, produced by failures in adjacent cell wall. Involves the development of shear planes, buckling of whole fibres and is normally preceded by slip plane development and microscopic compression creases Deformation visible to the naked eye. Multilayered accumulation of the structural deformation pattern

22

V. Bucur Table 2.1 (continued)

Term

Author

Macroscopic compression lines

Kisser and Observable with Steininger (1952) light microscopy Dinwoodie (1966)

Microscopic compression crease

Occurrence

Microscopic compression crease or line

Dinwoodie (1966)

Microscopic compression failure Microscopic compression failure lines or creases

Keith and Côté (1968)

Observable with SEM

Kucera and Bariska (1982)

Observable with SEM

Description Enlarged microscopic compression line which is visible to the naked eye Severe crinkling of the cell walls, produced by increased loading following microscopic compression crease formation Distinct rows of dislocations. The second stage in cell wall failure following slip plane formation. May develop independent to slip planes. Closely associated converging or crossing slip line

Microscopically visible changes in cell shape as buckling and/or telescopic shortening, type S or U. Type S : double bending, type U – triple bending Microscopic Kisser and Common in most Progression of slip lines compression lines Steininger (1952) wood species horizontally, from fibres to fiber Minute compression Dadswell and Common in brittle A lining up of failures in failure Langlands (1934) heart adjacent cell wall, and produced by incipient decay Minute dislocation Dinwoodie (1966) Common in wood Slip line species Multiple slip plane Scurfield et al. Synonymous with “creases” (1972) defined by Dinwoodie 1968 as areas where varying numbers of slip planes are concentrated. The next stage following slip plane formation. Severe type of cell wall fold. Node Green (1962) Common in Point of flexing in pupl pulped tracheids tracheids and stressed wood Bending of fibres from the Offset Bienfait (1926) In latewood during early original axial line gross failure Pre-crack Boatright and Macroscopic Crack propagation normal to the Garrett (1983) crack extension plane of the pre-crack in LT occurring parallel to the grain Predominant fracture Donaldson (1995) Radiate pine Differences in fracturing at S1 /S2 boundary behaviour

2

Terms for Delamination in Wood Science and Technology

23

Table 2.1 (continued) Term

Author

Radial dislocations

Chafe (1977)

Slip planes, and slip lines

Kucera and Bariska (1982)

Occurrence

Observable with SEM

Common in most Robinson (1920), Slip lines, slip wood species Bienfait (1926), planes, as defined Dadswell and between 1920 and Langlands (1934), 1952 (observed Wardrop and with light Dadswell (1947); microscope) Kisser and Steininger (1952) Slip plane in relation Dinwoodie (1968) to microfibrils

Slip plane with SEM Wilkins (1986) Slip plane, defined using SEM, after 1960

Keith and Côté (1968), Keith (1971)

Slip planes

Wilkins (1986)

Stress line and microfibrils

Scurfield et al. (1972)

Stress lines, or thrust Dinwoodie (1968) lines

Observable with SEM Observable with SEM

Observable with light microscope or with SEM

Description Radial dislocations found in the inner S2 and extending to the cell lumen. Removal of growth stress causes cell shortening and the closer packing of microfibrils changes the lumen surface from smooth to convoluted Local deformation or crinkling of cellulose fibrils in the whole of the cell wall of one or two neighbouring cells, without prominent change in shape of cell Fine crack lines in the cell wall, preceding buckling, crinkling and tension failure. A crinkle in the cell wall. Fine streaks intercrossing at a certain angle and extending through the secondary wall Dislocation or crinkling of the fibrils comprising the cell wall occurring either singly or in pairs All deformation observed as wrinkled transverse lines Non-crossing single line cell wall deformation. Possibly a stage in microscopic compression failure and not always a sectioning artefact. The scale of observation and the type of microscope used must be defined Pre-cursor of slip plane. Barely detectable cellulose microfibril deformation Precursor of slip planes. Slight dislocation virtually unobservable by polarization microscopy or staining, but observable by electron microscope

24

V. Bucur Table 2.1 (continued)

Term

Author

Occurrence

Description

Stress lines or thrust lines

Kucera and Bariska (1982)

Observable with SEM

Telescopic shortening Thrust line

Kucera and Bariska (1982) Kisser and Frenzel (1950)

Mainly thick walled cells Common in most wood species

Wrinkling in the cell wall towards the lumen, affecting “cell wall lining” A diversion of cells from their natural axial orientation Slight local thickenings of the cell walls due to small deformation of the fibrils. Pre-slip plane May be considered morphologically similar

Thrust lines, slip Wilkins (1986) planes, compression creases Thrust-line or stress Wilkins (1986) lines describe only those slip planes not observable with light microscopy Transverse fracture Sell and surface Zimmermann (1998) Wrinkling of cell wall

Scurfield et al. (1972)

Observable with light microscope or with SEM Observable only with SEM

High resolution FE -SEM

Pre-slip plans deformation, which are not distinguishable from slip planes when using SEM .

Poly-laminated concentric structure of the cell wall layers observed in transverse surface Involves only the covering lining the lumina of fibres. It is a stage after multiple slip plane formation in the sequence of events occurring during axial compression

Delamination effects in plywood, as defined in ASTM D 1038 are noted below: – Blister in plywood is an elevation of the surface of an adherend (separation between plies) somewhat resembling in shape a blister on the human skin; its boundaries may be indefinitely outlined and it may have burst or become flattened. – Broken Grain (shelling, leafing, grain separation) a separation on veneer surface between annual rings. – Closed Surface Checks – Delamination – the separation of layers in a laminate because of failure of the adhesive, either in the adhesive itself or at the interface between the adhesive and the adherent – Durability as applied to the glue bond – its resistance to deterioration related to exposure conditions – see also delamination – Gap – an open joint or split in the inner plies which results when cross band or centre veneers are broken or not tightly butted – Open Joint – failure of bond or separation of two adjacent pieces of veneer so as to leave an opening, usually applied to edge joints between venerers

2

Terms for Delamination in Wood Science and Technology

25

– Rupturing of some of wood structural elements which forms cavities of various sizes in radial plane – Skips or Voids in the glueline of plywood – Starved Joints – a glue joint that is poorly bonded because of an insufficient quantity of glue. – Sunken joint in plywood – a depression in the surface of the face ply directly above the edge joint in a lumber core or crossband. Usually the result of localized shrinkage in the edge-jointed layer. – Wood Failure ◦ as applied to plywood glueline testing – the area of wood fiber remaining at the glueline following completion of the specific shear test. Determination is by means of visual examination and expressed as a percent of the test area ◦ As applied to failure in plywood not directly associated with the adhesive, is a rupture, shelling, tearing or breaking of wood itself. 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: – – – – – –

the shear through the thickness of structural panels (ASTM D 2719) the shear modulus of wood based laminated structural panels (ASTM D 3044) the toughness of wood based structural panels (ASTM D 3499) the stresses for structural glued-laminated timber (ASTM D 3737) the stresses for structural composite lumber products (ASTM D 5456) the accelerated aging test (ASTM 1037, Chapter 7) for the ability of the material to withstand severe environmental exposure conditions .

2.5 Terms for the Delamination in Wood-Based Fibre and Particle Panels ASTM D 1554 gives the terms related to wood-based fibre and particle panels defined “as a group of board materials manufactured from wood or other lignocellulosic fibres or particles to which binding agents and other materials may be added during manufacture to obtain or improve certain properties”. Under the generic name of wood-based fibre and particle panels, two types of panels are included: the fibrous – felted panels and the particleboards. Fibreboard panels – is “a board generic term encompassing sheet materials of widely varying densities

26

V. Bucur

manufactured from refined or partially refined wood fibres”(for example: medium density fibreboard (MDF) having the density between 400 and 800 kg/m3). Particle board is composed of particles such as: chips, flakes, strands, wood – wool, etc. “The particle is the aggregate component of a particleboard manufactured by mechanical means from wood or other lignocelullosic material, comparable to the aggregate in concrete”. The delamination effect in wood-based fibre and particle panels is noticed in: – ASTM D 1037 and is related to the shear test in compression, to the interlaminar shear test and to the edgewise shear test which is a shear test normal to the plane of the board – ASTM D 1038 which recommends the accelerated aging test “used to obtain a measure of the inherent ability of a material to withstand severe exposure conditions. The cycling exposure to which the material shall be subjected is a simulated condition developed to determine relatively how a material will stand up under aging conditions” of high temperature and high relative humidity. The determination of the cohesive bond strength of the fibres or particles on the surface of wood – base fibre and particle panels in the direction perpendicular to the plane of the panel is regulated by ASTM D 5651.

2.6 General Classification of Delamination The myriad of terms related to delamination in wood and wood–based composites, presented previously required a new classification, which can support a more general mechanical approach related to delamination initiation and growth. In the following we propose to follow the classification suggested by Bolotin (1996) for engineering artificial composites. The criterion of this classification is the position of delamination into the member, such as: internal delamination, near surface delamination and delamination producing multi-cracking of the member (Fig. 2.1).

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

In the context of structural health monitoring, the mechanical behaviour of composites with delamination can be studied with linear or non linear Fracture Mechanics, with nondestructive ultrasonic methods and with model dependent methods implemented by finite element analysis, which are able to provide local and global damage information. The internal delamination will be referred to as delamination observed as cracks and studied with Fracture Mechanics. Internal delamination can be detected in solid wood as well in as in wood-based composites at submicroscopic, microscopic and macroscopic scale. For example: in solid wood, between the middle lamella and the other cell wall layers or between the S1 and S2 layers or between S3 and G layers as frequently observed for compression wood or tension wood (Fig. 2.2). At macroscopic scale the delamination occurs in the annual ring between zones of different densities such as earlywood and latewood, or earlywood and medullary rays. In wood-based composites such as fibreboards, the fibre adhesive interaction during manufacturing is random and sometime the fibres remain attached in bundles, the middle lamella is degraded and large voids between the fibres can be observed (Fig. 2.3).

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 middle 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 members 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 multiple 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 In Material Science, delamination is defined as a sub critical damage to the interfaces between the plies in a laminate composite that causes a reduction in the load carrying capacity of composite (Morris 1992). The terms which describe delamination in wood and wood- based composites are numerous and often confusing for multiple reasons (the use of terms which were previously considered inappropriate, new technologies related to microscopic observation of the structure, etc). A comprehensive understanding of these terms is essential for the uses of wood products under competitive conditions of modern technology. A new classification of the delamination in wood, wood products and wood-based composites is proposed, depending on its position in the member, such as: internal delamination, near surface delamination and delamination producing multi-cracking of the member. Fracture Mechanics is an useful tool for the study of initiation of cracks and growth of delamination in wood and wood-based composites. In the context of structural health monitoring, the detection of damage induced by delamination in wood-based composites can be achieved with nondestructive ultrasonic methods and with model dependent methods implemented by finite element analysis. These methods are able to provide local and global damage information, as can be seen in Chapter 3.

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 mechanical 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 deformations 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 3.2 3.3

Introduction . . . . . . . . . . . . . . . . . . . . Delamination Detection with an Ultrasonic Technique Delamination Detection with a Model-Based Method 3.3.1 Linear Behavior . . . . . . . . . . . . . . . 3.3.2 Nonlinear Behaviour . . . . . . . . . . . . . 3.4 Some Practical Aspects . . . . . . . . . . . . . . 3.5 Summary . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

33 33 38 38 40 44 45 46

3.1 Introduction In this chapter we shall describe the aspects related to delamination in composites revealed by a vibration-based approach and related to the local and global damage detection. The local damage detection is performed with an ultrasonic technique, with Lamb waves, while the global damage detection is based on a model – based method using low frequency vibrations and undertaking the analysis of structural models implemented by finite element analysis. In this chapter the delamination detection studies are commented in the context of structural health monitoring, which is referred as the process of implementing a damage detection strategy for mechanical engineering infrastructures or for other purposes.

3.2 Delamination Detection with an Ultrasonic Technique Interfaces play an important role in determining the performance of laminated composite materials on a wide variety of scales, from interlaminar bonds to adhesive V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_3, 

33

34

V. Bucur

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 purpose 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 properties, 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 characteristics of the structure by measuring the stiffness matrix, the viscoelastic parameters, etc. determined with ultrasonic waves. Theoretical models for plane wave propagation 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 propagation, 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, porosity, 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 variation 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, second, 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 possibility 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 conversion 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)

3.3 Delamination Detection with a Model-Based Method Successful application of damage detection and health monitoring of structures using the measured structural dynamic response and mathematical models has been possible with the advance in computer science and technology. Compared with the nondestructive testing and evaluation procedures, the model-based methods using low frequency vibrations have a more rigorous mathematical background, but also several limitations related to the interpretation of the physical meanings of the detected results and the precise numerical representation of the structures. The mechanical behavior of a damaged structure can be studied in two hypotheses, the linear or the non linear mechanical behavior. In that follows both aspects will be succinctly described.

3.3.1 Linear Behavior Model – based methods implemented by finite element analysis under static or dynamic loading, assume that the linear monitored structure responds can be accurately described by finite element analysis. It is assumed that the behavior of the structure is linear before and after damage. The composites are usually modeled as beams (Euler beam, Timoshenko beam) with through-width delaminations parallel to the beam surface located arbitrarily, or shells. Kim et al. (1997) proposed an analytical solution for predicting delamination buckling and growth of a thin fiber reinforced plastic layer in laminated wood beams under static bending. It was noted that the delamination growth is related to an explicit form of strain-energy release rate and the critical load can be accurately predicted. Simulation of the delamination indicated an unstable growth of the delamination after buckling of the delaminated sub-laminate, followed by arrested delamination growth. For the vibrating beams, the foundation of linear analysis is based on the concept of linear normal mode and the principle of superposition. Linear normal modes

3

Delamination Detection – A Vibration-Based Approach

39

are synchronous harmonic particular solutions of the homogeneous linear system (Vakakis 1996). The vibration phenomena of beams have been studied for the case of a single or multiple delamination through the beam thickness. Model-based methods under dynamic loading have been extensively commented and remarkably reviewed, periodically, by numerous authors (Salawu 1997; Doebling et al. 1997, 1998; Zou et al. 2000; Sohn et al. 2003; Montalvao et al. 2006). The dynamic model - based methods use changes in vibrational modal properties (i.e. modal frequencies, modal damping values and mode shapes) to infer changes in mechanical properties of the structure. The impulse or continuous excitation techniques can be used for vibrating the structure. Commonly hammer technique is used for impulse excitation (Fig. 3.6).The utilization of a non-contact scanning laser vibrometer system allows acquisition of a large number of measurement points for a better definition of the mode shapes. Continuous sine excitation can be produced by using PZT – lead-zirconate-titanate - ceramic wafers as actuator (Fig 3.7). The dynamic model – dependent methods can be subdivided into: modal analysis, frequency domain, time domain and impedance domain, according to the dynamic response parameters analyzed. Frequencies, mode shapes, curvature mode shapes and modal damping, which are function of the physical properties of the structure (mass, damping and stiffness), are the most commonly measured parameters, when the dynamic model - based methods are used. Modification of physical properties of the structure, such as for example reduction of stiffness resulting from cracks or delamination, will implicitly cause detectable changes in modal parameters. Furthermore, these changes must be used as indicators of damage, and the process of vibration - based damage detection reduced to some form of pattern recognition problem, as can be seen from the references cited below and extracted from a huge literature (Adams et al. 1978; Cawley 1990; Cawley and Adams 1979, 1987; Wang et al. 1982; Tracy and Pardoen 1989; Nagesh and Hanagud 1990;

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 frequencies 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. For damage identification and health monitoring of structures, many different issues are critical, such as: the excitation and measurement configurations, the selection of the type of sensors and their location, the signal processing performing such as: Fast Fourier analysis, time – frequency analysis, or wavelet analysis (Castro et al. 2007).

3.3.2 Nonlinear Behaviour 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

are defined as some synchronous periodic particular solutions of the homogeneous nonlinear system under which all degrees of freedom undergo the extreme position at the same time (Vakakis 1996). The most frequent nonlinearities in a delaminated beam are introduced by: the nonlinear geometric effects such as axial stretch effects; the deflection – dependent interactions in both longitudinal and transverse directions; the intermittent contacts between the segments during vibration, the delaminated segments constraining the movement of each other or the growth of delamination. From the existing literature in this field several articles were selected to illustrate the approach with nonparametric – models (Tseng and Dugundji 1971; Abhyankar et al. 1993; Hanagud and Luo 1994, 1997; Gammadi and Hanagud 1995; Nayfeh et al. 1995; Luo and Hanagud 1997a, b, c, 2000; Lestari and Hanagud 2001; Lu et al. 2001; Caron et al. 2006; Wang and Yu 2006; Perel 2006; Wang and Yu 2006; Friswell 2007; Wang and He 2007). The existing of the “delamination modes” was demonstrated by Hanagud and Luo (1994) and Luo and Hanagud (2000). Modeling of the delamination effects is shown in Fig. 3.8. After delamination the composite beam is represented as a combination of four beams connected at the delamination boundaries, having the characteristics denoted in the previous figure. Luo and Hanagud (2000) noted that the effect between the laminated surfaces depend on the relative position between the sublaminates during vibration. Some constraints between the upper and lower delamination surfaces still exist. Under a small amplitude vibration of the delaminated beam at a frequency corresponding to a delamination opening mode, the effect between delaminated sublaminates can be modeled as a distributed soft spring between them. When the amplitude exceeds a certain level, the spring effect becomes zero because the delamination opens beyond the small amplitude constraints. On the other hand, when the vibration mode does not tend to open

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 delaminated 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 displacement 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: – Gummadi and Hanagud (1995) for vibration characteristics of beams with multiple delaminations – Lestari and Hanagud (1999) for multiple delamination dynamics in composite beams, using the Euler – Bernoulli beam theory in connection with piecewise – linear springs to simulate the open and closed behaviour between the delaminated layers. – Della and Shu (2005), which studied the case of a beam with overlapping delaminations.

3

Delamination Detection – A Vibration-Based Approach

43

Fig. 3.10 Comparison between the experimental data and the prediction data based on the nonlinear 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

– Sridharan (2008) – for delamination behaviour of composites. (Please note that these recommendations reflect only my opinion when these pages have been written). Chattopadhyay et al. (1999) reported the nonlinear response of a delaminated smart composite cross – ply beams. The theory is implemented through finite element method including nonlinear induced strain effects. The numerical results indicate changes in the dynamic responses of the beam due to dilamination.

44

V. Bucur

Fig. 3.11 The nonlinear response of a smart composite cross – ply cantilever beam with delamination at the first mode of vibration (Chattopadhyay et al. 1999, Figure 9)

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.

3.4 Some Practical Aspects As noted by Sohn et al. (2003) the implementation of a structural health monitoring systems must answer questions, related to the presence of damage and to the operational evaluation, such as: – 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. The structural health monitoring process of big wood laminated structures, in light of normal aging and degradation resulting from operational environments, must involve the periodic inspection of the system using:

3

Delamination Detection – A Vibration-Based Approach

45

– sampled dynamic response measurements from an array of transducers, establishing their number, resolution, bandwidth, data acquisition (periodically or continuously), storage and transmittal hardware; – extraction of the damage – sensitive features, normalization of data by the measured inputs or by environmental cycles (summer, winter); – statistical analysis of data to determine the current state of the system. 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 summarized 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: – the nondestructive evaluation method using an ultrasonic technique with Lamb waves, which is an experimental method able to provide local damage information – the model dependent method, undertaken analysis of structural models implemented by finite element analysis and able to provide global damage information, for linear and non-linear mechanical behavior of the system The structural health monitoring process of big wood laminated structures, in light of normal aging and degradation resulting from operational environments, must involve the periodic inspection of the system using: – sampled dynamic response measurements from an array of transducers, establishing their number, resolution, bandwidth, data acquisition (periodically or continuously), storage and transmittal hardware; – extraction of the damage – sensitive features, normalization of data by the measured inputs or by environmental cycles (summer, winter); – statistical analysis of data to determine the current state of the system.

46

V. Bucur

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 delaminated 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 modelling 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 delaminations. 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 dynamics. 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

Kazys R, Svilainis L (1997) Ultrasonic detection and characterization of delaminations in thin composite plates using signal processing techniques. Ultrasonics 35:367–383 Kessler SS, Spearing SM, Atalla MJ (2002a) In situ damage detection of composite structures using Lamb waves methods. Proceedings of the1st european workshop on structural health monitoring, pp 374–381 Kessler SS, Spearing SM, Atalla MJ, Cesnik CES, Soutis C (2002b) Damage detection in composite materials using frequency response methods. Compos:Part B 33:87–95 Kim Y, Davalos JE, Barbero EJ (1997) Delamination buckling of FRP layer in laminated wood beams. Compos Struct 37(3/4):311–320 Konstantinidis G, Wilcox P, Drinkwater B (2005) Damage detection using a distributed array of guided wave sensors. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 265 Krohn N, Stoessel R, Busse G (2002) Acoustic non – linearity for defect selective imaging. Ultrasonics 40:633–637 Kundu T, Maji A, Ghosh T, Maslov K (1998) Detection of kissing bonds by Lamb waves. Ultrasonics 35:573–580 Ladevèze P, Lubineau G, Marsal D (2006) Towards a bridge between the micro – and mesomechanics of delamination for laminated composites. Compos Sci Technol 66:698–712 Lavrentyev A, Rokhlin S (1998) Ultrasonic study of environmental damage initiation and evolution in adhesive joints. RNDE-Research in Nondestructive Evaluation 10, 1, 26 pages Lee S, Park T, Voyiadjis GZ (2003) Vibration analysis of multi – delaminated beams. Compos Part B:Eng 34:647–659 Lestari W, Hanagud S (1999) Health monitoring of structures: Multiple delamination dynamics in composite beams. Proceedings of the 40th AIAA/ASME/ASCE/AHS structures, structural dynamics and materials conference Lestari W, Hanagud S (2001) Nonlinear vibration of buckled beams: Some exact solutions. Int J Solids Struct 38:4741–4757 Lestari W, Qiao P, Hanagud S (2007) Curvature mode shape-based damage assessment of carbon/epoxy composite beams. J Intell Mater Syst Struct 18(March):189–208 Lu X, Lestari W, Hanagud S (2001) Nonlinear vibrations of a delaminated beam. J Vibr Control 7:803–831 Lucero J, Taha MMR (2005) A wavelet aided fuzzy damage detection algorithm for structural health monitoring. Proceedings of the 23rd international. modal analysis conference. (IMAX XXIII), Paper no 78 Luo H, Hanagud S (1996) Delamination modes in composite plates. J Aerospace Eng 9(4):106–113 Luo H, Hanagud S (1997a) An integrated equation for changes with structural dynamics of damaged structure. Int J Solids Struct, December:4557–4579 Luo H, Hanagud S (1997b) Dynamic learning rate neural network training and composite structural damage detection. AIAA J 35:1522–1527 Luo H, Hanagud S (1997c) Delaminated beam nonlinear dynamic response calculation and visualisation. Proceedings of the 38th AIAA/ASME/ASCE/AHS SDM Conference 1: 490–499 Luo H, Hanagud S (2000) Dynamics of delaminated beams. Int J Solids Struct 37(10):1501–1519 Maslov K, Kundu T (1997) Selection of Lamb modes for detecting internal defects in composite laminates. Ultrasonics 35:141–150 Messina A, Williams EJ, Contursi T (1998) Structural damage detection by a sensitivity and statistical based method. J Sound Vibr 216:791–808 Montalvao D, Maia NMM, Ribeiro AMR (2006) A review if vibration – based structural health monitoring with special emphasis on composite materials (2006) Shocks Vib Dig 38(4):1–6 Nagesh Babu GL, Hanagud S (1990) Delamination in smart structures – A parametric study on vibration. Proceedings of the 31st AIAA/ASME/ASCE/ AHS SDM Conference, pp 2417–2426 Nagy P (1992) Ultrasonic classification of imperfect interfaces. J Nondestr Eval 11: 127–139

3

Delamination Detection – A Vibration-Based Approach

49

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 monitoring. 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 beamforming 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 4.2

Introduction . . . . . . . . . . . . . . . . . . . . . . Links with Fracture Mechanics . . . . . . . . . . . . . 4.2.1 Linear Elastic Fracture Mechanics . . . . . . . . 4.2.2 Nonlinear Fracture Mechanics . . . . . . . . . . 4.3 Micro-structural Aspects in Wood . . . . . . . . . . . 4.4 Micro-structural Aspects in Wood-Based Composites . . 4.5 Fracture Mechanics Parameters for Ecological Relevance 4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

51 52 60 72 79 84 91 91 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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_4, 

51

52

V. Bucur

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 structural 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 process 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.

4.2 Links with Fracture Mechanics The initiation of delamination is due to the initiation and growth of cracks. As described by Williams (1989) the crack is “a planar discontinuity which is not capable of transmitting a load normal to its faces. When it grows, new surface area is created, which is of fundamental importance in determining behaviour”. The conditions for crack growth have been studied with Griffith theory (Griffith 1920) and with modern fracture theory using the concepts of linear or nonlinear Fracture Mechanics. Figure 4.1 gives the schematic representation of a crack, located in a plate (2D representation) and the corresponding two dimensional stress states. Any deformation of the crack can be described through a combination of three fracture modes (Fig. 4.2): – the opening mode in tension – Mode I – opening – the in plane shear mode – Mode II – sliding – the out of plane shear mode – Mode III – tearing shear

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

Fig. 4.2 Three fracture modes: Mode I – tension; Mode II – in plane shear; Mode III out of plane shear

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 orientations 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 intensity 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 mechanism 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 relationships 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: 1 dp = nπ



K σY

2

54

V. Bucur

where n = 1 for plane stress and n = 3 for plane strain; K is the stress intensity factor and is defined as: a √ K = σ π a.F W where: σ is the representative stress, a is the crack length and F(a/W) is a function of the geometry of the specimen LEFM can be applied if the size of the plastic zone is small compared with the dimensions of the specimen. The final scope of LEFM approach is the prediction of crack propagation conditions under the hypothesis that the material exhibits a linear elastic behaviour right up to the point where fracture occurs. Two relevant parameters for fracture phenomena studies were developed: – 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 notations from Fig. 4.1, the stress field in the immediate vicinity of a crack tip, for an isotropic solid, can be written such as:       θ 3θ KI θ 1 − sin sin cos σx = √ 2 2 2 2π r       θ 3θ KII θ 2 + cos cos −√ sin 2 2 2 2π r       θ 3θ KI θ 1 + sin sin σy = √ cos 2 2 2 2π r       θ 3θ KII θ cos cos −√ sin 2 2 2 2π r       θ 3θ KI θ sin cos τxy = √ cos 2 2 2 2π r       θ 3θ KII θ 1 − sin sin +√ cos 2 2 2 2π r

(4.2)

σz = 0 for plane stress

(4.4)

(4.1)

(4.3)

4

Initiation and Growth of Delamination

55

and  σz = ν σx + σy for plane strain conditions   θ KIII τyz = √ cos 2 2π r   KIII θ τzx = √ sin 2 2π r

(4.5) (4.6) (4.7)

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:

    θ θ r cos k − 1 + 2 sin2 2π 2 2

    θ r KII 2 θ sin k + 1 + 2 cos + 2μ 2π 2 2

    θ r KI 2 θ sin k − 1 + 2 cos v= 2μ 2π 2 2

    θ KII θ r − cos k − 1 − 2 sin2 2μ 2π 2 2    θ r 2KIII sin w= μ 2π 2

KI u= 2μ

(4.8)

(4.9)

(4.10)

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: G = GI + GII + GIII =

KI2 K2 K2 + II + (1 + ν) III  E E E

(4.11)

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 viscoelastic 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 process 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: δ=

4 K2 π EσY

(4.12)

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 microcracks. 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) crack tip opening displacement (b) crack growth resistance curve or R-curve, the energy required for the propagation of a crack of unit area [J/m2 ] (Yoshihara 2001, 2003, 2004, 2005, 2006a, b; Morel et al. 2002, 2003, 2005; Coureau et al. 2006). Figure 4.3 explains theoretical behaviour of materials exhibiting bridging zone. Bridging zone can extend from the initial crack tip x0 to the notch root at xroot

a

b

c

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, increasing 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 softening 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 softening 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 displacement 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 propagates 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 fracture 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.

4.2.1 Linear Elastic Fracture Mechanics The development of the concepts related to the application of LEFM required several 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 substantial 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 LT RL TR RT TL KIc· > KIc· > KIc· > KIc· = KIc > KIc·

4

Initiation and Growth of Delamination

b1

Fig. 4.5 (continued)

61

b2

62

V. Bucur

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 Species

Crack system Fracture toughness Plane

kNm–3/2

TL RL TR RT LT LR TL TL TL TL TL TL TL TL TL TL and RL

309 410 355 355 2417 2692 190 185 494 492 564 407 478 505 681 112

TL TL TL TL RL TL RL TL

1890 2187 1626 2143 2230 2190 1665 159

References

Mode I Douglas fir

Western white pine Western red cedar Hoop pine Hard maple Paper birch Red oak Lauan Messymate stringbark Maiden’s gum Balsa

Schniewind and Centeno (1971)

Johnson (1973) Johnson (1973) Walsh (1971) Johnson (1973) Johnson (1973) Johnson (1973) Johnson (1973) Walsh (1971) Walsh (1971) Wu (1963) cited in Wood handbook (1999)

Mode II White spruce Lodgepole pine Amabilis fir Douglas fir Red spruce Balsa

Barrett (1981) Barrett (1981) Barrett (1981) Barrett (1981) Wood handbook (1999) Wood handbook (1999) Wood handbook (1999) 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). TL RL TL It was observed that always GRL C > GC and the ratio GC /GC is between 1.26 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)

Spruce Alder Oak Ash

Young’s moduli (GPa)

Kinitial (N/m)

√ KIc (MPa m )

Gf (J/m2 )

EL

ER

ET

RL

TL

RL

TL

RL

TL

10 11.7 13 15.8

0.8 1.1 1.6 1.5

0.45 0.6 0.9 0.8

1.44 2.33 2.58 3.57

1.01 0.95 1.31 1.60

0.49 0.67 0.83 1.16

0.31 0.33 0.41 0.65

337 244 348 551

213 155 271 345

It is to note that for real structures the Mixed Mode is dominant and consequently 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 process 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 opening 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)

of fracture energy Gf in tension perpendicular to grain in different orientations, for spruce, fir and Sitka spruce. The ratio between the values in RL system and TL system is between 1.33 and 1.37. The fracture parameters are strongly influences by the density (Forest Products Laboratory 1999; Reiter et al. 2002). Figure 4.9a shows the variation of fracture toughness for different species with density. Fracture toughness in Mode I increases with density. The density range was between 100 kg/m3 and 800 kg/m3 and the fracture toughness was measured parallel and perpendicular to the grain. More refined studies were reported by Donaldson (1997) related to the variation of microdensity in fractured zones. Figure 4.9b, c shows the microdensity variation in a transwall fracture zone in the middle lamella region. There is a linear decrease of microdensity in the vicinity of the fractured surface. Studies related to Mode III in wood were possible with the development of a specific experimental devices as for example those proposed by Tschegg (1986), Ehart et al. (1998, 1999), Tschegg et al. (2001). It was point out by Tschegg et al. (2001) that for larch and beech, crack initiation energy in Mode III is over twice as high as Mode I in RL and TL fracture systems, because of a much larger fracture process zone in Mode III than in Mode I. Moreover the Mode III crack has

66

V. Bucur

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 presence 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 thoroughly reviewed by Wang et al. (2003). The maximum fracture toughness was

4

Initiation and Growth of Delamination

67

Fig. 4.8 Mode I, Load-displacement curves obtained by the wedge splitting test in RL and TL system for ash, oak, alder and spruce (Reiterer et al. 2002, Figure 5)

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 bending 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] Orientation Fracture energy Gc (J /m2 ) Species

Plan

Mean

Minimum

Maximum

Coeff. Variation (%)

Spruce

RL TL RL/TL RL TL RL/TL RL TL RL/TL

220 160 1.37 210 157 1.33 220 164 1.34

159 100 1.59 126 97 1.29 157 136 1.15

345 247 1.39 367 236 1.55 248 196 1.26

19 29 − 26 37 − 16 16 −

Ratio Fir Ratio Sitka spruce Ratio

68

Fig. 4.9 (continued)

V. Bucur

4

Initiation and Growth of Delamination

69

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 bridging 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. – region between 5% and 12%, in which ◦ Gf decrease for beech and oak and increased for pine and spruce ◦ KIc decreased for beech, oak spruce and pine 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

Mode I

Dry Wet Dry/wet Dry Wet Dry/wet

0.486 0.214 ∗∗∗ 2.826 2.328 ∗∗∗

0.351 0.236 ∗∗∗ 1.088 0.458 ∗∗∗

0.282 0.270 ns 2.664 1.905 ∗∗∗

0.195 0.235 ∗∗∗ 1.228 0.443 ∗∗∗

2.69 2.21 ∗∗∗ – – –

2.39 1.88 ∗∗∗ – – –

Test Student( t) Mode II Test Student (t)

NB: Test Student ∗∗∗ confidence level 90%

70

V. Bucur

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

– region between 12% and 18 %, in which ◦ Gf increased for all species ◦ KIc increased for beech and oak but decreased for spruce and pine

4

Initiation and Growth of Delamination

71

– region between 18% and 30%, in which ◦ Gf is relatively constant for oak and spruce and decreased for beech ◦ KIc decreased for spruce and pine 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 StanzlTschegg 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) Linear elastic fracture mechanics 1

2

3

4

5 6

7

Wood is a homogeneous linear elastic medium (isotropic or orthotropic)

Comments

Wood is heterogeneous, cylindrically orthotropic with discontinuities on macro and micro structural levels. Brittle fracture occurs in an elastic range The pre-existing crack always propagates The crack does not grow along the original along the original crack direction orientation. The initial crack extension is always parallel to the grain, even when starter crack lies across the grain. At microscopic level fractured surfaces are irregular and tortuous Crack tip displacements can be separated into Only displacements can be separated into three different modes (Mode I – in plane three independent modes tension, Mode II – in plane shear, Mode III – out of plane shear) The stress intensity factors KI, KII KIII fully The simplicity of K characterization with characterize the intensity of stress only one parameter for all complex fracture distribution in the vicinity of the single phenomena is no more acceptable crack tip The inelastic zone is confined to a small The inelastic zone is not small volume of crack tip Crack surface are traction free at all stages of The crack surfaces are not traction free and loading, and the crack tip is anatomically not anatomically sharp. See the sharp “ligamentary bridging “of fibres. Crack bridging is part of toughening mechanisms The crack propagates dynamically at some The stability and rate of crack velocity can be terminal velocity controlled through appropriate choice of the rate of loading and experimental configuration.

