See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/301355302

Wood-Based Panels: An Introduction for Specialists

Book · August 2010

CITATIONS
READS
37
6,216

15 authors, including:

Mark Irle
Peter Niemz
Ecole Supérieure du Bois
ETH Zurich
99 PUBLICATIONS 936
CITATIONS 477 PUBLICATIONS 5,351 CITATIONS

Luisa Hora Carvalho

Jorge Manuel Martins
Polytechnic Institute of Viseu
Polytechnic Institute of Viseu
91 PUBLICATIONS 674 CITATIONS
73 PUBLICATIONS 362 CITATIONS

Some of the authors of this publication are also working on these related projects:

FCT: PTDC/EQU-EQU/111571/2009 “ECOUF Development of a new generation of urea-formaldehyde resins with low formaldehyde emission View project

Material analysis of wood-based products using computed tomography View project

All content following this page was uploaded by Mark Irle on 18 April 2016.

The user has requested enhancement of the downloaded file.

 Wood-Based Panels
An Introduction for Specialists

Wood-Based Panels
An Introduction for Specialists
Edited by

Heiko Thoemen
Mark Irle
Milan Sernek

Brunel University Press

Edited by
Heiko Thoemen
Mark Irle
Milan Sernek
 Contents

Contents

List of Contributors....................................................................................ix
Preface.....................................................................................................xiii
1 Wood-Based Panel Technology............................................................1
Chapter Summary...................................................................................1
1.1 Introduction.....................................................................................1
1.2 Why Make Wood-Based Panels?....................................................3
This publication is supported by COST
1.3 The Manufacture of Particleboards: A Short Overview..................5
1.3.1 Defining Particleboard........................................................5
Published by Brunel University Press, London, UB8 3PH. England
1.3.2 Wood as a Raw Material.....................................................5
This book may be cited as: Wood-Based Panels - An Introduction for 1.3.3 Particle Production............................................................12
Specialists
1.3.4 Particle Drying..................................................................18
1.3.5 Particle Sorting.................................................................22
ISBN 978-1-902316-82-6
1.3.6 Resin Metering and Blending...........................................25
1.3.7 Common Adhesives..........................................................31
No permission to reproduce or utilise the contents of this book by any means
is necessary, other than in the case of images, diagrammes or other material 1.3.8 Mattress Forming..............................................................33
from other copyright holders. In such cases, permission of the copyright 1.3.9 Mattress Pre-Pressing and Pre-Heating............................36
holders is required.
1.3.10 Pressing.............................................................................38
Neither the COST Office nor any person acting on its behalf is responsible 1.3.11 Processing Steps Immediately after Pressing....................47
for the use which might be made of the information contained in this
publication. The COST Office is not responsible for the external websites 1.4 The Manufacture of Oriented Strand Board: A Short Overview . 55
referred to in this publication.
1.4.1 Introduction.......................................................................55
© COST Office, 2010 1.4.2 Manufacture of OSB.........................................................56
Cover photos: Courtesy of Dieffenbacher GmbH, Eppingen, Germany
Layout: Henrik Schmidt, Wentorf, Germany
1.5 The Manufacture of Medium Density Fibreboard: A Short
Overview.......................................................................................61
1.5.1 Introduction.......................................................................61
1.5.2 Manufacture of MDF........................................................62
1.6 The Manufacturing of Plywood: A Short Overview....................74
1.6.1 Introduction.......................................................................74
1.6.2 The Manufacturing Steps..................................................75
1.6.3 Peeling...............................................................................78
1.6.4 Dryers................................................................................82
 iii
Contents Contents

1.6.5 Veneer Preparation...........................................................82 3.3 Heat and Mass Transfer Mechanisms in Hot Pressing................138
1.6.6 Board Layup......................................................................83 3.3.1 Introduction.....................................................................138
1.6.7 Pressing.............................................................................84 3.3.2 Heat Transfer..................................................................140
1.6.8 Finishing............................................................................85 3.3.3 Mass Transfer.................................................................142
1.7 A Potted History of Wood-Based Composites.............................85 3.3.4 Internal Mat Conditions..................................................144
1.8 References.....................................................................................90 3.3.5 Transport Properties........................................................150
2 Water Absorption of Wood and Wood-Based Panels – 3.4 Heat and Moisture Transfer in Conditioning..............................160
Significant Influencing Factors..............................................................95
3.4.1 Introduction.....................................................................160
Chapter Summary.................................................................................95
3.4.2 Internal Conditions during Conditioning.......................160
2.1 Introduction...................................................................................95
3.4.3 Relevant Heat and Mass Transfer Mechanisms..............162
2.2 Sorption Behaviour and Capillary Water Absorption...................98
3.5 Diffusion of chemicals................................................................163
2.2.1 Ultimate State of Wood-Water.........................................98
3.5.1 Introduction.....................................................................163
2.2.2 Sorption Behaviour...........................................................99
3.5.2 Diffusion of Resins.........................................................163
2.2.3 Water Absorption by Capillary Forces...........................102
3.5.3 Diffusion of Wax and Other Additives...........................164
2.2.4 Swelling and Shrinkage..................................................103
3.6 References...................................................................................167
2.2.5 Influence of Wood Moisture on the Surface Roughness
4 Advanced Imaging Techniques in Wood-Based Panels
of Wood-Based Materials...............................................108
Research.............................................................................................177
2.3 The Relationship between Moisture Content and Properties......110
Chapter Summary...............................................................................177
2.4 References...................................................................................114
4.1 Introduction.................................................................................177
2.5 Annex: Tables Listing Selected Properties of Wood and
4.2 Virtual Prototyping: Opportunities and Challenges....................180
Wooden Materials.......................................................................116
4.2.1 Levels of Complexity in the Internal Microstructure
3 Transport Phenomena..........................................................................123
of the Composites............................................................180
Chapter Summary...............................................................................123
4.2.2 Compatibility between Testing Approaches and
3.1 Heat and Mass Transfer Mechanisms in Porous Media..............123 Numerical Models...........................................................181
3.1.1 Introduction.....................................................................123 4.3 The Advanced Imaging Techniques...........................................182
3.1.2 The Heat Transfer Mechanisms......................................124 4.3.1 Digital Image Analysis...................................................183
3.1.3 The Mass Transfer Mechanisms.....................................126 4.3.2 Multi-Scale and Multi-Modal Correlation.....................189
3.2 Heat and Moisture Transfer in the Drying of Particles/Fibres....129 4.3.3 Optical Measurement of Deformations and Strains........191
3.2.1 Introduction.....................................................................129 4.3.4 Inverse Problem Approach.............................................191
3.2.2 Drying Regimes in Convective Drying..........................131 4.3.5 New Developments in Modeling....................................193
3.2.3 Heat Transfer..................................................................132 4.4 References...................................................................................194
3.2.4 Moisture Movement........................................................133
3.2.5 Effects of Wood Chip Properties....................................136
iv v
Contents
Contents

5 Adhesive Bond Strength Development............................................203

6.4.3 Applications of NMR in the Wood Based Panel
Chapter Summary...............................................................................203 Industry...........................................................................245
5.1 Introduction.................................................................................203 6.5 References...................................................................................248
5.2 Fundamentals...............................................................................204 7 Carbon Materials and SiC-Ceramics made from Wood-Based
5.3 Monitoring the Development of Adhesive Bond Strength..........208 Panels..................................................................................................251
5.3.1 Thermomechanical Analysis (TMA)..............................209 Chapter Summary...............................................................................251
5.3.2 Dynamic Mechanical Analysis (DMA)..........................210 7.1 Introduction.................................................................................251
5.3.3 Torsional Braid Analysis (TBA)....................................211 7.2 Specific Wood Based Materials..................................................252
5.3.4 Automated Bonding Evaluation System (ABES)...........213 7.2.1 Raw Materials and the Manufacturing Process...............253
5.3.5 Integrated Pressing and Testing System (IPATES)........216 7.2.2 Characterisation...............................................................253
5.3.6 Other Techniques............................................................217 7.3 Carbon Materials.........................................................................254
5.4 Open Questions...........................................................................219 7.3.1 Cabonization...................................................................255
5.5 References...................................................................................221 7.3.2 Characterisation...............................................................256
6 Innovative Methods for Quality Control in the Wood-Based 7.3.3 Potential Applications.....................................................256
Panel Industry....................................................................................225 7.4 SiC-Ceramics...............................................................................257
Chapter Summary...............................................................................225 7.4.1 Siliconisation...................................................................258
6.1 Introduction.................................................................................225 7.4.2 Characterisation...............................................................259
6.2 Infrared Thermography...............................................................226 7.4.3 Potential Applications.....................................................259
6.2.1 Principles of Infrared Thermography.............................226 7.5 References...................................................................................261
6.2.2 Infrared Cameras.............................................................228 8 Thermally and Chemically Modified Wood-Based Panels............265
6.2.3 Thermography for Non-Destructive Testing...................229 Chapter Summary...............................................................................265
6.2.4 Application and Examples..............................................231 8.1 Introduction.................................................................................265
6.3 Nar Infrared Reflectometry.........................................................236 8.2 Chemical Modification...............................................................267
6.3.1 Origin of NIR Spectra.....................................................236 8.3 Thermal Modification.................................................................271
6.3.2 NIR Reflectometry..........................................................237 8.4 Modification of Panels................................................................275
6.3.3 Panel Properties and NIR Spectra...................................238 8.5 Conclusion...................................................................................276
6.3.4 Multivariate Techniques for Data Analysis....................239 8.6 References...................................................................................278
6.3.5 Technical Aspects...........................................................239
6.3.6 Applications and Examples............................................240
6.4 Nuclear Magnetic Resonance.....................................................244
6.4.1 Nuclear Spin...................................................................244
6.4.2 Panel Properties and Nuclear Spin.................................245
vi vii
 Contributors

List of Contributors
Chapter 1 Wood-Based Panel Technology
Mark Irle
Ecole Supérieure du Bois
BP 10605 - Rue Christian PAUC
44306 Nantes
France
mark.irle@ecoledubois.fr
Marius C. Barbu
Faculty of Wood Industry
University "Transilvania" Brasov
Str. Colina Universitatu nr. 1
500084 Brasov
Romania
cmbarbu@unitbv.ro

Chapter 2 Wood-Water-Interaction

Peter Niemz
Institute for Building Materials (IfB), Wood Physics
ETH Zurich,
Schafmattstr. 6
8093 Zurich
Switzerland
niemzp@ethz.ch

Chapter 3 Transport Phenomena

Luisa Carvalho
Department of Wood Engineering
Polytechnic Institute of Viseu
Campus Politécnico de Repeses
3504 510 Viseu
Portugal
lhcarvalho@demad.estv.ipv.pt
Jorge M. S. Martins
Department of Wood Engineering
Polytechnic Institute of Viseu
Campus Politécnico de Repeses
3504 510 Viseu
Portugal
jmmartins@demad.estv.ipv.pt

viii ix
Contributors
 Austria manfred.dunky@boku.ac.at
Carlos A.V. Costa
Laboratory of Process, Environment and Energy Engineering
Faculty of Engineering
University of Porto
Rua Dr. Roberto Frias s/n
4200-465 Porto
Portugal
ccosta@fe.up.pt

Chapter 4 Imaging Techniques

Lech Muszynski
Department of Wood Science and Engineering
Oregon State University
119 Richardson Hall
Corvallis, OR 97331
USA
lech.muszynski@oregonstate.edu
Maximilien E. Launey
Lawrence Berkeley National Laboratory Materials Sciences Division
University of California Berkeley
1 Cyclotron Road, MS 62 203
Berkeley, CA 94720
USA
melauney@lbl.gov

Chapter 5 Bond Strength Development

Milan Sernek
Biotechnical Faculty
University of Ljubljana
Rozna Dolina, Cesta VIII/34
1000 Ljubljana
Slovenia
milan.sernek@bf.uni-lj.si
Manfred Dunky
Institute of Wood Science and Technology
BOKU-Vienna
Peter Jordan Strasse 82
1190 Vienna
 Contributors Wessling Group
Forstenrieder Str. 8-14,
Chapter 6 Qualitiy Control 82061 Neuried
Germany
Jochen Aderhold olaf.treusch@wessling.de
Fraunhofer Institute for Wood
Research Wilhelm-Klauditz-Institut, Chapter 8 Modified Wood-Based Panels
WKI Bienroder Weg 54E
Martin Ohlmeyer
38108 Braunschweig
Institute of Wood Technology and Wood Biology (HTB)
Germany
Johann Heinrich von Thuenen-Institute (vTI)
jochen.aderhold@wki.fraunhofer.de
Leuschnerstrase 91
Burkhard Plinke 21031 Hamburg
Fraunhofer Institute for Wood Germany
Research Wilhelm-Klauditz-Institut, martin.ohlmeyer@vti.bund.de
WKI Bienroder Weg 54E
Wulf Paul
38108 Braunschweig
Sonae Industria (UK) Ltd
Germany
burkhard.plinke@wki.fraunhofer.de Moss Lane, Knowsley Industrial Park
L33 7XQ Knowsley
Chapter 7 Wood Panel-Based Ceramics United Kingdom
w.paul@sonae.co.uk
Olaf Treusch

x xi
 Preface

Preface
Wood-based panels is a general term for a variety of different board
products, which have an impressive range of engineering properties.
While some panel types are relatively new on the market, others have
been developed and successfully introduced more than hundred years ago.
However, even those panel types having a long history of continuous
optimization are still a long way from being fully developed and they
probably never will be. Technological developments on the one hand and
new market and regulative requirements, combined with a steadily
changing raw material situation, drive continuous improvements of wood-
based panels and their manufacturing processes.
Advances particularly in the fields of adhesive formulations, production
technology, as well as online measuring and control techniques, have
triggered a technology push. The adaptation of these technologies to the
wood-based panels industry has been motivated by the requirement to
improve product quality and reduce manufacturing costs at the same time,
or, in other words, to secure the competitiveness of the wood-based panel
producers. But there is also a continuous market pull that drives the panel
manufacturers and research institutes towards product innovations.
Examples of such developments are the increasing demand for light
furniture, or the need to adapt panel properties so that new coating
technologies can be applied.
Considerable endeavours have been made to ensure that wood-based
panels have no negative effects on human health. In particular, panel
product formaldehyde emissions have been dramatically reduced over the
last decades, and further reduction is still the subject of huge efforts
applied by panel manufacturers, adhesive suppliers and researchers. In
addition, a relatively new issue which is the detection and reduction of
volatile organic compounds (VOCs) emitted from wood-based panels has
come to the agenda. Again, product and process adaptations are needed to
meet the new challenges.
The third major driving force for the permanent further development of
wood-based panels and the respective production processes is the
continuously changing raw material situation. The raw material used in an
individual production line usually depends on what is available in a
relatively small catchment area, and therefore may vary considerably
between different sites. However, there are not only regional variations of
the raw material supply, but also changes over time caused by several
factors. For example, forest management schemes have altered and will
continue to do so. Moreover, the demand from other industries and of the

xii xiii
Preface
 Preface
energy sector for wood previously used mainly for panel manufacturing
has dramatically increased in many regions. These changes force the
panel producers to shift towards alternative sources, including recovered A Word from the Chairman
wood, and permanently push the panel industries to modify and optimize
COST Actions are initiatives funded by the European Commission. An
their processes in order to maintain a consistent quality level. Clearly, the
Action is essentially a network of scientists and industrialists who are
high variability of the wood raw material constitutes a challenge not
interested in a particular topic. The principal aim of an Action is to
known by many other industries.
provide coordination at the European level of research funded at a
The challenges listed here are considerable. On the other hand, a fibre- or national level with a view to avoid repetition and encourage synergies
particleboard is a hierarchically organized product, with quite elaborate between existing and future research projects.
structures on the different size scales from the molecular to the
COST Action E49 is entitled Processes and Performance of Wood-
macroscopic level. Improving the engineering properties of panels
Based Panels and is concerned with all aspects of the wood-based panel
requires several of these hierarchical levels to be considered
(WBP) sector including: raw materials (wood, adhesives, coatings,
simultaneously. This complexity of the material structure is what makes it
coverings, etc.) manufacture of all panel types, their end-use and
so challenging to manipulate the panel properties. Or in other words,
properties.
innovations are difficult without understanding the fundamental
mechanisms at each of the hierarchical levels. To differentiate WBP products from other wood composite materials, E49
defines a panel product as one in which the thickness of the product is
The intention of this book is to give a general description of modern panel
considerably smaller than either its width or length and its manufacture
manufacture, but to also provide some state of the art information on a
includes a flat-pressing step. The broad categories of products covered
selected list of fundamental topics. Of course, it is not possible to include
are: particleboards; oriented strand boards (OSB); fibreboards,
all important aspects of panel manufacture into such a book. Therefore, it
particularly hardboards and medium density boards (MDF); and the
is our intent to give examples and to stimulate the interested reader to
veneer-based products including plywood and laminated veneer lumber
continue his studies by means of the respective technical literature. This
(LVL).
book may be used as a textbook for undergraduate and graduate students,
as assistance for practitioners, and as reference work for scientists E49 has three Working Groups: 1. Process optimisation and process
working in the field. It is an introduction into wood-based panel innovation; 2. Fundamentals and modelling; and 3. Performance in use
manufacture, but hopefully provides new insights into the fundamentals and new products. This state of the art report has been prepared by both
of the production technology even for specialists. WG1 and 2.
This state of the art report has been written by a number of experts who
have freely given up their time to write a range of interesting and
Heiko Thoemen
informative chapters. These experts are to be applauded as the main
Professor at Bern University of Applied Sciences (BFH) beneficiaries of this work are the scientific community and the WBP
sector. Certainly, this report may enhance the reputations and notoriety of
Biel/Bienne, Switzerland, in June 2010 its contributors, but, this is not what drives them; I truly believe that the
authors of this report genuinely wish to help the WBP sector and for
people to better understand WBP products. None of the authors will
receive any financial reward for their contribution and so I hope that, like
me, you will appreciate their efforts.
The principal editor of this report is Heiko Thoemen, who is also the
leader of WG2. He has been the driving force behind this work from the
beginning and his dedication has finally borne fruit with the publication
of this report. Heiko Thoemen has been assisted by the leader of WG1,
Marius Barbu. The first chapter aims to set the scene for the chapters that

xiv xv
Preface
 Wood-Based Panel Technology
follow by providing an overview of WBP manufacturing technology and
this is the main contribution from WG1; the other chapters have come
from members of WG2. Chapter 1
All members of E49 have contributed to this report in some way and I Wood-Based Panel Technology
thank each and every one of them for their efforts. Special mention must
go to the members of the core group: the two leaders mentioned above,
Mizi Fan (WG3) and vice chair of E49, Eleftheria Athanasiadou. Mark Irle and Marius C. Barbu

I wish to acknowledge the magnificent support E49 has received from the CHAPTER SUMMARY
COST Office in Brussels. In particular, E49 has been successfully guided
by Günter Siegel and Melae Langbein and their assistants – many thanks This chapter provides an overview of the manufacture of particleboards,
to you all. oriented strand boards, dry process fibreboards and plywood. The main
aim of this chapter is to provide enough background information so that
Mark Irle the reader may more readily understand the chapters that follow. It is
Chairman of E49 and further hoped that this chapter encourages researchers to contribute
research effort to the WBP sector.
Deputy Director of Ecole Supérieure du Bois
Nantes, France, in June 2010 1.1 INTRODUCTION

The principal aim of this chapter is to provide the reader with a brief
overview of the manufacturing technologies used to wood-based panel
(WBP) products so that they may better understand the more in-depth
chapters that follow.
There are some excellent English reviews on WBP manufacture for
example Maloney (1993), Moslemi (1974), Schniewind et al (1989),
Walker (1993 and 2006) and Youngquist (1999). Unfortunately, the most
recent of these is over 10 years old and technology moves fast in response
to ever changing markets. Consequently, it is hope too that this chapter
will serve as a supplement to existing reviews.
WBP products are made with fibres, particles or veneers, see Figure 1.1.
Each of these three raw materials is discussed in the following chapters
on particleboard, fibreboard and plywood manufacture.

xvi 1
Irle, Barbu
Wood-Based Panel Technology

1.2 WHY MAKE WOOD-BASED PANELS?

1.1: A map summarising the wide range of wood composites that can be made. This chapter is concerned with the panel products in the
It is often quoted that the manufacture of WBP products has been brought
about by the ever increasing cost of logs and lumber, which in turn, has
caused the managers of the world's forest resource to investigate ways and
means of using trees more efficiently. This is certainly true, as many
wood composites can utilise low grade logs such as thinnings, bowed and
twisted logs. They can also use wood by-products and recycled materials.
All sawmills produce large quantities of residues in the form of chips,
sawdust, and slabs. Even the most efficient sawmiller is unlikely to
convert more than 65% of each log in his yard. These residues can be
used to manufacture some of the many kinds of particleboards and
fibreboards that are made today.
Even disregarding the economic advantages, panel products would still be
manufactured because of man's general desire for better building
materials. As a building material, wood has a great number of advantages
but also some disadvantages; the main one being its variability. Wood is a
very variable material both between and within species, and not just in
appearance but, more importantly, in density, strength, and durability.
Table 1.1 and Table 1.2 show that wood composites can be manufactured
to have very much more uniform properties.
Table 1.1: Dimensional stability of timber and boards. Change in
dimensions from 30% to 90% relative humidity (Dinwoodie
dark boxes.

1981).
Direction to grain or board length
Parallel (%) Perpendicular (%) Thickness (%)
Solid Timber
Douglas fir negligible 2.0-2.4 2.0-2.4
Beech negligible 2.6-5.2 2.6-5.2
Plywood
Douglas fir 0.24 0.24 2.0
Particleboard
UF bonded 0.33 0.33 4.7
PF bonded 0.25 0.25 3.9
MF/UF bonded 0.21 0.21 3.3
Fibre-building board
Tempered 0.21 0.27 7-11
Standard 0.28 0.31 4-9
MDF 0.24 0.25 4-8

Although the strength properties of wood composites are generally lower

than solid timber they are more consistent. This means that they can

2 3
Irle, Barbu
 Wood-Based Panel Technology
support loads with smaller safety margins, which in effect reduces the
apparent difference in strength between solid wood and composites. 1.3 THE MANUFACTURE OF PARTICLEBOARDS: A SHORT
Bacteria, fungi and insects, readily decay wood, especially when it is wet. OVERVIEW
Some panel products are better in this respect, particularly in the case of
insect attack. Cement bonded composites have been found to be 1.3.1 Defining Particleboard
extremely resistant to degradation by fungi and even termites.
Particleboard is used as a generic term for any panel product that is
Other benefits of wood composites come from the fact that their made with wood particles. Of course, there is a great range of particle
properties can be engineered. Lumber is limited to a large extent by size, shapes and size used to make particleboards. The type of particle is
width in particular. It is difficult to obtain wood wider than 225 mm and therefore used to define the type of particleboard product. For example,
thicker than 100 mm. The dimensions of typical panels vary from market following English terminology, chipboard is made with chips, a
to market but are usually 2 – 2.5 m long and 1 – 1.5 m wide, but it is flakeboard with flakes, oriented strand board (OSB) with strands and so
possible to buy panels in much larger sizes if necessary. The majority of on (see Figure 1.1 for examples). Most European countries use the term
houses built today have particleboard floors because wooden floors are particle rather than chip and therefore particleboard as a term for
more expensive to buy and lay. The comparatively large size of a tongue chipboard. To avoid confusion over whether the text is referring to
and grooved particleboard floor panel (2440 x 660 mm in the UK) particleboard in the generic sense or particleboard in the specific product
enables a floor to be laid down far faster and produces a less "creaky" and sense, the name chipboard is being used in this text for the specific
flatter result than a traditional timber floor. Wood composites can be product.
made to have special properties such low thermal conductivity, fire
resistance, better bio-resistance or have their surfaces improved for Another aspect of particleboards is that the wood particles are bonded
decorative purposes. together by adding a synthetic adhesive and then pressing them at high
pressures and temperatures. This is important as the manufacture of these
Table 1.2: Strength properties of timber and boards (Dinwoodie 1981). panel products has a marked influence on their subsequent properties.
Bending Strength Bending Stiffness
1.3.2 Wood as a Raw Material
Thickness Density
(MPa) (MPa)
(mm) (kg/m3)
 Par. Perp. Par. Perp.
Solid Timber
Douglas fir 20 500 80 2.2 12700 800
Plywood
Douglas fir 4.8 520 73 16 12090 890
Douglas fir 19 600 60 33 10750 3310
Chipboard
UF bonded 18.6 720 11.5 11.5 1930 1930
PF bonded 19.2 680 18.0 18.0 2830 2830
MF/UF
bonded 18.1 660 27.1 27.1 3460 3460
Approximately 95
% of the ligno-
cellulosic material
used for particle-
board production is
wood. The rest
consists mainly of
seasonal crops such
as flax, bagasse,
and cereal straw. A
discussion of these
various seasonal
crops is beyond the
scope of this
chapter.
Some
wood
specie
s are
more
suitab
le for
particl
eboar
d
produ
ction
than
others
.

1.3.2.1 Workab
ility

A material that is
difficult and costly
Fibre-building boardto Once produced, the chips should have smooth surfaces and a minimum of
break into particles
Tempered 3.2 is 1030 69 65 4600 4600 end grain otherwise the particle will absorb too much adhesive to be cost
not Standard
suitable. 3.2 1000 54 52 --- --- effective. The characteristics of a chip are, to a certain extent, dependent
MDF 9-10 680 18.7 19.2 --- --- on the anatomy of the wood. Softwoods are preferred to hardwoods
because they tend to be easier to cut and the vessels present in hardwoods
cause the chip to have a rough surface.

4 5
Irle, Barbu
Wood-Based Panel Technology

1.3.2.2 Density

In order to manufacture a board of adequate strength the particles must be

compressed to at least 5 % above their natural density. In practice the raw
material is usually compressed to nearer 50 % of its natural density; so if
a raw material of about 400 kg/m³ is used then the finished board will
have a density of approximately 600 kg/m³. This degree of compression is
needed to achieve good chip-to-chip contact.
Figure 1.3 shows the relationship between raw material density and

Figure 1.2: The main process stations in a Particleboard production line (Metso Panelboard).
particleboard physical properties. From this graph it is clear that for a
single specie bending strength increases with compaction ratio. This is as
expected because the more the particles are squashed the greater the
contact between them. The graph also shows that for a specific bending
strength the required particleboard density increases with the density of
the raw material. Most regulations specify minimum strengths,
consequently a manufacturer using say Birch (Betula spp) will have to
produce boards of much higher density than another using Spruce (Picea
spp) to attain these minimums. The heavy boards will require a larger
press for manufacture, incur greater transport costs, and be more difficult
to cut and handle. Therefore, it is not surprising that low density woods
are preferred.

Figure 1.3: Relationship between raw material density and particleboard

bending strength, as predicted by a computer model.

Practically all the physical and mechanical properties of particleboards

are related to density. A manufacturer who attempts to use a number of
different density species over a production period, will produce boards

6 7
Irle, Barbu
with inconsistent properties unless he mixes the chips of the different Wood-Based Panel Technology
species to form a uniform furnish for the press.
particleboards generally compete with each other on price, because there
1.3.2.3 pH of wood
are many manufacturers producing very similar products. Therefore, to be
The curing rates of formaldehyde-based resins, e.g. urea formaldehyde competitive cheap low grade raw materials must be used. Other forms of
(UF) which is used to make most of the world’s particleboard production, particleboard such as oriented strand board (OSB) require much higher
are very dependent on the pH of the environment in which they cure. All quality raw materials because they need engineered particles (strands).
wood species have a pH, if a near neutral species is used then a resin may The extra cost involved in manufacturing these chips is offset by the
not cure sufficiently, or if an acidic species is used then precure may greater selling price of the finished product. The seasonal availability of
result. When adhesive precure occurs the board's surface layer is weak agricultural crops has to a large extent prevented the wide spread
and flaky. This is because the adhesive cures before the particles have exploitation of the enormous volume of suitable ligno-cellulosic material
been compressed and so when the press closes the precured resin bonds produced on farms around the world.
are broken. The pH of the raw material is not usually a problem if it It has already been said that the vast majority of particleboards are made
remains fairly constant, but if it fluctuates then the quantities of hardeners from wood. Wood is obtainable in three forms; as round wood, which is
and buffers added to the adhesives would have to be continually altered to being used less due to costs, as residues from other processes and as
suit the wood in use. Differences occur between species, within species recovered wood (normally processed in the form of large chips).
depending on where the tree grew, and within the tree (principally a
difference between heart and sapwood). In addition, the pH and buffer Round wood
capacity of wood can change with storage time and conditions (Elias and
Irle, 1996). The best particleboard furnish can be produced from round wood. A
particleboard manufacturer has as much control over particle size, shape,
and surface quality as is possible. This situation also allows the decision
1.3.2.4 Permeability
as to leave the bark on the wood prior to chipping or not.
It is postulated that very permeable species of wood will produce poor The disadvantage with this raw material is cost. In addition, there is the
quality particleboard because chips from such wood will absorb the cost of actually chipping the wood. The cost of this is split between the
adhesive applied to it thus creating starved joints. The way a wood breaks capital cost of the equipment together with running costs. On top of this
down is probably more important than its permeability as this will there is the cost of drying the particles.
determine the amount of exposed end grain. The amount and type of
extractives present might be even more important still as these will Wood residues
influence the contact angle that the adhesive makes with the wood after
blending. It is practically impossible to show the effect of wood The success of the particleboard industry stems from its ability to utilise
permeability on particleboard properties, because when two boards made wood residues. Forest residues may take the form of treetops and
with different species are compared many more factors other than just branches, or particles from chipped stumps. The former have not proved
wood permeability will be different. For example, anatomy will affect popular because they contain a high quantity of bark a needles.
surface roughness and possibly the geometry of the particles and these Sawmill residues are preferred since the slabs, edge trimmings, or chips,
two factors could be more important than particle permeability. if a chipping head rig is used, are usually debarked. The larger residues
have to be chipped, so they have some of the advantages and
1.3.2.5 Wood sources disadvantages of the round wood described above. However, sawmill
residues are likely to have a lower moisture content than round wood, so
In addition to having the physical characteristics described above the raw the chips produced should require less drying. Further downstream there
material should also be inexpensive and available in sufficient quantities are the joinery manufacturers who produce vast amounts of shavings and
to support sustained production over many years. Conventional sawdust. Of the two, shavings are preferred, however, since they are used
for pellet manufacture and animal bedding the demand for them is high,
so the cost of shavings is usually prohibitive. The prime advantage of
sawdust is that it is cheap. Joinery mill residues are usually dry and in a

8 9
Irle, Barbu is possible and the residue is likely to contain a range of different wood
species.
particulate form so only secondary break down is needed to produce Sawdust, which many particleboard manufacturers buy and use, is increasing
particleboard furnish. However, much less control over particle geometry in cost rapidly as it is being used for pellet manufacture. When used as a
surface layer furnish it helps to produce a hard smooth dense surface which Wood-Based Panel Technology
many furniture manufacturers prefer. Tensile strength perpendicular to
plane of board, often called internal bond (IB) strength, is improved by
the addition of sawdust. This is probably due to increased inter-particle e.g. stones, concrete, soil, etc.; ferrous metals, e.g. iron and steel; non-
contact. Other physical properties are lowered by sawdust addition. ferrous metals, e.g. aluminium, lead, brass, etc.; and organic materials
such as plastics, paints, rubber, and fabrics. Sophisticated cleaning
Plywood mill veneer cores are an excellent raw material for systems are available but these require significant capital expenditure.
particleboard manufacture. They are all the same size, inexpensive, and of Generally, the economic benefits of using recovered wood justify the
course bark free. Of course, the volumes available are limited. investment required and so a considerable increase in the use of recovered
wood is anticipated for particleboard manufacture over the next couple of
Recovered wood years.
Competition is very fierce in most sectors of the panels industry and so The limitation in the use of recovered wood in many countries is the lack
many manufacturers concentrate on reducing manufacturing costs as of infrastructure to collect, process and deliver it. Another potential
much as possible. Most manufacturers of particleboard use recovered limitation is competition for this resource from the new bioenergy
wood to reduce their manufacturing costs because it is often a cheap generation plants, which are often established with the help of state
alternative to other sources of wood and it is generally drier than other grants and subsequently supported through the receipt of higher than
sources and so there is a considerable saving in energy during the drying market prices for each unit of energy produced.
stage of panel production.
The use of recovered wood in particleboard provides manufacturers with
Using recovered wood seems to be an environmentally friendly thing to an opportunity to market their products as being environmentally friendly;
do and makes economic sense, but it does not come without its own in much the same way as the paper industry has successfully promoted
problems. The particleboard industry has always used a lot of “wastes” as recycled paper products.
raw materials for its products. These have included using trimmings,
sanding dust and reject boards within the production line, sawmill wastes, Bark
secondary processing residues, e.g. off-cuts, shavings and sawdust, and
forest residues. Many of these are classified as being pre-consumer Ideally bark should not be included in a particleboard furnish as it reduces
sources of recovered wood. In other words, they are residues generated as board strength properties and increases resin demand. Bark is most often
a result of making a product, e.g. wooden furniture, and have not been removed from small diameter logs with a drum debarker (Figure 1.4).
used for a particular purpose prior to being used as a raw material for The inclination and the dimensions of the drum (length 6 to 60 m and
another product, in this case, particleboard. diameter 3 to 4.5 m) ensure the out-feed but also the retention time of the
logs. Keeping a ratio of 0.7 between the log length and drum diameter this
The trend of greater use of recovered wood has come from a greater use debarker can process up to 350 m³/h with a bark removal efficiency of up
of post-consumer sources of recovered wood such as demolition timbers, to 99 %.
old furniture and pallets and packaging.
During wood chipping, and subsequent drying and sorting operations, the
Many sources of recovered wood, but especially post-consumer sources, bark is reduced to a fine dust. It is thought that the adhesive is absorbed
contain contaminants that must be removed. These include: minerals, by the dust, because of its high surface area to volume ratio, thus lowering
the amount of adhesive available for inter-particle bonding. Including
bark also tends to increase thickness swelling, linear expansion and
decrease IB strength.
Some people regard the darker board colour which results from including
bark depreciates its marketability. Despite these disadvantages most
manufacturers do not debark their logs before further processing for cost
reasons. About half of the bark is removed from the wood furnish,
however, at two stages: primary chipping, where large pieces of bark fall
off and are removed; after drying where bark dust makes up a high

10 11
Irle, Barbu
Wood-Based Panel Technology

proportion of the dust collected at the air-cleaning cyclones. The bark is

mills. Although this reference is rather old the general principles are still
usually burnt in the factory’s boilers as it has a high calorific value; often
relevant in today’s factories.
providing up to 30 % of a factory’s thermal energy need.
There are many different ways to generate particleboard furnish. For
example, if the raw material were round wood a manufacturer may decide
to use a drum flaker only, sort the produced flakes and sending those that
are too small to a boiler. Alternatively he may decide to use a hacker to
initially break the wood down, and then use a knife ring flaker to reduce
the hacker chips to particles.

Figure 1.5: Different methods of producing particleboard furnish.

1.3.3.1 Primary Breakdown Machines

Hackers tend to be large and robust with drum diameters up to 2.4 m.

They can chip logs of random length and diameter; the logs can even be
twisted or bent. The logs are fed end on to the drum so chip size is largely
controlled by adjusting the feed speed (20-36 m/min). However, large
particles are held in the cutting zone by a heavy breaker screen, which
Figure 1.4: Drum debarker for thin logs, thinnings and branches (Metso
Panelboard). typically has square meshes twice the chip length (<100 mm) and are
therefore broken down by subsequent passes of the knives. The particles
produced are thick (<10 mm) and long (<40-80 mm) with a wide size
1.3.3 Particle Production distribution: 15 % fine (<5 mm) and 10 % oversize (>50 mm). The two to
five knives large and are spring loaded into the drum so that if a large
The way particleboard furnish is produced is dependent on the raw stone enters the cutting zone then damage to the knife is minimised, see
material. Figure 1.5 illustrates some of the many routes which are Figure 1.6. Particle surfaces tend to be rough and fractured since they are
possible. Details of the machines themselves are shown in Figure 1.6 produced more by a splitting than a cutting action. They must be broken
through to Figure 1.11. Fischer (1972) covers the main operational down further and this is often done with knife ring flakers, see below, the
aspects of many of the wood reductionisers used in modern particleboard resultant furnish tends to be splinter like.

12 13
Irle, Barbu Wood-Based Panel Technology

Figure 1.6: The inside of a large drum hacker (Bruks-Klöckner and

Pallmann).

A hacker with a 1 MW motor and a shaft rotating at 300-400 rpm is

capable of converting logs up to 1 m in diameter. It can produce around
150 tons wet (or 70 tons o.d.) of chips 35-50 mm long per hour. A hacker
Figure 1.8: Vertical disc flakers that can cut random length logs (Metso
provides versatility as it can process a wide range of woods in various
Panelboard).
forms. Not all chipboard manufacturers have chosen not to install
hacker/knife ring combinations because the quality of furnish produced is The development of disc flakers has followed a similar line to that of
not as good as that possible with flakers. In addition, recent developments drum flakers in that early versions were limited to logs of a specific
in flaking technology which now allow some versions of both disc and length. For example, the first flaking system installed at the OSB factory
drum flakers to utilise random length and diameter logs, has reduced the in Inverness in 1986, used horizontal disc flakers which accepted logs cut
versatility gap between hackers and flakers. to approximately 1 m in length. Cutting the logs into short lengths reduces
the problems associated with bent and twisted logs. In 1994 the horizontal
disc flakers were replaced with vertical disc flakers similar to that shown
in Figure 1.8. These flakers can convert logs of mixed lengths and
diameters to flakes of predefined lengths and thicknesses. The installed
power of 500 kW allows using a 2 m diameter disc with six knives an
output of 50 tons wood/hour at a belt conveyer in feed speed of 76 m/min.

1.3.3.2 Secondary breakdown machines

A knife ring flaker is so called because of its outer ring of knifes (Figure
1.9). Two versions of this machine exist: one with a stationary outer ring
and the other with a counter rotating ring (Figure 1.9 left). The latter
Figure 1.7: Vertical drum flaker capable of cutting fixed length, random machine has a higher throughput and is more appropriate for wet material.
diameter logs and softwood chips, (Pallmann) The outer rings (10-55 kW) can be replaced in 5 to 25 minutes which
significantly reduces down time. The diagram to the right illustrates how
Round wood can also be chipped with drum or disc flakers. Early drum the inner impeller (100-630 kW) forces the wood particles against the
flakers were limited to logs that were cut to length (dependent on length outer ring of knives. The thickness of the resultant chip is determined by
of drum, see Figure 1.7) and fairly straight. Current drum and disc flakers the protrusion of the blades. Ring diameter can be between 600 to 2000
can use random length logs of poorer quality. In all drum flakers, mm including 28 to 92 knives and ring width from 140 to 600 mm,
regardless of the type, the knives are set at an oblique angle to the axis of depending on the throughput required. The circulated air by the impeller
the drum. This is to reduce vibration and strain applied to the drum varies between 4000 and 18000 m³/h allowing the processing of 2 to 30
bearing as the knife impacts the log; it also causes a slicing action. tons/hour o.d. chips.

14 15
Irle, Barbu
Wood-Based Panel Technology

through the holes of the mesh and not be reduced further. These machines
are efficient as in general only those particles which require reducing are
broken down. Such machines are often used for the preparation of surface
particles. Typically, these machines are 800 to 1800 mm in diameter,
have a grinding track width of 150 to 500 mm, powered by motors of 100
to 1000 kW, creating an airflow of 6000 to 30000 m³/h and have an
output of 1 to 10 tons chips o.d./hour.
Wing beaters are similar to impact mills except that the outer screen does
not have a solid region. Particles which are too large to pass through the
screen are ground down to the appropriate size by successive strikes from
the rotating impellor. Again larger particles will accumulate more
Figure 1.9: Knife ring flaker (Pallmann). moment and strike the outer mesh ring with more force.

The machines of the type shown in Figure 1.9 are known as impact mills
because the particles are reduced by them striking against the solid
“anvil” in the centre of the outer ring. The larger and heavier a particle is
the greater the moment it accumulates from the spinning inner propeller,
therefore, the greater chance that it will strike the anvil with sufficient
force to break it.

Figure 1.11: A hammer mill (Maier and Pallmann).

The hammers of a hammer mill (see Figure 1.11) are attached to the
central shaft by hinges which allow the hammers to swing back if they
collide with a large particle. Large particles are therefore broken down by
a series of blows. Particle size is determined by the size and shape of the
holes in the screen. The robust construction of these machines enables
them to break down solid wood into splinter like particles. Hammer mills
are available in a very wide range of sizes from laboratory-scale to huge
machines like that shown in Figure 1.11, which are used to produce
particles from recovered wood. Consequently, rotor diameter can vary
between 230 to 1800 mm, rotor lengths from 250 to 2000 mm, motor size
from 160 to 500 kW, producing 30-80 m³ o.d. chips/h.

Figure 1.10: An impact mill double stream (Pallmann).

Small particles will not accumulate much momentum from the spinning
propeller and are much more likely to follow the air flow (100 m/s)

16 17
Irle, Barbu
Wood-Based Panel Technology

1.3.3.3 Particle geometry

Wood composite properties can be engineered, to a certain extent, by

adjusting particle geometry. Unfortunately there is a lot of contradictory
evidence in this area. The clearest patterns emerge when board properties
are compared to the slenderness ratio of particles. Slenderness ratio is
calculated by dividing the particle's length by its thickness:

For the majority of properties long thin chips seem best. However, surface
quality (smoothness and hardness) and IB strength are enhanced small
particles. This is why most manufactures attempt to classify their furnish
using the fine particles for the surface layers and the larger particles for
the core.

1.3.4 Particle Drying

Figure 1.12: The characteristics of different chip dryer types (from Ressel
Once the particles have been cut their moisture content must be reduced 2008, acc. to Deppe and Ernst 2000).
to between 2 and 8 %, depending on the adhesive system to be used to
make the WBP. Such low moisture contents are required because residual
moisture is converted to steam in the hot pressing stage, if too much
steam is generated then, when the press opens, the board is likely to be
delaminated by the sudden release of steam pressure.
There are a number of different dryer types on the market as shown in
Figure 1.12. However, chapter will just consider the two of the most
common, the three pass and single pass dryers.
Figure 1.13 illustrates the basics of a three pass dryer. The particles are
introduced into the central tube which will often be heated by a direct
flame fuelled by gas, oil or wood residue. The majority of dryers are
capable of being heated by more than one heat source thus ensuring that
particle drying can occur all year round. The conditions in the central tube
are quite harsh; temperatures range from 250 to 850 °C (700 °C is typical)
and the air speeds are often as high as 8 m/s. In the second tube the air
flow is reversed. A combination of water evaporation and greater tube
volume causes the air temperature and speed to fall. In the final outer tube
the air flow is again reversed and the air temperature will have fallen to
between 60 and 100 °C.

Figure 1.13: A three pass dryer (Metso Panelboard).

18 19
Irle, Barbu The outer tube rotates, typically at 8 rpm, thus causing the particles to tumble,
which aids their passage through the dryer. These dryers are good in that they
are compact for a given evaporation rate. Typical dryers of this type Wood-Based Panel Technology
maybe 30 m long, 4.5 m in diameter and have evaporation rates of 7 tons
per hour, reaching a drying capacity of 25 tons furnish per hour.
the moisture content of the wood. Heat transfer by conduction is more
rapid than by convection, so those dryers that heat the particles mainly by
conduction, i.e. by touching the dryer sides or plates in the dryer, are
likely to use less energy, see Table 1.3. Water evaporates more readily at
high wood moisture contents, particularly above fibre saturation point
(FSP). Below FSP the water is physically bound to the wood cell wall
thus increasing the energy required to evaporate it.
Table 1.3: The specific energy required to evaporate 1 kg of water by
different dryer types.
Specific Heat Requirement
Dryer Type
(kJ/kg H2O evaporated)
Three pass 3350 to 3675
Single pass 3255 to 3550
Contact dryer 3150

Operational particleboard factories are often easy to find because there is

Figure 1.14: A single-pass dryer with dust extraction system (from Ressel usually a large white cloud coming from the dryer exhaust stack. Many
2008, acc. to Buettner2003). people incorrectly assume this to be smoke but in fact it is mainly made
up of steam. Small dust particles and Volatile Organic Compounds
The conditions inside a single pass dryer (see Figure 1.14) tend to be
(VOCs) are also present in the cloud and the emission of these is
moderate in comparison to the three-pass dryer. For example the inlet
restricted. The VOCs can have a strong smell and can be irritating. They
temperature will usually be around 500 °C and the average particle dwell
emanate from the wood, e.g. terpenes, waxes, other organic extractives,
time is increased to compensate for the lower temperatures. The dwell
etc. and also from the sander-dust if this is used as a fuel.
time can range from 20 minutes to as much as 60 minutes. The particles
are helped through the dryer by heated paddles and the rotation of the Emissions from dryers are limited by regulation. In general,
drum itself. manufacturers are not permitted to achieve these limits by dilution,
instead they must limit the production of emissions or install a treatment
An increase in demand for particleboard has caused many factories to
process, e.g. a scrubber (WESP –wet electrostatic precipitator). These
expand which, in turn, has created the need for higher throughputs. The
restrictions may cause a move toward lower drying temperatures (inlet
current trend is towards the installation of a single large dryer including a
temperatures of <400°C) coupled with additional exhaust cleaning.
flash tube pre-dryer (500 °C) with a high evaporation rate of maybe 50
tons/hour whereas previously 5-10 tons/hour would have been the norm. Drying small particles of wood to very low moisture contents is obviously
In recent years, most new dryers that have been installed are of the single a hazardous operation. Consequently all modern dryers have sophisticated
pass type with a predrying step and all are large with drying capacity rates fire detection and extinguishing systems. Many dryers also have spark
around the 70 tons of furnish or more per hour. detectors and automatic controlled sprinkler nozzles that are designed to
detect a potential fire or explosion hazard before either occurs.
The amount of energy required to evaporate a unit of water from a wood
chip is dependent on the way the chip is heated, the particle geometry and For efficient drying it is very important to determine the moisture content
of the furnish as it enters and exits the dryer. The moisture content of the
input material is likely to vary between 12-150 %. If dry material were to
be allowed to enter the dryer without appropriate control (lowering
termperature and increasing gas circulation) an explosion may result.
Consequently, moisture meters for the input need to be accurate at

20 21
Irle, Barbu
Wood-Based Panel Technology

estimating moisture contents above FSP. The output meters, on the other
hand, should accurate in the range of 0-20 % ± 0.2 %. In addition, the
Gyratory screen
meters should be capable of measuring the moisture content of moving Chips
Vibrating screen
material from different species, having different bulk densities and
geometries. Most systems rely on the use of infrared light or microwaves Recycled Gyratory screen
to measure moisture content. wood chips
Vibrating screen Tumbler screen
Shavings Gyratory screen
1.3.5 Particle Sorting
Vibrating screen
Some manufacturers sort their particles before drying, so those outside the
desired range are not dried, which saves energy Another goal of wet
screening in the particleboard lines is to adjust the ratio between face and
Gyratory screen coarse (1mm)
core particle before drying. The drying of particles process can be Tumbler screen fine (0.1mm)
Sawdust Sanding dust
Tumbler screen coarse (0.1mm)
controlled more accurately when particle size distribution is known
before. But wet particles are difficult to sort efficiently as they tend to
stick together. Consequently, in most particleboard factories the particles
are classified after drying. 0 50100150200250300 350
Capacity / Trougput (m³/h)
1st screening
Figure 1.16: Performances of different screening systems (from Ressel,
Flexible
section 2nd screening 2008, acc. to Allgaier).
hose section
3rd screening
section Air classifiers sort particles by air flow. Particles are introduced into a
counter flowing air stream so that the small particles are taken away by
Ball
the air flow and the heavier particles fall to the bottom where they are
cleaning removed by mechanical means, see Figure 1.17. A number of these may
Slide-in be joined together in series, each with a different air flow calculated to sift
screen
out a particular size. These are most commonly used in conjunction with a
Spring struts sieve system.

Figure 1.15: Gyratory rectangular and tumbling screen sieve (Allgaier).

There are two methods of sorting particles: mechanical sieves and air
classifiers. There are three types of the mechanical sieves found in
industry: vibrating inclined screen, vibrating horizontal screen, and
gyratory screen, which are readily recognisable from their housings. All
of these work on the same principle in that the particles are fed over a
series of wire meshes, the particles either fall through or are passed to a
collecting bin, see Figure 1.15.

Figure 1.17: An air classifier, wind sifter (Metso Panelboard).

22 23
Irle, Barbu
Wood-Based Panel Technology

For the forming of the particle mattress consisting of face and core layers
heavy particles, ferrous and partially non-ferrous matter and mineral
the sorting of dried particles into size and shape categories (fractions) like
impurities and dust (Figure 1.18).
dust, fines, coarse, over-size is necessary. The efficiency of the separation
and number of fractions depend on the number of screens. This
equipment needs additional cleaning, maintenance and noise reduction
measures.

Figure 1.19: A roller system for sorting particles generated from

recovered raw material (Acrowood).

Particle size classification can also be achieved using dynamic screens of

rotating rollers as shown in Figure 1.19. The advantage of this method is
that the screens are largely self-cleaning. The texture of the roller surface
and the distance in-between them permits highly accurate separation of
particles into face and core fractions and also the elimination of sand, soil
and foreign particles.