72

V. Bucur

4.2.2 Nonlinear Fracture Mechanics Nonlinear Fracture Mechanics (NLEFM) is more appropriate the LEFM to describe wood behaviour in several practical situations such as: fracture beyond initial cracking, creep rupture effects, size-effect on small clear specimens, adhesive joints, etc. To estimate the level of non linearity either fracture energy methods or stress-based concepts can be applied. Comparative studies on linear and non linear fracture mechanics on wood have been performed by Patton-Mallory and Cremer (1987), Boström (1992), StanzlTschegg et al. (1995), Tan et al. (1995), Reiter and Sinn (2002), and Vasic and Smith (2002), using non linear energy based fracture theory in order to quantify the relevance of deviations from the theoretical brittle response. It was admitted that wood has a softening behaviour (Vasic and Smith 1996a, b, 1998, 1999a, b). The apparent non linearity in the fracture response beyond the peak load is attributed to the gradual development of damage and microcraking in a fracture process zone around the crack tip. Stanzl-Tschegg et al. (1995) noted also that the fracture mechanism in wood is not purely brittle. Vasic and Smith (2002) explained the non linear behaviour of spruce in Mode I, by fibres bridging behind the crack tip, in the presence of stress concentrations. The bridging crack model propose by Vasic and Smith (2002) assumed that the sharp crack tip coexist with a bridging zone behind the tip crack (Fig. 4.11a, b). The variation of energy release rate and fracture toughness versus the crack length is shown in Figs. 4.12 and 4.13 on which “the influence of bridging stresses clearly increases with any increase in the crack length, if the maximum bridging stress is kept constant“. The fracture parameters reach a maximum at 4 mm which corresponds to the tracheids length of spruce. Vasic and Smith (2002) demonstrated that bridging of the fibres behind the crack tip is a major factor in toughening mechanism in wood. They confirmed the previous statements of Boström (1992) and Tan et al. (1995) that wood in fracture has a non linear behaviour similar to concrete. “The nonlinearity beyond the peak load was attributed to gradual development of damage in a fracture process zone around the crack tip”.

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 Model description

1

2

3

4

5

6

Comments

– fictitious model – existing of a cohesive zone – Cracks closing stresses act – FEA as a function of crack – the model is simple and separation distance incorporates the – Stress and crack opening are nonlinearities in to the related by Gf closing stress function – the fracture energy is predicted from the area under the closing stress versus opening function – post peak softening – pine specimens parameter and fracture – FEA energy – separate the microcracking – wedge-splitting tests – FEA from crack bridging – softening curve with contribution to fracture bilinear representation of energy stress-crack opening – the bridging components in relationship RL > TL, due to rays normal to the crack plane – bridging model – Canadian spruce specimens, – bridging stress occur on the Mode I crack faces close to the – in situ ESEM crack tip – the strength of the bridging – bridging zone length = stress determine whether 4 mm = tracheids length, fracture is brittle, which is the intrinsic quasi-ductile or ductile material length scale to a – combination on FEA and continuum fracture model ESEM observations – pine and spruce specimens, – crack propagation occurs Mode I when peak tensile stress is – bilinear and trilinear reached constitutive relationship for – crack interface element crack interface element represents the closing – in trilinear model the stresses softening is broken down – whole fracture process zone into microcracking and is lumped into a crack line bridging phenomena and is characterized by the – load displacement curves stress-crack opening law and R curves which exhibits softening – FEA – cohesive crack simulations – crack face friction is – Mode II and Mixed Mode negligible – measured values for GIc and GIIc – bilinear constitutive relationship for crack interface element -

References Hillerborg et al. (1976) Hillerborg (1991) Homberg et al. (1999)

Boström (1988, 1992)

Stanzl-Tschegg et al. (1995)

Vasic and Smith (2002)

Dourado et al. (2004, 2008)

Silva et al. (2006)

4

Initiation and Growth of Delamination

75

Table 4.6 (continued) Model description 7

8

Comments

References

Coureau et al. (2006a, b) – spruce specimens, crack and RL, Mode I – stress opening relationship Lespine (2007) is initial linear elastic regime – brittle law – the peak load of a lod –crack opening displacement curve is strongly affected by the slope of softening behaviour – the roughness is related to the microstructural features and toughness – the effect of crack closing stress function, the load crack opening displacement and R curve are related quantitatively – FEA Schmidt and Kallske (2007) – 3D anisotropic constitutive – nonlinear, continuous softening stress- crack law is implemented, which opening function – realistic could loaded and unloaded incorporate damage and load behaviour history – FEA

– cohesive crack model – elastic layer model – tensile strength σt and overall fracture energy Gf , are the most important properties of the cohesive zone – the fracture energy related to the constitutive law and must correspond to the plateau value of the R-curve – crack resistance of R-curve behaviour is related to the roughness of fracture surface

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 converted 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 acoustic 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. 4.14 Material point numerical modelling of wood structure (Nairn 2007a, Figure 2)

4

Initiation and Growth of Delamination

77

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 properties of the elements to be statistical quantities. Lattice models have focused on longitudinal properties of wood where the rods are wood fibres and springs represent 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 Model 1

2

3

4

5

Characteristics

Comments

Transverse isotropic – described by 5 elastic constants, – the simplest model material – Young’s moduli are as – valid mostly for tropical trees EL >> (10 . . . 20)ER ; . ER = ET . – miss low transverse shear – Finite Element Analysis modulus GRT – data from Nairn 2007 b – miss ring curvature and structure, EW, LW properties – miss ring curvature and Rectilinear – described by 9 elastic constants structure EW, LW properties orthotropic – simplify the analysis by aligning material coordinates of the anisotropy with – mesh generation relatively simple the rectilinear natural axes L,R,T – allow a low transverse shear modulus GRT – can describe the off-axis loading – Finite Element Analysis – data from Nairn 2007 b Homogenized – accounts for growth rings curvature – complicate mesh generation, compared to rectilinear cylindrical within a specimen, orthotropic material orthotropy – simplifies the analysis by using homogenized properties in transversal plane. – one can use large elements where stress gradients are small – small elements are required throughout the specimen in order to resolve orientation of material axes along curved growth rings. – can approximate effective mechanical properties, account for differences between pith and periphery boards, and account for size effects. – Finite Element Analysis – data from Nairn 2007 b – the closest approximation of – accounts for both growth ring Heterogeneous real wood structure curvature within a specimen and cylindrical – needs new methodology for variation in material properties orthotropy reliable values of such as EW and LW – fine mesh is required to resolve the EW, LW mechanical properties (acoustic microscopy Bucur structure of wood et al. 1995; Clair et al. 2000, – is recommended for modelling Bucur 2003; Bucur 2005 or failure processes induced by other ultrasonic techniques localized stresses – EW, LW Bucur et al. 1994) – Finite Element Analysis – data from Nairn 2007 b Monoclinc symmetry – described by 21 constants (Bucur – never used until now – 2009) and Rasolofosaon 1998)

4

Initiation and Growth of Delamination

79

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: – the facility to discretize realistic wood structures via a process of digitization of an image into pixels. – automatically handles contact and thus can be extended to high strain without numerical difficulty – can handle the specificity of the internal configuration of wood structure, the large deformations and the large calculations in relatively short time. – 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 acoustic microscopy

4.3 Micro-structural Aspects in Wood In this section detailed micro-structural aspects of delamination in solid wood will be discussed. We have seen that the space – time multi-scale nature of the delamination process in wood can be related to the prediction of crack nucleation, growth and arrest. The initiation of failure can be marked by the first acoustic signal (Lee et al. 1995; Dill – Langer and Aicher 2000; Dill – Langer et al. 2002; Reiter et al. 2000; Bucur 2005), or by the non-linearity point on the load/displacement curve (Tschegg 2001; Frühmann et al. 2003). Crack initiation, crack growth and crack arrest emerge as natural outcomes of the imposed load. In all these processes wood microstructure plays a very important role. Studies on wood fracture in relation to its structure using optical microscopy have been published since several decades (Mark 1967; Debaise et al. 1966, 1972; Dinwoodie 1966, 1968, 1974; Jeronimidis 1976, 1980). Gordon and Jeronimidis (1980) suggested that the cells in the vicinity of the fracture zone can absorb a great quantity of energy before breaking. The helical structure of the cellulose microfibrils in the S2 and the helically wound pattern of the microfibrils induces a specific form of buckling failure in tension, which causes a high energy absorption during fracture (Jeronimidis 1980a, b). In his previous work, Jeronimidis (1976) emphasises the essential part played by the S2 layer in the fracture process upon longitudinal tension. Keith and Côté (1968) described the layer boundary S1–S2 of the secondary wall as the place where intra-wall failure arises as a result of shear strains. Experimental studies on hollow cylindrical tubes scale models with helical fibres at different winding angles showed that the optimal trade-off between stiffness and toughness can be observed at a microfibril angle of about 15◦ (Gordon and Jeronimidis 1980). Kucera and Bariska (1982) using tube multilayered model specimens for direct observation of the formation of failure in longitudinal compression noted that “cracks do not occur until the reduction in specimen length reaches a stage where the wall folds fill the whole lumen of the tube. They always arise parallel to the axis across one or more folds in the longitudinal direction, and

80

V. Bucur

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 delamination 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 fracture 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) fracture 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 specimens 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 deformation 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 others 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

Fig. 4.18 Tangential fractures of two adjacent cells in Pinus radiata loaded in tension (scanning electron micrograph). A delamination is observed where intra-wall fracturing undergoes a transition between cell walls layers (Donaldson 1997, Figure 1)

82

V. Bucur

intra-wall fracturing undergoes a transition between cell walls layers. Donaldson (1997) noted that the transwall fractures are on tangential surfaces and are more common outside the juvenile wood rings, This fractures are due to changes in cell dimensions and microfibril angle with cambial age (Donaldson 1996). “Transwall fractures are favoured by large cell size and by low microfibril angles, and occur only on tangential wood fractures in Pinus radiata. Radial wood fractures are exclusively intrawall due to the favourable energy conditions provided by the alignment of tracheids in radial files”. Most of the transwall fractures observed had an irregular aspect. In contrast, the intrawall paths tend to follow the lamellate structure of the cell wall matrix, producing smooth surfaces. These aspects are described in Fig. 4.19 and in Fig. 4.20. More details related to fracture and wood anatomy are given in the outstanding contribution of Donaldson et al. (1996) – Rotorua Laboratory, New Zealand. The development of the equipment for in-situ studies with ESEM (Bodner et al. 1996; Vasic et al. 2002; Smith and Vasic 2003; Turkulin et al. 2005; Vasic and Stanzl-Tschegg 2008) has been a big step towards the understanding of crack initiation and propagation in wood during loading. The equipment allowed loading operation very precisely. Small inconvenient can be introduced by the action of electronic beam which weakened the cell wall in S3 (Hoffmeyer and Hanna 1989). Electronic beam damage induced fractures have very characteristic patterns, different from the other mechanical fractures. Bodner et al. (1996), for tension tests on Norway spruce observed that” in samples with parallel growth rings cracks propagated with jumps”, (probably the system LR was tested). In specimens with perpendicular growth rings, “the initial cracks developed into the final fracture eruptively and, without intermission”. Serrated (saw tooth) fracture pattern occurs in S2 and S3 , the microfibrils are pulled out. Thuvander and Berglund (2000) described the micromechanics of fracture in radial growth cracks in green pine (Pinus sylvestris) specimens with in-situ optical microscope. Figure 4.21 shows the morphology of the radial cracks. At the cells

Fig. 4.19 Higher magnification (transmission electron micrograph) of the cross section of the transwall fracture. S2 follows the line of least resistance while S1 and S3 layers protrude from the fracture zone. The intrawall fracture is seen along the ML surface (Donaldson 1997, Figure 2)

4

Initiation and Growth of Delamination

83

Fig. 4.20 Transwall fracture in cross section with transmission electron technique (Donaldson 1997, Figure 3). (a) the smoothness of the intrawall fracture compared to the roughness of the transwall fracture. Transwall fractures (short arrows) are separated by intervening intrawall fractures (long arrows) between S1 and S2 and S1 and ML. (b) the irregular sawtooth fracture surface in a transwall fracture

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 latewood zone, its growth rate decreases and could be arrested in earlywood. Latewood failure occurred mostly by cell splitting because of weak middle lamella. DillLanger 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 tangential 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)

4.4 Micro-structural Aspects in Wood-Based Composites 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 woodbased 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

Fig. 4.22 (continued)

V. Bucur

4

Initiation and Growth of Delamination

87

Fig. 4.23 MDF specimens size and orientation (Matsumoto and Nairn 2008, Figure 1); (a) compact 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

modelling techniques to predict the influence of species, size, engineering properties of the constituents, etc., on wood-based composites properties. Under normal conditions, strengths of wood-based composites are directionally dependent on the structural features and, on the other hand, are time dependent and sensitive to moisture content. Their heterogeneity is also their main source of weakness, irrespective of the nature of the constituents. As en example, in glulam, the interface between fibers and adhesive layer is critical for damage onset and development. The damage mechanisms themselves are numerous and closely connected. In many situations the most critical damage mechanism for composites design is the delamination between adjacent layers. Whatever the cause the delaminations, they can be very dangerous and can easily lead to a premature collapse of the structures. The physical phenomena behind delamination onset in wood-based composites can have the following causes: – the residual stress induced during manufacturing – the environmental conditions, such as moisture content and temperature gradients – machining and drilling producing peel-off of the uppermost plies and heat generation; the angle of penetration, the drill geometry, the fibre orientation, lay-up sequence are factors of influence – the geometrical configuration – free edge interlaminar stress, skin debonding, joints, tapered structures – the inclusions such as bolt, holes and notches The physical phenomena behind delaminations growth are induced by the application 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 designing 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 measured, 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 interference. “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. 4.24 Axial strain as a function of the position along the crack line with DIC and measurement of increment crack growth a (Matsumoto and Nairn 2008, Figure 2) Curve 1 – prior to crack growth, Curves 2 . . . 7 profiles after subsequent increments in crack growth, a = the crack growth between two point in the test, the shift between curves

4

Initiation and Growth of Delamination

89

Fig. 4.25 R-curve revised method (Matsumoto and Nairn 2008, Figure 4) (a) – integral transformation 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

the bridging material is crushed, causing the unloading compliance to be lower, resulting in a residual displacement”. To take into consideration the bridging zone, material point method and CRAMP – cracks in the material point method – (Nairn 2003, 2006; Guo and Nairn 2004, 2006) have been developed as summarized in Fig. 4.25. R-curve with fiber bridging process zone was calculated from the slope of energy area. The slop of the cumulative energy released per unit thickness is deduced by integrating force displacement data up to some displacement and subtracting the area under an assumed elastic return to the origin. Figure 4.26 shows the

Fig. 4.26 R-curve (J/m2 ) versus a (mm), the crack growth between two point in the test for LT fracture with discrete and revised analysis, for MDF specimen 12 mm and 609 kg/m3 (Matsumoto and Nairn 2008, Figure 5)

90

V. Bucur

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) Crack

Panel thickness 12 mm

Panel density

Panel thickness 19 mm t

Type

Gc

Slope

σc

Gc

Slope

σc

(kg/m3 ) 609

– //

737

//

J/m2 2062 54 4153 75.3

J/m3 21700 222 59600 814

MPa 0.78 0.038 2.55 0.14

J/m2 2233 48.2 4452 48.4

J/m3 10500 296 18400 303

MPa 0.43 0.056 0.66 0.10

good agreement between discrete and revised analyses of the R-curve for LT fracture 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: – – – –

the introduction of an explicit crack, the crack tip energy release rate was calculated including bridging effect, the total fracture energy released was calculated at the time of crack propagation, the COD – crack opening displacement – along the crack surface was calculated, the bridging fibers failed whenever COD > δc , the calculation can be performed until crack length reached the end of the specimen, – fracture mechanics methods were used to model crack tip processes and traction law methods were used to model the bridging zone, inserted only as the crack propagated. Another experimental approach for crack propagation detection was proposed by Watanabe and Landis (2007) with 3D micro tomography. Cutting and drilling of MDF are often required for boards used in the manufacture of furniture, cabinets and flooring. Delamination is one of the major defects observed with cutting and drilling. Digital image analysis was used by Davim et al. (2007) to study the delamination in an MDF plate with coating layers induced by drilling. An empirical factor has been proposed to characterize the delamination at the entrance and exit surfaces of the hole produced by drilling. The digital image analysis shows a typically brittle fracture on the coating layer. The damage area at the drill entrance is slightly larger than at the drill exit. Higher cutting speed should be used to induce minimal delamination and to obtain greater material removal during drilling.

4

Initiation and Growth of Delamination

91

4.5 Fracture Mechanics Parameters for Ecological Relevance Fracture mechanics criteria have been developed by Mattheck (1996, 2004), Mattheck and Bethge (1991), Matthek et al. (1995), Mattheck and Kubler (1997) to describe failure modes of trees. Strong winds (hurricanes, cyclones, etc) cause extensive damage to forests, world-vide every year. Considerable research has been carried out to understand the physical processes involved in order to improve the sylvicultural practice. Spectral analysis was used to identify the dynamic behaviour of trees (Guitard and Castera 1995; Gardiner 1995). “Wind affects the growth rate of the tree and determines the occurrence of windthrow in the later years of the development of a forest. Consequently the forecast of the financial viability of any forest project is dependent upon an accurate assessment of the wind speed” and of the modelling of trees behaviour (Gardiner 1995). Fracture mechanics parameters have been used to put in evidence potential influence of air pollution on wood quality in Europe and Canada (Grosser et al. 1985; Bondietti et al. 1990; Niemz et al. 1990; Koch et al. 1996; Stanzl-Tschegg, Filion et al. 1999; Beismann et al. 2000; Beismann et al. 2002). Unfortunately no unified methodology was used and the results have been ambiguous. Only tendencies have been observed. In spruce, exposed to SO2 emission, the fracture morphology (Koch et al. (1996) has shown short fibres, and crack initiation with bent tracheids in the vicinity of rays. Stanzl-Tschegg, Filion et al. (1999) described the SO2 pollution in spruce with the notch-tensile strength via ring width and density. A pronounced influence was observed on trees grown between 1970 and 1985, and a subsequent recovery in trees that had survived this period. Beismann et al. (2002) noted the response of stems of 6 to 7 year old spruce and beech trees studied after 4 years growth in elevated atmospheric CO2 in combination with a nitrogen treatment and on two different soil types. The fracture toughness, modulus of elasticity (EL ) and wood density were strongly influenced. Smith and Chui (1994) observed differences in Mode I fracture energy of premature plantation grown red pine for crack growth in the L direction. Differences in bending properties of plantation grown white spruce have been reported by Zho and Smith (1991). Differences in fracture energy of Pinus radiata wood from different plantations were reported by King and Vincent (1998). Donaldson (1995) put in evidence cell wall fracture properties in relation to lignin distribution and cell dimensions among three genetic groups of radiate pine. As a conclusion it can be suggested that the environmental influences on wood quality require the development of specific techniques for wood microstructural studies.

4.6 Summary Reliable prediction of delamination growth is still proving to be problematic, leading to the use of large safety factors and reticence in using wood-based composites

92

V. Bucur

in safety critical applications. This has led to composite structures being perceived as expensive to fabricate and needing frequent inspection and repair. The recognise 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 structural 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 elastic 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 fracture 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 timbers, 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 predictive (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 evaluated 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 emission. 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 R – Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1998) Fracture characterization of PARALAM 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 simply 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

Morel S, Mourot G, Schmittbubl J (2003) Influence of the specimen geometry on R – curve behaviour and roughening of fracture surfaces. Int J Fract 121(1/2):23–42 Morel S, Bouchard E, Schmittbuhl J, Valentin G (2002) R-curve behaviour and roughness development of fracture surfaces. Int J Fract 114:307–325 Morel S, Dourado N, Valentin G, Morais J (2005) Wood a quasi-brittle material R-curve behavior and peak load evaluation. Int J Fracture 131(4):385–400 Morris VL, Adam JM, Hunt DG, Bonfield PW (1996) Fracture mechanics in wood and woodbased panel products at different testing speeds and relative humidity. COST Action – wood mechanics. Workshop on mechanical properties of panel products, 22 – 23 March Watford, pp 336–351 Morris VL, Hunt DG, Adam JM (1999) The effect of experimental parameters on fracture energy of wood-based panels. J Inst Wood Sci 15(1): 85 Moses DM, Prion HGL (2004) Stress and failure analysis of wood composites: a new model. Compos: Part B, 35:251–261 Nairn JA (2009) Analytical and numerical modelling of R curves for cracks with bridging zones Int J Fract 155(2):167–181 Nairn JA, Matsumoto N (2009) Fracture modelling of crack propagation in wood and wood composites including crack tip process and fiber bridging mechanics. Proceedings of the 12th International Conference on Fracture, Ottawa, July 12–17 Nairn JA (2007a) Material point simulation of transverse fracture in wood with realistic morphologies. Holzforschung 61:375–381 Nairn JA (2007b) A numerical study of the transverse modulus of wood as a function of grain orientation and properties. Holzforschung 61:406–413 Nairn JA (2007c) Numerical modelling of deformation and fracture of wood including heterogeneity and anisotropy. Cost Action 35, Lausanne, May 21–23. Nairn JA (2006) Numerical simulation of the transverse compression and densification in wood. Wood Fiber Sci 38:576–591 Nairn JA (2003) Material Point Method calculations with explicit cracks. Comput Model Eng Sci 4:649–664 Niemz P, Diener M (1999) Vergleichende Untersuchungen zur Ermittlung der Bruchzähigkeit an Holzwerkstoffen. Holz Roh Werkst 57:222–224 Niemz P, Diener M, Pöhler E (1997) Untersuchungen zur Ermittlung der Bruchzähigkeit an MDF – Platten. Holz Roh Werkst 55:327–330 Niemz P, Bemmann A, Wagebführ R (1990) Untersuchungen zu den Eigenschaften von immissionsgeschädigtem Fichtenholz. Holztechnologie 30(2):70–71 O’Brien TK (1998) Interlaminar fracture toughness: the long and winding road to standardisation. Compos B 29(1):57–62 Patton – Mallory M, Cramer SM (1987) Fracture mechanics: a tool for predicting wood component strength. For Prod J 37(7/8):39–47 Pearson RG (1974) Application of Fracture mechanics to the study of tensile strength of structural lumber. Holzforschung 29(1):11–19 Petterson RW, Bodig J (1983) Prediction of fracture toughness of conifers. Wood Fiber Sci 15: 302–316 Porter AW (1964) On the mechanics of fracture in wood. For Prod J 14(8):325–331 Raju IS, O’Brien TK (2008) fracture mechanics concepts, stress fields, strain energy release rates, delamination initiation and growth criteria. In Sridharan S (ed) Delamination behaviour of composites, Woodhead Publishing, Cambridge, England Renaud M, Rueff M, Rocaboy AC (1996) Mechanical behaviour of saturated wood under compression. Part 2: Behaviour of wood at low rates of strain some effects of compression on wood structure. Wood Sci Technol 30:237–243 Reiterer A, Stanzl- Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of softwood and hardwood. Wood SciTechnol 34:417–430

98

V. Bucur

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 perpendicular 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 threedimensional 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 fracture 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: micromechanical 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 radially 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 5.2

Introduction . . . . . . . . . . . . . . . . . . . Repeating Cell Unit Model with Cyclical Constrains 5.2.1 Cell Wall Layer Properties . . . . . . . . . 5.2.2 Circular Based Cell Model . . . . . . . . . 5.2.3 Squared Based Cell Model . . . . . . . . . 5.3 Summary . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

101 102 102 104 106 118 119

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 particularly 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 modelling 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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_5, 

101

102

P. Blakemore

The intention of this chapter is to outline the development of a numerical model to assess the importance of the different secondary cell wall layers for both collapse and recovery, to demonstrate if the proposed collapse recovery mechanism is plausible. The model presented was developed using a model that Innes (1995) originally created. As the name implies, collapse is essentially a structural or mechanical phenomenon. Currently, the most appropriate tool for numerically modelling this type of problem is to use Finite Element Analysis (FEA) or Modelling (FEM). A repeating cell unit model with cylindrical constraints is proposed, firstly with a circular based cell model and secondly with a square based cell model, and is discussed.

5.2 Repeating Cell Unit Model with Cyclical Constrains 5.2.1 Cell Wall Layer Properties 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 possible. 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

Fig. 5.1 Three-dimensional representation of microfibril orientation in the primary cell wall (P) and the secondary cell wall (S) of a typical fibre or tracheid (Wardrop and Bland, 1959)

5

A Theoretical Model of Collapse Recovery

103

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). 5.2.1.1 Viscoelastic Material Model (S1 , and S3 ) The FEM package used for this modelling was MSC.Marc, which is a good solver for non-linear problems. The viscoelastic material model it uses is based on a de Prony-series that relates E, K or G against time at a certain temperature. For example, the de Prony-Series for E is shown in Eq. (5.2) The basic effect of temperature on E∞ was based on Eq. (5.1) (Innes 1996). A shift function is then used to adjust the creep curves for other temperatures. One of the more common forms of the shift function for polymers is the Williams-Landel-Ferry (WLF) equation (Williams et al. 1955) (Eq. (5.3)). In this instance, the shift factor and estimates of C1 and C2 were fitted such that the creep curves produced in a simple uniaxial tension model matched the creep model developed by Oliver (1991). Full details of the Oliver (1991) model, and how the appropriate MSC. Marc parameters were fitted to provide a match, are outlined in Chapter 6 of Blakemore (2008). Egreen = exp(4.206 + 0.003265BD − 0.03029T) E(t) = E∞ +

N

 En exp

n=1

−t λn

(5.1)

 (5.2)

where En = Modulus constant for series n λn = Relaxation time constant for series n – h t = Time – h log [αT ] =

−C1 (T − Ts ) C2 + (T − Ts )

(5.3)

where αT = Shift Factor C1 = Constant (Specific to Ts ) C2 = Constant (Specific to Ts ) T = Measurement temperature – K Ts = Reference temperature – K αT =

λnT λnTs

(5.4)

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 broken 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 easily, 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 elasticplastic 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 nonlinear elastic stress-strain relationship can be found in Chapter 6 of Blakemore (2008).

5.2.2 Circular Based Cell Model For any cell scale model to be representative of macroscopic behaviour, it needs to be based on a geometrically representative unit with cyclically repeating boundary conditions. This ensures that many such models could be joined together in both directions and the behaviour would still be consistent at a macro scale. Based on the perfectly cylindrical nature of the single cell model initially used by Innes (1995, 1996), the simplest repeating unit, based on this, is a hexagon (Fig. 5.2). Given the circular simplicity of the secondary cell wall layers, it is geometrically implausible that the cell would collapse flat on its own under a uniform hydrostatic tension pressure alone. Hence, a lateral displacement was applied to the model to force the cell into a non-circular shape. This was done by displacing (3 μm) the left and right edges inwards towards each other. Node displacements, and not edge forces, were used as otherwise the edge forces would have to be specified so as to maintain the straight edges required for the model to be cyclically repeating. There is some basis for this compressive force in reality as drying stresses can, and do, contribute to the occurrence of collapse, and for many cells these stresses will occur primarily in one axis only. To maintain cyclical constraints in the y-axis direction, it was necessary that the horizontal lines above and below the central whole cell should remain as a straight edge and horizontal (to prevent free body rotation). The simplest way of doing this was to tie the Y degree of freedom (DOF) for all the nodes on each line to the Y DOF to that of the central node (coloured red in Fig. 5.4). The horizontal compression was applied with a load case whereby all the nodes on the left and right edges were incrementally and linearly displaced towards each other over a

5

A Theoretical Model of Collapse Recovery

105

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 converge 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 temperature 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) highlighting 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.

5.2.3 Squared Based Cell Model 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 displacement 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 unsupported 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

Fig. 5.8 Refinement of mesh in the high stress corners of the square based multiple cell model. (Blakemore 2008)

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) displacements 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 material 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)

the mechanical properties of secondary cell wall layers directly. An alternative to measurement though, as outlined in (Blakemore 2008, Chapter 1) is the significant progress that has been achieved in modelling the mechanical properties of timber based on knowledge of the chemical microstructure and the mechanical properties of extracted chemical constituents. For example, Harrington et al. (1998) calculated the following cell wall elastic constants for Pinus radiata at 12% moisture content (Table 5.2 ). Many of the values shown in this table are an order of magnitude greater than the equivalent values shown in Table 5.3. Although, it should be noted that in this table the longitudinal direction (l) is parallel to the microfibril direction, and not parallel to the longitudinal orientation of the cell as is the “z” direction as indicated in Table 5.3. This is because in the model that used these data (Astley et al. 1998) the mean microfibril angle and known random variation around the mean, used in the different cell wall layers, could be varied to analyse its effect on board scale properties. To some extent, if the values used here are low it may have to some degree be compensated for the inherent stability of the simple and regular geometry of the model being used. While an attempt at generating similar values as shown in Table 5.2 might produce much more realistic values for an ash-type eucalypt

114

P. Blakemore

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)

than those shown in Table 5.3 it is unlikely to lead to a significant improvement in the model at this stage. This is because this approach is still not able to provide essential information on the non-linear large deformation behaviour, the time dependent behaviour, or the temperature-dependent behaviour of the cell wall properties. All of which are critical for a significant improvement in the collapse and recovery behaviour of the model being attempted here. The other most obvious reason for explaining the differences in the two tables (Tables 5.2 and 5.3) is the stated moisture content for the two tables; respectively 12% moisture content and green. This highlights another problem with attempting to replicating (Table 5.2) for an ash-type eucalypt species, and that is the need to generate a table for green moisture 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 T (◦ C)

Ei (MPa)

λ1 (h)

E1 (MPa)

Eα (MPa)

21.5

358

2.45535

12.1385

345.583

5

A Theoretical Model of Collapse Recovery

115

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)

content. As Harrington et al. (1998; Harrington et al. 1998) acknowledged, “while it is expected that at the lower moisture contents – such as at 12% – the properties of extracted constituents will be much the same as those in situ, this may not hold true at higher moisture contents”. Hence, this approach may not be as useful for predicting green cell wall properties as it is for dried material properties. Related to this discussion about the effect of moisture content, the next major limitation of Table 5.2 Calculated elastic constants for the cell-wall layers at 12% moisture content in Pinus radiata (Harrington et al. 1998) Wall Layer

Et (GPa)

Er (GPa)

El (GPa)

υrt

υlt

υlr

Grl (GPa)

Glt (GPa)

Gtr (GPa)

S3 S2 S1 CML

8.43 9.85 8.54 5.07

7.98 9.16 8.02 5.12

50.36 63.96 53.10 18.43

0.39 0.39 0.38 0.38

0.33 0.33 0.33 0.31

0.32 0.33 0.32 0.31

2.65 3.02 2.66 1.78

3.00 3.38 3.02 2.11

2.68 2.96 2.66 1.88

NB: “l” refers to a direction parallel to the length of the microfibres, and not the longitudinal direction of the cell

116 Table 5.3 Material properties for different cell wall layers

P. Blakemore

Er (E11 ) Eθ (E22 ) Ez (E33 ) νrθ (ν12 ) νθ z (ν23 ) νzr (ν31 )

S1 and S3

S2

620 6,200 620 0.05 0.5 0.38

620 620 6200 0.38 0.05 0.5

the recovery model is that it was attempted with green material properties maintained 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 moisture 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 formulate 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 transfer 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 properties 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 problem. However, even if a combined heat transfer and moisture diffusion model could be easily implemented, the collapse and recovery model will not be progressed significantly 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).

−1 node

−0.00969 −0.47345 −11.85242 −0.77253

5.2

−0.00981 −0.12648 −12.02427 −0.64927

r (μm)

U (μm) σr (MPa) σt (MPa) σz (MPa)

S3

−0.00967 −0.54445 −6.57728 −0.87351

5.37 −0.00966 −0.55488 −1.32599 −0.94043

+1 node

S2

−0.00982 −0.61094 −1.14803 −0.87948

−1 node

−0.00986 −0.61525 −5.11821 −0.78025

7.2

Table 5.4 Results from FEM solution. Three layer orthotropic case (Blakemore 2008)

−0.00984 −0.68686 −9.06328 −0.71417

+1 node

S1

−0.00987 −0.98577 −9.03168 −0.82618

7.5

5 A Theoretical Model of Collapse Recovery 117

118

P. Blakemore

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 recovery 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 inelastic 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 incomplete, 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 estimate these properties has been to use a range of homogenisation and finite element modelling techniques based on the generalised knowledge of the cell wall ultrastructure 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 simple, 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% moisture 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 possible 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 collapse 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 computationally 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 plantation 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 6.1

Anatomical Features of Wood Delamination . . . . . . . . . 6.1.1 Weathering and Decay . . . . . . . . . . . . . . . . 6.1.2 Internal and Intra-Ring Checking . . . . . . . . . . . 6.1.3 Resin Pockets . . . . . . . . . . . . . . . . . . . . 6.1.4 Shelling . . . . . . . . . . . . . . . . . . . . . . . 6.1.5 Reaction Wood . . . . . . . . . . . . . . . . . . . 6.1.6 Induced Delamination . . . . . . . . . . . . . . . . 6.2 Ultrastructural Features of Cell Wall Delamination . . . . . 6.2.1 Ultrastructure of Wood Cell Walls . . . . . . . . . . 6.2.2 Location of Cell Wall Delamination . . . . . . . . . . 6.2.3 Mechanism of Delamination . . . . . . . . . . . . . 6.2.4 Influence of Microfibril Angle . . . . . . . . . . . . 6.2.5 Influence of Delignification and Pulp Refining . . . . . 6.2.6 Influence of Species . . . . . . . . . . . . . . . . . 6.2.7 Influence of Moisture Content . . . . . . . . . . . . 6.2.8 Influence of Temperature . . . . . . . . . . . . . . . 6.3 Microscopic Methods for Evaluation of Delamination in Wood 6.3.1 Light Microscopy . . . . . . . . . . . . . . . . . . 6.3.2 Confocal Microscopy . . . . . . . . . . . . . . . . 6.3.3 Electron Microscopy . . . . . . . . . . . . . . . . . 6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

124 126 128 128 128 129 130 130 130 131 134 134 135 135 136 137 137 137 138 138 139 139

L. Donaldson (B) Bioproduct Development, Scion - Next Generation Biomaterials, 49 Sala St. Rotorua, Private Bag 3020, Rotorua 3046, New Zealand e-mail: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_6, 

123

124

L. Donaldson

6.1 Anatomical Features of Wood Delamination At the microscopic level, wood delamination can be defined as the separation or disintegration of fibres as a result of physical or chemical processes resulting in fracturing. Wood shows complex anisotropic behavior related to its microscopic structure, and this is also reflected in its fracture behaviour (Figs. 6.1 and 6.2). Delamination of wood can occur by intrawall fracture between adjacent tracheids

Fig. 6.1 Diagram of cell wall delamination for both radial and tangential planes, showing transwall 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. 6.2 Diagram illustrating the phenomenon of fibre bridging where single fibres or short rows of fibres span the developing fracture. This feature is more common in some species than others and is thought to prevent abrupt crack propagation resulting in a stepwise failure process

6

Delamination of Wood at the Microscopic Scale

125

or fibres, or in association with rays, or less commonly by transwall fracture, where the cell lumen is exposed (Koran 1967, 1968; Jeronimidis 1976; Kucera and Bariska 1982; Boatright and Garrett 1983; Côté and Hanna 1983; Zink et al. 1994; Donaldson 1997) (Figs. 6.2 and 6.3). Interwall fracture is a special case of intrawall fracture that occurs within the middle lamella (Côté and Hanna 1983; Zink et al. 1994). Delamination can occur naturally from weathering or decay, growth stress or drying stress (including brittleheart, surface, intra-ring and internal checking), physical damage to the living tree (resin pockets), and physiological stress (resin pockets, traumatic resin canals, shelling). These phenomena can be divided into radial delamination (intra-ring checking, internal checking, weathering, decay) and tangential delamination (resin pockets, traumatic resin canals, shelling). Brittleheart is a special case of transverse delamination. These cases of natural delamination often occur as a result of changes to the anatomical or chemical properties of the wood. Because wood tracheids and fibres are aligned in radial files, delamination may occur preferentially in the radial longitudinal plane between the rows of cells. Thus, wood splits easily in the radial longitudinal plane by crack propagation along the radial files of tracheids, and/or along the rays (Thuvander and Berglund 2000; Thuvander et al. 2000). There are differences between radial cracks that grow radially from a tangential surface, and those that grow longitudinally from a transverse surface. The former often shows deflections of the propagation when encountering a latewood boundary, with a stick-slip method of propagation due to the stress distribution produced by the alternating layers of soft earlywood and stiff latewood (Thuvander and Berglund 2000; Thuvander et al. 2000). Delamination by intrawall fracture is more difficult in the tangential plane because the tracheids or fibres are randomly arranged in this direction. Crack propagation must therefore follow an irregular course, which requires more energy, and may propagate more slowly. There are also fewer potential failure points, such as bordered pits, on the tangential

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

walls. Delamination in the transverse plane is difficult because there is usually no defined path for crack propagation. Thus transverse delamination occurs mainly by transwall fracture but with some intrawall stepping (Zimmermann and Sell 1997). This type of fracture may be associated with brittleheart in some hardwood species, where compression failures in the cell wall provide the pathway for crack propagation (Dadswell and Langlands 1934; Green 1962; Wilkins 1986a, b). Of course, delamination can also occur in wood-based products as a result of applied stresses during service or manufacture and may be associated with adhesion problems when failure occurs at glue lines, for example. Wood structure can also influence these types of delamination. For example, penetration of adhesive into the wood structure, or even into the cell wall, will result in stronger bonds. The presence of any delamination at the glue line as a result of inadequate surface preparation will weaken the adhesion (Jokerst and Stewart 1976; Murmanis et al. 1983; Singh et al. 2002). Various other types of delamination occur on wood surfaces as a result of sawing or planing and these micro-cracks can be penetrated by various surface coatings resulting in improved adhesion of the coating (Singh and Dawson 2004, 2006; Singh et al. 2007).

6.1.1 Weathering and Decay Delamination associated with weathering and/or decay occurs as a result of chemical modification of the wood cell walls, primarily the breakdown and removal of lignin by the action of UV radiation in sunlight, and the stresses caused by wetting/drying cycles (Borgin 1971a; Sell and Leukens 1971; Bamber and Summerville 1981; Voulgaridis and Banks 1981; Feist and Hon 1984; Singh et al. 1995; Evans et al. 2000; Turkulin et al. 2001; Singh and Dawson 2003; Kim et al. 2008), or by the activities of microorganisms (Sandberg 1999). In addition to wetting and drying cycles, under some conditions freezing and thawing cycles also contribute. Delamination as a result of weathering usually occurs by simple cell separation at the middle lamella (Voulgaridis and Banks 1981; Evans et al. 2000), as a result of breakdown of lignin (Bamber and Summerville 1981). The wood breaks up into individual fibres or fibre bundles (Singh and Dawson 2003), which may also show thinning of the cell wall as a result of erosion of the secondary wall from exposed lumen surfaces. Bundles of microfibrils are peeled away due to degradation of the cell wall resulting from chemical and physical processes (Kim et al. 2008). In some species such as pine (Pinus spp), ray cells are mostly unlignified, and hence may break down well before tracheids and fibres, resulting in radial delamination. When tissues other than rays are involved, the separation of cells may result in mixed radial and tangential delamination (Bamber and Summerville 1981). In cases where mainly polysaccharides are degraded, leaving an intact lignin residue (e.g., brown rot), delamination may occur also in the transverse direction because of the brittleness of dry lignin, allowing transwall fracture to easily take place (Irbe et al. 2006). The presence of cellulose microfibrils tends to resist transwall failure in the transverse plane in undegraded tracheids or fibres.

6

Delamination of Wood at the Microscopic Scale

127

Radial and tangential surfaces may show different responses to weathering (Sandberg 1999; Sandberg and Söderström 2006). Sandberg and Söderström (2006) found that tangential surfaces show a greater number and width of cracks in pine and spruce (Picea spp), compared to radial surfaces. Cracks on the tangential surface occur in both earlywood and latewood, but on radial surfaces, cracks are more common at the ring boundaries (Sandberg and Söderström 2006). Delamination of the middle lamella occurs frequently in the latewood on tangential surfaces (Sandberg and Söderström 2006). The tracheid or fibre cell walls exposed on weathered or decayed surfaces may show checks that follow the alignment of the microfibrils in the S2 layer. These checks may be related to softrot cavities, or may result from breakdown of the cell wall matrix and the influence of wetting and drying cycles (Borgin 1971a; Nilsson and Daniel 1990; Blanchette et al. 1994). Delamination of pit membranes as a result of weathering has been described by Turkulin and Sell (1997). Checks in the secondary wall induced by weathering, show bridging involving macrofibrils (bundles of microfibrils), a behavior that mirrors the bridging of whole tracheids in radial delamination at a larger scale (Stanzl-Tschegg 2006; Keunecke et al. 2007) (Fig. 6.2). It seems likely that such bridging behavior will also influence fracture toughness (resistance to fracture) at the nanostructural level, as it does at the anatomical level. Borgin et al. (1975) examined a range of ancient wood samples with ages ranging from 900 to 4400 years. Delaminations were found at the middle lamella/S1 interface with cracks and fissures also present in other parts of the cell wall. Blanchette et al. (1994) found intrawall cracks and fissures within the secondary walls of archaeological wood from ancient Egypt. This form of physical degradation was associated with exposure to limestone, gypsum, sodium chloride and moisture. Donaldson (1993) found similar delamination associated with sodium chloride deposits in the cell walls of Podocarpus tracheids from wood that had been buried on the sea floor. In ancient wood samples, cracks and intrawall delaminations are often observed, even in apparently sound wood. Borgin (1971a) found intrawall delamination within both the secondary wall and the middle lamella in the absence of microbial degradation. Similar delaminations were observed in wood from ancient tombs by Nilsson and Daniel (1990), and in at least one case, this was associated with softrot attack. Daniel et al. (2004) used cryo-FESEM to study delamination in white-rot decayed birch wood (Betula verrucosa Ehr.). The decay progressively removed matrix material between macrofibrils, resulting in concentric delamination of layers of macrofibrils within the secondary wall. Ando et al. (2006) compared fracturing properties of old Japanese red pine (Pinus densiflora Sieb. Et Zucc.) wood from a Buddhist temple and new wood within 3 years of felling. They found more uneven and complicated surfaces dominated by transwall fracture in the old wood (270 years old), suggesting that there was a prolonged formation of microcracks before fracture in these samples. Many of these transwall fractures were initiated from bordered pits, suggesting that these structures have a role in concentrating stress. Microcracks may have accumulated at the

128

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 delamination 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.

6.1.2 Internal and Intra-Ring Checking 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 perpendicular 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 follows 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).

6.1.3 Resin Pockets Resin pockets are localised regions of tangential delamination that result from mechanical or physiological damage, and are often associated with wind and or drought conditions (Temnerud et al. 1999). There are probably many different causes associated with the different types of resin pockets that occur in softwoods, but for typical type 1 resin pockets, there is a tangential delamination that opens to form a lens-shaped cavity. This cavity subsequently fills with resin and callus tissue, which proliferates from the damaged ray cells exposed at the surface (Somerville 1980). The exact mechanism is unknown, but in some cases the delamination is associated with bands of traumatic resin canals that form parenchymatic bands, that clearly represent a structural weakness (Glerum and Farrar 1966; Lee et al. 2007). Such resin canals are associated with unseasonal frost or insect attack (Lee et al. 2007). Delamination resulting in resin pocket formation may be associated with bending stress as a result of wind sway (Temnerud et al. 1999).

6.1.4 Shelling Tangential delamination during drying of logs or sometimes sawn timber, which may or may not be associated with growth ring boundaries, is known as shelling.

6

Delamination of Wood at the Microscopic Scale

129

Fig. 6.4 Confocal projection of drought-grown radiata pine showing intrawall delamination at the middle lamella (arrows) associated with reduced lignification. Field of view = 159×159 μm

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).

6.1.5 Reaction Wood 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

difference may be associated with reduced lignification at this transition in normal wood (Maurer and Fengel 1991). Tension wood is the gravitational reaction wood formed on the upper side of leaning stems and branches in hardwoods. Tension wood fibres are characterised by an unlignified inner wall layer known as the gelatinous or g-layer. Radial intrawall delamination has been described in eucalypt (Eucalyptus spp) tension wood, possibly associated with high levels of growth stress (Chafe 1977). These radial delaminations were characterised by local disruptions to the microfibril orientation, appearing remarkably similar to the radial checks found in severe compression wood. In many hardwood species, the G-layer of tension wood delaminates as an artefact during sectioning for microscopy (Clair et al. 2005). This separation of the G-layer from the secondary wall is considered to be caused by the sectioning procedure but is indicative of an inherent zone of weakness between the lignified secondary wall and the unlignified G-layer (Donaldson 2001; Clair et al. 2005). By serially sectioning embedded tissue, where further damage due to sectioning is prevented by the embedding resin, this delamination can be shown to occur up to 30 μm below the cut surface (Clair et al. 2005).

6.1.6 Induced Delamination Huang (1995) used ultrasonic treatment to induce delamination of wood cell walls in order to observe microfibril orientation. This was most effective on large diameter tracheids with high microfibril angles. Treatment with Congo red was found to enhance checking. This type of delamination appears to follow cellulose microfibrils 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).

6.2 Ultrastructural Features of Cell Wall Delamination 6.2.1 Ultrastructure of Wood Cell Walls The wood cell wall has a complex hierarchical organization, starting with cellulose 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 primary cell wall, which contains randomly oriented cellulose microfibrils and is highly lignified. The secondary cell wall, which characterises wood cells, contains 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

orientation of microfibrils, and an inner S3 layer with transversely oriented microfibrils and a relatively high degree of lignification. The secondary cell wall thus has a crossed-grain, laminate-type structure (Donaldson and Xu 2005; Donaldson 2007). Transverse fracture surfaces can themselves reveal ultrastructural details of the cell wall that are not readily observed by other methods (Sell and Zimmermann 1993; Zimmermann and Sell 1997). Spruce fractured in the transverse plane by applying a tensile load, was used to reveal macrofibrils, clusters of cellulose microfibrils in wood, which may sometimes form radial lamellae. It is still not clear if such radial lamellae represent a natural structure of wood cell walls, or an artifact induced by the fracturing process (Singh et al. 1998; Donaldson and Frankland 2004). These studies also demonstrate the complexity of transverse fracture showing both transwall and intrawall delamination producing a complex three-dimensional surface of tracheid bundles with macrofibril lamellae and cell wall fragments (Zimmermann and Sell 1997).

6.2.2 Location of Cell Wall Delamination Wood shows anisotropic fracture properties depending on the method of stress application (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 different 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 longitudinal 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 orientation 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

Fig. 6.5 Stereo scanning electron micrograph showing a tangential longitudinal fracture surface in radiata pine wood. This fracture shows both intrawall (IW) and transwall (TW) fracture. This stereo image should be viewed with red/green or red/blue glasses with red to the right eye. Scale bar = 100 μm

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 material, because of the highly oriented cellulose microfibrils, and this seems to favor intrawall fracture. Putoczki et al. (2007) have used TEM to study delamination associated 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 microfibrils 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 tracheid 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 primary wall and S1 layer (Schmitt et al. 1996; Donaldson 1997). The middle lamella

6

Delamination of Wood at the Microscopic Scale

133

Fig. 6.6 Transmission electron micrographs of radiata pine cell walls showing (a) intrawall, and (b) transwall fracture. The intrawall fracture is at the compound middle lamella/S1 interface while the transwall fracture shows stepping from the S2 to the S1 , and then to the middle lamella. Scale bar = 1 μm

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 longitudinal 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 specimens 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 process, 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

intact, forming an energy-absorbing sheath on the outside of the fibre (Fernando 2007; Fernando et al. 2007). Wood which has undergone radial compression in an attempt to improve the development of fibre flexibility by inducing cell wall delamination, shows small regions of intrawall fracture at the corners of cells in the outer secondary wall (Bergander and Salmén 1997). Singh et al. (2002) have used TEM to study delamination of planed surfaces in relation to PVA glue penetration, showing the complex mixture of intracellular and intercellular fracture which influences the penetration of adhesive into the wood surface. In more damaged surfaces, adhesive penetration was more restricted, especially as a result of closure of ray openings at the surface, and crushing of surface cells.

6.2.3 Mechanism of Delamination Using an environmental SEM with an in-situ deformation stage, Stanzl-Tschegg (2006) was able to observe the fracture process, while simultaneously acquiring load displacement curves. In samples under tensile load, there was a difference in crack propagation mechanism between earlywood and latewood. The weaker earlywood fibres fractured transversely, whereas the stronger latewood fibres separated by shear, resulting in a longitudinal fracture path. Using confocal microscopy, Dill-Langer et al. (2002) examined the mechanism of fracture in tension perpendicular to the fibre axis, comparing tangential, radial and oblique directions. Two mechanisms were identified, with transwall fracture of earlywood cells in the tangential plane, intercellular delamination between tracheids along radial files with crack progression in the radial plane, and a combination of both mechanisms in the oblique direction.

6.2.4 Influence of Microfibril Angle 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 longitudinal 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 irreversible 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

6.2.5 Influence of Delignification and Pulp Refining Chemically delignified pulp fibres show delamination as a result of refining. This delamination is typically within the S2 region, forming concentric lamellae, and may result in loss of most or all of the S1 layer, and separation of the S3 layer (Page and De Grâce 1967; McIntosh 1967; Kibblewhite 1972; Molin and Daniel 2004; Billosta et al. 2006). Such delamination is often more apparent in latewood tracheids than in thin-walled earlywood tracheids (Molin and Daniel 2004). Delamination varies with intrinsic viscosity, with pulps of high viscosity showing little or no delamination, the fibres instead fracturing into fragments (Molin and Daniel 2004).

6.2.6 Influence of Species Variation among species is the consequence of differences in wood properties such as density. Comparing spruce, oak (Quercus spp), ash (Fraxinus spp) and alder (Alnus spp), Stanzl-Tschegg (2006) found that species differences in critical stress intensity factor, and specific fracture energy (both measures of energy consumption during fracturing), could be attributed to differences in density (r = 0.98) confirming earlier studies (Schniewind et al. 1982; Petterson and Bodig 1983). Spruce showed a different relationship between fracture properties and density, indicating that other factors are also involved in determining the differences in energy consumption of the fracture process (Stanzl-Tschegg 2006). Spruce shows stable crack propagation, while oak, ash and elm show unstable crack propagation, with a stop/start behavior indicating greater brittleness in the hardwoods as compared to spruce. This was attributed to shorter fibres in hardwoods, with consequently less fibre bridging behavior, and resulting greater energy dissipation (Stanzl-Tschegg 2006) (Fig. 6.2). In thermo-mechanical pulping, differences in lignin distribution and cell wall structure between species may account for differences in fracture behavior. For example, radiata pine shows fibre separation at the middle lamella as well as at the middle lamella/S1 boundary, while in rubber wood (Hevea braziliensis) fibre separation occurs exclusively in the middle lamella (Singh and McDonald 2000; Singh et al. 2003). Comparing oak with ash, both of which are ring-porous, but have different ray anatomy, Stanzl-Tschegg (2006) determined the elastic behavior, critical stress intensity factor, and specific fracture energy. In both species, stiffness was greater in the radial longitudinal than the tangential longitudinal direction, and this can be attributed to the effect of rays. Ash has higher radial stiffness as a result of multiseriate rays, compared to oak, which has mostly uniseriate and sparse large multiseriate rays. Values are comparable in the tangential longitudinal direction for both species. The multiseriate rays of ash provide greater resistance to crack initiation, and to a lesser extent, crack propagation. The amount of ray tissue can influence the ability to withstand torsional loads in radially oriented samples, with the higher amount of ray tissue in beech (Fagus spp), apparently explaining its resistance to torsional load compared to spruce (Loidl et al.

136

L. Donaldson

2008). In the longitudinal orientation, beech and spruce had similar specific fracture energy under torsional load.

6.2.7 Influence of Moisture Content Moisture content (MC) influences delamination of cell walls by influencing ductility. Reiterer and Tschegg (2002) found a difference in fracture surface structure between samples at 7% MC and 55% MC. At 7% MC, longitudinal fracture surfaces are relatively smooth, but at 55% MC, the fracture surface shows more deformation, with many long cell wall fragments, which may be responsible for bridging behaviour during crack propagation (Fig. 6.2). Using environmental scanning electron microscopy (ESEM), Vasic and StanzlTschegg (2007) found that fracture path was determined by local density variations that were dependent on species. Fracture toughness was attributed to bridging effects caused by cell wall “ligaments” that were abundant at high moisture content, but were reduced at humidity values of 65% or 30%. Similar results were shown earlier using conventional SEM by Vasic et al. (2002), and Smith and Vasic (2003). In green wood, water was observed to be squeezed out of the cell lumens at the crack tip. Variations in load deformation curves demonstrated the role of microstructure in the fracture response of different species. Softwoods were found to be more ductile than hardwoods, with lower maximum load and specific fracture energy in pine and spruce, compared to oak (Vasic and Stanzl-Tschegg 2007). Stanzl-Tschegg (2006) compared fracture properties in spruce for a range of MC’s, on samples covering a range of densities. Critical stress intensity factor decreased from 7 to 12% MC with less decrease from 12 to 55% MC. In contrast, specific fracture energy increased with MC. The increase in specific fracture energy was found to be due to formation of microcracks and irreversible deformations at the fracture initiation site, which required more energy at higher moisture contents. Increased fibre bridging during crack propagation used more energy as the material became more ductile with increasing moisture content. Comparing green and oven dried/re-soaked samples fractured under tension, Kifetew et al. (1998) found rough fracture surfaces in green early- and latewood samples. Fracture in latewood was mainly intrawall failure. Samples that were dried and re-soaked showed transwall failure with smooth fracture surfaces, indicating a more brittle behavior. These authors concluded that drying of cell walls causes irreversible damage, which influences the fracture mechanism at the ultrastructural level. Delamination of cell walls on abrasive-planed and knife-planed surfaces in wood adhesive bonds has been studied because of the effect of surface structure on bond strength (Jokerst and Stewart 1976; Murmanis et al. 1983). Abrasive-planed samples showed radial cracks in the S2 layer, and delamination between S1 and S2 layers. The knife-planed samples had far fewer delaminations. When these samples were subjected to a soak/dry treatment, the bond strength of the abrasive-planed samples was much lower than the knife-planed samples. This difference was explained by

6

Delamination of Wood at the Microscopic Scale

137

the presence of cell wall delaminations, including rupturing and cracking within the S2 layer in the abrasive planed sample (Murmanis et al. 1983).

6.2.8 Influence of Temperature Koran (1967) examined the effect of temperature on the structure of radial longitudinal 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 producing 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 stability 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.

6.3 Microscopic Methods for Evaluation of Delamination in Wood 6.3.1 Light Microscopy While it is possible to examine samples for delamination directly by light microscopy, it is usually important to preserve the structural integrity very carefully, as sample preparation methods such as sectioning can easily induce further delamination, as well as distorting any delamination already present. In addition, if examination of a fracture surface in cross-section is required to determine the cell wall layers exposed at the surface, then quite thin sections are needed to clearly visualize the wall layers (Fig. 6.3). Both of these constraints mean that embedding of

138

L. Donaldson

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 sectioning 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 identifying 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 combination with light microscopy, to examine the complex delamination associated with sawn surfaces (Donaldson et al. 2007).

6.3.2 Confocal Microscopy In some cases, where delamination is associated with chemically altered samples, such as weathered or decayed material, it may be useful to use high-resolution confocal fluorescence microscopy to show the relationship between the chemical degradation and the delamination. This approach has been used to study the effect of reduced lignification on the occurrence of shelling in drought-grown radiata pine (Donaldson 2002) (Fig. 6.4). In some cases, examination of both unembedded and embedded material may be an advantage. Confocal microscopy has the added advantage of allowing some depth perspective, which is particularly useful for imaging rough fracture surfaces (Donaldson et al. 2007). Using confocal reflectance microscopy can be a useful alternative to scanning electron microscopy for identifying the exposure of cell wall layers on the fracture surface, allowing identification of layers by their texture (Donaldson and Frankland 2004).

6.3.3 Electron Microscopy Both scanning (SEM) and to a lesser extent, transmission electron microscopy (TEM) have been used to study fracture surfaces and wood deformation processes in a wide variety of wood-based materials (Borgin 1971b; Côté and Hanna 1983; Sell and Zimmermann 1993; Donaldson 1997; Zimmermann and Sell 1997;

6

Delamination of Wood at the Microscopic Scale

139

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 reexamination 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 understanding the delamination properties of wood and wood-based materials, as shown by recent attempts to model this behavior, which have shown the need to incorporate 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 shearing 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 deterioration 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

Donaldson LA, Bardage S, Daniel G (2007) Three-dimensional imaging of a sawn surface – A comparison of, confocal microscopy, FESEM and serial sectioning methods. Wood Sci Technol 41:551–564 Donaldson LA, Frankland A (2004) Ultrastructure of iodine treated wood. Holzforschung 58: 219–225 Donaldson LA, Singh AP (1998) Bridge-like structures between cellulose microfibrils in radiata pine (Pinus radiata) kraft pulp. Holzforschung 52:449–454 Donaldson LA, Xu P (2005) Microfibril orientation across the secondary cell wall of radiata pine tracheids. Trees 19:644–653 Evans PD, Wallis AFA, Owen NL (2000) Weathering of chemically modified wood surfaces. Natural weathering of Scots pine acetylated to different weight gains. Wood Sci Technol 34:151–165 Feist WC, Hon DN-S (1984) Chemistry of weathering and protection. In Rowell RM (ed) The chemistry of solid wood. Advances in chemistry series 207. American Chemical Society, Washington, DC, pp 401–451 Fengel D (1978) On the fibrillar structure of cellulose from wood. Holzforschung 32:37–44 Fernando D (2007) Ultrastructural characterization (morphological and topochemical) of wood pulp fibres; Effects of mechanical and Kraft processes. Doctoral thesis, Swedish University of Agricultural Sciences, p 56 Fernando D, Rosenberg P, Persson E, Daniel G (2007) Ultrastructural aspects of fibre development during the stone groundwood process: new insights into derived pulp properties. Holzforschung 61:532–538 Fruhmann K, Burgert I, Tschegg EK, Stanzl-Tschegg SE (2003a) Detection of the fracture path under tensile loads through in situ tests in an ESEM chamber. Holzforschung 57:326–332 Fruhmann K, Burgert I, Stanzl-Tschegg SE, Tschegg EK (2003b) 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 Gindl W, Teischinger A (2003) Comparison of the TL-shear strength of normal and compression wood of European larch. Holzforschung 57:421–426 Glerum C, Farrar JL (1966) Frost ring formation in the stem of some coniferous species. Can J Bot 44:879–886 Green H (1962) Compression-caused transverse discontinuities in tracheids. Pulp Paper Mag Can 62:155–168 Huang CL (1995) Revealing fibril angle in wood sections by ultrasonic treatment. Wood Fibre Sci 27:49–54 Irbe I, Andersons B, Chirkova J, Kallavus U, Andersone I, Faix O (2006) On the changes of pinewood (Pinus sylvestris L.): chemical composition and ultrastructure during the attack by brown-rot fungi Postia placenta and Coniophora puteana. Int Biodeterior Biodegradation 57:99–106 Jeronimidis G (1976) The fracture of wood in relation to its structure. In: Baas P, Bolton AJ, Catling DM (eds) Wood structure in biological and technological research. Leiden Bot Ser 3:253–265 Jokerst RW, Stewart HA (1976) Knife- versus abrasive-planed wood: quality of adhesive bonds. Wood Fibre 8:107–113 Keunecke D, Stanzl-Tschegg S, Niemz P (2007) Fracture characterization 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 Kibblewhite RP (1972) Effect of beating on fibre morphology and fibre surface structure. Appita 26:196–202 Kifetew G, Thuvander F, Berglund L, Lindberg H (1998) The effect of drying on wood fracture surfaces from specimens loaded in wet condition. Wood Sci Technol 32:83–94 Kim JS, Singh AP, Wi SG, Koch G, Kim YS (2008) Ultrastructural characteristics of cell wall disintegration of Pinus sp. in the windows of an old Buddhist temple exposed to natural weathering. Int Biodeterior Biodegradation 61:194–198

142

L. Donaldson

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 threedimensional 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 microscopic 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 woodcoating 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 Adya P. Singh and Bernard S.W. Dawson

Contents 7.1 7.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Confocal Laser Scanning Microscopy (CLSM) . . . . . . . . . . . . . 7.2.1 High Resolution CLSM Examination of Wood-Coating Interface: Coating Penetration into Cell Wall Micro-Cracks . . . . . . . . . 7.2.2 Combined Light Microscopy (LM), Confocal Laser Scanning Microscopy (CLSM) and Scanning Electron Microscopy (SEM) Reveals a Complex Wood-Coating Interaction in a Highly Textured Wood Surface-Coating Interface . . . . . . . . . . . . . . . . . 7.3 Wood-Coating Interface Examined by Field Emission Scanning Electron Microscope (FE-SEM) in Combination with Backscattered Electron Imaging (BEI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

145 147

. . . . .

147

. . . . .

149

. . . . . . . . . . . . . . .

152 155 156

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 microorganisms. Application of coatings to the exposed surfaces of wood products, that can prevent solar radiation and water from reaching wood tissues, can provide protection 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. A.P. Singh (B) Wood and Biofibre Technologies, Scion Te Papa Tipu Innovation Park, Rotorua 3010, New Zealand e-mail: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_7, 

145

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 interaction and coating performance (Richter et al. 1995) because of the differences in the introduced morphological and structural changes, which can occur at both cellular 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), abrasive 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 (bandsawn) 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 physical 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 penetrate 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

more clearly understanding the physical aspects of wood-coating interactions, and the information presented here is based on the high resolution work undertaken in three recent studies (Singh and Dawson 2004, 2006; Singh et al. 2007).

7.2 Confocal Laser Scanning Microscopy (CLSM) Most studies aimed to understand coating penetration characteristics and woodcoating interaction are based largely on light microscopy. While useful in revealing coating penetration into larger voids within wood tissues, such as lumens of vessels, tracheids and ray cells and in resin canals, and assessing the depth of coating penetration into both soft and hard woods (Côté and Robinson 1968; de Meijer et al. 1998, 2001; Nussbaum et al. 1998; Rijckaert et al. 2001) these studies did not provide information on coating penetration into the voids of smaller dimensions, such as cell wall cracks and delaminations that are often present in machined, sanded and saw-textured surfaces (Singh and Dawson 2004, 2006). A recent study aimed to probe the wood-coating interface at resolution greater than obtainable with light microscope demonstrated the usefulness of CLSM in obtaining images much superior in definition than obtainable using light microscope (LM), more clearly revealing the pattern of coating distribution within the surface tissue layers of coated panels (Singh and Dawson 2004). This work will now be considered in greater detail.

7.2.1 High Resolution CLSM Examination of Wood-Coating Interface: Coating Penetration into Cell Wall Micro-Cracks The work of Singh and Dawson (2004) involved CLSM examination of the woodcoating 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 achieving 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 sections 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

The greatest advantage of LM is in a more rapid screening of sections, which is particularly useful in obtaining quantitative information on a feature based on screening of a sufficiently large number of sections, such as determining depths to which coatings can penetrate into wood, an aspect that has been widely investigated using LM (de Meijer et al. 2001). In the LM illustration provided here (Fig. 7.1) from the work of Singh and Dawson (2004), the coating (stained red with Sudan IV) is well differentiated from wood cell walls (stained light blue with toluidine blue), enabling coating penetration into the lumens of tracheids and rays to be clearly resolved. However, the wood-coating interface appears fuzzy, making it difficult to determine whether or not the coating has also penetrated into cell walls. The high resolution CLSM image illustrated in Fig. 7.2 from the work of Singh and Dawson (2004) provides evidence that in addition to filling larger voids the polyurethane coating has also penetrated into fine cracks and delaminations that

Fig. 7.2 CLSM view of machined wood-polyurethane interface in a section. The penetration of coating (crimson colour) into cell lumens as well as tiny cell wall (purple colour) cracks (arrowheads) can be clearly resolved. Bar=20 μm

7

Probing the Wood Coating Interface at High Resolution

149

had formed in the surface tissue layers of the panel during machining. This information 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 coating 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.

7.2.2 Combined Light Microscopy (LM), Confocal Laser Scanning Microscopy (CLSM) and Scanning Electron Microscopy (SEM) Reveals a Complex Wood-Coating Interaction in a Highly Textured Wood Surface-Coating Interface 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, particularly 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 information on the micromorphology of surface wood tissues in relation to coating distribution. This new knowledge formed the basis for understanding why stain coatings 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 woodcoating 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 woodcoating 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 curling. 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

a

151

b

c

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 surface 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, apparently 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.

7.3 Wood-Coating Interface Examined by Field Emission Scanning Electron Microscope (FE-SEM) in Combination with Backscattered Electron Imaging (BEI) 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 mechanical 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 textured 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 coating 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 temperature. 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 differentiate 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 contributory 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 composites 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 differentiation. 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

However in future studies aimed to image wood-coating interface at high resolution using the FE-SEM-BEI approach one should be mindful that not all coating types may react with OsO4 , and it may be necessary to experiment with other high atomic number stains. Additionally, the optimization of backscattered electron signals requires consideration also of other factors, such as the concentration of stain and length of treatment, and operating conditions of the SEM itself, with particular attention given to setting the accelerating voltage suitable for maximizing the output of backscattered electrons. The excellent brightness contrast of the coating relative to wood cell walls achieved in the work of Singh et al. (2007) was a result of optimization of all these factors, while ensuring protection of the sections against damages from the electron beam, as high beam intensities can distort wood tissues, and damage the

Fig. 7.6 High magnification FE-SEM-SEI (a) and FE-SEM-BEI (b) of the same region of bandsawn 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 distinguishable 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

wood-coating interface. In this connection, use of a FE-SEM was important in this work, as this instrument has the ability to produce high resolution images at relatively low kV, which is not possible with a conventional low resolution SEM. The technique developed by Singh et al. (2007) for backscattered imaging clearly differentiated the coating material from wood cell walls, enabling the coating distribution at the wood-coating interface to be clearly visualized. A comparison of identical areas of the wood-coating interface imaged both in SEI and BEI modes, as shown in Figs. 7.5a and b, 7.6a and b, illustrates this point. In the low magnification BEI image shown in Fig. 7.5b, an intricate pattern of coating distribution within the highly distorted masses of surface tissues is readily observable, because the coating appears much brighter than wood cell walls. The coating is present within the large surface depressions and voids formed from cracking and dislocation of surface tissue masses, and also fills the lumens of individual cells, which appear greatly distorted. In comparison, in the SEI image shown in Fig. 7.5a of the same tissues region, the coating-penetrated surface tissues regions appear as a homogenous mass, because the brightness contrast of the coating material is similar to that of wood cell walls, and thus the pattern of coating distribution is not clearly resolvable. The usefulness of the BEI method of imaging wood-coating interface proved to be even more remarkable when the interface was examined at high magnifications (Singh et al. 2007). A high magnification BEI image of the wood-coating interface mode illustrated in Fig. 7.6b reveals that in addition to filling the lumens of tracheids the coating has penetrated the spaces of much smaller dimensions, present in cell walls as cracks and delaminations. In the SEI image of the same region of wood-coating interface and taken at the same magnification (Fig. 7.6a) these features are not resolvable.

7.4 Summary The information presented in this article highlights the advances made in understanding 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 distribution 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 dimensions, 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

understanding of wood-coating interactions, as it became evident that the coatings had penetrated into cell wall cracks and delaminations varying in dimensions from micrometer to nanometer scales, in addition to cell lumens. It is widely regarded that coating adhesion plays a crucial role in coating performance, and in this regard mechanical interlocking of coatings into the wood surface is important. The high resolution images of the wood-coating interface, which show widespread presence of tiny voids in cell walls in the surface and subsurface tissues, lead us to better appreciate the extent to which the voids can increase the space for coating penetration and entaglement and the surface area for the adhesion of coatings to cell walls. Acknowledgements We thank Dr. Ray Dickie, Editor of the Journal of Coatings Technology and Research, for permission to reproduce the figures illustrated in this article.

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 electron 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 comparison 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 woodcoating interface. J Coat Technol Res 1(3):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. 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 8.2 8.3 8.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . The Mechanism of Wood Delamination Induced by Microwaves Dielectric Properties of Wood . . . . . . . . . . . . . . . . Solid Wood Delaminations in Various Microwave Applications . 8.4.1 Impregnation . . . . . . . . . . . . . . . . . . . . . 8.4.2 Drying . . . . . . . . . . . . . . . . . . . . . . . . 8.4.3 Bending . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Controlling the Microwave Delamination Zone in Solid Timber 8.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 8.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

159 160 162 163 164 165 165 167 169 169 170

8.1 Introduction High intensity microwave treatment creates changes in wood structure. The application 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 macroelements (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 applications, 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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_8, 

159

160

G. Daian

Delamination occurrence and the extent of cell rupture depend mainly on the intensity and exposure time to the microwave’s electric field (Vinden and Torgovnikov, 2000). The intensity of microwave power should be carefully selected and controllably distributed to wood structure (by using optimal microwave applicators design) to achieve an aimed modification level and avoid wood structural deterioration. Optimal microwave applicators design can be theoretically accomplished through 3D electromagnetic simulations (Daian et al. 2009). The characterization of dielectric properties of wood enables electromagnetic simulators to model the electromagnetic field interactions with the wood (Daian 2005) and hence to control the microwave delamination process.

8.2 The Mechanism of Wood Delamination Induced by Microwaves To understand the mechanism of wood delamination induced by microwave energy, one has to recognize how the microwaves work when irradiating wood. Comprehensive and scientific details on the microwave-wood interaction are reviewed by Daian (2005). Basically, microwaves are very high-frequency radio waves, processing an electric field which rapidly reverses direction at about 2 billion cycles per second. Polar molecules (such as water) try to orient in the direction of the field, and the rapid cycling of the field in opposite directions causes the water molecules to flip backwards and forward. As a result, the wood (containing water) will heat up when placed in a microwave field (CRC Wood Innovation 2004). The generation of heat in wood under electromagnetic radiation is given by the wood ability to absorb and store electrical potential energy (dielectric properties) and establishes a temperature distribution in the material, which ultimatelly leads to changes in wood moisture and density. Heat losses, heat diffusion and specific heat phenomena create dynamic changes in the temperature field (Tinga 1993) and internal pressure, implicitly. Consequently, a non-linear change of the dielectric properties of wood occurs during microwave heating process. Figure 8.1 depicts the inherent feedback paths of the microwave heating process of wood. Microwave heating of wood is described as four-phase phenomena (Perre and Turner 1999): – 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

Fig. 8.1 Physical model of microwave heating of wood (Daian 2005)

– Thermal runaway phase – commences when the wood becomes dry in certain 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 pressure 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 boiling 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 (tracheids, 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 investigated microwave processes and technologies for achieving various forms of microwave modification of wood, ultimately proven to have industrial applicability (Australian Forests and Timber News 2007; 2008). Three license agreements were completed with industry: with TimTech Chemicals – an advanced preservation 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.

8.3 Dielectric Properties of Wood The response of wood at the microwave field oscillations depends on the wood ability 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 dielectric 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 properties 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 temperatures 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 technique, 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 suggested 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 opposite 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 therefore, estimations were determined based on available data from literature, extrapolations, 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 Solid Wood Delaminations in Various Microwave Applications In this section, three microwave applications are discussed: impregnation, drying and bending.

164

G. Daian

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 preservative 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 uniform 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

Fig. 8.2 Two characteristic features of a transversal section of a specimen (90×60 mm2 ) with a high degree of structural microwave modification of Eucalyptus regnans impregnated with resin (Przewloka et al. 2007, Figure 1)

Fig. 8.3 Macroscopic aspect of a specimen of Vintog composite material (60×45 mm2 ), produced from Eucalyptus oblique modified by microwave treatment (a) and resin impregnated (b) (Przewloka et al. 2007, Figure 2)

8

Delamination in Timber Induced by Microwave Energy

165

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 considered 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 information 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 tension 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 compression 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 component 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 properties 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 tangential 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.