1.3.6 Resin Metering and Blending

(A significant part of this section is taken from Ressel 2008)

Liquid raw adhesives are often purchased as water-based solutions,

containing approximately 50% (PF) to 65% (UF) solids. These are
thermosetting adhesives in which the curing process (condensation
reaction) has been interrupted before delivery of the solution, thus storage
duration is limited to several weeks depending on season, transportation
and storage temperature. Adhesives can be purchased in powder form for
longer storage, but, this rarely used in Europe
Figure 1.18: Compact cleaning system for particle screening and foreign
matter separation (Metso Panelboard). Before application on the dried furnish the adhesive solution must be
blended, according to proven recipes, with water and additional additives,
The combination of a screening conveyer, air sifter, magnetic drum and e.g. hardeners, colours, fire retardants, preservatives, etc.. Different
heavy particle separators in one machine provides a compact cleaning adhesive formulations may be used for the surface and core layers. The
system that is particularly effective when recovered wood is used as raw amount of adhesive mix to be added is calculated on a solid adhesive
material for the particleboard production. This system separates oversize, substance to oven dry wood basis.

24 25
Irle, Barbu
Wood-Based Panel Technology

The hardener solution, which is added to catalyse the resin curing reaction
Table 1.4: continued
is introduced as a percentage of solid hardener substance to solid resin
basis, is added as late as possible to avoid premature curing due to
production line stoppages. The surface and core layers experience
different curing conditions during hot pressing and so different adhesive
mixes are used for surface and core layer furnishes.
Resination of the furnish is a continuous process, which requires
constant and continuous mass and/or volume flow control (furnish
weight, resin weight/volume) to guarantee accurate blending and uniform
panel properties. The mixing of the adhesive recipe on the other hand Batch-wise adhesive blending is used for small and medium-capacity
before application to the furnish fractions, is done either continuously or particleboard plants, operating with long production series without
batch wise. The type and amount of adhesive depends on panel type changing the adhesive formulation. During batch-wise adhesive blending
(interior or exterior application), particle size, hot pressing conditions etc. all of the adhesive formulation components, with exception of the
The data in Table 1.4 are just for general information and may vary. hardener, are mixed in a tank. Hardener solution is first added to the
furnish in the particle blender
Table 1.4: Typical resin addition levels for different panel types (Ressel,
2008).

Figure 1.20: Batch wise adhesive blending of particle (Metso Panelboard).

Continuous or in-line adhesive blending is used for large capacity

particleboard plants (Figure 1.21). Adhesive blending and the recipes are
normally controlled and managed by a computer system. Facilities that
use MDI are required to take special precautions. With in-line blending
all components are added simultaneously into the particle blender. Pre-
conditions for reliable adhesive blending and thus uniform product
quality, but also high flexibility in changing the production programme,
are reliable and accurate measurement and metering. The benefits of such
highly sophisticated systems are: significant raw material savings of up to

26 27
Irle, Barbu
5%, optimum adjustment to any type of raw material and requirements Wood-Based Panel Technology
through integrated metering on the basis of flow meters and metering
scales, high flexibility due to individual simultaneous metering.
and every particle. The additives may be atomised by either hydraulic or
Different methods for measuring and metering of the adhesive air spraying nozzles. For the former, the droplet size is dependent on
components: nozzle design, liquid viscosity and liquid pressure (5-17 MPa). The
disadvantages of these sprayers are the high pressures required and nozzle
 Mass flow measurement according to the Coriolis-principle, which clogging. The droplet size for air atomising spraying nozzles is also
is suitable for all media with variable density, viscosity or affected by the above factors (although the liquid pressures are very much
conductivity. The system measures the real mass flow. lower). In addition, the pressure of the air impacting the liquid jet also
 Magneto-inductive flow measurement uses the electric conductivity influences droplet size. The ideal droplet size is reported to be less than
of the medium as its measuring method and therefore is applicable 35 μm in diameter (Moslemi, 1974). In practice the size varies in the
to aqueous solutions with constant conductivity. range of 30-100 μm. The main reason for this is that formaldehyde
adhesives are colloidal suspensions, and suspensions do not atomise
 Mass flow measurement with differential metering scales is a readily.
proven and highly accurate measuring and metering method. It is
Viscosity and ambient temperature affect resin penetration in wood.
equally suitable for liquid and powdery adhesives and additives. It
Ideally, the resin formulation should remain on the particle surface so that
is easy to calibrate and to check as well as a uniform technology,
it is available for bonding. The smallest particles absorb five times more
suitable for all media and flow rates
resin than the largest on a weight for weight basis. Some mills add the
finest particles just at the end of the blender in order to minimize resin
absorption
Particle dosing is required to ensure a controlled furnish flow into the
blender and to achieve reliable constant resination. Particle dosing bins
are built with either one or two belt tables. Additionally they are equipped
with dust suction nozzles and belts with scrapers to keep the dosing bin
belt clean. The weight scale controls furnish discharge volume. If two
separate belt tables operate, the bin itself is separated from the metering
belt to allow high bin volume and to ensure high measuring accuracy. On
the first belt table, the level of the bin is measured and furnish flow
adjusted. The second belt conveys the flakes via online weight scale,
controlling the furnish discharge mass (Figure 1.22).
Most modern mills use continuous blenders to add resin and other
additives such as wax to the furnish. The spraying of adhesives onto dried
particles is often termed blending. Blenders are classified in two types:
short and long retention time.
From 1950s onwards, the short retention time (2-3 seconds) blenders have
been preferred. These are 2-3 m long and 600 mm diameter. They have a
short concentrated spraying area, so some particles inevitably receive
Figure 1.21: Continuous adhesive blending of particle (Metso Panelboard). more resin than others. Resin redistribution is improved in the rest of the
blender by violent tumbling (600-1000 rpm) in the bulk of the blender
Blending is the mixing of dried wood particles and adhesive formulation
(Figure 1.23).
with the aim of achieving a uniform distribution of drops of resin on each

28 29
Irle, Barbu
 Wood-Based Panel Technology

Manufacturers of single layer and graduated density boards, see section

3.8, will often use a single blender. Two blenders will be used for the
production of multi-layered boards; one for the core and one for the
surface furnish. This allows the possibility of optimising the additive
levels for the different layers. For example, more hardener may be added
to the core layer to increase cure rate or more water for the faces to cause
a steam shock effect during hot pressing.

1.3.7 Common Adhesives

For economic production of particleboards the adhesive must cure in the

press very quickly, say within 1-5 minutes, but it must also have a potlife
Figure 1.22: Furnish dosing bin and blender (Metso Panelboard). of something in excess of 20-30 minutes so that the adhesive does not
cure before entering the press. Longer potlives are desired, and achieved
by many, to allow for line stoppages.
The most commonly used particleboard adhesive is UF, this is followed by
melamine formaldehyde (MF) and then phenol formaldehyde (PF). Other
adhesives have only minor importance in the global view.
Urea Melamine Phenol Formaldehyde
Mono methylol
Mono methylol melamine phenol
Mono
The methylol of UF are that it is relatively cheap, cures to a clear or
advantages
urea HOH2CHN N NH OH is much production
white glueline, provides goodC dry
C strength,
2 and there H
H N CH2OH CO
experience with this resin system.
N N In order to achieve a high ratio of
CO
potlife to high temperature gelC time latent hardeners are usedH (Meyer,
H N H
1981). The most popular of theseNH
used
2 to be ammonium chloride (NH 4Cl).
CH2OH
When added to UF resin it produces hydrochloric acid by two different
mechanisms. In one, it reacts with free formaldehyde and chain end
methylol groups thus:

Figure 1.23: A short retention time blender (from Ressel 2008, acc. to 4 2 2 2
Binos).
Like similar chemical reactions the reaction rate increases with
A long retention blender is a large drum (L<5m) requiring minutes for temperature. So when the ammonium chloride is first added there is a
the same throughput and job as the short one. It treats the furnish gently sudden drop in pH, as result of the high availability of free formaldehyde,
and reach an efficient resin distribution for inhomogeneous one. but then the pH falls at a much slower rate until heat is applied. The

30 31
Irle, Barbu operations to help heat the chip dryers. Chlorides are therefore released when
the dust is burnt because it contains adhesive and the chlorides attack the
additional source of acid arises from the dissociation of ammonium metal components of the dryers leading to severe corrosion.
chloride to ammonia (NH3) and hydrochloric acid on heating. Ammonium chloride has been largely replaced with ammonium sulphates
A fast transition from liquid to solid in the press can be obtained by and ammonium nitrates; both of which avoid corrosion problems in the dryer.
adding 1 g of ammonium chloride to 100 g of resin (solids basis). Such Using UF resins is not without its disadvantages. UF is not weather resistant,
addition levels, however, can cause precure, so NH 3 or hexamine are which precludes exterior uses. It also releases formaldehyde and since there
added. The former is better in that it reduces formaldehyde release. The are regulations which limit the maximum concentration of formaldehyde in
use of ammonium chloride and other halogen containing compounds has the air, this can restrict the number of interior uses for boards bonded with
fallen. This is because many mills burn the dust from their sanding
this resin. Wood-Based Panel Technology
The moisture resistance of UF can be improved by fortifying it with
melamine to form a melamine-urea formaldehyde resin (MUF). These Isocyanate-based adhesives (MDI) have been used for commercial
adhesives are clear and strong, but they are more expensive. The price production of particleboards, MDF and OSB. Relative to the volumes of
difference between a fortified and pure UF resin depends on the amount UF adhesives used, however, the isocyanate adhesives are small. These
of melamine added, but a pure MF resin is about three times more adhesives have become known as MDI, which stands for methylene
expensive than pure UF. Although not quite such a problem, MF and diphenyl diisocyanate. In actual fact the isocyanates used consist of
MUF resins still emit formaldehyde. polymeric forms of MDI. Although more expensive than formaldehyde
based adhesives, MDI performs so well that a particleboard with adequate
Phenol formaldehyde resins on the other hand are much more weather properties can be made with much less resin than is possible with
resistant and do not have a formaldehyde problem. Admittedly, they are
formaldehyde resins. The first isocyanates caused production problems as
more expensive, being approximately twice the price of UF resins. Some they tended to cause the board to stick to the metal platens. This has
people consider the characteristic dark red colour of cured PF resin
largely been solved using release agents, see Galbraith et al (1983). There
detracts from the board's appearance, but since most boards are covered in are versions of isocyanates that can be mixed with water to form an
use, e.g. melamine or veneer coverings, this is probably not a major
emulsion which considerably eases the difficulty of spraying very small
disadvantage. However, in use some additional production problems are quantities of resin onto the furnish.
encountered with this adhesive. Higher temperatures and longer time
periods are required to cure PF. Not only does this reduce productivity but The binder cost as a proportion of total manufacturing costs is around 15-
it can also lead to significant penetration of the adhesive into the wood 25 % of total. When the amount used is considered (2-10 % of wood
chip; if the glue has been absorbed by the particles then it is not available weight), then it can be seen that a small change in use or cost can have a
to bond them together and poor board strength results. significant effect on profit. A great deal of work has been done on wood
adhesives. Much of this information has been coherently reviewed by
Pizzi (1983) and in the two state of the art reports written by members of
COST Action E31 (Dunky et al, 2002 and Johansson et al, 2002).

1.3.8 Mattress Forming

It is vital that this stage is carried out properly. Board density directly
affects properties, consequently, the preparation of a consistent mattress is
critical in producing a board with suitable properties. The mattress must
be laid down uniformly across its length and width, but not necessarily
through its thickness. Instead the density profile through the board's
thickness must be symmetrical about the board's centre. If this is not
achieved then the board will probably warp in service.
Mattress formers are either segregating or non-segregating. They may
operate as either batch or continuous formers, the latter are preferred in
modern mills. In some small and old mills the mattress is laid onto caul
plates, which are metal plates 4 to 6 mm in thickness and the size of the
press platens. The caul plate is used to transfer the mattress to other
process stages, including pressing. The high capacity mills of today do
not use caul plates because of the capacity limitations they impose, capital
outlay, maintenance costs, and additional handling equipment required
e.g. caul separators and return lines. Instead, in these mills the mattress is
laid directly onto a conveyor belt.

32 33
Irle, Barbu
Wood-Based Panel Technology

Non-segregating formers tend to be used in the production of single or

roller; the heavy particles having more momentum than the small, tend to
multi-layered boards. There are many different designs from various
travel further. The overall effect of both machines is to produce a board
manufacturers but they all aim to randomly orientate the particles so that
with small fine particles on the surfaces (good for smoothness and
the board strength properties are similar with and across the machine
hardness) and larger chips in the middle (good for bending and impact
direction. The motion of the conveyor belt, or the forming machine, does,
strength).
however, tend to cause some particle orientation in favour of the machine
direction. Figure 1.25 shows an example of the relatively new textured roller
system that uses a bed of rollers to classify particles. The roller surfaces
have a range of patterned surfaces, of which there are many designs with
new patterns appearing with further development. The regular patterns
have 3D shapes like diamonds/pyramids of 0.5 to 3.0 mm height which
facilitate separation whilst minimising further particle size reduction. The
diameter of the rollers is around 80 mm and some rotate rapidly (80 rpm)
and have a coarse pattern whereas others rotate more slowly (30 rpm) and
have a fine pattern. This helps to uniformly distribute the furnish across
the whole machine width. The gap between the rollers at one end of the
bed is about 0.1 mm and this increases to 2 mm along the length of the
bed. Consequently, fine particles are able to pass through at the start of
the bed and the larger particles are transported down the bed where the
gap increases thus allowing more particles to fall through. Therefore, one
bed of rollers is able to lay one half of a panel or one layer if the panel is
of a 3-layer structure. A significant advantage is that with proper
adjustment of the beds there are no abrupt changes in particle size and
densities between layers and this leads to better IB values.

Figure 1.24: Principle of classifying and forming of particle mattresses

(Ressel 2008).

Segregating formers work by one of three principles: air jets; throwing

rollers and textured separating rolls (see Figure 1.24). In the first
machine, particles are dropped in front of a column of air jets. The
horizontal air flow created by the air jets causes the small light particles to Figure 1.25: Roller (crown) former line for high production rate 3,100
land on the conveyor belt away from the column; heavier particles tend to m³/day (Metso Panelboard).
settle at the column base. The other former classifies the particles by
The roller system is very compact compared to traditional formers. This is
dropping them onto a spinning ribbed roller. The chips are thrown off the
most obvious when a roller system is retro-fitted to an existing line where,

34 35
Irle, Barbu
Wood-Based Panel Technology

because it requires less space than previous former, there are large gaps
does not have to close so much to compress the mattress to the desired
around the new former. The roller system does not require air flow and so
thickness, which also helps to reduce the chance of precure. A press can
there is less dust emission and particle drying so trimmings from mattress
compress a pre-pressed mattress far quicker than one that has not been pre-
can be recirculated back in to the former and still have similar properties
pressed because the prepress squeezes out much of the air in the mattress.
to fresh particles. Also it is more able to form low density mattresses and
This air can sometimes cause particles to be blown out of the side of the
fine and uniform surfaces.
mattress if it is pressed too quickly. Consequently, the overall press cycle
A significant advantage is that, unlike sieve-based systems, the roller is reduced. A good example of how pre-presses may increase production
system is largely self cleaning. Compared to conventional former the was shown by Kronospan who raised their production from 180,000
investment is low because the building height is lower, the platforms and m³/annum to 216,000 m³, a 20% increase, just by adding pre- presses to
supportive structure are simple and no fans and exhausting pipe are their manufacturing lines at the end of the 1970s. All modern lines include
necessary. The equipment can eliminate oversize foreign matter and a pre-press.
particles and thus protect the steel belts of the continuous press.
Pre-presses are also used on production lines that have continuous
presses. Many of the benefits given above are also of relevance to
1.3.9 Mattress Pre-Pressing and Pre-Heating continuous presses.
The press is the most expensive single piece of equipment in a
particleboard factory (approximately 15% of investment) so it must be
operated as efficiently as possible. Many presses produce boards in
batches whereas the rest of the mill, such as flakers, dryers and formers,
operate continuously. As a result, most modern mills have cold pre-
presses that compress the mattress to about 50-70% of its formed height
(see Figure 1.26).
Inlet section Main nip section

Forming
Conveyer belt
Side walls

Figure 1.27: Preheating of mattresses using steam or hot air via surface or
(only for MDF) via core (Courtesy of M. Gruchot).
Main hydraulic
cylinders After cold pre-pressing, and especially for thicker mattresses, mattress

Figure 1.26: A pre-press (Metso Panelboard). pre-heating is used to increase mat temperature (ideally the core) is
sometimes used. The most appropriate choice of heating method is
Since a pre-pressed mattress is thinner the press does not have to open dependent on furnish type, moisture content, energy price and board type.
so much in order to accommodate it. It therefore follows that the press Mattress preheating methods include: electro-magnetic energy (high

36 37
Irle, Barbu
Wood-Based Panel Technology

frequency <45 MHz and microwave >2,000 MHz), heat of condensation

reduce the amount of, edge effects. These edge effects are caused by the
(saturated steam, supersaturated steam, high humid hot air) and hot dry
inherent temperature and pressure variations experienced by particles in
air. The high costs of power today largely preclude the use of electro-
the outer edges of a mattress. As a result the outer portions tend to have
magnetic energy for WBP, but it is still used for the preheating for veneer
inferior physical properties to the bulk of the board, see section 0. The
based products (moulded plywood, LVL, PSL etc.) and for heating during
amount of "waste" removed as a proportion of board size is much smaller
pressing of GLT. For particleboard and especially for dry process
for large boards than for small ones. A similar reduction in waste tends to
fibreboard, preheating typically uses steam or hot air (see Figure 1.27).
occur when non-standard sizes are being cut from large production
panels.
1.3.10 Pressing

Particleboard can be made with batch or continuous presses. Batch

presses are available in single or multi-daylight forms. A single daylight
press (up to 52 m) makes a single board at a time, whereas a multi-
daylight press produces more than one board per cycle (typical 6-10 for
particleboard, but, much larger numbers can also be found).
The majority of commercial size presses are heated by steam, hot water or
hot oil, the latter two being preferred because consistent platen wide
temperatures are easier to obtain. The use of electric heating elements is Figure 1.28: A single daylight press (Dieffenbacher).
restricted to laboratory scale presses because of the high cost and large
temperature variations likely in large platens. Other heating systems have Better thickness tolerances and shorter pressureless waiting time have
been proposed (Moslemi, 1974; Axer, 1975). contributed to improved sanding allowances of 0.7 – 1.4 mm. The steam
injection press is a special single daylight press that can make very thick
Typical platen temperatures range from 200 to 220°C. Such high panels (>50 mm).
temperatures are required to achieve rapid cure of the adhesive. The
specific pressing pressure ranges from 2 to 4 MPa, principally depending Multidaylight Presses
on final panel density, but, raw material density and panel thickness also Some multi-daylight presses make as many as 48 boards per cycle.
have an impact. Typical multi-daylight presses produce between 4 and 8 boards, 5 to 7 m
long and 2.5 m wide. It can be seen that multi-daylight presses have much
1.3.10.1 Batch Presses higher capacities than single daylight presses; for example, a multi-
daylight press producing 4 boards 5 m long per cycle will produce the
Single Daylight Presses
equivalent of a single daylight press 20 m long.
In order to attain economic capacities and to maximise the advantages of
An essential part of using a multi-daylight press efficiently is that all the
single daylight, these presses tend to be designed to manufacture large
boards of a press load must enter and exit the press at the same time.
boards. Most presses installed today are 25 m or longer. Egger (UK) Ltd.
Whilst in the press they must also be subjected to the same temperature
had until 2009 a 52 m press, which has now been replaced by a
and pressure regimes. If this is not done then there will be significant
continuous press. Although long, the width of the boards that these
variation between the boards of a single press load. Multi-daylight presses
presses are invariably no more than 2.8 m. This is to allow steam to
therefore have loading and unloading cages attached to them which
escape from the mattress during pressing. Similar widths are found in
introduce and extract the boards. Figure 1.29 illustrates the system of
multi-daylight presses for the same reason.
arms and levers employed to ensure the simultaneous closing and opening
The main advantage of single daylight type presses is related to the size of of all daylights. This equipment is not required for single daylight presses.
product produced. All densified wood composites must be trimmed after
pressing, this is not just to provide a straight edge, but also to remove, or

38 39
Irle, Barbu
Wood-Based Panel Technology

calculated by multiplying its width by the panel exit speed. Thin panels
travel through the press much faster (i.e. 3 mm at 120 m/min) than thick
ones (i.e. 38 mm at 5 m/min), because there is less material to heat. In
theory then, the production capacity of the press is the same regardless of
panel thickness. This does not quite happen in practice. They are,
however, more efficient than batch presses at making thin panels. A batch
press must open, be loaded and then close. This takes a finite amount of
time. So when making thin panels, this opening and closing takes a
significant proportion of the production time and overall capacity falls.

Figure 1.29: The simultaneous opening and closing mechanism of a multi-

daylight press.

A major disadvantage of multi-daylight pressing is the large sanding

allowance of 1.0 - 2.5 mm depending on thickness and product type. Low
thickness tolerances and softer surfaces due to some pressureless closing
time are the main causes of this.
Recent OSB projects have been based on multi-daylight systems, have
reached a design capacity of 700,000 m³/a with 12 daylights of 3.66 x Figure 1.30: A diagram of the principles of a continuous press (Thoemen
and Humphrey 1999).
10.37 m. The main reason for the re-gained success of multi-daylight
presses for OSB manufacture is that these presses are less affected by Another advantage caused by the fact that a continuous press remains
glue spots and mat inaccuracies. When continuous presses are used they “closed” the whole time, is that the panels produced have very close
tend to require significant steel belt repair when used for OSB. thickness tolerances. This in turn, reduces the amount of sanding required.
Continuous Presses Tremendous amounts of money can be saved by minimising sanding
losses.
Single and multi-daylight presses make boards in batches and yet the rest
of production line operates continuously, i.e. blending, forming, pre-
pressing. Consequently, equipment is required to transfer mats from the
continuous to the batch pressing stages. Continuous presses, on the other
hand, do not require such equipment and so simplify production flow, see
Figure 1.30.
The popularity of continuous presses can be seen in the statistics
published in Wood Based Panels International (Wadsworth, 2001 and
Natus, 2008) on particleboard production. These show that continuous
presses account for 36 % of the manufacturing lines in Europe and around
about 50 % of the production volume and these figures will increase over Figure 1.31: A diagram of the mechanism employed to minimise the
the next few years. friction between the stationary hot platens and the moving
steel belts that carry the mat through the press (Thoemen
A continuous press can be used to make a wide range of product
and Humphrey 1999).
thicknesses and because the panel leaves the press as one long piece, it
can make a wide range of sizes too. Its production capacity is easily

40 41
Irle, Barbu
Wood-Based Panel Technology

The use of continuous presses for the production of particleboard, OSB

Advantage Disadvantage
and MDF around the world is a real success story.
Raw Materials: Reduced trimming High initial cost.
Typical press widths are between 1.8 – 3.0 m. Only a few plants exist
and sanding losses
with widths greater than 3.0 m. Pressure distribution, frame design, and
mat de-aeration are the main reasons why presses wider are not made Production: Lower energy Steel belt: Need for careful belt
frequently. requirement per unit volume tracking control. Belt easily
produced. Similar production damaged by hard objects on/near
There is a continuing trend for greater press lengths and speeds. The
capacity at different product mattress surface, so excellent metal
longest continuous presses for particleboard and MDF are in the 50 m
thicknesses production capacity of detection in mattress required.
range and for OSB in the 60 m range (the current record is 77 m). Press
batch presses is reduced when Extensive wear of belt requires
line speeds have been increased with the growing demand for thin MDF
making thin boards. Thickness expensive replacement after 3-5
(< 3 mm) and the ongoing lowering of minimum thickness of this product
changes are easy to perform and years of operation.
(<1.5 mm). Line speeds of 120 m/min are now possible. As consequence
quick
of these developments, a modern single production line for particleboard
can make 2,500 m³/day and 1,500 m³/day for OSB. Properties: Reduced density Poor heat transfer from platen to
The hydraulic systems are characterized by precise pressure control by profile that gives a better IB for a mattress. Not so appropriate for
proportional valve technology. The high line speeds demand fast control given density and reduced tool pressing thick (>40 mm) boards.
systems, which are realized by PLC, local numerical control systems, or ware. Reduced precure and no Equipment construction is limited
by central micro-processor control. asymmetry so there is scope for regarding high compression rates
resin development nor thin (<2 mm) boards.
For press gap control, specialized thickness metering systems are needed.
Depending on the basic press design these instruments may have to The calendaring presses were the first systems able to make WBP
withstand quite high working temperatures. The product thickness profile continuously. They were introduced to the market in 1971. Due to the
is fed back to automated control loops to achieve good accuracy. limited heat transfer and pressure capabilities the calendar presses are
Continuous presses have controllable heating zones down their length. limited to thin particleboard and MDF (<4 mm). They often have
This adds another variable for manufacturers to use to manipulate board microwave pre-heaters (Figure 1.32). Around hundred lines of this type
properties so that the product is more suited to a particular end-use, e.g. are still operating. Limitations in capacity (300 m³/day) caused by the
flooring as opposed to furniture applications, all on the same press. Roll drum diameter (3-4 m) and length (< 2.5 m) and of the density of the
et al. (2001) presented some results from a continuous press that has a product limits development possibilities.
cooling zone at the end of the pressing section. It would appear that this
helps reduce “blows” and, consequently, the mattress can enter the press
with a higher moisture content which helps heat transfer through the
mattress and speeds production.
The advantages and disadvantages of continuous presses are summarised
as:

Figure 1.32: Calendar press with microwave pre-heater and simultaneous

one face coating, detail with a modern calendar press (Bison
and Binos).

42 43
Irle, Barbu
Wood-Based Panel Technology

1.3.10.2 Mattress conditions during pressing

Prior to pressing a mattress should have the same temperature through out
and have a uniform density (symmetrical for multilayer boards). The
board's moisture content should also be consistent throughout unless a
manufacturer is using the steam shock technique, more of which later. In
addition, there will not be any vapour pressure or internal stresses present
in the mattress before pressing. Once pressing starts, however, all of these
factors will begin to change at various rates in different parts of the
mattress. Consequently, the geographical position of a wood particle will
Figure 1.33: Observed thickness swelling across a daylight (4 by 2.5 m) of
determine the conditions it is subjected to, which in turn will give rise to OSB. Note swelling at lowest point is approximately 10% and
the board possessing inconsistent physical properties. 24% at highest point (Bolton, et al 1989)
The minimum press time possible is determined by the time required to
heat and cure the adhesive in the core layer. Heat is transferred in the When a board is first placed into a press its surfaces begin to warm. Since
mattress is by radiation, conduction and convection (water vapour wood temperature is negatively correlated to strength then this surface
movement). The latter transfer mechanism is predominant, particularly in heating will cause the surface layers to squash more readily than the core.
the early stages of pressing (Bolton et al, 1989). The main reasons for this As a result even single layer commercially pressed boards will have
are that wood is a good insulator and vapour flow will occur readily as higher density surfaces. This is often seen as an advantage because a
soon as a vapour pressure gradient is created (by surface heating). Note board with high density surfaces will have a higher bending strength than
that the surface moisture does not have to turn to steam, which is water a board which does not. In addition, the high density makes the surfaces
above its boiling point, in order to move through mattress since any scratch resistant and less prone to absorb paints and adhesives applied to
vapour pressure gradient will induce some flow. the surface. However, particleboards are usually made to a set thickness
and density, so if the surface layers are dense then it follows that the core
Humphrey and Bolton (1989) and others (see Chapter 3), have developed layers must have a low density. It has already been mentioned that board
a computer model which predicts changes in mattress environment physical properties are closely related to the density of the product, see
during pressing. Their model is unlike others in that it is a three section 3.2.2. A low density core will therefore equate to a low IB
dimensional model; predicting conditions across the board's width and strength.
length and not just through its thickness like most models. Their model
predicts that particles at the edges of a mattress will experience lower Studies have shown that this variation in density profile can be reduced
vapour pressures and temperatures during the press cycle. These two by closing the press very slowly e.g. seven or more minutes (Moslemi,
factors are likely to reduce particle stress relaxation and hence increase 1974). Such long closing times inevitably lead to precure in the faces and
the internal stresses remaining in the board after pressing. The effect of loss of production. The alternative method would be to close the press
this is clearly shown in Figure 1.33, which shows thickness swell across a extremely quickly, say within a few seconds, so that the surfaces do not
daylight (2.5 by 4.5 m) of an OSB. Note how the edges swell have sufficient time to heat up and plasticise, but this would require very
considerably more than the central region. large and powerful hydraulic pumps. The relatively new technique of
steam injection pressing (Geimer, 1982) significantly reduces density
The way a board is pressed will determine its density profile, which is the variation and, at the same time, greatly reduces total press time. This is
change in density from one surface to the other. A board's density profile achieved by injecting steam through a perforated platen as the board is
can significantly affect its strength properties and therefore its end use. being pressed. The density variation is reduced by the fact that the board
is more uniformly heated through out during compression. Whilst the
press cycle may be reduced because the core layer is heated more quickly
and so the adhesive starts to cure sooner. Many people see this as steam
injection's main advantage, but there are other benefits. Instead of

44 45
Irle, Barbu
Wood-Based Panel Technology

producing particleboards with conventional properties in half the time,

cycle. The level of stress remaining in the board once the press opens will
maybe this technique could be used to make much more stable boards
be dependent on temperature, moisture content, particle geometry, density
using standard press time cycles. Much of the inherent instability of
of raw material and final product, and total press time.
densified wood composites comes from particle densification reversal.
When particleboard mattresses are pressed for long enough the stresses
induced by compression are relaxed through wood creep, consequently, if 1.3.11 Processing Steps Immediately after Pressing
stress relaxation is sufficient then spring back through stress reversal All densified wood composites must be trimmed after pressing, this is
becomes much less important. The development of steam injection not just to provide a straight edge, but also to remove, or reduce the
pressing will also allow the production of much thicker products, 100 mm amount of, edge effects (Figure 1.33). These edge effects are caused by
or more. Such thicknesses cannot be made by conventional means the inherent temperature and pressure variations experienced by particles
because of the difficulty in heating the core layer. in the outer edges of a mattress. As a result the outer portions tend to have
inferior physical properties to the bulk of the board, see section 3.10.2.
The amount of "waste" removed as a proportion of board size is much
smaller for large boards than for small ones. A similar reduction in waste
tends to occur when non-standard sizes are being cut from large
production panels. This explains why, initially large single-daylight
presses, and then continuous presses gained in popularity.

Figure 1.34: The change of parameters during continuous hot pressing, Figure 1.35: Primary board finishing after hot pressing: blister detection,
example refers to MDF (Thoemen et al 2006). trimming, cutting to length, thickness determination,
weighing, and cooling (GreCon).
Mattress moisture content and distribution will also affect density profile.
Some manufacturers apply a fine spray of water to the mattress faces with
the idea that during pressing this extra water will evaporate, migrate to the 1.3.11.1 Blister detection
core and therefore heat the centre more quickly. This is known as the If the internal gas pressure in the mattress at the end of the press time
steam shock technique. There is evidence that this does indeed happen. exceeds the internal bond strength in the hot state (at 100°C in core),
The addition of water to the faces has a secondary effect; it plasticises the blisters (delaminations) occur. They tend to be concentrated in the middle
surface layers so that surface density is increased and density profile because this is where gas pressure is highest. Blisters can be found using
variation is enlarged (Maloney, 1993). ultrasound transmitters and receivers distributed across the width of the
As a press closes stresses are induced in the wood particles. These production line immediately after the boards exit the press (Figure 1.36).
stresses act in opposition to the closing of the press and so the level of The earlier this information is available, the easier it is to adapt
stress is often termed counter pressure. A simple press will operate by production parameters. The press operator initially changes the press
pumping hydraulic oil into the press rams until a certain pressure is speed (lowering), then may try reducing the moisture content of the
reached. Board thickness is governed by metal stops. When such a press mattress at press entrance or improve the furnish bond by increasing
has "closed to stops" the stops will bear the load from the press ram. The adhesive content; all of which add to production costs. If no blisters are
stresses in the mattress will begin to fall off with time. Panel thickness detected and the IB stays within the safety margins, then the operators
can be controlled by using distance transducers and adjusting hydraulic usually try to gradually increase the line speed to maximise production
pressure in the rams. Again, the hydraulic pressure will fall during the capacity.

46 47
Irle, Barbu Wood-Based Panel Technology

Figure 1.37: Diagonal saw for cut to length panels after continuous hot
pressing (Metso Panelboard).

1.3.11.3 Weighing and panel thickness measurement

Each individual trimmed panel is controlled in terms of thickness and

weight. Usually the thickness is measured continuously at three or more
traces over the whole panel using roller pair gauges (Figure 1.38).
Between measurements the roller pairs contact each other and calibrate
themselves in order to assure a high measurement precision (0.02 mm).
For high line speeds (>100 m/min) this type of calibration is not possible
and a movable C-frame like that for the ultrasound equipment permits
Figure 1.36:Blister detection in the hot panels after press (GreCon).
external calibration of the equipment.

If the hot panels contain blister they are immediately eliminated from the
flow. They may be used as packaging material or rechipped to particles
with a hammermill and reused as furnish or as fuel.

1.3.11.2 Trimming and cutting to length

Using a diagonal or flying saw composed of two circular saws the

continuous press endless hot board is cut to length (normally no more
than 6.5 m) to a so called master panel. The correlation between the
press speed and the traversing circular saws is computer controlled. For
thin panels, which are pressed at high speeds, both circular saws operate.
Also this sawing equipment has to fulfil high precision for the panel
Figure 1.38: Thickness gauge based on contact roller pairs (GreCon and
length ±2 mm, squareness ±1.5 mm/m and straightness ±1.5 mm/m. EWS).

Normally, the weight of each panel is measured by a scale installed under

a separate belt conveyer. The weight cells need 2 seconds to obtain an
accurate value. For new production lines with high line speeds classic

48 49
Irle, Barbu
Wood-Based Panel Technology

scales do not have enough time to accurately weigh the panel. Also the Temperature Variation in a 20 mm Thick Mattress Pressed at 120° C for 10 mins.
accuracy of this weighing system is not very good for thin panels because
the weight of the panel is low relative to the weight of the scale itself. 140
Surface
Dimensional data from the saws and thickness gauges and the weight of 120
2 mm
6 mm

each panel is used to calculate panel density. This can also be achieved 10 mm

100
using an X-ray system, which directly measures the density and can also

Temperature (° C)
detect also foreign matters, furnish balls and other defects, but it is 80

relatively expensive. 60

40

1.3.11.4 Cooling 20

Particleboards bonded with UF or MUF must be cooled after they have 0

0 5 10 15 20 25 30 35 40
been taken out of the press. If the boards were stacked together Time (minutes)

immediately, the residual heat would cause thermal degradation of the

glue. On exiting the press the boards are placed in a star cooler, which Figure 1.40: Temperature change in a chipboard mattress during and
after pressing.
looks like a paddle boat wheel (Figure 1.39). The cooler is usually large
enough to accommodate a sufficient number of boards so that as boards
leave the cooler their surface temperature will have fallen to about 40 °C.
1.3.11.5 Intermediary storage
From the graph shown in Figure 1.40, it can be seen that the internal
temperatures are likely to be much higher. Depending on the stack size After cooling and stacking, the panels are stabilized through additional
and season, the temperature in the centre can be more than 55 °C after 3 cooling, curing and rehumidification in an intermediary storage.
days of storage. Although transportation based on fork lift trucks is still common place, it
larger mills it has been replaced with fully automated systems (robots),
which are able to handle the big stack sizes with the master formats (press
width x 6.5 m x 1.5 m, weighing approximately 20 t) to an intermediary
storage of some 0.5 ha or more.
One type of such system in based on stack care crane bridges, rolling on
overhead rails installed on building frame. Both movements cross- and
lengthwise are performed simultaneously, which confer a high
transferring rate capacities (transferring speed 3 m/s).

Figure 1.39: A star cooler.

Phenolic bonded boards can be hot stacked. In fact, hot stacking is

recommended for PF bonded boards as their properties are usually seen to
improve by this process. It is said that the improvement is due to
increased cure of the adhesive, which seems very plausible. However, Figure 1.41: Full automatic intermediary crane based storage system
some of the improvement may also be due to continued internal stress (Metso Panelboard).
relaxation permitted by the high temperatures in the stack.

50 51
Irle, Barbu
The systems are keeping the evidence of all piles and towers in a chaotic Wood-Based Panel Technology
way. Some advantages of such modern and full automatic intermediary
storage system are: the perfect horizontal position of panels in piles,
complete protection of edges and surfaces, continuous and precise records burning chamber of the power plant, generating about one third of the
of stored volume, minimal energy consumption, low pollution and dust thermal energy.
generated relative to the traditional fork lift trucks and no supervision The sanding equipment consists in one calibration machine with
needed by operators. The disadvantages are high investment costs, high superposed cylinders and coarse paper grit (60-80) and a finishing
skills required for the maintenance team, in case of accidental stop, machine with two or three pairs of oscilating sanding shoes using fine
complete downtime for the whole line. paper grit (100-150) (Figure 1.43).
Panel stacks after 2-3 days intermediary storage have a temperature below
40 °C and their surfaces firm and the resin completely cured. The panel
conveyers and rollers feed the panels at 40 to 100 m/min depending on
the sanding allowance, thickness and sanding quality target. Generally,
the thinner the panels, the higher the in-feed speed. The sanding paper has
the same width as the panels for this reason for different width are
available different paper sets. The speed of the sanding paper is above 60
m/s. The wear of the sanding paper depends on the resin type and amount,
the furnish quality (recycled wood amount), sanding temperature and dust
Figure 1.42: Full automatic intermediary storage system based on chariots exhaustion performance. The power increase of the sander motors
(Metso Panelboard). together with surface quality changes (e.g. discoloration) are the first
signs of paper wear (in hours). For coating with melamine paper, a grit of
120 is sufficient, for lacquering or direct printing 150 grit or even higher
1.3.11.6 Sanding and Cut-to-Size is needed. The high in-feed speed makes it impossible to control by eye
alone. Typically defects after sanding are dark spots (caused by glue
There are two reasons why the panels are sanded after hot pressing and balls, foreign matters), light spots (caused by dust agglomeration and non-
intermediary storage: resinated furnish), longitudinal traces (caused by broken grits or
1. To remove procured surfaces that are often weak, fibrous and agglomerations on the paper), transverse traces (caused by grit
porous delamination from the paper or deterioration of the paper joint area).
2. To calibrate panel thickness because the panels springback on Major sanding defects are differences in thickness in the cross sections
release from the press and this varies from panel to panel and also (caused by cylinder misalignment) and along the panels (caused by the
within a panel vibration of cylinders due to machine construction or foundations).

The sanding allowances depend on panel thickness, press precision and The thickness of each panels should be measured on-line in at two points
sanding machine (finishing quality). For thin panels (<6 mm), which or more before and after sanding in order to allow a continuous machine
require a high tolerance, 0.3-0.5 mm sanding allowance for both sides is control and setting. Only big dark spots or major defects can be
common. The thicker the panel is, the higher the requested sanding recognized by the operator. Modern on-line self learning video control
allowance (0.8-1.2 mm). Calculated for the whole panel production (PB of both surfaces simultaneously can assure a continuous and reliable
and MDF) approximately 3 % of the line output is processed to sanding check. An important aspect for the process security are the spark
dust. Generally this dust results from the high densified (600-1000 kg/m³) detection resulting during sanding by the overheating of the grit or
and high resinated (i.e. 10-12 % UF resin) face layers of the boards. The contact with metals. The mixture of dry fine dust and air in the exhausting
sanding dust will be air conveyed to a specialised bunker together with pipes or air filter bags increases the explosion risks. The dust filtering
trimmings and cut to size off cuts and injected in to the upper part of the from the exhaust air remains an expensive and time intensive
maintenance task.

52 53
Irle, Barbu
Wood-Based Panel Technology

this target, an optimization of master panel size based on the order of

small parts is necessary before production. Small size parts obtain a better
price, access an important market segment and allow a flexible use of
stored panels. It needs a well prepared production plan, a good the production. Main producers of such systems are Anton, Giben and
organization of the storage areas and many intermediary packet handling Schelling.
zones. The capacity of such a cut to size centre could cover 30 to 50% of

1.4 THE MANUFACTURE OF ORIENTED STRAND BOARD: A

SHORT OVERVIEW
Figure 1.43: Sanding line with calibration (2 heads) and finishing (4 to 6
heads) and grading system (Metso Panelboard). 1.4.1 Introduction
Cut to size equipment consists of many tables with conveyers and pile Oriented Strand Boards (OSB) are multi-layered panels made from
positioners, stationary and mobile circular double saws and racks for strands of wood of a predetermined shape (typically, 15 - 25 mm wide, 75
intermediary storage of small size piles, pilling and packaging units. The - 150 mm long, and 0.3 - 0.7 mm thick) bonded together with a binder
theoretical minimum size could be an A4 paper, but in practice, are (often water resistant) under pressure and heat. The strands in the outer
multiples of doors, furniture size (table, walls or fronts) and flooring layers are aligned parallel to the long board edge and to the production
elements. The height of a mini pile which can be processed in one cut line. Whereas the strands in the core layer are often smaller and can be
does not normally exceed 20 cm and needs paired circular saws (under randomly oriented, or aligned, generally at right angles to the strands of
and over the table) so that relatively thin blades can be used to reduce the face layers. The European specification for OSB is EN300.
material loss.
OSB was first developed in USA based on patents dating from 1935 and
later wood panels based on “veneer strips crosswise oriented”. The 1st
pilot plant in USA started experimental production in 1963. The 1st plant
in Europe operated only in 1978. The growth in demand for OSB is
second only to MDF. In only 35 years, the market acceptance is complete
and around 100 production lines with a capacity above 40 millions
m³/year have been installed around the world. North America is by far the
largest producer of OSB; 85 % of the world’s production in concentrated
in USA and Canada. Europe operates 15 new factories with a total
capacity of 4 millions m³/year.
About 75 % of OSB is used in construction, 20 % for packaging and the
rest for decorative and diverse purposes. The standards EN 300 and EN
13986 classify OSB in four classes: OSB/1 is for general purposes in
indoor (dry), OSB/2 is load-bearing in dry conditions, OSB/3 is also for
Figure 1.44: Cut to size centre (MDF Hallein). load-bearing but in humid conditions and OSB/4 is for heavy duty
construction in humid conditions.
The amount of chips including the panel margins which results by cut to
size the master panels in small parts should not exceed 1-3 %. To reach

54 55
Irle, Barbu
Wood-Based Panel Technology

OSB is usually made with thicknesses ranging from 10 to 32 mm. Often

the most difficult test to pass for OSB/3 and 4 is the IB and MOR after
cyclic boiling and so manufacturers use moisture resistant resins like
isocyanates (PMDI), phenolic based resins (PF, MUPF) or melamine
reinforced UF resins (MUF). The MOR and MOE values observed
parallel to the long panel edge are normally double those observed across
the panel. This is an effect of the strand orientation in the face layer.

1.4.2 Manufacture of OSB

OSB can be produced only from fresh wood and not recovered wood.
Softwoods in form of thinnings, tree tops, cores from veneering etc. are
the most common sources of wood. A typical production line (Figure

Figure 1.45: The main process stations in an OSB production line (Metso Panelboard).
1.45) is very similar to a particleboard line. The differences being:
 The stranding of debarked logs is made in one step via disc flakers
(Figure 1.8) or knife ring flaker (Figure 1.9).
 Some manufacturers use two separate one pass dryers (Figure
1.12) for the core and faces strands generated with two specialized
flakers. This avoids supplementary work and facilitates a precise
drying of the different layers (face m.c. > m.c. core), but requires
more investment.
 Strand grading is achieved in large rotating drums or shacking
boxes (Figure 1.46) that rotate at low speeds in order to avoid
further strand breakage. Fines and shorts are used in the core
whereas long strands kept for the faces, which is the opposite of the
particleboard process.
 Strand blending is carried out in large separate drums without a
central shaft and paddles by atomizing the resins (Figure 1.47)
 The orientation of the strands in the faces and core of mattress
requires specialized formers (Figure 1.48)
 Generally 80 % of OSB panels are sold unsanded, therefore small
capacity sanders are required.

56 57
Irle, Barbu
Wood-Based Panel Technology

1.4.2.1 Strand grading

The strands can be graded whilst still wet and fines and dust can also be
eliminated before entering the dryer thus saving on energy costs. A disc
screen is based on interlocked rotating discs which can separates the wet
strands into fines (dust), face and core layer fractions (Figure 1.46). Such
equipment minimises further strand breakage compared to classical
screens, needs no air cleaning and requires less energy but it is relatively
expensive.

Figure 1.47: Drum resination of strands using atomizing discs (Metso

Panelboard).

Figure 1.46: Disc screens for the cleaning of wet strands (PAL and
Acrowood)

1.4.2.2 Strand resination

Resination of the strands is carried out in large drums (diameter > 2.5 m
and length > 8 m) that rotate at low speed (about 100 rpm) without a
central shaft (Figure 1.47). The drum filling ratio, rotation speed and
retention time (inclination) depends on the production line speed,
indirectly on the panel thickness and resin type. Normally, two resination
drums (for core and faces) operate in parallel. In place of air nozzles
(usually working at high pressure to assure a fog like resin distribution),
many atomizing discs (wheels spinning at >10,000 rpm) fixed in two
rows on the glue pipes are used. Some years ago the face layers were
resinated with MUPF resins and the cores with PMDI, thus avoiding the
risk of steel belts sticking and high cyanide emission during hot pressing.
The new trend in Europe is to glue the faces and core with 3-6 % and 4-
10% PMDI respectively (depending on OSB type). The advantages are:
very low formaldehyde emission, good weathering resistance and high
line speeds (4-8 s/mm). Whereas the disadvantages are: high costs, the
need for permanent application of release agent spraying by nozzles or Figure 1.48: The forming principle for OSB face and core layer
roller application on the steel belt or mat surface and controlled (Dieffenbacher, Siempelkamp and Metso Panelboard).
exhaustion in the press and cooling star area.

58 59
Irle, Barbu
Wood-Based Panel Technology

1.4.2.3 Strand forming

The orientation and proportion of long strands in the face layers,

particularly at the faces, has amarked effect on parallel bending
properties. In the faces, about 80 % of the strands are oriented parallel to
forming belt. The space between the discs can be adjusted so that smaller
strands can fall between the near the core whereas larger ones fall on to
the face (Figure 1.48). In the core the strands are laid across the machine
direction, but, the orientation is less well defined and some cases random.
The strands are fed in to pockets that rotate and because the strands are
long they tend to orientate themselves in the “V” of the pocket. The Figure 1.49: An OSB levelling roller as prepress unit and trimming/metal
height of pockets and discs above the mattress can also be used to control detection in the mattress before pressing in a multidays press
the degree of orientation; the higher they are form the mattress the lower (Dieffenbacher).
the orientation.

1.4.2.4 Strand forming 1.5 THE MANUFACTURE OF MEDIUM DENSITY

The orientation and proportion of long strands in the face layers, FIBREBOARD: A SHORT OVERVIEW
particularly at the faces, has amarked effect on parallel bending
properties. In the faces, about 80 % of the strands are oriented parallel to 1.5.1 Introduction
forming belt. The space between the discs can be adjusted so that smaller
Medium Density Fibreboard (MDF) was first developed in USA from
strands can fall between the near the core whereas larger ones fall on to
hardboard manufacture. Hardboard is made via a wet process that is
the face (Figure 1.48). In the core the strands are laid across the machine
similar to paper manufacture and therefore produces a lot of polluted
direction, but, the orientation is less well defined and some cases random.
water that requires treatment before disposal. Semi-dry processes were
The strands are fed in to pockets that rotate and because the strands are
developed in the 1950s and this led to a fully dry-process method. The
long they tend to orientate themselves in the “V” of the pocket. The
first dry-process MDF factory was built in Deposit, USA, in 1965 and the
height of pockets and discs above the mattress can also be used to control
first MDF factory in Europe is thought to be that built in the former
the degree of orientation; the higher they are form the mattress the lower
German Democratic Republic at Ribnitz-Damgarten in 1973 (Williams,
the orientation.
1995).
1.4.2.5 Strand mattress levelling Since these small beginnings there has been exponential growth in the
installed production capacity of MDF across the world. By the early
Due to the strand shape they tend to naturally lay flat in the mattress and 1990s it was being hailed as a “… one of the most exciting new products
so no prepress is required. A heavy steel drum is enough to level the to come along over the last 50 years.” (Anon, 1995). There was
mattress before hot pressing (Figure 1.49). Trimming mattress remains tremendous optimism in the MDF industry in the early 1990s, which was
are recycled in the core. An electromagnet checked out the ferrous metals fuelled by the rapid expansion of both the market and supply. The
before entering the hot press. optimism turned out to be founded and growth continues at an
exponential rate.
MDF is an engineered wood product composed of fine lingo-cellulosic
fibres, combined with a synthetic resin and joined together under heat
and pressure to form panels. The most commonly used lingo-cellulosic
fibres are wood, but other plant fibres can be used, e.g. bagasse and cereal
straws.