8.5 Controlling the Microwave Delamination Zone in Solid Timber The size, distribution and frequency of micro- and macro-checks, generated by the application of microwave power in the wood structure, can be controlled and optimized according to each microwave application’s objectives. The occurrence of uniformly distributed internal checks may be intended for some microwave applications while for others, the microwave treatment effect on wood structure has to be minimized. For instance, the optimal pattern of wood delamination zone in microwave pre-drying would be targeted to minimal in terms of size, with very narrow internal voids and acceptable surface checks (Torgovnikov et al. 2003). A series of major factors and parameters influences the microwave delamination zone in hardwood species and these include (Torgovnikov et al. 2003): – – – – –

Microwave frequency, power and the intensity of the supplied energy, Microwave applicator design, Power pulsation, Treatment time or timber board speed within the applicator, Annual rings orientation of the timber boards.

In the microwave processing of wood, each application involves adjustments in the intensity of the microwave energy supplied. A lack of careful consideration of this aspect frequently leads to structural deterioration of the wood during processing. Cracking, distorting and warping are problems usually caused by the high internal vapour pressure generated by fast heating. As a result, the intensity of microwave power should be selected in such way to limit the amount of moisture removed during processing (Metaxas and Meredith 1993). The design of the microwave applicators (i.e. geometry and size) affects the intensity and distribution of energy and hence, the profile of the delamination zone. The microwave applicators can provide irradiation to selected areas of the wood: e.g. single-sided modified zone, full cross-section modification, etc. (Torgovnikov et al. 2003; Daian et al. 2009). The extent to which the wood modification by microwaves occurs depends on the treatment time and the supplied power schedule as well. Using optimized pulsing power schedules (i.e. time-on and time-off) the cracking in the direction parallel to the annual rings can be reduced (Torgovnikov et al. 2003). Timber boards are differently modified by microwaves in terms of the size of delaminated surface and internal checks, as well as appearance of ring-shakes, when considering the profile of cutting – flat-sawn or/and back-sawn (Torgovnikov et al. 2003; Daian et al. 2009).

168

G. Daian

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 treatment 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 configurations 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 addition, 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 applicator 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 revolutionary 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 optimized. 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 experimental 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: • microwave may increase considerably the permeability of wood for impregnation applications, by forming micro and macro checks in both radial-longitudinal and radial-tangential direction; • microwave pretreatment may reduce significantly the formation of internal/honeycombing and surface checking in drying applications; • microwave softening may provide enhanced quality in bending applications when wood moisture content is high. Commercialization of the microwave-wood technologies is limited to date.

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

The existing research indicates that microwave can be applied to wood in pre-set conditions to considerably modify the wood structure so that to improve performances in applications such as wood preservation, drying or bending. The exploratory efforts have been extended up to scale-up processes, equipment prototyping and pilot-scale production, but isolated and limited number of industrial applications are available to date.

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 processingFact 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, EastPetersburg, 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 9.2

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Delamination Induced by Weathering in Solid Wood . . . . . . . . . . 9.2.1 Methods of Measurements . . . . . . . . . . . . . . . . . . . 9.2.2 Factors of Influence . . . . . . . . . . . . . . . . . . . . . . 9.2.3 Structural Aspects . . . . . . . . . . . . . . . . . . . . . . . 9.3 Delamination Induced by Weathering in Wood-Based Composites Panels 9.3.1 Natural Outdoor Exposure . . . . . . . . . . . . . . . . . . . 9.3.2 Artificial Exposure . . . . . . . . . . . . . . . . . . . . . . . 9.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

173 175 176 177 178 191 191 192 192 193

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 degradation induced by the weather (sunlight, wind, precipitations, diurnal and seasonal changes in relative humidity, atmospheric pollution, etc). Knowledge about weathering 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- ultraviolet 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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_9, 

173

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) Specific gravity

Erosion (μm) after various exposure times 4 years

10 years

16 years

Species

(kg/m3 )

Latewood Earlywood Latewood Earlywood Latewood Earlywood

Douglas fir Southern pine Western redcedar Redwood

460 450

105 135

270 320

285 315

905 710

500 525

1405 1355

310

200

500

765

1320

1380

1945

360

165

405

440

835

835

1385

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 earlywood 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 different 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 different 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) Erosion (μm) after various exposure times 4 years

10 years

16 years

Products

Latewood

Earlywood

Latewood

Earlywood

Latewood

Earlywood

Western redcedar plywood Redwood plywood Douglas fir plywood

170

580

455

1090

910

1475

125

440

475

800

515

1250

110

270

255

500

515

905

9

Delaminations Induced by Weathering in Wood

175

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 wetting 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 retarding 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 fissures or cracks and the development of delaminations induced by weathering on non coated wood and wood based composites. The aspects related to photochemical and biological processes of natural or treated wood are not discussed in this chapter.

9.2 Delamination Induced by Weathering in Solid Wood 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 standards (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

9.2.1 Methods of Measurements The checks at wood surface can be detected visually or automatically. Visual detection is the oldest existing method and refers to the crack length, width, depth, number or measuring only the largest check found in the specimen or board (as noted in different standards for softwood or hardwood). The manual measurements were performed with rules, calipers, micrometers and feeler gauges. Flaete et al. (2000) used a visual detection and registered the cracks when deeper than 75% of the board thickness. Three length categories of cracks depth were registered such as: 1–50 mm, 51–120 mm and > 120 mm. The length of each crack was measured. A more refined system was proposed by Rietz (1961) who used a categorical rating system based on the total length of cracks on a scale ranging from 0 to 5. Evans et al. (1997, 2003) modeled the cracks as rhombus or two right prisms back to back. Szymany and McDonald (2004) used the reflection of laser and fluorescent light to detect checks at wood surface. For visual inspection during weathering Sudiyani et al. (1999) proposed a classification of cracking in 5 classes such as: checking occurred at less than 10% to checking occurred at more than 70% of the surface area. Wahl et al. (2001) used laser light to measure the width and depth of cracks. Kamden and Zhang (2000) used stylus profilometry to measure the width of cracks. The manual measurements are very laborious. Lopez et al. (1998) developed an automatic fibre-optic reflectometric technique for the automatic detection and measurement of surface cracks of different materials. Christy et al. (2005) automatically identified checks at wood surface in southern pine and Douglas fir using “an algorithm embedded in waveform analysis software that searches line scans within black and white digital images of the surface looking for brightness minima with certain features that are characteristics of checks”. The checks are shown black on a white background. Figure 9.1a shows the selected area of a southern pine specimen, the brightness profile along a line across the centre of the specimen and the checks detection on the brightness profile. Checks position, size and shape can be quantified rapidly and automatically. On the brightness profile along a line across the centre of the specimen (Fig. 9.1b) it is possible to detect the presence of checks as minima on the brightness profile (Fig. 9.1c). It is generally recognized today that the ultrasonic methods are perhaps the most versatile of the techniques for non-destructive testing available. Different ultrasonic techniques have been developed for crack depth measurements in metals and composites, based on analytical and numerical investigations of the reflection and transmission coefficients for normal as well as oblique incidence of ultrasonic waves (Achenbach et al. 1980; Angel and Achenbach 1984; Lidington and Silk 1975; Dong and Adler 1984; Achenbach et al. 1992; Li et al. 1992). With this methodology, some problems arisen such as the coupling between the transducers and specimens who required sophisticated calibration procedures. The modern-air coupled transducers should avoid these problems. The achievements cited in previous references could inspire in the future new methodology for crack depth measurement on wood and wood-based composites.

9

Delaminations Induced by Weathering in Wood

177

a

b

c

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

9.2.2 Factors of Influence Table 9.3 summarises the possible interactions between the weathering and the development of delaminations in non coated wood. The erosion process at wood surface is produced by rain and wind. The fissures and the delaminations are produced by the sun light – UV radiation. The dynamics of moisture content and temperature are related to the precipitations, sun radiation and air humidity and temperature. Differences between climatic regions ranging from northern temperate, Mediterranean, etc, to tropical are substantial in terms of annual duration of sunshine, annual sum of global radiation, annual sum of precipitation –rain, snow,

178

V. Bucur

Table 9.3 Possible interactions between the weathering and the development of delaminations in non coated wood No

Climate elements

Processes

1 2

Sun light, UV radiation Precipitations, sun radiation, air humidity Rain, snow, wind, driving rain

Photochemical degradation Fissures and delaminations Moisture and temperature Dimensional instability, dynamics cracks and delaminations Erosion processes Superficial wood erosion

3

Effects on non coated wood

wind, etc. The meteorological conditions of each site are typical. The severity of climate 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 + (9.1) C.I = 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 radiation 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 correlation 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.

9.2.3 Structural Aspects In this section will be discussed some wood structural aspects, at macroscopic and microscopic scale, related to delamination induced by weathering for natural outdoor exposure and for artificial exposure. 9.2.3.1 Natural Outdoor Exposure After outdoor long term exposure all wood species develop a grayish appearance due to the fact that wood extractibles are removed and the delignified fibres are

9

Delaminations Induced by Weathering in Wood

179

exposed to atmospheric conditions. The combined effect of sunlight (photochemical degradation) and water (precipitations, raining water, snow) degrades wood main components and transform the wood surface into a network strongly contaminated by microorganisms (Sell and Wälchi 1969). (a) Macroscopic Aspects 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 commonly accepted that wood anisotropy plays an important role in crack formation. For the structural uses it is recommended, to display the boards in vertical position to avoid the development of checks and splits (US Forest Products Lab 1974; Sandberg 1997). Sandberg D (1999) studied the development of cracks at macroscopic and microscopic level in radial and tangential surfaces of pine (Pinus silvestris) and spruce (Picea abies) specimens after 33 months outdoor exposure in Stockholm from September 1993 to May 1996. For these experiments the specimens have been produced with a star-sawing patter which allows the perfect delimitation of wood radial and tangential surfaces. The specimens have been planed. Figure 9.2 shows the specimens before and after exposure. The minimum crack assessed visually had 0.25 mm width. Table 9.4 gives the average of total crack length per unit area after 33 months outdoor exposure and subsequent conditioning at 12% moisture content. The ratio of crack length in tangential versus radial planes is 12.1 for pine and 9.4 for spruce. This ratio is of the same order of magnitude as the ratio between the tangential and radial shrinkage. The mass density expressed as the ration between the weight and the volume at 12% moisture content has no effect on crack development in this specific case of pine and spruce outdoor exposure to North European climate. It was noted that the degradation started on tangential surface. Surface degradation has developed a corrugated appearance which follows the pattern of the annual ring. The earlywood degrades more rapidly than the latewood. The corrugated appearance is more important on radial surfaces. The appearance of the cracks at macroscopic level is described such as: – 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. Depending on species, sometimes the durability of wood can be or not affected by harvesting time, origin and stand characteristics and growth conditions of trees. Rydell et al. (2005) reported Scots pine behavior during 9 years of weathering in

180

V. Bucur

Fig. 9.2 Specimens before and after 33 days outdoor exposure (Sandberg 1999) (a) radial surface of spruce before exposure (b) the same surface after 33 days outdoor exposure (c) tangential surface of pine before exposure (d) the same surface after 33 days outdoor exposure, with evident cracks (Sandberg 1999, Figure 5)

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) Density

Total crack length per unit are (m/m2 )

Ratio of crack length

Shrinkage ratios (T/R)∗

Species

(kg/m3 )

Radial

Tangential

–

–

Pine Spruce

475 415

2.1 2.9

25.5 27.3

12.1 9.4

Note: Data from Kollman and Côté (1968)

9

Delaminations Induced by Weathering in Wood

181

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. (b) Microscopic Aspects In that follows we discuss the microscopic aspects of delamination in some softwoods, extensively used in building industry, such as: pine (Pinus sylvestris), spruce (Picea abies Karst.) and sugi (Cryptomeria japonica D. Don). Borgin (1970) was, to our knowledge one of the first to use the scanning electron microscope for the study of weathered wood. He selected specimens from wooden structures of pine (Pinus sylvestris) in Norway exposed to the weather up to 1000 years, from Norwegian lofts and stave churches. The wood was exposed to weathering without any protective treatment. The eroded surfaces were silver grey colour and had the pattern of the year rings was corrugated, of dense and less dens parts. The destructive process was limited to 2. . . 3 mm. The very slow deterioration of this layer protected the bulk of the interior structure. ”The slow breakdown of wood followed the reverse pattern of how the fibres are built up with a fibrillar, laminated structure cemented together as a fibre – reinforced high polymer composite material”. The micrographs showed bundle of fibres eroded from the surface, partly or completely loosened surface fibres, fractured fibres with middle lamellae partially removed, bordered pits with the torus completely destroyed, eroded microfibrilar structure, etc. the middle lamella, the primary wall, the outer layer of S1 have been removed by the long time weathering. The complete destruction of tracheids was scarcely found. The most damaged sections were those of the transversal section, were fibres were originally cut off. The individual fibres were remarkably stable and durable. The microfibrils were the most stable. The adhesion between the structural elements was gradually destroyed. The sculpturing in fibres walls were gradually enlarged causing a weakening of the fibre structure. The weakest element was the middle lamellae, composed mostly from lignin. The mechanisms of wood structure breakdown due to the environmental factors were largely discussed by Borgin (1971), Borgin et al. (1975). For the solid wood exposed to outdoor conditions, as for the very old wood (extracted from the Egyptian pyramids) which was stored very long time under stable conditions of temperature and moisture content, the weakest part of the structure are the border between the middle lamella and S1 and the interfibrillar matrix. Figure 9.3 shows delaminations between tracheids induced by the weathering removal of middle lamella in pine (Pinus sylvestris) extracted from wooden structures in Norway exposed for 300 years to the cold climate of North Europe. Systematic studies on the effect of outdoor exposure on wood microstructure were reported by Sandberg D (1999). The dynamics of delamination at microscopic level in pine and spruce after 33 months outdoor exposure to North European climate was studied, It was observed that the fragments of tracheids are broken and

182

V. Bucur

Fig. 9.3 Delaminations between tracheids induced by the weathering removal of middle lamella in pine (Pinus sylvestris) from wooden structures in Norway exposed 3 centuries to the cold climate of North Europe (Borgin 1970, Figure 3)

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 dynamics, 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 border. 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 weathering (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 moisture 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 sunlight 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 earlywood. 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 orientation -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 formation 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 orientation 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

Imamura Y (1993) reported the morphological changes induced in sugi wood (Cryptomeria japonica D. Don) after exposure to natural weathering (sunlight, wind and rain) during 12 weeks in winter. The specimens exposed outdoor undergo photochemical degradation primarily in the lignin components. The most evident deterioration of the microscopic structure appeared around the pits in the cell wall. The margo fibril of the pit membranes are the first to be broken. The apertures of bordered pits were enlarged in the earlywood tracheids, “to the limit of the pit chambers, and diagonal checks through the pits were visible, which probably followed the fibril angle of the S2 layer”. During sunlight irradiation the cell wall shrinks. This shrinkage generates local concentration of stresses in cell wall, which determine the appearance of microcracks along the middle lamellae, along the border between S1 and S2 , at the limit between the latewood and earlywood. The most characteristic pattern was the erosion of the middle lamellae in which lignin reaches its maximum concentration, showing round appearance of tracheids cross section (Fig. 9.7). Hon and Feist (1986) studied the behaviour of four hardwood species – red oak, white oak, yellow poplar and sweetgum to weathering during 30 days from September1981 to January 1982 at Blacksburg, Virginia, USA. Figure 9.8 shows the transversal section of yellow-poplar before and after outdoor exposure for 30 day. In transversal section the erosion of middle lamella was observed and the roughening at the surface of the microscopic section. The erosion of the middle lamellae can be considered as the structural initiation of a crack. On the tangential section after 30 days of exposure, slight separation of the procumbent cells in rays and roughening of cell walls were observed. In radial microscopic section a roughening of cell was also observed.

Fig. 9.7 Round appearance of transversal section of eroded tracheids in sugi after 8 weeks exposure to natural weathering (Imamura Y 1993, Figure 4A) Note the delamination of the cell wall structure at S1 and S2 boundary (see arrows)

186

V. Bucur

a

b

c

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 exposure. 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

9.2.3.2 Artificial Exposure (a) Macroscopic Aspects The factors which influence the delamination produced by the artificial exposure are the same as those for outdoor exposure. Sandberg D (1996) studied the influence of pith and juvenile wood on proportion of cracks in sawn timber when kiln dried and exposed to wetting cycles on boards of Norway spruce and Scots pine and have shown that cracks are related to the juvenile wood, but the crack formation is not related with the annual ring orientation. Sandberg D (1996, 1997) observed in his experiments that the juvenile wood increased crack formation. This can be explained by the properties of juvenile wood which shrinks in longitudinal direction more then adult wood when drying. Flaete et al. (2000) studied the crack formation in aspen (Populus tremula L.) and Norway spruce (Picea abies (L) Karst.) during accelerated weathering. The boards were exposed to four climate regimes, 1 h/regime such as: light and heat radiation, water spray, frost and room temperature. The weathering device worked continuously 112 days. The cracks developed more in Scots pine than in aspen. The

9

Delaminations Induced by Weathering in Wood

187

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 periphery 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 direction. 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. (b) Microscopic Aspects The technological advancement with scanning electronic microscopy allowed the developments of studies related to the fine structure of wood. Numerous micrographs related to the photodegradation and weathering of wood have been published in the last three decades (Sell and von Luekens 1971; von Luekens and Sell 1972; Raczkowski 1980; Kucera and Sell 1987; Kuo and Hu 1991). In that follows we selected several to illustrate the crack formation and delaminations in fine structure of wood induced by artificial exposure to weathering for species from temperate zone (European) and tropical. The delamination effect in the transversal section of Southern pine cross section after exposure to UV (λ > 200 nm) during 1000 h can be observed in Fig. 9.9. The middle lamella was completely eroded, UV radiation producing delamination between the tracheids. The deterioration of the pits after 1000 h UV exposure generated delamination at the border of pits which extends in alignment with the microfibril orientation (Fig. 9.10). The high energy protons degraded the lignin and the cohesion between wood anatomical elements. Hon and Feist (1986) reported the effect of UV (λ > 220 nm) irradiation on yellow poplar after 500 h, 1000 h and 2000 h. Figure 9.11 shows the corresponding

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

deterioration of transversal microscopic structure (middle lamellae, checking and roughening of the cell wall). After 2000 h of exposure the delamination is evident through the separation of cells. It was reported that the tangential section the development of the very rough surface is due to the degradation of lignin and consequently the microfibrils emerged to the surface. The pits were severely damaged generating the delamination into the adjacent cell wall (Fig. 9.12). Kishino and Nakano (2004) reported the development of delamination produced by artificial weathering of eight tropical species (auri – Acacia auriculiformis; bangkai – Shorea spp, cumaru – Amnurana acreana, ipe – Tabebula spp, jahhra – Eucalyptus marginate, keruing – Dipterocarpus spp, robusta – Eucalyptus robusta). The development of cracks has been studied in relationship with the wettability from the prospective of chemical and structural modifications of wood surface. Each 120 min weathering cycle was composed from 120 min light irradiation (300 nm<λ<700 nm, incident light intensity 390 W/m2 ) followed by 120 min of water spray. The irradiation time started with 20 min and finished with 600 min. For further measurements the specimens were conditioned at 20◦ C and 65% relative humidity. For stereoscopic microscopic observations sections (2 ×25×25 mm) were prepared. Microscopic image analysis was used to determine cracks average length and width. Ratio of cracks area to section area was determined. Figure 9.13 shows stereoscopic micrographs of three species (auri – Acacia auriculiformis, robusta – Eucalyptus robusta and ipe – Tabebula spp) before and after 600 h artificial weathering. Figure 9.14 shows the variation of the contact angle versus the parameters of cracks (area, average length and width). Small cracks were observed in ipe – Tabebula spp for specimens which having a relatively low specific gravity (0.90 g/m3 ) and showing the highest contact angle of 88.9◦ . The largest cracks were observed in cumaru – Amnurana acreana, – 930 kg/m3 for which the smallest mean contact angle of of 50.5◦ was measured. It was concluded that the differences

190

V. Bucur

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.

9.3 Delamination Induced by Weathering in Wood-Based Composites Panels The aim of section is to review the studies focused on the development delaminations induced by weathering in wood-based composites panels produced with natural, untreated wood.

9.3.1 Natural Outdoor Exposure 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 yellowpoplar 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 evident 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 recommended 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.

9.3.2 Artificial Exposure The artificial exposure to accelerated aging of wood based composites is regulated by numerous international and national standards (ASTM D 1037; ASTM D 2017; ASTM D 2481; ASTM 2898). Aging tests have been performed to observe by indirect destructive or nondestructive methods the internal bond characteristics of wood – based composites. The internal bond strength is a parameter which is directly related to delamination. The methodology for accelerated aging test is standardized and is not treated in this chapter. Generally speaking this approach describes only quantitatively the delamination without details on structural modification. The main purpose of this methodology is related to the quality of products and it was conceived to provide valuable practical information to those consumers and builders who use and install wood based composite panels and products

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 depolymerization. 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 automatically. 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 commonly 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 microscopic 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 engineering 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 10.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Influence of Temperature, Relative Humidity and Rate of Air Circulation on Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.1 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.2 Relative Humidity . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2.3 Air Circulation Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Wood Drying Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.1 Air Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3.2 Kiln Drying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4 Kiln Drying Schedules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 Drying Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6 Prediction of Drying Stress and Strain Using Mathematical Models . . . . . . . . 10.6.1 Hardwood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6.2 Softwood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

197 199 199 200 200 200 201 201 203 204 205 207 208 211 212

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: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_10, 

197

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 processing. 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 moisture 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 product, 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 movement 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 without 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 timber 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 percentage 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 characteristics 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. Prompt drying of wood immediately after felling therefore results in significant upgrading of, and value adding to, the raw timber. Drying enables substantial long term economy in timber utilization by rationalizing the utilization of timber resources. The drying of wood is thus an area for research and development, which concerns many researchers and timber companies around the world.

10.2 Influence of Temperature, Relative Humidity and Rate of Air Circulation on 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 holding 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 unsaturated 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 indicates 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.

10.2.2 Relative Humidity 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).

10.2.3 Air Circulation Rate Drying time and timber quality depend on the air velocity and its uniform circulation. At a constant temperature and relative humidity, the highest possible drying rate is obtained by rapid circulation of air across the surface of wood, giving rapid removal of moisture evaporating from the wood. However, a higher drying rate is not always desirable, particularly for impermeable hardwoods, because higher drying rates develop greater stresses that may cause the timber to crack (i.e. delamination) or distort. At very low fan speeds, less than 1 m s−1 , the air flow through the stack is often laminar flow, and the heat transfer between the timber surface and the moving air stream is not particularly effective (Walker et al. 1993). The low effectiveness (externally) of heat transfer is not necessarily a problem if internal moisture movement is the key limitation to the movement of moisture, as it is for most hardwoods (Pordage and Langrish 1999).

10.3 Wood Drying Methods 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 drying (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 drying 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.

10.3.1 Air Drying 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 consists 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 available. 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 drying, 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. 10.1 The air drying yard at Boral Timber’s Herons Creek site

10.3.2 Kiln Drying Kiln drying is a method of drying in which air temperature, humidity and velocity can be adjusted to control the loss of moisture from wood and enable particular or specified MC to be achieved. This is generally used in the context of conventional kiln drying. The process of kiln drying consists primarily of drying wood using introduced heat sources (directly, using natural gas and/or electricity; indirectly, through steam-heated heat exchangers, although solar energy is also possible). In

202

N. Haque

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 timber 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 manipulation 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. Compartment-type kilns are most commonly used by timber companies. A compartment kiln is filled with a static batch of timber through which air is circulated. In these types of kiln, the timber remains stationary. The drying conditions are successively varied from time to time in such a way that the kilns provide control over the entire charge of timber being dried. This drying method is well suited to the needs of timber companies, which have to dry timbers of varied species and thickness, including refractory hardwoods that are more liable than other species to check and split (i.e. delamination).

10.4 Kiln Drying Schedules Satisfactory kiln drying can usually be accomplished by regulating the temperature and humidity of the circulating air to suit the state of the timber at any given time. This condition is achieved by applying kiln-drying schedules. The desired objective of an appropriate schedule is to ensure drying timber at the fastest possible rate without causing objectionable degrade such as delamination. The following factors have a considerable bearing on the schedules. • The species; because of the variations in physical, mechanical and moisture transport properties between species. • The thickness of the timber; because the drying time is approximately inversely related to thickness and, to some extent, is also influenced by the width of the timber. • Whether the timber boards are quarter-sawn, back-sawn or mixed-sawn (e.g. bastard sawn); because sawing pattern influences the distortion due to shrinkage anisotropy. • Permissible drying degrade; because aggressive drying schedules can cause timber to crack, delaminate and distort. • Intended use of timber; because the required appearance of the timber surface and the target final MCs are different depending on the uses of timber. • Considering each of the factors, no one schedule is necessarily appropriate, even for similar loads of the same species. This is why there is so much timber drying research focused on the development of effective drying schedules.

204

N. Haque

10.5 Drying Defects Drying defects are the most common form of degrade in timber, next to natural defects such as knots (Desch and Dinwoodie 1996). Drying degrades can be divided into two broad categories: • 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 diamonding. 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 generally 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.

Fig. 10.3 Some defects due to uncontrolled drying

10

Delamination in Timber Induced by Drying

205

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 quality 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.

10.6 Prediction of Drying Stress and Strain Using Mathematical Models 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 dryinginduced 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 longitudinal 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 predict 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 schedule). 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 tangential 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 arising 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 timber, including different cell wall thicknesses, mean that there is a range of shrinkages for various timbers. If all other timber properties (mechanical, physical, and transport 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 gradient 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 optimized 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 significantly 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 Drybulb

Wetbulb

Strain 5

Moisture Drybulb Wetbulb

70

Strain

4

60 50

3

40 2

30 20

1

10 0

0 0

10

20

30

40 50 Time (days)

60

70

80

90

Fig. 10.4 Predicted strain and MC for a typical optimized drying schedule for hardwood

Instantaneous strain (%)

Temperature (°C) & Moisture content (%)

Moisture 80

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 consideration 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 collapse 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 sample, 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 properties 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 drying, 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 60/40 60/30 70/40 70/30 90/60 120/70 140/90

Moisture content (%)

140 120 100 80 60 40 20 0 0

20

40

60 80 Time (hours)

100

120

140

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. 10.6.2.1 Low Temperature (LT) to Accelerated Conventional Temperature (ACT) The predicted stresses for range of drying schedules from 50/40 (i.e. dry-bulb 50◦ C and wet-bulb 40◦ C) to 90/60 are shown in Figs. 10.6 and 10.7. Figure 10.6 shows predicted surface stress and Fig. 10.7 shows predicted stress at the centre of a 40 mm board. 2

Surface stress (MPa)

1 0 0

20

40

60

80

100

120

–1 –2 –3 –4 Time (hours)

Fig. 10.6 Predicted surface stress of a 40 mm board for a range of schedules

90/60 70/40 70/30 60/30 60/40 50/40

210

N. Haque 1.2

Stress at centre (MPa)

1 0.8 0.6 0.4 0.2 0 –0.2

0

20

40

60

80

100

120

90/60 70/40 70/30 60/40 60/30 50/40

–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 gradient 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 compression 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, delamination will occur either internally or on the surface resulting in internal or surface checks on the radiata pine timber board. 10.6.2.2 High Temperature Drying (HT) Drying above 100◦ C is termed as high temperature (HT) drying. Predicted drying stress during high temperature drying at surface and centre of the board is shown. in Fig. 10.8 For two HT schedules, the predicted stress results are similar to ACT

10

Delamination in Timber Induced by Drying

211

2

120/70 Surface 140/90 Surface 120/70 Centre 140/90 Centre

Surface stress (MPa)

1.5 1 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 product 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 condition 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 hightemperature 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 Chih Lin Huang

Contents 11.1 11.2 11.3 11.4

Introduction . . . . . . . . . . . . . . The Unique Wood Structure . . . . . . The Raw Materials: Tree and Stem/Log . The Manufacturing Processes . . . . . . 11.4.1 Solid Wood Products (SWP) . . . 11.4.2 Engineered Wood Products (EWP) 11.5 Product Failure and Product Durability . 11.5.1 Product Failure . . . . . . . . . 11.5.2 Product Durability . . . . . . . 11.6 Summary . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

215 217 220 223 223 226 230 231 232 236 236

11.1 Introduction The highly-competitive forest industry is facing mounting pressures from both substitutions and imports. The North America industry is not only adjusting to the changing wood quality and timberland ownership but is also coping with the cyclical 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 couple 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 operationally feasible. For example, while it is possible to sort raw materials at various

C.L. Huang (B) Weyerhaeuser Technology Center, 32901 Weyerhaeuser Way S, Federal Way, WA 98001, USA e-mail: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_11, 

215

216

C.L. Huang

stages throughout the value chain to reduce lumber variability, such efforts may not be cost-effective for the following reasons: • added plant floor space needed to accommodate the additional sorts, • labor costs for additional handling, as well as • extensive changes to mill processes. 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: • ban on CCA (Chromated Copper Arsenate), • boiler MACT (Maximum Achievable Control Technology), • government’s preferential procurement policy, and other incentives on recycling and biofuels, • timber tax, and • natural-disaster related changes to building codes. 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 lumber, 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 products. 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 manufacturing. 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 supply 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: • • • • •

Adhesives and wax Connectors Fasteners Plastics Fiber-reinforced polymers Each of these could contribute to the delamination of the EWP product. Wood has its own set of contributors to delamination failures:

• micro-fissures on the cell wall, • breakage such as checks, splits, shakes, • and cracks across the grain and cell walls 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 delaminations that originate in the unique wood structure of the living tree, as well as, through manufacturing, and finally, the products in-service.

11.2 The Unique Wood Structure Poems were made by fools like me, But only God can make a tree. Joyce Kilmer American poet (1886 – 1918)

218

C.L. Huang

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 environment, 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 (earlywood 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 density. Besides easy splitting characteristics along the LR planes, ring curvature and the discontinuity of density from the first formed earlywood are additional macrocharacteristics 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 glucose or fibrils are tightly bundled or braided into microfibrils which are priminarily crystalline. The microfibril bundles are embedded in a matrix of amorphous hemicellulose 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 hemicellulose 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 secondary 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 microfibrils. 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 cambium, 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 tangential 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 hemicellulose, 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 tension 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 longitudinal 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 shrinkage 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.

11.3 The Raw Materials: Tree and Stem/Log Various types of fiber separations and breakages may happen in a tree’s life. For example, cracks and splits can be caused by: • frost, lightening; or wind, • internal checking and ring shake associated with stresses or bacterial infection; and, • fractures/breakages from mechanical handling damage. 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 neighboring 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 hemlock, 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

Fig. 11.2 Severe star checks and intra-ring checks of an oven-dried Eucalyptus disc

222

C.L. Huang

Fig. 11.3 Failure due to ring shake of a piece of glulam

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 harvesting and handling techniques future saw log values can be improved, especially for species susceptible to split, crack, shake, and check. Properly controlled timing, 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 soundness 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 manufacturing processes where “the wood really meets the steel”.

11.4 The Manufacturing Processes 11.4.1 Solid Wood Products (SWP) Lumber grade recovery (#2 and better) and mill throughput (MBF/hour) are essential 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 misaligned 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 temperature 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 proficiency, 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 temperature 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 personnel. 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 process 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 collection 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 center 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 experienced 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 modern 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 example, 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 grading 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 straightforward, 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

Engineered Wood Products (EWP) randomizes the defects, so concentrated strength-reducing areas are eliminated resulting in a more uniform product. Through highly-aligned fibers and panel densification, strand products such as laminated strand lumber (LSL) and oriented strand board (OSB) convert low-value wood into high-value stiff products. With the declining of wood quality, there is growing demand for product stiffness, stability and uniformity.

11.4.2 Engineered Wood Products (EWP) Glulam is one of the oldest EWP. With two or more layers of finger-jointed lumber 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 adhesive, 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 glulam 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 randomization 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 underdried 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 technology 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 reactivity 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 compounds (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 curing 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: • Phenol formaldehyde (PF) is a water-resistant adhesive commonly used in plywood, LVL, PSL, and OSB production. • Urea formaldehyde (UF) is a low-cost adhesive used for interior non-structural products such as particleboard and medium density fiberboard (MDF). • Resorcinol formaldehyde (RF) cures in room temperature or ambient conditions, but has a short use time or pot life. • Melamine formaldehyde (MF) has good durability and low-formaldehyde emissions, but it is more expensive than PF. • Diphenylmethane-diisocyanate (pMDI) is expensive, but it has excellent reactivity with water and wood and is often used in the core of OSB. It is also suitable for the steam injection press in LSL production. 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 having 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 content 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 manufacturing 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 delaminations. 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 online, 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: • The molecule size, distribution, and the reactivity of the adhesive, • MC and grain orientation of wood, • curing temperature, pressure, and press time. 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 general, 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 pretreatment are ways to facilitate a strong bond that will prevent the delamination of the products.

11.5 Product Failure and Product Durability Designing a structure to resist failure in any service condition is ideal; however, extreme stresses will cause failures whether the stresses are induced by natural catastrophes, or by excessive loads, beyond the design limits of the structure. Although extreme stress loads such as these are not covered in the limited warranty of the product, failures caused by natural disasters may trigger product recalls or sometimes aggressive changes to code requirements. In some cases, these code changes may be based on emotional reactions to a natural disaster such as hurricane and exceed reasonable standards. Regardless, the industry is required to adapt processes and products to meet these new code changes. It is in the best interest for the manufacturer to actively participate in the code-changing processes because once changes are made to code standards they are difficult to undo. Similar to QC programs for the manufacturing processes, quality assurance (QA) programs make sure products are within performance specifications for their intended in-service use. Most QA programs are proprietary to each manufacturer; therefore, the QA can be done internally or by an independent agency that meets industry standards for sample testing and frequencies. ASTM International (American Society for Testing and Materials) is an example of testing standards that laboratories use to conduct their tests. Other QA programs such as a well-established product performance tracking system can help the manufacturer capture data and pinpoint the extent and causes of failed products. RFID (Radio Frequency Identification) and bar code are examples of a product tracking system, but there are other technologies that can be used to collect performance data.

11

Industry Prospective of Delamination in Wood and Wood Products

231

11.5.1 Product Failure Intrinsic anisotropic shrinkage and growth stresses contribute to the delamination of wood products, as well as the product’s in-service MC gradients due to the fluctuating environment. These variables create swelling-shrinking stresses that either initiate or worsen the delamination of the wood product. Failures of most in service SWP are due to the ever swelling-shrinking stresses combined with mechanical stresses (load, tensions) and product fatigue. Wood is high in tension strength; however, knots on the tension side of a structural member in bending, become the concentrated nucleus for failure. This is because there are swirling-grain patterns around knots which makes it easier for the knots to separate from the wood surrounding it, or to open when forces are normal to the grain (Fig. 11.5). The in service failure of SWP can also be initiated at other types of natural discontinuities of strength in clear wood regions such as: • • • •

the LR planes and the rays, the first formed earlywood, cell wall checks of CW, and the CML-S1 or S1/S2 boundaries.

Depending on the amount of the defect-randomization, the cause of the EWP in-service failure has at least two aspects of complexity: first, the size, shape, and orientation of the wood element; and secondly, the adhesive curing and its interactions to the wood element, FRP, or both. Once cured, some adhesives are stronger than wood, so the strength-discontinuities of the wood are toughened. Changes in the wood structure after densification make the study of the EWP delamination

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 identifying 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 concern 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 evaluating the in-service product applications, the type and frequency of support loads. Bracing and connectors are also important factors regarding the load or stresstransfer 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 propagations. 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 excessive 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 homeowners, 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 personal injuries and property damage. As expected, the durability of wood products is more demanding for outdoor applications.

11.5.2 Product Durability The creep and relaxation phenomena of wood are well-known: strain increases under long-term constant loads and internal stress gradually relaxes under long-term constant strain. Long-term behaviors of adhesives demonstrate these characteristics as well. Deformation rate with time is expressed as a U shaped curve where the rate is high initially (primary creep), then maintains a constant level (secondary creep), and finally increases sharply before failure (tertiary creep). The duration

11

Industry Prospective of Delamination in Wood and Wood Products

233

of each phase is determined by the magnitude of the load. These time-dependent stress-strain relationships are also related to in-service temperature and MC. Creep adjustment factors, provided by the manufacturer, should be used when designing the members under heavy loads to prevent excessive creep deflection. Installing the crowned-edge up of a floor joist is one example of a way to off-set creep. These time-dependent, mechanical stress-strain relationships are also influenced by the environment as well as chemical agents. Environmental factors that influence creep are: temperature, moisture, and sunlight. Chemical agents that influence creep are: salt-water exposure, alkaline from petroleum products, and iron-tannin reactions to metal connectors such as bolts. Thermal expansion of dry wood is anisotropic: tangential direction expands more than radial direction, and expansion in the longitudinal direction is very small. If the wood is wet, the effect of wood shrinkage dominates the dimensional changes. Loss of strength in wood is accelerated and prolonged in high-humidity and high-temperature conditions. Vacuum-pressure-soak-dry, boiling, and field testing are the common methods for evaluating the durability of products. Besides fluctuating MC and temperature, oxidization and ultra-violate (UV) radiation not only discolors the wood, but also breaks down the lignin of the wood. Loose grains on a piece of weathered flat-sawn decking lumber are often initiated at the tips of the annual rings on the pith side. Like the tip of a quill pen, tips of annual rings have a tendency to curl inwards. Exposing the vertical-grain side of the product to the weather, or installing the pith-side down, are common practices to protect the wood when building an out-door deck. The swelling and shrinking stresses on weatheredwood surfaces results in microscopic fissures that develop into checks and splits (Fig. 11.6). Delaminations from mechanical stresses, physical changes, chemical changes, and weathering are considered normal wear-and-tear of wood products. Bio-deterioration like stain, mold, mildew, decay fungi, and insects such as termites, carpenter ants, and wood-borers can lead to serious durability problems of wood products. Molds, mildew, and staining fungi do not feed on cell walls. They mostly affect aesthetic appearances, and occasionally become a health concern, but rarely have an effect on mechanical properties. Fungus thrives on wood with high MC (greater than 20% MC), so kiln-dried lumber (MC 19% and less) should be free of fungal growth. There are three types of wood decaying fungi: brown-rot, white-rot, and soft-rot. Brown-rot destroys carbohydrates leaving behind a brown-color lignin and making the wood appear flaky from cubicle checks. Dry-rot is one kind of brown-rot with an extensive network of hyphae spreading into “dry” areas. White-rot attacks both cellulose and lignin, leaving the deteriorated wood with a stringy texture. Certain types of white-rot can be used to enhance pulping (Akhtar et al. 1993). Soft-rot is a different kind of fungi that attacks from the surface of the wood. The unique microscopic feature of soft-rot is its attack along the microfibril in the cell wall layers. Because of this, the orientation of soft-rot cavities can be used to study MFA (Brändström et al. 2002).

234

C.L. Huang

Fig. 11.6 Examples of checks and splits in weathered utility poles, note the checks and splits are not developed in the incised and treated portion near the ground line

Wood-decaying fungi, (especially soft-rot), need a humid condition to grow. Consequently, outdoor wood structures and exterior wood applications are most vulnerable. However, wood wetted by condensation, whether from poorly ventilated attics or crawl spaces, leaky roofs, ground contact, or moisture from broken pipes, make these areas most susceptible to rot. Wood decay can cause significant strength loss in the infected structural member. Replacements from rot are expensive; if left unnoticed, the decay can progress to a point in the structure such that it is beyond repair or more seriously causes the loss of lives or property. If the in-service wood product cannot be kept dry, preservative treatments are ways to extend the service life of the product. Heartwood of softwood species has a natural resistance to bio-deterioration: untreated heartwood lasts about 5–10 years and 10–20 years in warm and cool climate regions respectively (Wang and DeGroot 1996). Due to the aspiration (cell wall pit closure that block the flow of vapor/liquid) and encrustation of pit membrane by extractives, heartwood is refractory for preservative treatment. Treated sapwood lasts much longer than heartwood. The permeability of dried earlywood is lower than the latewood due to high percentage of aspiration in earlywood areas. The amounts of ray tracheids and the collapse of ray parenchyma are closely related to the radial flow of vapor/liquid in treating round wood or flat sawn lumber. Some preservatives are effective fungicides but are ineffective on insects. Termites not only excavate the wood but also feed on the wood using enzymes from the protozoan symbionts in their stomachs. Large colonies of termites can cause serious damage to wooden structures. Carpenter ants only tunnel the wood for nesting and feeding on dead insects or the honeydew excreted by aphids in nearby woods.

11

Industry Prospective of Delamination in Wood and Wood Products

235

Powderpost-beetles lay eggs in large vessel elements, so they infest hardwood species with large vessels like oak, hickory, ash, and mahogany. Delamination, checks, and cracks are normal wear-and-tear of wood products, but they provide openings and microenvironments for wood-deteriorating organisms to enter, grow, and lay eggs. Wood preservatives are grouped by three types of applications: oil-born, waterborne, and organic. The oil-borne preservatives are used for industrial applications; for instance, railroad ties, utility poles, and bridge timbers. The traditional waterborne, copper-based preservatives are effective for non-industrial applications; but the associated heavy metals used in the preservative formulations are becoming safety and environmental concerns. Organic fungicides and insecticides used in agriculture, as well as nano-silver biocides, are being considered as woodpreservatives; however, the degradation of the organic compounds as well as their cost are the major hurdles for their applications. Wood preservatives are commonly applied to wood, in an industrial process, through a vacuum/pressure treatment in a large cylindrical pressure vessel. The treatibility of wood is species- and agedependent. For instance, Douglas-fir is difficult to treat, while plantation loblolly pine (which has little heartwood), is easy to treat. Incising, beating, vibrating, or rupturing the aspirated pit and the parenchyma in the rays by microwave heating may increase the permeability of wood for easy treating. Treating with bacteriainoculated water to create opening in the aspirated pits can improve the permeability, but it is most likely unfeasible for commercial applications. The waterproof adhesives used in EWP products make it durable, but EWP products for outdoor applications also need to be treated with preservatives. Highly densified EWP such as OSB and LSL will somewhat relax or spring back at the end of a manufacturing press cycle. The OSB siding panel-swelling related lawsuits, after the 1993 hurricane Andrew in Florida, raised concerns on the durability of OSB. When exposed to high humidities in service conditions, such as in poorly ventilated attics, edge-swelling from roof-sheathing panels may telegraph to roof shingles. Spring back of highly-densified EWP can be minimized in steam injection press. This is because lignin softens at the glass transition temperature (Tg), and this Tg is considerably lower at high MC. This means that the higher the moisture content, the lower the Tg of lignin becomes. The spatial and temporal profiles of the temperature in a regular press, lag behind the high Tg of the low-MC strands. While in a steam-injection press, the Tg stays within reach of the press temperature so lignin plasticizing is possible to prevent spring back. However, the cost of the steam-injection press is not cost-effective for modern OSB manufacturing, high throughput. Edge sealing is one way to delay the edge swelling of OSB in-service. Uneven shrinking/swelling, as well as creep, may cause buckling, thereby creating gaps at truss connections or delaminations between the joist and the floor sheathing. NDE methods are very useful in detecting the initiation of delaminations and monitoring the progress of propagations of the delamination within a wooden structure frame. Durability of wood products may not be covered by building codes, yet it can be a very important issue for the consumer and should be considered by the industry when developing a new product.

236

C.L. Huang

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 highquality 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 . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Why is Drying Slow and Collapse High in Eucalypts . . . . . . . 12.3 Causes of Surface and Internal Checking in Eucalypts . . . . . . 12.4 Eucalypts Species Prone to Collapse . . . . . . . . . . . . . . . 12.5 Collapse Recovery and Check Closure . . . . . . . . . . . . . . 12.6 Dimensional Analysis and Predictions from Heat-Transfer Theory 12.7 Effect of Reconditioning on Internal Checking . . . . . . . . . . 12.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

237 238 241 246 248 250 252 252 253

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: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_12, 

237

238

P. Blakemore

12.2 Why is Drying Slow and Collapse High in Eucalypts Many eucalypt species are well known for their slow drying and propensity for collapse (Campbell and Hartley 1984). Indeed, Kauman (1964) notes that both of these features were clearly recognised in the eucalypts as far back as 1826. Collapse and slow drying are intrinsically related. The most accepted theory for explaining collapse is the liquid tension theory. The first person to propose this theory was Tiemann (1915). However, Kauman (1964) provides the most comprehensive development of the theory, and an excellent detailed discussion of collapse in general. A more recent and concise summary is provided by Chafe et al. (1992). The theory is essentially concerned with the removal of liquid water from the lumens of saturated cells. As liquid water is removed from wood it is replaced by a mixture of water vapour and air. Where a saturated cell neighbours one filled with water vapour and air, a meniscus forms in the interstitial gaps or capillaries in the pits. Surface tension and capillary forces in these capillaries create a hydrostatic or liquid tension throughout the water in the saturated lumen. If these liquid-tension stresses are strong enough to overcome the strength properties of the cell wall, the cell wall will buckle or collapse, and the cell will flatten. There are two fundamental physics equations relating to the properties of the largest meniscus involved in the system, which allow the liquid-tension stresses that are developed to be approximated. The first is Laplace’s equation, which relates the total liquid tension to the radii of the curved liquid surface.  P=σ

1 1 + r1 r2

 (12.1)

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:  P=

ρRT M



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)

loge

po p

(12.2)

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 Radius of meniscus (nm)

Liquid-tension force (MPa)

Relative vapour pressure (p/po ) (%)

15 60 150 1500

9.70667 2.42667 0.97067 0.09707

93.1 98.2 99.3 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 materials, 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 extractives 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 collapse 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 assumptions. 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 determinate of whether or not collapse will occur is the strength of the cell walls. Unfortunately, the mechanical properties of individual cell walls, or more importantly, the mechanical properties of the various secondary layers that make up the cell wall, are almost impossible to measure directly. Hence, as a first approximation, 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 example, 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 predicting 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 layers 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 educated 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 overcome 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 displacement 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 displacement 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 temperatures 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.

12.3 Causes of Surface and Internal Checking in Eucalypts All forms of checking in timber, both surface and internal, are due to differential shrinkages that result in internal drying stresses. This variation in shrinkage is due to both variations in cell wall properties, and the timing and location of the shrinkage as drying progresses. In the part of the board that is trying to shrink more, some of that shrinkage is being resisted or restrained by the parts of the board around it that are shrinking less or not at all. Hence, a tension stress is developed in the higher shrinking area, and a compressive stress in the lower shrinking area. The magnitudes of the drying stresses that are developed in this way depend on both the shrinkage potentials of the different areas of wood and the stiffness properties of the wood. An important dynamic with regard to the stiffness properties is that this property increases greatly as moisture is removed from the cell walls below the local fibre saturation point. When the local tension stresses are greater than the local tensile strength limits, checking occurs. The shrinkage that occurs in a board can be either caused by normal shrinkage, due to the loss of moisture from the cell walls below the fibre saturation point, or

242

P. Blakemore

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 arrangements 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 shrinkage 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 quartersawn 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 gradients 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 combination 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 weakness 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 eucalypts 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 gradients 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 progression 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 conventional 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

Area (Radial × Tangential) Shrinkage (%)

20 18 16 14 12 10 8 6 4 2 0

0

10

20

30

40

50

60

70 80 MC (%)

90

100 110 120 130 140

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, collapse 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 intraring 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,

Fig. 12.6 Diagram of how tension set in the outer surface layers can lead to a sunken face in the wide faces after stress reversal has occurred

246

P. Blakemore

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.

12.4 Eucalypts Species Prone to Collapse 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, measurements 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 shrinkage 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 drying, 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 spurious positive collapse recovery values in low collapse specimens (e.g. Blakemore and Langrish 2007). Nevertheless, this general technique has been used repeatedly 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

20

900

18

800

16

700

14

600

12

500

10

400

8

300

6

200

4

100

2

0

0

Recovered Collapse (%)

1000

E. regnans (35 y.o.) E. regnans E. pilularis E. delegatensis E. nitens E. dalrympleana E. rubida E. obliqua E. rubida E. radiata E. robertsonii E. lindleyana E. grandis E. dives E. viminalis E. fastigata E. globulus E. stuartiana E. ovata E. viminalis E. jacksoni E. australiana E. ovata E. grandis E. obliqua E. dives E. mannifera E. baxteri E. dunnii E. haemastoma E. globoidea E. rossii E. saligna E. marginata E. globulus E. calophylla E. cypellocarpa E. sieberi E. globulus E. papuana E. wilkinsoniana E. diversicolor E. muelleriana E. macrorrhyncha E. pilularis E. campanulata E. botryoides E. camaldulensis E. consideniana E. robusta E. bridgesiana E. stjohnii E. goniocalyx E. carnea E. maidenii E. cladocalyx E. astringens E. tereticornis E. maculata E. resinifera E. microcorys E. cloeziana E. leucoxylon E. longifolia E. gomphocephala E. tetrodonta E. polyanthemos E. paniculata E. bosistoana E. sideroxylon E. molucanna E. melliodora E. largiflorens E. albens E. decorticans E. tessellaris E. wandoo E. cambageana E. spenceriana

Basic Density (kg m–3)

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 previous 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) reconfirmed 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 content 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

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).

12.5 Collapse Recovery and Check Closure 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 number 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 collapse 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 questionable. 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

15%, collapse recovery started to be reduced significantly. An optimal mean moisture content of about 18% was identified, although there was little difference in recovery in the range of 15–25% moisture content. One of the most important considerations in this is the core moisture content. It is important that the core moisture content be below about 30% moisture content. Above this level collapse recovery in the core of the boards was incomplete and in particular the closure of internal checks incomplete. This is shown visually in Fig. 12.9 below. It was also thought that in high moisture content cores, the collapse simply was not recovered, rather than it recovering and then subsequently re-collapsing. Historically, a darkening of the core wood, and severe inter-ring checking has been observed when reconditioning has been undertaken with a high core moisture content. This was not observed in these studies suggesting that for this to occur the core temperature must be much lower. Blakemore and Langrish (2008b) and Blakemore (2008) also found that, while acknowledging the importance of mean moisture content, the uptake or movement of moisture was not thought to be critical for collapse recovery. In this sense then, it was identified that, provided the board moisture content was in the right range, the length of steaming only needed to be long enough to raise the temperature of the core to close to 100◦ C (See also Fig. 12.10). 20% 32.8%

35.2%

BEFORE RECONDITIONING

AFTER RECONDITIONING

15% 16.9%

14.2%

AFTER FINAL DRYING TO 12%

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

9 mm bottom

90

1020

Chamber

80

1010

Side Loadcell

70

Datalogger Temp

1000

Width

60

990

Mass

50

980

40

970

30

960

20

950

10

940

0

Mass (g)

Temperature (°C) Width Swelling (× 0.1 = %)

middle

930 0

1

2

3

4 Time (hrs)

5

6

7

8

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)

12.6 Dimensional Analysis and Predictions from Heat-Transfer Theory 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): kτ ατ (12.2) Fo = 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 = heat capacity of dry wood (kJ kg−1 K−1 ) θ = temperature (K) c=

(Co + 0.01 · MC · Cw ) (1 + 0.01 · MC) 

+MC −0.06191 + 2.36 × 10−4 · θ − 1.3310−4 · MC

(12.4)

where: c = heat capacity of wet wood (kJ kg−1 K−1 ) Cw = heat capacity of water (kJ kg−1 K−1 ) MC = percentage moisture content (%) An approximate value for k was obtained from Oliver (1991): 

k = (7.23ρ + 13.6 · MC) × 10−4 + 0.086 × 0.277778

(12.5)

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 approximately 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 parallel 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 material 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 purpose, 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

12.7 Effect of Reconditioning on Internal Checking In the studies undertaken by Blakemore and Langrish (2008b) and Blakemore (2008), providing that the guidelines with regard to mean and core moisture content was meet, most of intra-ring checking that was present before steam reconditioning was not visible after reconditioning (Fig. 12.9). Conventional industry wisdom is that such closed checks pose a problem, in that if exposed when machining or moulding is undertaken, they will result in a feathering affect on the surface of the product. This may not be that significant problem for products such as quartersawn flooring, but is a problem for backsawn products cabinetry such as high value kitchen cupboard doors. However, how much of this effect that is due to completely closed checks, or incompletely closed checks, is not clear.

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 density 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 control. 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, recovery 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 engineering 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, microfibril 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 13.2

Introduction . . . . . . . . . . . . . . . . . . . . . . Ultrasonic Methods . . . . . . . . . . . . . . . . . . 13.2.1 Materials . . . . . . . . . . . . . . . . . . . . 13.2.2 Ultrasonic Methods . . . . . . . . . . . . . . . 13.3 Tension Wood Detection with Ultrasonic Velocity Method 13.4 Acoustic Imaging with Stress Wave Technique . . . . . . 13.4.1 Imaging with Radial Stress Waves Velocities . . . 13.4.2 Imaging with Tangential Stress Waves Velocities . 13.5 Summary . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

255 257 257 257 260 261 262 262 266 266

13.1 Introduction The last two decades of the 20th century have been characterized by an extraordinary 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 properties 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: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_13, 

255

256

V. Bucur

types of waves (electromagnetic or mechanical ). Given the hierarchical structure of wood, it is obvious that one should seek multi-scale characterization tools. The ultrasonic waves have a resolution between 10−3 and 10−2 m, or more, depending on frequency (> 20 kHz) used for wood inspection. The ultrasonic waves propagated in wood are related to its elastic properties. The most common waves for ultrasonic imaging are the longitudinal waves. The resolution of ultrasonic imaging techniques is very much limited by the frequency, the wave length and by the size of the transducers. The main benefits of using acoustic techniques for wood quality assessment are that they are non-invasive and safe at a relatively low energy levels. The ultrasonic techniques based on the measurements of the velocity of propagation of a pulse, are used for sorting logs, for lumber quality and veneer quality assessment, or for pulp and paper quality. Log segregation through velocity measurements, using the time of flight method, is a current technique for detecting higher possible proportion of high-grade lumber. In Switzerland spruce logs of very high quality for musical instruments are detected in this way. Sorting logs and stems for structural uses is a current technique in US, Australia, New Zealand and Brazil. Assuming a price difference between structural and ordinary lumber of 200 NZ$ /m3 , it was reported an increase of about 1.8 million NZ$ for a mill processing 300,000 m3 of logs per year (Wang et al. 2007). The detection and the location of reaction wood (compression or tension wood) in trees and logs is a major interest for wood industry in Australia and all over the world. The S2 layer of compression wood tracheids often shows helical checks, which can introduce radial delminations (Boyd 1972). Tension wood has an important thick gelatinous layer (or G – layer) in the cell wall. The microfibrils in the G layer are much thicker (Müller et al. 2006) than in normal wood and are laid at very small angles at the fiber axis (Fengel and Wegener 1984). In G layer radial delaminations have been observed in Eucalyptus spp. by Chafe (1977) as local disruptions to the microfibril orientation (Fig. 13.1). Chafe (1977) noted “ . . . dislocations were evident as longitudinal striations running approximately parallel to the major extinction position of the cell wall and close to parallel with the cell axis”. Furthermore he pointed out “Radial striations were abundant in the wall of many earlywood fibers and were detected to a lesser extent in latewood cells”. “It would seem from these observations that there may exists an association between the presence of high growth stress in living trees and the occurrence of radially oriented dislocations in their fiber walls”. The tension wood is weak because of morphological difference between the lignified secondary wall S2 and the unlignified G –layer (Donaldson 2001). This structural weakness as well as the disruption of microfibril orientation in tension wood can be put in evidence with ultrasonic techniques – acoustic tomography and ultrasonic velocity method. The main advantage of ultrasonic techniques is related to the fact that wood behavior can be inspected in all three anisotropic directions (L, R and T) on the same specimen.

13

Acoustic Tomography for Tension Wood Detection in Eucalypts

257

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 Ultrasonic Methods 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.

13.2.2 Ultrasonic Methods 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 11

9

11B

8 7

12 12B 1

6

5

2 3 4

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 longitudinal 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 transducers 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 frequency 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 transducers 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 assuming 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 measurements (for 16 transducers, 120 independent acquisitions). The time uncertainty is 1 μm and the calculated limit of the spatial resolution is 25 mm.

13.3 Tension Wood Detection with Ultrasonic Velocity Method Figure 13.6 shows the variation of ultrasonic velocity VLL on a disc following four rays. To avoid cluttering, only these four rays were represented on the previous figure. The ray 1 corresponds to the wider side of the disc, where the tension wood is located. The ray 6 corresponds to the opposite wood, while the rays 4 and 9 correspond to lateral wood. It can be clearly seen that VLL values are the highest along the ray 1 (average 3847 m/s), the lowest along the ray 6 (average 3187 m/s), and in between, along the rays 4 and 9, corresponding to the lateral wood (average 3544 m/s and 3435 m/s respectively). These results are in line with those for beech samples published by Bucur et al. (1991). In poplar tension wood Coutand et al. (2004) found higher Young’s modulus in L direction (EL ), than in normal wood. The authors argued that this was due to the smaller microfibril angle in the S2 and G layers. The EL values are positively associated with VLL . The highest VLL values are related with the same smaller microfibril angle. Analyzing the previous data, it can be stated that in Eucaluptus spp. the velocity VLL has potential in differentiating the tension wood from normal wood.

13

Acoustic Tomography for Tension Wood Detection in Eucalypts

261

Longitudinal velocity (m/s)

5000 Ray 1 Ray 6

Ray 4 Ray 9

4500 4000 3500 3000 2500 0

50 100 150 Distance from cambium (mm)

200

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)

13.4 Acoustic Imaging with Stress Wave Technique Acoustic tomography allows reconstruction of the distribution of the velocity of ultrasonic waves as it propagates within the investigating section of a tree. Acoustic tomography is a fast, reliable, safe and inexpensive testing procedure, and has become increasingly popular for in situ examination of trees, logs, boards, etc as noted in some reference articles published since 1990 (Tomikawa et al. 1990; Biagi et al. 1994; Socco et al. 2000, 2004a, b; Lawday and Hodges 2000; Andrews 2003; Martinis 2002; Martinis et al. 2004; Maurer et al. 2006; Divos and Divos 2005; Bucur 2005; Gun et al. 2005; Attia 2007; Lee et al. 2007; Sandoz and Benoit 2007; Yang et al. 2007a, b; Lin et al. 2008; Schubert et al. 2009). The described techniques can provide two or three dimensional spatial location of the internal structure of the tree under test. The theory behind the acoustic tomography which utilizes the scattering of acoustic plane waves is largely discussed by Zhang and Lu (1996) or by Berryman (2000). The investigation of the relationship between wood properties, products quality and growth stress in trees from plantations has a long research history (Jacobs 1938; Boyd 1950, 1972; Chafe 1977; Polge and Thiercelin 1979; Ferrand 1981; Kubler 1987; Fengel and Wegener 1984; Archer 1986; Timell 1986; Zobel and van Buijten 1989; Bamber 2001; Blakemore 2008). High growth stress in trees are related to the presence of reaction wood (compression or tension wood). The delaminations in tension wood are similar to radial checks in compression wood. Chafe (1977) noted that the radial dislocations represent “a kind of localized failure of the cell wall, it seems possible that, if sufficiently abundant, they could affect the strength of converted wood products”. Today, acoustic imaging with the stress wave technique is able to put in evidence the radial dislocations observed by Chafe in 1977. This will be demonstrated in the next sections.

262

V. Bucur

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)

13.4.1 Imaging with Radial Stress Waves Velocities Figure 13.7 shows the velocity map on the eccentric and concentric discs with 8 and 16 probes obtained with radial stress velocity (measured in R direction). Divos and Divos (2005) reported that the resolution of the system with 8 transducers is 2.5% (velocity difference 100 m/s). The color of the diagram gradually changed from green to red as the proportion of tension wood increases. The red color indicates lower ultrasonic velocity while the green color indicates higher ultrasonic velocity. The range of ultrasonic values was between 1630 m/s (green) and 1542 m/s (red). The velocity map for the eccentric disc show regions of considerable lower velocities on the wider side of the disc, where tension wood is likely to be present, than on the opposite side. In comparison, the velocity map for the concentric disc was far more uniform indicating greater homogeneity of wood inspected transversal structure. As expected, the velocity maps obtained using 16 transducers appeared to give a better representation of discs structure, than for those using only 8 transducers.

13.4.2 Imaging with Tangential Stress Waves Velocities 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 eccentric 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. 13.8 (continued)

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

1

265

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)

a

b

c

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 secondary 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 corresponding 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 reconstruction. 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 direction 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 different 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 velocities 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 technique 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 contraintes 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 resolution 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 14.2

Introduction . . . . . . . . . . Material and Methods . . . . . 14.2.1 Moisture Measurements . 14.2.2 Warp Measurements . . 14.3 Material Data . . . . . . . . . 14.3.1 Moisture Model . . . . . 14.3.2 Mechanical Model . . . 14.4 Results and Discussion . . . . . 14.4.1 Moisture . . . . . . . . 14.4.2 Deformations and Stresses 14.5 Summary . . . . . . . . . . . References . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

269 271 271 272 274 274 276 277 277 278 283 284

14.1 Introduction This chapter focuses on moisture-induced deformations in three-layered crosslaminated timber with a symmetrical build-up, where the fibre direction of the middle layer is oriented perpendicular to that of the outer layers. Dimensional stability, 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 approximately 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: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_14, 

269

270

T. Gereke et al.

Fig. 14.1 The types of warp deformation

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 simulation, 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 literature 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)

where h [m s−1 ] symbolizes a mass transfer coefficient. Diffusion is expressed in terms of concentration of water relative to the dry volume of wood. Thus, the effect of swelling is omitted from the mass balance. The total strain rate is assumed to be the sum of elastic strain rate ε˙ el , moistureinduced swelling ε˙ ω , and mechano-sorptive deformation ε˙ ωσ :  ˙ + αω ε˙ = ε˙ el + ε˙ ω + ε˙ ωσ = Sσ˙ + Sσ ˙ + mσ ω ˙

(14.2)

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

⎢ ⎢ −ν E−1 E−1 −ν E−1 RT T ⎢ RL L R ⎢ ⎢ ⎢ −νTL EL−1 −νTR ER−1 ET−1 S=⎢ ⎢ ⎢ 0 0 0 ⎢ ⎢ ⎢ 0 0 0 ⎣ 0 0 0

0

0

0

0

0

⎤

⎥ 0 ⎥ ⎥ ⎥ ⎥ 0 0 0 ⎥ ⎥ ⎥ −1 GLR 0 0 ⎥ ⎥ ⎥ ⎥ 0 G−1 0 LT ⎦ 0 0 G−1 RT

(14.3)

Due to moisture dependency of the moduli, the rate of the compliance matrix S˙ 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 mechano-sorptive strain rate follows an expression proposed by Takemura (1967) and Leicester (1971), see also Ranta-Maunus (1990). The mechano-sorptive property matrix is defined as (Ranta-Maunus 1990; Santaoja et al. 1991; Ormarsson 1999). ⎡ ⎤ mL −μRL mR −μTL mT 0 0 0 ⎢ ⎥ ⎢ −μLR mL mR −μTR mT 0 0 0 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ −μLT mL −μRT mR mT 0 0 0 ⎥ ⎢ ⎥ (14.5) m=⎢ ⎥. ⎢ 0 0 0 mLR 0 0 ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ 0 0 0 0 mLT 0 ⎥ ⎣ ⎦ 0

0

0

0

0 mRT

14.2 Material and Methods 14.2.1 Moisture Measurements 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.

14.2.2 Warp Measurements 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 influence 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 orientations 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 LR =

2aOL , atot

(14.6)

Table 14.1 Panel characteristics Group

ID

Annual ring AR0 orientation AR45 AR90 Layer ratio LR37 LR52 LR57 aA

Production a αBL = aTL (mm)b aML (mm)b atot (mm)b θ (◦ ) LR (–) A A A B B A

10 10 10 7 7 10

10 10 10 24 13 15

30 30 30 38 27 35

0 45 90 90 90 90

0.67 0.67 0.67 0.37 0.52 0.57

– laboratory, 1C polyurethane, board width 100 mm; B – industry, urea resin, board width 26±1 mm b a – thickness, BL – bottom layer, TL – top layer, ML – middle layer, tot – total

14

The Hygroscopic Warping of Cross-Laminated Timber

273

Fig. 14.2 Panel coordinate system and definition of ring angle θ

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 conditioned 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 specimens 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 rubber 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. 14.3 Test setup of the hygroscopic warping experiments

274

T. Gereke et al.

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 cupxz =

 1  A,D,G u¯ z − u¯ B,E,H + u¯ C,F,J z z 2

(14.7)

and the cup deformation in the yz-plane as cupyz =

 1  A,B,C u¯ z − u¯ D,E,F + u¯ G,H,J . z z 2

(14.8)

i,j,k

The notation u¯ z 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.

14.3 Material Data 14.3.1 Moisture Model A one-dimensional diffusion model with a total of 340 eight-node volume elements along the thickness was applied (100 elements in each 10 mm thick wood layer and 20 elements in each 0.1 mm thick adhesive layer). The material parameter necessary for modelling diffusion according to Fick’s law is the diffusion coefficient D [m2 s−1 ]. It was implemented into the material model by DT (ω) = 8.0 · 10−11 e4ω , t ≥ t∗

 t D∗T (ω, t) = DT (ω) (1 − κ) ∗ + κ t < t∗ t

(14.9) (14.10)

14

The Hygroscopic Warping of Cross-Laminated Timber

275

The subscript T refers to diffusion in the transverse direction where identical properties in radial and tangential direction were assumed. Equation (14.9) is an expression found by Toratti (1992) but the amount of dependency on moisture content (MC) was increased by Hanhijärvi (1995). The diffusion coefficient of wood strongly depends on the MC. With increasing MC the resistance to diffusion decreases and, thus, D increases. If t < t∗ , the diffusion coefficient increases linearly from κDT (ω) at the initial state (t=0) to DT (ω) at t = t∗ . This reduction of DT in the beginning of the analysis refers to the effect that water diffuses into wood as water vapour and is then absorbed by the cell walls. During this process the diffusion slows down. Time-dependent sorption, i.e. the coupling between the two phases of water diffusion in wood, plays a significant role (Frandsen et al. 2007). During unsteady-state diffusion, DT is assumed to depend not only on MC but also on time. The barrier between unsteady-state and steady-state diffusion is referred to as t∗ . It has been determined from continuous weighing of the samples. A specific concentration cˆ [kg m−3 ] was calculated by cˆ =

m(t) − m0 , V0

(14.11)

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 ∂ cˆ = 0. ∂t

(14.12)

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: Dadh (ω) = A1 · ω−A2 + A3

(14.13)

In an optimisation process, the error between the real physical (experimental) data and the numerical approximation was minimised. The shape factors were determined to A1 = 9.17 · 10−13 m2 s−1 , A2 = 0.51 [−] and A3 = −2.39 · 10−12 m2 s−1 . These parameters are an acceptable compromise, reflecting the measured moisture contents in all layers at all times considered. The dependency of the concentration c on the relative humidity ϕ under isothermal conditions of T=20◦ C 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.

and f6 = −1.09 10−3 . Consequentially, the concentrations corresponding to the conditions in ambient air and the initial conditions are cinitial = ca,top = 5.60 kg m−3 relating to ϕ=65% and ca,bottom = 13.01 kg m−3 relating to ϕ=100% (Eq. (14.1)). The mass transfer coefficient h [m s−1 ] of the boundary layer has been proposed by Hanhijärvi (1995) as h (ω) = 3.2 · 10−8 e4ω .

(14.15)

14.3.2 Mechanical Model Eight-node volume elements were used for the modelling. The fineness of the element mesh was chosen in order to obtain accurate results. The mesh density (number of elements) (14.16) NOE = Nx · Ny · (3Nz + 6) 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 moduli G, the Poisson’s ratios ν and moisture content (Eq. (14.13)) according to Neuhaus (1981), 0% ≤ ω ≤ 28% Parameter

a0

a1

a2

a3

EL (MPa) ER (MPa) ET (MPa) GLR (MPa) GLT (MPa) GRT (MPa) ν LR (–) ν LT (–) ν RT (–)

12792 1000 506 763 881 61 0.046 0.021 0.153

1522 361 500 593 139 –107 0.136 0.257 1.075

–90073 –20917 –13499 –19861 –13925 –617 –0.456 –1.435 3.980

188504 46665 29733 47671 27691 1725 –0.214 2.190 –19.1

14

The Hygroscopic Warping of Cross-Laminated Timber

277

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 directions 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 Results and Discussion 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

Fig. 14.5 Measured and calculated moisture profiles

278

T. Gereke et al.

Fig. 14.6 History plots of simulated moisture contents in the individual layers compared to the measurements, the studied material points of the simulations were located in the middle of each layer

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.

14.4.2 Deformations and Stresses 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 direction 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 properties 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) ID

n

ρ0 (kg m−3 )

cupxz (mm)

cupyz (mm)

AR0 AR45 AR90 LR37 LR52 LR57

6 6 6 6 6 3

411 407 431 459 461 451

0.06 (0.05) 0.06 (0.06) 0.10 (0.15) 0.11 (0.19) 0.07 (0.09) 0.04 (0.27)

0.42 (0.05) 0.22 (0.08) 0.27 (0.08) 0.08 (0.02) 0.14 (0.04) 0.12 (0.08)

14

The Hygroscopic Warping of Cross-Laminated Timber

279

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 simulated 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 maximum 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 modification 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

The Hygroscopic Warping of Cross-Laminated Timber

Difference to reference (%)

14

50

cupxz

40

cupyz

281

30 20 10 all OL ML

all OL ML

0

all OL ML

all OL ML

αL

αR

all OL ML

all OL ML

αL

αR

–10 –20 –30

Difference to reference (%)

EL

ER

50

cupxz

40

cupyz

m L mR

ER* αR*

m L mR

ER* αR*

30 20 10 0 all OL ML

all OL ML

EL

ER

–10 –20 –30

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

of EL , ER , α L , α R , mL and mR in the individual layers and in the complete panel. The cupping cupyz is significantly influenced by a change of the longitudinal parameters EL and α L in the middle layer and of the radial parameters ER and α R in the outer layers. The cup deformation cupxz is significantly influenced only by a change of the parameters in the outer layers.