60 61
Irle, Barbu
Wood-Based Panel Technology

MDF is available in a range of thicknesses from 2 to 100 mm and a very

wide range of panel sizes. The density of these panels varies from about
500 to 900 kg/m3; panel density tends to increase as panel thickness
decreases. Panels have smooth, high density faces and are pink-brown to
dark brown in colour, unless a dye has been added during manufacture.
The European definition of MDF is provided by EN 316. The main
grades of MDF are: General purpose, dry (MDF), General purpose, humid
(MDF.H), Load-bearing, dry (MDF.LA), Load-bearing, humid
(MDF.HLS). The minimum requirements for these panels are specified in

Figure 1.50: The main process stations in an MDF production line (Metso Panelboard).
EN 622-5.

1.5.2 Manufacture of MDF

MDF can be produced from a wide range of lingo-cellulosic fibres

including agrofibres (Suchsland & Woodson, 1986) and recycled wood
(Anon., 1995). MDF fibres are normally made by using a thermo-
mechanical pulping (TMP) process. This process uses the combined
action of heat and mechanical energy to break the bonds between the cells
that make up wood. Wood cells are joined by a region called the middle
lamella, which is rich in lignin. Lignin is an amorphous polymer that can
adsorb small quantities of water, and so, its softening temperature is
moisture content dependent. The high temperatures (170-195°C) and
humidities (60-120%) used in the TMP process therefore cause significant
reductions in the strength of the middle lamella region and this increases
the likelihood of failure occurring in the middle lamella when mechanical
energy is applied during the refining process.
The main process steps for the production of MDF are described in more
detail below:

1.5.2.1 Chipping

Wod is chipped in to square particles with sides of approximately 25 mm

and 5 mm thick using drum chipper (Figure 1.7) or as residues from
modern saw-mills using chipper canter. Many other sizes are used,
including sawdust, depending on the source of the raw material.
The chips are screened to remove under- (<2 mm) and over-sized (>50
mm) particles.

62 63
Irle, Barbu
Wood-Based Panel Technology

1.5.2.2 Chip storage

A linear storage system can be closed or open. The closed type is for
Depending on climate region and availability of raw materials there are volumes up to 50,000 m³ (i.e. bark as fuel for energy generation) and
many ways to buffer the amount of chips for 2 weeks to 2 months open type for volumes up to 200,000 m³. The chips are reclaimed by a
production time, see Figure 1.51. The easiest and cheapest option is to conical screw of 10 to 18 m. Also a boom type stacker with crane type
store the chips outdoors on platforms, but, this brings the potential of belt reclaimer and reversible conveyer for in and out feed typical for the
further “contamination” with soil, stones and foreign mater, freezing, or paper industry could be used for high capacity plants.
bio-degradation. Another is to use, round steel or concrete silos with or
without insulation having a volume between 5,000 to 30,000 m³ and 1.5.2.3 Washing
diameters of 25 to 42 m. The advantages compared to the out-door
platforms are: reliable and uniform line feeding (based on screw reclaimer Today chip washing is seen as being a compulsory step in order to
discharge unit), first in – first out discharge, lower moisture content remove the bark, soil, sand and other abrasive contaminants (Figure 1.52).
because the stock is protected from water and snow and no wind blow of In this way the life time of the refiner discs is increased and doubled in
the stock. some cases. The water can be heated to better clean species with high
resin or gum content and to defrost chips stored outdoors in winter. In this
way also conveyer, screws and dryers are protected and energy saved.
The water is kept in a closed loop and cleaned in the same circuit as the
water emanating from the chip plug screw. The surface wetting caused by
washing is thought to improve steaming and refining.
Screw drainer
Scrap separator

Surge bin

Rotary screen

Collecting vessel

Scrap drainer
 To scrap draining unit

Figure 1.52: Chip washing (Metso Panelboard).

Figure 1.51: Raw material preparation and chips handling/storage for the
MDF production (Metso Panelboard).

64 65
Irle, Barbu

a) b)
1.5.2.4 Pre-heating

One decade ago was an optional stage where the washed chips are heated
to 40-60°C at atmospheric pressure in a surge bin (Figure 1.53b). This
softens the particles so that a better plug is formed in the conical screw
feeder (Figure 1.53c), which is a device that allows the continuous
compression (2:1 rate) of chips into the main steaming tube (Figure
1.53a), which is also known as a preheater or digester, whilst still c) d)
maintaining a high pressure in the tube. The plug screw squeezes out
some of the free water in the chips. The volume removed can be some
m³/h depending on line capacities and raw material. This water and a part
of steam condensates must be treated.

e) f)
Figure 1.53: The main parts of a defibrator: a) tube, b) window, c) pluging
screw, d) main engine, e) blow pipe, f) resin injection.
 Wood-Based Panel Technology are fed into the gap (about 1 mm) between two discs via a hole in the
middle of the stationary disc. The surfaces of the discs have a series of
1.5.2.5 Steaming raised bars. At the centre of the disc the breaker bar pattern is coarse
and at the periphery of the disc the bars are much finer (Figure 1.54).
The compressed and squeezed chips are heated with saturated steam at Consequently, as the wood is driven across the radius of the disc by
6 to 10 bar, which creates an internal temperature of 175 – 195 °C for 3 centrifugal forces, it is gradually broken down into its constituent fibres
to 7 minutes depending of the digester size (5-18 m³), wood species, and fibre bundles. Pattern design depends on wood species and chip size.
chip size and required fibre quality. The retention time of the chips in The adjustable gap between the two discs determines the refiner energy
the preheater influences fibre colour and quality and is determined by consumption (150-400 kWh/ton fibre) and fibre quality. The throughput is
the flow rate (screw feeder speed). The digester level is controlled by a typically 15…70 t/h depending on disc and engine size and chip geometry.
radio-active measuring device. The discharge is accomplished by ribbon
screw unit (Figure 1.55). The fibres exit via a discharge opening positioned in the refiner housing
and controlled by a blow valve (Figure 1.53d and e). The purpose of this
device is to continuously adjust the flow rate of steam and fibres in to the
Refining
blowline and thereby control the pressure in the refiner housing. A
The refiner converts steamed chips into fibre bundles. The refiner diverter valve (Figure 1.53f) after the blow valve can divert the fibre
housing is pressurised with saturated steam (8-10 bar). Most refiners flow to the start-up cyclone or to the blowline. The start-up cyclone
have two discs (diameter 44 – 72 inch), one stationary and another that guarantees a safely start-up and shut-down of the system (under steam
rotate at about 1500 rpm by powerful motors (typically 5-12 MW pressure). Only if the fibre quality and resination are satisfactory and the
depending on the diameter of the discs), see Figure 1.55. Wood chips process parameters constant (dryer) the valve may be switched to
production (blowline).

66 67
Irle, Barbu
Wood-Based Panel Technology

Surge Bin
1.5.2.6 Resinating (blowline)

Agitator The fibres are discharged from the refiner down a pressurised blow line, a
Steam pipe of 80 to 120 mm diameter, which conveys the hot, wet fibres and
steam to the dryer. Blending of the fibre in the blowline under high
Discharge
pressure facilitates a uniform resin distribution due to the steam
screw
Preheater expansion, which cause high turbulent flow, rapid accelerations (that
Screw feeder separate fibres) to high velocities of up to 100 m/s. The adhesive, usually
Preheater Defibrator
a formaldehyde-based resin (UF), is injected via water cooled 3-5 mm
Main motor Defibrator nozzles at high pressure (12-14 bars) in to the blow line with other
feeder
additives, e.g. water, hardener, repellents, colours, etc. The amount of
Steam Foundation
block adhesive added will vary between 8 and 15% (resin solids/oven dry wood
basis), depending on grade of panel being made. Disadvantageous of this
blowline blending is the fact that the resinated fibres have to pass
Steam
connection
through the flash dryer (110…140°C) which tends to precure some of the
Defibrator
Fibre to dryer to preheater
Fibre to
resin. To compensate for this a higher resin amount is required (compared
Top
dryer to dry resination).
Figure 1.54: A typical pressurised disc refiner based system with
preheating unit and digester tube suitable for the production 1.5.2.7 Dry blending
of thermo-mechanical pulp (TMP) (Metso Panelboard).
A new approach to fibre blending has two stages. In the 1 st stage the resin
is applied via the blowline at an addition level of only 3-5% adhesive.
The wax is applied as usual via refiner discharge screw. The 2 nd resination
stage takes place (like 30 years ago) in a separate blender (Figure 1.56)
after fibre drying. Resin is applied by air-nozzles, mounted on a ring in
front of the open in-feed (negative air pressure because of exhausting
funs) of the drum blender. The internal paddles generate a turbulent
fibre flow ensuring effective blending. The blender and out-feed pipe are
water cooled to avoid resin and fibre deposits. After 2 nd dry fibre
resination a gentle drying to the target moisture content is recommended
by IMAL. It is claimed that resin savings of up to 35 % compared to
conventional blowline are possible.

Figure 1.55: Photographs of the discs and segments found inside of a

Figure 1.56: Dry fibre blender with open in-feed and spraying nozzles ring
typical pressurised refiner used to make MDF fibre.
(IMAL).

68 69
Irle, Barbu
Wood-Based Panel Technology

1.5.2.8 Drying

The wet resinated fibres are blown through a flash tube dryer at around
30 m/s. The dryer is often 1 to 3 m in diameter, over 100 m long.
Cyclones (3 – 5 m diameter) at the end of the dryer separate the dried
fibre from the steam. The moisture content of the fibres after drying is
approximately 8 – 12 %. The flash dryers are often heated directly with
cooled gases (<200 °C) from the energy plant, in order to avoid an
extended precurring of the resin. They are designed as one stage dryer
(longer and less temperature control) or two stages dryers, which
requires two cyclones, two fans (600 kWh and 300 kWh), more energy,
but improved drying and lower precurring (Figure 1.57). The moisture
content of fibres at the end of 1st stage is around 40 % and the gas
temperature in 2nd stage is low (<100°C). Waste air from 2 nd stage can be
recirculated to 1st stage and mixed with fresh hot gases due its lower
humidity.

Figure 1.57: A two stage dryer with waste air recirculation (Metso Panelboards).
70 71
Irle, Barbu
Wood-Based Panel Technology

1.5.2.9 Forming
1.5.2.10 Pre-pressing
The dry fibre is pneumatically conveyed to mat formers via a system of
The mat is then passed to a continuous prepress (Figure 1.58). Here the
classifiers (sifter) and filters (Figure 1.58); these latter devices are
mat is squashed to reduce its air content and to increase its density. This
designed to remove any fibre clumps. The mat formers are designed to
reduces the time required for hot pressing and avoids fibre dislocations at
distribute an even layer of fibres (no layering) onto a continuously
the press in-feed. Before pre-pressing a 38 m thick MDF has a mattress
moving belt. The belt speed varies with the panel thickness being made.
height of about 1 m and after the mat height is around 350 mm.
The fibre mat forming requires a dosing bin with integrated weight scale,
Mattresses over 200 mm height (>18 mm panels) are often preheated
and a spreader system. Due to the fibre consistency, handling and
(section 3.9).
spreading is different to that for particleboards. Only one forming head is
used to spread the homogenous fibre mattress. The dosing and metering
bin has to assure a constant fibre flow to avoid compressing at the 1.5.2.11 Hot-pressing
spreader head. A back-rake conveyor built on the top of the bin keeps In principle, the hot-pressing process is similar to that for particleboard
the fibres at a constant height when they reach the doffing rollers. described in 1.3.10. The hot-press applies a combination of heat (180-
A typical mat for an 18 mm thick MDF is 680 mm high and has a bulk 210 °C) and pressure (0.5-5.0 MPa) to consolidate the mat and convert it
density of 23 kg/m3. to MDF. The pressing stage can be used to manipulate panel properties by
altering the panel’s density profile, which is the variation in density
through the thickness of the board, see Figure 1.59.

Fibre infeed
system

Mat
Fibre Z-sifterTM Dosing and UniformerTM Pre-press Hot-
storage Reject press
Metering
bin system
bin

Figure 1.59: A typical density profile for an 18 mm thick MDF panel.

Immediately after pressing the panels are stood on their edges to expose
their faces to ambient air to facilitate rapid cooling.
Finishing entails trimming, sanding and cut to size. Various added-value
steps such as laminating, profiling and painting may be applied.

Figure 1.58: Fibre weighing, sifting, mat forming (mechanical) and

prepressing until continuous press in-feed and air forming
with spike roll (Metso Panelboards and Binos).

72 73
Irle, Barbu
Wood-Based Panel Technology

1.6 THE MANUFACTURING OF PLYWOOD: A SHORT

%, and so plywood panels are very stable. Additional veneers must be
OVERVIEW
added in pairs so as to maintain panel symmetry.
1.6.1 Introduction Figure 1.60 shows a cross-section of a 7-ply plywood with the line of
symmetry marked. Any dimensional movement in the top half should be
Plywood panels usually have an odd number of veneers. The terms 3-ply counteracted by similar movement in the bottom half.
and 5-ply are commonly used and refer to the number of veneers used to
make a panel. Plywoods with seven or more veneers are also made. The There are two ways of making thicker plywood panels. One is to use
need for an odd number of veneers is caused by the fact that wood is an thicker veneers and the other is to use more veneers. The advantage of the
anisotropic material, i.e. it has different properties in its three main latter method is that it produces more homogeneous panels. This is
directions, longitudinal, radial and tangential. because the ratio of veneers in the two directions tends to one as the
number of veneers increases. For example, in a 3-ply the ratio is 2:1, in a
An example of this anisotropic behaviour is the high strength of wood 5-ply it is 3:2, in a 7-ply 4:3 and so on. Unfortunately, increasing the
parallel to the grain, e.g. longitudinal compression, compared to strength number of veneers also increases manufacturing costs.
perpendicular to the grain, e.g. tangential or radial compression. Another
is the difference in the dimensional stabilities between the three grain
directions. If a dry piece of wood is soaked in water it will swell. It will
swell most in the tangential direction and least in the longitudinal
direction, with the radial directional movement between the two but
nearer the tangential movement. The relative movement between the three
directions varies from species to species but is typically around 20:12:1
(tangential:radial:longitudinal). It is this dimensional anisotropy that
makes it necessary for plywoods to be made with an odd number of Figure 1.60: The cross-section of a 7-ply plywood with the line of
symmetry shown.
veneers.
Each veneer in a plywood is laid down with its grain at right angles to its
neighbour. This is done to minimise the strength anisotropy in the panels. 1.6.2 The Manufacturing Steps
In other words, the longitudinal grain of one veneer reinforces the
tangential grain of adjacent veneers. Bonding just two veneers together, 1.6.2.1 Log preparation
so that their grains are at right angles to one another as in a plywood,
dramatically reduces strength anisotropy but, such a panel would be The quality of plywood is largely dependent on the quality of veneer.
dimensionally unstable. To visualise this, imagine the cross-section of a 2- Cutting smooth-faced veneers requires well maintained equipment and
ply panel, with the top veneer showing longitudinal grain and the bottom the wood to be as easy to cut as possible. The strength of wood is
transverse grain. When placed in water the bottom veneer, in this cross- dependent on its moisture content and so the logs should be kept wet and
section, will swell 20 times more than the top veneer, because of the ratio certainly above their fibre saturation point. Another reason to keep the
given above, and so the panel will cup. logs wet is to prevent them from cracking as a split in a log will not allow
a continuous ribbon of veneer to be produced.
To make a stable panel, therefore, a third veneer must be added. This
creates a panel that is symmetrical through its thickness, i.e. the bottom Before veneer can be cut the logs must be prepared so that the wood can
half is a mirror image of the top half. So when a 3-ply panel is placed in be cut efficiently and produce a smooth, even veneer. The peeling step
water, the veneers still try to swell, but this time the swelling forces are uses a sharp knife that is easily damaged by hard objects like stones, grit,
balanced and the panel remains flat. Plus, the amount of swelling is nails, etc. The bark often has such hard objects in it as a result of tree
limited by the longitudinal swelling of adjacent veneers, i.e. generally <1 felling. Since the bark cannot be cut in to useful veneer it is best to
remove the bark before the peeling step so as to protect the knife.

74 75
Irle, Barbu Wood-Based Panel Technology

Figure 1.62: Knife ring debarkers are often used to clean logs prior to
peeling.

Wood strength can be further reduced by heating the log and this can
help produce good veneer but the wood should not be heated too much,
making it too soft, as fuzzy grain will result in which fibres are pulled out
instead of being cut.
The wood is not always heated before peeling, but, it is necessary for high
density species, or when cutting thick veneers, and when the logs are
frozen. Warming a log lowers its strength and reduces wear on knife and
especially knots which can cause nicks in knife. As will be shown later,
warming the wood also increases its flexibility enabling it to survive some
of the strains that are applied to it during peeling.
Logs are heated with steam, hot water or a combination of the two.
Figure 1.61: An example of a plywood production line flow (Zudrags). Steaming carries the risk that the log will begin to dry and split, hot water
reduces this. The temperature that a log is heated to depends on the
Most plywood factories will cut their logs in to two lengths; a longer one species. Generally, the higher a log’s specific gravity the higher the
of around 2.7 m and a shorter one of around 1.3 m. The longer logs temperature. Certainly, low density species, e.g. poplar, do need to be
provide veneer for the plywood faces, where the grain direction is heated. Whereas Okumé, Beech and Oak are heated to 65°C, 80°C and
normally parallel to the longest edge of the panel, and internal veneers 85°C respectively.
where necessary. Lower quality logs are generally cut to shorter lengths,
especially if the log is curved, so as to reduce the amount of veneer A side effect of heating is colour change. Perhaps that most well known
discarded in forming the perfect cylinder needed to produced full size is the pink colour that is caused when beech is steamed whereas
sheets. unsteamed beech has yellow/straw colour. The colour change is
associated with chemical changes in the extractives present. Nearly all
sapwood of species darken on heating as do many heartwoods and

76 77
Irle, Barbu
Wood-Based Panel Technology

sometimes the change may not be desirable, especially if it is non-

In general, the thicker the veneer, the greater the chance of tension checks
uniform and the veneer is wanted for face veneers.
being formed. This is because the minimum radius of curvature for a thick
The heat should be applied carefully. High water temperatures combined veneer is larger than that for a thin veneer or in other words, thin veneer is
with short heating periods cause temperature differentials where the more pliable than thick veneer. If the checking is very bad, then the
outside and ends of the log are softened but inside is still cold and hard. veneer can break.
This situation will cause the veneer to break when the knife hits the cold
To help minimise tension checks, the wood is kept saturated and often
zone of the log. In extreme cases it can cause chuck spin-out, where the
heated to ensure that it is as pliable as possible. A compression force
chucks which spin the log at its ends continue to turn when the log stops
applied just ahead of the knife tip with a nose bar can also reduce tension
on the knife edge (see photo).
checking.
The heating time should be long enough to give an even temperature
throughout log, i.e. the temperature differential between the outside and
the innermost part of the log should not exceed 6 °C. Clearly the heating
time is proportional to the diameter squared of the log, so if the diameter
is doubled then heating time is quadrupled.

1.6.3 Peeling
Figure 1.64: A magnified view of edge of a plywood showing tension checks
The principal aim of veneering is to form continuous sheets. Veneering (small splits).
knives tend to have high rake angles (typically around 70°) and
sharpness angles of around 20°, see Figure 1.63. It goes without saying The peeling of veneer is achieved by rotating a log, or more correctly a
that the knife should be sharp to ensure that the knife geometry is bolt, against a sharp knife. The bolt is rotated by “chucks” inserted in its
maintained all the way to the cutting tip. ends. The chuck teeth, of which there are many designs, provide the
torsion resistance needed to rotate the bolt with enough force to cut the
veneer from it.

Figure 1.65: Left one of many chuck designs and right the chucks inserted
Figure 1.63: A diagram showing the main aspects of veneering. in the log ends.
The veneer is subject to tremendous stresses as it is cut away from the Profitable plywood manufacture is dependent on maximising the yield of
bolt. These stresses cause the veneer to split at regular intervals on the veneer, and preferably in full-size sheet form, from the bolt. Since logs
knife side of the veneer. These splits are called tension or lathe checks. are not perfectly round and often contain defects, simply inserting the
Their presence changes the properties of the surface, particularly in terms chucks that spin the bolt in to the centre might not be the most efficient
of permeability. Consequently, the face without the tension checks is solution.
called the “tight side” and the face with the checks is the “loose side”.

78 79
Irle, Barbu
Many mills still rely on the trained eye of a skilled operator to align a log Wood-Based Panel Technology
in the lathe. A shadow mask shone on to the log ends is often used to
assist decisions (see Figure 1.66). Modern mills tend to use laser
scanning technology which produces a 3D image of the log that a 1.6.3.1 Veneer processing
software optimisation program uses to decide where best to place the The veneer ribbon is often wound on to a bobbin for processing, i.e.
chucks. clipping to size, dry, etc. at a later date. This method is often known as
The size of the chuck also has an effect on yield. On the one hand, a large “peel and reel”. Although still very common in traditional mills around
chuck is needed to avoid chuck spinout and at the start when a log is the world, modern mills tend to favour processing the veneer ribbon
heavy. On the other, a small chuck diameter allows the log to be peeled to immediately after it leaves the lathe (Baldwin, 1981).
a small spindle. Most modern mills achieve both by using retractable Mills that tend to peel small diameter logs, e.g. birch in Scandinavia, send
chucks which consist of a central chuck surrounded by an outer ring. the veneer directly to a continuous dryer that is close coupled to the lathe.
Initially, both the ring and chuck are inserted thus providing a large The advantage of this approach is that the veneer is subject to only one
diameter chuck. As the log diameter diminishes the outer ring is retracted clipping and grading step, whereas, for all other methods there is a
thus permitting peeling to a smaller diameter. clipping and grading step before drying followed by a regarding and
possible reclipping step. This approach is not used widely because it is
technically difficult to dry a continuous ribbon of veneer. The main
difficulty is to ensure that speed of the belt carrying the veneer through
the dryer adjusts for the shrinkage that takes place. The shrinkage
effectively shortens the length of the veneer and therefore the speed that it
should be conveyed through the dryer.
Table 1.5: The three different approaches to veneer processing.
Ribbon Processed
Ribbon Processed Later
Ribbon Dried
Immediately Immediately
1. Peel and reel 1. Peel to ribbon 1. Peel to ribbon
2. Clip 2. Clip 2. Dry
3. Sort 3. Sort 3. Clipping
Figure 1.66: Two methods of aligning logs: left a shadow mask and right a 4. Dry 4. Dry 4. Veneer grading
laser scan of log as it is slowly rotated. 5. Veneer grading 5. Veneer grading 5. Jointing
6. Clipping 6. Clipping
As a bolt’s diameter decreases then its resistance to the bending forces 7. Jointing 7. Jointing
exerted by the knife and nose bar falls rapidly because bending resistance
is proportional to the diameter cubed. If the log is not supported in some
1.6.3.2 Drying
way then the bolt will deflect away from the knife, causing thicker veneer
to be cut from the log ends than in the middle. The resultant spindle will Whatever the method of veneer, eventually the veneer must be dried. Its
be barrel shaped resulting in the loss of veneer. moisture content must be reduced to between 6 and 12 %. Considering
This situation can be mitigated by supporting the back of the bolt with a that the veneer has a moisture content well above FSP, this step requires a
backing roll. Motorised backing rolls exist that also help to rotate the log lot of energy.
permitting even smaller chuck diameters. Various drying defects will occur in veneer unless it is dried carefully.
Spindless peelers can peel down to 50 mm cores, which can significantly Wood is very weak perpendicular to the grain and so veneers often split.
reduce waste when peeling small diameter logs. The technology is not It is surprising to learn, perhaps, that splitting is not generally caused by
widely used however. drying but by rough handling as they are splits initiated in green state,
perhaps in the tree, or during veneering.

80 81
Irle, Barbu reaction wood or improper cutting. It is undesirable because of the difficulty
of laying up a board with veneer that does not lie flat.
Veneer often has a buckled or wavy appearance. This may be caused by Even though veneer is relatively thin collapse and honeycombing defects are
still possible if the veneer is dried too fast and at too high a temperature. Wood-Based Panel Technology
The rippled surface can sometimes be removed by steam reconditioning.
Likewise, case hardening, again caused by drying too fast which generates
stresses in the veneer surface. Many veneer sheets will require patching. This is achieved by stamping
out the defect, e.g. a knot, included bark, dark stain, etc, with a standard
Drying can also alter the veneer surface so that glue does not wet it so sized punch. The hole is then filled with a patch with the same shape and
well. For example, extractives can migrate to the surface and alter the size that has been previous punched out of clear veneer. Even more
contact angle between the glue and veneer surface. Alternatively, the veneer is jointed together to form full size sheets. There are many
thermal energy can cause cross-linking of cellulose. different joint methods: adhesive tape, hot-melt impregnated string,
stitching, edge gluing, scarf jointing to name a few. Modern mills use
1.6.4 Dryers compositors that combine veneer pieces of varying widths in to one long
piece that is then cut in to full size sheets.
The two most common dryers are the roller dryers and belt dryers. The
latter allows veneer to be fed in to the dryer with its grain angle Many people are surprised to learn that board strength is not significantly
perpendicular to the sense of feed. Consequently, they can be used to dry affected by the use of veneer sheets that have been joined together. This is
continuous ribbons of veneer. Roller dryers are energetically efficient because the strength of wood perpendicular to the grain is low and so the
because the rollers are heated and since these carry the veneer through the presence of a jointed piece is little different to a whole piece.
dryer the direct contact with the dryer permits heating through
conduction, which is more efficient than by convection. The rollers are 1.6.6 Board Layup
typically 100 mm in diameter.
Despite huge advances in computer technology and automation, plywood
Very high air speeds of 1000 - 1500 m/min are used in the dryers to panels are still essentially compiled by hand. This is because of the large
minimise the thickness of the boundary layer of air surrounding wood variation seen in the form and colour of veneers; they are not flat, not
because it is the thickness of this layer that determines drying rate. The air particularly flexible and are often split. These factors make it difficult for
is blown vertically onto the veneer faces facilitating fast drying. an automated system to cope with veneer. Humans can recognise veneer
defects and can modify the layup accordingly.
Dryers are split into sections. The temperatures and humidities can be
controlled independently and used to help ensure uniform drying. The The most common method of applying adhesive is with a roller coater
first stage tends to have the highest temperature of around 150-200 °C. which applies a known amount of adhesive to both faces. Adhesive is
This is acceptable during initial drying because of the high moisture applied to both faces of alternate veneers and not every veneer. For
content of the veneer. Subsequent stages are cooler to produce uniform example, in a 3-ply panel adhesive is only applied to the middle veneer
dried veneer. and for a 5-ply only to veneers 2 and 4. Other adhesive application
systems are used including: curtain coating, spraying and extrusion.
1.6.5 Veneer Preparation Discussion of these is beyond the scope of this short review and readers
are recommended to read Baldwin (1981).
Veneer grading follows drying. In simple terms, sheets are graded as
being best face quality, back face quality and interior quality, i.e. veneers The grammage of adhesive applied varies with grade of panel, type of
to be used within the panel. Actual grading rules contain many more adhesive, wood species and, in particular, the surface smoothness of the
grades and vary from species to species and type of plywood being made. veneer. Less adhesive is necessary for smooth faced veneers. A typical
grammage will range from 140 to 240 g/m² per face (or glueline). It is
sometimes quoted as a double face measure (and so double the figures
given in the previous sentence) because the roller coaters apply adhesive
to both faces at once.
When laying-up plywood, the veneers placed on the outer faces should
be the orientated so that their tight faces are exposed. If this is not done,
then there is a risk of excessive checking on the surface with time as the
veneer moves with changing atmospheric conditions. Some mills have

82 83
Irle, Barbu
1.6.7 Pressing
semi-automatic lay-up lines where surface veneers are handled by
machine. Panels are generally cold prepressed after lay-up. The aim is to flatten the
panels and allow the adhesive to develop tack. Both factors make it easier Wood-Based Panel Technology
to feed the panels in to the hot-press. The prepress operates cold and the
panels are pressed as a bundle in a single-daylight press.
For example, 10 mm thick panel would have a press cycle of around 6
Prepressing is synchronised with the hot press. Modern mills with large minutes. As mentioned above, this is much longer than for particle and
multi-daylight presses often use an automatic loading and unloading fibreboards, which are pressed between 4 and 10 seconds per millimetre
cages to minimise dead time between cycles and improve overall of panel thickness. So a 10 mm thick particleboard may be pressed every
production capacity. The cages also help to ensure that all panels are 50 s.
subjected to the same temperature histories. The presses in such mills
also tend to have simultaneous daylight closing mechanisms to ensure The specific pressure applied to the panels is around 1 MPa; a little
that each panel is pressed to the same pressure and for the same time. This lower for low density plywoods, e.g. poplar and higher for high density
results in more consistent panels. plywoods e.g. beech. Once again, this is quite different of particle and
fibreboards, which are pressed at around 3 MPa. Pressure is applied in
Panels are fed into presses by hand in older mills with smaller presses and any gluing situation to ensure adequate contact between the bonding
these presses tend to close by pressing one day-light against another. So surfaces. Since plywood is made of “flat” sheets a low pressure is
for an upward stroke press, the bottom day-light will close first followed sufficient to achieve the contact required; particle and fibreboards on the
by the second and so on. On opening it is the top daylight that opens first other hand are made of elements with varying sizes and shapes that are
and so it is clear that the panel in the top daylight is pressed for a shorter randomly orientated so much higher pressure is required to force the
time than the bottom panel. Plywood press cycles are much longer than elements close enough for the adhesive to bond them together.
those for particle- and fibreboards and so the difference in the pressing
cycle of the plywood in the top daylight compared to the one in the The press pressure peaks near the start of the cycle when the wood is
bottom is relatively small in percentage terms, but, it is different and does cool. The pressure is reduced during the cycle to minimise induced
not help panel consistency. compression and to maintain panel thickness.

The press temperature for panels made using amide-based glues, e.g. 1.6.8 Finishing
UF and MF, is between 100 to 120 °C. This is much lower than the
temperatures used for particle- and fibreboards using similar adhesives. Boards are trimmed so that they have straight edges and 90° corners.
The lower temperatures are associated with lower internal steam pressures Often the panels are sanded to ensure panel calibration.
and so less risk of blisters, lower energy costs but also longer press times.
Repairs may be necessary to panel faces and edges. Face repair is most
The panels are normally cooled briefly before being stacked ready for
often needed when using a species with dark knots. These can be routered
further processing.
out and the hole filled with a coloured hot-melt plastic. The plastic cools
Many types of plywood are made with phenolic-based adhesives, which and solidifies very quickly and once sanded it is difficult to see the repair.
require more energy to initiate cure. This is achieved using higher press
Plywood manufacturers have developed a wide range of finishes for their
temperatures, typically 160 °C and “hot stacking” immediately after
products in order to increase their value and to differentiate themselves
pressing so that the panels remain hot for many hours.
from other producers. The finishes include: non-slip surfaces, phenolic
The length of a cycle can be calculated using a simple “rule of thumb” impregnated paper finishes for concrete form work and various colours.
which is 1 minute plus 30 seconds for each millimetre of panel thickness.
1.7 A POTTED HISTORY OF WOOD-BASED COMPOSITES

Veneers were initially produced manually, by sawing, and later,

mechanized, by means of large diameter rotating saws and special veneer
frame saws. It was only in 1818, in France, that the first rotary cutter
(lathe) was patented for the production of veneers. The first slicer for
decorative veneers was patented somewhat later in 1870. These two
inventions have been modernized and upgraded ever since, but they are

84 85
Irle, Barbu applications.
Particleboards originated in Germany. The first mention of the manufacture
still in use today. In 1934 the waterproof synthetic resins (PF glues) were of such boards dates back to 1887, when Hubbard made a so-called “artificial
produced, which made possible the production of plywood for exterior
wood” of wood flour and albumin-based glues, consolidated under high Wood-Based Panel Technology
temperature and pressure. In 1889 Kramer obtained a German patent for
his method of gluing of wood shavings onto a flax fabric. Then he layered
the fabrics in a similar way to plywood (alternatively cross-oriented). In particleboards were made. Between 1938 and 1940, the German company
1905 Watson (USA) developed a method to produce boards of thin square Torfit obtained two patents for the production of particleboards; liquid
particles (flakes). Today, his patent is still at the centre of flakeboard and resins were used for gluing particles and formed in to a particle mat which
OSB manufacturing. was subsequently hot-pressed.

Beckmann (Germany) suggested in 1918 a new technique for the The same company built the first plant for the industrial manufacturing of
production of a layered board with a core of compressed particles or particleboards in 1941 in Bremen (Germany). Fahrni obtained in 1943 a
wood flour and veneer faces; it was the forerunner of the products known French patent for the production of particleboards. He later developed
at present as Com-Ply, a veneered particleboard or plywood with a equipment (Novopan) for particleboard plants. Kreibaum (Germany)
particle core. Freudenberg (Germany) mentions in 1926 the use of planer produced between 1947 and 1949 the first extruded, as opposed to flat-
shavings glued with resins of the epoch, to produce boards. In 1934 Nevin pressed, particleboards.
(USA) recommended the mixing of coarse sawdust and shavings with Although the idea of producing wood fibres was one and a half century
resins and to harden them by hot-pressing for board manufacturing. old, it was only in 1844 in Germany that a practical method for producing
Antoni (France) obtained, in the same year, boards from a combination of wood fibres by mechanical processing in the so-called “mills” could be
wood fibres and particles glued with urea- or phenol-formaldehyde resins. developed. In England, in 1851, a chemical system was invented for
1935 was a very innovative year as researchers in France, Germany, wood refining, in order to obtain fibres for paper production. In its
Japan and USA suggested a range of new WBP products. Samsonow competition with other raw materials used to produce paper, wood
(France) suggested the use of cross-alternating veneer strips for a became the most economically profitable material, due to its low cost and
particleboard, similar to the present products OSB. Satow (Japan) availability. Fibreboard structure is closer to that of paper products than
obtained an American patent for the manufacture of boards made of 75 of other wood-based products.
mm long chips, randomly arranged in order to reduce board warpage. The first insulation board plant built in 1898 in England answered the
Roher (Germany) outlined the possibility of gluing wood particles onto efforts to capitalize the large quantities of oversize fibre bundles removed
the surfaces of a plywood core within a single pressing operation. from pulp.
In 1936 Loetscher (USA) presented a patent concerning the automated In 1914 Carl Muench made a softboard from wood fibres, which had
manufacturing of particleboards. In 1937 Chappuis (Switzerland) insulating properties and this explains why this type of product is often
described the manufacture of particleboards from dry particles, by called insulation board. “Insulite” is the trade name of this fibreboard,
applying powder resins (Bakelite). produced in a large plant on the same location since 1916 (USA). A
Another Swiss, Phol, presented in 1936 his patent for the alternative use second plant for insulation fibreboards was built in 1931. The
of long veneer strips (50-200 mm), now used for making load-bearing manufacturing process is a modified paper making method. A large tank
structures from veneer based composites, i.e. LVL. stored the mixture of fibres and water. This mixture was then
“dehydrated” on a sloping and continuously moving screen/sieve, thus
During World War II, when the production of synthetic resins was forming a fibre mat. By drying the wood fibre mat in an oven, porous
improved, the first attempts for the industrial manufacturing of fibreboards with densities between 160 and 400 kg/m³ could be made.
This wet fibre-board process has been improved over the years and is an
established manufacturing method for fibreboards. It has relatively low
production costs because of the relatively small amounts of binders
present in the product. Muench investigated alternative sources of fibres
and also used agriculture by-products like corn, wheat straw, bagasse, etc.
The first plant for insulation fibreboards made with non-wood ligno-
cellulose raw materials was built in 1920.

86 87
Irle, Barbu pressurised with steam to pressures as high as 8 MPa causing a temperature
of around 290 °C. The pressure is suddenly released causing the softened
An alternative method for pulping wood using the steam explosion chips to pass rapidly through a grid which breaks up the chips in to coarse
technique invented by William Mason was introduced in 1924. Wood fibres. In 1926 in the USA, the Mason Fibre Company started manufacturing
chips are introduced in to a vessel called a "gun", which was then the first hardboard using the Masonite process.
Asplund (Sweden) proposed the thermo-mechanical pulping process in Wood-Based Panel Technology
1934. This technology has been subsequently widely adopted for making
fibres both in the paper and building board industries. The principal steps
of this method are described in section 5.2. The first two dry-process hardboard plants were built in 1952 in US as an
adaptation to industrial production of the semi-dry process. Dry-process
The excellent properties of the fibreboards are the result of the bonds hardboards have lower mechanical properties than those made using the
created due to lignin reactivation under high pressure, temperature and wet process, due to the high water absorption capacity of the remaining
moisture conditions. In order to improve the fibreboard properties when “unbound” lignin and hemicellulose, but have the advantages of smooth
used in contact with water, additional water-repellent substances faces on both sides, lower densities, thickness variability and a high
(paraffin/wax) or synthetic binders (phenol-formaldehyde resins < 3%) workability.
can be added. This technology is widely known as the wet process and
the fibreboards thus manufactured are internationally known as The first MDF plant was built in 1965 in Deposit, USA. The first MDF
hardboard when panel densities exceed 800 kg/m³ or softboards if the factory in Europe is thought to be that built in the former German
density is less than 400 kg/m³. The particularity of this process lies in that Democratic Republic at Ribnitz-Damgarten in 1973. Other factories were
water is used to convey fibres and to form the fibre mat. Another specific built in Eastern Europe: 1975 Busovaca (today Bosnia-Herzegovina) and
aspect of the wet process is that a screen is used in the press, at the bottom 1978 Illirska Bistrica (today Slovenia).
of the fibre mat, so that water and steam can readily escape during hot During the last decade, a distinct category of products have appeared,
pressing. As a result, the screen pattern is embossed on the backside of namely the Low-Density Fibreboards (400–600 kg/m3).
the finished hardboard. The production of fibreboards using the wet
process has decreased drastically during recent decades. The main reason The use of multi-opening and today, continuous presses, was the next
for this is that the process is inefficient for making thick panels stage in developing the manufacturing technologies for particleboards,
(essentially they are limited to panels of 12 mm or less). In addition, the MDF and OSB. These three categories of WBP now dominate the sector
strict environmental requirements imposed on water pollution in highly in terms of production volume (world 70%, Europe 90%). Equipment
industrialized countries have significantly increased manufacturing costs. production and technology improvement, as well as WBP development,
This technology has patent environmental shortcomings regarding the has remained until the late 1990s in German hands, which has fitted out
recycling and treatment of waste waters, which contain large amounts of almost the entire WBP factories all over the world. The first continuous
fibres and organic substances (cellulose, lignin, wax, etc.). Nevertheless, press made outside of Germany was produced in China and installed in
softboards still hold an important position in the market of insulation the Czech Republic.
boards (i.e. made of polyurethane foams or of mineral-bonded fibres).
The field of wood-based composites has dramatic changes ever since the
Research dating back to 1945 aimed at replacing the wet forming system 1950s in terms of production capacities and technological developments.
with a dry forming process using an airflow. Much of the research was The prospect of producing particles and fibres with various dimensions
conducted in the USA and it ultimately resulted in the development of and shapes, the use of new types of synthetic resins, of modern
Medium-Density Fibreboard (MDF). technologies and specially designed, reliable, partially or completely
automated equipment have given a boost to WBP development. It was the
beginning of a new age, not only in the production of WBP, but of other
wood-based products as well.

88 89
Irle, Barbu Balwin, R.F. (1981). Plywood Manufacturing Practices. Miller Freeman
Publications Inc., San Francisco.
1.8 REFERENCES Barbu, M.C. (2009): Fibre generation and processing. Presentation during the
4th International Wood Academy, Walailak University, Nakhon Si
Anonymus. (1995). Magic number in sight for world production. Wood Thammarat, 05-10.10.09
Based Panels International 17(October):10-11.
Barbu, M.C. (2009): Panel finishing. Presentation during the 4th International
Anonymus (1997). Wood Based Panels – Transition from British Wood Academy, Walailak University, Nakhon Si Thammarat, 05-10.10.09
Standards to European Standards. Wood Panel Industries Federation,
Edition 3, July 1997. Behrens, J. (2009): Pressing of thin panels. Presentation during the 4th
International Wood Academy, Walailak University, Nakhon Si Thammarat,
Axer, J. (1975). New Concepts for the Production of Particleboard. In 05-10.10.09
Proceedings of the 9th International Particleboard Symposium, W.S.U.
Bolton, A.J., Humphrey, P.E. & Kavvouras, P.K. (1989). The Hot Wood-Based Panel Technology
Pressing of Dry-formed Wood-based Composites. Part III. Predicted
Vapour Pressure and Temperature Variation with Time, Compared with
Experimental Data for Laboratory Boards. Holzforschung 43. Elias, R. and Irle, M.A. (1996). The Acidity of Stored Sitka Spruce Chips.
Holz als Roh-und Werkstoff 54(1):65-68.
Bolton, A. J., P. E. Humphrey, and P. K. Kavvouras (1989): The hot
pressing of dry-formed wood-based composites. Part VI: The importance Fischer, K. (1972). Modern Flaking and Particle Reductionizing. In
of stresses in the pressed mattress and their relevance to the minimisation Proceedings of the 6th International Particleboard Symposium, W.S.U.
of pressing time, and the variability of board properties. Holzforschung Fruehwald, A., Thoemen, H. (2008): The rise of wood based panels.
43(6): 406-410. Presentation during the 3rd International Wood Academy, University of
Deppe, H-J, Ernst, K. (2000): Taschenbuch der Spanplattentechnik, 4. Hamburg, 25.02-07.03.08
Auflage, DRW- Verlag, Leinfelden- Echterdingen Galbraith, C.J., Cohen, S.C., Ball, G.W. (1983). Self-Releasing
Dinwoodie, J.M. (1981). Timber Its Nature and Behaviour. Van Nostrad Emulsifiable MDI Isocyanate: An Easy Approach for All-Isocyanate
Reinhold Company, London. Bonded Boards. In Proceedings of the 17th International Particleboard
Symposium, W.S.U.
Dunky, M. Pizzi, A. and Van Lemmput, M. (2002). Wood adhesion and
glued products. Working group 1: wood adhesives. State-of-the-art report Geimer, R.L. (1982). Steam Injection Pressing. In Proceedings of the 16th
Volume 1, COST Action E31. International Particleboard Symposium, W.S.U.
Gretten, K. (2008): Sensor technology for wood based panels.
Presentation during the 2nd International Wood Academy, University of
Hamburg, 18 -29.09.06
Hasener, J.; Barbu, M.C. (2009): Overview on NDT technologies for on-
line control in the wood-based panel industry and an outlook for future
trends. Proceeding of COST E49 Workshop on “Processes and
Performance of Wood-Based Panels”, 28-29 April, Istanbul, pag. 2-14
Humphrey, P.E. & Bolton, A.J. (1989). The Hot Pressing of Dry-formed
Wood-based Composites. Part II. A Simulation Model for Heat and
Moisture Transfer, and Typical Results. Holzforschung 43(3):199-206
Johansson, J. Pizzi, A. and Van Leemput, M. (Editors) (2002). Wood
adhesion and glued products. Working group 2: glued wood products.
State-of-the-art report Volume 2, COST Action E31 (ISBN 92-894-4892-
X).
Jain, N.C., Gupta, R.C. & Jain, D.K. (1967). Particleboard from
Groundnut Shells. Proceedings 11th Silviculture Conference, May, 1967,
India.
Maloney, T.M. (1993). Modern Particleboard & Dry-Process Fibreboard
Manufacturing. Miller Freeman Publications, San Francisco. (TS 875.M3)
Meyer, B. (1981). Urea Formaldehyde Resins. Addison-Wesley
Publishing Co., Inc. Massachusetts.
Meinert, K. (2008): OSB production. Presentation during the 3rd
International Wood Academy, University of Hamburg, 25.02-07.03.08

90 91
Irle, Barbu Mosesson, J.G. (1980). The Processing and Use of Waste Straw as a
Constructional Material. Conservation and Recycling, Vol. 3, pp. 389- 412.
Mitlin, L. (1968). Particleboard Manufacture and Application. Novello & Moslemi, A.A. (1974). Particleboard. Vol. 1 - Materials. Particleboard. Vol. 2
Co. Ltd., Kent.
– Technology. Southern Illinois University Press. (TS 875.M6) Wood-Based Panel Technology
Natus, G. (2008): Comparison of pressing systems. Presentation during
the 3rd International Wood Academy, University of Hamburg, 25.02- Schniewind, A.P., Cahn, R.W., Bever, M.B. (1989). Wood and wood-
07.03.08 based materials - Concise Encyclopedia. Pergamon Press Plc, Headington
Hill Hall, Oxford OX3 0BW, England.
Pizzi, A. (1983). Wood Adhesives Chemistry and Technology. Marcel
Dekker, Inc., New York. Shumate, R.D., Schenkmann, A.H. & Sloop, J.E. (1976). Experiences
with Air Suspension Classifiers in Particleboard Manufacture.
Ressel, J. (2008): Raw materials. Presentation during the 3rd International Proceedings of the 10th International Particleboard Symposium, W.S.U.
Wood Academy, University of Hamburg, 25.02-07.03.08
Steffen, A. (2008): Particleboard mat forming. Presentation during the 3rd
Ressel, J. (2008): Particle generation and screening. Presentation during International Wood Academy, University of Hamburg, 25.02-07.03.08
the 3rd International Wood Academy, University of Hamburg, 25.02-
07.03.08 Steinwender, M.; Barbu, M.C. (2009): Environment Impact of the Wood
based Panels Industry. Proceeding of ICWSE, 4-5 June, Brasov, pag. 767-
Ressel, J. (2008): Particle drying. Presentation during the 3rd 775
International Wood Academy, University of Hamburg, 25.02-07.03.08
Stipek, J.W. (Translated by R.J. Vance) (1982). The Present and Future
Ressel, J. (2008): Adhesive application. Presentation during the 3rd Technology in Wood Particle Drying. Proceedings of the 16th
International Wood Academy, University of Hamburg, 25.02-07.03.08 International Particleboard Symposium, W.S.U.
Ressel, J. (2008): Mat formation. Presentation during the 3rd International Subramanian, R.V., Somasekharan, K.N., Johns, W.E. (1982). Acidity of
Wood Academy, University of Hamburg, 25.02-07.03.08 Wood. Holzforschung 37(3):117-120
Rexen, F. (1975). Straw as a Raw Material for Particleboard and Paper Suchsland, O. & Woodson, G.E. (1986). Fibreboard Manufacturing
Production. Report on the First World Straw Conference. ed T. R. Miles, Practices in the United States. USDA, Forest Service, Agriculture
Eugene, Oregon. Handbook No. 640.
Roll, H., Barbu, M., Beck, P., Hoepner, D., Kaiser, U. & Lerach, K. Suchsland, O. & Woodson, G.E. (1986). Fiberboard Manufacturing
(2001). Continuous Hot Press with Cooling Section for MDF. In Practices in the United States. USDA, Forest Service Agriculture
Proceedings of the 5th European Panel Products Symposium, October, Handbook No. 640.
2001.
Thoemen, H. (2008): Pressing of panels. Presentation during the 3rd
Rowell, R.M. & Norimoto, M. (1988). Dimensional Stability of Bamboo International Wood Academy, University of Hamburg, 25.02-07.03.08
Particleboards made from Acetylated Particles. Mokuzai Gakkaishi
34(7):627-629. Thoemen, H. & Humphrey, P.E. (1999). The Continuous Pressing Process
for Wood-Based Panels: An Analytical Simulation Model. In Proceedings
of the Third European Panel Products Symposium, The BioComposites
Centre, University of Wales, Bangor.
Thoemen, H., Meyer, N., Barbu, M. (2006): Pressing from a fundamental
point of view: actions to increase press performance. Proceeding of 4th
“Press Users Club” Seminar, Shanghai, China, 23-26 October.
Vasisth, R.C. & Chandramouli, P. (1975). New Panel Boards from Rice
Husks. FAO Background Paper, FO/WCWBP/75.
Wadsworth, J. (2001). Focus on Particleboard. Part 1: Europe and North
America Revealed. Wood Based Panels International 21(5):9-25.

92 93
Irle, Barbu Walker, J.C.F. (2006). Primary wood processing: Principles and Practice 2nd
Edition. Springer, Amsterdam.
Walker, J.C.F. (1993). Primary wood processing: Principles and Practice. Williams, W. (1995). The panel pioneers: an historical perspective. Wood
Chapman & Hall. London.
Based Panels International 17(September):6. Wood-Water-Interaction
Youngquist, J.A. (1999). Wood-based Composites and Panel Products. In
Wood handbook - Wood as an engineering material. Gen. Tech. Rep. FPL– Chapter 2
GTR–113. Madison, WI: U.S. Department of Agriculture, Forest Service,
Forest Products Laboratory. 463 p.
Water Absorption of Wood and Wood-Based
Panels – Significant Influencing Factors
Peter Niemz

CHAPTER SUMMARY

Wood and wood-based materials absorb water from the air by sorption
and from a fluid by capillary forces. The sorption behaviour of solid wood
depends, amongst other factors, on the amount of extractives. The
velocity of water absorption strongly depends on the cutting direction and
on the anatomical structure of wood. It is clearly higher in the fibre
direction than perpendicular to it. The formation of tyloses by hardwood
species and the closed pits of softwoods considerably reduce the capillary
water absorption. Wood based materials tend to absorb more water
through their edges than through their faces.
The equilibrium moisture content (EMC) for wood based materials such
as fibre boards, MDF or OSB is lower than for solid wood. This seems to
be caused by the processing method. For wood based materials glued with
phenol resin, however, the strong hygroscopic behaviour of alkali leads to
higher EMCs than when other adhesives are used. The EMC has a strong
influence on all physical and mechanical properties of wooden materials
(MOE, MOR, IB, hardness, thermal conductivity etc.).
The swelling of particleboard and OSB in the direction of the plane is
higher than that of solid wood in the fibre direction. The thickness
swelling of particle board, OSB and MDF is much higher that of solid
wood perpendicular to the grain. This correlates with the densification of
the particles.