14

The Hygroscopic Warping of Cross-Laminated Timber

283

The mechano-sorptive strain was found to be very important for the cup deformation 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 cupping 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 coefficients 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 contrast, 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 coefficient 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 significantly 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 moduli 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 stiffness 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 layers 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 determination 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

Santaoja K, Leino T, Ranta-Maunus A, Hanhijärvi A (1991) Mechano-sorptive structural analysis of wood by the ABAQUS finite element program. Research Notes 1276, Technical Research Center of Finland, Espoo Sell J (1997) Eigenschaften und Kenngrössen von Holzarten. Baufachverlag AG, Dietikon Siau JF (1995) Wood: Influence of moisture on physical properties. Department of Wood Science and Forest Products, Virginia Polytechnic Institute and State University, Blacksburg, USA Takemura T (1967) Plastic properties of wood in relation to the non equilibrium states of moisture content. Part II. Mokuzai Gakkaishi 13(3):77–81 Teischinger A, Vanek M (1987) Eignung verschiedener Auftrennmethoden zur Bestimmung eines Feuchtegradienten im Holz nach der Darrmethode. Holzforschung und Holzverwertung 39:5–8 Toratti T (1992) Creep of timber beams in a variable environment. Helsinki University of Technology, Espoo Vanek M, Teischinger A (1989) Diffusionskoeffizienten und Diffusionswiderstandszahlen von verschiedenen Holzarten. Holzforschung und Holzverwertung 1:3–6

Chapter 15

Acoustic Emission Activity Induced by Delamination and Fracture of Wood Structure Voichita Bucur

Contents 15.1 15.2 15.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . Parameters of the Acoustic Emission Signals . . . . . . . . . . . . . Parameter-Based Acoustic Emission Techniques . . . . . . . . . . . 15.3.1 Species Effect . . . . . . . . . . . . . . . . . . . . . . . 15.3.2 Grain Angle Effect . . . . . . . . . . . . . . . . . . . . . 15.3.3 Annual Ring Structure . . . . . . . . . . . . . . . . . . . 15.3.4 Tension Wood . . . . . . . . . . . . . . . . . . . . . . . 15.3.5 Moisture Content . . . . . . . . . . . . . . . . . . . . . . 15.3.6 The Reused Old Wood . . . . . . . . . . . . . . . . . . . 15.4 Some Aspects Related to the Energy of the Acoustic Emission Signals 15.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

287 288 290 290 292 294 297 297 299 301 302 304

15.1 Introduction The damage evolution related to fracture phenomena in wood can be expressed at several internal length scales such as atomic, micro, meso and macro scales. Damage initiates on the atomic scale and reaches relevance for larger scales while it propagates, leading to failure when reaching the macro scale. In this section we focus our attention on aspects related to the micro and macro scales. The dominating structure on micro scale is composed from the anatomic constituents while on meso scale the dominating structure is composed from the annual rings. Acoustic emission technique can be used to study the fracture damage mechanisms on micro and meso scales. Acoustic emission analysis requires the analysis of mechanical data and the acoustic emission rate, the localization of the acoustic emission source, the V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_15, 

287

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) generate 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.

15.2 Parameters of the Acoustic Emission Signals Acoustic emission technique provides the opportunity for continuous monitoring of specimens under test, in real time. The literature for nondestructive evaluation and testing of materials is very abundant in describing the acoustic emission techniques and the parameters of the acoustic emission signals (Stephens and Leventhal 1974; Sachse and Kim 1987; Scott IG 1991; Ando 1993; Green 2004; Minozzi et al. 2003; Landis 2008; ASNT 2005; Muravin 2009 and journals such as JASA, Ultrasonics, J. Acoustic Emissioin, Materials Evaluation, etc). The acoustic emission activity and signal strength release by the materials depend on the nature, microfracture characteristics and processes involved. In this chapter several basic aspects are summarized for better understanding of the requirements for specific aspects related to wood and wood-based composites. The approaches in analysing AE signals are: parameterbased techniques and signal-based quantitative techniques. The parameter – based techniques are the most popular and are able to analyse large acoustic events, to store data and to facilitate data fast visualisation using equipment of reasonable cost. The signal-based quantitative techniques require high costs and long time for fine signal processing. These techniques are recommended in advanced fracture analysis (Grosse and Ohtsu 2008). Figure 15.1 illustrates a typical acoustic event. The conventional classical parameters as defined by ISO 12716 are: hit, counting/ring-down count/emission count, amplitude, duration, rise time, energy, average frequency, initial frequency, reverberation frequency, frequency centroid, peak frequency, rise time divided by amplitude called RA value, RMS (root mean squared values), the threshold voltage. A combination of some of them in an empirical relationship inspired from seismic signals analysis gives N(V) = kV m where k and m are calculated from data generated at the proportional limit on load – COD curve. The m coefficient is effective for acoustic emission studies and confirms the progressive deterioration of the materials. Other parameters can be calculated such as for example the acoustic emission

15

Acoustic Emission Activity Induced by Delamination and Fracture

289

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 emission characteristics, propagation effects and transducers excitation effects. The use of acoustic emission and Weibull statistics to characterize the specimen is an effective 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 transducer in contact or not with the specimens (Murphy et al. 1990; Green 2004). Furthermore the signal is amplified and processed in real time with appropriate 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 analysis 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 conditions 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.

15.3 Parameter-Based Acoustic Emission Techniques 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, technological 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.

15.3.1 Species Effect The effect of species on acoustic emission activity and fracture in RL system, Mode I has been studied by Reiter et al. (2000, 2002). Cubic specimens have been used for splitting test. The horizontal force component of the splitting test FH , called simply horizontal force was correlated with acoustic emission activity and wood fracture. Figure 15.2 illustrates the variation of the number of AE counts versus

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

a

291

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 formation 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 formation 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) Brittleness (mm)

Species

AE counts (number)

B=

Spruce Alder Oak Ash

9567 1367 867 433

1.1 2.1 2.0 2.1

2 1 FH,max L kinit Gf

KIc MPa 0.49 0.67 0.83 1.16

√

m

Gf J/m2 337 244 348 551

Note: L is the ligament length, k init is the slop is proportional with the effective modulus of elasticity, 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 difference 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.

15.3.2 Grain Angle Effect 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 displacement (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 amplitude, 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 acoustic 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

a

b

c

d

293

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)

The literature is very scarce in data on acoustic activity on specimens subjected to static torsion and fatigue. The shear mode of rupture under torsional loading is one of the most complex possible modes. Chen et al. (2006) investigated the behavior of Red lauan (called also Philipine mahogany, Shorea spp.) and Sitka spruce under torsional fatigue experiments. Table 15.2 shows the complexity of the fracture modes developed under torsional loading. In static torsional testing microcrack initiation was observed through the acoustic activity prior to maximum loading. Acoustic emission activity (events number) increases as the grain angle increases from 45◦ to 90◦ . The red lauan produced more counts than Sitka spruce. Keiser effect before cracking, in fatigue testing was reported for both species. Specimens under fatigue testing produced more events than under static tests.

294

V. Bucur

Fig. 15.5 Amplitude, acoustic events number and relative load, up to the proportional limit, on notched specimens in tension. (a) for small angle (0◦ and 15◦ ) and (b) for other angles between 30◦ and 90◦ . Mokuzai Gakkaishi (1992, Figure 7)

Table 15.2 Fracture modes at different grain angle orientation of specimens under torsional loading (data from Chen et al. 2006) Fracture modes at different grain angle orientation under torsion Species Red lauan Sitka Spruce

0◦

45◦

90◦

II RL, III RT

I RL, I RT

II RL, III RT

II TL, III TR

I TL, I TR

II TL, III TR

and III TR; II TL, III TR and III RT and III RT; II RL, III RT and III TR

15.3.3 Annual Ring Structure Ansell (1982) was probably the first to demonstrate the influence of the earlywoodlatewood 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 densiflora) 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 japonica), 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 activity 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 studied 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 adjacent 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 consists predominantly of ruptured cell walls comparable to the rupture in RT system. Having reached half the specimen with the crack turns 90◦ anticlockwise propagating 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)

in case of pure TR configuration, caused by the debonding of adjacent fibres with hardly any cell wall damage”. The 45◦ specimen was used to put in evidence the relationship between structural damage and acoustic emission activity. The onset of acoustic emission was prior to the first in-situ microscopic visible crack. More that 95% of the acoustic emission events and visible crack propagation were observed during a plateau of load displacement curve. The acoustic emission plot exhibited peaks in the same time as the load deformation plot. Dill –Langer and Aicher (2000) suggested using these observations as support for further theoretical modeling of damage in wood.

15.3.4 Tension Wood The presence of reaction wood in general (compression or tension wood) in lumber submitted to drying induce quality degradation. Cunderlik et al. (1996) monitored the drying cracks in the tension and opposite wood by acoustic emission and SEM. The opposite wood generates higher numbers of acoustic events than tension wood (Fig. 15.8). Tension wood activity decreases with increase of gelationous fibres proportion in wood tissue. In opposite wood dominated the cracks of high acoustic activity in the middle lamella (intercell), while in tension wood the delamination of gelatinous layer (G-layer) from the secondary wall (S2-layer) is characterized by low acoustic activity.

15.3.5 Moisture Content Acoustic emission technique from the early 1980 was related to checking detection and wood drying (Kawamoto and Williams 2002). Acoustic emission signals during 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 crystalline segments of cellulose (Booker 1994). The checking occurs when the rate of

298

V. Bucur

Fig. 15.8 Acoustic emission count rate during drying of tension and opposite wood in beech (Cunderlik et al. 1996, Figure 1)

slips exceeded a critical value. The slip lines generate slow AE events. Brittle microcracks 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

Pattern classification of acoustic emission signals during wood drying by principal component analysis and artificial neural network for oak (Quercus variabilis) was proposed by Kim et al. (2005). The acoustic emission parameters selected for this study were: peak amplitude, ring-down count, event duration, frequency, energy, rise time and peak amplitude/rise time. The sources of acoustic emission events during drying were due to water movement and to surface check. The cumulative AE hits versus drying time and the waveforms of AE signals corresponding to surface check and water movement are shown in Fig. 15.9. AE signal produced by water movement shows lower in peak amplitude, longer in rise time and lower in peak frequency, then the signals caused by surface check. 96% of the variability of AE signals was accounted in the principal components plane 12. Kim et al. (2005) noted that when the value of the first principal component is greater than 1, the number of AE events cause by water movement are higher then AE events caused by surface checking. The classification of AE signals with artificial neural network (ANN) was performed in two hypotheses. The first ANN classifier has six input nodes for six AE parameters. The second ANN classifier has only two inputs for two principal components. All classifiers have two output nodes, the surface check pattern and the water movement pattern. Eight nodes (hidden layer) were chosen for optimum pre processing of ANN classifier. The neurons between layers were activated with tangent sigmoid function. The recognition rate was used to evaluate the performance of AE and ANN classifiers (Table 15.3). Surface checking was recognised at 83% with ANN classifier and AE parameter inputs and by 85% with ANN classifier and principal components inputs. This technique seemed to be promising for improving wood drying technology.

15.3.6 The Reused Old Wood 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) Water Inputs

Pattern

Learning (number)

AE parameters

Water Check Total Water Check Total

458 66 − 471 44 −

Principal components

Check Validation (number)

Learning (number)

396 83 − 448 74 −

42 434 − 29 456 −

Recognition Validation (number) 104 417 − 52 426 −

Rate

%

91.6 86.8 89.2 94.2 91.2 92.7

72.9 83.4 81.3 89.6 85.3 87.4

is appropriated for examining the differences between new and old wood in shearing 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.

15.4 Some Aspects Related to the Energy of the Acoustic Emission Signals The acoustic emission energy, monitored with a high speed waveform acquisition 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 loadCMOD 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 waveform. From data plotted in Fig. 15.12 (which gives the variation of the measured

Fig. 15.11 Load (N), strain energy release (GIc ) versus fracture time (seconds) of a notched specimen of eastern hemlock (Tsuga Canadensis) (Landis and Whittaker 2001, Figure 2)

Fig. 15.12 The measured load and the cumulative acoustic emission energy release during fracture. (Landis and Whittaker 2001, Figure 4)

302

V. Bucur

Fig. 15.13 The linear relationship Between the acoustic emission energy rate and fracture energy release (Landis and Whittaker 2001, Figure 5)

load and cumulative acoustic emission energy release during fracture) it can be noted that: – an internal damage is produced before the growth of the crack – a constant rate at which the AE energy is released after the starting of crack growth – a constant rate of fracture energy release 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 during 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 Delamination and fracture phenomena in wood can be monitored non-destructively, continuously and in real time with acoustic emission technique. The conventional classical parameters of acoustic emission are: hit, counting/ring-down count/emission count, amplitude, duration, rise time, energy, average frequency, initial frequency, reverberation frequency, frequency centroid, peak frequency, rise time divided by amplitude called RA value, RMS (root mean squared values), the threshold voltage (ISO 12716). The approaches in analysing AE signals

15

Acoustic Emission Activity Induced by Delamination and Fracture

303

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)

are: parameter-based techniques and signal-based quantitative techniques. The parameter-based techniques are the most popular for wood material studies. Acoustic emission technique operates using resonant or wideband transducers in ultrasonic frequency range (100 kHz–1 MHz). The most appropriated technique for wood is with piezoelectric transducers between 100 and 200 kHz. Acoustic emission is highly sensitive to the initiation and growth of delamination in wood and has advantages over conventional ultrasonic and radiographic methods. Transverse failure is one of the most important damage mechanisms controlling the loss of stiffness in wood which may be lifetime limiting in for structural members. Factors such as species, grain angle orientation, annual ring structure, moisture content, tension wood, etc effects the acoustic emission activity related to crack propagation, delamination and fracture phenomena. Acoustic emission technique provides a sensitive approach for real time detection of cracking, and also an unique view into the micromechanics of crack initiation and growth of delamination. The damage processes in the material under test can be observed during the entire load history without any deterioration of the specimen. The final objective of monitoring acoustic emission phenomena in wood is to provide beneficial information to prevent deterioration during processing (drying, etc) and catastrophic failure of the material. For the future, it could be suggested the development of new signal-based procedures, for wood and wood based composites, with a more quantitative analysis of the acoustic emission signals based on a 3D localization of AE sources and the recordings obtained from a sensor network. Another research field to be developed is related to the acoustic emission activity produced by micro-cracks with a two or

304

V. Bucur

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 emission. 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 signalbased 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 during 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 mechanosorptive 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 NDEapplications. 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 Voichita Bucur and Saeed Kazemi-Najafi

Contents 16.1 16.2

Introduction . . . . . . . . . . . . . . . . . . . . . . Basic Aspects . . . . . . . . . . . . . . . . . . . . . 16.2.1 Waves Propagation Paths . . . . . . . . . . . . 16.2.2 Linear Ultrasonic Inspection Techniques . . . . . 16.2.3 Ultrasonic Transducers and Scanning Procedures . 16.2.4 Non-linear Ultrasonic Inspection Techniques . . . 16.3 Delamination Detection in Wood-Based Composite Panels 16.3.1 Through Transmission Technique . . . . . . . . 16.3.2 Plate Wave Technique . . . . . . . . . . . . . . 16.3.3 Industrial Applications of Non-contact Technique . 16.4 Summary . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

307 309 309 311 320 327 329 329 339 343 344 345

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: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_16, 

307

308

V. Bucur and S. Kazemi-Najafi

– particle based boards – oriented strandboard, particleboard, fibre board, – mineral bonded particleboard and fibreboard, – wood – plastic composites. Wood-based panel products have become increasingly specialized in recent years and are used in a wide range of applications. The demand for panels is forecast to increase in the next decades as quality logs for traditional products become increasingly scarce and as designers and consumers gain experience with positive product attributes and new applications. Wood based composite panels will probably in the future, substitute some metallic structural elements and plastics. The strength and durability of wood based materials are determined by their homogeneity and flaw presence. Similar to other composite materials, flaws and damage in WBCPs are: – inherent processing flaws (voids, debonding, fibre breakage, non-uniform fibre and matrix distribution, fibre misalignment, foreign inclusions, ply gaps, delaminations, and matrix cracking). These defects tend to degrade certain structural design properties and remain in the material throughout its life cycle – in-service damage, induced by the exposition to various environmental and mechanical loading conditions (impact events, hygrothermal cycles and fatigue in very variable environmental conditions of temperature and humidity, etc). The main defects observed in service are: matrix cracking and crazing, ply gaps and delamination, fibre pullout, fibre fracture, degradations by aging and environmental aggressive conditions. Delamination is one of the most important defect which occurs in WBCP. In order to design and use WBCP with confidence, it is important to assess their integrity and evaluate their tolerance to flaws. Unfortunately, some flows cannot be detected until failure occurs. Therefore, the structural integrity of WBCP should be assessed by means of nondestructive evaluation methods at the earliest possible stage of damage existence. Nondestructive evaluation techniques have been developed to minimize the effects of defects and to insure the quality control of the final product. Non-destructive inspection techniques (Bucur 2003a, b) with different sensitivity levels can be used for non-destructive evaluation of wood-based composites, such as: X-ray and γ-ray radiography, thermography, ultrasonic techniques, acoustic emission and acousto-ultrasonics. Among these techniques, ultrasonic techniques are the most popular, due to the ease of integration into the production line, of the relatively low cost and inherent safety. The advantages or the disadvantages of different available techniques depend on the type of damage to be detected and on the testing conditions. Sophisticated laboratory techniques can give highly accurate results, but may not be able to assess the state of the structure under in-service conditions. The aim of this chapter is to review the existing ultrasonic techniques for delamination detection in WBCP.

16

Delamination Detection in Wood – Based Composites Panel Products

309

16.2 Basic Aspects Theoretical aspects related to the propagation of elastic waves in anisotropic solids have been discussed in numerous reference books and articles (Kolsky 1963; Viktorov 1967; Musgrave 1970; Auld 1973; Green 1973; Graff 1975; Krautkrämer and Krautkrämer 1990; Birks and Green 1991; Nayeh 1995; Schmerr 1998; Rose 1999). It was stated that the real part of the elastic constants of the materials can be determined by measuring two main parameters, the velocity of the elastic waves and the density of the material. The presence of delamination induces modification of materials mechanical and elastical properties which can be detected with ultrasonic techniques. Ultrasonic inspection involves the utilization of stress waves having a frequency higher than 20 kHz. Linear and non linear ultrasonic techniques have been developed for ultrasonic inspection of wood and wood – based composite materials.

16.2.1 Waves Propagation Paths Bulk waves or surface waves can be used for the characterization of the mechanical 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

Fig. 16.2 Plate modes (Tucker 2001, Figure 1, page 7). https://research.wsulibs. wsu.edu:8443/dspace/bitstream/ 2376/53/1/b_tucker_052101.pdf. (a) Schematic representation of the displacement with plate wave modes; (b) plate mode signal composed from extensional and flexural waves

propagating in solid plates with free boundaries and representing a combination of both compression waves and shear waves. There are two distinct modes of Lamb waves, discernible by their particle displacement patterns and velocities (Fig. 16.2), – 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 antisymmetric 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 velocity or attenuation. Numerous reference books and articles (Bland 1988; Birks and Green 1991; Nayeh 1995; Rose et al. 1987, 1992; Schmerr 1998; Rose 1999; BarCohen and Chimenti 1986; Tang 1988; Tang and Henneke 1989a; Huang et al. 1998) reported the utilization of velocity and attenuation measurements for the evaluation 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 measurements. 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.

16.2.2 Linear Ultrasonic Inspection Techniques 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 transmitter 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

16.2.2.1 Contact Techniques A typical ultrasonic inspection system consists, very roughly, of three main units: the signal generator system with a corresponding preamplifier and amplifier, the transducers and the data acquisition system (PC and oscilloscope). The signal generating system or the pulser, generates short, large amplitude electric pulses of controlled energy, which are converted into short ultrasonic pulses when applied to an ultrasonic transducer. The transducer is the core of all non-destructive ultrasonic inspection procedures. The basic requirements for ultrasonic transducers are: good sensitivity and resolution, controlled beam pattern, reproducible performance under various testing condition and high signal to noise ratio. In that follows pulse –echo technique and through transmission technique will be described. Pulse Echo Technique The typical pulse echo inspection configuration is shown in Fig. 16.3a.The operation principle of the pulse echo technique is to excite the test sample into a mechanical vibration with an ultrasonic transducer driven by a pulse, and to measure the time of wave propagation through the material using the same transducer acting

Fig. 16.3 Pulse echo technique with one transducer. (a) transducer in contact with the specimen (Olympus Panametrics 2009); (b) schematics of the pulse echo technique and signal display in a sound and defective zone ( Stoessel 2004, Figure 19). http://elib.uni-stuttgart.de/ opus/volltexte/2004/1622/pdf/ Dis_Stoessel.pdf

16

Delamination Detection in Wood – Based Composites Panel Products

313

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. delamination), 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 velocity 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. Through Transmission Technique The operation principle of the through transmission technique is to excite the test sample into a mechanical vibration with an ultrasonic transducer driven by a short pulse, and to measure the time of propagation with a second transducer acting as a receiver and disposed to the opposite side of the sample (Fig. 16.5). In this case the bulk wave energy is transmitted through the panel thickness and is received by a second transducer on the opposite side of the specimen This technique is much easier in application than the pulse echo technique because the signal travels only once through the thickness. Changes observed in received signal amplitude or other signal characteristics are induced by the internal structure of the material. This response is then used to measure the velocity of propagation of the ultrasonic wave and furthermore for the characterization of the mechanical behaviour of the specimen. By measuring the time-of-flight, the amplitude of the ultrasonic signal or other parameters of the signal (i.e the RMS voltage), the location and the size of the defects can be estimated. The technique using the normal incidence of the transducers to the surface of the specimen is most sensitive to flaws parallel to the surface (delaminations) (Smith et al. 1989; Wooh and Daniel 1994; Wooh and Wei 1999; Hosur et al. 1998; Žukauskas et al. 2005) while defects lying perpendicular to the surface (cracks in

314

V. Bucur and S. Kazemi-Najafi

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

Fig. 16.5 Through transmission method. (a) Schematic representation of the principle of the through transmission method; (b) transmission of acoustic waves in two zone, a sound zone and a zone with defects (Stoessel 2004, Figure 20, page 75). http://elib.unistuttgart.de/opus/volltexte/2004/ 1622/pdf/Dis_Stoessel.pdf

315

a

b

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

16.2.2.2 Non-contact Techniques The non contact ultrasonic technique and the utilisation of air-coupled ultrasound transducers have gained considerable attention in the last 20 years. The elimination of the physical contact (with a gel) between the transducers and the material under inspection has been an enormous step towards the successful implementation of ultrasonic techniques in numerous applications and industrial processes (Hayward 1997; Castaings et al. 1998; Buckley 1999, 2000; Bharadwaj 2002; Bharadwaj et al. 2000; Blomme et al. 2002; Stoessel et al. 2001; Vun et al. 2003; Kleinschmidt 2003; Döring et al. 2006; Solodov et al. 2006a, b; Lionetto et al. 2007). Technological advancements with air-coupled transducers have made possible the study of the non linear acoustical behaviour of many materials and on the other hand permitted the development of automated nondestructive techniques for materials quality assessment. The non contact technique is a sensitive tool for testing materials’ structure and strength, for monitoring the degradations produced by the environmental conditions, and for quantifying the damage of structural elements. The air-coupled transducers can be configured to work in through transmission or to generate guided plate waves (Fig. 16.7). In conventional through transmission configuration of normal incidence, a beam of air coupled ultrasound excites longitudinal waves which can detect defects in materials. The flexural wave velocity can be detected using focused slanted transmission of air-coupled ultrasound. This transmission mode is used to generate and detect locally the flexural waves in wood and to measure their velocities. Figure 16.8 shows the methodology proposed by Solodov et al. (2004b). To excite the flexural waves in plate specimen with air coupled ultrasound, the “resonance” values of the angle

Fig. 16.7 Possible configurations of air-coupled transducers (Airstar 2001). http://airstar1.com/ air-coupled%20us.htm – air star coupled ultrasound, 2009-06-22. (a) through transmission, longitudinal wave; (b) shear wave; (c) plate wave two sided inspection; (d) plate wave one-sided inspection

16

Delamination Detection in Wood – Based Composites Panel Products

317

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

of incidence need to be determined. This is obtained by sample rotation. The “resonance” is manifested by important rise of the amplitude of the transmitted plane wave. In this case, the angle of incidence θ must satisfy the following condition sin θ =

Vair Vplate

(16.1)

where V are the velocities in air and in plate. In this case, the ultrasonic wave in air and the propagating plane wave in the material move in phase along the surface (Fig. 16.8b). This phase coincidence determines the “resonance”. The “resonance” is understood as the synchronous excitation of the plate wave and its re-radiation into the air from the opposite side of the plate. The acoustic coupling between the plate surface, its thickness t is obtained if t ≤ λs , where λs is the wavelength of the shear wave of the plate material. By measuring θ 0 the velocity of the flexural mode of the plate wave can be measured, along an arbitrary direction of propagation. The wavelength in air is small, about 0.7–0.35 mm in ultrasonic frequency range between 500 kHz and 1 MHz. this provides a localized quasi-plane wave spot of 2–4 mm in the focused area. The advantage of this technique is to provide a single point excitation of the flexural plate waves for use in the experiments for remote monitoring of in-plane local anisotropy in plate type specimens. As noted by Haberger et al. (1979), for Lamb waves propagating in symmetrical directions of thin orthotropic plates, the following relation can be used Vplatei =

√



Ei ωt 3ρ(1 − νik .υki )

1/4 (16.2)

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

a

V. Bucur and S. Kazemi-Najafi

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

To qualify the anisotropy of wood specimens, in an anisotropic plane (i.e LT) an anisotropic factor was calculated as the ratio between the velocities Vplatei in L and T directions. Figure 16.9 shows wood anisotropy in beech and in spruce veneer laminae with flexural waves with air-coupled probes. Solodov and Busse (2006) refined the previous methodological approach and proposed a new methodology based on the mode conversion of focused air-coupled ultrasound into plate and surface waves, by combining fully air coupled and hybrid air –coupled – optic configuration (Fig. 16.10). The principle of this method is based on the observation that an improvement of the weak penetration of air coupled ultrasound in solids (produced by the impedance mismatch at the air-solid interface) can be obtained by using mode conversion in slated configurations. In this case a

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 various 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 ultrasound 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 interior 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

a

b

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 evidence 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) Velocity of surface waves (m/s) Species

Plane

Angle (◦ )

400 kHz

200 kHz

Fir

RL

Beech

RL

Balsa

TL

Oak

RL

0 (axis L) 90 (axis R) 0 (axis L) 22.5 45 67.5 90 (axis R) 0 (axis L) 90 (axis R) 0 (axis L) 90 (axis R)

1320±20 Not reported 1260±20 Not reported 890±90 Not reported (580–640)±20 1360±20 (800–1000) ±40 1570±30 Not reported

1270±20 Not reported 1216±15 1048±15 815±12 727±12 697±12 (1320–1370)±20 570±15 1470±20 846±25

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 interrogation 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.

16.2.3 Ultrasonic Transducers and Scanning Procedures 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 piezoelectric transducers, the airborne transducers can be either piezoelectric or capacitive transducers. 16.2.3.1 Contact Transducers A piezoelectric transducer is schematically illustrated in Fig.16.13. The transducer incorporates a piezoelectric element, which converts electrical signals from pulser/receiver into mechanical vibrations (transmitted mode) which vibration travel through the sample and then, is converted by the receiver into an electrical signals (received mode). The transducers are available in different shape and sizes. The commonly used materials are the ceramics: PZT (barium titanate and lead zirconate titanate), and the polymer polyvinylidenr difluorride (PVDF). For the PZT at a frequency of 1 MHz, the thickness of the ceramics is of about 2 mm and the velocity is of 4170 m/s. The backing layer is epoxy resin containing 1 part in 20 of tungsten powder. This layer absorbs back radiation and provide coupling to the ceramics. A series of matching materials of different impedances may be used to provide a more gradual change in impedance and to reduce reflected radiation at the interface with the specimen. The ultrasonic beam produced by the transducer has two zones, the near field (Fresnel Region) and the far field (Fraunhofer region). In the near field all the components of the beam propagate in parallel. In the far field the beam diverges. The ultrasonic field is diffracted, scattered and absorbed by the material under test. The performance of the transducers is determined by the following parameters: radiation surface area, mechanical damping, physical parameters of the piezoelectric and backing materials and connections of electrical and acoustical

322

V. Bucur and S. Kazemi-Najafi

a

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

components of the system. The main characteristic of an ultrasonic transducer is its frequency. The frequency noted on the metallic capsule of each ultrasonic transducer is the central or centre frequency and depends primarily on the backing material. Transducers of frequencies, ranging between 0.2 and 2.25 MHz provide greater energy and penetration into material while transducers of high frequency ranging between 15.0 and 25.0 MHz provides greater sensitivity to small discontinuities but have reduced penetration. Higher frequency transducers are more sensitive to small discontinuities, because of their shorter wavelength. In addition, higher frequency transducers tend to have better resolution due to shorter energy bursts and the shorter wavelength. On the other hand, in higher frequency range the signal energy attenuates more and tends to scatter, causing a loss of sensitivity for the thicker sections of the material under test. The utilization of transducers operating at different frequencies is recommended. In this way, the desired illumination of the selected features (adhesive, fibres, etc) can be obtained. Proper ultrasonic testing requires careful selection of the transducers frequency range, for obtaining an appropriate balance between sensitivity and penetration for a specific material to be

16

Delamination Detection in Wood – Based Composites Panel Products

323

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 experiments. 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.

16.2.3.2 Non-contact Transducers Non contact transducers or air – coupled ultrasonic transducers can be piezoelectric or capacitive. The non contact probes made out of piezoelectric composites have an active phase composed from piezoelectric rods aligned in parallel along the thickness direction and embedded in a 3D passive polymer matrix (Hayward and Gachagan 1996). The transmitter transducers can be focused and in this way the acoustic matching to air is better than with plain piezoelectric ceramics. For the capacitive transducers no impedance matching is required and, on the other hand, they can produce a wide bandwidth in air. The capacitive transducers are composed of a thin, metallized, membrane film, attached to a direct current bias voltage to a rigid contoured conducted backplate to form a capacitor (Schindel et al. 1995). Applied voltage signals induce vibration into the membrane, and generate ultrasound. The ultrasound is detected by a change in gap between the membrane and the back plate (Fig. 16.14). The wave modes produced by air coupled ultrasound were described in details by several authors (Hutchins and Schindel 1994; Strycek et al. 1997; Castaings et al. 1998; Green 1999, 2000; Dang et al. 2002; Gan et al. 2005). The successful implementation of the non – contact ultrasonic transducers in nondestructive quality control of wood-based panels depends on the ability of the transducers to detect the interaction between ultrasound and the characteristics of the panel under test. An effective transducer must satisfy several conditions by which the measurement system may be assessed, such as: the range over which the measurements can be performed, the smallest intensity of the signal that can be measured, the sensitivity to the typical defects, the reproducibility and accuracy of the measurements, or the amenability to rapid testing in a production line (Grandia and Fortunko 1995; Green 1990; Green 1999, 2000).

16.2.3.3 Scanning Procedures The existing scanning procedures with the modern technology are summarized in Table 16.2. Panel scanning can be performed manually or automatically. Manual scanning is carried out by specific transducer movements by hand in an orbital, swivel, lateral and transversal scanning plan. This requires an experienced operator. The automatic scanning system integrates the ultrasonic instrumentation, the scanning bridge, and

324

V. Bucur and S. Kazemi-Najafi

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 techniques: 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 transducers. 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) Scanning mode

Description

Display the signal for x = time, y = amplitude B scann Image showing a cross section of the tested object perpendicular to the scanned surface and parallel to a reference direction. The cross section will normally be the plane through which the individual A-scans were collected Rotation Orbital scanning B scann Modified B-scan: amplitude is recorded versus rotation angle and time Polar Modified B-scan: scann maximum amplitude is recorded versus rotation angle C scann Image shows a cross section of the test object parallel to the scanning surface Volume 3D representation of the scann inspected volume. At each inspection point of scanning surface a complete A-scan is recorded P scan Projection view of several B or C scans

x

y

Amplitude A

Time of flight t

Data acquisition

yes

yes

A(x0 , y0 , t)

yes

yes

A(β, x0 , y0 , t)

yes

yes

yes

Amax (β x0 , y0 , t)

yes

yes

Amax (β, x0 , y0 )

A(x, y, tn)

β

A scann

D scan

F scan

yes

yes

yes

yes

yes

yes

yes

yes

A(x, y, t)

yes

yes

yes

yes

yes image showing a cross section of the tested object perpendicular to the projection of the beam axis on the scanning surface. The D scan will be perpendicular to the B scan Modified C scan : values yes of some features (centre frequency or phase angle) are recorded and displayed instead of amplitudes

yes

Am(xm , y0 , t) Am(x0 , ym , t) A(x, y, tn ) t(x, y)

yes

yes

F(x, y)

326

V. Bucur and S. Kazemi-Najafi

An A-scan image is a display in which the received pulse amplitude is represented as a displacement along the y-axis, (the amplitude axis) and the travel time of the ultrasonic pulse is represented as a displacement along the x-axis (the time axis) (Fig. 16.4b), or in other words the acquisition of the amplitude A is A(x0 , y0 , t) This is the simplest scan mode. The resolution in the x and y directions depends on the beam width. In a linear amplification system the vertical excursion is proportional to the amplitude of the signal. With the use of logarithmic scale amplifiers the y-axis amplitude – presents a logarithmic scale. A B-scan image is a 2D imaging and is the most common method of ultrasonic imaging Fig. 16.15. The reflectivity is driven from a two dimensional slice through a portion of the specimen. The acquisition of the amplitude A(x0 , y, t). The B-scan image is composed from numerous A-scan measurements, and shows the time of flight versus the position of the measurement. From this B scan image is also possible to read the thickness of the specimen. B scan image do not gives information about the depth of the defect. A C- scan image is a 2D graphical representation and the amplitude acquisition is A(x, y, tn ). Figure 16.16 shows schematically a C – scan image compared with B-scan image. The C-scan image as explained by DIN EN 1330 – 4 shows a cross section of the sample and a defect parallel to the scanning surface, as a projection of the internal structure regardless of their depth. The image reconstruction with C-scan mode integrates the amplitudes of the signals. The defect becomes visible through the change in signal amplitude. The C-scan image illustrates the sample as seen from “above”. The C-scan imaging becomes more and more popular because of the recent improvement of the ultrasonic equipment and the advances in microprocessor technology. Hillger (1997) noted: “Ultrasonic imaging of internal defects with a high degree of validity in composite components requires an optimisation of the pulse

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 frequency 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 Cscan 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).

16.2.4 Non-linear Ultrasonic Inspection Techniques The non-linear acoustical behaviour of different media, observed mainly through the distortion of propagating waves, was study from several decades (Bjorno 2001; Delsanto 2007). The non-linear waves distortion generate fundamentals and higher harmonics or subharmonics which can be found in the frequency spectrum. During wave propagation an increasing number of higher harmonics are generated and contrary to expectation, their amplitude increases with distance. It is to note that the amplitude of the fundamental frequency decreases with the distance because of energy transfer to the higher harmonics. “Classical” non-linear acoustics describes the wave form distortion in homogeneous media. In nonhomogeneous

328

V. Bucur and S. Kazemi-Najafi

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 heterogeneities 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 ultrasonic 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,

Fig. 16.18 Delamination in a wooden rod of 21 cm long, by imaging the second harmonic with B scan. The delamination is at 12 cm from the origin (Solodov et al. 2004b, Figure 10, page 509

16

Delamination Detection in Wood – Based Composites Panel Products

329

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.

16.3 Delamination Detection in Wood-Based Composite Panels 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 references 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 techniques. Special emphasis was put on the study of the influence of experimental conditions (effect of frequency, testing configuration with contact and non contact 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 techniques will be discussed: the through transmission technique and the plate wave technique.

16.3.1 Through Transmission Technique Through transmission technique can be performed with the transducers in contact with the specimen or, with non-contact ultrasonic transducers.