2.1 INTRODUCTION

Wood is a macromolecular material. It has, on average, the following

composition: 50 % carbon, 43 % oxygen, 6 % hydrogen, 1 % nitrogen.
The main components of wood are:
 Cellulose (particularly in fibrils): 40-60 %
 Hemicelluloses: 6-27 %

94 95
Niemz
Wood-Water-Interaction

 Lignin (matrix substance): 18-41 % ( 41 % for compression wood; 25-

a) b)
32 % for softwoods; 18-25 % for hardwoods)
 Secondary wood constituents (for example extractives): 0.3-20 %

c) d)

Figure 2.2: SEM images: a) Softwood (Picea abies), image: ETH Zurich,
Wood Physics, b) Hardwood (Nothofagus alpina), image:
ETH Zurich, Wood Physics, c) MDF, image: TU Dresden;
Bäucker
d) Particle board, image: ETH Zurich, Wood Physics.

Water can adhere in macroscopic as well as in microscopic (see Figure

2.2 a) and b) and sub-microscopic cavities of the cell wall. In a wooden
material, water can also be in the cavities between the particles (see
Figure 2.1: Wood structure (Schmitt, University of Hamburg). Figure 2.2c and d). Extractives are deposited in the cell wall and in the
lumens (as in the case of oak where the heart wood is clearly darker than
The cellulose in wood acts as a reinforcement and lignin behaves as a the sapwood). As a protective mechanism, hardwoods form tyloses
matrix, similar to fibre-reinforced synthetics. Molecules of cellulose form whereas softwoods close their pits. Thereby, living trees are protected
micelles and these in turn form fibrils that are crystalline and amorphous against fungi, preventing the transfer of water. The closing of pits, the
(Figure 2.1). The middle lamellae of the cell wall consist mainly of formation of tyloses and the extractives content vary greatly between
lignin, the S2 of cellulose. The alignment of the fibrils in the S 2 various kinds of trees, especially amongst many tropical trees, but also
determines the mechanical properties and the swelling and shrinkage of amongst temperate trees such as acacia (Acacia ssp.) or oak (Quercus
wood in the fibre direction. Robur L.). These are distinct tylosis formers and produce a high amount
of extractives. Spruce species, which are softwoods, are characterised by
a high amount of closed pits.

96 97
Niemz
 Wood-Water-Interaction
Wood-based panels (plywood, particle boards, fibre boards, etc.) are
composed of chopped or cut elements of wood (planks, veneer, chips, The equilibrium moisture content (EMC) is calculated as follows:
fibres) that are glued together. Synthetic bonding agents are frequently
used (urea resin, phenolic resin, isocyanates). The amount of adhesive
present varies with the panel type but does not often exceed 10 %, for (1)
floors up to 16 %. Wood based materials in static use often have a clearly
higher density than solid wood (40-80 % higher). The kind of adhesive For oven drying we used a temperature of 103±2oC.
and the structure of the material influence the properties of the wood
based material as well as the technology of the production (defibration, 2.2.2 Sorption Behaviour
drying, gluing).
2.2.2.1 Solid wood

Solid wood (the basic raw material for particleboards, OSB and MDF)
2.2 SORPTION BEHAVIOUR AND CAPILLARY WATER has a large specific inner surface. For spruce it is about 220 m²/g
ABSORPTION (calculated according to the theory of Hailwood-Horrobin). Wood
behaves hygroscopically; it takes up water from the air and gives it off by
2.2.1 Ultimate State of Wood-Water desorption. In Figure 2.3 (for solid wood) and 2.4 (for wooden materials)
the different sorption stages can be seen. This works up to a relative air
Wood is a capillary porous system. Water can be adsorbed in both, macro humidity of circa 100 % where the fibre saturation point is reached. The
and micro pores (pores in the cell wall). Pores that are conditional on the EMC of wood is correlated to a certain temperature and relative air
anatomical structure have a diameter of 10-1 to 10-5 cm, whereas those humidity and depends on the wood species and the type of wooden
conditional on the molecular structure are 10-5 to 10-7 cm in diameter. material. In general, adsorption (increasing humidity) and desorption
(decreasing humidity) can be distinguished. A hysteresis effect occurs
The wood-water system can be divided into three different states:
between both; for this reason, the wood moisture content will be 1-2 %
 Absolutely dry = oven-dry (water is inexistent, wood moisture = higher for desorption than for adsorption. There are three stages of water
0 %). absorption:

 Fibre saturation (the complete micro pore system, i.e.  Chemisorption (build up of a monomolecular layer of water)
intermicellar and interfibrillar cavities of the cell wall, is filled with
 Physisorption (build up of a polymolecular layer of water)
water; the wood moisture content is approximately 28 % which can
vary between wood species).  Capillary condensation (condensation of water on the capillary
caused by a saturation pressure that is smaller for capillaries than
 Water saturation (both the micro and macro system are filled with
for plain surfaces (for example for a capillary radius of r = 1,06.10-
water; wood moisture depends on the wood density and lies
4 cm, the relative vapour pressure is 99,9 %; for r = 0,86.10-7 cm,
between 770 % (for Balsa, Ochroma lagopus SW) and 31 % (for
the relative vapour pressure is 30 %)
Lignum vitae, Guaiacum officinale L. Data according to
Trendelenburg and Mayer-Wegelin 1955). Below the FSP, the equalisation of moisture takes long periods of time as
the moisture transport happens by diffusion (see moisture distribution in a
Water adsorbed from absolute dry to the fibre saturation point (FSP) is
board as a function of time; Figure 2.8).
called bound water because it is hydrogen-bonded to the wood cell wall
polymers; water absorbed above the FSP is called free water. Using a thermal or hygrothermal pre-treatment (such as high temperature
drying) can reduce the EMC (Figure 2.3). Treatment with heat and
pressure leads to a reduced amount of hemicelluloses and therefore to
lower wood moisture percentage and a higher dimension stability.

98 99
Niemz
Wood-Water-Interaction

be explained with the presence of adhesives, which have different

sorption behaviours compared to wood. For example, alkali-containing
phenolic resin is highly hygroscopic (Figure 2.4). The wooden materials
themselves also differ between each other. Furthermore, there is a strong
influence of the density (Figure 2.5b).

Figure 2.3: Sorption of Pinus radiata (influence of thermal treatment

(unb = untreated).

The EMC and swelling of wood are reduced to 50 % by using a thermal

treatment at temperatures between 180 and 240 °C (in the case of solid
wood). Densified wood has a slightly lower EMC in the range of capillary
condensation and FSP than un-densified wood, while in the ranges of
chemisorption and physisorption the EMC is slightly higher (Popper et al.
2002).
Simultaneous thermal or hygrothermal treatment and densification leads
to a significant reduction of EMC compared to normal wood.
Densification nevertheless is reversible by exposure to water (memory
Figure 2.4: Comparison of the sorption behaviour of softwood, particle
effect). boards with phenolic resin (PF) and particle boards with urea
The relative humidity of air in buildings can be considerably influenced resin (HF) (Seifert, cited in Kehr 1974).
by the hygroscopic behaviour of wooden objects within them. Living
The EMC of MDF is slightly lower than that of particleboards (Figure
spaces with a high amount of wood have smaller variations in relative air
2.5a). This is caused by the hygrothermal processes used to produce the
humidity than those with non-hygroscopic materials present. Wood adds a
fibres. The EMC of MDF is correlated to the vapour pressure used during
measurable contribution towards greater comfort in living spaces.
the defibration process, the temperature and the defibration duration
(Krug and Kehr 2001). With high pressures and long defibration
2.2.2.2 Wood-based panels durations, the thickness swelling of MDF decreases.
Wood-based panels (WBPs) are strongly influenced by adhesives, The sorption behaviour of WBP can be described for instance by the
density and the production technology (drying, defibration, pressing at Hailwood-Horrobin (HH) or the Brunauer-Emmet-Teller (BET)
high temperatures). The EMCs of WBPs are lower than those of solid method (see DIN ISO 9277:2003-05).
wood of the same species under the same conditions. There is a strong
influence of particle drying (thermal modification), defibration
(producing fibres) and hotpressing (thermal or hydrothermal treatment,
densification). Partially, the moisture-related behaviour of WBPs can also

100 101
Niemz
 Wood-Water-Interaction

a lower uptake of immersion liquids. The velocity of water absorption is

significantly influenced by:
 the density of the material (the higher the density, the lower the
absorption velocity)
 the anatomical direction of wood (clearly higher velocity in the
fibre direction than perpendicular to it)
 the wood species
 the content of paraffin

a)  the surface coating

 the sample size (dimensions)
Table 2.1 shows the coefficients of water absorption (absorption of
liquid water by immersion (DIN EN ISO 15148) in kg/(m2.s) for spruce
and beech.
Table 2.1: Coefficients of absorption of liquid water by immersion (DIN
EN ISO 15148).
Spruce Beech
kg/(m2s) kg/(m2s)
Longitudinal 0.017 0.044
Radial 0.003 0.005
Tangential 0.004 0.004

For particle boards (density 670 kg/m3) the coefficient in direction of the
plain is 0.025 and perpendicular to the plain 0.0014.
b) The water absorption per unit of time is clearly higher in the direction of
Figure 2.5: Sorption of different wood based panels: a) Sorption of the fibre than in the radial or tangential directions, in direction of the
different wood based panels (ETH Zurich) , b) Influence of plane higher then perpendicular to it. This relationship is also valid for
the density on the moisture content of fibreboards (Kiessl and water absorption from air humidity. This is the reason why large cross-
Möller 1989). sections (as used for building materials, e.g. Glulam) require long storage
periods to reach the EMC over the complete cross section.
2.2.3 Water Absorption by Capillary Forces
2.2.4 Swelling and Shrinkage
Mainly beyond the FSP, wood absorbs water (liquid water) by capillary
forces. Moisture transport happens according to the laws of capillary 2.2.4.1 Solid wood
physics (from wide to small capillaries). Closed pits (softwoods) or
Changes of dimension occur with changes of moisture within the
tyloses (hardwoods) strongly reduce the water uptake which also leads to
hygroscopic area (up to the FSP, Figure 2.6, left side). Swelling occurs

102 103
Niemz
Wood-Water-Interaction

when water is taken up; shrinkage occurs when water is released. Table 2.3: Swelling of wooden materials (%/% EMC)
Swelling and shrinkage are different for the three principle directions of
 Swelling/Shrinkage (%/%)
wood. It is small in longitudinal direction, whereas wood swells 10–20 Material In the direction of the Perpendicular to the plane
times more in the radial direction (direction of the rays) and even 15–30 plane (thickness)
times more in the tangential direction (see Table 2.2).
Plywood 0.02 0.30
The maximum swelling increases with wood density, although the time to Phenol resin 0.025 0.45
Particle-
achieve this increases too. board Other 0.015 0.70
Table 2.2: Swelling and shrinkage of wood according to DIN 52184. Glulam 0.01 0.24
Max. swelling (%) Rel. swelling (%/%) MDF 0.15..0.20 0.80
Species
Longitudinal Radial Tangential Radial Tangential
When zones of different moisture contents have developed in a board, a
Spruce 0.2 – 0.4 3.7 8.5 0.19 0.36 warping can be observed (Figure 2.7). This effect also applies to
Pine 0.2 – 0.4 4.2 8.3 0.19 0.36 particleboards, MDF and OSB. In the case of particleboards and
Beech 0.2 – 0.6 6.2 13.4 0.20 0.41 fibreboards, the density profiles (homogeneity, different densities at the
Oak 0.3 – 0.6 4.6 10.9 0.18 0.34 upper and lower side of the board) and board thicknesses play an
important role. The higher the thickness, the lower the warping. Figure
2.7 shows the warping of a solid wood panel with different middle layers
2.2.4.2 Wood based materials made of wooden materials. There is a strong influence of the middle
layer’s MOE: the lower the MOE of the middle layer, the higher is the
In wood based materials, swelling in the direction of the plain is slightly warping of the whole board. In elements with big dimensions, an
higher than parallel to the grain in solid wood. It is also clearly higher equilibrium moisture content for the whole element will never be reached.
perpendicular to the plain than for solid wood perpendicular to the grain Instead, a moisture profile is formed (Figure 2.8).
(see Table 2.3). This is due to the reversal of densification of the particles
that were compressed during the production of the board (Figure 2.6b). Apart from stresses inherent to the material, a considerable swelling
Swelling is largely reversible for solid wood but this is not the case for pressure builds up when the component is firmly mounted. In fact, the
particleboards, fibreboards and compressed wood where some of the pressure exerted by swollen wood was applied in ancient times to bust
swelling is irreversible. Swelling increases with the density. The rel. rocks. The measured swelling pressure in the case of hindered swelling is
swelling is calculated in % swelling/% change of the EMC. The swelling significantly smaller than the theoretically calculated one. This is because
rate of particleboard and MDF is not linearly correlated with the RH a substantial amount of the stresses is reduced due to (plastic)
(from oven dry to fibre saturation; Sonderegger und Niemz (2006), see deformations and relaxation. The swelling pressure is higher in humid air
also annex Table 2.12). than it is for immersions in water. The density of wood and the swelling
pressure have proven to be correlated, with the one increasing with the
Figure 2.6c shows the thickness swelling as a function of moisture other. A higher swelling pressure was observed parallel to the grain than
content. If the boards expand only by the volume of water absorbed, as perpendicular to it. A swelling pressure of 0.2-0.4 N/mm 2 was measured
solid wood does, thickness swelling would theoretically follow the dashed for MDF (density: 600 kg/m3). The swelling pressure in water is lower
line. The particleboard swells considerably more which means that voids than in water vapour; it increases with the density.
are created within board. The MDF does not. It behaves essentially like a
solid wood (Suchland 2004). Internal stresses also occur in glued elements with different properties
(different element thicknesses, different fibre and grain angles,
For large cross-sections, the EMC can only be reached in the outer differences in the EMC). These stresses can cause delamination, cracks
regions. As a result, a moisture profile will build up, causing internal in the elements or destroy the structure of wood based panels after a
stresses and possibly cracks. natural weathering (see Figure 2.10).

104 105
Niemz
Wood-Water-Interaction

a) b)

a)

b)

c)
c)
Figure 2.6: Swelling of wood and wooden materials: a) solid wood Figure 2.7: Warping of a three layer solid wood panel under different
(Mörath 1932) ; l - longitudinal, r - radial, t - tangential, b) climatic conditions (upper side 1: 20oC/65%RH, lower side 2:
wood based materials (Teichgräber, cited in Kehr 1974); 20oC/100%RH) and middle layer from different wooden
comparison of longitudinal swelling in relation to solid wood materials: (a) solid wood, (b) MDF density 500kg/m 3 and (c)
(1) and swelling of thickness (2) of particleboards with solid MDF density 700kg/m3 (ETH Zurich, IfB, Wood Physics
wood (pine), c) thickness swelling from particleboard and 2009).
MDF during cyclic exposure (Suchsland 2004).

106 107
Niemz Wood-Water-Interaction

Figure 2.8: Moisture profiles in a wood panel after changing the climatic
conditions from 20oC/65% RH to 2oC/90%RH, (Dunky and OSB / July 2004 MDF / July 2004
Niemz 2002).

2.2.5 Influence of Wood Moisture on the Surface Roughness

of Wood-Based Materials

During manufacturing of wooden particle boards, the particles are

strongly compressed. When the panels get moist (e.g. while aqueous
adhesives are applied), thickness swelling of the densified particles occurs
(spring-back effect), accompanied by an increase in surface roughness.
The thicker the particles, the higher is the surface roughness. The surface
roughness for particleboards is 60-80µm (veneer) and 40-60 µm (finish
lamination). For high-quality surfaces (e.g. those to be coated with OSB / March 2006 MDF / March 2006
overlay papers), very small particles have to be used in the top layers.
Thick particles become apparent after coating. Moreover, density
variations of the panels (in the direction parallel to the panel surface) can
cause a corrugated surface. Examples are shown in Figures 2.9 and 2.10.
Weathering is strongly influenced by adhesive type and particle
dimensions as well. We have a strong swelling from the particles in the
surface, the roughness increased after weathering, for OSB the roughness
is higher than for MDF (Figure 2.9).

OSB / June 2007 MDF / June 2007

Figure 2.9: OSB and MDF after a natural weathering, (ETH Zurich, IfB,
Wood Physics).

108 109
Niemz Wood-Water-Interaction

100

y = -4,1322x + 113,68
R2 = 0,79 LDF
80
MDF

Bending strength (MPa)

HDF
60

y = -1,8283x + 53,862
R2 = 0,98
40

20
y = -0,9419x + 27,008
R2 = 0,72
0
0 5 10 15 20 25
Moisture content (%)

Figure 2.11: Relationship between EMC and bending strength for MDF
with different densities (ETH Wood Physics, Bektha and
Niemz 2009).
a) b)

Table 2.4: Influence of the EMC on the Brinell hardness of wooden

materials for different EMCs (ETH Wood Physics, Niemz
Figure 2.10: Surface roughness of particle board in dry (a) and wet (b) 2009) (Load: 500N, v – coefficient of variation).
condition (Dunky and Niemz 2002).
Particleboard Particleboard
19 mm Solid wood panel
16 mm
(HF – resin) (spruce) 25 mm
(PF – resin)
EMC Brinell Brinell Brinell
(%) EMC EMC
Hardness Hardness Hardness
(%) (%)
(N/mm²) (N/mm²) (N/mm²)
0
2.3 THE RELATIONSHIP BETWEEN MOISTURE CONTENT 0
45.6 85.1 19.4
0
AND PROPERTIES v=18.0 % v=17.4 % 8.8 v=31.6 %
33.5 14.6
7.7
v=14.6 % v=23.1 %
The moisture content has a strong influence on most properties of wood 10.5
44.0 38.5 15.6
(up to the FSP) and WBPs. Strength (Figure 2.11), MOE and hardness 8.2
v=18.9 %
9.4
v=24.4 % v=24.4 %
(Table 2.4) decrease with increasing EMC; the thermal conductivity 24.6 26.7 12.5
increases (Table 2.5) as well as the diffusion coefficient (Figure 2.12 and 12.7 16.3 16.9
v=12.2 % v=14.2% v=16.8 %
2.13). Both are also strongly influenced by density. 18.0 19.7 10.4
17.45 27.5 24.12
v=17.2 % v=15.8 % v=14.4 %

110 111
Niemz
Wood-Water-Interaction

3.00E-10
Table 2.5: Thermal conductivity (Z) depending on the moisture content.
y = -6.27E-13x + 5.75E-10
Plywood Mean values of 5 specimens per type (at dry condition only 3
OSB
R2 = 0.986
Particle boards
specimens per type); Z10.dry.reg(+) = Z at dry condition and
2.50E-10
Fibre boards 10°C; OZ+ = change of Z per percent moisture content in
Coated boards W/(m•K) and in percent (Sonderegger and Niemz 2009).
2.00E-10 Linear regression (Particle boards)
Material Thickness Density L 10.dry.reg(m) OL OL m
Dwet-cup [m

Linear regression (Fibre boards)

3
m
Linear regression (Coated boards) (mm) (kg/m ) (W/(m·K)) (W/(m·K·%)) (%/%)
1.50E-10 Plywood (beech), 25 25.7 679 0.1304 0.00255 1.96
y = -2.76E-13x + 2.70E-10 mm
2 R2 = 0.621
1.00E-10 OSB 3, 18mm 18.4 562 0.0959 0.00074 0.77
Particle board (V20), 16.4 597 0.0965 0.00128 1.32
5.00E-11 16 mm
y = -4.23E-14x + 4.41E-11 MDF (V20), 3 mm 2.9 802 0.1104 0.00115 1.04
0.00E+00 R2 = 0.717
MDF (V20), 16 mm 16.2 696 0.0974 0.00121 1.24
450 550 650 750 850 950 1050
Substrate of laminate 6.6 785 0.1138 0.00151 1.32
Density [kg/m3] flooring (HDF), 7 mm

Figure 2.12: Diffusion coefficients (D) determined with wet cup tests In summary, wood moisture influences all properties of wood and wood
depending on the density (mean values) (Sonderegger and based materials. The EMC of WBPs is slightly smaller than that of solid
Niemz 2009). wood. The swelling in the direction of the panels is slightly higher than in
the longitudinal direction of solid wood, whereas the thickness swelling is
6.00E-10
Plywood clearly higher than that of wood perpendicular to the grain due to the
y = -1.34E-12x + 1.27E-09
R2 = 0.87
OSB spring-back effect of the densified particles.
5.00E-10 Particle boards
Fibre boards In the annex, various important properties of wood and wooden materials
Coated boards
4.00E-10 Linear regression (Particle boards)
related to the moisture are listed.
Ddry-cup [m

Linear regression (Fibre boards)

Linear regression (Coated boards)
3.00E-10
y = -7.13E-13x + 6.20E-10
2 2.00E-10
R2 = 0.69

1.00E-10
y = -2.00E-13x + 1.99E-10
R2 = 0.68
0.00E+00
450 550 650 750 850 950 1050
Density [kg/m3]

Figure 2.13: Diffusion coefficients (D) determined with dry cup tests
depending on the density (mean values) (Sonderegger and
Niemz 2009).
112 113
Niemz
Wood-Water-Interaction

2.4 REFERENCES
Niemz, P.: Physik des Holzes und der Holzwerkstoffe. DRW, Leinfelden-
Echterdingen: 1993
Diverse authors: Holzlexikon (4th edition). DRW, Leinfelden-
Echterdingen: 2003 Niemz, P.; Sonderegger, W.: Untersuchungen zum Sorptionsverhalten
von Holzwerkstoffen. Bauphysik, Berlin 31(2009): 244-249
Bektha, P.; Niemz, P.: Effect of relative humidity on some physical and
mechanical properties of different types of fibreboard. Euro. J. Wood Popper, R.; Niemz, P.; Torres, M.: Einfluss des Extraktstoffanteils
Prod. (2009 on line) ausgewählter fremdländischer Holzarten auf deren
Gleichgewichtsfeuchte. Holz als Roh- und Werkstoff 64 (2006), 491-496
Bodig, J; Jayne, B.A.: Mechanics of Wood and Wood Composites.
Krieger Publishing Company, Malibar, Florida: 1993 Popper, R.; Niemz, P.; Eberle, G.: Sorptions- und Quellungseigenschaften
von verdichtetem Holz. Holzforschung und Holzverwertung, Wien (2002)
Burmester, A.: Formbeständigkeit von Holz gegenüber Feuchtigkeit. 6, S.114-116
Grundlagen und Vergütungsverfahren. Berlin, BAM report No. 4: 1970
Sonderegger, W.; Niemz, P.: Untersuchungen zur Quellung und
Dunky, M.; Niemz, P: Holzwerkstoffe und Leime. Berlin, Springer Wärmedehnung
Verlag 2002
von Faser-, Span- und Sperrholzplatten Holz als Roh- und Werkstoff
Hailwood ,A.J. HorrobinS.: Absorprtion of water by polymers. Analyssis 64(2006) 64: 11–20
in term of single model. Trans. Faraday Soc. 42B (1946)p.84-102
Sonderegger, W.; Niemz, P.: Thermal conductivity and water vapour
Kehr, E. (Chapter 2) in „Werkstoffe aus Holz“; Fachbuchverlag, Leipzig transmission properties of wood based materials.
1974
Euro. J. Wood Prod. 67(2009):313-321Suchsland, O. The swelling and
Kiessl, K.; Möller, U.: Zur Berechnung des Feuchteverhaltens von shrinking of Wood. Forest Prod. Society, Madison: 2004
Bauteilen aus Holz und Holzwerkstoffen. Selektion feuchtetechnischer
Parameter. Holz als Roh- und Werkstoff, Berlin 47 (1989) p. 317-322 Trendelenburg, R.; Mayer-Wegelin, H.: Das Holz als Rohstoff (2nd
edition). Carl Hanser, München: 1955
Krug, D.; Kehr, E.: Einfluss des Aufschlussdruckes bei der
Faserstoffherstellung auf die Quellungsvergütung. Holz als Roh und Walker, J.C.F.: Primary Wood Processing. Principles and Practice. 2nd
Werkstoff, Berlin 59(2001) 342-343 edition, Springer 2006
Künzel, H.: Risse in bewittertem Holz. Einfluss auf die Willeitner, H.; Schwab, E.: Holz - Aussenverwendung im Hochbau.
Feuchteverhältnisse, Abhilfe. Bauen mit Holz, Karlsruhe (1993); 12, p. Verlagsanstalt Alexander Koch, Stuttgart: 1981
1018-1022
Kollmann, F.: Technologie des Holzes und der Holzwerkstoffe (volume
1, 2nd edition). Springer, Berlin: 1951
Kollmann, F.; Coté, W.: Principles of Wood Science and Technology
(volume 1). Springer, Berlin, Heidelberg: 1968
Maloney, T.: Modern Particleboard and Dry Process Fibreboard
Manufacturing. Miller Freemann, San Francisco: 1993
Mörath, E.: Studien über die hygroskopischen Eigenschaften und die
Härte der Hölzer. Darmstadt, 1932

114 115
 Niemz
Wood-Water-Interaction

2.5 ANNEX: TABLES LISTING SELECTED PROPERTIES OF

WOOD AND WOODEN MATERIALS
Table 2.9: EMC of wood-based materials (adsorption) at different (Niemz
and Sonderegger 2009).
Table 2.6: Sorption of different wood species (ETH, Wood Physics 2010).
(EMC = equilibrium moisture content, RH = relative humidity, density EMC (%) at RH %
x = mean, s = standard deviation). Material and thickness
kg/m3 35% 50% 65% 80% 93%
EMC (%) Plywood, beech, 25 mm 742 6.70 7.74 9.75 13.58 20.62
Species
35 % RH 50 % RH 65 % RH 80 % RH 90 % RH Plywood, beech, 35 mm 777 7.30 9.09 11.00 13.56 21.21
Plywood, beech, 50 mm 752 7.10 9.01 11.31 14.27 20.17
x
Fir 7.16 8.84 10.18 13.42 17.63
OSB, 12 mm 663 6.46 7.99 9.47 13.15 19.36
s 0.83 0.90 1.00 0.86 0.92
x OSB, 15 mm 634 6.56 8.09 9.53 13.62 19.81
Spruce 7.69 9.65 11.31 15.18 19.07
s 0.23 0.15 0.26 0.42 0.44 OSB, 18 mm 619 6.66 8.34 10.10 12.91 19.16
x 7.57 8.92 10.03 13.26 17.29
s Pine 0.62 OSB, 22 mm 627 6.55 8.05 9.49 12.86 19.28
0.64 0.64 0.36 0.29
OSB, 25 mm 622 6.53 8.03 9.63 12.90 19.04
s Beech x 7.17 8.92 10.33 14.61 18.78
0.08 0.07 0.07 0.08 0.22 Particleboard, 10 mm 724 6.90 8.14 9.86 12.68 19.90
x
Chestnut 8.03 9.66 11.12 14.50 17.45 Particleboard, 16 mm 657 7.00 8.43 10.40 13.28 20.47
s 0.29 0.31 0.30 0.33 0.45 Particleboard, 16 mm, coated 683 6.53 7.72 9.27 12.23 20.45
Particleboard, 19 mm 630 7.27 8.53 10.40 13.55 21.16
Table 2.7: EMC of wood-based materials (after production) (Niemz Particleboard, 25 mm 615 7.59 8.11 9.75 12.93 20.72
1993). particleboard, 25 mm, coated 636 6.53 7.85 9.49 12.26 20.37
Moisture Particleboard, 40 mm 633 6.79 8.14 9.79 13.22 21.44
Material
content %
Particleboard, 40 mm, coated 635 8.00 8.14 9.64 12.98 20.74
Plywood 5-15 Particleboard, 6 mm 769 7.38 8.64 10.42 13.76 22.34
Particleboard 94 Particleboard, 7 mm 790 7.32 8.56 10.34 12.99 20.34
Fibre hardboard 53 Particleboard, 7 mm, coated 826 6.97 8.27 9.91 12.36 18.59
MDF 94 MDF, 3 mm 825 6.37 7.74 9.31 11.40 19.18
Spruce (20oC/65% RH) 12 MDF, 6 mm 843 5.98 7.08 8.07 10.77 18.93
MDF, 10 mm 803 5.98 7.15 8.22 10.89 18.84
Glulam 102
MDF, 15 mm 532 6.11 7.67 9.40 12.50 18.18
MDF, 16 mm 728 5.50 6.83 8.47 11.92 19.13
Table 2.8: EMC of fibre board (wet process) at different RH (Niemz and MDF, 16 mm, coated 774 6.25 7.22 8.41 10.19 18.14
Sonderegger 2009).
MDF, 19 mm 809 6.26 7.37 8.23 10.10 17.98
density EMC (%) at RH
Material MDF, 25 mm 742 6.16 7.34 8.28 10.51 18.59
kg/m3 35% 50% 65% 80% 90% MDF, 25 mm, coated 777 6.28 7.32 8.66 11.04 18.44
Fibreboard (wet process) 980 5.66 6.83 9.10 12.72 15.69 MDF, 40 mm 763 5.70 6.73 7.60 9.97 16.98
Fibreboard (wet process) HDF, 7 mm 871 5.29 6.38 8.09 11.28 17.56
for insulation (without 150-200 6.55 8.34 11.1 15.37 19.72 HDF, 7 mm, coated 920 6.34 7.27 8.35 10.25 17.04
adhesive)
HDF, 8 mm, coated 934 6.23 7.25 8.35 10.43 17.23
Fibreboard minerally
320 1.11 1.30 1.59 2.28 3.17
bounded

116 117
Niemz Wood-Water-Interaction

Table 2.10: Coefficient of capillary water uptake (AW-coefficient of

Table 2.12: Differential swelling for wood-based materials (Sonderegger
capillary water uptake DIN EN ISO 15148) (ETH Zurich,
and Niemz 2006) (p. - perpendicular, l. – longitudinal).
Wood Physics 2009).
Swelling in (%/%)
Parallel to the board Perpendicular to the Perpend.
Thickness surface board surface Material/ Density in the direction of the plane to the
Material
Density AW in Density in (AW) Thickness (kg/m³) plane
mm
(g/cm3) (kg/m2s0,5) (g/cm3) kg/m2s0,5 35 – 95% 35 – 80% 80 – 95% 65 – 95%
Solid wood 27 0.44 0.0115 0.42 0.0022 RH RH RH RH
panel
MDF 756 0.016 0.040 0.010 1.07
Plywood 15 0.51 0.0381 0.50 0.0026 18 mm v = 0.4 v = 3.6 v = 2.8 v = 7.4 v = 2.1
Particle- 630 0.016 0.028 0.011 0.98
19 0.67 0.0254 0.69 0.0014 l.
board(V20) OSB 3, v = 4.4 v = 17.3 v = 9.0 v = 29.5 v = 10.7
MDF 22 0.69 0.0556 0.69 0.0125 18 mm 649 0.018 0.034 0.011 0.91
p. v = 2.5 v = 9.3 v = 8.4 v = 18.8 v = 10.7
OSB 0.65 0.0234 0.68 0.0018 Particleboard 733 0.040 0.049 0.036 1.22
V20, 8 mm v = 1.1 v = 2.5 v = 3.6 v = 3.1 v = 3.9
Particleboard 649 0.027 0.036 0.023 0.93
Table 2.11: Diffusion properties of wooden materials (Niemz 1993). V20, 18 mm v = 1.9 v = 11.3 v = 8.0 v =18.8 v = 4.6
Particleboard 596 0.032 0.040 0.029 0.87
Density Water vapour V20, 30 mm v = 0.7 v = 4.0 v = 5.6 v = 5.7 v = 3.1
Material
(kg/m3 ) resistance factor μ
Particleboard 649 0.035 0.041 0.031 0.92
Radial 470 55 V313, 19mm v = 1.6 v = 5.4 v = 5.2 v = 8.1 v = 4.8
Spruce
Tangential - 100 665 0.010 0.017 0.006 0.22
l.
MDF 470 20 Plywood larch v = 1.6 v = 23.1 v = 17.2 v = 42.9 v = 10.2
12,5mm 685 0.021 0.038 0.013 0.22
900 50 p.
v = 3.6 v = 32.5 v = 31.7 v = 33.8 v = 9.3
780 0.012 0.024 0.006 0.42
Particleboard 470 20 l.
v = 1.4 v = 29.4 v = 7.1 v = 36.9 v = 7.9
900 360 3Plywood:
beech p. 783 0.023 0.037 0.016 0.40
Strandboard 470 65 20mm v = 1.0 v = 19.3 v = 12.7 v = 28.0 v = 5.4
p. 766 0.022 0.034 0.015 0.41
900 1400 . v = 1.2 v = 13.5 v = 7.7 v = 22.9 v = 3.4
Solid wood panel 450 40/400
Fibre insulation board 175 5
Hardboard (fibreboard) 1000 120

118 119
12
0 Table 2.13: Water vapour resistance factors (µ) and diffusion coefficients (D) derived from dry and wet cup tests Ni
(Sonderegger and Niemz 2009) (MC - moisture content, µ mf - Water vapour resistance factor of the em
melamine face) z
Thickness Density Dry cup Wet cup
Material Required Measured MC COV µmf.d Ddr MC COV µmf. Dwet
[kg/m³]
[mm] [mm] [%] [-] [%] ry[-] [m²/s]
y [%] [-] [%] [-]
wet [m²/s]
25 26.0 738 7.06 97.8 18.4 5.16 x10-11 15.95 44.1 23.2 3.92x10-11
Plywood µdry µwet
35 35.4 778 8.15 100.8 18.8 4.38x 10-11 14.95 66.0 19.9 2.62x10-11
(beech)
50 50.1 756 9.50 97.2 4.4 4.22 x10-11 15.01 48.8 15.2 3.82x10-11
12 11.9 659 7.11 100.5 4.2 5.59 x10-11 18.02 42.8 14.8 4.46x10-11
15 15.0 638 6.98 116.8 0.1 4.97x10 -11 18.28 47.3 20.3 4.24 x10-11
OSB 3 18 18.3 618 7.53 112.6 16.6 5.06 x10-11 18.15 47.6 5.7 4.54x10-11
22 22.0 644 7.18 98.8 17.7 5.97x10-11 16.70 75.3 7.2 2.54 x10-11
25 25.6 629 7.25 139.1 16.4 4.34x10-11 15.91 93.3 26.3 2.20 x10-11
6 5.9 776 7.59 65.1 7.8 7.39 x10-11 16.50 27.1 4.2 5.81x10-11
10 10.3 710 8.15 56.9 8.6 8.29x10-11 15.09 25.1 15.5 7.58 x10-11
16 16.4 654 7.45 29.7 3.2 1.89x10-10 16.21 16.8 9.9 1.12x-10
Particleboard
19 18.9 636 7.75 35.1 3.3 1.55x10-10 17.52 18.5 5.5 1.09 x10-10
(V20)
25 25.1 612 7.05 48.2 15.7 1.31x10 -10 15.52 26.4 13.6 7.44x10-11
40 38.3 626 7.03 27.8 2.1 2.25 x10-10 14.21 20.2 10.7 9.21x10-11
16 15.5 700 7.80 146.3 20.4 8932 3.88 x10-11 10.52 162.8 22.2 11270 1.05x10-11
Melamine faced 25 25.3 631 7.33 111.8 18.3 8117 5.50 x10-11 10.77 120.2 9.6 11925 1.55 x10-11
board 40 38.6 635 7.46 57.9 14.5 5854 1.05 x10-10 10.41 78.0 14.7 11162 2.37x10-11
(particle board)
Substrate of 7 6.6 799 7.58 94.8 3.1 4.90x10 -11 15.95 38.4 3.3 3.99 x10-11
laminate flooring
(particle board)
Laminate flooring 7 6.8 840 8.34 200.5 5.4 3689 2.11x10-11 13.24 149.8 18.7 3862 1.01x-11
(particle board)

Table 2.13: continued

Thickness Density Dry cup Wet cup W
Material Required Measured MC COV µmf.d Ddr MC COV µmf.w Dwet oo
[kg/m³]
[mm] [mm] [%] [-] [%] ry[-] [m²/s]
y [%] [-] [%] et[-] [m²/s]
d-
3 2.9 856 8.01 58.9 3.6 7.41x10-11 11.88 31.1 8.5 4.55x10-11
6 6.0 848
µdry
6.73 42.4 4.1 1.25x10-10
µwet
12.15 27.1 7.5 4.71x10-11
Wa
10 9.9 811 6.61 39.0 4.0 1.39 x10-10 14.70 24.8 1.0 5.51 x10-11 ter-
MDF (V20) 16 16.7 726 6.98 20.4 6.5 2.78 x10-10 14.11 13.4 5.8 1.17x10-10 Int
19 18.9 810 6.69 33.4 4.3 1.62x10-10 12.82 22.9 1.6 5.87x10-11 era
25 25.0 749 6.57 20.4 1.3 2.83x10-10 12.91 15.4 1.8 9.49x10-11 cti
on

40 40.2 766 6.04 16.6 7.1 3.69x10-10 11.29 13.9 6.2 9.64 x10-11
Melamine faced 16 16.0 779 6.36 132.8 30.5 8900 4.52 x10-11 9.03 145.9 5.0 10556 9.39x10-12
board (MDF) 25 24.4 783 6.81 120.4 5.3 12085 4.47 x10-11 8.73 155.4 16.0 17059 9.03x10-12
Substrate of
laminate flooring 7 6.4 876 6.49 54.5 13.6 9.36 x10-11 12.81 36.7 3.6 3.31x10-11
(HDF)
Laminate flooring 7 6.5 925 6.73 289.0 5.1 5060 1.62x10-11 9.50 251.3 4.4 4656 4.59x10-12
(HDF) 8 7.6 953 6.47 349.8 11.9 7525 1.33x10-11 10.95 207.6 18.1 4397 5.52x10-12
MDF wall panel 15 14.9 536 7.76 13.5 1.3 5.21x10-10 12.51 8.9 2.0 2.44 x10-10

12
1
 Transport Phenomena

Chapter 3

Transport Phenomena
Luisa Carvalho, Jorge Martins, Carlos Costa

CHAPTER SUMMARY

Transport phenomena, as heat and mass transfer, are involved in several

processes in the manufacture of wood-based products, such as drying of
particles/fibres, hot-pressing and conditioning. The drying of
fibres/particles and the hot-pressing operations involve simultaneous, heat
and mass transfer. This chapter starts to focus on the basic mechanisms of
transport phenomena that occur in porous media, as heat conduction,
convection and radiation and also mass diffusion and convection. Then,
the main operations involving these mechanisms in the production of
wood-based panels are described: the drying of particles/fibres and the hot-
pressing process. The drying regimes in convective drying, heat transfer
and moisture movement are important issues treated in this chapter. The
monitoring of the internal conditions in the mattress during the hot-
pressing and the prediction of transport properties are also treated. Heat
and mass transfer in conditioning (hot-stacking and normal stacking) is
also covered. This chapter finishes with the diffusion of chemicals, resins,
wax and other additives.

3.1 HEAT AND MASS TRANSFER MECHANISMS IN POROUS

MEDIA

3.1.1 Introduction

Wood is a natural material and can be characterised as a capillary porous,

cellular, hygroscopic, anisotropic and viscoelastic material. The
particle/fibre mat is a complex structure at multiple levels, including the
cellular structure of wood and the pseudo cellular structure of the mat
(Kamke, 2004). So, fibre and particle mats are at the micro-level,
heterogeneous and hygroscopic porous media. The description of
transport mechanisms in a rigid porous media poses many problems, but
the situation is even more difficult if the medium is compressible
(Aguilar, 2006).

122 123
Carvalho, Martins, Costa
 Transport Phenomena
3.1.2 The Heat Transfer Mechanisms

There are three basic heat transfer mechanisms: conduction, convection where g is the gas density (in kg/m3), Cp the gas specific heat (in J/(kgK))
and radiation (see Figure 3.1). Heat can also be generated inside the and vg is the velocity of the gas phase (in m/s). In certain conditions, as in
material if a source of energy is present. combustion (gas generation) and in drying (phase change of water) the
convection can be enhanced.
The convective heat flux at the boundary, from a surface to the ambient
T1>T2 can be given by:
Surface, T1
T1 Moving fluid
. Ts>Tœ (3)
q Tœ boundary c œ
. 1
q
T2 Surface, T2
Ts where h is the external heat transfer coefficient in W/(mK) that is
q function of the system geometry, fluid properties, flow velocity and
q2 temperature difference between the surface (Ts) and the ambient (Tœ).
T1

Figure 3.1: Conduction, convection and radiation heat transfer modes.

A third mechanism is thermal radiation, which is the transfer of heat by
Heat conduction is defined as energy transferred by temperature electromagnetic radiation. Externally generated radiation from a hot body
difference or gradient and can occur in solids, liquids or gases. Heat flows travels at the speed of light to the object and is absorbed by both liquid
from higher-temperature regions to those of lower temperature (Aguilar, and solid material. The wavelength of the radiant energy must match the
2006). In conduction, energy is transferred by atomic or molecular product absorptivity for the best energy coupling and drying rates. Some
excitation from a higher to a lower-energy zone (Smith, 1994). Most materials act as a window to radiant energy at specific electromagnetic
metals are good conductors because their atoms are more readily set in a wavelengths and in many cases, hotter radiant heaters do not produce
vibration mode by thermal excitation. Most inorganic compounds and better energy transfer rates (Aguilar, 2006). Heat transfer by radiation (q
organic molecules, such as polymers and wood are more difficult to set in in W) can be modelled by the Stefan-Boltzmann law for a grey body, at
a vibration mode and are poor conductors. The constitutive equation for an absolute temperature T1 and the surrounding enclosure at T2:
conductive heat flux is known as Fourier’s first law and can be written 4
1 2
4
(4)
as:
(1) where A is the grey body area (body with an emissivity >1 which is
independent of the wavelength),  is the emissivity of this body,  the
Where q˙ is heat flux in W/m, k is the thermal conductivity of the Stefan-Boltzmann constant (5.67x10-8 W/(m2K4)), with the emissivity of
material in W/(mK), T is the temperature in K and  the gradient operator wood being equal to 0.9 (Incropera and DeWitt, 1990). This value,
(m-1). however is subject to several conditions like surface properties and
wavelength and ranges from 0.85 to 0.95.
A second heat transfer mechanism is convection. For engineering
purposes it is a subclass of conduction in a fluid where bulk flow or flow Thermal energy can also be generated internally by an exothermic
of the thermally excited molecules is the mechanism for transferring chemical reaction, but also by radio-frequency or microwave energy
heated fluid to a colder region (Smith, 1994). The hot fluid must still produced outside the material and converted to thermal energy inside.
transfer energy to a solid surface by molecular collision and then transfer These electromagnetic wavelengths only heat dielectric products. The
it into solid by conduction. Convection may be augmented by forced thermal excitation is selective to dipolar molecules such as water, and
flow to increase the rate of molecular collisions. For instance, in a gas therefore, if the initial conditions are uniform, heating will be uniform,
phase, the convective heat flux (Wm-2) can be expressed as: and consequently the temperature and moisture profiles after pressing.
The heat generated in the medium depends primarily on three variables:
g pg g (2) the intensity of the electromagnetic field applied, its frequency and the
material dielectric properties (Torgovnikov, 1993). Wood fibres are good
electric insulators but interact with electromagnetic radiation due to their
124 125
Carvalho, Martins, Costa
Transport Phenomena

composition of asymmetric molecules and water. This heating method is Table 3.1: Moisture transport mechanisms through a fibrous porous
very capital intensive and nowadays is less viable because, energy costs media (Nilsson et al., 1993).
are very high. Electromagnetic radiation heating has been used in the
Transport properties
forest products industry for over 60 years and modern-day equipment Transport
Phase Local Temperature Relative humidity
Mechanisms
makes this technique especially useful for wood drying and polymer dependency (RH) dependency
curing in the manufacturing of wood composite products, such as Independent (not
particleboard, MDF and OSB, and solid-wood panels (Pereira et al., Gaseous Proportional considering
Gas diffusion Pores
2004). phase to T1,75 swelling/ shrinking
effects)

3.1.3 The Mass Transfer Mechanisms Gaseous

Pores with Independent (not
Knudsen diameter Proportional considering
phase in very
In a fluid, molecular diffusion and convection (bulk flow) are the two diffusion
small pores
less than to T0,5 swelling/ shrinking
10 nm effects)
basic mechanisms of mass transfer. In a porous web structure, additional Surface Adsorbed Fibres
Increase with T Increase with RH
mechanisms of capillary action, surface tension forces, wicking diffusion phase surface
(movement along the fibre) and vaporization are also important (Aguilar, Interior of
Solid internal Adsorbed
2006). In addition, if a phase change takes place, a large quantity of diffusion phase
cell wall Increase with T Increase with RH
thermal energy is involved (latent heat of vaporisation). For a material material

that is drying, this energy comes from the surface layers and therefore the Capillary Condensed
Only when the
Pores pores contain
transport phase
surface temperature decreases. The effect, called evaporative cooling, water
occurs in all drying operations and can be compensated by heat transfer.
Convective mass flux (bulk flow) of water vapour in the gas phase can
The rate of drying in many practical situations is controlled by internal be expressed as:
mass transfer. In porous solids like wood or wood particle/fibre
mattresses, internal mass transfer may occur within the solid phase or v v g (6)
within the void space. Several mechanisms of internal mass transfer have
been proposed in the literature including liquid and vapour diffusion, where v is the water vapour density (in kg/m ) and vg is the velocity of
3

hydrodynamic or bulk flow and capillary flow. Table 3.1 lists the most the gas phase (in m/s).
important mechanisms of moisture transport in porous media. When a fluid, for instance water vapour, is flowing outside a solid surface
For instance, in a mixture of air and steam, the gaseous phase is in convection motion, the rate of convective mass transfer is driven by a
transferred by convection and each component is transferred by diffusion concentration difference of vapour between the solid surface (Cs) and the
and convection in the whole phase. The diffusive flux can be obtained by ambient air stream (Cœ):
Fick’s first law (1833). Fick’s first law established that the rate of
boundary c c œ (7)
diffusion is proportional to a concentration gradient, e.g. water vapour
concentration in air: where kc is the convective mass transfer coefficient (m/s).
v v (5) Simultaneous heat and mass transfer occurs in all drying systems. In a

where n˙ is the diffusive flux in kg/(sm2), Dv is the water vapour simultaneous mass and heat transfer, the temperature gradient can cause
mass transfer (Soret effect) and a mass concentration gradient can
diffusivity in air in m2/s, C the concentration of steam in kg/m3. generate a heat flux (Dufour effect). Luikov (1966) was the first to use
this approach, having applied the irreversible thermodynamics principles
to describe the mechanisms of moisture transport in porous solids by
thermal diffusion.
126 127
Carvalho, Martins, Costa Transport Phenomena

In a porous media, three types of flow can be considered: viscous or

laminar, turbulent and molecular slip flow or Knudsen flow. v
(10)
Laminar flow occurs for Reynolds numbers below 2000. Turbulent cat
flow is present for Reynolds numbers greater than 2300 (Siau, 1984). The where Pv is the vapour pressure above the meniscus, Psat the saturated
Knudsen flow consists of molecular diffusion through a capillary due to vapour pressure, both in Pa,  the liquid surface tension (N/m), MM is the
the pressure gradient arising from the applied pressure differential. When molecular mass of water (kg/kgmol),  its density (kg/m3),  the contact
the capillary radius is smaller or in the same order of the mean free path angle (°), r the capillary radius (m) of the gas-liquid interface, R the gas
of the molecules Knudsen diffusion or split flow occurs. Flow in the constant (8.314 J/(kgmolK)) and T the temperature in K. From this
interstices of a porous media is usually of low Reynolds number, i.e, equation, it is possible to infer that if the meniscus is concave, which
Re<<1, and in this case the pressure gradients drive the flow to balance means that the liquid wets the solid (<90º i.e. 0 <cos <1), the capillary
the viscous stress gradients. The steady-state laminar bulk flow of fluids condensation in a pore of radius r will occur at a pressure below the
through porous media can be described macroscopically by Darcy’s law saturation vapour pressure.
(1856) (with no gravitational effect):

(8)
3.2 HEAT AND MOISTURE TRANSFER IN THE DRYING OF
where v is the velocity of the fluid in m/s,  is the viscosity of the fluid in PARTICLES/FIBRES
Pa s, K is the permeability of the medium in m 2, which is a measure of the
resistance of a porous medium to flow through it, and P is the pressure 3.2.1 Introduction
gradient (Pa). Capillary flow occurs as a result of pressure gradient
induced by capillary forces. The Young-Laplace equation permits the Drying is an important operation in the production of wood-based panels
calculation of the capillary pressure across the interface liquid-gas at (WBP), which consumes a large amount of energy, affects product quality
equilibrium: and without appropriate control, causes environmental concerns (Pang,
2001). The drying of furnish (particles/fibres) is very different from the
c g 1 (9) drying of solid wood. In solid wood drying the operation can take days,
while in drying of WPB furnish takes minutes or seconds at relatively
where Pg is the pressure in gaseous phase and P l the pressure in the liquid high temperatures (Maloney, 1989). Although the fundamental
phase below the meniscus, both in Pa, r the capillary radius (m),  the mechanisms are the same, the drying operation is a little different in the
manufacture process of particleboard, fibreboard and OSB. The operating
contact angle between the liquid and the capillary wall (º) and  the liquid
conditions (temperature, applied pressure) and the moisture content of
surface tension (N/m). In order to re-establish the equilibrium, there will
wood at dryer entry and exit are also different. In the case of
be migration of water from the larger pores to the small ones, i.e. from the
particleboard, furnish arrives to the dryer with moisture content ranging
zones of low capillary potential to the zones of high capillary potential.
from 10 to 200 %. Typical moisture contents are around 30-40 % during
Generally, this kind of transport only is considered for local moisture
summer, and 60-70 % during winter. If the particles or fibres are to be
values above the fibre saturation point (FSP), when the liquid phase is
used with a liquid resin, then they must be dried to about 2-7 %. In MDF,
continuous. However, at the hygroscopic domain, capillary
drying temperatures reach 160 °C for drying times between 2 and 5
condensation can occur in tiny pores, which results from the proximity of
seconds that leads to final resinated fibre moisture contents between 5 and
the solid surface (adsorption effect), but also from the curvature of the
8 % (Ntalos and Grigoriou, 2001). In OSB strands have moisture contents
meniscus (Kelvin effect) (Gregg and Sing, 1991). The first effect is
of around 2 % after drying. The drying temperature depends on furnish
restricted to a distance close to the diameter of fluid molecules. The
moisture content and can vary from 260 to 870 ºC (Maloney, 1989).
relative vapour pressure through a liquid-vapour interface can be
calculated, assuming cylindrical pores, using the Kelvin equation: The most common dryers for particles/flakes are rotary single-pass or
three-pass dryers. Modern single-pass dryers are long-retention dryers

128 129
Carvalho, Martins, Costa
Transport Phenomena

incorporating pneumatic-mechanical conveying of the particles/flakes and

separation (Pang, 2001). A number of operating conditions are associated
operating at low inlet temperatures (400-500 °C). The three-pass dryer
with this operation, such as initial moisture content of furnish and its
provides a pre-drying (interior pass) and a final drying (second and third
variation, particle geometry of furnish, variable ambient conditions, the
passes) and the particles are dried directly with hot gases. The rotary
feeding system to the dryer, contamination and discoloration and fire and
dryer consists of a large horizontally oriented, rotating drum (typically 3
explosion problems (Maloney, 1989).
to 6 m in diameter and 20 to 30 m in length) (see Figure 3.2).
The drying phenomena have been extensively studied for solid wood and
several models have been proposed (Whitaker, 1977; Plumb et al., 1985;
Stanish et al., 1986; Ben Nasrallah and Perré, 1988; Liu et al., 1994;
Kocaefe et al., 2006). However, articles about the drying of particles,
strands or fibres in WBP manufacture are scarce. Although the drying of
solids in a rotary dryer has been extensively covered in the literature,
drying of wood furnish has been studied by only a few researchers.
Kamke and Wilson (1986a, 1986b) analysed the factors affecting the
retention time of particles within a rotary dryer, as well as the heat and
mass transfer with aim of developing a computer model to describe the
drying behaviour. There are also few studies that deal with wood chips
drying in superheated steam in pneumatic conveying (Fyhr and
Rasmuson, 1996, 1997; Johansson et al., 1997a).

3.2.2 Drying Regimes in Convective Drying

The rate of drying for lumber depends on the relative humidity of air, the
air temperature, and the air velocity across the surfaces. The drying
process is usually divided into two different regimes: the period of
Figure 3.2: Dryer in a particleboard plant (Sonae Indústria, Oliveira do “constant drying rate”, where the process is determined by external
Hospital plant). conditions and the period of “falling drying rate”, where the internal
The wet wood particles are mixed directly with hot combustion gases, in a moisture migration limits the drying rate (Johansson et al., 1997a). In
co-current flow pattern, at the inlet to the rotating drum. The gas flow hygroscopic materials like wood, the drying by forced convection can be
provides the thermal energy for drying, as well as the medium for described as follows:
pneumatic transport of the particles through the length of the drum.  First phase: After establishing the equilibrium and while the
Interior lifting flanges serve to agitate and produce a cascade of particles moisture content is above the FSP, the temperature of the wood
through the hot gases (Shu et al., 2006). So, heat is transferred from the does not change. So, all the heat transferred to wood is used to
gas to the wet solids as a result of a temperature driving force, and the evaporate the water. The vapour pressure at the wood surface is
particles increase in temperature and lose moisture (Kamke and Wilson, equal to the saturation vapour pressure; drying is controlled by the
1986). The water vapour is then transferred to the gas stream under a external resistances to heat and mass transfer. This is the phase of
vapour pressure gradient. “constant drying rate” and it cannot be observed if the external
In the case of MDF, the mixture of wet fibres and steam flows through a conditions are severe (high temperature and low relative humidity
blowline, where the adhesive and other additives are added to the fibres of air).
and then into a tube dryer. A tube dryer typically has a horizontal part  Second phase: This phase occurs when the capillary water
followed by a vertical part and the total length is normally around 100 m. migration is lower than the evaporation flux. So, the moisture
The mixture of air, steam and resinated fibres enters a cyclone for content at surface will fall below the FSP and the vapour pressure

130 131
Carvalho, Martins, Costa
Transport Phenomena

is then lower than the saturated vapour pressure, according to the

drying rate is dominated by internal moisture movement within the fibre
sorption isotherms. As drying progresses the moisture
walls (Pang, 2001).
concentration gradient in wood becomes lower and so the rate of
moisture migration, leading to a decrease of drying rate: it is called
the “falling drying rate”. The resistance to the transfer of moisture
inside the material becomes the most important mechanism.
 Third phase: When all the material is in the hygroscopic domain,
drying continues slowly until the equilibrium between the material
and the external air stream is reached. The drying rate is controlled
by heat transfer, because the moisture adsorption decreases as
temperature increases.
Due to wood anisotropy, the three phases can coexist, depending on the
size of material and properties as water diffusivity in the transversal and
longitudinal directions.

3.2.3 Heat Transfer

In drying of particles, it is important to consider the transport of heat

inside the particles (internal flux) and also at the surface of the particles Figure 3.3: Example of a typical evolution of temperature and relative
(external flux). So, heat is transported internally by convection and humidity (RH) along the fibre dryer in the production of
conduction whereas external heat transfer from the surface occurs via MDF (Mangualde plant (Portugal) from Sonae Indústria,
convection and radiation. data collected by Pedro Reis).

In the conventional drying of solid wood, the heat transfer by convection

is less important than the transport by conduction (Liu et al., 1994). The 3.2.4 Moisture Movement
convection mechanism is only considered in drying by forced
convection (Ben Nasrallah and Perré, 1988), by vacuum at moderate Moisture in wood can exist in three forms: water vapour in the pores,
temperature (Fohr et al., 1995) and convective drying at high capillary or free (liquid) water in the pores and hygroscopic or bound
temperature (Pang and Keey, 1995). In drying of wood particle/fibres, water in the cell walls (Simpson, 1983-84a, b). Considering the
heat is transferred to the material primarily by convection air currents and mechanisms of moisture movement in wood, two situations must be
so this mechanism is the most important. Therefore, the drying medium considered:
properties (air and water vapour) will affect the process. The period of  Below FSP: Moisture movement as vapour though the cellular
constant drying rate which does not exist (or is very short) using air cavities and pit openings; adsorbed or bound water movement
drying, becomes more significant with decreasing amounts of air in the through the cell walls. Molecules are fixed at sorption sites by van
drying medium, in particular when using pure superheated steam der Waals forces and hydrogen bonds. Water molecules can only
(Johansson et al., 1997a,b). In case of MDF, because the fibre move if they have sufficient energy to escape from their adsorption
temperature at the blowline outlet is normally higher than the wet-bulb sites and become activated molecules. (see Figure 3.4c)
temperature at the dryer inlet, flash drying occurs immediately after the
wet fibres and steam enter the dryer. Flash drying occurs in a very short  Above FSP: Free water movement due to a capillary pressure
distance, and so heat and also mass transfer can be evaluated using only gradient or due to the increase of gas volume inside the cell lumens
the length of a single fibre (see Figure 3.3). Because heat and mass with the increase of temperature. (see Figures 3.4a and 3.4b)
transfer coefficients between the fibre and the air stream are high, the

132 133
Carvalho, Martins, Costa So the mechanisms of mass transfer through wood during drying can be
divided into two main classes according to Siau (1984):
  Diffusion; intergas diffusion, which includes the transfer of water Transport Phenomena
vapour through the air in the lumens of the cells and bound water,
within the cells walls;
The drying process of fibres in MDF production was described by Pang
 Bulk flow of fluids through the interconnected voids under the (2001) as follows. In a micro-structural scale, the fibre wall contains both
influence of a static or capillary pressure gradient. holes (pit openings) and fine pores due to the mechanical action of the
refining. At the dryer entrance, fibres are very wet and covered with
In the first phase of wood drying, water in the liquid state migrates easily liquid water (from the refining operation and from the adhesive spraying)
longitudinally and founds a transversal path (see Figure 3.4a). During the and therefore the fibre surface is vapour saturated. During drying, the
third phase, the water migration proceeds essentially along the thickness liquid flows outwards through pit openings (with a diameter of about 4
direction (see Figure 3.4c). During the second phase, there is a transition µm) and then through the large pores. The process persists until the fibre
from a mechanism mainly longitudinal to a transversal mechanism for moisture content is decreased to below 50%. At this point, some liquid
moisture transfer (see Figure 3.4b). The migration purely in the thickness still remains in very fine pores and this liquid will evaporate within the
direction settles when all the material is in the hygroscopic domain (see material. At the same time the vapour diffuses through the fibre material
Figure 3.4c). towards the outer surface. Once the remaining liquid water has been
For drying processes at low temperature (conventional drying of solid removed drying slows down and become controlled by bound water
wood), the migration of water to the surface is due to a difference of diffusion and water vapour movement within the cell walls (Pang, 2001)
chemical potential, whereas at high temperature (drying of
particles/fibres), the migration of water is mostly due to a pressure 3.2.4.1 Moisture diffusion
gradient in wood, due to a high temperature gradient. These statements
Water vapour diffusion in air can be described by Fick’s law, driven by
are valid for both solid wood and for particles/fibres drying.
a partial pressure gradient and is much more rapid than bound water
diffusion. Bound water diffusion in wood is sometimes described by
Fick’s law, but this has generated some controversy, regarding the driving
force used (moisture content, partial pressure of water vapour and
chemical potential). The traditional approach assumes the moisture
gradient as the driving force for diffusion (Stamm, 1959; Siau, 1984;
Simpson, 1993). Another line of thought considers bound water diffusion
as a response to a vapour pressure gradient (Bramhall, 1976). The vapour
pressure gradient presents more consistency with the experimental results
and has a more universal application, as in the non-isothermal drying. At
low moisture content, diffusion is probably best described by chemical
potential driving forces (Siau, 1984). Below FSP (hygroscopic domain),
water movement can follow a continuous path though cell wall or through
a combination of cell walls and cell cavities. Stamm (1964) analysed this
situation with an electrical analogy using parallel and series networks of
Figure 3.4: Mechanism of moisture movement during wood drying a-first diffusion paths.
phase (moisture above PSF), b-second phase c-third phase (all
the material is in the hygroscopic domain) (from Moyne, 3.2.4.2 Bulk flow of gaseous phase
1983).
In case of drying, only bulk flow of the gaseous phase has interest. Bulk
flow of liquids has practical applications in wood preservation (pressure
or vacuum treatment) and the impregnation of wood with chemicals, e.g.
pulping, wood polymer composites. Darcy’s law is considered a good
approximation to describe the one-phase flow through wood (Plumb et
al., 1985; Siau, 1984), although some special situations in which Darcy’s

134 135
Carvalho, Martins, Costa of red oak and the presence of nonlinear flow due to kinetic energy losses
where fluids enter pit openings from vessel or tracheid lumens. It can occur
law would not be obeyed are reported by Siau (1984). This is the case of at Re numbers between 0.04 and 16, assuming radii from 0.005 to 2 mm
turbulent flow that perhaps could occur in the largest earlywood vessels and a pit membrane thickness approximately 0.1 m.
 Transport Phenomena
3.2.4.3 Capillary flow of liquid

The flow of free water above FSP requires both a continuous passageway of liquid and gas permeabilities were also performed for some species
for flow and a driving force. The continuous passageway is provided (Siau, 1984). Longitudinal permeabilities range from 10 -10 for red oak to
mainly by cell cavities and interconnecting pits (pit openings range from 10-15 m2 for Douglas-fir heartwood. Transverse permeabilities range from
0.1 to 1 μm in diameter (Simpson, 1983-84a). The driving force is the 10-16 to 10-18 respectively (Siau, 1984). For softwood, the ratio transverse-
capillary pressure (difference between the pressure in the gas phase and to-longitudinal is around 1/20000 (Comstock, 1970). For hardwoods they
in liquid phase). Bulk flow of liquid water can be described using Darcy’s are even lower due to the high connectivity of vessels. Wood permeability
law and is determined by liquid permeability. affects drying behaviour: the period of constant drying is longer for high
permeability species, because the capillary forces tend to keep the surface
The capillary pressure is calculated using the Young-Laplace equation. above FSP; the falling rate period is faster for higher permeability species
For a 10 m radius ( =73 mJ/m2 and complete wetability  =0 º), which in the longitudinal direction.
could be the typical value of a softwood tracheid, the capillary pressure
will be 14.6 kPa, while for a 1 m radius, typical of a large pit opening, 3.2.5.2 Geometry
the capillary pressure will be 146 kPa. If the gas pressure is 1 bar (100
kPa), then the pressure below the meniscus is –46 kPa. This negative The size and the shape of the individual particles in a furnish are
pressure is called capillary tension (Siau, 1984). These elevated capillary significant factors that influence wood drying. With existing technology it
tensions that are generated in wood during drying are responsible for is impossible to achieve uniform particles (Maloney, 1989). The total
phenomena as the collapse of wood and pit aspiration that reduces drying time increases with the chip thickness: if the thickness is halved,
permeability. As it was mentioned above, even below FSP, capillary the total drying time is approximately halved too. In thinner chips, the
condensation can occur in pores In case of wood, and considering for maximum overpressure in the centre of material is higher, because the
instance a relative humidity of 90 %, i.e. a relative pressure of 0.9, the distance for the transport of heat from the surface is halved and the main
maximum radius for which capillary condensation will occur at ambient pressure release occurs in the longitudinal direction. When the
temperature, is 0.01 m ( =73 mJ/m² and complete wettability  =0 º), a longitudinal dimension is doubled the maximum overpressure increases
little lower than the range of pit membranes openings (around 0.05 m). by a factor of more than two because of the larger length of longitudinal
flow necessary for pressure equalisation. The total drying time is not as
3.2.5 Effects of Wood Chip Properties sensitive to the wood chip length. So for equal volumes, a thin chip is
better for drying purposes than a short one, due to the highest resistance
3.2.5.1 Permeability to flow in the thickness direction (Fhyr and Rasmuson, 1997).
Suggested R&D Topics include:
For wood, permeability is dependent on wood structure. The
permeability of wood cannot be solely related to its porosity but also to  Influence of the drying conditions on the subsequent wetability and
the availability of interconnecting pits and perforation plates between its absorption of the resin
cells (Siau, 1995). Several authors proposed models to estimate the
permeability of wood considering its anatomical aspects (Comstock,  Influence of the drying conditions (moisture internal profile) on the
1970; Spolek and Plumb, 1980; Siau, 1984). Experimental measurements industrial press operation
 Influence of the drying conditions on the final product VOC
emissions

136 137
Carvalho, Martins, Costa
Transport Phenomena

3.3 HEAT AND MASS TRANSFER MECHANISMS IN HOT

PRESSING

3.3.1 Introduction

Mattress consolidation of a WBP is achieved through hot-pressing. The

thermal energy is used to promote the cure of the thermosetting adhesive
and soften the wood elements. The mechanical compression is needed to
increase the area of contact between the wood elements to allow the
possibility of adhesive bond formation. This operation has a major effect
on the balance of properties of the resulting panel; a rigorous control of
all processing variables is necessary to achieve appropriate product
quality and to minimise pressing time. Several coupled physico-chemical-
mechanical phenomena are involved during hot-pressing, making this
operation very complex (see Figure. 3.6). Besides heat and mass transfer,
other mechanisms are also involved as the polymerization of the adhesive
and the rheological behaviour of the wood elements. Several researchers
have described these mechanisms with the aim of modelling the hot-
pressing process. These include Bolton and Humphrey (1988), Humphrey Figure 3.5: Continuous press in a particleboard plant (Sonae Indústria,
and Bolton (1989a), Thoemen (2000), Zombori (2001), Fenton et al. Oliveira do Hospital plant).
(2003), Carvalho et al. (2003), Dai and Yu (2004), Pereira et al. (2006),
and Thoemen and Humphrey (2001, 2006). The mat of wood particles/fibres is a capillary porous material in which
the voids between and within particles/fibres contain a gas mixture of air
Several types of presses can be used: batch or continuous, steam injection, and steam. Globally, the most important mechanisms of heat and mass
plate and/or radio-frequency or micro-wave heated. The most common transfer are (Pereira et al., 2006):
type is the batch press with heated plates (multi-opening), but in the last
decade, continuous presses (see Figure 3.5) with moving heated belts are  Heat is transferred by conduction due to temperature gradients and
substituting batch presses, particularly for particleboard and MDF by convection due to the bulk flow of gas: conduction follows
manufacture. Regardless of the press design, the mechanisms of heat and Fourier’s law;
mass transfer in the mat are the same, but vary by degree of importance
and direction of flow (Kamke, 1994). In this operation a number of  The gaseous phase (air and water vapour) is transferred by
factors are involved, related not only with the material itself but also the convection; each component is transferred by diffusion and
operating conditions, including resin type, catalyst, press temperature, convection in the whole phase. Diffusion follows Fick’s law and
wood species and fibre/particle geometry, mat moisture level content and the gas convective flow follows Darcy’s law: the driving force for
distribution, pressing time (batch process), pressing speed (continuous gas flow is the total pressure gradient and for diffusive flow is the
process), mat volume and its relation to the board density profile, steam partial pressure gradient;
pressure within the board during pressing and pre-cure and post-cure of  The migration of water in the adsorbed phase occurs by molecular
the resin (Maloney, 1989). diffusion due to the chemical potential gradient of water molecules
within the adsorbed phase;
 Phase change of water from the adsorbed to the vapour state and
vice-versa.

138 139
Carvalho, Martins, Costa
 Polymerisation Transport Phenomena
reaction
Moisture content, temperature, water
vapour pressure, É
pressure gradient that will create a flux of heat by convection to the
edges. When the temperature of the medium exceeds the ebullition point
of water, imposed by the external pressure, the horizontal pressure
gradient becomes the more important driving force (Constant et al.,
1996). However, it is not necessary to attain the ebullition point of free
water to have a vapour flow. Any change in temperature will affect the
EMC of wood and so the vapour partial pressure in the voids (Humphrey
Mechanical properties
(viscoelastic behaviour) Mass and heat transport
and Bolton, 1989a). Also, if the vapour is cooled, it will condense,
Hot-pressing operation properties liberating the latent heat and a rapid rise of temperature will occur. So,
there is also a phase change associated with the bulk flow, which imparts
the temperature change (Kamke, 2004). This condensation will happen
continuously from the surface to the core and not as a discrete event,
Densification, stress relaxation which complicates the modelling of this system.

3.3.2.3 Heat transfer by radiation

Transverse compression
Heat transfer by radiation is usually neglected, since for the relatively
lower range of temperatures (< 200 ºC), it would be insignificant
Figure 3.6: Mechanisms involved during the hot-pressing of wood-based compared with conduction and convection. However, during press closing
panels and their interaction. and before the platen makes contact with the mat, as well as during the
first instants of pressing while mat density is relatively low, heat transfer
by radiation can be a significant part of the total heat transferred
3.3.2 Heat Transfer (Humphrey and Bolton, 1989a). On the other hand, on the exposed edges
the heat is continuously transported to the surroundings by radiation
3.3.2.1 Heat transfer by conduction (Zombori, 2001).
The conduction is one of the more important mechanisms of heat transfer Heat conduction Heat convection
in the hot-pressing of WBP, mostly in case of batch pressing, in which
Hot-plate Hot-plate
almost all the heat supplied to the mat comes from the heated platens.
Heat is transferred through the interface plate/mat to the interior by
conduction and will be used to resin polymerisation and to remove water
presented in the mat as bound water. To remove this water it is necessary
to supply energy equal to the sum of the water latent heat of vaporization
and the heat of wetting (or sorption) sufficient to break hydrogen bonds
between water and wood constituents.
Core

3.3.2.2 Heat transfer by convection

Cell wall substance Voids

Convection occurs because the heat transferred from the hot platens Figure 3.7: Heat transfer mechanisms in a WBP mat (adapted from
causes the vaporization of moisture, increasing the water vapour pressure. Thoemen and Humphrey, 2001).
A gradient of vapour partial pressure is formed across the board
thickness, causing a convective flux of vapour towards the mat centre. On
the other hand, the increase of gas pressure will cause a horizontal

140 141
Carvalho, Martins, Costa The other possible sources are the exothermical reaction of the resin cure
and the heat of compression. The contribution of the heat of compression is
3.3.2.4 Other sources generally neglected. Bowen (1970) estimated that its contribution for heat
transfer was around 2 %. The contribution of the exothermic polymerisation
of the resin depends on the reaction rate and condensation enthalpy. Bowen Transport Phenomena
(1970) showed that as much as 22 % of the total heat came from this
reaction, although 10 to 15 % may be a more reasonable level (Maloney,
1989). Diffusion inside the mat during hot-pressing includes vapour diffusion
and bound water diffusion. The driving force for the diffusive flow of
Suggested R&D Topics include: vapour is the partial pressure gradient. The convective and diffusive
fluxes occur simultaneously, but it is widely accepted that convective gas
 Amount of heat that originates from conduction and from the flow is the predominant mass transfer mechanisms during hot-pressing
condensation of vapour (Denisov et al., 1975; Thoemen and Humphrey, 2006).
 Percentage of heat transfer by convection and by conduction
3.3.3.2 Mass transfer by diffusion
 How can heat transfer be maximized?
The migration of water in the adsorbed phase occurs by molecular
3.3.3 Mass Transfer diffusion and follows Fick’s first law with the chemical potential gradient
of water molecules within the adsorbed phase as the driving force to
3.3.3.1 Mass transfer by convection diffusive flux. This is a slow process and thus it is often considered
negligible by some authors (Carvalho et al., 2003) in comparison with
In WBP hot-pressing, it is generally assumed that moisture content is steam diffusion. Zombori and others (2003) studied the relative
below the FSP and so the water is present as vapour in cell lumens and significance of these mechanisms and they found that the diffusion is
voids between particles/fibres and bound water in cell wall (Kavvouras, negligible during the short time associated to the hot-pressing.
1977; Humphrey, 1982; Carvalho et al., 1998; Carvalho et al., 2003; Local equilibrium is in general assumed to describe the mass transfer. The
Zombori, 2001; Thoemen and Humprey, 2006; Pereira et al., 2006). local moisture content is equal to the EMC. The adsorbed water and
Two main mass phases are then considered, the gaseous phase (air + steam are related by a sorption equilibrium isotherm. Any change in
water vapour) and the bound water and local thermodynamic equilibrium temperature will affect the EMC and thus the partial pressure of the
is also assumed. The gaseous phase is transferred by convection due to a vapour in the surrounding air (Humphrey and Bolton, 1989).
gas pressure gradient (bulk flow) and the vapour is transferred by
Gas convection Gas diffusion
diffusion.
Hot-plate Hot-plate
The bulk flow occurs in response to a gas pressure gradient caused by the

Core
vaporisation of moisture present in mat. In conventional hot-pressing, this Cell wall substance Voids, filled with water
containing bound water vapour-air mixture
bulk flow is generally assumed to be laminar and the contributions of
other mechanisms (turbulent and Knudsen) of gas flow are neglected Figure 3.8: Mass transfer mechanisms in a WBP mat (adapted from
(Sokunbi, 1978; Humphrey, 1982, Denisov et al., 1975; von Haas et al., Thoemen and Humphrey, 2001).
1998). So, the gas convective flow follows Darcy’s law (Carvalho et al.,
2003; Zombori, 2001; Fenton et al., 2003; Thoemen and Humprey, 2006).
In the case of steam injection pressing, this assumption is not valid, since
turbulent flow is likely to be of some importance (Hata et al., 1990).

142 143
Carvalho, Martins, Costa
Transport Phenomena

3.3.3.3 Capillary transport

level of gas affects heat convection. A high level of gas pressure can
At press entry the moisture content of the furnish is relatively low cause local or even a complete delamination of the mat, when the pressure
(generally below 14 %) and although a possible presence of liquid water is released during press opening and the resin bond strength is not
brought by the adhesive and capillary condensation in some tiny pores, it sufficient to resist this pressure drop.
is generally assumed that the whole mat is below the FSP (Kavvouras,
1977; Humphrey, 1982; Zombori, 2001; Thoemen and Humprey; 2006). 3.3.4.1 Monitoring the internal conditions
In case of particleboard, the moisture content at the press entry might be
11 %, while the particle moisture content before resin blending could be With the aim of studying the effects of several hot-pressing variables on
around 2-4 %. During blending, considerable quantities of water are the final board properties, several researchers have attempted to monitor
added with the resin (water content around 50 %), and so unless the this operation by measuring the temperature and gas pressure inside the
equilibrium is achieved by the furnish before entering the press (in that mat during lab trials, but also in industrial presses (Maku et al., 1959;
case, the water will be adsorbed in the cell walls of wood) some capillary Strickler, 1959, Kamke and Casey, 1988a,b; Humphrey, 1991; Bolton et
translation might occur (Humphrey and Bolton, 1989). In case of MDF, al., 1989b; García, 2002; García et al., 2001, 2003).
the fibre drying after the resin spraying in the blow-line results in the
decrease of moisture and it is reasonable to consider that the equilibrium
will be attained before the hot-pressing, and thus the water will be
adsorbed in the fibres (Carvalho, 1999).
There is also a possibility of capillary condensation in tiny pores. In case
of WBPs, the relative humidity does not exceed 90 % (Humphrey, 1984,
Kamke and Casey, 1988) and considering a temperature of 115°C, inside
the mat, the maximum pore diameter filled with water will be 0.007 µm.
This will correspond to capillary pressures of 14.6 to 20 kPa, which are an
order of magnitude less than the predicted maximum vapour pressure
differential between the centre and the edges of board (at atmospheric
pressure). So, even if some fine capillaries do fill by capillary
condensation, it is unlikely that capillary translation of liquid will occur
(Carvalho, 1999). Figure 3.9: Temperature and gas pressure within a mat for a 16 mm
particleboard during continuous hot pressing ■=surface
3.3.4 Internal Mat Conditions layer, ^=intermediate layer, ◆=core layer. Top three
lines=temperature curves (from Meyer and Thoemen, 2007).
During the hot-pressing cycle, the internal conditions of the mat change in
space and time. The transient and simultaneous phenomena of Temperature was measured by using thermocouples and the total gas
momentum, heat and mass transfer inside the mat do not allow a pressure was measured by using a pressure probe consisting of a small
straightforward prediction of the internal mat environment. Although diameter stainless steel tube connected to a pressure transducer. The tube
experimental measurements of these conditions are useful, its prediction and the transducer may be filled with a low vapour pressure fluid, as
demands mathematical modelling, which also provides knowledge for a silicon oil, to reduce the dead volume in the hydraulic line (avoiding
better understanding of the phenomena involved. condensation of water vapour inside the tube) and also transfer the
pressure change at the open end of the microbore tubing to the pressure
Temperature and gas pressure affect the hot-pressing process and so the sensitive diaphragm in the transducer well (Kamke and Casey, 1988a).
properties of WBPs. High temperatures are needed to achieve rapid cure This technique was developed and demonstrated by Kavvouras (1977)
of the adhesive and assist mat compaction by wood plasticization. Due to and Kamke and Casey (1988). They measured the gas pressure inside a
the compaction and the vaporisation of water, gas pressure builds up. The flakeboard with 510 x 160 x 19 mm and target density of 720 kg/m 3, a
platen temperature of 190 °C and press closure time of 1 minute and they

144 145
Carvalho, Martins, Costa
obtained maximum values from 40 kPa to 200 kPa for initial moisture Transport Phenomena
contents between 6 and 15 %. Bolton et al. (1989a) observed that for a
particleboard with 283 x 283 x 15 mm, platen temperature of 160 ºC and
final density from 450 to 850 kg/m 3 vapour pressure attained values of the Suggested R&D topics include:
same magnitude, from 20 to 200 kPa. They found significant differences  Development of new on-line sensors for monitoring the mat
in vapour pressure inside an industrial board and inside a laboratory board internal conditions, especially in continuous pressing
made with the same furnish and press conditions (Humphrey and Bolton,
1989a). In a 15 mm thick board (1.8 x 3.65 m) of 650 kg/m 3 density,  Development of systems for the correct positioning the probes
initial moisture content 16 % for surface layers particles and 12 % for during continuous pressing
core layer particles, the maximum core vapour pressure was at least
200 kPa at the end of press cycle, while in the laboratory board, they  How to treat the collected data and deal with its uncertainty?
obtained a maximum of 60 kPa, 150 s after press closure. Steffen et al.  How to overcome the problems encountered in the measurement of
(1999) monitored temperature and gas pressure inside a mat in movement gas pressure originated by the rheological transformations in the
during continuous pressing. The entire measurement equipment including mat during pressing?
the cable travelled through the hot press. A miniature pressure transducer
covering a range from 0 to 700 kPa was imbedded in a plastic case. The  How to overcome the difficulties to measure moisture content at
sensor was prepared for sealed connection with air-filled stainless tubes several positions inside the mat?
(diameter 2 mm). A portable system PressMAN® was developed by the
Alberta Research Council and measures mat thickness, hydraulic 3.3.4.2 Evolution of the internal conditions during hot-
pressure, core temperature and gas pressure (Alexopoulos, 1999). pressing
Meyer and Thoemen (2007) monitored the continuous hot-pressing of A typical trend of the evolution of temperature and vapour pressure
particleboard using this system. Deng et al. (2006) also used the during the hot-pressing can be described as follows. Heat is transported
PressMAN® to monitor the internal conditions during the hot-pressing of by conduction from the hot platen to the surface. This leads to a rapid
laboratory MDF mats that were pre-heated with microwaves. During the rise in temperature, vaporising the adsorbed water in the surface and thus
preheating with microwaves or HF, temperature is measured with fibre increasing the total gas pressure. The gradient between the surface and the
optic sensors (Deng et al., 2006; Celeste et al., 2004) (see Figure 3.10). core results in the flow of heat and vapour towards the core of the
mattress, therefore increasing the gas pressure. As a consequence, a
positive pressure differential is established from the interior towards the
lateral edges, and then a mixture of steam and air will flow through the
edges. The rate of temperature rise is larger at the superficial layers,
because the core temperature is less affected by platen temperature
(Kamke and Casey, 1988a,b) (see Figure 3.11). The temperature rise at
the core is the balance of two mechanisms: the gradient of gas pressure
that causes the flow of vapour to the centre and the effect of water loss
from the board. This temperature rise at the core is thus linear. The ratio
between the amount of heat lost by moisture escaping from the board
edges and the input of heat from press platens is smaller in an industrial
press than in a laboratory one. So, due to a higher steam production at
board centre in an industrial press, a higher build up of gas pressure is
Figure 3.10: Temperature time evolution at board centre during the pre- observed (Humphrey and Bolton, 1989b). If the water vapour gives up
heating of MDF for mats with thicknesses of 50, 70 and 80 enough heat to the surrounding mat during its journey to the centre it will
mm (adapted from Pereira et al., 2004).
condense (Kamke, 2004). If the water vapour reaches the equilibrium
with the moisture content at the prevailing temperature, it will be
adsorbed and become bound water, if not it will remain as liquid water.

146 147
Carvalho, Martins, Costa system as the temperature gradients near the surface of the board decrease
and also to the increase of vapour loss from the edges of the board.
A progressive decrease in the rate of temperature rise occurs after this Humphrey (1982) also attributes this fact to the increase of heat of wetting
period, which may be attributed to a decrease in heat transfer to the as the moisture content falls, thus resulting in a reduction of water
evaporated per unit of energy input from the heated platens. The change Transport Phenomena
of temperature at the surface attains almost a plateau, when the rate of
vapour gain from the surface becomes smaller than the rate of lateral
vapour loss from surface layers, which implies a decrease of the gas 3.3.4.3 Influence of pressing parameters on the internal
pressure. When venting begins (prior to press opening), water vapour conditions
pressure decreases and equalizes through the mat, while temperature stays The levels of temperature and gas pressure are affected by pressing
high. The venting gas may move to the edges and toward the surfaces of parameters, namely press closing time, initial moisture content and platen
the mat. temperature. A faster press closing time results in a larger rate of mat
densification. This leads to a rapid build-up of steam pressure in the
surface region and thus a faster rate of convective heat transfer to the core
(Kamke and Casey, 1988b). Low moisture content and low platen
temperature result in a low gas pressure. High initial moisture content
increases the rate of temperature rise in the core region but the maximum
temperature may not be higher than the temperature found in a lower
moisture content mat. Kamke and Casey (1988a) showed that high platen
temperature and high moisture content mats results in a large deviation of
gas pressure between the surface and core layers after press closing.
140
2.5 Temperature and gas pressure will also influence the viscoelastic
temperature
120 behaviour of the mat and so the formation of a vertical density profile.
Compressiv
2.0
100 Wolcott et al. (1990) investigated the interaction of temperature, moisture
e
content, and compaction on the formation of a vertical density gradient in
80 D 1.5stres
A s the board. On the other hand, particle geometry, wood density and the
60 stress B (MPa compression of wood particles will affect the porosity and permeability.
1.0
40 ) A larger permeability allows for a larger rate of gas flow through the
edges, which will retard the pressure build-up and thus lead to a lower
20 0.5
plateau temperature. Winistorfer et al. (2000) developed a radiation-based
0 system for the in-situ measurement of density during board consolidation.
0.0
This system coupled with former ones for temperature and gas pressure
0 40 80 120 160 200
measurements should lead to better insight of the consolidation and
time (s) formation of vertical profile. García et al. (2003) measured the internal
Figure 3.11: Compressive stress and temperature time evolution at board temperature and vapour pressure responses to flake alignment during
centre during the hot-pressing of MDF mats for three platen OSB hot-pressing. The results indicated that lower density mats will heat
temperature: A-140 °C; B-150 ºC and D-160 ºC (from faster and have lower internal gas pressure. Less aligned mats will
Pereira, 2002). initially heat faster and have higher internal gas pressure, but towards the
end of pressing they will take slightly longer to heat.
Suggested R&D topics include:
The measurement of the evolution of moisture content throughout the mat
 How to control the steam flux through the edges to have a uniform during the hot-pressing is also important as it affects the rheology of the
density profile? wood and the cure of the adhesive. There is evidence that the cure of
 How to measure gradients of steam pressure and temperature from resins that polymerize by condensation reactions is affected by the
board centre to edges? relative humidity of the surrounding air (Carvalho, 1999). This kind of
data would give also information for a qualitative understanding of the
 How the change in water vapour pressure affects the resin dimensional stability of the panels. Other transport properties, such as
polymerisation reaction? conductivity, are also influenced by moisture content. Unfortunately,
there are still difficulties of measuring moisture content on line and thus

148 149
Carvalho, Martins, Costa importance.
Suggested R&D topics include:
the prediction of this variable by mathematical modelling is of great
  A quantitative understanding of responses of pressing variables is Transport Phenomena
needed
 More lab-scale trials at limit conditions (high moisture content) and or fibreboard, where the particles tend to lay randomly in a plane parallel
different wood furnishes to the plane of the board (Kamke, 2004), there will be a difference
between the out-of-plane (thickness) and the in-plane (width and length)
 Need for correlations between lab and industrial presses data thermal conductivity. The transversal conductivity was considered by
Siau (1984, 1995) as a combination of the thermal conductivity of cell
3.3.5 Transport Properties walls, the conductivity of air and the conductivity of moisture. He
considered a geometrical model for a single wood cell with cubic
geometry. To derive the transverse conductivity, he considered the
conductance of the cross walls (cell-wall substance) in series with the
conductance of the lumen (air) and the side walls (cell-wall substance).
Based on this approach, Zombori (2001, 2002) considered for a strand
board the following expression for the thermal conductivity in the
thickness direction:
Transport properties are dependent on changing physical properties that
vary with respect to space and time during hot-pressing. They are a a
a T
(11a)
function of temperature, moisture content and vapour pressure. a CN T CN
Additionally, the physical properties are dependent on mat structure and
so, the transport properties are direction-dependent. Although all WBP
are comprised mainly of wood their very different in structures and T cw w a (11b)
compressibilities causes the transport properties of plywood,
particleboard, MDF, and OSB to be very different from one another. On where ka is the conductivity of air (0.024 W/(mK)), kT the conductivity of
the other hand, they are not well known in some cases, considering the wood in transverse direction (W/(mK)), kcw the conductivity of cell wall
environment of high temperature and gas pressure. So, they are generally substance (0.217 W/(mK)), kw the conductivity of water (0.4 W/(mK)),
estimated using empirical correlations that sometimes were not obtained sm the space fraction of the mat,  the porosity of wood, M the moisture
for the kind of WBP in study, but for another kind (mostly particleboard) content of wood (fraction) and SG the specific gravity of wood. The
or even for solid wood. thermal conductivity of the mat in the lateral direction was estimated in
a similar manner, considering the orientation of the strands quantified
Suggested R&D topics include: through a degree of alignment. The longitudinal thermal conductivity of
solid wood was considered as 2.5 times higher than the transverse
 More data on physical and mechanical properties of mats during hot-
conductivity (Siau, 1995). For the thermal conductivity of a particleboard
pressing are needed
in the thickness direction, Humphrey (1982) derived an experimental
 How to deal with changing wood raw material mix (wood species relationship and included a correction term for temperature (Kuhlmann,
mix and geometry of wood particles) and their influence on 1962; Humphrey, 1982) and another for moisture content (Kollmann and
material and transport properties? Malmquist, 1956; Humphrey, 1982):

z (12a)
3.3.5.1 Thermal conductivity

Like solid wood, WBPs are relatively good insulators, especially in the
direction perpendicular to the plane of the board, where there is high
resistance to flow due to the interruption of the path by the poorly (12a)
conducting air-filled pores. In particular, softboard or ultra-light MDF
that are low-density fibreboards, are especially designed to be insulators
in construction applications. In any flat-pressed WBP, as particleboard

150 151
Carvalho, Martins, Costa experimental work of Ward and Skaar (1963) with an UF (urea-
formaldehyde) bonded particleboard. The same expression was used by
For the conductivity parallel to the board plane (kx), Humphrey (1982) Carvalho et al. (1998, 2003), Pereira et al. (2006) and Nigro and Storti
assumed a value given by equation (12a) multiplied by 1.5, based on (2001). Fenton et al. (2003) and Lee et al. (2006) used for an OSB mat, the
equation 11a combined with the correction for temperature in 11b. Transport Phenomena
Thoemen (2000) and Thoemen and Humphrey (2006) used the following
expression for MDF, where the thermal conductivity
0.30
perpendicular to the
k
p (14)

This expression was used by Carvalho et al. (1998, 2003), Pereira et al.
(2006) and Nigro and Storti (2001). Zombori (2001) used a similar
expression given by Skaar (1972):

b v
d d (15)
p
plane in W/(m·K) is calculated by the sum of T , which denotes the
thermal conductivity measured at 0 % moisture content and 30 °C in where d , v and.b are respectively, the density of dry-wood, air and
function of density  in kg/m3 (von Haas, 1998) and a correction factor bound water in kg/m3. Thoemen (2000) and Thoemen and Humphrey
derived by Haselein (1998) to account the moisture content M (%) and (2006) used an expression from Haselein (1998) derived using a mixture
temperature T (ºC) effects: rule to account for moisture:
0.30
T T (13a)
p
T (16)

Fenton et al. (2003) and Lee et al. (2006) calculated the specific heat
0.30
T –8 2 –5 –2
(13b) (J/kg.ºC) using an expression from Simpson (1999) only in function of
temperature (ºC):

p (17)
–3 –4 –5
T esp (13c)
Suggested R&D topics include:
with Texp being the average experimental temperature used in the
 New correlations are needed for mat specific heat for different
measurements of thermal conductivity (here, 30 ºC).
types of furnish
Suggested R&D topics include:
3.3.5.3 Permeability
 What is the contribution of thermal conductivity to the increase of
temperature inside the mat? The rate of convective heat transfer to the panel core is controlled by its
permeability in thickness direction. Permeability in the plane of the
 What is the thermal conductivity of modern-day fibre and particle board also controls the flow of vapour to the panel edges (Hood, 2004). In
mattresses? the transverse direction, the permeability of WBPs is highly dependent on
panel densification. So, this property will change during the hot-pressing
3.3.5.2 Specific heat operation; during the press closure the permeability is high, after press
closure the permeability will be considerably lower and when the venting
The specific heat of the mat is the amount of heat necessary to increase
begins, the press opens slowly, generating additional gaps between
the temperature of one unit of mass by one degree in (C or K). It is
particle/fibres and thus increasing the permeability.
generally estimated by adding the specific heat of dry wood and that of
water according to the material porosity and assuming full saturation. The This property is dependent on panel structure which in turn is a function
specific heat Cp (J/(kg-·K)) increases with temperature T(K) and moisture of the size and form of particles/fibres, density of wood and the degree of
content M(%) as derived by Siau (1984): the compression of the mat. Due to the pore structure, the transverse
permeability is usually very much higher than the longitudinal

152 153
Carvalho, Martins, Costa wood from which the board is made, which seems to indicate that
permeability is mainly determined by the pores between the wood particles
permeability and very much higher than the transverse permeability of the in the mat, than by the particles themselves (Denisov et al., 1975; Bolton
and Humphrey, 1994). In a strand mat there are relatively wide particles, Transport Phenomena
and the gas must move around the particles and not through them. It will
follow the path of least resistance and so in the transverse direction the
OSB strands tend to hinder gas flow because they lie flat in the plane of the to the board plane. In the absence of a more reliable relationship (the
panel, whereas gas flow parallel to the plane can follow the edges of parallel data was based on permeability measurements on extruded
strands rather than having to go around them. Hood (2004) measured the particleboard), other researchers have used this ratio (Carvalho and Costa,
permeability through the mat thickness and in the plane of OSB mats 1998; Carvalho et al., 2003; Nigro and Storti, 2001; Zombori, 2001;
(with no vertical density profile) as a function of compaction ratio and Zombori et al., 2003; Fenton et al., 2003; Lee et al., 2006).
flake thickness. He found that the permeability through the thickness of the Thoemen (2000) and Thoemen and Humphrey (2006) used two
mat was highly dependent on compaction ratio (mat density to wood alternative sets of permeability data obtained by Haselein (1998) and von
density) and to a lesser extent on flake thickness (Hood et al., 2005). A Haas et al. (1998) for MDF mats. Both sets of data give permeability (in
high compaction ratio implies a larger compression strain of the particles m2) as a function of density (in kg/m3) and expressed also by an
and thus smaller and fewer gaps between particles (Kamke, 2004). exponential equation of the form:
Permeability in the plane of the mat decreases also with increasing
compaction ratio, but in a less severe manner than through the mat gz (20)
thickness. Transverse and in-plane permeabilities were higher for mats
comprised of thicker flakes, which is thought to be caused by a reduction The coefficients a, b and c are presented in Table 3.2.
in flow-path tortuosity. Table 3.2. Coefficients a, b and c of equation 20.
The density of wood species can also influence permeability; for the Flow direction a b c
same target panel density, high density wood species lead to high mat
permeabilities than low-density wood species. Several authors presented Haselein (1998) in-plane -0.006 2.95x10-6 -0.199
data for the specific permeability of particleboard in the thickness cross-sectional -0.026 4.98x10-6 -0.074
direction as a function of board density (Bowen, 1970; Sokunbi, 1978; von Haas et al. (1998) in-plane -0.041 9.51x10-6 -0.015
Humphrey, 1982). Carvalho and Costa (1998) fitted an exponential curve -5
cross-sectional -0.037 1.10x10 -0.037
to the data presented by Humphrey and Bolton (1989) for transverse
permeability of particleboard (in m2) as a function of mat density (in
kg/m3): Pereira (2002) and Pereira et al. (2006) used the same expression with the
coefficients of von Haas et al. (1998) to calculate the in-plane and the
–12 –3
(18) cross-sectional permeabilities of a MDF mat.
gz NAT

Suggested R&D topics include:

This expression was also considered by Zombori (2001) and also by
Nigro and Storti (2001). Fenton et al. (2003) used the following  How do the geometry, alignment and size of particles (fines,
expression derived by Marceau (2001) to calculate the cross-sectional strands) affect mat permeabilities?
permeability of an OSB mat:
–15  When wood crosses the glass transition temperature, what changes
in mat permeability will take place?

3.3.5.4 Diffusivity of water vapour

Water vapour diffusion occurs due to a partial pressure in the voids of the
mat. The interdiffusion coefficient of an air-vapour mixture (binary
diffusion coefficient) is calculated by the following semi-empirical
gz OSB (19) equation (Stanish, 1986), in m2/s:
To estimate the longitudinal permeability, Humphrey and Bolton (1989)
established a ratio of 59:1 between the permeabilities parallel and normal

154 155
Carvalho, Martins, Costa
Transport Phenomena
1.75
Zombori (2001) set this attenuation factor (a at a value of 0.5 in both the
 va
–5 (21) vertical and horizontal directions, assuming that the pathway is similar for
steam diffusion horizontally and vertically in the mat structure. Fenton et
with the gas pressure P in Pa and temperature T in K. This equation was al. (2003) and Lee et al. (2006) used the same value.
used by several researchers, who studied the heat and mass transfer in
wood-composite mats (Carvalho and Costa, 1998; Carvalho et al., 2003; The diffusivity of steam decreases during the consolidation of the mat
Nigro and Storti, 2001; Zombori, 2001; Zombori et al., 2003; Fenton et due to an increase in board density. Jensen and Emmler (1996) obtained
al., 2003; Lee et al., 2006). Thoemen and Humphrey (2006) used a the diffusivity of steam in the transversal direction of MDF mats with
similar expression based on data presented by Cussler (1984). different densities, at ambient temperature. Carvalho (1999) fitted this
data and obtained the following equation:
To estimate the effective diffusivity of steam in air within the voids of a
wood composite mat, it is necessary to account for the porous structure –4 –8 –10 2
v c c (25)
and the tortuous path. The following expression was then considered, in
Thoemen and Humphrey (2006) considered that a porous medium offers a
which the diffusivity is reduced by the square of porosity (s) and is
resistance to molecular gas diffusion that can be expressed by an
multiplied by a tortuosity factor ():
obstruction factor kd. They used experimental data obtained in steady-
2 state experiments on MDF samples (with homogeneous density profiles)
eff va (22) fitted by an exponential equation, relating the dimensionless obstruction
factor in the cross-sectional direction with the material density (kg/m3):
Pereira et al. (2006) considered an empirical tortuosity factor equal to 2 –3
in both the vertical and horizontal directions. Hata et al. (1990) used the d (26)
following expression for the effective diffusivity of steam in the thickness
direction for a particleboard:
3.3.5.5 Diffusivity of bound water
eff va (23)
It is believed that bound water flux by diffusion through the particles and
In the horizontal directions, Hata et al. (1990) considered that the fibres is small compared to the movement of water by convection. So,
diffusivity of steam among the particles is equal to that in open air: many researchers have neglected this mechanism (Carvalho et al., 2003;
Dvx = Dvy = Da. Stanish et al. (1986) used an attenuation factor a to account Nigro and Storti, 2001).
for closed pores resulting from the cellular nature of the solid: Bound water diffusivity was determined by Stanish et al. (1986). With
Zombori (2001) set this attenuation value to 0.5 in both the vertical and the lack of more reliable data, several researchers (Zombori, 2001;
horizontal directions, assuming that the pathway is similar for steam Pereira, 2002; Pereira et al., 2006) considered the bound water diffusivity
diffusion horizontally and vertically in the mat structure. Fenton et al. to be constant and equal to the same value given by Stanish et al. (1986),
(2003) and Lee et al. (2006) used the same value. for the temperature and moisture content range expected during hot-
pressing:
The diffusivity of steam decreases during the consolidation of the mat
due to an increase in board density. Jensen and Emmler (1996) obtained b
–13 3
(27)
the diffusivity of steam in the transversal direction of MDF mats with
different densities, at ambient temperature. Carvalho (1999) fitted this
data and obtained the following equation: 3.3.5.6 Thermodynamic relationships
2 The description of heat and mass transfer during drying and hot-pressing
eff va (24) needs some thermodynamic relationships. The most important of these as

156 157
Carvalho, Martins, Costa 6 3
(28)

used by several researchers for modelling these operations, are given in As an alternative, Stanish et al. (1986) fitted a polynomial to steam tables
the following pages. and obtained:
6 2
The heat of vaporisation of water (in J/kg) as a function of temperature (in (29)
K) can be given by the Clausius-Clapeyron equation:
The enthalpy of air, steam and bound water can be calculated using the Transport Phenomena
following expressions (Stanish et al., 1986; Zombori et al., 2003). The
enthalpy of air as a function of temperature is:
depleted. Fenton et al. (2003), Lee et al. (2006) and Zombori et al. (2003)
a P air (30) used this relationship.

where Cp air is air heat capacity (1000 J/(kgK)). The enthalpy of dry wood (in J/kg) can be estimated as (Stanish et al.,
1986):
The enthalpy of water vapour can be calculated as (Stanish et al., 1986;
Zombori et al., 2003): wood (34)

Fenton et al. (2003) and Lee et al. (2006) used an expression by Simpson
and Tenwolde (1999).
2
wood (35)

The differential heat of desorption (in J/kg) for the wood-water system
depends on the moisture content (M) according to the equation of
Bramhall (1979) and this expression was used by Humphrey (1982),
Carvalho and Costa (1998), Carvalho et al. (2003), Pereira et al. (2006):
6
S (36)

Thoemen et al. (2006) used an equation, presented by Humphrey and

Bolton (1989a) which includes both the heat of sorption (equation 36) and
the latent heat of vaporisation (equation 28).
6
2
v dp (31) The saturation vapour pressure can be adequately described by the
dp
where the temperature of the dew point as a function of the vapour partial following expression fitted to the data of Keenan and Keyes (1936):
pressure (Pv in Pa) is described by:
cat (37)
–4 3 (32)
dp v v v The viscosity of water vapour (in Pa s) as a function of temperature (in
K) can be calculated using an expression derived by Humphrey and
The differential enthalpy of bound water at any concentration is equal to Bolton (1989a) who fitted an equation from Sutherland (1893) to
the free water enthalpy minus the differential heat of sorption. For wood, experimental data from Keenan and Keyes (1966):
Stanish et al. (1986) assumed that the differential heat of sorption varies
quadratically with bound water content and at zero bound water content 1.5

is equal to 40 % of the heat of vaporization: –5

v (38)

b This expression was used by several researchers (Pereira et al., 2006;

p pw fcp
(33) Thoemen and Humphrey, 2006). Thoemen and Humphrey (2006) used
b the same approach and derived the viscosity of air:
where Cpw is the specific heat of water (4180 J/(kgK)) and b is the
fsp

1.5
bound water density at FSP. At full saturation, the differential enthalpy of –6
a (39)
bound water reaches a maximum and decreases as the bound water is

158 159
Carvalho, Martins, Costa
Transport Phenomena

Fenton et al. (2003) and Lee et al. (2006) used expressions from the
Perry’s Chemical Engineers handbook. The viscosity of the air-vapour
mixture was obtained from a linear combination of the component
viscosities weighted by the mole fraction in the mixture (Thoemen and
Humphrey, 2006; Fenton et al. 2003; Lee et al., 2006).