16.3.1.1 Technique with the Transducers in Contact with the Specimen The techniques with the transducers in contact with the specimen where developed in time and frequency domain. Dill-Langer et al. (2005) studied in laboratory conditions, in time domain, the behaviour of a specimen composed from two lamellae bonded with PRF adhesive, having a central adhesion deficient zone (Fig. 16.19a). Inspection with through transmission technique, B – mode scanning – with dry contact transducers, was performed in the direction parallel to the fibres. The following parameters of the ultrasonic signal were recorded: the time of flight (TOF), the initial amplitude of the first recorded oscillation (labelled as initial amplitude) and the total peak-to-peak amplitude (labelled global pp amplitude) of the entire recorded signal (Fig. 16.19b). The experimental results shown in Fig. 16.19c point out that in

330

V. Bucur and S. Kazemi-Najafi

a

b

c

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 timeof-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 specimens 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 defective 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 frequency 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 D

(16.3)

The “masking effect ” increases with diffraction angle such as:   V/f Dmin ≥ 2d.tg (arcsin 1.22 D

(16.4)

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



Dmin



V/f ≥ 2.(e − d).tg arcsin 1.22 D

 (16.5)

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. 16.3.1.2 Technique with Non-contact Ultrasonic Transducers In this section several applications of non contact techniques will be discussed, using transducers in conventional configuration, which is normal to the surface

16

Delamination Detection in Wood – Based Composites Panel Products

335

a

b

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

of the inspected specimen. In laboratory conditions, we discuss the inspection of specimens of oriented strandboards and of a 3 layered composite. Delaminations in specimens of pinus radita will be also commented. The utilisation of the non contact ultrasonic transducers for the detection of voids in (OSB) with high, low and random alignment, was reported by Vun (2003) and Vun et al. (2003). Figure 16.24 shows the experimental configuration for OSB testing in conventional direct contact technique and in non contact technique. In OSB panels in-plane density variation contributes to the formation of delamination and voids. These defects are strong scatters of elastic waves. The validity of the non contact technique was proved in comparison with contact technique in a through transmission mode. For the direct contact technique two transducers of 100 kHz frequency were used, coupled with a thin silicon gel. To characterize the OSB in direct contact configuration the following parameters were used: the velocity of ultrasonic wave propagation, the material impedance, the attenuation given by the peak amplitude and the RMS voltage which represents the signal intensity of the acquired signal. The non contact transducers used 250 kHz frequency. In this case the following variables were calculated: the velocity, the integrated response which is the

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 transmission 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 transmission 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 technique 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 parameters. 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 variables 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 measurement 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 piezocomposite 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 previous 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 delamination 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 transducers, 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 thickness. 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-0708103163628/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

16.3.2 Plate Wave Technique 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

As reported by the abundant literature for different types of composites, several techniques have been proposed to excite and receive plate waves such as: – immersion technique (Hosten 1991; Bar-Cohen and Chimenti 1986; Chimenti 1997; Karim et al. 1990; Martin and Chimenti 1987) – acousto-ultrasonics (Huang et al. 1998; Rodgers et al. 1991; Rose et al. 1987; Stiffler 1986; Tang et al. 1988; Tang and Henneke 1989a, b), – laser generated ultrasound (Han et al. 1999), – acousto ultrasonic angled beam technique (Rogers 1995) – air-coupled ultrasonic transducers (Castaings and Cawley 1996; Potel et al. 2008). The immersion technique involves excitation and reception of leaky Lamb waves by immersing the specimen into a liquid (water). This technique is not recommended for wood-based composites which are very hygroscopic. The Laser generated Lamb waves technique requires a specific equipment. Acousto – ultrasonic and angled beam ultrasonic techniques seem to be appropriated the most appropriated techniques for the inspection of wood-based composites panels. In that follows, we will discuss the capability of plate wave technique to detect delamination and debonding in wood-based composite panels, using the contact technique as well as the non contact technique and as reported by Grimberg et al. (2000), Tucker (2001), and Tucker et al. (2003). 16.3.2.1 Technique with the Transducers in Contact with the Specimen To study the capability of plate wave technique to detect the delamination in wood – based laminated. Tucker (2001) and Tucker et al. (2003) utilised specimens with simulated delamination in a MDF panels. Two individual MDF 3.2-mm panels were boned together with polyvinyl acetate resin. The delaminated area was a square of 60 mm2 . The plate wave test setup is described in Fig. 16.27. The transducers were put in direct contact with the specimen. For the determination of the phase velocity, the transducers were kept at a constant distance (178-mm). The transducer pair was moved along the plate at with a step of 15-mm. The excitation frequency was successively 10, 15, 20, 25, and 30 kHz. Figure 16.27c shows the location of the delamination area with 25 kHz transducers in a MDF board. For each frequency the time of flight was recorded. It was concluded that the lower frequencies are more sensitive to the delamination detection than higher frequencies and that the plate wave technique is sensitive to delamination detection and location in MDF panels. 16.3.2.2 Technique with Non-contact Ultrasonic Transducers Air coupled ultrasound C –scan image of delaminations between a veneer lamina and a solid wood layer is shown in Fig. 16.28. The thickness of the specimen is 30 mm and the scanned area was 30 80 mm. Dark zones horizontally displayed, of different thickness of missing glue are see. In the upper right side of the image other defects can be seen. The same specimen was inspected with optical lockin

16

Delamination Detection in Wood – Based Composites Panel Products

a

341

b

c

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-tucker052101.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 composite 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 ultrasound – 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 particleboard core and a 1 mm beech veneer layer on each side of the core. The

342

V. Bucur and S. Kazemi-Najafi

Fig. 16.28 Air coupled ultrasound C –scan image of delaminations between a veneer lamina and a solid wood layer. (Stöessel 2004 ) http://elib.unistuttgart.de/opus/ volltexte/2004/1622/pdf/Dis_Stoessel.pdf. Permission from (Stöessel 2004, Figure 41, page 100). The image was obtained equipment Hillscan 4000. with focused transducers (focal length 40 mm) central frequency 450 kHz, transmitter input up to 1400 Vpp, optimal rectangular pulse – burst mode, preamplifier 40 dB

delamination was simulated by inserting a regular Teflon frame of 80 μm thickness. The frequency of the ultrasonic beam (induced by a Hertzian contact) was 60 kHz. The measured velocity in the sound zone was in average 4301 m/s rather in the zone with delamination was considerably lower, between 1103 and 2128 m/s. It was concluded that the velocity of the plate waves generated in the multi-layer wood based specimen by means of the low frequency ultrasound transducers with Hertzian contact is an efficient parameter for delamination detection in laboratory.

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

16.3.3 Industrial Applications of Non-contact Technique The non contact techniques have been primarily used for testing composites for aeronautical and space industry. Furthermore the applications have been extended very quickly to large a variety of materials, and especially to hydrophobic materials such as wood and wood based composites (Buckley 2000; Vun et al. 2005, 2008). In wood industry the air coupled ultrasound transducers have been used mostly in transmission configuration for: the detection of decay in wood prior to processing, of delaminations in wood-based composites panels, and for in service inspection of wood products. Electronic Wood System (2006), using ultrasonic transmission mode, patented and put on the market an equipment for delamination detection with air coupled ultrasound in thick MDF (>120 mm) and in OSB, OSL, LVL, LVP and plywood having the thickness up to 200 mm. The system detects blows, delaminations and air pockets using a resonance technique. Papadakis and Kovacs (1980) described in detail this technique used firstly for quality assurance of iron parts. The principle of this method is based on the fact that a resonance occurs at the frequency at which an integral number of half-wavelengths fit into some dimensions of the workpiece. The ultrasonic energy is imparted to the workpieces which vibrates at its natural frequencies. In a zone with delaminations, the natural frequencies are altered. The analysis of the resonance frequencies permits the detection of the defective zone. The main advantage of the resonance technique is the capability to sample almost the entire workpiece (except the areas around nodes that are not stressed). “The question of corrective action in the presence of a reflection from a potential flaw that may or may be not of a detrimental size or shape is still a management decision rather then a purely scientific one” (Papadakis 1999). The patent developed by Electronic Wood System (2006) used non contact ultrasonic transducers installed across the panel width, in transmission technique configuration (Fig. 16.30). The minimum size of detectable delamination is 1 cm (Kleinschmidt 2003). Several thresholds allow display of a multicoloured ultrasound picture of the panel. The picture recognizes variations of moisture content, density, thickness and temperature using ultrasound, microwave and X ray techniques. The air coupled ultrasonic transducers are encapsulated and the system is protected against noise produced by sander, saws, compressed air, heat and dust. A similar system exists in Australia at Heyfield, Victoria (Hurley 2008) for the detection of the internal checking within dried eucalyptus boards. The non contact scanning system is equipped with 16 emitters and 16 receivers. The transducers are manufactured by Airstar Inc. US (Loertscher et al. 1996). The detection is performed on boards travelling at 1800 m/min in production line. A real time coloured picture is displayed on each board and is used for further operations (trimming, etc). The non contact ultrasonic technology is very robust, inherent safety and cost effective; the transducers are protected against external noise, heat and dust, humidity and high temperature; the maintenance time is significantly reduced.

344

a

V. Bucur and S. Kazemi-Najafi

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 testing 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 particle, 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 inspection 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 transmission 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 characterization. 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 harmonics 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 dimensions 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 workpiece 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 composite 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 characterization 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/dilllanger1.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 properties 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 transducers. 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 debonding 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 transient 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 denoising 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 technology 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 ultrasonic 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 technology 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 monitoring 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 determination 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

Neuenschwander J, Niemz P, Kucera LJ (1997) Orientierende Untersuchungen zur Anwendung der bildgebenden Ultraschallprüfug zur Fehlererkennung in Holz. Holz als Roh-und Werkst 55:339–340 Olympus, Panametrics (2009) http://www.olympus-ims.com/data/File/panametrics-UT.en.pdf. Accessed 16 June 2009 Papadakis EP (1999) Ultrasonic instruments and devices. Reference for modern instrumentation, techniques and technology. Academic, San Diego, CA Papadakis EP, Kovacs BV (1980) Theoretical model for comparison of sonic-resonance and ultrasonic velocity for assuring quality in instant nodular iron parts. Mater Eval 38(6):25–30 Potel C, de Belleval KF (1993) Propagation in a periodically anisotropic multilayered media. J Acoust Soc Am 93(5):2669–2677 Potel C, de Belleval JF, Genay E, Gatignol Ph (1996) Behavior of Lamb waves and multilayered Rayleigh waves in an anisotropic periodically multilayered medium. Application to the longwave length domain, Acustica-Acta Acustica 82(5):738–748 Potel C, Leduc D, Morvan B et al. (2008) Lamb wave attenuation in a rough plate. I. Analytical and experimental results in an anisotropic plate. J Appl Phys 104:074908–074908–10 Rogers WP (1995) Elastic property measurement using Rayleigh-Lamb waves. Res Nondestruct Eval 6:185–208 Rodgers JM, Green AT, Borup SW (1991) Acousto-ultrasonic measurement of internal bond strength in composite wood products. Mater Eval 49(5):566–571 Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge Rose JL, Zhu W, Cho Y (1992) Boundary element modelling for guided wave reflection and transmission factor analyses in defect classification. IEEE Ultrason Proc Symp 1:885–888 Rose WR, Rokhlin SI, Alder L (1987) Evaluation of anisotropic properties of graphite – epoxy composites using Lamb waves. In: Review of Progress in Quantitative Nondestructive Evaluation, vol 6B, Plenum Press, New York, pp 1111–1118 Schindel DW, Hutchins DA, Zou L, Sayer M (1995) The design and characterization of micromachined air-coupled capacitive transducers. IEEE Trans Ultrason Ferroelect Freq Contr 42:42–50 Schmerr LW Jr (1998) Fundamentals of ultrasonic nondestructive evaluation, a modelling approach. Plenum Press, New York Schmoldt D L, Ross RJ, Nelson R M (1996) Ultrasonic defect detection in wooden pallet parts for quality sorting. SPIE Proc Series 2944:285–295 Solodov IY (1998) Ultrasonics of non-linear contacts: propagation, reflection and NDE applications. Ultrasonics 36:383–390 Solodov IY (2001) CAN: an example of nonclassical acoustic nonlinearity in solids. Ultrasonics 40:621–624 Solodov I, Strössel R, Busse G (2004a) Material characterization and NDT using focused slanted transmission mode of air-coupled ultrasound. Res NonDestruct Eval 15:1–21 Solodov I, Pfleiderer K, Busse G (2004b) Nondestructive characterization of wood by monitoring of locall elastic anisotropy and dynamic nonlinearity. Holzforschung 58:504–510 Solodov I, Busse G (2006) New advances in air-coupled ultrasonic NDT using acoustic mode conversion. ECNDT 2006 Berlin – We.2.4.2 http://www.ultrasonic.de/article/ ecndt2006/doc/We.2.4.2.pdf. Accessed 7 January 2007 Solodov I Y, Doering D, Pfleiderer K, Busse G (2006a) Linear and nonlinear NDE using aircoupled Lamb waves. AIP Conf Proc 820:1492 Solodov I Y, Pfleiderer K, Gerhard H, Predak S, Busse G (2006b) New opportunities for NDE with air-coupled ultrasound NDT Int 39:176–183 Stiffler RC (1986) Wave propagation in composite plates. Ph.D. Dissertation, College of Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA Smith BT, Heyman JS, Buoncristiani AM, Blodgett ED, Miller JG, Freeman SM (1989) Correlation of the deply technique with ultrasonic imaging of impact damage in graphite-epoxy composites. Mater Eval 47(12):1408–1416

350

V. Bucur and S. Kazemi-Najafi

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 composites. 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 strandboard. 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 properties 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, Cic˙enas 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 17.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crack Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.1 Crack Between Lamellae . . . . . . . . . . . . . . . . . . . . 17.2.2 Crack Inside the Lamella Material . . . . . . . . . . . . . . . . 17.3 Crack Depth Detetrmination with Ultrasonics . . . . . . . . . . . . . . 17.4 Strength Prediction of Lamellas In situ . . . . . . . . . . . . . . . . . 17.5 Detection of Other Internal Defects in Glulam . . . . . . . . . . . . . . 17.6 Shear Strength Determination Between the Lamellae . . . . . . . . . . . 17.7 Delamination and Other Defects in Structural Element of Historic Buildings 17.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

353 355 355 355 356 358 358 358 360 363 364

17.1 Introduction The importance of nondestructive techniques for assessing the internal conditions of structural members is generally accepted for old or more recent buildings (Ross and Pellerin 1994; Pellerin and Ross 2002, Tanaka et al. 1998; Brashaw et al. 2005; Wacker et al. 2007). Efforts to preserve buildings that have nationally significant importance have been based on the development of nondestructive techniques able to assess their mechanical performances (Wang X and Wacker 2006; Sandoz and Benoit 2007; Lee et al. 2007; Divos et al. 2007). The overall goal of these efforts is to preserve to the maximum possible extent of the historical materials used for the monuments. On the other hand, in recent years, the stability of existing glulam structures became an important issue after the collapse of some of the early structures. For example, in Germany on Jan.2, 2006, 15 people, most of them children

F. Divos (B) Faculty of Wood Science, University of West Hungary, Sopron, Hungary e-mail: [email protected]; [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_17, 

353

354

F. Divos

Fig. 17.1 Crack depth determination with 0.1 mm thick feller gauge

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 characteristic 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 technique 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,

Fig. 17.2 Crack depth determination by feller gauge penetration is limited by the crack path, which is not straight

17

Delamination Evaluation of in-Service Glulam Beams

355

ultrasonic technique can be an alternative when applied to crack depth determination. Typically the depth measured by ultrasonic technique is a bit higher, than feller gauge plate penetration depth.

17.2 Crack Types In glulam structures, two main crack types are recognised, the crack between the lamellae and the cracks inside the lamellae material.

17.2.1 Crack Between Lamellae Adhesive failure or other technological failures cause partial separation between lamellas. It is to note that after relatively few years of service (ex 6 years), the glulam beams exposed to sun and rain suffer from delamination, as shown in Fig. 17.3. Fig. 17.3 Cracks in glulam beam in service for 6 years. Crack depth is 24 mm

Figure 17.4 shows a typical example of delamination in a glulam structure, long time in service, on Robinia pseudoacacia. This sample has been taken from a demolished structure after 34 years of service. The structure has been demolished due to severe delamination and shape change.

17.2.2 Crack Inside the Lamella Material In the past, glulam elements containing pith have often been used by some manufacturers, 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. 17.4 Severe delamination in a glulam beam made of Robinia pseudoacacia after 34 years of service. Lamella thickness is 20 mm

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 delamination 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.

17.3 Crack Depth Detetrmination with Ultrasonics 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

Fig. 17.6 Cracks at the glulam beam end due to the fastening screws

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) L = 0.5∗ sqrt(V 2 (Tc − Ti )2 + 2V(Tc − Ti )D)

(17.1)

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.

17.4 Strength Prediction of Lamellas In situ Re-evaluation after many years of service of an existing glulam structure needs strength and stiffness data. In the case of old beams, in function for more than 30 years, often no strength grading was applied to produce the beams. Hence wave velocity determination in the fiber direction can provides the lamella modulus of elasticity (MOE) according to the following equation: 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 relatively 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.

17.5 Detection of Other Internal Defects in Glulam Detecting internal decay, holes or other defects in glulam structures is possible by measuring the transit time of P or S-waves between the two faces of the beam. This type of defect was simulated on the sample used for this experiment. The test set up using P waves is shown in Fig. 17.9. The defective zone showing a longer transit time is located clearely by the red color on the graph.

17.6 Shear Strength Determination Between the Lamellae The shear strength between the lamellae provides critical information about the glulam 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

Fig. 17.8 Velocity determination in a lamella along the fibers, using a special device for keeping a constant distance between the transducers

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 attenuation measurements via shear wave amplitude can be a potential tool for the

360

a)

F. Divos

b)

Fig. 17.10 Shear transducers. (a) The position of shear transducers on the transversal section of the laminar structure; (b) the probes

Fig. 17.11 Relationship between the measured shear strength and signal amplitude

evaluation of delamination between glued layers in beams. More research and technical development is necessary for in situ applications.

17.7 Delamination and Other Defects in Structural Element of Historic Buildings 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 renovation. 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 structure of an old baroque castle – Esterhazy Castle – in Pápa in Hungary (Fig. 17.12).

17

a)

Delamination Evaluation of in-Service Glulam Beams

361

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 ultrasonic 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 consistency 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 demonstrated 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 modulus 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)

Fig. 17.13 In situ measurements

applied units in the equations are: MORest [MPa], Fscrew [kN] and v, the velocity [km/s]: MORest = 0.809Fscrew · v2 + 26.8 A similar MOR predictor formula applies for hardwoods: MORest = 1.258Fscrew · v2 + 36.9 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

Fig. 17.14 Relationship between the modulus of rupture and the predictor parameters.  represents coniferous, ∇ represents hardwood specimens

MOR [Mpa]

140 120 100 80 60 40 40

60

80 100 Predictor [MPa]

120

140

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 individual wooden beams have been estimated with ± 9 MPa accuracy. Figure 17.15 shows the strength distribution in the larch ceiling.

17.8 Summary In this chapter are discussed several procedures for delamination evaluation of in-service glulam beams and other structural members via ultrasonics Efforts to preserve buildings that have nationally significant importance have been based on the development of nondestructive techniques able to assess their mechanical performances. The stability of existing glulam structures became an important issue after the collapse of some of the early structures. One of the most frequent observed defects are the cracks. The primary characteristic of a crack is the depth. The common test for crack depth determination is the penetration depth of 0.1 mm thick feller gauge. Ultrasonic technique can be an alternative when applied to crack depth determination. Typically the depth measured by ultrasonic technique is a bit higher, than feller gauge plate penetration depth. In glulam structures, two main crack types are recognized, the crack between the lamellae and the cracks inside the lamellae material. Cracks form a material discontinuity and as such are a vibration propagation barrier. Re-evaluation after many years of service of an existing glulam structure needs strength and stiffness data. In the case of old beams, in function for more than 30 years, often no strength grading was applied to produce those beams. Wave velocity determination with P waves, in the fiber direction can provides the modulus of elasticity of the lamellae. The shear strength between the lamellae provides critical information about the glulam beam mechanical capacities. Predicting

364

F. Divos

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 determines 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 structures. 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 18.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . Material and Methods . . . . . . . . . . . . . . . . . . . 18.2.1 Tests on the Basic Material . . . . . . . . . . . . . 18.2.2 Test on Parquet Planks . . . . . . . . . . . . . . . 18.2.3 Analytical Model A: Calibration Model . . . . . . . 18.2.4 Analytical Model B: Distortional Effects . . . . . . . 18.2.5 Analytical Model C: Gap Opening . . . . . . . . . . 18.3 Deformations in Parquet Floors . . . . . . . . . . . . . . . 18.3.1 Model A: Calibration and Comparison to Test Series 2 18.3.2 Model B: Cupping of the Parquet . . . . . . . . . . 18.3.3 Model C: Stresses in the Glue Line and Gap Opening . 18.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 18.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

365 366 366 367 368 370 370 371 371 373 374 376 377 378

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: [email protected]

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_18, 

365

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 complemented 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.

18.2 Material and Methods 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

Half element = 94 mm

X Y

click joint

local coordinate system

2 mm

14.2 mm

Z

Surface layer: Oak

(Quercus robur L.)

Core layer: Pine

(Pinus sylvestris L.)

8.6 mm

Half element = 94 mm

Backing layer: Veneer pine

Fig. 18.1 Geometry and consistency of the parquet floor

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.

18.2.1 Tests on the Basic Material Wood species used in these experiments are: pine, oak, beech and ash. The adsorption 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.

18.2.2 Test on Parquet Planks The influences of different parameters such as material properties, material orientation, properties of the glue line and the geometry of the product on the stresses and deformations were tested. 30 square samples with a side length of 150 mm were cut out of parquet planks and conditioned in a standard climate of 20◦ C/65% relative humidity until reaching equilibrium moisture content. Ten specimens had a lacquered surface layer (denoted test series 2, A1–A10) and ten specimens were non-lacquered (denotes test series 1, and A11–A20). In addition, the transport of moisture in one dimension only was enforced on 5 lacquered and 5 non-lacquered specimens by applying moisture insulation on the edges. The samples were dried in a climate of 20◦ C for a period of 28 days. The bending deformation of parquet specimens was measured at four points on the surface layer of the specimens (Fig. 18.2). The vertical deformation was measured along two orthogonal directions of the plate. In position A the grain direction of the core layer and thus x axis was parallel to the primary axis of the global

Position A H

G D A

2 (Z) 1 (X)

3 (Y)

I

E C

B

measurement points for horizontal deformation

measurement points for horizontal deformation

F

direction of x axis

Position B G' = I

A' = G

2 (Z)

I' = C

H'

F' = B

E' = E

D' = H B' = D

C' = A direction of x axis

3 (X)

1 (Y)

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, curvature 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 parquet 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.

18.2.3 Analytical Model A: Calibration Model The calibration model has been constructed for calibration and verification of the numerical analysis applied to specimens with moisture-isolated edges. Some assumptions have been made in estimating the effective coefficient of diffusion, These are: • 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 continuative 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))   ∂ λ(T) ∂T ∂T = ∂t ∂x cρ ∂x   ∂u ∂ ∂u = DW (u) ∂t ∂x ∂x Where the parameters are expressed in the following units: – diffusion coefficient [m2 s−1 ] – temperature T – specific heat c [Jkg−1 K −1 ]

(18.1) (18.2)

18

Moisture Induced Stresses and Deformations in Parquet Floors

369

– 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 surface 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 determined 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 direction 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 specimens 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

R (variable) surface layer L

425

65

425

SL: left strip

SL: middle strip

SL: right strip

T (variable) H

G D A

2 (Z) 3 (Y)

I

E

F

B

C

1 (X) core layer

T 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.

18.2.4 Analytical Model B: Distortional Effects 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 component. 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.

A (0/14.2/13)

Moisture exchange limited to top surface

B (94/14.2/13) u1 = 0

2 (z) u3 = 0

125 + 1 + 125

1(x) 94

3 (y)

u2 = 0

A v

B

Fig. 18.4 Analytical model A: Geometry, static system and boundary conditions used for the modelling of the distortional effects

18.2.5 Analytical Model C: Gap Opening 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

2 8.6 3.6

1 (X)

A

B B.3

A.3

uA.1

uB.1

Fig. 18.5 Geometry, static system and boundary conditions of model C

behaviour of the click joint and the gaps of the surface layer. The model corresponded 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 dimensions. 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.

18.3 Deformations in Parquet Floors In the following the moisture induced stresses and deformation in parquet floors determined with the models A, B and C will be discussed.

18.3.1 Model A: Calibration and Comparison to Test Series 2 The estimation of the moisture transport using different effective diffusion coefficients 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 Z Y

L

X v

L L

w

Fig. 18.6 Bending v in xz plane

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

Fig. 18.7 Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%. Bending deformation of lacquered specimens (A1–A10 are the specifications of the samples of test series 2)

Bending v in xz plane [mm]

–0.05

Numerical simulation Measurements

–0.1 –0.15 –0.2 –0.25 –0.3

Mean values

–0.35 A1

A2 A3 A4 A5 A6 A7 A8 A9 A10 Specification of lacquered specimens

0

Fig. 18.8 Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%. Bending deformation of non lacquered specimens (A11–A12 are the specifications of the samples of test series 2)

Bending v in xz plane [mm]

–0.05

Numerical simulation Measurements

–0.1 –0.15 –0.2 –0.25 –0.3 Mean values –0.35 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 Specification of non lacquered specimens

Moisture Induced Stresses and Deformations in Parquet Floors

Fig. 18.9 Overall comparison of Analytical model A: Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%

373

0 Bending v, measurements [mm]

18

Laquered specimens Non lacquered specimens

–0.05 –0.1 –0.15 –0.2 –0.25 –0.3 –0.35

–0.4 –0.4 –0.35 –0.3 –0.25 –0.2 –0.15 –0.1 –0.05 Bending v, numerical analyse [mm]

0

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.

18.3.2 Model B: Cupping of the Parquet 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 calculation, 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 Angle = 45 deg Angle = 90 deg Angle = 0 deg

0.7

Fig. 18.10 Influence of the surface layer’s angle of growth rings after reduction of the moisture content from 7.5 down to 5%. (Local tangential direction parallel to horizontal plane at α = 0◦ )

Cupping v [mm]

0.6 0.5 0.4 0.3 Beech Beech SL 3.6mm Angle ==Beech 30deg Angle 30deg Angle =SL 45 deg Angle Oak Oak =Oak 45 3.6mm deg Angle =SL 90 deg Angle =Ash 03.6mm deg Ash Ash Angle deg Angle==090 deg

0.2 0.1 0

0

20

40

60 80 Time [days]

Beech Oak Ash

100

120

374 0.8 0.7 0.6

Cupping v [mm]

Fig. 18.11 Influence of the surface layer’s thickness after reduction of the moisture content from 7.5 down to 5%

S. Blumer et al.

0.5 0.4 0.3 4.8 mm Surfacelayer layer2.6mm 2.6mm Surface 3.6 mm Surfacelayer layer3.6mm 3.6mm Surface 2.6 mm Surfacelayer layer4.8mm 4.8mm Surface

0.2 0.1 0

20

40

60 80 Time [days]

100

120

100

120

0.8 0.7 0.6

Cupping v [mm]

Fig. 18.12 Influence of the surface layer’s material after reduction of the moisture content from 7.5 down to 5%

0

0.5 0.4 0.3 Beech SL 3.6mm Beech Oak Oak SL 3.6mm Ash SL Ash3.6mm

0.2 0.1 0

0

20

40

60 80 Time [days]

shown in Fig. 18.11, for moisture content decreasing from 7.5 to 5%. The influences 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.

18.3.3 Model C: Stresses in the Glue Line and Gap Opening 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 vertical 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 parametric 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 vertical 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 surface 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.

Cut through the parquet plank [mm]

15 SL 4.8 mm/CL 8.6 mm SL 3.6 mm/CL 8.6 mm SL 2.6 mm/CL 8.6 mm

Surface layer (SL)

10

Groove Core layer (CL)

5

Backing layer (BL) 0

–30

–20

–10 0 10 20 Horizontal stresses S11 [MPa]

30

Fig. 18.15 Horizontal stresses S11 in CUT B Influence of the geometry of the surface layer

3

φR,T = 0.8 φR,T = 1.6 φR,T = 3 hoak hpine

Vertical stresses S22 [MPa]

2.5 2 1.5

ΦR,T = 0

1

ΦR,T = 0.8..1.6..3 (gradient)

0.5 0 –0.5 –1 –1.5 –2

0

5 10 15 Horizontal distance to the gap [mm]

20

Fig. 18.16 Influence of creeping gradient in the surface layer

18.4 Conclusion As concluding remarks it is noted that a model that includes the whole parquet system 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 techniques 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 conditions 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 TVBK1031, 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 Sergio J. Sanabria, Christian Müller, Jürg Neuenschwander, Peter Niemz, and Urs Sennhauser

Contents 19.1 19.2 19.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical Considerations . . . . . . . . . . . . . . . . . . . . Material and Methods . . . . . . . . . . . . . . . . . . . . . . 19.3.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . 19.3.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . 19.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . 19.4.1 Imaging of Glue Presence and Repeatability of Measurements 19.4.2 Influence of Natural Variability and Anisotropy of Wood . . 19.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

379 380 381 381 382 383 383 385 385 386 386

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 specimens 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 S.J. Sanabria (B) Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland e-mail: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_19, 

379

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 delaminations 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 surface 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 inspection 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 precise imaging of areas with and without adhesive. Advantages and limitations of this method are discussed.

19.2 Theoretical Considerations The interpretation of the measurements is based on the theory of plane waves in homogeneous isotropic layered media (Brekhovskikh 1980). The sample is modeled as a three layers system, i.e. wood/glue/wood for glued material and wood/air/wood in the case of non-glued material. Due to the high acoustic impedance mismatch between air and solids the pressure level of an ACU signal which propagates through non-glued material is significantly lower than the level for glued material. Only a single echo of a longitudinal wave propagating through the three layers is considered. The acoustic attenuation in the glue line is neglected. It is assumed that the voltage level measured with an ACU transducer is proportional to the force exerted on its surface by ultrasonic waves (Schmerr and Song 2007). A simplified expression for the level ratio is given in Eq. (19.1): Lglued/non glued = 20 · log10

Vglued Vnon glued

= Twood→glue→wood − Twood→air→wood

 1 ·Z2 Zi = ρi · ci T1→2→1 = 20 · log10 (Z4·Z+Z 2 ) 1 2 (19.1) Where: Lglued/non glued (dB) is the amplitude level ratio between ACU signals propagating 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 material 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 propagation 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.

19.3 Material and Methods 19.3.1 Sample Preparation A total of 46 samples of common spruce (Picea abies Karst.) were manufactured in the Wood Physics Laboratory of ETH Zurich; each consisting of two 5 mm thick solid wood lamellas glued together except for some defined areas (Table 19.1). The R HB 110) applied adhesive is a one-component polyurethane resin (PURBOND 2 to one side of the boards with an amount of 200 g/m . The boards were pressed together hydraulically during 3 h with a stress of 0.8 N/mm2 . Before the gluing the wood was conditioned to normalized climatic conditions (T = 20◦ C and RH = 65%), which were afterwards also used for storage. Only solid wood lamellas with a small percentage of knots, resin pockets, grain distortion, etc. were used in order to analyze the variability of ultrasonic signals propagating through defect free glued timber. The cross-section of the samples is approximately in the orthotropic R-T plane and the curvature of the year rings is negligible (Fig. 19.1). After ultrasonic measurement, samples of type C and D were broken up and the profile of the transition between glued and non-glued areas was recorded with optical means.

Table 19.1 Geometry of glued timber samples manufactured for ACU measurement Type

Description

A B

Single solid wood lamella of dimensions 500×100×10 mm3 Two lamellas of dimensions 500×100×5 mm3 glued together to form a glued timber object of 500×100×10 mm3 Same geometry as B. No adhesive applied in the left half area (250×100 mm2 ) of the lamellas Same sample type as B. Adhesive only applied in two small areas (about 30×100 mm2 ) on the left and right edges of the lamellas

C D

382

S.J. Sanabria et al.

Fig. 19.1 Photographs of external surfaces of a typical C sample. An aluminum spacer of 1 mm thickness on the edge of the non-glued part allows control of the gap thickness. The year ring angle varies between 90º (propagation in orthotropic T direction) and 45º. Small defects like a 26 mm long resin pocket were allowed

19.3.2 Experimental Setup The measurement setup is shown in Fig. 19.2. Two ACU broadband planar transducers (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 sample, 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 diamR moves eter of the sound field penetrating the latter. A three-axis system from ISEL 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 generated 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 ]:

VPEAK = max V(t) [t1, t2]

VRMS

    =

1 t2 − t1

t2 V(t)2 dt

(19.2)

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.

19.4 Results and Discussion 19.4.1 Imaging of Glue Presence and Repeatability of Measurements Figure 19.3 demonstrates successful ultrasonic imaging of absence and presence of adhesive of a typical glued timber sample of type C, which corresponds to the object photographed in Fig. 19.1. As expected, there is a strong voltage reduction in the left area of the board surface, corresponding to the non-glued region. The transition between glued and non-glued areas could be imaged accurately. Figure 19.4 shows an ultrasonic image of a sample type D, in this case the two glued areas on the left

384

S.J. Sanabria et al.

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 amplitude 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 predicted 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.

19.4.2 Influence of Natural Variability and Anisotropy of Wood Wood inhomogeneity introduces variations of up to 8 dB in voltage measurements of glued material without compromising the detectability of non-glued areas. Due to the small uncertainty of ACU measurements specific wood structure features can be visualized in the C-scans. Regions with highest latewood concentration show lowest voltage levels. A possible reason is the fact that latewood has higher acoustic impedance than earlywood and therefore larger impedance mismatch with air. Small defects in the material decrease the measured voltage, since they scatter partially the ultrasonic field; for instance, a resin pocket can be visualized in Fig. 19.3. Variations of the year ring angle could not be correlated to voltage amplitude changes in a clear fashion, an indication that the influence of anisotropy is not large for the inspected objects.

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.

and precise spatial data acquisition. Moreover it is a fully non-invasive method, since no couplant is required between transducers and sample. Current state-of-theart transducers plus moderate pulser voltage and receiver gain allow transmission through 10 mm thick glued timber samples with a signal-to-noise ratio of 55 dB; therefore inspection of thicker objects is promising. The repeatability error is smaller than 1%. A through-transmission measurement set-up achieves level variations of up to 50 dB between glued and non-glued material, which ensures a reliable glue line assessment despite amplitude variations of up to 8 dB in bonded regions, due to the heterogeneity of the wood structure. Future research work is planned to inspect thicker (over 10 cm) multiple laminated glued timber. The main challenge is to resolve small amplitude level variations between bonded and disbonded areas from larger level variations within bonded material (higher influence of natural variability and anisotropy).

19.6 Summary Wood is a sustainable construction material. Glued timber products make efficient use of the strength properties of solid wood; moreover, structural members of expanded dimensional and geometrical properties can be produced. The integrity of the glue lines of timber laminates needs to be assessed during the full life cycle of the product; therefore, a non-destructive reproducible inspection method is required. As part of an ongoing project, we performed air-coupled ultrasound (ACU) measurements in glued timber laminates. A normal transmission setup with 120 kHz commercial transducers was used to analyze samples consisting of two spruce solid wood lamellas glued together with polyurethane adhesive introducing defined delaminated areas. Ultrasonic scanning with high resolution was performed to successfully image the presence or absence of glue. The geometry of the delaminated regions and features of the wood structure could also be visualized. We have demonstrated that ACU is a sensitive, accurate, reproducible and non-invasive inspection alternative with respect to conventional contact techniques; therefore, it is wellsuited for glued timber inspection. Future work is planned for the inspection of more complex glued timber structures. Acknowledgements This research has been supported by the Swiss National Science Foundation under contract 200021-115920. The authors acknowledge the work of Oliver Tolar and Fabian Binkert in the analysis of optical images and ultrasonic data.