3.4 HEAT AND MOISTURE TRANSFER IN CONDITIONING

3.4.1 Introduction

Upon leaving the press, most boards are cooled, generally in a star cooler.
This is the case for WBP bonded with amide-based resins because cooling
minimises hydrolysis of the adhesive, but also to facilitate the subsequent
sanding operation. Cooling is also thought to help balance moisture
content distribution and stresses. After a pre-calibration in sanders, boards Figure 3.12: Temperature variation in a mattress pressed to 20 mm at
are stored as packs for conditioning. Other boards, generally bonded with 200°C (data collected by Mark Irle).
phenolic resins, but also with special formulations of UF resins are hot-
stacked. This operation serves to equalize differences of temperature and Measurement of the internal conditions of boards during conditioning
moisture content, but also to complete the cure of the resin. WBPs like could be useful not only for predicting the final board properties through
MDF, particleboard or OSB leave the hot-press with surface temperatures modelling and simulation, but also to the development of new resin
of about 210 ºC and above. However, resin is not completely cured formulations. However, few studies have focused on this step. Ohlmeyer
especially at the core layers and for PF resins, which need more time to and Kruse (1999) measured the temperature and moisture content in
cure compared with UF or MUF. different layers for normal stacked (initial temperature around 60 ºC)
and hot-stacked UF-panels (initial temperature around 90 ºC) and for hot-
Existing know-how about the effects of this operation is based on 30 year- stacked PF boards (initial temperature around 100 ºC). They observed
old investigations, but nowadays this knowledge is of limited use, because that when storing UF-panels in a stack, the temperature decreases slowly
significant advances have occurred with respect to process technology and and it took up to 10 days to cool the panels in the centre of the stack to
resin systems (Ohlmeyer and Kruse, 1999). Moreover, there is rather little room temperature. The temperature in the outer parts of the stack
fundamental literature research that could be useful to understand the decreased at a much faster rate. When the boards were separated for
mechanisms of heat and mass transfer during the conditioning (hot sanding and re-stacked in smaller stacks, the cooling rate increased. On
stacking and normal stacking). the other hand, they observed that moisture content changed only
marginally when boards are stored in a stack and that the differences in
3.4.2 Internal Conditions during Conditioning moisture content between the layers of hot-stacked panels was lower
when compared with normal stacked panels. During the transport from
The temperature, moisture content and vapour pressure changes
the press outlet via the cross-cut saw to the cooling zone convection due
dynamically during hot pressing, but it continues during the hot-stacking
to the movement causes a surface temperature decrease. When boards are
or normal stacking (after cooling). When leaving the hot press, board
lying in the cooling zone the surface temperature can increase again due
temperature decreases rapidly, especially at the surface layers so that very
to heat flux from the core.
quickly the temperature of the surface layers is lower than the core layers
(see Figure 3.12). In addition, the moisture content of the surface layers is
considerably lower than that of the core layers.

160 161
Carvalho, Martins, Costa 3.4.3 Relevant Heat and Mass Transfer Mechanisms
During hot stacking and normal stacking, heat will be transferred by Transport Phenomena
convection driven by a temperature gradient between and within panels in
the stack and the environment. Inside the stacks, heat will be mainly
transferred by conduction. The use of controlled temperature air could 3.5 DIFFUSION OF CHEMICALS
accelerate heat transfer by convection and thus panel cooling. Handling
of the panels to smaller stacks could result in a faster heat transfer, due to 3.5.1 Introduction
an increase of the ratio surface area to volume of the stack. Initial stack
The diffusion of chemicals, for example adhesives, wax and other
temperature will also affect heat transfer. In hot stacking, heat transfer is
additives during the hot-pressing of WBPs will certainly affect the
faster at surface than at core layers, while in normal stacking, differences
properties of the final products. However, there is a lack of information,
between the layers on temperature drop rate are not significant.
not only in the evaluation on transport properties of these substances but
Mechanisms for mass transfer during hot stacking and normal stacking also about the interaction of these substances with wood during this
include bulk flow of gas (air and water vapour) from the board to the process. The diffusion of these chemicals will depend on their physical
environment and moisture diffusion inside the stack. For hot-stacking, state. In general, resin and wax are added as aqueous solutions, as well as
when the total pressure gradients are significant, convection is still an wood preservatives and fire retardants, but in some cases they can be
important mechanism for moisture movement, while for normal stacking added in powder form. On the other hand, under the conditions they are
when the total pressure gradients have decreased, diffusion will become subjected during drying, blending and hot-pressing, phase changes and
more significant. The migration of water towards the surface will be reactions will produce other products, namely volatile substances such as
mostly due to a chemical potential gradient. So, hot-stacking will result formaldehyde and other volatile organic compounds (VOC).
also in a faster moisture movement inside the panel, due to the decreasing
resistance to diffusion with increasing temperatures and the increased rate 3.5.2 Diffusion of Resins
of water desorption at higher temperatures (Haas and Fruehwald, 1999).
The amount of resin in wood composites is relatively low (usually less
Suggested R&D topics include: than 14 % by oven-dry weight of wood), but the development of bond
strength due to resin cure and interaction with wood is crucial for physical
 Monitoring of mat conditions during hot-stacking and normal
and mechanical performance of the final boards. Ideally, resin cure should
stacking
be sufficient by the end of the press cycle and it may continue during
 More quantitative understanding of responses to these conditions subsequent stages. However before gelation, the mobility of resin might
have importance to its possible redistribution and penetration in wood
 Post-curing of adhesives during hot-stacking cells during the pressing cycle.
 Influence of hot-stacking physical and mechanical properties on the There are some investigations about the distribution of resin in WBPs, but
final boards no fundamental studies exist about their transport processes. Brady and
Kamke (1988) have studied the effect of the hot-pressing parameters on
phenol-formaldehyde resin penetration by using fluorescence
microscopy and a manual digitalisation technique. They concluded that
the uniformity of resin penetration is more influenced by the natural
variability of wood than by pressing conditions (temperature, moisture
content, time, pressure). However, these conditions influence resin
penetration by controlling the viscosity of resin: pressure is the driving
force for bulk flow; moisture content and temperature have direct effect
on resin viscosity. Gindl et al. (2002) studied the diffusion of MUF resin
in cell walls of spruce wood using UV-microscope. Thumm and Grigsby
(2002) achieved the simultaneous visualisation of wax and resin UF in
MDF boards using a labelling technique for wax and resin and a confocal
laser scanning microscope. They provide some clues about the
movement of resin into the fibre wall. Xing et al. (2005) used the same

162 163
Carvalho, Martins, Costa resin is very easy under high moisture and temperature conditions and it is
at a maximum during the second drying stage. The migration of phenol-
technique with a toluidine blue O staining system to study the penetration formaldehyde resin was also studied by DMA (dynamic mechanical
of UF resin into MDF fibres and they concluded that the penetration of analysis) (Laborie et al., 2002). Xu (2009) investigated resin penetration into
wood particles under press conditions using FPA-FTIR spectroscopy at Transport Phenomena
microscopic scale. She found that PMDI penetrates much deeper into the
cell structure than UF.
capillary forces. Capillary pressure increases with decreasing capillary
The movement of the resin inside the board during the hot pressing radius and decreases with increasing liquid contact angle. So, the sizing
operation is linked to resin properties such as resin “flow” and viscosity, agents will decrease the surface wetability, thus decreasing the capillary
but also with mat properties such as permeability, density and moisture forces.
content. Resin “flow” is the ability of a resin to remain fluid under heat
and pressure, wetting new surfaces and accommodating particle During the hot-pressing and stacking, the flow of wax can occur if the
movement before being immobilized by loss of solvent and/or polymer temperature is above the melting point of the wax for a sufficiently long
growth (Maloney, 1989). The resin curing should normally take place time. In general, longer press times and hot-stacking helps the wax
after the platens have achieved target thickness. However, in some boards efficiency. However, it is also necessary that the wax particles adhere to
resin pre-cure can occur at board surfaces leading to low density surfaces. wood. For this, it is necessary that the water present in the emulsion is
An optimum value of resin flow permits maximum board consolidation absorbed by wood, leaving the wax surrounded by the surfactant at the
and internal bond strength, with a minimum loss of surface hardness due surface. Secondly, it is necessary that the surfactant releases wax particles
to resin pre-cure. On the other hand, a high resin flow can cause loss of to permit that the wax adheres to wood. This mechanism will depend
board strength due to the excessive resin penetration into furnish. Resin upon the more or less complete destruction of the surfactant in the hot
viscosity is crucial in the application process (spraying and blending), but press cycle (Maloney, 1989). Thumm and Grigsby (2002) observed that if
will also influence the permeability. Some adhesives are sensitive to the the wax is applied first and then the resin in MDF fibres, the wax will
relationship between the liquid carrier content and the cross-linking reduce the movement of resin into the fibre wall if it forms a layer over
reaction: if the carrier is removed too rapidly the adhesive molecules will part of fibre surface.
not have the mobility necessary for optimum cure. If the carrier is An optimal wax emulsion should minimize water absorption and
removed too slowly full cure may not be achieved. So permeability will thickness swelling, but also hydrocarbon emissions without affecting
affect the transport of resin carrier and thus the bond quality. other properties as internal bond (Eckert and Edwardson, 1998). So, the
Suggested R&D topics include: evaporation of hydrocarbons from wax will occur during the hot-pressing
and could lead to fires above the press due to accumulation of
 How much resin penetrates in wood during the blending operation hydrocarbons in the ventilation system. The diffusion of these chemicals
or during hot-pressing? should be similar to other volatile compounds from wood (terpenes,
alcohols) or from resins (formaldehyde).
 How does resin mobility affect curing and bond strength
development? The mass transport of volatile compounds occurs either by internal
diffusion within the gas phase (air and water vapour), driven by a
3.5.3 Diffusion of Wax and Other Additives concentration gradient, by convection (bulk flow of gas) inside the mat
and by convection across the surface boundary layer, driven by a
Wax only affects the absorption of liquid water and not of water vapour. difference between the concentration of the component at the surface and
The initial penetration of liquids in board is through the voids between the its concentration on bulk air. These mechanisms involve transport
particles or fibres, which can be considered as irregularly shaped properties, such as the diffusivity of a given component, e.g
capillaries. So, the liquid water will enter the board mainly due to formaldehyde, into other component (as air), permeability and convective
mass transfer coefficients. Mass transport coefficients from the mat to the
surroundings will depend on local air flow parameters (velocity, flow
regime and temperature). In near steady state conditions, e.g. board
conditioning, it would be expected that the controlling mechanisms are
the internal diffusion and the transport of vapour through the board-air
inter-phase. However, during hot-pressing, first the evaporation and then
the bulk flow (the water vapour will transport the volatile compounds)
will be the controlling mechanisms. These volatile compounds obviously

164 165
Carvalho, Martins, Costa transfer to the core, increasing the mat temperature (Wang and Gardner,
1999). The mat temperature will promote the formation and the diffusion of
will contribute to an increase in gas pressure inside the mat. Mat these compounds. It was observed that the total VOCs and formaldehyde
conditions will affect their transport. Moisture is an agent that will extract emissions increase with platen temperature (Makowski and Ohlmeyer,
and transport the volatile compounds and also contribute to a faster heat 2006).
Other additives that are used include preservatives and fire retardants. Transport Phenomena
The diffusion of these chemicals will depend on their characteristics,
mostly if they are aqueous solutions. In general, preservatives are not
used widely, an if they are then only for exterior boards. The recent 3.6 REFERENCES
systems are aqueous solutions and are mixed with resin, which permits an
even distribution of the preservatives over the particles. Aguilar, H. (2006) Study of the rheology of deformable porous media:
application to paper production. PhD Thesis in Chemical Engineering,
Fire retardants are used in some boards. In general, they are added to the Faculty of Engineering from University of Porto, Portugal.
furnish, in the blender, in powered or liquid form in small quantities.
They can be mixed with the liquid resin. Some, such as the borates, have Alexopoulos, J. (1999) OSB Pressing optimisation and problem solving
high diffusivities. The salts are diffused by concentration gradients into using monitoring of temperature and gas pressure. In: Proc. of the third
the interior of the particles. The blender conditions but also the pressing European panel Products Symposium. Llandudno, Wales, pp.190-208.
conditions can affect the diffusion of these chemicals, in particular in Ben Nasrallah, S., Perré, P. (1988) Detailed study of a model of heat and
systems, which contain water-soluble monomers that polymerize to an mass transfer during convective drying of porous media. Int. J. of Heat
insoluble state at elevated temperature. Geimer et al. (1994) investigated and Mass Transfer, 31(5), 957-967.
the injection of three different gases into wood composites: a fire
retardant (methyl borate), a resin catalyst (methyl formate) and an Bird, R.B., Stewart, W.E., Lightfoot, E.N. (1966) Transport Phenomena.
accelerator for cement-bonded panels (carbon dioxide). John Wiley and Sons, New York.
Suggested R&D topics include: Bloch, J.F. (1992) Tranfers de Masse et de Chaleur dans les Milieux
Poreux deformables Non Saturés: application au Pressage du Papier.
 Contribution of volatile compounds (e.g. formaldehyde) to the Thèse Doctorat, Institut National Polytechnique de Grenoble, Grenoble,
increase in total gas pressure inside the mat relatively to water France.
vaporization
Bolton, A.J. and Humphrey, P.E. (1988) The Hot Pressing of Dry-formed
 Is the diffusion of wax really important in affecting water sorption Wood-based Composites. Part I. A Review of the Literature, Identifying
and resin penetration? the Primary Physical Processes. Holzforschung, 42(6), 403-406.
Bolton, A.J., Humphrey P.E. and Kavvouras, P.K. (1989a) The Hot
Pressing of Dry-formed Wood-based Composites. Part III. Predicted
Vapour Pressure and Temperature Variation with Time, Compared with
Experimental Data for Laboratory Boards. Holzforschung, 43(4), 265-
274.
Bolton, A.J., Humphrey P.E. and Kavvouras, P.K. (1989b) The Hot
Pressing of Dry-formed Wood-based Composites. Part IV. Predicted
Variations of Mattress Moisture Content with Time. Holzforschung,
43(5), 345-349.
Bolton, A.J., Humphrey P.E. and Kavvouras, P.K. (1989c) The Hot
Pressing of Dry-formed Wood-based Composites. Part VI. The
Importance of Stresses in the Pressed Mattress and their Relevance to the
Minimisation of Pressing Time, and the Variability of Board Properties.
Holzforschung, 43(6), 406-410.
Bolton, A.J., and Humphrey, P.E. (1994) The permeability of WBPs. Part
1. A Review of the Literature and some unpublished work. Holzforshung,
48 (Suppl), 95-100.

166 167
Carvalho, Martins, Costa Brady, D.A., and Kamke F.A. (1988) Effects of hot-pressing parameters on
resin penetration. For. Prod. J., 38(11/12), 63-68.
Bowen, M.E. (1970) Heat transfer in particleboard during hot-pressing. Bramhall, G. (1976) Fick’s law and bound-water diffusion. Wood Sci. and
PhD dissertation, Colorado State University, Corvallis, USA.
Technol., 8(3), 153-161. Transport Phenomena
Bramhall, G. (1979) Mathematical model for lumber drying I-Principles
involved. Wood Sci. and Technol., 12(1), 14-21 Denisov, O.B., Anisov, P., Zuban P.E. (1975) Untersuchung der
Permeabilität von Spanvliesen. Holztechnologie, 16(1), 10-14.
Carvalho L. (1999) Estudo da Operação de prensagem do Aglomerado de
Fibras de Média Densidade: Prensa descontínua de pratos quentes (Study Dorri, B., Emery A.F., Malte P.C. (1985) Drying rate of wood particles
of the hot-pressing of Medium Density Fiberboard (MDF): batch platen with longitudinal mass transfer. J. of Heat Tranfer, 107, 12.
press). PhD Thesis in Chemical Engineering, Faculty of Engineering of
University of Porto, Portugal. Droin-Josserand, A., Taverdet, J.L., Vergnaud, J.M. (1989) Modeling the
process of moisture absorption in three dimensions by wood samples of
Carvalho, L.H., Costa, M.R.N. and Costa, C.A.V. (2003) A Global Model various shapes: cubic, parallelepipedic. Wood Sci. and Technol., 23(3),
for the Hot-Pressing of Medium Density Fiberboard (MDF). Wood Sci. 259-271.
and Technol., 37, 241-258.
Fenton, T., Budman, H., Pritzker, M., Bernard, E., Broderick, G. (2003)
Carvalho, L.M., Costa, C.A.V. (1998) Modeling and simulation of the hot- Modeling of Oriented Strand Board Pressing. Ind. Eng. Chem. Res., 42,
pressing process in the production of medium density fiberboard (MDF). 5229-5238.
Chem. Eng. Comm. 170, 1-21.
Fhyr, C., Rasmuson (1997) Some aspects of the modelling of wood chips
Comstock, G.L. (1970) Direction permeability of softwoods. Wood Fibre, drying in superheated steam. Int. J. Heat Mass Transfer, 40(12), 2825-
1, 283-289 (1970) 2842.
Constant, T., Moyne, C., Perré, P. (1996) Drying with internal heat Fhyr, C., Rasmuson (1996) Mathematical model of steam drying of wood
generation: theoretical aspects and application to microwave heating. chips and other hygroscopic porous media. AIChE J., 42(9), 2497-2502.
AIChE J., 42(2), 359-368.
Fick, A (1833) On diffusion. Annalen der Physic, Ser 2, 94, 59-85.
Cussler, E.L. (1984) Diffusion. Mass transfer in fluid systems. Cambridge
Univ. Press, UK. Fohr, J.P., (1995) Vacuum drying of oak wood. Drying Technology,
13(8&9), 1675-1693.
Dai, C., Yu, C. (2004) Heat and mass transfer in wood composite panels
durting hot-pressing: Part I. A physical-mathematical model. Wood and García, P.J. (2002) Three-dimensional heat and mass transfer during
Fibre Sci., 36(4), 585-597. oriented strand board hot-pressing, PhD Thesis, Faculty of Forestry, The
Univ. of British Columbia, Vancouver, British Columbia, Canada.
Darcy, H.P.G. (1856) Les Fontaines Publiques de la Ville de Dijon,
Victor Dalmont, Paris. García, P.J., Avramidis, S., Lam, F. (2001) Internal temperature and
pressure response to flake alignment during ho-pressing. Holz Roh-
Deng, J., Xie, Y., Feng M. (2006) An experimental study of microwave Werkst, 59, 272-275.
pre-heating of an MDF fibre mat: Moisture and temperature distribution
and the impact on hot-pressing. For. Prod. J., 56(6), 76-81. García, P.J., Avramidis, S., Lam, F. (2003) Horizontal gas pressure and
temperature distribution responses to OSB flake alignment during hot-
pressing. Holz Roh- Werkst, 61, 425-431.
Geimer, G.L., Leão, A., Armebruster, D., Pablo, A. (1994) Property
Enhancement of wood composites using gas injection. In; Proceedings of
the 28th International Particleboard/Composite Materials Symposium,
Washington State Univ., Pullman, WA, USA, pp.243-259
Gindl, W., Dessipri, E., Wimmer, R. (2002) Using UV-Microscopy to
Study the diffusion of Melamine-Urea-Formaldehyde Resin in Cell Walls
of Spruce Wood. Holzforshung, 56(1), 103-107

168 169
Carvalho, Martins, Costa Haselein, CR, (1998) Numerical Simulation of pressing wood-fibre
composites. Ph.D. Thesis, Oregon State Univ., U.S.A.
Gregg, S.J., Sing, J.S. (1991) Adsorption, Surface Area and Porosity, von Haas, G., Steffen A. and Fruehwald, A. (1998) Untersuchunger zur
Academic Press, London.
Permeabilität von Faser-, Span- und OSB-Matten für Gase. Holz Roh- Transport Phenomena
Werkst. 56(6), 386-392.
von Haas, G., and Fruehwald, A. (1998) Einfluß der Temperatur auf die Jensen, U., Emmler, R. (1996) Diffusion resistance of MDF and
Geschwindigkeit der sorptiven Wasseraufnahme von Furnieren. Holz particleboard. In: Proceedings of the 1996 International Conference on
Roh- Werkst, 57, 77-78. Wood Mechanics, FMPA, Stuttgart, Germany, pp. 345.
Hata T., Kawai, S., Sasaki, H. (1990) Computer simulation of temperature Johansson, A., Fhyr, C., Rasmuson, A. (1997a) High temperature
behavior in particle mat during hotpressing and steam injection pressing. convective drying of wood chips with air and superheated steam. Int. J.
Wood Sci. and Technol., 24, 65 Heat Mass Transfer, 40(12), 2843-2858.
Hood, J., (2004) Changes in Oriented Strandboard Permeability during hot- Johansson, A., Fhyr, C., Rasmuson, A. (1997b) Influence of the drying
pressing. MSc Thesis, Faculty of Virginia Polytechnic Institute and State medium on high temperature convective drying of single wood chips. Int.
University, Blacksburg, Virginia, U.S.A. J. Heat Mass Transfer, 15(6-8), 1801-1813.
Hood, J.P., Kamke, F.A., Fuller, J. (2005) Permeability of oriented Kamke F.A., Wolcott, M.P. (1991) Fundamentals of flakeboard
strandboard mats. For. Prod. J., 55(12), 194-199. manufacture: wood-moisture relationships. Wood Sci. and Technol., 25,
57.
Humphrey PE, (1991) Pressing issues in panel manufacture: Internal
behavior during pressing and its impact on time minimization, properties Kamke, F.A., Wilson, J.B. (1986a) Computer simulation of a rotary dryer.
and profit. In; Proceedings of the International Particleboard Symposium, Part I: Retention time. AIChE J., 32(2), 263-268.
Pullman, WA, USA, pp.99-108.
Kamke, F.A., Wilson, J.B. (1986b) Computer simulation of a rotary
Humphrey, P.E. (1982) Physical aspects of wood particleboard dryer. Part II: Heat and Mass transfer. AIChE J., 32(2), 269-275.
manufacture. Ph.D. Thesis, Univ. of Wales, U.K.
Kamke, F.A., Casey, L.J. (1988a) Fundamentals of flakeboard
Humphrey, P.E. and A.J. Bolton (1989a) The Hot Pressing of Dry-formed manufacture: internal-mat conditions. For. Prod. J., 38(6), 38- 44.
Wood-based Composites. Part II: A Simulation Model for Heat and
Moisture Transfer, and Typical Results. Holzforschung, 43(3), 199-206. Kamke, F.A. and Casey, L.J. (1988b) Gas pressure and temperature in the
mat during flakeboard manufacture. For. Prod. J., 38(3) 41-43.
Humphrey, P.E. and A.J. Bolton (1989b) The Hot Pressing of Dry-formed
Wood-based Composites. Part 5 :The Effect of Board Size: Comparability Kamke, F.A., (2004) Physics of Hot-pressing. In: Fundamentals of
of Laboratory and Industrial Pressing. Holzforschung, 43(6), 401-405. Composite Processing: Proceedings of a Workshop, pp. 3-18, USDA
Forest Service, General Technical Report FPL-GRT-149, Madison WI,
Hunter, A. J. (1993) On movement of water through wood-The diffusion U.S.A.
coefficient", Wood Sci. and Technol., 27, 401-408.
Kamke, F.A., Watson, L.T., Lee, J-N, Shu, J. (2006) A Web-based
Incropera F.P., Witt, D.P. (1990) Fundamentals of heat and mass transfer, problem solving environment for the wood-based composites industry.
John Wiley & Sons, New York. For. Prod. J., 56(9), 26-32.
Kavvouras, P.K. (1977) Fundamental process variable in particleboard
manufacture, Ph.D. Thesis, Univ. of Wales, U.K.
Keenan, J.H., Keyes, F.G. (1936) Thermodynamic properties of steam.
John Wiley and Sons, London.
Keey, R., Langrish, T., Walker, J.C. (1999) Kiln-Drying of Lumber.
Springer Series in Wood Sicence, Springer Verlag, Berlin.

170 171
Carvalho, Martins, Costa Kollmann, F., Malmquis, L., (1956) Über die Wärmeleizahl von Holz und
Holzwerkstoff. Holz Roh-Werkst. 14(16), 201-204.
Kocaefe, Younsi, R., Chandry, B., Kocaefe, Y. (2006) Modeling of heat Krischer, O., Kroll, K., (1956) Die wissenschaftlichen Grunlagen der
and Mass transfer during high temperature treatment of aspen. Wood Sci. Trocknungstechnik. Springer Verlag, Berlin.
and Technol., 40, 371-391.
Kuhlmann, G. (1962) Untersuchung der technischen Eigenschaften von Transport Phenomena
Holz und Spanplatten in Abhängigkeit von Feuchtigkeit und Temperatur
im hygroskopishen Bereich. Holz Roh-Werkst. 20(7), 259-270.
Meyer, N., Thoemen, H. (2007) Gas pressure measurements during
Laborie, M.P., Salmén, L., Frazier, C.E. (2002) The cooperativity analysis continuous hot pressing of particleboard, Holz Roh Wekst., 65(1), 49-55.
of segmental motion in wood adhesive interphase: A probe of nanascale
morphologie. In: Proceedings, 6th Pacific- rim Bio-based composites Moyne, C. (1983) Séchage. In: Le matériau bois: Propriétés, Tecnhologie,
Symposium & Workshop on the chemical modification of cellulosics, Mise en oeuvre, A.R.B.O.L.O.R., Nancy.
Portland, Oregon, pp. 18-25. Nigro N., Storti, M. (2001) Hot-pressing process modelling for medium
Lee, C., Budman, H. Pritzker, M. (2006) Simulation and Optimisation of density fibreboard (MDF). Int. J. Math and Math Sci., 26(12), 713-7129.
the Continuous Oriented Strand Board Pressing Process. Ind. Eng. Chem. Nilsson, L., Wilhelmsson, B., Stenström, S. (1993) The diffusion of water
Res., 45(6), 1974-1988. vapour through pulp and paper. Drying Technology, 11(6), 1205-1225.
Lee, J., Kamke, F., Watson, L. (2006) Simulation of the hot-pressing of a Ntalos, G., Grigoriou, (2001) Particleboards. In: State of the Art of Report
multi-layered wood strand composite. J. of Comp. Mat. published on line. of COST E13 (Wood Adhesion and Glued Products), Johansson, Pizzi,
Liu, J., Avramidis, S., Ellis, S. (1994) Simulation of Heat and Moisture Van Leemput eds.
Transfer in Wood During Drying under Constant Ambient Conditions. Ohlmeyer, M., Kruse K. (1999) Hot Stacking and its effect on panel
Holzforschung, 48(3), 236-240. properties, In: Proceedings of the 3rd European Panel Symposium, pp.,
Luikov, A.V. (1966) Heat and mass transfer in capillary-porous bodies. Llandudno, Wales.
Pergamon Press, London. Pang, S. (2001) Improving MDF fibre drying operation by application of
Maku, T., Sasaki, R, Hamada, H. (1959) Studies on the particle board. a mathematical model. Drying Technology, 13(5-7), 1395-1409.
Repor IV. Temperature and moisture distributions in particle boardduring Pang, S. and R.B. Keey (1995) Drying Kinetics of Pinus Radiata Boards
hot-pressing. Wood Res. 21, 34-46. at Elevated Temperatures, Drying Technology, 19(8), 1789-1085.
Makowski, M., Ohlmeyer, M. (2006) Influences of hot pressing Pereira, C. (2002) Estudos da Operação de Prensagem e Cura do
temperature and surface structure on VOC emissions from OSB made of Aglomerado de Fibras de Media Densidade (Studies of the Hot-pressing
Scots pine. Holzforshung, 60, 533-538 and Curing of Medium Density Fiberboard. PhD Thesis, Faculdade de
Maloney, T. (1989) Modern Particleboard and Dry-Process Fiberboard Engenharia da Universidade do Porto, Porto, Portugal.
Manufacturing. 3rd Edition, Miller Freeman Publications, San Francisco. Pereira, C., Carvalho, L.H., Costa, C.A.V. (2004) High Frequency
Marceau, P. (2001) Rapport préliminaire: permeabilité au gaz des Heating of Medium Density Fiberboard (MDF): Theory and Experiment.
panneaux, unpublished report. Chem. Eng. Sci., 59(4), 735-745.
Pereira, C., Carvalho, L.H., Costa, C.A.V. (2006) Modeling the
continuous hot-pressing of MDF. Wood Sci. and Technol., 40, 308-326.
Perré, P., Degiovani, A. (1990) Simulation par volumes finis des
transferts couplés en milieu poreux anisotropes. J. of Heat and Mass
Transfer, 33(11), 2463-2478.
Perry’s Chemical Engineers’Handbook, 7th ed.; Perry, R.H., Green,
D.W., Maloney, J.O. eds.; Mc Graw Hill, New York, 1984, pp.2-320-2-
322.

172 173
Carvalho, Martins, Costa Reis, P. (2009) Estudos da Operação de Prensagem e Cura do Aglomerado
de Fibras de Media Densidade (Studies of the Hot-pressing and Curing of
Plumb, O.A., Spolek, G.A., Olmstead, B.A. (1985) Heat and mass Medium Density Fiberboard. PhD Thesis, Faculdade de Engenharia da
transfer in wood during drying. Int. J. of Heat and Mass Transfer, 28(9), Universidade do Porto, Porto, Portugal.
1669. Shu, J., Watson, L.T., Zombori, B.G. (2006) WBCSim: an environment for
modelling WBPs manufacture. Engineering with Computers, 21, 259- 271. Transport Phenomena
Shusheng, P., Keey, R.B. (1995) Modelling the temperature profiles
within boards during the high-temperature drying of Pinus radiata timber: Second International Symposium on Drying, Mac Graw Hill University,
the influence of airflow reversals. Int. J. Heat and Mass Transfer, 38(2), pp. 84-92.
189-205.
Stamm, A.J. (1959) Bound-water diffusion into wood in the fiber
Siau, J.F. (1984) Transport process in wood, Springer-Verlag, Berlin. direction. For. Prod. J., 9(27), 27-32.
Siau, J.F. (1995) Wood. Influence of moisture on physical properties. Stamm, A.J. (1964) Wood and Cellulose Science. Ronald Press, New
Virginia Polytechnic Institute and State University, NY. York.
Simpson, W. T. (1993) Determination and use of moisture diffusion Stanish, M. A., Schajer G. S., Kayihan, F. (1986) A mathematical model
coefficient to characterize drying of northern Red oak (Quercus rubra). of drying for hygroscopic porous media, AIChE J., 32(8), 1301.
Wood Sci. and Technol., 27, 409-420.
Steffen A, Haas, G., Rapp, A., Humphrey, P., Thoemen, H. (1999)
Simpson, W., TenWolde, A. (1999) Physical Properties and Moisture Temperature and Gas Pressure in MDF-mats during Industrial Continuous
Relations of Wood. In: Wood Handbook-Wood as an Engineering Hot-Pressing. Holz Roh- Werkst., 57, 154-155.
material, USDA Forest Service, Forest Products Laboratory, Madison,
WI, USA. Strickler, M.D. (1959) The effect of press cycles and moisture content on
properties of Douglas fir flakeboard, For. Prod. J., 9(7), 203-207.
Simpson, W.T. (1983-84a) Drying wood: a review. I. Drying Technology,
2(2), 235-264. Sushland, O. (1967) Behaviour of particleboard mat during the press
cycle. For. Prod. J., 17(2), 51-57.
Simpson, W.T. (1983-84b) Drying wood: a review. II. Drying
Technology, 2(3), 353-368. Sutherland, W. (1893) The viscosity of gases and molecular force. The
London, Edinburgh, and Dublin Philosophical Mag. Journ. Sci, 36 (fifth
Skaar, C. (1972) Water in wood. Syracuse University Press, Syracuse, series)(223), 507-531.
New York.
Thoemen H., Humprey P.E. (2003) Modeling the continuous pressing
Smith, T.M. (1994) Heat transfer dynamics. Tappi J., 77(8), 239-245. process for wood-based composites. Wood and Fibre Sci., 35, 456-468.
Sokunbi, O.K. (1978) Aspects of particleboard permeability. M.Sc. Thoemen, H., Humprey, P.E. (2006) Modeling the physical processes
Thesis, Univ. of Wales, U.K. relevant during hot pressing of wood-based composites. Part I. Heat and
Mass transfer. Holz Roh- Werkst., 35, 456-468.
Spolek, G.A., Plumb, O.A. (1964) A numerical model of heat and mass
transfer in wood during drying, Drying’80, vol 2, In: Proceedings of the Thoemen, H. (2000) Modeling the Physical Processes in Natural Fibre
Composites During Batch and Continuous Pressing. Ph.D. Thesis, Oregon
State Univ., USA.
Thoemen, H., Humprey, P. (2001) Hot pressing of wood-based
composites: selected aspects of the physics investigated by means of a
simulation. In Proceedings of The Fifth European Panel Products
Symposium, Llandudno, Wales, UK, pp. 18-30.
Thoemen, H., Humprey, P.E. (1999) The Continuous pressing process for
WBPs: an analytical model. In: Proceedings of The Third European Panel
Products Symposium, Llandudno, Wales, UK, pp. 18-30.
Thumm, A., W. Grigsby (2002) Interaction of wax and UF resin in MDF:
quantitative analysis of the relationships between wax and resin on MDF

174 175
Carvalho, Martins, Costa Torgovnikov, G.I. (1993) Dielectric Properties of Wood and Wood-Based
Materials. Springer-Verlag, Germany.
fibres. In: Proc. of the Sixth European Panel Products Symposium, Wang, W. and Gardner, D. (1999) Investigation of volatile organic
Llandudno , Wales, UK, pp. 51-59.
compound press emissions during particleboard production. For. Prod. J., Imaging Techniques
49(3), 65-72.
Ward, R.J., Skaar, C. (1963) Specific heat and conductivity of Chapter 4
particleboard as a function of temperature. For. Prod. J., 13(1), 32.
Whitaker, S. (1977) Simultaneous heat, mass and momentum transfer in Advanced Imaging Techniques in Wood-Based
porous media: a theory of drying. Advances in Heat Transfer, 13, 119- Panels Research
203.
Winistorfer, P.M. (2000) Fundamentals of vertical density profile Lech Muszynski and Maximilien E. Launey
formation in wood composites. Part I. In-situ density measurement of the
consolidation process. Wood and Fibre Sci., 32(2), 209-219. CHAPTER SUMMARY
Wolcott, M.P., F.A. Kamke, D.A. Dillard (1990) Fundamentals of Although numerical modelling seems to be a promising tool in virtual
flakeboard manufacture: viscoelastic behavior of the wood component. prototyping of new composite materials, the prevailing empirical practice
Wood and Fibre Sci., 22(4), 345-361. does not adequately address the complexity of bio-based materials such
as wood composites, and the outputs are generally incompatible with the
Xing, C., Riedl, B., Cloutier, A., Shaler, S. (2005) Characterization of
requirements of contemporary modelling tools. The innovative material
urea-formaldehyde resin penetration into medium density fibarboad
characterization techniques based on advanced imaging technologies and
fibres. Wood Sci. and Technol., 39, 374-384.
inverse problem methodology seem particularly suitable for complex
Xu, Yan (2009) Investigations on adhesive bonding of wood and particle heterogeneous composites.
boards using UF resins and PMDI (in german). PhD Thesis, Technical
Advanced image analysis techniques provide new means for quantitative
University, Faculty of Life Sciences, Braunschweig, Germany
characterization of wood-based composite materials. Particularly non-
Yrjölä, J., Saatamoinen, J.J. (2002) Modelling and practical operation destructive methods based on computed tomography yield an
results of a dryer for wood chips. Drying technology, 20(6), 1077-1099. unprecedented insight in the morphology and micromechanics of
particulate composites, which enables a new comprehensive approach to
Zombori B. (2001) Modeling the transient effects during the Hot-Pressing experimental determination of material characteristics. Combination of
of Wood-based Composites. PhD Thesis, Faculty of the Virginia full-field measurements with inverse problem methodology brings the
Polytechnic Institute and State University, Blacksburg, Virginia, USA. material testing on new level of efficiency by removing many limitations
Zombori B, Kamke, F.A., Watson, L.T. (2002) Simulation of the internal of the traditional test methods. Such integrated approach to material
conditions during the hot-pressing process. Wood and Fibre Sci., 35 (1), 2- characterization supports development of morphology-based numerical
23. modelling for rapid prototyping of new enhanced materials and
manufacturing processes.
Zombori B, Kamke, F.A., Watson, L.T. (2002) Sensitivity analysis of
internal mat environment during hot-pressing. Wood and Fibre Sci., 35
4.1 INTRODUCTION
(1), 2-23.
Wood-based composites, and prominently panel products are examples
of complex, anisotropic, and heterogeneous materials, which allow
various levels of flexibility in engineering their properties to the
requirements of the final use. A key to securing the leading position in the
market is not only developing stronger and more durable materials,
characterized by mechanical performance carefully tailored to the
requirements of the end use, but ability to bring breakthrough innovations
e.g. in the technology, which would allow widening sustainable raw
material base. In developing enhanced manufacturing processes, new

176 177
Muszynski, Launey performance. However, despite long history of research in this area
relatively little is known on the actual micro-mechanical interaction
adhesive formulations and bonding techniques the focus is on the between the adhesive and wood within the bond interphase. While
internal bond, providing composite integrity and mechanical contemporary adhesive formulators are capable of engineering the
interaction between the adhesive and constitutive wood polymers on the Imaging Techniques
molecular level, the micro-mechanics of the bond on the scale of wood
anatomical features is rarely considered as a diagnostic tool or even
investigated. Across the industry, bond performance is assessed through a common denominator is the format of the output: features visible and
number of standardized tests on relatively bulk product specimens. The hidden for human eye are represented as visual maps of light intensity
principle qualifying criterion may in most cases be reduced to the general and/or colour for the researchers to behold and analyze.
requirement that the bond be stronger than the wood substrate. Only few For many years, increasingly advanced imaging techniques have been
of these tests return true qualitative data and none addresses the actual used as a primary or supporting reference tool providing invaluable
micromechanics of the interphase. Significant progress in this area insight in the internal structure of biological systems and complex
requires a more holistic approach, and is hard to imagine this progress materials. With time, cameras fitted on light or fluorescent microscopes
occurring without better understanding of the composite performance and became a commonplace. Electron scanning microscopes and CT scanning
internal bond durability on the micro-mechanical level, as well as devices are becoming more accessible, less expensive and more common,
reliable modelling based on that understanding (Wolcott and Muszynski too. In most cases, however the actual use of images generated by those
2008). In this respect, wood-based composites may become a model for techniques had not extended beyond visual inspection and qualitative
a larger family of other bio-particulate composites. However, despite assessment of features within the field of view by an expert eye. This
substantial research effort in material characterization of wood-based approach is quite effective in fields of research where conclusions can be
composites, reliable modelling poses a significant challenge. The drawn based on comparison of certain visible aspects or characteristics in
principle obstacles are the inherent complexity of the raw material the analyzed images to a standard image of the same object, or where
(various grades of wood: veneer sheets, strands, particles, fibres) and of existence (or not) of such characteristics can inform the decision making
the composite interaction in the internal bond on the micro-mechanical process. This is for instance the case when finished surfaces are visually
level, as well as that the body of quantitative knowledge generated in the compared to a standard, or when a presence of visible checks, knots and
field is hardly compatible with the required inputs of available modelling repairs is a base of disqualification of plywood sheets from decorative
tools. grades. Visual inspection is however much less effective when
The complexity of the internal structure and the compatibility of the quantification of the findings (counting multiple objects, measurements
experimental output with the requirements of the contemporary modelling etc.) is of interest, like for instance when the level of discoloration or the
techniques can be effectively addressed by application of advanced size of knots and checks is to be measured, and their number within the
imaging techniques coupled with numerical modelling of the composite field of view counted.
structure. The value of the output generated by imaging techniques goes far
“Imaging techniques” is a collective name used to refer to all kind of beyond what is available for visual inspection. Information coded in
experimental techniques recording data in form of two- or three- digital images can and is being harnessed more effectively by application
dimensional images. In most general sense, this category includes of advanced, quantitative imaging techniques coupled with tools for
micrographs from light, electron or atomic force microscopes, automatic image processing and analysis. These techniques seem to be
radiographs, images acquired from thermographic cameras, particularly suitable to address the complexity of the internal structure of
ultrasonographs, as well as three-dimensional data acquired via medical wood-based products. They also have a potential for bridging the gap
and industrial computed tomography (CT) based on x-ray or gamma between experimentation and modelling in the area of wood-based panel
radiation, nuclear magnetic resonance and other techniques. Their products.
In this chapter, opportunities and principal challenges to bridging
experimentation and modelling with advanced imaging techniques in
the area of wood-based composites research are discussed. Several of
promising material characterization techniques based on advanced
imaging technologies and inverse problem methodology, which seem
particularly suitable for providing necessary input data for numerical
models are also presented.

178 179
Muszynski, Launey Numerical models are important tools for discovery. Scientific models are
being created to predict behaviour of physical objects and phenomena of
4.2 VIRTUAL PROTOTYPING: OPPORTUNITIES AND various level of complexity. Numerical models are used for prototyping and
CHALLENGES virtual testing of hypotheses and are particularly desirable whenever physical
tests are too complex, impractical, or too expensive. Accurate models
capable of correlating our understanding of how various processing regimes
and treatments affect morphology and micro- mechanics of the Imaging Techniques
composites with their bulk properties and service performance would be
critical for prototyping and developing new, advanced bio-based
materials and products. Such models would be also important tools for Another level of complexity comes from the new arrangement of these
improving properties of those already present on the market. particles in the composite, which although not entirely random, is
nevertheless far from deterministic (Zhou et al. 2008). The rearranged
It is not surprising then, that modelling many aspects of physical and microstructure will also include additional voids in inter-particle spaces,
mechanical nature of wood-based materials has been in focus of quite cured adhesive or, in case of wood plastic composites, a continuous
vigorous research activity. In his bibliography of finite element (FEA) polymer matrix. In the most general case the resulting size, shape,
modelling in wood research Mackerle (2005) lists more than 260 papers orientation and distribution of the component elements (strands, particles,
published just in one decade between 1995 and 2004. fibres, adhesive and voids) are of interest for predicting the bulk
properties of the composites. All of the morphological characteristics
However, actual implementation of virtual prototyping based on material
listed above are best described in statistical terms, and the description
modelling to development of new and improved wood-based panels still
must be based on large number of measurements. Because of the
seems to be a concept from the future. Two principal obstacles seem to be
metamorphosis the particles undergo in the consolidation process, in situ
the inherent complexity of the internal structure of wood-based
characterization would be preferred as input for reliable numerical
composites and the incompatibility between the outputs of contemporary
models.
testing approaches commonly used in wood composites research and the
input requirements of most numerical models. These models must also address the internal bonding, which provides
integrity to the reconstituted composites and facilitates load transfer
4.2.1 Levels of Complexity in the Internal Microstructure of the between the components. In most wood-based composites, the bonding
Composites is provided by polymer adhesives, which do not form a distinct
continuous matrix assumed by most composite theories. In addition,
Modelling of complex heterogeneous particulate bio-based composites is models developed for synthetic and mineral composites assume that the
not a trivial task. In the case of wood-based composites, with one bond is formed on a well defined interface between the impermeable
exception for plywood, the naturally variable wood structure is reduced to components. This is certainly not the case with wood-based composites.
small irregular units (wafers, strands, particles, fibres), henceforth Instead, microscopic evidence shows a substantial inter-phase where the
addressed collectively as particles, characterized by wide distribution of adhesive penetrates the porous cell structure on the particle perimeter
geometries and sizes. The original cellular structure of wood is disrupted (Marra 1992; Kamke and Lee 2007). Although there are many
and modified to various degrees during the comminution process, and publications reporting tests on bonds between isolated strands (e.g. Smith
further during the consolidation, which particularly for panel products 2005), fibres and properties of fibre-polymer interfaces (Mott et al. 1996;
includes hot-pressing under dynamic temperature and changing moisture Shaler et al. 1997; Egan and Shaler 2000; Tze et al. 2003), the
content conditions. contribution of the interphase to the micro-mechanical performance of
the internal bond, and the performance of the composite as a whole are
not completely understood.
Such complexity may be best addressed through multi-scale modelling,
which refers to correlation of phenomena and properties obseRVEd on
various levels of material organization (load transfer through the
interphase, contributions of individual particles, and properties of bulk
composite).

4.2.2 Compatibility between Testing Approaches and

Numerical Models

The predictive power of numerical models depends as much on sound

constitutive theory as on reliability of the input data: morphology,

180 181
Muszynski, Launey prevailing approach in material characterization in the area of wood
composites is that of simplified comparative tests responding to the
boundary conditions, and bulk characteristics of the modelled object industry’s need for inexpensive tools to assess their quality against an
acquired through measurement and empirical tests. In the same time, the accepted standard. In most cases however, this approach does not produce
characteristics compatible with the inputs of modelling software based on Imaging Techniques
the principles of theoretical mechanics. The traditional testing and
measurement methods, where the bulk mechanical characteristics are
determined from relatively simple analytical solutions derived for small 4.3.1 Digital Image Analysis
deformations in idealized homogeneous and isotropic solids, are generally High resolution digital microscopy, scanning electron microscopy (SEM),
not adequate for the level of complexity found in particulate bio- computed tomography (CT) based on x-ray, gamma and neutron
composites. Collecting viable input data for modelling of such radiation (Wellington and Vinegar 1987; ASTM 2000; Macedo et al.
particulate composites requires better understanding and cooperation 2002; Richards et al. 2004), nuclear magnetic resonance (NMR)
between modellers and experimentalists and may require revision of relaxometry, and other advanced instruments (Park et al. 2003; Pétraud et
traditional testing practices and approach to material testing. al. 2003) return digital images carrying a wealth of spatial information
coded in discrete color or greyscale values. Virtually all have been
4.3 THE ADVANCED IMAGING TECHNIQUES successfully used for visualization of the internal structure of solid wood
(e.g. Mannes et al. 2010), and various morphological aspects in wood-
The complexity of the internal structure and the compatibility of the based composites. These values representing method-specific quantities
experimental output with the requirements of the contemporary modelling (e.g. real colours, densities, x-ray attenuation, temperatures) assigned to
techniques can be effectively addressed by application of advanced millions of pixels or voxels arranged in two- or three-dimensional arrays.
imaging techniques and coupling them with three-dimensional The data may be manipulated and effectively analyzed by employment of
numerical modelling of the composite structure. By contrast to the robust algorithms for automated image enhancement, pattern
traditional methods, recent progress in modern high-resolution non- recognition, and quantitative characterization of spatial distribution of
contact imaging and full-field measurement techniques makes it possible various physical features: e.g. components, phases, inhomogeneities and
to explore enhanced approaches to experimental procedures. Although void spaces resulting from various processing and loading regimes.
many of the imaging techniques discussed below are by no means new,
the amount and value of the information carried by images is often 4.3.1.1 Image enhancement
underestimated.
Presence of noise in raw images obtained from various imaging
There are generally two ways to approach the geometrical aspects of the techniques blurs the boundaries between regions and objects and may
composite microstructure in morphology-based modelling. One, often interfere with automatic processing. Therefore, prior to analysis digital
applied in finite element method (FEM), is to generate the element mesh images are often subjected to image enhancing procedures like noise
based on generalized knowledge of the micro-structure of the removal, edge and contrast enhancement. Despite the conceptual
composite, where various morphological characteristics are either created similarity of image enhancement to one-dimensional signal filtering not
explicitly using idealized average parameters, or generated using all one-dimensional signal processing techniques do naturally extend to
stochastic methods in order to reflect the variability of the real material in higher dimensions (Mallat 1989; Coifman and Donoho 1995; Candes and
a more realistic manner. The other approach, often used in the material Donoho 1999). Among the alternative methods the most prominent is
point method (to be discussed later in this chapter) is to use a digitized nonlinear anisotropic diffusion algorithm first proposed by Perona and
structure of an actual material sample as captured with one of the Malik (Perona and Malik 1990; Catte et al. 1992; Weickert 1999), which
imaging techniques. Note, that also in the first approach a reliable allows effective noise removal in two- and three-dimensional image data
knowledge of the variability of the morphological structure is assumed. while preserving edges between regions of consistently different
intensities. A coherence-enhanced anisotropic diffusion particularly
suitable for relatively low resolution CT reconstructions was proposed by
Frangakis and Hergel (Frangakis and Hegerl 2001; Sheppard et al. 2004).
Comprehensive lists of image enhancement techniques are available in
the literature (Gonzalez and Wintz 1987; Pratt 1991).