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 timber 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 measurements. Springer, New York, NY Siddiolo AM, D’Acquisto L, Maeva AR, Maev RG (2007) Wooden panel paintings investigation: 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

Delamination phenomena in manmade composites (Tay 2003, Sridharan 2008) as well as in wood and wood-based composites have received much attention from scientists and practitioners due to serious technological implications and obvious scientific curiosity for this subject. The long term objectives for the application of current research on delamination in wood and wood-based composites reported in this book are to develop robust design/modelling tools for minimizing the potential failures of both conventional and new products. Topics such as the theoretical aspects, the methodology for delamination detection and the factors inducing and affecting delamination were reviewed. The industry prospective of delamination in different products was also presented. The theoretical aspects have been related to physical understanding of phenomena for delamination initiation and growth. For structural health monitoring and damage detection techniques two approaches were used: the vibration – based monitoring and the fracture mechanics. The vibration – based approach involves model-based methods using low frequency vibrations, the fracture mechanics approach requires linear elastic and nonlinear concepts. In the vibration-based approach, the specimens are assumed to be free of defects: however, in numerical approaches, the stress concentration near to a notch or a flaw leads to mesh dependency. Stress criteria are needed in order to evaluate the occurrence of failure. Stress criteria require the definition of a critical crack dimension which depends on the material and stacking sequence. The main purpose of the model is to predict the deviation in materials properties if damage occurs (cracks, voids etc). The availability of a reliable model has many benefits such as the design and optimization of efficient testing configuration, the correct interpretation V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail: [email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_20, 

389

390

V. Bucur

of experimental data, the development of an inverse technique based on quantitative data and the generation of training set for neural network. Fracture mechanics approach is based on the concept of strain energy release rate and assumes the presence of an inherent defect in the specimen (a notch). Crack propagation occurs when the strain energy at the crack front is equal to the critical strain energy rate, which is a material property. Fracture mechanics approach has demonstrated satisfactory accuracy in modelling the propagation of delamination; however, in many structural applications the locus of damage initiation is not obvious. In order to overcome the referred drawbacks cohesive damage models combined with continuum damage mechanics emerge as a suitable option which is not necessary to take into account an initial defect and in the same time, mesh dependency problems are minimized. In addition to these two approaches, the analysis of cumulative damage is fundamental in life prediction of components and structure under loading (Lemaitre and Desmorat 2005). Test procedures related to delamination of wood and wood based composites were commented. The behaviour of wood and wood-based composites at different length scales was discussed. The mechanism of delamination under static loading is well understood. Models were developed to explain the delamination growth and propagation under static loading or induced by other stresses such as microwave, drying and weathering. The development of ultrasonic techniques for non-destructive inspection of structural members was emphssized. Due to the great potential of incorporating novel biomaterials or integrating nanofibers/nanoparticles advanced wood-based composites are very attractive for new structural designs and applications. However the heterogeneity of wood – based composites is the main challenge irrespective to the unique nature of the constituents in the advance-composite design. In order to avoid over – dimensioning in wood or wood-based composites design it is necessary to develop theories and analytical/numerical tools that can take into account the initiation and growth of delamination in these new advanced wood-based composites. The delamination resistant design concept applied to wood-based composites can strongly influence their performance and cost. However more research is needed in order to achieve a fully mature methodology for use in design and certification of such wood-based composites structures. If delamination onset has been successfully predicted in laboratory samples using different codes (finite elements, etc), delamination predictions using these codes need to be validated on full size structural elements by comparing field data with experimental data. Changes in temperature and hygrometry can result in significant properties variations, so laboratory simulated testing is essential both to check fabrication quality of wood-based composites and to validate design data. The anisotropic nature of wood and wood-based composites as well as their multiple failure modes, have caused major difficulties in testing procedures for products strength. With respect to initiation of delamination, their stable propagation or unstable growth, the effect of load type (tensile, compressive, biaxial, etc) the rate of load application (monotonic, quasi static, dynamic, combined, etc) and of environmental conditions (temperature, hygrometry and pressure) should be investigated.

20

From Present Researches to Future Developments

391

Damage involves mechanisms at all scales, from the molecular level revealed by sophisticated instruments to massive scale during physical inspections. The damages 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 experimental characterization as well as for modelling of delaminations. Implementation of more sophisticated approaches for mechanical behaviour of wood-based composites 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 fracture 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 particular 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 composites, the key issues are the prediction of delamination in different products and the improvement of the design of new advanced wood - based composites and structures, to prevent and minimize the products from delaminations. Without doubt, these fascinating challenges will be solved in the future by scientists and engineers 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

A Accuracy, 88, 225, 323, 363, 390 Acoustic coupling, 317 emission, 287–303, 308 imaging, 261–265 impedance, 380–381, 385 tomography, 255–266 waves, 259, 266, 311, 315, 318–320, 344 Active control, 391 Adhesion, 9, 34, 126, 145, 149, 152, 181, 329, 334 Adhesive joints, 72 Advanced materials, 3 Aging, 25–26, 44, 192 Air circulation rate, 200 Air-coupled transducers, 316, 324 American society for testing and materials (ASTM), 18–19, 24–26, 85, 87, 175, 192, 230, 289, 337 Amplitude, 41, 259, 288–289, 291–292, 294–295, 299, 301, 311–314, 317, 319–320, 324–331, 334, 336, 341, 359–360, 380, 382, 384–386 distribution, 295 Anatomic features, 124–130 Anisotropic, 3, 5–7, 9, 34, 53, 63, 75, 83, 92, 124, 131–132, 139, 187, 205–206, 211, 231, 233, 240, 242, 256, 258, 309, 318, 328, 339, 344, 390 Anisotropy, 60, 75, 78, 175, 179, 184, 193, 203–204, 206, 242, 292, 295, 317–320, 339, 385–386 Annual rings, 6, 24, 27, 78, 83, 167, 179, 181–184, 186, 193, 233, 271–272, 276, 279, 287, 290, 294–297, 303, 318, 337 Array, 6, 45, 75–76, 324 Attenuation, 34, 289, 298, 311, 323, 331, 333–339, 359, 364, 380

Automatic, 79, 176, 193, 323–324 Average frequency, 288, 302 values, 90 Axial strains, 88 Axial tension, 103 B Beam, 11, 38–44, 82, 154, 232, 244, 312, 316, 321, 325–326, 331, 334–335, 340, 342, 353–364, 366–368 Bending stress, 128, 226 Bending test, 67 Bernoulli theory, 42, 366–368 Biomaterials, 5, 145, 390 Biomechanical, 236 Boundary conditions, 104, 111–112, 199, 369–371, 373 Brittle fracture, 56, 71, 80–81, 90 Buckling, 20–23, 26, 28, 38, 76, 79, 166, 222, 235 C Calibrating, 176, 366, 368–371, 377 Cantilever, 39, 44 Capacitive sensors, transducers, 320–321, 323, 337, 344 Cellulose, 5–8, 11, 23, 63, 79, 126, 130–132, 183, 218, 233, 237, 240, 297 Chemical, 7, 9, 11, 29, 113, 118, 124–126, 138, 145, 169, 173, 175, 189, 191, 198–200, 233, 242 Chipboard, 223 Clear wood, 221, 225, 231 Climate, 6, 174–175, 178–179, 181–182, 186, 234, 270, 272–274, 277, 282, 365, 367, 372, 377 climatic cycling, 376 Coefficient of diffusion, 368–369

V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3, 

393

394 Cohesive strength, 9, 240 Collapse, 18, 87, 101–118, 159, 165, 204–205, 208, 234, 237–249, 333, 353, 363 Components, 3, 6, 29, 57, 74, 153, 161, 163, 166, 179, 182, 185, 198, 217–218, 242, 246, 299–300, 302, 321–322, 326, 390 Composites fiber reinforced, 9 wood-based, 3–5, 9–12, 17, 26–27, 29, 45, 51–92, 173–193, 288, 308–309, 311, 323, 329, 339–340, 343–344, 389–391 Compression, 5, 7, 10, 22–24, 26–27, 79, 87, 104, 125–126, 129–131, 133–134, 146, 166, 210, 219–220, 222–223, 225, 256, 261, 266, 297, 310, 332, 337 Concept, 37–38, 51, 53, 55, 152, 169, 208, 390 Constitutive law, 76 Constraints fracture toughness, 54, 60, 63, 65–67, 69–70, 72–73, 91, 136, 292, 391 mechanical, 102 Continuous emission, 288, 291–292, 302, 311 Contrast of the image, 150, 153–154, 385 Correlations, 88, 147, 149–150, 152, 155, 178, 362 Costs, 88, 199, 215–216, 222, 288, 377 Count, 288, 298–299, 302 Coupling, 176, 275, 277, 311, 317, 321, 331–332, 379 coupling media, 331–332 Crack arrest, 20, 53, 60, 64–65, 79 propagation, 7, 21–22, 54, 59–60, 63, 66, 70, 74, 87–88, 90, 124, 126, 134–136, 232, 292, 294, 297, 301–303 size, 73 Crack growth rate applied stress, 126 average crack length, 90, 179–180 crack propagation, 83, 87 damage, 83, 91 environmental effects, 233 fracture control, 88 fracture initiation, 81, 136 initiation, 91 length, 88 orientation, 85 stress intensity factor, 53–56, 60, 71, 88, 135–136, 291, 294 stress intensity rate, 88

Index Crack mouth opening displacements (CMOD), 301 Crack opening displacements (COD), 57–58, 75, 90, 292 Creeping, 366, 370, 375–377 Curing, 12, 227–232 cure, 226–229, 231 Curvature, 19, 39, 48–49, 78, 170, 218, 276, 367 Curve dispersion, 34, 37 R-curve, 57–59, 75, 89–90 D Damage evolution, 287 mechanics, 59, 390 Damping, 39–40, 321 Debonding, 4–5, 83, 87, 295, 297, 308, 337, 339–340 Decay, 22, 125, 126–128, 138, 145, 175, 178, 198, 223, 225, 232–234, 343, 358, 361 Decibels (dB), 292, 295, 328–329, 335, 342, 380–382, 385–386 Defect, 18–19, 26, 29, 34–36, 57, 90, 159, 165–166, 168, 197, 204–206, 217, 221, 223–228, 231, 259, 308, 311, 313–316, 320, 323, 326–328, 332–334, 336, 339–341, 344, 354, 358–363, 379, 380–382, 385, 389–390 size, 259, 334 Density, 9–11, 25–27, 28, 41, 53, 58, 60, 65, 69, 90–91, 112, 114–115, 133, 135–136, 160, 162–163, 179–180, 183, 205, 208, 218, 221–223, 227, 229, 237–238, 242, 246–247, 250–252, 276, 278, 309, 317, 323, 336–338, 343, 358, 361, 366, 368–369, 380–381 Dielectric properties, 160, 162–163, 169 Disbonds, 311, 328, 380, 385–386 Dislocation, 21–23, 151–152, 155, 256–257, 261, 288 Distortional effect, 368, 370, 377 Double cantilever beam (DCB), 73 Drying air, 200–203, 208 kiln, 200–203 Ductile fracture, 80–81 Duration, 56, 165, 175, 177, 217, 232, 288–289, 299 Dynamic fiber reinforced composites, 9 fracture, 391 toughness, 391

Index E Ecological relevance, 91 Effective crack length, 88 Elastic constants, 8–9, 55, 78, 113, 115, 309, 311 Elasticity, 10, 91, 116, 270, 276, 283, 291, 337, 339, 358, 361, 366 Elliptical, 242 Energy balance, 54 Energy release rate, 38, 54–55, 58–59, 72–73, 88, 90, 390 Engineered wood products (EWP), 216–217, 226–232, 235 Engineering materials, 26, 45 Errors, 4, 8 Event, 24, 45, 64, 288–289, 291–294, 297–301, 308 Experimental determination of fracture toughness, 391 stress intensity factor, 391 Exposure artificial, 178, 186–192 outdoor, 149, 174–175, 178–186, 191–192 F Failure modes, 91, 288, 366, 390 stress, 211, 389 Fatigue tests, 293 Fiber board, 229, 308 pullouts, 308 Finite ABAQUS, 365 element method, 43, 116, 366, 376 finite element analysis (FEM), 27, 33, 38, 45, 76, 78, 102–103, 112, 117, 372–373 Flaws, 34, 53, 308, 313, 320 Flexural, 39, 42, 310–311, 316–319 Flexural vibration, 39 Focused air-coupled transducers, 324 Fourier Transform Technique, 35, 37 Four-point-bending test, 367 Fractography, 80, 301 Fracture process zone, 56–57, 65, 72, 74, 92 Frequency natural, 343 spectrum, 327 Fundamental frequency, 327

395 G Gain, 34, 201, 217, 308, 320, 382, 386 Gap opening, 368, 370–371, 374–376 Geometry, 11, 54, 60, 62, 87, 105, 111, 113, 163, 167, 244, 329, 333–334, 337, 339, 341, 366–371, 373–374, 376–377, 381 Glue, 9–10, 24–25, 126, 134, 138, 192, 199, 227, 277, 283, 331, 340, 358–359, 367–368, 371, 374–377, 379–386 Glued laminated timber – glulam, crosslam, glulam slabs, 307 Grading, 221, 224–225, 358, 363 Grain angle effect, 89 slope, 166 Griffith, 52 Griffith theory, 52 Growth rings, 18, 78, 82, 128, 206, 222, 225, 242, 252, 272, 278, 283, 369, 373, 377 Guided waves, 320 H Hardness, 138, 178, 252, 297 Hardwood, 6, 11, 126, 130, 161–162, 164, 176, 185, 207–208, 211, 220, 222–223, 230, 235, 291–292, 362, 380 Harmonic, 39, 290, 328–329, 341–342 Health monitoring, 3, 12, 27, 29, 33–34, 37–38, 40, 44–45, 64, 389, 391 Heat transport, 160 Hemicelluloses, 7 Hidden specimens, 220–221, 225, 299–300 High resolution, 24, 138–139, 145–155, 386 Hook’s Law, 311, 344 Humidity, 4, 12, 26, 80, 136, 173, 177–178, 189, 198–203, 206, 233, 239, 244, 270, 273, 275, 308, 343, 366–367, 369, 377 I Impact testing, 223, 308 Impregnation, 159, 163–166, 169, 199 Infrared spectroscopy, 332 Inhomogeneities, 385 Initiation, 5, 20, 26, 29, 51–91, 135–136, 183–185, 218, 232, 235, 242, 291, 293, 295, 303, 356, 389–391 In-plane, 317, 320, 336 In-service damage, 308 In-situ, 64, 69, 82–83, 134, 288, 295, 297, 358, 391 Inspection continuous, 45 one side, 316 two side, 316

396 Integrity, 3, 25, 29, 34, 45, 137–138, 232, 308, 354, 386 Interface, 4, 9, 11–12, 24, 34–35, 42–43, 58–59, 74, 87, 127, 132–133, 145–156, 165, 295, 318, 321 Interface wood-cement, 12 Interfacial cracking, 59 Interfacial layers, 368 Interlaminar fracture toughness, 391 Inverse Fourier transform, 35, 37 Irwin lamellar tearing, 55 Iterative, 110 J J-integral, 57–59 Joints, 24–25, 72, 84, 87, 192, 226 K Knots, 25, 166, 204, 217, 221, 223, 225–226, 231, 380–381 L Lamb wave, 33–35, 37–38, 45, 309–311, 317, 339–341, 344 Laminated veneer lumber (LVL), 3, 9–10, 12, 19, 29, 227, 229, 307, 343–344 Laminates, 4, 19, 355, 379–391 Laser scanning microscopy, 145, 147–152, 295 Laser vibromerty, 39, 319, 380 Lateral, 104, 166, 220, 260, 262, 266, 323, 327 Leaky surface acoustic waves (SAW), 318–320 Levels, release rate, 38, 54–55, 58–59, 72–73, 88, 90, 390 Life prediction, 390 Lignin, 6–7, 63, 91, 126, 130, 135, 173, 181, 185, 187, 189, 193, 218–220, 233, 235 Linear elastic fracture mechanics (LEFM), 53, 56, 60, 74, 92 Load displacement curves, 59, 64, 67, 74–75, 79, 134, 297 modes, 64 rate, 38, 60, 71–72 Local, 20, 23–24, 27, 29, 33, 37, 40, 54, 71, 130, 136, 185, 241, 256, 276, 288, 317, 328, 341, 366, 370, 373 residual stress, 3, 87–88, 112 Localization, 287, 289, 292, 303, 313 Longitudinal, 6–7, 18, 20, 41, 53, 77, 79, 113, 115, 125, 129, 131–137, 159, 161–162, 164–165, 169, 186, 205–206, 217–220,

Index 233, 256, 258, 261, 269, 282–283, 309, 313, 316, 322, 366–367, 369–370, 380 Low frequency vibrations, 33, 38, 389 Low temperature, 137, 208–210 Lumber, 3–4, 9–10, 18–19, 25, 191, 215–217, 221–227, 231, 233–234, 255–256, 297, 300, 307, 324, 380 M Machining, 87, 149, 252, 391 Macrocracks, 56, 64, 291 Main, 19, 35, 45, 56, 64, 69, 87, 112, 116, 139, 147, 152–153, 162–163, 179, 192, 206, 220–221, 228, 240, 242, 246, 251, 256, 269, 289, 302, 308–309, 311–312, 322, 333, 337, 343, 355, 365–366, 375, 383, 386, 389–390 amplifier, 312 Manufacturing defects, 223–224 process, 28–29, 223–230 Mass production, 3 Material properties, 77–78, 101, 103–105, 111–112, 115–116, 251, 367, 377 Maximum, 64, 66–67, 72, 134, 136–137, 168, 178, 185, 198, 202, 206–208, 210–211, 216, 221, 239, 246, 279, 291, 293, 300, 325, 329, 353, 374 Mechanical performance, 353 properties, 3, 6–11, 34, 39, 78, 113, 116, 166, 175, 193, 205, 225, 233, 240, 292, 307, 329 Mechanical properties of cell wall, 7, 113, 240 Medium density fiberboard (MDF), 26, 85, 87–90, 229, 340–341, 343 Microcracks, 20, 56, 59, 64, 69, 127, 133, 136, 185, 291–293, 295, 298, 337, 380 Microdefects, 35 Microfibril angle, 6, 79, 82, 113, 129–131, 134, 219, 240, 242, 260 Microscopy confocal, 134, 138 electron, 81, 131, 136, 138–139, 146, 149–152 light, 9, 21–22, 24, 131, 137–138, 147, 149–152 Microstructural, 7, 56, 75, 91 Micro-voids, 149, 159, 165 Microwaves, 159–170, 200, 227, 229, 235, 343, 390 Mineral bonded particleboard and fiberboard, 308

Index Modal analysis, 39, 42 Mode I, 52–53, 60, 62–67, 69, 71–72, 74–75, 80, 84–85, 91, 290, 301, 303 Mode II, 52–53, 60, 62–65, 69, 71, 74 Model analysis of structural, 33 circular based, 102, 104–106 collapse recovery, 101–119 linear behaviour, 38–40 local and global information, 27 mathematical, 38, 205–211, 240 model-based methods, 38–44, 389 nonlinear behaviour, 40–44 squared based, 106–117 Modeling, 8–9, 35, 41, 75, 77, 92, 168–169, 205–206, 297 computational modelling, 5, 9 Modulus of elasticity, 91, 116, 291, 337, 358, 361, 366 Moisture, 136–137, 174, 178, 207, 209, 270–272, 274–278, 297–299, 365–377 Monotonic loading, 59 Monte Carlo simulation, 10 Morphology, 76, 80–82, 84, 91, 134 Multi-layered materials, 6 Multiple delaminations, 29, 39, 42 N Natural defects of wood knots, 204, 225 slop of grain, 225 Neural network, 44, 289, 299, 390 Nominal stress, 53, 295 Non-contact, 39, 311, 316–320, 323–324, 329, 334–344 ultrasonic transducers, 320, 323, 329, 334–343 Nonlinear acoustic modulation, 290 nonlinear behaviour, 40–44, 72, 328, 344 Nonlinearity, 50, 349 Nonparametric models methods, 44 Notch effect, 129 Nuclear magnetic resonance (NMR), 9 Numerical analysis, 365, 368, 372 calculation, 377 O Opening, 24–25, 41, 52, 56–59, 64, 73–75, 90, 92, 134, 228, 235, 239, 292, 368, 370–371, 374–377 mode stress intensity, 56

397 Operational, 44–45, 153, 215 Orientation, 5, 11, 24, 53, 60, 64–67, 71, 75, 78, 85, 87, 90, 92, 102, 106, 113, 128, 130–131, 136, 152, 167, 184, 186–187, 218–219, 223, 230–231, 233, 243, 256, 266, 272, 276, 278, 283, 289, 294, 303, 313, 318, 367–368 crack, 64, 90 OSB oriented strandboards, 336 Out of plane, 52–53, 71, 92 Overall mechanical characterization, 34 P Parameters, 3, 5, 34, 39, 54, 55, 59, 64–65, 70, 72, 76, 79, 91–92, 103, 112, 162–163, 167, 169, 189, 191, 275–277, 280, 282–283, 288–293, 299–300, 302, 309, 313, 321, 324, 327, 329–330, 332, 336–337, 339, 361–363, 367–368 linear fracture mechanics, 72, 92 Parquet cupping, 373–374 deformation, 377 floors, 365–377 gap opening, 370, 374, 376–377 geometry, 366–371, 373–374, 376, 377 parquet lacquered specimens, 367, 373 parquet planks, 366–368, 370, 376–377 simulate the behaviour of parquet planks, 368 Particle based boards – oriented strandboards, particle board, fiber board, 308 Pattern recognition, 39, 297 Peak amplitude, 289, 299, 329, 336 Periodical, 39, 45, 175 Phase velocity, 34–35, 340 Physical methods, 45, 52, 87–88 Piezoelectric actuators, 40 sensors, 40 Pin contact forces, 42 Planar, 52, 56, 382 Plane stress, 54–55 Plane wave propagation, 34 Plastic deformation, 56, 59, 131–132 strains, 104–106, 137, 337 zone, 53–54, 56 Plasticity, 88 Plate wave technique, 329, 339–342, 345 Plywood, 9–10, 12, 19, 24–25, 28, 85, 149–155, 174, 191–192, 216, 227, 229, 339, 343

398 Point source, 36, 201 Poisson’s ratios, 8, 271, 276, 280 Polymers, 103, 119, 217–218 Poor cure, 227 Porosity, 34 Prediction of life, 390 Pressure, 18, 28, 104–105, 108, 111–112, 159–163, 167, 192, 200–201, 215, 218, 227–228, 230, 233, 235, 238–239, 272, 379, 380, 390 sound, 380 Principle of superposition, 38 Processing, 12, 40, 119, 150, 160, 162, 167, 169, 198, 206, 220–221, 224, 228, 237–253, 256, 288–289, 299, 303, 308, 343, 391 internal checking, 220, 237–253, 343 Production process, 11, 169, 217, 227, 229, 236, 308, 323, 343 Propagation, 5, 7, 20–22, 34–35, 53–54, 56–60, 63, 64, 66, 70, 74, 82–83, 87–88, 90, 92, 124–126, 134–136, 175, 183–184, 217–218, 221, 232, 235, 255–256, 289, 291–292, 294–297, 301–303, 309–313, 317, 319, 322, 327, 336, 339, 356, 363, 375, 381–383, 385, 390 Pulp, 11, 22, 92, 133, 135, 216–217, 233, 246, 256, 391 P wave, 314 Q Q factor, 193 Quality assessment, 3, 256, 316, 379 Quantitative, 3, 34, 75, 148, 192, 246, 288, 303, 390 R Radial, 6–7, 18–19, 23, 25, 53, 75, 77, 81–84, 103, 124–131, 134–138, 159, 161–162, 164–165, 167, 169, 179–180, 183–185, 188, 205–206, 216–262, 217–221, 233–234, 242, 244, 256–259, 265–266, 269–270, 275, 279, 282–283, 295, 302, 367–369 Radiations, 126, 145, 160, 167–168, 173, 177–178, 185–189, 233, 290, 317, 321 Radiographic, 303 Ratios, 8, 180, 271, 276, 279–280, 294, 317 Raw material, 9, 11, 215–217, 220–224, 227, 236, 307 Rayleigh, 320 R-curve, 57–59, 75, 89–90 Reaction wood, 129–130, 179, 193, 220, 256, 261, 266, 297

Index Reconditioning, 101, 112, 118, 165, 227–228, 247–250, 252 Reference, 11, 18, 20, 34, 39, 51, 54, 63, 74–75, 103–104, 175–176, 191, 240, 255, 261, 281–282, 309, 311, 320, 325, 329, 371 stress, 18, 20, 34, 54, 74–75, 92, 104, 175, 240, 261, 282, 309, 311, 320, 371 Reflection, 34, 176, 193, 311, 315, 318, 326, 343 elastic waves, 34, 311 Refraction, elastic waves, 311 Relative humidity, 26, 80, 173, 189, 199–200, 202, 206, 239, 244–245, 270, 273, 275, 366–367, 369, 377 Reproducibility, 323, 332, 385–386 Residual strength, 363 stress, 3, 87–88, 112, 363, 391 Resin pockets, 125, 128, 166, 381–382, 384–385 Resistance, 3, 11, 24, 54, 57–60, 66, 75, 82, 88, 127–128, 132–133, 135, 191, 232, 234, 244, 270, 275, 307, 361 Resonance, 316–317, 343–345 Restraint cracking, 220 Review, 39, 45, 51, 66, 155, 160, 169, 175, 191, 217, 242, 290, 308, 344, 389 Rise time, 288, 299, 302 Rock, 219 Rods, 77, 323, 328 Root mean square (RMS) voltage, 288, 302, 382 Roughness, 75, 83, 146, 153, 292 Round wood, 234 Rupture, 7, 25, 53, 56, 59, 72, 83, 160–161, 166, 168–169, 204, 231, 293, 295–296, 337, 361–362 S Safe design, safety, safe performance, 5, 9, 44, 91–92, 229, 232, 235, 308, 343, 345, 391 Safety factors, 91–92 Samples, 82, 112, 127, 133–139, 165, 168, 198, 225, 240–241, 246–248, 251, 260, 271–273, 275, 278, 366–367, 372–373, 381–382, 385–386, 390 Sampling rate, 382 Scanning electron microscopy (SEM), 21–24, 62, 72, 131, 134, 136–139, 146, 149–155, 219, 297 Scanning modes, 325

Index Scattering, 261 Serviceability of structures, 3 Shape change, 355 Shear bands, 129 modulus of elasticity, 116, 358, 361, 363 waves, 309–310, 316–317, 359, 364 strains, 79, 105 Shrinkage, 18, 25, 29, 175, 179–180, 183, 185, 192, 198, 203–207, 219–220, 231–233, 237, 241–244, 246–252, 271, 367 Simulation, simulated, 7, 9–10, 13, 26, 35, 38, 42, 58–59, 64, 74–75, 77, 88, 112, 160, 162, 168–169, 175, 206, 270, 278–280, 282–283, 290, 331, 337, 339–340, 342, 358, 368–372, 376, 390 Size, 10, 53–54, 66, 72–73, 78, 82, 87, 146, 165–169, 176, 218, 225–226, 230–231, 239, 247–248, 256, 259–260, 313, 323, 332–335, 343, 345, 361, 363, 366, 374, 390 Smart composites, 43–44 Softwood, 6–7, 11, 53, 63, 76, 128–129, 132–133, 136, 161–162, 164, 176, 181, 192–193, 198, 202, 208, 211, 218, 220, 230, 234, 279, 291–292 Solid state adhesive layer, 12 Species, 6, 21–24, 53, 60, 63–65, 67, 69–70, 87, 114, 124, 126, 130, 133, 135–136, 159, 161–162, 164, 166–167, 174–175, 178–180, 184–185, 187, 189–191, 198, 203, 205, 207, 216, 218, 221–223, 234–235, 237–240, 242, 246–248, 290–294, 303, 320, 358, 361, 366, 374 Spectral analysis, 91 Spectroscopy, 332 Split, 19, 24, 62–64, 67, 74, 83–85, 88, 92, 125, 159, 165–166, 168, 174, 179, 191, 197, 203, 204, 208, 211, 217–218, 220–226, 231, 233–234, 236, 272, 290, 380 Spring, 6, 41–42, 77, 204, 206, 219–220, 235 Stability, 26, 56, 71, 113, 137, 149, 198, 226, 241, 269, 279, 307, 353, 363, 370–371 Static curves, 293, 337 loading, 390 tests, 293 Statistical based methods, 44 Stiffness, 5, 7, 10, 34–35, 39, 41, 52, 59, 75, 77, 79, 101–102, 135, 225–226, 228, 241–242, 252, 278–279, 283, 292, 303, 358, 363, 369

399 Strain basic concept, 211 curves, 88, 134, 232, 294 energy, 38, 54–55, 58, 60, 301, 390 energy release rate, 38, 54, 55, 390 field, 54–55, 233, 390 hardening, 337 Strength, 5–7, 9–10, 26, 52–53, 59, 74–75, 77, 87, 91, 116, 129, 134, 136, 149, 152, 192, 199, 205, 210–211, 217, 222, 225–226, 228, 231, 233–234, 238, 240–241, 261, 288, 290–291, 308, 316, 358–364 Stress concentration, 72, 225–226, 295, 297, 389 distribution, 54, 56, 60, 71, 83, 107, 110, 112–115, 118, 125, 232, 282–283, 377 distribution in glue line, 377 field, 54–55, 183 stress intensity factor, 53–56, 60, 71, 88, 135–136, 291–294 Structural integrity, 137–138, 232, 308 Structures, 5–7, 9, 12, 18–19, 24, 26, 29, 34, 38–40, 42, 44–45, 53, 59, 63–64, 69, 75–79, 81–84, 102, 124, 126, 131, 133, 135–137, 139, 159–161, 163–164, 166–170, 179, 181, 183, 185, 187, 189, 193, 207, 217–220, 230–232, 234–235, 255–256, 261–262, 284, 287–303, 308, 313, 316, 326, 337, 344, 354–355, 358, 360–361, 363, 370, 379, 385–386, 390 Substrate, 9, 59, 152 Surface, 18, 24, 149–152, 179, 209, 211, 238, 241–246, 250, 262, 299, 366, 374, 376 T Tangential, 6–7, 18, 53, 57–59, 81–83, 103, 124–129, 131–132, 134–135, 137–138, 166–167, 169, 179–180, 183–187, 189, 205–206, 219–220, 233, 242, 244, 258, 260, 262–266, 269, 271–272, 275, 279, 292, 295, 328, 367–369, 373, 377 Temperature, 4, 12, 26, 87, 101–105, 110, 112–115, 118, 137, 152, 160–163, 174, 177–178, 181, 198–203, 206–211, 218, 223–224, 227–230, 232–233, 235, 238–241, 246, 248–252, 297, 308, 343, 367–368, 370–371, 377, 390 Tensile loads, 131, 134, 294 stress, 74, 81, 175, 182, 205, 210, 297, 328

400 Tension, 5, 18, 22–23, 27, 41, 52–53, 63, 65, 67, 71, 79, 81–85, 87, 103–104, 110, 130–131, 134, 136, 165–166, 204, 210, 219–220, 225–226, 231, 238–241, 244–245, 255–266, 290, 292, 294–295, 297–298, 303, 310–311 Tests, 25–26, 62, 64, 67, 69, 84, 88–89, 192, 198, 208, 223, 240, 261, 271–275, 279, 288, 290, 301–303, 312–313, 321, 322–323, 325, 332, 340, 354 testing configuration, 329–331, 389–390 Theoretical, 3, 7, 34, 45, 52, 57–58, 64–65, 72, 92, 101–119, 219, 289, 297, 309, 320, 334, 380–381, 389 Thermal stress, 297 waves, 161, 297 Thermo mechanical analysis, 135, 391 Thickness, 19, 25, 28–29, 34, 36, 39, 89–90, 150–152, 174, 176, 193, 203, 206–207, 227–228, 244, 248, 250–252, 271–274, 276, 279–280, 284, 295, 310–311, 313, 317–318, 320–321, 323, 326, 331–334, 337, 340–343, 356, 358, 367–368, 373–374, 382–383, 385, 391 Three dimensional systems, 270 three or Tri or 3D non-linear stochastic finite element model, 10 three point bending test method, 67, 366 Threshold, 87, 208, 211, 241, 288–289, 302, 343, 391 Timber, 9, 12, 25, 29, 92, 113, 128, 159–170, 186, 192, 197–211, 215–216, 221, 235, 237, 241, 246–248, 252, 266, 269–284, 307, 331–332, 379–386 Time reversal concept, 37 Tomography, 90, 255–266 Toughness, 5, 25, 53–54, 56, 59–60, 63, 65–67, 69–70, 72–73, 75, 79, 90–91, 127, 136, 222, 232, 292, 391 Transform Fourier transform, 35, 37 Transient, 35, 288 effect, 288 Transition, 60, 81–82, 129–130, 132, 137, 229, 235, 296, 381, 383–384 Transmission electron microscopy (TEM), 81, 128, 131–134, 138–140 Travel time, 326

Index U Ultrasonic nondestructive evaluation - local damage information, 33, 45, 288, 328 path, 260, 333 Ultrasonic techniques contact techniques, 334–340, 343, 379, 386 non-contact, 329, 334–342 pulse echo, 312–314, 344 through transmission, 312–314, 324, 329, 380, 386 Ultrasound, 165, 229, 316–319, 323, 337, 340–343, 380, 385–386 Ultrastructural features, 130–137 Uncertainties, 260, 385 Uniaxial loading, 103 Uniform stress field, 83 V Validity tests, 336 Velocity plate wave, 317–318 stress wave, 258, 262–264, 266, 361 surface wave, 320 ultrasonic, 256, 258, 260–262, 266, 332 Veneer, 3, 9–10, 12, 19, 24, 192, 221, 227–229, 256, 307, 318–320, 337, 339–342, 344, 366, 380 Veneer based panels – plywood, laminated veneer lumber, 307 Verification tests, 368 Vibration vibration based damage identification, 34 Viscoelasticity, 339 Viscoelastic properties of wood, 56, 60, 116 Voids, 25, 27–28, 147–149, 151–153, 155–156, 159, 161, 165, 167, 226, 229, 308, 336–337, 389 Volume, 60, 71, 163, 169, 179, 206–207, 220–221, 270, 274–276, 278, 314, 325 Volumetric strain, 206 W Warping, 167, 174, 191, 204, 206, 269–284 Waves modes, 34, 309–310, 323 plane, 34, 261, 317, 380 surface, 260, 309, 311, 318–320, 357 ultrasonic, 34, 176, 256–257, 261, 311, 313, 317, 336, 380, 385 Weak bond, 34, 229 Weathering, 125–128, 145, 173–193, 233, 390 Wedge loads, 64, 67, 74, 88

Index Wood based composite panels, 192, 308, 311, 329–344 Wood coating interface, 145–156 Wood flooring systems, 365, 377 Wood-plastic composites, 311 Wood species, 6, 22–24, 53, 60, 126, 130, 159, 161–162, 164, 166, 167, 175, 178, 185, 205, 223, 234–235, 292, 361, 366

401 Y Yield, 53, 92, 146, 153, 206 strength, 53 Young’s modulus, 7–8, 10, 55, 260, 317, 337 structural modifications, 7 Z Zone plastic deformation, 56, 59