182 183
Muszynski, Launey
Imaging Techniques

4.3.1.2 Image segmentation

lumens appear penetrated by the matrix material, is clearly visible in the
Image segmentation usually refers to processing of images with the goal original colour micrograph, but only the two-stage colours segmentation
of automatic detection of objects or phases presented in the image. The technique allowed automatic recognition of the zone and calculation of
difficulty of this task depends largely on the number and uniformity of relative areas of clear wood, matrix, and the interphase visible as the red
objects or phases being analyzed as well as on the definition of their zone in the bottom right picture (Wang 2007).
boundaries. While phase segmentation, or division of the image into Similar approaches may be employed for segmentation of X-ray
regions of common characteristics can be achieved using relatively computed tomography (CT) data generated through computational
elementary operations like colour separation or thresholding the greyscale reconstruction of internal feature of heterogeneous samples from a series
intensity (e.g. Sahoo et al. 1988; Chen et al. 2010), definition and of X-ray projections of the samples recorded at different angles. In the
characterization of individual particles within the image may pose a following example, a small sample of OSB shown in Figure 4.2a was
serious challenge. This is certainly true in case of particulate wood-based scanned at resolution of ~6 µm/voxel. Images in Figures 4.2b and c show
composites (OSB, particleboards, fibreboards), where irregular particles the central vertical and horizontal sections of the 3D data. The histogram
are tightly packed together with multiple contact points. represents the intensity distribution in the specimen.

c)

a) b)

Figure 4.1: Colour segmentation of a microscopic image of wood particles

embedded in PVC matrix reveals extensive inter-phase (red
zone in the bottom right picture) where cell walls and lumens
appear penetrated by the matrix material (Wang 2007).
d)
For instance, in a recent study the density distribution in commercial
panel products is evaluated in terms of greyscale intensity frequency Figure 4.2: XCT scans of an OSB sample (a) at resolution of ~6 µm/voxel:
distributions in combined radiographs of the panels (Chen et al. 2010). A central vertical section (b), and central horizontal section (c).
series of images presented in Figure 4.1 illustrate how phase The histogram (d) represents the intensity distribution in the
segmentation based on RGB colour aspects of a micrograph generated in specimen.
a fluorescent microscope is used to separate and quantify extensive inter-
Three-dimensional renderings of a particle board specimen in Figure 4.3
phase between wood flour particles and polymer matrix in a wood-
demonstrate the comparison of the initial volume (a) with three
polymer composite sample. The interphase, where cell walls and

184 185
Muszynski, Launey
consecutive stages of phase segmentation based on thresholding of the Imaging Techniques
volume intensity range (b-d).
growing (Adams and Bischof 1994) while other use fuzzy set theoretic
approaches (Bezdeck 1981; Bezdeck 1992). Most of these techniques are
not suitable for noisy data, which stresses the importance of effective
image enhancement procedures. More computationally involved
procedures are needed to work with noisy data (e.g. Hansen and Elliott
1982; Hertz et al. 1991; Kosko 1991).
However, there is no single method which can be considered good for all
images, nor are all methods equally good for a particular type of image.
Most of these sophisticated methods have been developed in the medical
a) b)
field because of the simplicity of patterns studied and the thorough
knowledge of their characteristics (Ozkan et al. 1993; Chen et al. 1998;
Pardo et al. 2001; Hibbard 2004; Zhang et al. 2006). In heterogeneous
particulate composite materials, the geometry of the particulate
component is more complex and irregular; therefore the segmentation is
more complicated.
Currently many commonly accessible image-processing software bundles
offer various machine vision tools for 2-dimensional images, including
efficient edge detection and particle analysis algorithms capable of
isolating and quantitatively describing multiple particles. Standard
c) d) particle analysis returns distribution of sizes, areas and orientation of
irregular particles.
Figure 4.3: Three-dimensional renderings of a particle board specimen:
the initial volume (a) is compared with three consecutive Numerous generalized algorithms have also been proposed for the
stages of phase segmentation based on thresholding of the analysis of 3-dimensional data (Price 1995; Garboczi 2002; Thompson et
volume intensity range (b-d). al. 2006). An example in Figure 4.4 demonstrates two stages of particle
analysis of a closely packed particulate composite performed on a
While phase segmentation may be very helpful in assessing relative binarized X-ray CT image of a sample reported by Thompson et al.
volumes of voids, woody particles and adhesive in the composite, it (2006). The segmentation allows automated statistical analysis of the
provides only visual clues as to the number, shape and position of the axial dimensions, volumes, and orientation of all the particles within the
individual particles. These characteristics can be quantified using a group analyzed sample. Recently a similar approach was attempted to segment,
of procedures collectively referred to as particle analysis. visualize and analyze random cellulosic fibrous networks based on X-ray
A great variety of particle analysis algorithms have been developed in CT data (Faessel et al. 2005; Walther and Thoemen 2009).
the last few decades and this number continually increases each year.
Several review papers for segmentation techniques have been presented
in the literature (Fu and Mui 1981; Haralick and Shapiro 1985; Borisenko
et al. 1987; Pal and Pal 1993; Freixenet et al. 2002). Some of these
techniques use spatial details such as boundary (Davis 1975) and region
(Zucker 1976) based methods, some use hybrid techniques such as
watershed transform (Vincent and Soille 1991) and seeded region

186 187
Muszynski, Launey
 Imaging Techniques

particularly carefully checked in case of the wood-based composites,

a) where the scale of the component particles may differ by levels of
magnitude (for instance compare strands, particles and fibre bundles used
in panel products).
Effective properties are evaluated by methods of edge detection or
averaging of microscopic variables like stresses and strains over a
representative volume element (RVE) of the microstructure, which is
an important parameter on the micromechanical analysis of composite
materials. The concept of RVE was introduced by Hill (Hill 1963) as a
micro-structural sub-region that is representative of the entire
microstructure in an average sense. Hashin and Shtrikman (Hashin and
b) c) Shtrikman 1962, 1963) extended the concept and introduced
representative volume as a reference cube that is small compared to the
Figure 4.4: Segmented and binarized X-ray tomography images of (a) a entire body, for which the volume average of variables like strain, stress,
particulate material and visualization of the segmented
or phase volume fraction are the same as those for the whole body. The
particle-scale reconstruction of the volume (Thompson et al.
consequences for wood-based panel products are that the smallest
2006); (b-c) a random cellulosic fibrous network (Walther
and Thoemen 2009). volume to be meaningfully analyzed for morphological characteristics or
modelled has to include a good representation of the component elements,
that is: strands, particles or fibres. Obviously, RVEs for OSB,
4.3.1.3 Importance of the representative volume particleboards and fibreboards will be very different. It will also depend
on the specific morphological features being characterized. For instance,
Usually, mechanical testing is done on macroscopic samples with the size of representative volumes appropriate for characteristics
volumes extremely large when compared to order of magnitude of known describing local density variation, shape and size distributions of voids or
features of the composite micro-structure. Since modelling of such a cured adhesive spots in OSB, may be levels of magnitude smaller than
huge volume cannot be practically accomplished on contemporary these required to characterize the in-situ strand geometry.
computers, it is generally assumed that the ensemble average of a given
physical property obtained on smaller volumes is equal to its volume 4.3.2 Multi-Scale and Multi-Modal Correlation
average in the infinite-volume limit. However, the assumption makes
sense only if the medium is statistically homogeneous, i.e., the statistical One of the common problems in morphological analysis and
properties of all regions of space are similar. This assumption should be characterization of wood-based composites is the trade off between the
resolution and the field of view. The resolution of digital images is related
to the fixed number of elementary sensors (pixels) in the detector array,
which transforms the physical signal into the digital image. Consequently,
increasing the spatial resolution of the images is achieved by decreasing
the size of the field of view. Problems arise when morphological features
of different scales are of interest at the same time, as in the case of strand
geometry and adhesive spot distribution in OSB panels mentioned above.
The geometry, distribution and orientation of the strands have to be
considered at the centimetre scale, but then sub-millimetre or micron
scale features of the adhesive spots cannot be resolved. In such cases
image representations of the same sample at two different resolution
levels have to be generated and the position of the high resolution detail
should be registered with the related region in the low resolution image.

188 189
Muszynski, Launey Similarly, different imaging techniques (or modalities) may be used in
order to visualize different aspects of the same region of interest. This is a
In fact, this registration process can span several steps over multiple commonplace in fluorescent microscopy, where micrographs generated with
levels of magnitude (centi- to nano- meters). different light spectra are correlated with each other, but is also performed
between micrographs generated using imaging instruments based on Imaging Techniques
entirely different physical principles and returning images of different
scales. Multi-modal correlation may for instance allow precise
registration of images generated using light-, fluorescence-, electron- and 4.3.3 Optical Measurement of Deformations and Strains
atomic force microscopy, as well as two-dimensional micrographs of Digital images and volumetric data are also used for non-contact full-
physically exposed sections with three-dimensional CT data of the same field measurement of deformations and strains in heterogeneous
sample acquired before the cut up. Figure 4.5 illustrates an attempt at a anisotropic solids under various loading regimes with sub-pixel accuracy
manual registration of a detail recorded in fluorescent microscope on a by means of digital speckle photogrammetry (DSP) based on the digital
physical section of a WPC sample in a virtual section of three- image correlation (DIC) algorithms (Ranson et al. 1987; Sutton and
dimensional data for the same sample before cutting (Rahmati and Chao 1988; Bruck et al. 1989; Vendroux and Knauss 1998). DSP allows
Muszynski 2009). determination of displacements of a dense mesh of selected points on
surfaces of deformed specimens by comparing successive images
acquired during tests and cross correlating the gray intensity patterns of
the direct neighbourhood of the points (or the reference areas). DSP has
been already successfully applied to determine strains in specimens of
solid wood subjected to external loads and climate changes (Muszynski et
al. 2006), individual wood fibres and paper (Sutton and Chao 1988; Choi
et al. 1991; Mott et al. 1996), fibre reinforced plastics (FRP) (Russell and
Sutton 1989; Muszynski et al. 2000), concrete (Choi and Shah 1997), and
adhesive films (Muszynski et al. 2002).
Figure 4.5: A CT section (3.0 mm x 2.1 mm) of wood plastic composite (a) In late 1990s, a similar algorithm has been developed for volumetric data
compared with a fluorescence micrograph (b). Manual dubbed digital volume correlation (DVC: Bay et al. 1999; Smith et al.
extraction of subsamples and registration of the images was a 2002). DVC allows calculation of 3D internal strain fields in the analyzed
serious challenge and could not be accomplished with the volume by comparing data from x-ray tomographic scans of the same
desired precision (Rahmati and Muszynski 2009). specimen acquired in unloaded and loaded states. This method was found
Novel techniques for numerical multi-scale and Multi-modal to be very accurate in mapping the strain intensities in porous media such
correlation of image data, collectively referred to as correlative as bone tissue polymer and aluminium foams (Bay et al. 1999; Smith et
microscopy, are developed mainly in the field of biomedical research al. 2002; Sutton 2004), and recently similar approach was demonstrated
(e.g. Vicidomini et al. 2010). Correlative microscopy is considered a with wood sample by Forsberg et al. (Forsberg et al. 2008; Danvind et al.
new frontier of the advanced image analysis and each year brings another 2009).
wave of interesting techniques with multiple potential applications in
wood-based panel research. 4.3.4 Inverse Problem Approach

Two- and three- dimensional full-field methods provide displacement and

strain data equivalent to hundreds of strain gages and are capable of
capturing localized deformation gradients and strain concentrations that
could not be possibly detected through discrete measurements from
traditional instrumentation, such as extensometers, LVDTs, displacement
gages or strain gages. In addition, the output format is readily compatible
with many numerical modelling packages based on FEM, so that the
displacement and strain fields measured by means of the full-field
methods may be compared with the results of theoretical simulations of
the same test configurations using existing models. In this approach,
known as the inverse problem methodology, the measured and
theoretical strain fields are used as input data in order to determine

190 191
Muszynski, Launey Pierron and Grediac 2000; Lecompte et al. 2005). The general idea of this
method is shown on the schematic diagram in Figure 4.6.
localized material properties even for heterogeneous anisotropic materials
A. Experiment & imaging C. Morph. Modeling
and for specimens of complex geometries (Grediac and Pierron 1998;
 Imaging Techniques
Material
property input
statistical redundancy for even very complex constitutive models. It
follows quite naturally, that the existing models are perfectly suited tools
to assist in design and development of such complex test configurations,
while on the other hand the results of the tests may provide validation to
the models’ assumptions.

4.3.5 New Developments in Modeling

New developments may be also expected in the area of modelling. Some

time ago, a material point method (MPM) was proposed by Sulsky
(1994) as an alternative to FEM, which is particularly suited for
modelling heterogeneous solids based on their morphology. In this
method, digital images of heterogeneous surfaces may be used directly as
digitized input of the material geometry removing rather difficult task of
B. Experimental cyy Adjust generating complex FEM meshes. The MPM method is as suitable for
(DIC) D. cyy from model
property
application with the inverse problem methodology as FEM, and was
successfully used for modelling wood at various levels of material
organization (Nairn 2003; Nairn and Guo 2005; Nairn 2007). An MPM
algorithm is available in public-domain as a 3D parallel code (Parker
2002).
Sensitivity
analysis

Difference (A)

No Yes
Stop A

Prop.

Figure 4.6: Flow chart of the material parameter identification problem

by coupling full-field techniques in material testing, FEM
modelling and inverse problem methodology (Muszynski and
Nairn 2008).

In fact, the combination of full-field measurements with inverse

problem methodology brings the material testing on a whole different
level of efficiency by removing one of the limitations of the traditional
test methods, which is the requirement that the mechanical material tests
are reduced to simplest load cases. By combining the full-field
measurements with inverse problem methodology and careful design
of specimen geometries, loading, and boundary conditions, it is possible
to determine all involved terms of anisotropic compliance matrices in a
single test (Lecompte et al. 2005). The great number of virtual
measurement points returned by full-field methods provides enough

192 193
Muszynski, Launey
Adams, R. and L. Bischof (1994). Seeded Region Growing. Ieee
4.4 REFERENCES Transactions on Pattern Analysis and Machine Intelligence, 16(6): 641- 647.
ASTM (2000). ASTM E 1441-00: Standard Guide for Computed Imaging Techniques
tomography (CT) Imaging. in: AMERICAN SOCIETY FOR TESTING
AND MATERIALS. West Conshohocken, PA.
Choi, D., J. L. Thorpe and R. B. Hanna (1991). Image analysis to measure
Bay, B. K., T. S. Smith, D. P. Fyhrie and M. Saad (1999). Digital volume strain in wood and paper. Wood Science and Technology, 25(2): 251-262.
correlation: Three-dimensional Strain Mapping Using X-ray
Tomography. Experimental Mechanics, 30(3): 217-226. Choi, S. and S. P. Shah (1997). Measurement of Deformations on
Concrete Subjected to Compression Using Image Correlation.
Bezdeck, J. C. (1981). Pattern Recognition with Fuzzy Objective Experimental Mechanics, 37(3): 307-313.
Function Algorithms Springer, New-York.
Coifman, R. R. and D. L. Donoho (1995). Lecture Notes in Statistics. in:
Bezdeck, J. C. P., S.K. (1992). Fuzzy Models for Pattern Recognition: G. O. Anestis Antoniadis, Ed. Lecture Notes in Statistics. New-York,
Methods That Search for Structures in Data IEEE Press, New-York. Springer. 103: 225.
Borisenko, V. I., A. A. Zlatopolskii and I. B. Muchnik (1987). Image Danvind, J., F. Forsberg, R. Mooser, M. Arnold, E. Hack and P. Wyss
segmentation (State-of-the-Art SuRVEy). Automation and Remote (2009). An example on acquiring wood ultra-structure, density and strain
Control, 48(7): 837-879. in 3D using non-destructive Synchotron Radiation Micro Computed
tomography and Digital volume correlation. in: Proceedings of COST
Bruck, H. A., S. R. McNeill, M. A. Sutton and W. H. I. Peters (1989). Action FP0802 Workshop: “Experimental and computational methods in
Digital image correlation Using Newton-Raphson Method of Partial wood micromechanics”, Vienna, Austria, May 11-13, 2009. 123.
Differential Correction. Experimental Mechanics, 28(3): 261-267.
Davis, L. S. (1975). A SuRVEy of Edge detection Techniques. Computer
Candes, E. J. and D. L. Donoho (1999). Ridgelets: a key to higher- graphics and image processing, 4: 248-270.
dimensional intermittency? Philosophical Transactions of the Royal
Society of London Series a-Mathematical Physical and Engineering Egan, A. and S. M. Shaler (2000). Fracture and Mechanics of Fracture for
Sciences, 357(1760): 2495-2509. Resin Coated Single Wood Fibres. Fourth Panel Products Symposium.
Llandudno, Wales, UK: 95-103.
Catte, F., P. L. Lions, J. M. Morel and T. Coll (1992). Image Selective
Smoothing and Edge-Detection by Nonlinear Diffusion. Siam Journal on Faessel, M., C. Delisée, F. Bos and P. Castéra (2005). 3D Modelling of
Numerical Analysis, 29(1): 182-193. random cellulosic fibrous networks based on X-ray tomography and
image analysis. Composites Science and Technology, 65(13): 1931-1940.
Chen, C. W., J. B. Luo and K. J. Parker (1998). Image segmentation via
adaptive K-mean clustering and knowledge-based morphological Forsberg, F., R. Mooser, M. Arnold, E. Hack and P. Wyss (2008). 3D
operations with biomedical applications. Ieee Transactions on Image micro-scale deformations of wood in bending: Synchrotron radiation
Processing, 7(12): 1673-1683. [mu]CT data analyzed with digital volume correlation. Journal of
Structural Biology, 164(3): 255-262.
Chen, S., X. Liu, L. Fang and R. Wellwood (2010). Digital X-ray analysis
of density distribution characteristics of wood-based panels. Wood Frangakis, A. S. and R. Hegerl (2001). Noise reduction in electron
Science and Technology, 44(1): 85-93. tomographic reconstructions using nonlinear anisotropic diffusion.
Journal of Structural Biology, 135(3): 239-250.
Freixenet, J., X. Munoz, D. Raba, J. Marti and X. Cufi (2002). Yet
another suRVEy on image segmentation: Region and boundary
information integration. in: Computer Vision - Eccv 2002 Pt Iii. 2352: 408-
422.
Fu, K. S. and J. K. Mui (1981). A SuRVEy on Image segmentation.
Pattern Recognition, 13(1): 3-16.

194 195
Muszynski, Launey Application to aggregates used in concrete. Cement and Concrete Research,
32(10): 1621-1638.
Garboczi, E. J. (2002). Three-dimensional mathematical analysis of Gonzalez, R. C. and P. Wintz (1987). Digital Image Processing. Addison-
particle shape using X-ray tomography and spherical harmonics: Wesley, Reading, MA.
Grediac, M. and F. Pierron (1998). T-shaped specimen for the direct Imaging Techniques
characterization of orthotropic materials. International Journal for
Numerical Methods in Engineering, 41(2): 293-309.
Measurements. SEM Annual Conference & Exposition on Experimental
Hansen, F. R. and H. Elliott (1982). Image segmentation Using Simple and Applied Mechanics. Portland, OR: paper #: 119/045.
Markov Field Models. Computer Graphics and Image Processing, 20(2):
101-132. Macedo, A., C. M. P. Vaz, J. C. D. Pereira, J. M. Naime, P. E. Cruvinel
and S. Crestana (2002). Wood Density Determination by X- and Gamma-
Haralick, R. M. and L. G. Shapiro (1985). Image segmentation Ray Tomography. Holzforschung, 56: 535-540.
Techniques. Proceedings of the Society of Photo-Optical Instrumentation
Engineers, 548: 2-9. Mackerle, J. (2005). Finite element analyses in wood research: a
bibliography. Wood Science and Technology, 39: 579-600.
Hashin, Z. and S. Shtrikman (1962). On Some Variational Principles in
Anisotropic and Nonhomogeneous Elasticity. Journal of the Mechanics Mallat, S. G. (1989). A Theory for Multiresolution Signal Decomposition
and Physics of Solids, 10(4): 335-342. - the Wavelet Representation. Ieee Transactions on Pattern Analysis and
Machine Intelligence, 11(7): 674-693.
Hashin, Z. and S. Shtrikman (1963). A Variational Approach to the
Theory of the Elastic Behaviour of Multiphase Materials. Journal of the Mannes, D., F. Marone, E. Lehmann, M. Stampanoni and P. Niemz
Mechanics and Physics of Solids, 11(2): 127-140. (2010). Application areas of synchrotron radiation tomographic
microscopy for wood research. Wood Science and Technology, 44(1): 67-
Hertz, J., A. Krogh and R. G. Palmer (1991). Introduction to the Theory 84.
of Neural Computation Addison-Wesley, New York.
Marra, A. A. (1992). Technology of wood bonding : principles in
Hibbard, L. S. (2004). Region segmentation using information divergence practice. Van Nostrand Reinhold, New York. 454 p.
measures. Medical Image Analysis, 8(3): 233-244.
Mott, L., S. M. Shaler and L. H. Groom (1996). A novel technique to
Hill, R. (1963). Elastic Properties of Reinforced Solids - Some measure strain distributions in single wood fibers. Wood and Fiber
Theoretical Principles. Journal of the Mechanics and Physics of Solids, Science, 28(4): 429-437.
11(5): 357-372.
Muszynski, L., R. Lagana and S. M. Shaler (2006). Hygro-mechanical
Kamke, F. A. and J. N. Lee (2007). Adhesive penetration in wood - A Behavior of Red Spruce in Tension Parallel to the Grain. Wood and Fiber
review. Wood and Fiber Science, 39(2): 205-220. Sience, 38(1): 155-165.
Kosko, B. (1991). Neural Networks and Fuzzy Systems: A Dynamical Muszynski, L., R. Lopez-Anido and S. M. Shaler (2000). Image
Systems Approach to Machine Intelligence. Prentice Hall, Englewood Correlation Analysis Applied to Measurement of Shear Strains in
Cliffs, NJ. Laminated Composites. in: Proceedings of the SEM IX International
Congress on Experimental Mechanics, Orando, FL, June, 5-7, 2000. 163-
Lecompte, D., H. Sol, J. Vantomme and A. M. Habraken (2005). 166.
Identification of Elastic Orthotropic Material Parameters Based on ESPI
Muszynski, L. and J. Nairn (2008). Coupling Advanced imaging Analysis
and Morphology Based Modeling for Integrated Characterization of Micro-
mechanics of Wood And Wood-based Composites. International
Conference PHOTOMECHANICS 2008. Loughborough, UK (poster).
Muszynski, L., F. H. Wang and S. M. Shaler (2002). Short-term creep
tests on phenol-resorcinol-formaldehyde (PRF) resin undergoing moisture
content changes. Wood and Fiber Science, 34(4): 612-624.
Nairn, J. A. (2003). Material point method calculations with explicit
cracks. Computer Modeling in Engineering & Sciences, 4: 649–664.

196 197
Muszynski, Launey 61: 406-413.
Nairn, J. A. and Y. Guo (2005). Material point method Calculations with
Nairn, J. A. (2007). A Numerical Study of the Transverse Modulus of Explicit Cracks, Fracture Parameters, and Crack Propagation. in:
Wood as a Function of Grain Orientation and Properties. Holzforschung,
Proceedings of 11th Int. Conf. Fracture, Turin, Italy, Mar 20-25. Imaging Techniques
Ozkan, M., B. M. Dawant and R. J. Maciunas (1993). Neural-Network-
Based Segmentation of Multimodal Medical Images - a Comparative and Price, K. (1995). Annotated Computer Vision Bibliography. Institute for
Prospective-Study. Ieee Transactions on Medical Imaging, 12(3): 534- Robotic and Intelligent System School of Engineering (IRIS), University
544. of Southern California. http://iris.usc.edu/Vision-
Notes/bibliography/contents.html. Last update: April 2, 2009. Access
Pal, N. R. and S. K. Pal (1993). A Review on Image segmentation date: April 4, 2009.
Techniques. Pattern Recognition, 26(9): 1277-1294.
Rahmati, H. and L. Muszynski (2009). Correlating fluorescent
Pardo, X. M., M. J. Carreira, A. Mosquera and D. Cabello (2001). A micrographs with volumetric x-ray CT scans of wood plastic composite
snake for CT image segmentation integrating region and edge (WPC) samples. Unpublished data. Corvallis, OR, Oregon State
information. Image and Vision Computing, 19(7): 461-475. University.
Park, B.-D., S. G. Wi, K. H. Lee, A. P. Singh, T.-H. Yoon and Y. S. Kim Ranson, W. F., M. A. Sutton and W. H. Peters (1987). Holographic and
(2003). Characterization of RIce Husk Using Microscopic and Micro- Specle Interferometry. in: A. S. Kobayashi, Ed. SEM Handbook of
Analytical Techniques for Rice Husk/Thermoplastic Composites. the 7th Experimental Mechanics. Englewood Cliffs, New Jersey, Prentice-Hall,
International Conference on Woodfiber-Plastic Composites. Madison, Inc.: 388-429.
WI: 323-331.
Richards, W. J., M. R. Gibbons and K. C. Shields (2004). Neutron
Parker, S. G. (2002). A Component-based Architecture for Parallel Multi- tomography developments and applications. Applied Radiation and
Physics PDE Simulation. International Conference on Computational Isotopes, 61: 551-559.
Science (ICCS2002) Workshop on PDE Software. P. M. A. Sloot, C. J. K.
Tan, J. J. Dongarra and A. G. Hoekstra. Amsterdam, The Netherlands Russell, S. S. and M. A. Sutton (1989). Strain Field Analysis Acquired
Springer-Verlag, Berlin. 2331: 719−734. Through Correlation of X-Ray Radiographs of a Fiber Reinforced
Composite Laminate. Experimental Mechanics, 29: 237-240.
Perona, P. and J. Malik (1990). Scale-Space and Edge-Detection Using
Anisotropic Diffusion. Ieee Transactions on Pattern Analysis and Sahoo, P. K., S. Soltani, A. K. C. Wong and Y. C. Chen (1988). A
Machine Intelligence, 12(7): 629-639. SuRVEy of Thresholding Techniques. Computer Vision Graphics and
Image Processing, 41(2): 233-260.
Pétraud, M., N. Labbé, J.-C. Lartigue, B. De Jéso and G. Sébe (2003).
Nuclear Magnetic Resonance Relaxometry Applied to Wood Fibers. in: Shaler, S. M., A. Egan, L. Mott and E. N. Landis (1997). Fracture and
Proceedings of 7th International Conference on Woodfiber-Plastic Micromechanics of Resinated Fibers. in: Proceedings of 4th International
Composites, Madison, WI. 358. Conference on Woodfiber-Plastic Composites, Madison, WI. 32-39.
Pierron, F. and M. Grediac (2000). Identification of the through-thickness Sheppard, A. P., R. M. Sok and H. Averdunk (2004). Techniques for
moduli of thick composites from whole-field measurements using the image enhancement and segmentation of tomographic images of porous
Iosipescu fixture: theory and simulations. Composites - Part A: Applied materials. Physica a-Statistical Mechanics and Its Applications, 339(1-2):
Science and Manufacturing, 31(4): 309-318. 145-151.
Pratt, K. (1991). Digital Image Processing. Wiley, New-York. Smith, G. (2005). The lap-shear strength of bonds between Oriented
Strand Board (OSB) like strands coated with pMDI resin. European
Journal of Wood and Wood Products, 63(4): 311-312.
Smith, T. S., B. K. Bay and M. M. Rashid (2002). Digital volume
correlation Including Rotational Degrees of Freedom during
Minimization. Experimental Mechanics, 42(3): 272-278.

198 199
Muszynski, Launey rigid cellular polyurethane foam under uni-axial compression. Department
of Mechanical Engineering, Oregon State University. Corvallis, OR. 7pp.
Sulsky, D., Z. Chen and H. L. Schreyer (1994). A particle method for Sutton, M. A. and Y. J. Chao (1988). Measurement of strains in a paper
history-dependent materials. Comput. Methods Appl. Mech. Engrg., 118: tensile specimen using computer vision and digital image correlation. TAPPI
179–186. Journal, 71(3): 173-175.
Sutton, M. (2004). An attempt to track and quantify micro-damage within
Thompson, K. E., C. S. Willson and W. L. Zhang (2006). Quantitative Imaging Techniques
computer reconstruction of particulate materials from microtomography
images. Powder Technology, 163(3): 169-182.
Wellington, S. L. and H. J. Vinegar (1987). X-ray Computerized
Tze, W. T. Y., S. C. O'Neill, D. J. Gardner, C. P. Tripp and S. M. Shaler Tomography. Journal of Petroleum Technology, August: 885-898.
(2003). Coupling Polystyrene and Cellulose Fibers with Hydrophylic and
Hydrophobic Silanes: Effects on Interfacial Properties. in: Proceedings of Wolcott, M. and L. Muszynski (2008). Position paper: “Materials and
the 7th International Conference on Woodfiber-Plastic Composites, Wood-based Composites” in: “Wood Engineering Challenges in the New
Madison, WI, May 19-20, 2003. 29-35. Millennium - Critical Research Needs”. Materials of Pre-Conference
Workshop for ASCE Structures Congress. Vancouver, BC, Canada 41-50.
Vendroux, G. and W. G. Knauss (1998). Submicron Deformation Field
Measurements: Part 2. Improved Digital image correlation. Experimental Zhang, L., G. Wang and W. Wang (2006). A new fuzzy ART neural
Mechanics, 38(2): 86-92. network based on dual competition and resonance technique. in:
Advances in Neural Networks - Isnn 2006, Pt 1. 3971: 792-797.
Vicidomini, G., M. C. Gagliani, K. Cortese, J. Krieger, P. Buescher, P.
Bianchini, P. Boccacci, C. Tacchetti and A. Diaspro (2010). A novel Zhou, C., C. Dai and G. D. Smith (2008). A generalized mat
approach for correlative light electron microscopy analysis. Microscopy consolidation model for wood composites. Holzforschung, 62(2): 201-
Research and Technique, 73(3): 215-224. 208.

Vincent, L. and P. Soille (1991). Watersheds in Digital Spaces - an Zucker, S. W. (1976). Region Growing: Childhood and Adolescence.
Efficient Algorithm Based on Immersion Simulations. Ieee Transactions Computer graphics and image processing, 5: 382-399.
on Pattern Analysis and Machine Intelligence, 13(6): 583-598.
Walther, T. and H. Thoemen (2009). Synchrotron X-ray
microtomography and 3D image analysis of medium density fiberboard
(MDF). Holzforschung, 63(5): 581-587.
Wang, Y. (2007). Morphological Characterization of Wood plastic
Composite (WPC) with Advanced imaging Tools: Developing
Methodologies for Reliable Phase and Internal Damage Characterization.
PhD Thesis. Oregon State University. Corvallis, OR. 123 pp.
Weickert, J. (1999). Coherence-enhancing diffusion filtering.
International Journal of Computer Vision, 31(2-3): 111-127.

200 201
 Bond Strength Development

Chapter 5

Adhesive Bond Strength Development

Milan Sernek and Manfred Dunky

CHAPTER SUMMARY

The bond strength of thermosetting adhesives develops during the

hardening or curing process, which is usually carried out in a hot-press at
a defined pressure and temperature, and for a defined period of time. The
mechanical properties of the adhesive during curing can be examined by
any of the following methods: Thermomechanical analysis (TMA),
Dynamic mechanical analysis (DMA), Torsional braid analysis
(TBA), Integrated pressing and testing system (IPATES), and Automated
bonding evaluation system (ABES). Among these, ABES and IPATES
are to be preferred from the practical point of view since these techniques
provide data on the shear strength of the adhesive bond (ABES) or
internal bond (IPATES), whereas TMA, DMA and TBA measure the
changes of the different moduli. A short description and principles of the
mentioned methods is given. Examples of typical results obtained with
each technique are presented and briefly discussed. References for
detailed information on results from these methods are provided.

5.1 INTRODUCTION

Adequate bond strength and long-term performance of the wood-based

panels such as particleboard, fibreboard (MDF), oriented strand board
(OSB) and plywood is achieved with sufficient adhesive curing during
pressing. At the end of the press time – when the press opens or when the
board leaves the continuous press – a certain degree of bond strength (i.e.
certain mechanical hardening) is necessary in order to withstand the
various internal forces in the board (spring back) due to (i) mechanical
stresses caused by deformation of the wood substance (reduced by
relaxation processes within the board) and (ii) by the internal steam
pressure built up during the hot-press process. The full chemical curing
however can be completed afterwards outside of the press (e.g. during hot
stacking).

202 203
Sernek, Dunky
 Bond Strength Development
If the sufficient adhesive bond strength is not developed, the compressed
panel will exhibit either a low performance (internal bond, shear
 The further reactions of the thermosetting resins (chain
strength, etc.) or it will delaminate (blister) when the press pressure is
elongation, branching, and crosslinking) to a more or less three
released and the press opens. A stronger adhesive bond can be achieved to
dimensional network with a theoretical endless high molar mass,
a certain limit by a prolongation of the pressing time, but this increases
generating an insoluble resin which is not longer thermoformable
the cost of heat energy supplied, and reduces production capacity.
(thermoplastic). These reactions can also be followed by means of
Manufacturers of wood-based panels are thus continually trying to find an
various spectroscopic methods, taking as measure the changes in
optimal pressing time, which provides a good balance between the
the portion of various specific structural elements in the adhesive.
required performance of the panel and acceptable production costs. In
Suitable methods are Infrared (IR), Nuclear Magnetic Resonance
order to enable shorter pressing times and therefore reduced production
(NMR) 13C or solid state NMR.
costs an early, quick and instantaneous formation of the bond strength is
important.  The solidification of the adhesive (Figure 5.2) during curing by
building up the three-dimensional network, also described by the
achievable degree of cross-linking; usual test methods for this
mechanical curing are the Dynamic Mechanical Analysis (DMA),
5.2 FUNDAMENTALS the Thermal Mechanical Analysis (TMA), the Torsional Braid
Analysis (TBA), the Thermal Scanning Rheometry (Stefke and
bond strength between the wood substrate and the adhesive develops Dunky 2006) as well as various gel test methods at different
during the hardening or curing of the adhesives, which involves temperatures and using various hardeners (Stefke and Dunky
conversion of a liquid adhesive through gelation and vitrification to fully 2006), using the moment (or period) of formation of the gelled state
cured adhesive. Gelation marks the transition from liquid to a rubbery as measure.
state and it retards macroscopic flow. The gel corresponds to the
 The formation of the bond strength between two adherents
formation of an infinite network in which polymer molecules are cross-
(materials being bonded); this can be followed by methods like the
linked and form a macroscopic molecule. The viscosity and modulus of a
Automated Bonding Evaluation System (ABES) (Humphrey 1993),
polymer increase dramatically when the gel point is achieved.
the Composite Testing System (ComTeS) (Heinemann 2004), the
Vitrification occurs when the glass transition temperature (Tg) of the Integrated Pressing and Testing System (IPATES) (Heinemann
formed network rises to the temperature of cure. In adhesive curing it 2004, Roos 2000) or also any other type of tests where bonds are
marks the transformation from a rubber to a gelled glass. Tg is the critical created under certain conditions (especially in dependence of time)
temperature that separates glassy behavior from rubbery behavior. The and then tested (Sernek et al. 2005). The easiest procedure, even
strength of adhesion between adherents increases linearly as the amount time consuming, is to press a series of plywood samples with
of interlinking between the two adherent layers increases. Cross-linking different press times applied and to test the already generated and
(as effect within the adhesive resin) affects many of the physical and still generating bond strength in a suitable procedure like the shear
mechanical characteristics of thermosetting adhesives. It improves their mode. Similar tests can be performed with lab particleboards or lab
strength and durability, and also increases solvent and high temperature MDF/HDF, with following testing of the internal bond or other
resistance. Hardening and gelling of thermosetting adhesives and resins properties describing the formation of bon strength and material
can usually be monitored by: cohesion.
 The exo- or endothermic behaviour (Figure 5.1) during gelling
and hardening (chemical curing); suitable test methods are the
Differential Thermal Analysis (DTA) or the Differential Scanning
Calorimetry (DSC).

204 205
Sernek, Dunky
 Bond Strength Development

in the pressed board, which are caused by (i) the steam pressure in the
board, and (ii) the residual mechanical stresses due to the densification of
the particle or fibre mat and even the deformation of the particles and
fibre themselves.
Humphrey und Ren (1989) pressed plywood under nearly isothermal
conditions and constant moisture content. The bonding strength was
determined immediately after the end of the various press times and
increased with the press time up to a certain level when full curing
occured. By plotting the steepness of these curves for different
temperatures versus the reciprocal absolute temperature a so-called
Figure 5.1: DSC-plot of an UF-resin (Dunky 1996). reactivity index, similar to an activation energy, can be calculated.
The processes taking place during hardening depend on the basic
chemical type of the adhesive. Adhesives can consist of reactive
components with various molar masses, from low molecular weight
species as reactive monomers up to already more or less polymeric
structures still having reactive groups within their chains or at the end
points of the oligomeric or polymeric molecules; in the latter case
chemical crosslinking of rather thermoplastic structures can occur.
For thermosetting adhesives the usual reactions taking place are
polycondensation or polyaddition, with the hardened and, most
probably, three-dimensional network state as final structure. At the same
time, the development of the mechanical bond strength between the two
adherents as a consequence of the increase in cohesive bond strength by
these further reactions takes place.
The curing process usually is performed at higher temperatures depending
on the process used; but also hardening at room temperature or at slightly
elevated temperatures is possible. The pH conditions of this step vary
with the type of resin:
 With usually acidic condensation taking place for aminoplastic
resins,

Figure 5.2: DMA-plot of an aminoplastic adhesive resin (Kim et al. 2006).  Rather alkaline conditions with phenolic resins,

The bond strength generated and measured when testing lab  Resins based on resorcinol prefer even neutral conditions.
particleboards or lab MDF/HDF, however, does not only reflect the The reactions are very much accelerated by heat based on the well-known
mechanical degree of curing, but also is influenced by residual stresses temperature dependence of the rate of chemical reactions as described by
the equation of Arrhenius (Figure 5.3). In this example the inverse gel
time was taken as a measure for the rate of the hardening reaction. The
slope of the lines indicates the energy of activity of these chemical
processes.

206 207
Sernek, Dunky Bond Strength Development

 Others: rheological techniques, Composite Testing System (Com-

-3.5
TeS), etc.
-4.0
5.3.1 Thermomechanical Analysis (TMA)
-4.5
Thermomechanical analysis (TMA) involves measurements of the
-5.0 dimensional changes of material under controlled conditions of force,
ln(gel time-1)

atmosphere, time and temperature. Force can be applied in different

-5.5 modes such as compression, flexure or tension. TMA measures the
intrinsic properties of the material (Young’s modulus, glass transition,
-6.0
thermal expansion coefficient) and also processing or product
performance parameters. The viscoelastic properties of the material (creep
-6.5
or stress relaxation) can be observed. The degree of crosslinking in an
-7.0 adhesive can be determined through detecting the glass transition
temperature (Tg). The schematic principle of the TMA is shown in
-7.5 Figure 5.4. There are different experimental configurations for the TMA,
0.00265 0.0027 0.00275 0.0028 0.00285 0.0029 0.00295 0.003 with the three point flexion as the most useful one for monitoring the
1/T (K-1) curing of a wood adhesive.
Figure 5.3: Arrhenius plots of the logarithmic inverse gel time with the
gel times expressed in seconds plotted against the reciprocal
absolute temperature (1/K) for a straight UF resin as measure
for the hardening rate with various hardeners at three
temperatures. ■ 3.2% Acetic acid, ☐ 2.3% Acetic acid, ■
1.6% Acetic acid O 2.3% Ammonium sulphate, ▲1.2%
Ammonium sulphate ● 2.3% Ammonium chloride.

5.3 MONITORING THE DEVELOPMENT OF ADHESIVE BOND

STRENGTH
Figure 5.4: Schematic principle of TMA (left) and different type of
The development of the adhesive bond strength during curing can be mounting configuration (right) (http://www.anasys.co.uk
observed by different methods such as: /library/tma1.htm).

 Thermomechanical analysis (TMA), Both, non-isothermal and isothermal TMA are useful methods to forecast
the internal bond (IB) of a particleboard (Laigle et al. 1998) and to
 Dynamic mechanical analysis (DMA), investigate the influence of different mixture of adhesives on the cross-
 Torsional braid analysis (TBA), linking and hardening process of adhesives (Yin et al. 1995). This study is
an instructive example for application of TMA investigating some
 Automated bonding evaluation system (ABES), thermosetting wood adhesives (UF, MUF and PMUF). The TMA results
are presented in terms of the evaluation of relative elastic moduli as a
 Integrated pressing and testing system (IPATES),

208 209
Sernek, Dunky
Bond Strength Development

function of temperature. Three significant zones can be recognized related

(i.e. mechanical cure) are based on the monitoring of the adhesive
to the change of the relative modulus of the adhesive during curing:
stiffness and the tan as a function of temperature.
1. First zone with low modulus: at the onset of curing and before
gelling, the adhesive behaves as a liquid, which cannot yet transfer Many of the DMA studies on wood adhesives were conducted with a
glass filter as substrate, which was impregnated with adhesive
the stresses between the wood adherents.
(Christiansen et al. 1993, Umemura et al. 1996), whereas some of the
2. Second zone with increasing modulus: gelling of the adhesive DMA experiments were done using wood samples as substrates for the
creates threedimensional structures, by transforming the adhesive adhesives (Wang et al. 1996, Bučar and Tišler 1997, He and Yan 2005).
from a liquid to a rubbery state; the relative elastic modulus of the
adhesive and thus of the bond increases, assuming a sufficient
adhesion situation. The start of this second zone can be defined as a
gel temperature (Tgel).
3. Third zone with a slight decrease of the modulus after reaching a
maximum, which is due to thermal degradation processes in
aminoplastic adhesives as well as to the differences in the relative
expansion coefficient of wood and adhesive. Hardening
temperature (Tv) can be defined as a temperature where the rate of
increase of the elastic modulus attains its maximum.
TMA was also used in several other studies in order to evaluate the
adhesive bond strength development: Laigle et al. (1998), Pichelin et al.
(2000), Garnier et al. (2002), Zanetti and Pizzi (2002), Lecourt et al.
(2003), George et al. (2003).

5.3.2 Dynamic Mechanical Analysis (DMA)

Dynamic mechanical analysis (DMA) (sometimes referred to as Dynamic

mechanical thermal analysis (DMTA), is a widely used technique in order
to study the mechanical properties of materials. DMA analyses a response Figure 5.5: DMA study of amino resins (Courtesy: Jose Gomez-Bueso).
of a material subjected to a sinusoidal stress, which generates a
corresponding sinusoidal strain. Characteristic values like the modulus,
the viscosity, and the damping factor can be determined from 5.3.3 Torsional Braid Analysis (TBA)
measurements of the amplitude of the deformation at the peak of the sine
wave and the lag between the stress and strain sine waves (Menard 1999). Torsional braid analysis (TBA), the processor of DMA, is an excellent
Particularly, the storage modulus (E') (Figure 5.5) and the loss modulus technique to characterize the cure properties of polymeric materials such
(E'') can be determined from the measurements. The mechanical as adhesives. The techniques are useful in measuring the glass transition
temperature (Tg), the elastic modulus (E’) and the loss modulus (E’’).
damping term tan is defined as E''/E'. Hardening phenomena such as
These parameters are useful to characterize the transformation of a liquid
gelation and vitrification of an adhesive can be clearly identified from
to the solid state (hardening). A schematic model of a torsion pendulum is
these parameters.
shown in Figure 5.6. The adhesive sample is clamped at the two extremes.
DMA is very useful technique for observing the viscoelastic nature of The upper clamp is bound to an inertial rotating counterbalanced system.
polymers and also for examining the curing of adhesives. The cure results Torsional oscillations are induced by applying a torque.

210 211
Sernek, Dunky
 Bond Strength Development

A TTT cure diagram is a plot of the times, required to reach gelation and
vitrification functions during isothermal cure of adhesive as a function of
curing temperature. Several events occur through curing time. These
events include the onset of phase separation, gelation, vitrification, full
cure, and devitrification.
Much of the behaviour of a thermosetting material can be understood in
terms of the TTT diagram through the influence of gelation, vitrification,
and devitrification on the properties of the material investigated. Gelation
marks the transition from liquid to a rubbery state. It retards macroscopic
Figure 5.6: Scheme of an inverted torsion pendulum (Riande 2000). flow, and retards growth of a dispersed phase. Vitrification occurs when
the glass transition temperature rises to the temperature of cure. It marks
The temporary dependence of the angular displacement is measured. The the transformation from a rubber to a gelled glass or from a liquid to an
analyzed material is cured at different temperatures and the data are ungelled glass. Vitrification retards the possible further chemical
collected over time. Values of Tg are obtained in order to construct the reactions. Devitrification occurs when Tg decreases again due to excessive
time-temperature-transition (TTT) diagram (Figure 5.7). heat impact. Devitrification, due to thermal degradation, marks the
lifetime for the material. An excellent introduction to study the cure of
thermosetting systems by TBA is given by Gillham (1997); TBA was also
used in several studies on different wood adhesive types: Steiner and
Warren (1981), Steiner and Warren (1987), Ohyama et al. (1995), Tomita
et al. (1994).

5.3.4 Automated Bonding Evaluation System (ABES)

The so-called automated bonding evaluation system (ABES) has been

developed and patented by Humphrey (1993). The ABES enables the
determination of strength development characteristics of different
adhesives in combination with an adherend. The system includes bonding
and testing of a lap-shear specimen under controlled conditions (Figure
5.8).
The system uses a pair of relatively thin adherend strips (e.g. veneer,
wood or metals) and the adhesive, which is applied to the end of one strip.
The strips are put together to form a lap-shear specimen. The specimen is
put in the testing device and pressed by side like in a small hot-press.
After a certain press time, this small press opens and the specimen is
immediately (or after a defined cooling interval of few seconds) tested in
tension shear mode (Figure 5.9).
Figure 5.7: Schematic TTT diagram for a thermoseting system (Gillham
1997).

212 213
Sernek, Dunky Bond Strength Development

140oC 130oC 125oC 110oC

6
95oC

Shear strength (MPa)

5

1
0
0 100 200 300 400 500

Bond Pressing Time (Sec.)

Isothermal strength development rates (kPa/sec.)

200

150

Figure 5.8: Automated Bonding Evaluation System (Humphrey; US-patent

5176028). 100

50

90 100 110 120 130 140

o
Bond Pressing Temperature ( C)

Figure 5.10: A typical set of isothermal strength development plots for UF

adhesive-to-maple bonds (top), with a derived plot of
regressed bonding rate against temperature (bottom)
(Humphrey 2006).
Figure 5.9: Schematic picture of the ABES principle. ABES and similar adapted methods have been found to be useful for
The technique provides valuable data on the shear strength of the determining the development of bond strength for various adhesive types
adhesive bond as a function of the pressing parameters (e.g. time and under different pressing conditions (Kreber et al. 1993; Prasad et al.
temperature) and conditions (e.g. the cooling effect) (Humphrey 2006) 1994; Chowdhury and Humphrey 1999; Kim and Humphrey 2000;
(Figure 5.10). The measured time dependent shear strengths describe the Heinemann et al. 2002a; Heinemann et al. 2002b; Lecourt et al. 2003).
formation of the pure cohesive bonding strength, without wood failure. ABES is an excellent technique for strength development determination
As soon as wood failure occurs, the test series has to be stopped. because it reveals the mechanical properties (e.g. shear strength) of the
adhesive bond. For practice, mechanical testing of the adhesive bond line

214 215
Sernek, Dunky
Bond Strength Development

is the method which is used to determine the quality of the adhesive cure
and the effectiveness of the wood-adhesive interaction (Steiner and 2,0
160°C
Warren 1981). A shortcoming of ABES compared to DMA is that only

Internal Bond Strength [N/mm²]

one data point of shear strength per test is provided, since ABES is a 1,5 140°C
destructive test.
120°C
5.3.5 Integrated Pressing and Testing System (IPATES) 1,0
100°C
The so-called Integrated Pressing and Testing System (IPATES) was
developed for the determination of the adhesive cure in wood fibre and 0,5 80°C
particle mats. The principle is similar to the principle of ABES. A mat is
formed on a disc as steel press platen with 100 mm in diameter (Figure 0,0
5.11). The mat is heated by two electrically heated blocks and compressed 0 100 200 300 400
by the universal testing machine. A special adhesive is used to ensure the Pressing Time [s]
proper linkage between the press platens and the sample being tested.
After pressing, the specimen is destructively tested in tension mode for Figure 5.12: Development of internal bond strength obtained by IPATES
determination of the internal bond (Heinemann et al. 2002a). A similar (Courtesy: C. Heinemann).
method (ComTeS) has been used for testing wood composites
(Heinemann et al. 2002b).
5.3.6 Other Techniques

There are some other techniques for investigating the development of

adhesive bond strength such as acoustic test procedures (Biernacki and
Beall 1996, Chen and Beall 2000) and mechanical tests. A classical shear
test with an adapted ABES principle (heating) and with a possibility for
dielectric analysis (DEA) of the adhesive cure is shown in Figure 5.13
(Sernek et al. 2006).
This simple technique allows the evaluation of the development of
adhesive bond strength (measured as shear strength and wood failure) as a
function of the pressing time. An example of the development of adhesive
bond strength during phenol-formaldehyde resin curing is shown in
Figure 5.14. The strength development curve shows that the shear
Figure 5.11: Integrated pressing and testing system (IPATES). strength was zero at the beginning of the pressing (first stage), since the
adhesive was still in a liquid state. With longer pressing time and
Heinemann et al. (2002a) investigated the influence of temperature (80-
increased temperature of the adhesive, gelation occurred and the adhesive
160 °C), mat density (500-800 kg/m 3) and resin consumption (gluing
mechanical strength started to build up. In the second stage, the shear
factor) (7-13 %) on the internal bond (IB) of MDF. The strength
strength increased almost linearly with the pressing time. Intensive
development of the used UF adhesive resin as a function of pressing time
polycondensation and cross-linking yielded more and more linkages in
and temperature is shown in Figure 5.12. It is evident that the
the three dimensional network, which allowed the adhesive to withstand
development of IB is much slower at lower pressing temperatures
higher stresses. Further curing led to the last stage where the shear
(especially 80 °C), and it does not reach the same level as for the samples
strength curve leveled off and reached a maximum value. A model was
pressed at higher temperatures.
used to describe the development of adhesive bond strength as a function
of time (solid line).

216 217
Sernek, Dunky
 Bond Strength Development

5.4 OPEN QUESTIONS

Main task of a comprehensive evaluation of the ongoing processes

include reliable correlations between these parameters. Still a partly
unclear question is the description and prediction (i) of the suitability of
adhesives for bonding two surfaces and (ii) of the achievable properties of
the bond line and the bonded products. Hence valuable, but still
controversial information concerning the curing behaviour of various
resins is gained by comparing the progress of both, the chemical and the
mechanical curing; up to now only few examples in literature describe
especially this connection of these two hardening processes as they can be
monitored using different methods; one reason for this fact might be the
difficult experimental realization. The following types of presentations
are possible:
 Plotting the achieved partial degrees of curing after a certain time
span in an x-y-plot can reveal the comprehensive hardening pattern
of a resin (Figure 5.15) (Geimer et al. 1990). Such a correlation
plot of the degree of chemical cure (e.g. measured by DSC) and the
Figure 5.13: Device for investigation of the physical-chemical and
mechanical properties of a specimen during resin curing: (a) increase of mechanical strength (e.g. measured by DMA or ABES)
pressing of the specimen and monitoring the curing by DEA; can be regarded as a fingerprint of the curing behaviour of a resin.
(b) determination of the shear strength of the adhesive bond.
 Figure 5.15: Chemical-mechanical hardening plot for two different PF
resins; prehardening temperature 140 °C, times in minutes
(Geimer et al. 1990).
Figure 5.14: Development of the shear strength of a phenol-formaldehyde
adhesive bond.  Monitoring the course of both reactions over time (Geimer and
Christiansen 1991, 1996).

218 219
Sernek, Dunky  Type of the adhesive,
 Composition and cooking procedure,
 Monitoring the course of the chemical hardening during the press
time up to the moment where the press opens and therefore some  Type and amount of hardeners,
certain bond strength must be reached (Heinemann et al. 2004).
 Additives which might accelerate or retard the hardening process,
It has been recognized, that the various test methods to describe the
hardening behaviour and the formation of the bond strength not  Hardening temperature: press temperature, temperature in the bond
necessarily show the same course of the changes in the resins; this is quite line, temperature in the core layer,
obvious, because all methods are based on different changes in the
adhesive during curing.  Properties of the wood surfaces.

Especially the forming of the bond strength is influenced by much more Since no universal method exists describing all aspects of gelling and
parameters than just reflecting the chemical curing as seen (i) as the hardening as well as of formation of bond strength, still various methods
thermal reaction in DTA or DSC or (ii) by monitoring the increase in must be used side by side. Combining experiences of chemical reaction
molar mass or (iii) looking at the changes in the portion of various pathways and results of the various test procedures as described above give
structural elements in the resin. All these parameters rather describe the a good chance when further developing and tailor making adhesive resins
covalent behaviour of the adhesive in the bond line; bond strength and adjusting their behaviour to special bonding processes.
however includes parameters like wetting and penetration (including over-
penetration), since forming of adhesively acting bonds is mainly based on
secondary attraction forces.
Target of the development of reactive adhesives is to achieve a reactivity
as high as possible, however under consideration of e.g. the storage
stability of the adhesive, the pot life of the adhesive mix and the existing
process parameters. The reactivity of an adhesive (mix) hence is
determined by various parameters:
 Bond Strength Development Dunky, M. unpublished results (1996, 2004)
Garnier S., Pizzi A., Huang Z. 2002. Dry I.B. forescasting of commercial
5.5 REFERENCES
tannin adhesives-bonded particleboard by TMA bending. Holz als Roh-
und Werkstoff, 60, 5: 372
Biernacki J.M., Beall F.C. 1996. Acoustic monitoring of cold-setting
adhesive curing in wood laminates. International Journal of Adhesion and Geimer, R.L., Christiansen A.W. 1996. Critical values in the rapid cure
Adhesives, 16, 3: 165-172 and bonding of PF-resins. For.Prod.J. 46, 11/12: 67 - 72
Bučar D.G., Tišler V. 1997. Curing dynamics of tannin-based adhesives. Geimer, R.L., A.W.Christiansen. 1991. Adhesive curing and bonding:
Holzforschung und Holzverwertung, 48, 6: 101-104 response to real time conditions. Proceedings Adhesives and Bonded
Wood Products, Seattle, WA, 12 - 29
Chen L., Beall F.C. 2000. Monitoring bond strength development in
particleboard during pressing, using acousto-ultrasonics. Wood and Fiber Geimer R.L., Follensbee R.A., Christiansen A.W., Koutsky J.A., Myers
Science, 32, 4: 466-477 G.E. Resin characterization. Proceedings 24th Wash. State University Int.
Particleboard/ Composite Materials Symposium, Pullmann, WA, 1990, 65
Chowdhury M.J.A., Humphrey P.E. 1999. The effect of acetylation on - 83
the shear strength development kinetics of phenolic resin-to-wood bonds.
Wood and Fiber Science, 31, 3: 293-299 George B., Simon C., Properzi M., Pizzi A., Elbez G. 2003. Comparative
creep characteristics of structural glulam wood adhesives. Holz als Roh-
Christiansen A.W., Follensbee R.A., Geimer R.L., Koutsky J.A., Myers und Werkstoff, 61, 1: 79-80
G.E. 1993. Phenol-formaldehyde resin curing and bonding in steam-
injection pressing. Part II. Differences between rates of chemical and Gillham J.K. 1997. The TBA torsion pendulum: a technique for
mechanical responses to resin cure. Holzforschung, 47, 1: 76-82 characterizing the cure and properties of thermosetting systems. Polymer
International, 44(3): 262-276

220 221
Sernek, Dunky Adhesives 2005, San Diego, CA
Humphrey, P.E., S. Ren. 1989. Bonding kinetics of thermosetting adhesive
He G.B., Yan N. 2005. Effect of wood species and molecular weight of systems used in wood-based composites: the combined effect of temperature
phenolic resins on curing behavior and bonding development. and moisture conent. J.Adhesion Sci.Technol. 3: 397 - 413
Holzforschung, 59(6): 635-640
Humphrey, P.E. 1990. Device for testing adhesive bonds. USP 5176028
Heinemann C., Fruehwald A., Humphrey P.E. 2002a. Evaluation of
adhesive cure during hot-pressing of wood-based composites. COST E13, Kim J.W., Humphrey P.E. 2000. The effect of testing temperature on the
Workshop Proceedings, 3rd Workshop – Vienna, 19-20 September 2002: strength of partially cured phenol-formaldehyde adhesive bonds.
1-8. http://www.bfafh.de/bfh-pers/pdf/publ_heinemann_02_3.pdf Proceedings Wood Adhesives 2000, South Lake Tahoe, CA
Heinemann C., Lehnen R., Humphrey P.E. 2002b. Kinetic response of Kim S., Kim, H.-J., Kim, H.-S., LeeY.-K., YangH.-S. Thermal analysis
thermosetting adhesive systems to heat: physicio-chemical versus study of viscoelastic properties and activation energy of melamine- modified
mechanical responses. Proceedings of the 6th Pacific Rim Bio-Based urea-formaldehyde resins. J.Adhesion Sci. Technol. 20 (2006) 8,
composites symposium. Portland/USA, 10.-13.11.2002, Corvallis: 803 - 816
Oregon state university: 34-44
Kreber B., Humphrey P.E., Morrell J.J. 1993. Effect of polyborate pre-
Heinemann, C., R. Mitter, M. Dunky, Thermokinetic simulation of a hot treatment on the shear strength development of phenolic resin to Sitka spruce
press cycle in the production of particleboards. Eighth European Panel bonds. Holzforschung, 47, 5: 398-402
Products Symposium, Llandudno (North Wales) 2004
Laigle Y., Kamoun C., Pizzi A. 1998. Particleboard IB forecast by TMA
http://www.anasys.co.uk/library/tma1.htm bending in UF adhesives curing. Holz als Roh- und Werkstoff, 56, 3: 154
Humphrey 2006. Temperature and reactant injection effects on the
bonding kinetics of thermosetting adhesives. Proceedings Wood
 Bond Strength Development Sernek M., Kokalj A., Jošt M. 2006. The development of adhesive bond
strength during phenol-formaldehyde resin curing. Wood resources and
Lecourt M., Pizzi A., Humphrey P. 2003. Comparison of TMA and panel properties : conference proceedings : Cost Action E44-E49,
ABES as forecasting systems of wood bonding effectiveness. Holz als Valencia, Spain, 12-13 June 2006. Valencia: AIDIMA, Furniture, wood
Roh- und Werkstoff, 61, 1: 75-76 and packaging technology institute: 89-96

Menard, K.P. 1999. Dynamic mechanical analysis : a practical Stefke, B., M. Dunky, Catalytic influence of wood on gelling of
introduction. CRC press. Boca Raton, 208 p. formaldehyde based glue resins. J.Adhesion Sci. Technology 20 (2006) 8,
761 – 785
Ohyama M., Tomita B., Hse C.Y. 1995. Curing property and plywood
adhesive performance of resol-type phenol-urea-formaldehyde Steiner P.R., Warren S.R. 1981. Rheology of wood-adhesive cure by
cocondensed resins. Holzforschung, 49, 1: 87-91 torsional braid analysis. Holzforschung, 35, 6: 273-278

Pichelin F., Pizzi A., Fruehwald A. 2000. OSB adhesives rate of strength Steiner P.R., Warren S.R. 1987. Behavior of urea-formaldehyde wood
development on single strands couples. Holz als Roh- und Werkstoff, 58, adhesives during early stages of cure. Forest Products Journal, 37, 1: 20-
3: 182-183 22

Prasad T.R.N., Humphrey P.E., Morrell J.J. 1994. The effects of Tomita B., Ohyama M., Itoh A., Doi K., Hse Ch.-H. 1994. Analysis of
chromated copper arsenate and ammoniacal copper zinc arsenate on shear curing process and thermal properties of PUF cocondensation resins.
strength development of phenolic resin to Sitka spruce bonds. . Wood and Mokuzai Gakkaishi (J.Japan Wood Res.Inst.) 40, 2: 170-175
Fiber Science, 26, 2: 223-228 Umemura K., Kawai S., Mizuno Y., Sasaki H. 1996. Dynamic
Riande, E., Polymer Viscoelasticity, 86, 2000. mechanical properties of thermosetting resin adhesives II. Urea resin.
Mokuzai Gakkaishi, 42, 5: 489-496
Roos, T., Diploma Thesis, University of Hamburg, 2000

222 223
Sernek, Dunky
Quality Control

Wang X.M., Riedl B., Geimer R.L., Christiansen A.W. 1996. Phenol-
formaldehyde resin curing and bonding under dynamic conditions. Wood Chapter 6
Science and Technology, 30, 6: 423-442
Yin S., Deglise X., Masson D. 1995. Thermomechanical analysis of
Innovative Methods for Quality Control in the
wood/aminoplastic adhesives joints cross-linking - UF, MUF, PMUF. Wood-Based Panel Industry
Holzforschung, 49, 6: 575-580
Jochen Aderhold and Burkhard Plinke
Zanetti M., Pizzi A. 2002. Time delay effect in TMA methods for MUF
resins testing. Holz als Roh- und Werkstoff, 60, 5: 3
CHAPTER SUMMARY

The objective of this chapter is to review some innovative techniques for

quality control in the wood-based panels industry which are not yet
commercially available. The authors selected those techniques which they
think are most likely to find industrial application in the near future:
infrared thermography, near infrared reflectometry, and nuclear magnetic
resonance.

6.1 INTRODUCTION
 Wood based panel companies face their customer's increasing demand for
high and consistent quality of their products and for flexible production.
At the same time, the cost pressures intensify. Quality assurance systems
following standards such as ISO/EN/DIN 9000 become more and more
important. Just in time delivery concepts require the production of up to
20 different panel types on the same production line within 24 hours
(Nielsen 1994). In order to meet these requirements, a careful monitoring
of the production process and preventive measures to avoid production
failures are necessary (Deubel 1992a, Deubel 1992b). This includes
appropriate testing procedures for the final products, the panels. Panel
testing methods roughly fall into two categories: on-line methods and off-
line methods. On-line methods should be integrated into the production
process and work non-destructively. In many cases, every single panel
can be tested. Some points of interest are measurement of the raw density
and the mechanical strength as well as the detection of structural failures
such as delaminations. Off-line techniques on the other hand will always
be limited to spot samples which can be tested for properties not
accessible with on-line testing.
Regarding on-line techniques, a variety of testing methods have
developed over the years, many of which are already available on the
market, including ultrasonic, microwave, and x-ray techniques as well as

224 225
Aderhold, Plinke
Quality Control

image processing. These commercially available techniques are not œ

within the scope of this chapter since they are described in detail 0 0 4
fi (4)
elsewhere (e.g. Welling 1998).
0

Recently, some new techniques have evolved, including infrared with:

thermography, near infrared reflectometry, and nuclear magnetic
5 4
resonance. These three new methods will be described in detail in the
B
following sections since the authors believe that these techniques are most 32
–8 –2 –4
(5)
likely to find applications in the industry in the near future.

6.2 INFRARED THERMOGRAPHY wavelength is given by the famous Planck equation (Planck 1900):

M_0_Lambda [W/μm*m^2]
–5
0 2
6.2.1 Principles of Infrared Thermography

The basis of infrared thermography is the fact that every object having a
temperature above absolute zero emits electromagnetic radiation which is
called thermal or Planck radiation. At a given wavelength the radiated
power density (for a so-called black body) depends on the temperature
only so that the temperature can be calculated by measuring the radiated
power density. The dependence of the power density on temperature and
 Figure 6.1: Thermal radiation of a black radiator between 300 K and
fi (1) 700 K.
2
0
In this equation M Furthermore, the peak position of the power density max shifts to smaller
is the power density per3wavelength interval emitted
Z
wavelengths for increasing temperatures. This is described by Wien’s
by an ideal ("black") radiator (unit: W/m ), whereas Z stands for the
wavelength and T for the absolute temperature. The constants C 1 und C2 displacement law:
contain only natural constants such as h (Planck's constant), c (speed of
light) und kB (Boltzmann's constant): 2500
NAS (6)
300 K
2 –16 2 Historically,
2000
both the Stefan-Boltzmann law and Wien's400displacement
K
500 K
1 (2) law were discovered before Planck's law (Stefan 1879, Boltzmann
600 K 1884,
Wien 1893).
1500 700 K
2 B (3)
Figure 6.1 shows the thermal radiation of a black radiator for the Since M01000
depends on temperature only, one can estimate the temperature
Z
temperature range between 300 K and 700 K. It is apparent that the of an object
500
by measuring the power density if Planck's equation is valid,
overall power density (the areas under the curves) increases strongly with i. e. when the atoms or molecules of the object are in thermal equilibrium.
increasing temperature. This fact is mathematically described by the In most practical
0 cases this will be true. Gas discharges and similar
Stefan-Boltzmann law which can be obtained by integrating Planck's phenomena are 1 an exception from Planck's
10 law. Consequently,100 the
Lambda [μm]
equation: temperature of gas discharges cannot be measured in the described way.

226 227
Aderhold, Plinke hand have normally very low emissivities. In the infrared region of the
electromagnetic spectrum this is especially true for metallic surfaces. In
At temperatures between room temperature and 450°C, most of the these cases reflected thermal radiation from the environment can exceed the
radiated intensity is in the so-called thermal infrared (IR) region of the object's thermal radiation, which can lead to misinterpretations. In particular,
electromagnetic spectrum having wavelengths of 2.5 μm and longer. Of the infrared image can show a reflection of the camera itself, so that the
practical importance are the regions labelled as mid-wave IR (MWIR, cooled detector appears as a very cold spot. This is the so-called narcissus
3 μm … 5 μm) and long-wave IR (LWIR, 8 μm … 12 μm) where suitable effect. Examinations of metallic surfaces using thermography thus require
detectors are available, and where the atmosphere is sufficiently extreme care. In many cases, problems with low emissivity surfaces can be
transparent. solved by covering them with special high emissivity paints.

The "black" body describes the ideal, theoretical case. Real objects will 6.2.2 Infrared Cameras
under given conditions emit less power density than black bodies. This
fact is usually described by a factor which is called emissivity. It is A huge number of different infrared cameras having different detection
represented by the Greek letter s and can, by definition, have values principles are available. Today, almost all cameras have focal plane arrays,
between 0 and 1. i.e. matrices from single detector pixels as in normal digital cameras. From a
practical point of view, the most important difference is between cameras
The emissivity of an object depends on the nature of its surface.
having cooled photovoltaic detectors or microbolometer detectors,
Empirical values for many materials can be taken from the literature respectively. Cooled photovoltaic detectors are made from compound
(Gaussorgues 1994). Many organic materials, water and glass have
semiconductor materials. Infrared photons having an energy larger than the
emissivities of 0.9 or higher. Strongly reflecting surfaces on the other band gap can be absorbed, generating an electron-hole pair, and
consequently electric current. The number of semiconductors having a Quality Control
suitably small band gap is limited, including InSb (Indium Antimonide)
and HgCdTe (Mercury Cadmium Telluride, MCT or CMT) as the most
important materials. These semiconductors are difficult to grow in high causing high cost. Furthermore, to improve signal-to-noise ratio, they
quality, and production yield of focal plane arrays is low, have to be cooled down to 90 K or lower. In the past this was mainly done
by cooling with liquid nitrogen. Today, mechanical coolers, so-called
Stirling coolers, can do the cooling. However, these are again expensive.
Consequently, prices for cameras with cooled detectors are in the range of
70 k€. Their temperature resolution is around 15 mK. Frame rates can be
as high as 1 kHz.
The pixels of microbolometer detectors are essentially very thin silicon
plates with an absorbing layer. When an infrared photon is absorbed by
these plates their temperature and consequently their electric resistance
rises, which can be used for detecting the photons. Since the production
of these structures is similar to the well-developed silicon CMOS
technology, these detectors are much cheaper than InSb or MCT
detectors. Although the temperature of microbolometer detectors should
be well-controlled as well, and although Peltier coolers are often applied,
these devices do not require expensive Stirling coolers. Consequently,
prices for microbolometer cameras start from 10 k€. A temperature
resolution of 100 mK and frame rates up to 50 Hz can be achieved. Since
this is sufficient for many applications, it is expected that microbolometer
cameras will become more and more important and will allow new
applications of thermography in many areas for which cooled
photovoltaic cameras would be too expensive.
It should be noted that glass and quartz are not transparent in the thermal
IR so that expensive special lenses from Si, Ge, ZnSe or suchlike are
required for any camera type.

6.2.3 Thermography for Non-Destructive Testing

As explained before, an object's infrared radiation depends on its

emissivity as well as on its temperature. Both dependencies can be used
for infrared machine vision in the wood-based panel industry.
In classical machine vision in the visible spectral range, differences in
intensity of the light reflected from the object under study are utilized in
order to draw conclusions about possible defects or foreign bodies. In a
similar manner, emissivity differences can be used. Objects having
different emissivities radiate thermal radiation of different intensity even
if they are at the same temperature. Thus, they have different grey values
in the infrared image. Alternatively, the objects can be illuminated by
infrared radiators so that the light reflected by them can be detected. The
contrast in the infrared can be very different from that in the visible part
of the spectrum. For instance, metal and glass surfaces are highly

228 229
Aderhold, Plinke
reflective in the visible range, whereas in the infrared this is only true for Quality Control
metal surfaces. Glass has a high emissivity and consequently a low
reflectivity in the infrared range.
Cooling
Apart from emissivity differences, and more importantly, temperature
differences can be used for panel testing and process control. In systems
where heat and mass transfer is possible (which is true in most practical
cases) temperature differences are always connected with heat flow.
Temperature differences will cause heat flow, and a heat flow caused by
some reason will lead to temperature differences, the amount of which
will depend on the object's thermal properties. If no inner heat sources
(such as exothermic chemical reactions) and no bulk flows are present, as
will be the case for wood-based panels a certain time after pressing, heat
flow and consequently temperature differences will occur only when the
object is not in thermal equilibrium with its environment. This means that Cooling
the object must cool down or heat up in some way. In such a cooling or
heating process, regions of different thermal conductivity and/or capacity Figure 6.2: Principle of passive heat flow thermography.
can have different surface temperatures and can thus be differentiated in
If it is not possible to utilize cooling or heating processes caused by the
the infrared image. This can be used for panel testing since many defects
production process, the object can be exposed to an exterior heat pulse
differ in thermal conductivity (air inclusions, delaminations) or thermal
(active heat flow thermography, see Figure 6.3). In the easiest case, this
capacity (moisture) from the good regions. The defects need not be at the
can be done if the object passes an infrared heater on a conveyor belt.
surface in order to be detected but may be deeper in the material. The
This will cause a heat wave which penetrates into the object while the
exact penetration depth of the technique depends on many factors.
surface cools down. If the conduction of heat is retarded by defects of low
How can the necessary cooling or heating processes be generated? The thermal conductivity, the surface above the defect will stay warm for a
easiest case are objects which cool down (or heat up) anyway as a longer time, which can again be detected in the infrared image. As a rule
consequence of the production process. Wood-based panels produced by of thumb one can expect to see a defect if its distance from the surface
hot pressing are a good example. In this case, it is sufficient to simply does not exceed its diameter.
observe the object with an infrared camera (passive heat flow
thermography, see Figure 6.2). Surface regions under which defects with Heat pulse
a low thermal conductivity can be found will cool down more rapidly
since less heat can be supplied form the hot interior. Air inclusions and
delaminations are examples for such defects. In such a way, they can be
detected in the infrared image.

Figure 6.3: Principle of active heat flow thermography.

6.2.4 Application and Examples

Since wood-based panels are mostly manufactured by hot-pressing,

quality control is in many cases possible by passive heat flow

230 231
Aderhold, Plinke
Quality Control

thermography. Here the cooling of the objects to be tested is observed by

between 40 °C (coating with decorative papers) and more than 100 °C
an infrared camera. The camera is installed typically directly behind the
(OSB). Figure 6.5 shows infrared images of blisters in particleboards of
press (Figure 6.4).
22 mm thickness as an example. The images were taken after the
application of decorative papers.

Figure 6.5: Infrared images of blisters in 22 mm thick particleboards.

The dark (cool) spot in the left imager shows a blister in the
lower layer while the light (warm) spot in the right image
shows a blister in the upper layer.

Infrared images of particleboards can be easily evaluated automatically

because of their homogeneous structure. A bit more complicated is the
situation for plywood, where the natural wood structure is still
recognizable when applying decorative papers. Nevertheless, defects like
Figure 6.4: Set-up for passive thermography behind the press.
fallen out knots or defective gluing can be detected (Figure 6.6).
Many defects differ in thermal capacity and/or conductivity from the A further field of application of the passive thermography in the wood-
good areas and become thus apparent by different surface temperatures. based panel industry is the detection of fluctuations in density and
If, for example there is a delamination, the surface above this spot cools moisture content in the production of oriented strand boards (OSB).
down faster, because due to the lower heat conductivity of the defect less Such inhomogeneities involve corresponding differences in the thermal
heat follows from the inside than in good areas. capacity leading to a different cooling behaviour. Figure 6.7 shows a
Typical defects detectable this way include: freshly pressed OSB laboratory board with intentional defects such as
over-moistening (top right) and excess raw density (left below). The
 Blisters in particleboards infrared image (right) was taken at a time where the thermal contrast
between defect and good material was maximized.
 Delaminations between wood-based panels and decorative papers
After an appropriate calibration the density measurement can be done
 Defective gluing in plywood quantitatively. To this end a number of test boards was first characterized
by thermography and then cut into small pieces, the density of which was
 Fallen out knots in the internal veneers of plywood
determined by weighing. The density distribution obtained this way
 Variable glueline thicknesses corresponds both qualitatively (Figure 6.8) and quantitatively (Figure 6.9)
very good with the gray values of the infrared image.
Thermography has successfully been used for the testing of boards with
thicknesses between 3 mm and 38 mm and at production speeds of up to
30 m/min. The surface temperatures of the wood-based panels can vary

232 233
Aderhold, Plinke Quality Control

775
750
725
700
675

Raw density (kg/m³)

650
625
600
575
550
525
37710

Figure 6.6: Infrared image of plywood boards of 5 mm thickness. The

left image shows a fallen out knot in the middle layer, while 3773037750 37770
the right image shows a defective gluing of the veneers. Gray value

Figure 6.9: Quantitative comparison between density and grey value.

In case that no cooling processes inherent to the process can be used, the
active variant of the heat flow thermography (Figure 6.10) can be used
where the boards are transported on a conveyor belt passing first an
infrared heater and then the infrared camera. Important applications here
also are the testing of plywood regarding gluing defects and fallen out
knots as well as of high-quality solid wood (wood for musical instruments
or pencils) regarding density inhomogeneities.

Figure 6.7: Freshly pressed OSB laboratory board with intentional

defects.

3 800 -850
4
750 -800
700 -750
5 650 -700
6 600 -650
550 -600
7 500 -550
8 450 -500
400 -450
9

Figure 6.8: Comparison of gravimetrically determined density Figure 6.10: Principle of the online active heat thermography.
distribution (left) and infrared image (right) of an OSB
laboratory board. Density scale in kg/m³. Black knots in solid wood can also be detected by means of ultrasonic
excitation. Black knots cause problems because they can fall out easily.
The ultrasonic excitation causes frictional heat in the black knots which
can easily be detected by thermography. Good knots give no signal
(Figure 6.11).

234 235
Aderhold, Plinke
Quality Control
 infrared since the levels are equidistant. Near infrared spectral lines would
mechanics, only transitions between neighbouring levels would be not exist.
allowed which all would have the same frequency in the mid-wave Table 6.1: The infrared spectral region (Osborne 1993).
Characteristic
Spectral range Wavelength range [μm]
transitions
Molecular vibrations:
Figure 6.11: Solid wood piece with good knot (left top) and black knot Near Infrared (NIR) overtones and mixing 0,7 ... 2,5
(right above). The black knot appears as light area in the frequencies
infrared image. Molecular vibrations: 2,5 ... 50
Mid-wave Infrared
fundamental frequency
Suggested R&D topics include: Far Infrared Molecular rotation 50 ... 1000

 Automated defect detection in wood-based panels having an The existence of NIR spectral lines is due to deviations of the chemical
inhomogeneous structure such as massive wood or plywood bond from the harmonic oscillator model. If nonlinearities are included,
transitions between levels other than immediate neighbours are allowed as
 Techniques for compensating parasitic effects such as changing well, leading to so-called overtones. They are not necessarily integer
background conditions, heat sources in the surrounding, sunlight multiples of the fundamental frequency since nonlinearity also leads to
etc. non-equidistant levels. Furthermore, frequency mixing is possible.
Consequently, NIR spectra are more complicated, but also contain more
 Building a solid fundament for density measurements by means of information than spectra at longer wavelengths.
thermography
The position of NIR spectra lines for a given bond has been extensively
studied and can be found in the literature. Table 6.2 shows the position of
some spectral lines for some bonds which are important for adhesives
6.3 NAR INFRARED REFLECTOMETRY used in the wood-based panel industry. The exact positions depend on the
structure of the complete molecule.
6.3.1 Origin of NIR Spectra Table 6.2: Some NIR spectral lines relevant for adhesives used in the
wood-based panel industry (Osborne 1993).
The infrared spectral region contains the wavelength interval between
0.7 μm and 1000 μm and is subdivided in the near, mid-wave and far Type of bond Contained in … 1 [nm] 2 [nm] 3 [nm]
infrared (Table 6.1). Please note that this definition is a little different H2O Water ~1458 ~980 ~744
from that given in the thermography chapter.
-CH2 Formaldehyde 1738-1795 1170-1209 890-913
Many molecules have strong spectral lines in this region. The reasons for -NH2 Urea 1501-1535 1014-1031
this are molecular rotations and vibrations. Molecular rotations cause
-CH (aromatic) Phenol ~1684 ~1134 ~859
spectral lines in the far infrared and are not important here. Molecular
vibrations, i.e. stretching and bending oscillations of the chemical bond, ROH Phenol 1398-1421 ~979
lead to complicated spectra in the near and mid-wave infrared. If the
chemical bond is approximated by a harmonic oscillator and treated 6.3.2 NIR Reflectometry
quantum mechanically, a spectrum of equidistant energy levels is
obtained. Transitions between these levels are connected with emission or The appearance of spectral lines which are specific for a given molecule
absorption of infrared radiation. Following the rules of quantum in the infrared spectral region can be used to identify many (mostly

236 237
Aderhold, Plinke normally does not play a role and does not disturb the measurements. Cheap
conventional optical materials such as glass and quartz can be used whereas
organic) substances or even to measure their quantity in compounds. The expensive special optics made from Ge, ZnSe and suchlike are needed in the
NIR range has some advantages over other spectral ranges: mid-wave and far infrared. Due to the shorter wavelength the ratio of
scattered and absorbed radiation is larger than in the other infrared regions.
Since NIR spectra are due to anharmonicities, they are more differentiated This fact is utilised in NIR reflectometry. A ray of light striking a surface
than spectra in other infrared ranges. The object's thermal radiation can be reflected in a mirror like way (Figure 6.12a) or penetrate into the
sample. The light penetrating into the sample can be absorbed (Figure Quality Control
6.12c) or reflected diffusively (Figure 6.12b). It is the diffusively reflected
part of the light which contains spectral information of the sample. It can
easily be collected by a lens and subsequently analyzed. of adhesive as well as the mixing ratio of softwood and hardwood can be
measured in the mat of a particleboard line (Niemz 1994).
Unlike other spectroscopic techniques, NIR reflectometry does not
require sample surfaces of high optical quality. Elaborate sample Recovered wood can be recycled into particleboards. However, it is
preparation is not necessary. Furthermore, the samples need not to be important to sort out wood treated with paint or preservatives. For this
transparent. For this reasons NIR reflectometry is a suitable tool for purpose, NIR reflectometry has been proposed as well (Feldhoff 1998).
online process monitoring in the wood-based panel industry. However, Gindl et al. have shown a clear correlation between the mechanical
calibration procedures and multivariate statistical methods (chemometry) strength of European larch and the NIR reflection spectrum – with
are necessary for evaluating NIR data. possible applications in strength grading (Gindl 2001).
Since NIR spectra are temperature dependent (Thygesen 2000), the
spectra can in principle also be used for temperature measurement.

6.3.4 Multivariate Techniques for Data Analysis

The evaluation of NIR data can be done by modern multivariate

techniques such as principal components analysis (PCA) and related
techniques. Mathematically, PCA is a coordinate transformation. It
Figure 6.12: Mirror like reflection (a), diffuse reflection (b) and absorption assumes that any measured spectrum having N data points can be
(c) of a ray of light (Givens 1997). described by a point in a coordinate system having N dimensions. The
measured spectra for a given set of samples which differ in some relevant
property consequently form a point cloud in this coordinate system. PCA
6.3.3 Panel Properties and NIR Spectra now transforms this coordinate system into a new one which is
orthogonal, and where the variance of the data is largest for the first axis.
In the wood-based industry NIR reflectometry is already being used for The variance is a measure for the information content of a data set. The
the measurement of moisture content. Many other applications have second axis has the second largest variance and so on. This coordinate
been discussed in literature but they have not found industrial applications transformation is described by a transformation matrix whose column
in a broader frame. For example, Niemz et al. have shown that the amount vectors are the sought-after principal components. The idea is to consider
only the first few principal components whose variances describe a
sufficient part of the overall variance. This is equivalent to choosing a few
wavelengths at which the samples show the largest differences. By
evaluating only these selected wavelengths, the amount of data to be
processed can be dramatically reduced.

6.3.5 Technical Aspects

As in other spectroscopic techniques, NIR has three functional blocks:

excitation source, dispersive element, and detector. The easiest excitation
sources are thermal radiators such as halogen lamps.
The most popular dispersive elements are holographic gratings. More
advanced devices are acousto-optical tuneable filters (AOTFs) and line
spectrographs. A line spectrograph transforms a line from the object
under study onto the matrix detector of a normal camera in such a way

238 239
Aderhold, Plinke
Quality Control

that one dimension of the array contains the spatial information along the
line, whereas the other dimension contains the optical spectrum of a given
point on the line. By moving the object under test along a conveyor band
it can be scanned line by line, resulting in a complete set of spectra for
every single point on the surface, limited only by the pixel number of the
camera (spectral imaging). Line spectrographs are available for spectral
ranges from the ultraviolet to the thermal infrared. In the NIR range, of
course a NIR camera has to be used as a detector.

Figure 6.13: Principle of a line spectrograph.

On the detection side, semiconductor detectors are used most frequently Figure 6.14: Application of Spectral Imaging for resin detection on wood
particles, (Top left: Original scene showing strands with and
which utilize photoconductivity (e.g. PbS detectors) or photovoltaic
without UF application, Bottom left: Mean NIR intensity per
effects (e.g. InGaAs detectors). PbS detectors have a broader spectral
Pixel shown as gray value, Bottom right: local spectrum with
range (800 nm to 3.0 μm) than InGaAs detectors (900 nm to 2.0 μm), but characteristic UF peak, Top right: Image segmented by peak
they are less sensitive than InGaAs by about one order of magnitude. area, strands with UF application are marked blue).
Both detectors can be used to build NIR cameras. InGaAs focal plane
array cameras were introduced to the market for acceptable prices some The grey value image in Figure 6.14 (bottom left) was computed from the
years ago. mean intensity of each pixel in such a way that image structures and
strand shapes are roughly visible. Further steps are necessary to extract
6.3.6 Applications and Examples information from the hyperspectral image: The spectra of each pixel
(typically 25000 per image) are smoothed, normalized to absorption
Images which comprise complete reflexion spectra for each pixel (rather spectra and transformed to their 2 nd derivation. Then compounds such as
than three intensity values for red, green and blue like in conventional resins or water can be can be detected by their absorptions at typical
imaging) are called hyperspectral images. To achieve such an image a wavelengths. In Figure 6.14 (bottom left) the absorption spectrum of the
line spectrograph and a NIR camera are used to scan a moving surface. pixel marked to the left of the diagram is shown. Beside the peaks due to
Figure 6.14 (top left) shows a mat formed with OSB strands. It was surface moisture (around 1400 nm) the peak typical for UF-based resins
scanned using a demonstration set-up for NIR spectral imaging (Plinke, appears. In Figure 6.14 (top right) all pixels showing this peak in their
Ben-Yacov 2009) at a resolution of approximately 3 mm/Pixel in moving spectra are marked as blue. Thus, strands with an UF application can be
direction and of 1 mm/Pixel across. A hyperspectral image with a spectral distinguished from those without resin.
resolution of 316 intensity values for the wavelength range 1050 to
1650 nm is stored and evaluated in a specific image processing program. Classical chemometry offers more sophisticated evaluation methods for
spectra (Kessler 2007). With PCA, a set of spectra can be decomposed

240 241
Aderhold, Plinke
Quality Control

into their principal components, i.e. a new set of matrices of scores,

loadings and residues. During the decomposition process, the variation of
the spectra is transformed into a small set of principal components for
each spectrum (scores) and loadings which show the dependency from the
wavelengths. The elements in the score matrix can be used to form a
score space as a good expression of the similarity of spectra. If their
scores are in close vicinity in the score space then they probably comprise
NIR signals of the same compounds. However, this method can only be
applied successfully if other effects to the spectra than chemical
composition (e.g. scattering, measurement geometry, drift of illumination,
surface properties etc.) can be excluded or controlled by selection of
specific principal components. Then it is possible to build a mathematical
model from a set of training data and use it to predict the composition of
compounds from their spectra.
An application of PCA used to segment a hyperspectral image is
demonstrated in Figure 6.15: The top image is a scene with strands
without resin in the background (red) and with three types of resin (UF,
blue; PMDI, brown; MUPF, green). The colour classification of the pixels
is derived from the scores of their 1 st and 2nd principal components. Their
score space is shown in Figure 6.15 (bottom) as one dot for each pixel.
The scores of PC-1 and PC-2 show distinct clusters, and therefore the
pixels in the left image are coloured according to the membership of their
PC set to one of the clusters. Again, it must be emphasized that several
conditions must be fulfilled to use PCA for an evaluation of hyperspectral
images. The images in Figure 6.14 and 6.15 were generated with the
software tool “SpectraWalker” which has been developed for this
purpose.
Figure 6.15: Application of PCA in spectral imaging to detect different
Other than moisture measurement, there are few industrial applications of resin types (Top: Strands, coloured according to resin type
NIR reflectometry and especially NIR spectral imaging reported for the (red: no resin; blue: UF; brown: PMDI; green: MUPF),
wood-based panel industry. Hutter et al. (2003) published a study that Bottom: Score plot of two principal components of the pixels
aimed at improving fibreboard production by a design of experiments in the left image).
procedure. In this study wet-process fibreboard production was
systematically investigated in order to better understand the process and NIR reflectometry methods are considered to be an important tool for
to address ways for future improvements. Experimental Design was gathering process data. Problems such as fluctuations of the light source,
applied to examine the effects of different wood mixtures and processing “flutter” effects (i.e. a rising or falling surface, measuring out of focus) as
variables on visual and physical properties of the final fibreboard. All well as optical density fluctuations, e.g. on the fast moving conveyor belt,
designed experiments were done without adding any synthetic adhesive or were addressed. A control system ensured the functionality, performed
hydrophobic agent. Response Surface Methodology was used to visualise calibrations automatically and aligned the system with internal reference
fibreboard properties and to carry out multiple response optimisation. standards. The prototype was developed in close co-operation with the
industry. The on-line control system was complemented by a high level
multivariate data analysis including appropriate calibration models
(Kessler 2000).

242 243
Aderhold, Plinke
Metso Panelboard has used NIR technology as the basis for their Quality Control
PanelProTM system for monitoring important parameters in MDF
production. Engström (2008) of Casco Adhesives, Sweden, described real
time determination of formaldehyde emission from particleboards using 6.4.2 Panel Properties and Nuclear Spin
NIR spectroscopy. There is also some similar work on OSB (Taylor and The most commonly known application of NMR is in medicine where it
Via 2009). is used to represent and differentiate different kinds of tissue. Over the
Suggested R&D topics include: last years, it has become a more and more important tool for process
control and materials analysis. It works best for liquids, but can also be
 Adapting NIR reflectometry to harsh industrial environments applied for solids.
 Exploring potential applications of NIR Reflectometry in industry Since wood is an organic compound, it contains many of the
abovementioned nuclei with a spin of l=½, especially 1H (protons).
Consequently, wood gives a strong NMR signal. An important feature is
the fact that protons in different molecules such as water, wood
6.4 NUCLEAR MAGNETIC RESONANCE constituents, adhesives etc. give different NMR signals. In such a way it
is possible to measure the moisture content and the raw density of wood
6.4.1 Nuclear Spin or wood-based panels independently. Therefore, NMR has frequently
been used to monitor the drying of various wood products or to study the
Nuclear Magnetic Resonances (NMR) is a technique which is based on internal structure of wood. However, applications to the field of wood-
the fact that many atomic nuclei have intrinsic magnetic moments which based panels mostly refer to studies of adhesive curing. There are only
are characterized by the so-called magnetic quantum number m. It is few papers on non-destructive testing of wood based panels using NMR.
related to a nuclear property called spin which is characteristic for a given This is mainly due to the high cost of NMR equipment and to the limited
nucleus and which can take values that are integer multiples of one-half penetration depth of the technique. However, it can provide information
(0, ½, 1, 3/2, …). For a given spin, there are (2l+1) possible values of m not accessible by other techniques.
ranging from –l to +l in integer steps.
6.4.3 Applications of NMR in the Wood Based Panel Industry
Because of their relative simplicity, nuclei having a spin of l=1/2 such as
1H, 13C, 15N, 19F, 29Si und 31P are of particular importance for NMR In the wood sector NMR has been used mainly for fundamental research
since they can have only two magnetic states, namely ½ and +½. In the (drying of wood, reactions of adhesives in wood-based panels, cellular
absence of a magnetic field, these two states have the same energy. If, structure of wood etc.). There are only few publications concerning the
however, a magnetic field is applied, the energy of the states differs by an use of NMR for process monitoring and quality control in the wood-based
amount which is proportional to the magnetic field strength. Transitions panel industry. Possible applications are the measurement of raw density
between these states are possible by absorption or emission of and moisture content in particleboards, the estimation of the degree of
electromagnetic radiation with a characteristic frequency. curing of adhesives and the detection of adhesion defects.
The effective magnetic field at the nucleus can be different from the Among other quantities such as strength, stiffness, and dimensional
external field due to shielding by other nuclei and also by electrons. accuracy, the moisture content is an extremely important property of
Therefore, the transition frequency is different for different nuclei and logs, sawn timber, and wood-based products. For optimum processing
also for nuclei having a different chemical environment. Therefore, the conditions in the paper industry, the moisture variation of logs should be
nature of the nucleus and of the chemical bond of which it is a part can be as low as possible. Sorting the logs according to moisture content would
analyzed by measuring the transition frequency. This is generally done by be very helpful for a stable and economic production of paper products.
exciting the transition by a transient electromagnetic pulse and a The moisture content of sawn timber and wood-based panels has to meet
subsequent observation of resonance absorption. a certain target value in order to avoid problems with cracks, distortion,
and decay. It also influences the measurement of other important
properties including distortion. The lack of reliable, precise, and fast
techniques for moisture determination often causes wood products to fall

244 245
Aderhold, Plinke
Quality Control

short of the customer’s expectations. Finally, drying timber to the target

moisture is very energy intensive. A better control of this process will
diminish energy consumption. The precise estimation of moisture content
and of the raw density profile is very important in the wood-based panel
industry as well since they will have a strong effect on the practical value
of the product.
The moisture content can only be precisely determined if the local density
is also known. Unlike other techniques, NMR can measure these
quantities at the same time, giving more accurate and also more stable
results.
It is also possible to obtain moisture and density profiles perpendicular
to the surface and to achieve a certain lateral resolution.
This was demonstrated in a publication by Bloem et al. (1997). These
authors designed a special hand-held NMR device which works with one- Figure 6.17: Principle of simultaneous on-line moisture and density
sided access to the plate only (Figure 6.16). They were able to measure measurement (Wolter et al. 1997).
the moisture content and simultaneously the raw density of a 26 mm thick
particleboard (Figure 6.17). The system acquires NMR signals of a disc-
shaped measuring volume at a certain depth inside the board’s cross
section. The raw density measured by NMR was in excellent accordance
with data obtained from x-ray measurements (Figure 6.18).
Suggested R&D topics include:
 Adapting NMR to harsh industrial environments
 Increasing penetration depth by two-sided access

Figure 6.18: Typical raw density and moisture profile of a particleboard,

measured by NMR.

Figure 6.16: Hand-held NMR device for one-sided access.

246 247
Aderhold, Plinke
6.5 REFERENCES Quality Control

P. Bloem, D. Greubel, G. Dobmann, O. K. Lorentz, B. Wolter, NMR for

Kessler, W.: Multivariate Datenanalyse für die Pharma-, Bio- und
Non-Destructive Testing of Materials, Materials, functionality & design,
Prozessanalytik. Weinheim : Wiley-VCH, 2007
Proceedings of the 5th European Conference on Advanced Materials and
Processes and Applications. Vol. 4: Characterization and S. H. Nielsen, L. Jensen, Vollautomatisierte Produktion für Just-in-Time-
production/design : EUROMAT 97, Maastricht, NL, 21 - 23 April 1997 Lieferung, Schenck-Automatisierungssymposium für die Plattenindustrie,
Darmstadt, 13.-14. Oktober 1994, 1-6
L. Boltzmann, Ableitung des Stefan'schen Gesetzes, betreffend die
Abhängigkeit der Wärmestrahlung von der Temperatur aus der P. Niemz, F. Dutschmann, B. Stölken, Möglichkeiten zum Nachweis des
elektromagnetischen Lichttheorie, Annalen der Physik 22 , 291 (1884) Klebstoffanteils in beleimten Spänen, Holz als Roh- und Werkstoff 52, 6
(1994)
P. Deubel, Einführung moderner Qualitätssicherungssysteme in der
Möbelindustrie (1), Holz- und Kunststoffverarbeitung 27(9), 942 (1992) B. G. Osborne, T. Fearn, P. H. Hindle, Practical NIR Spectroscopy with
Applications in Food and Beverage Analysis, Longman Scientific and
P. Deubel, Einführung moderner Qualitätssicherungssysteme in der
Technical 1993
Möbelindustrie (2), Holz- und Kunststoffverarbeitung 27(10), 1104
(1992) M. Planck, Über eine Verbesserung der Wien’schen Spektralgleichung,
Verhandl. Dtsch. phys. Ges. 2, 202 (1900)
B. Engström, Real Time Determination of Formaldehyde Emission from
Particleboards Using NIR-Spectroscopy, International Panel Products B. Plinke, D. Ben-Yacov: Überwachung der Klebstoffverteilung im OSB-
Symposium 2008, 39 Vlies mit ortsauflösender Spektroskopie. 8. Holzwerkstoffkolloquium,
Dresden, 10.-11.12.2009
R. Feldhoff, T. Huth-Fehre, K. Cammann, Detection of inorganic wood
preservatives on timber by near infrared spectroscopy, J. Near Infrared J. Stefan, Über die Beziehung zwischen der Wärmestrahlung und
Spectrosc. 6, A171 (1998) Temperatur, Sitzungsberichte der Akademie der Wissenschaften II 79,
391 (1879)
G. Gaussorgues, Infrared Thermography, Chapman & Hall 1994
A. Taylor, B. K. Via, Potential of visible and near infrared spectroscopy
W. Gindl, A. Teischinger, M. Schwanninger, B. Hinterstoisser, The
to quantify phenol formaldehyde resin content in oriented strand board,
relationship between near infrared spectra of radial wood surfaces and
Eur. J. Wood Prod. 67, 3 (2009)
wood mechanical properties, J. Near Infrared Spectrosc. 9, 255 (2001)
L. G. Thygesen, S.-O. Lundqvist, NIR measurement of moisture content
D. I. Givens, J. L. De Boever, E. R. Deaville, The principles, practices
in wood under unstable temperature conditions. Part 1. Thermal effects in
and some future applications of near infrared spectroscopy for predicting
near infrared spectra of wood, J. Near Infrared Spectrosc. 8, 183 (2000)
the nutritive value of foods for animals and humans, Nutrition Research
Reviews 10, 83 (1997) J. Welling (Ed.), Directory of non-destructive testing methods for the
wood-based panel industry, Luxembourg : Off. for Off. Publ. of the
T. Hutter, R.W. Kessler, J. Zander, A. Ganz, Characterisation of modern
Europ. Communities, 1998
wet-process fibreboard production using experimental design, 7th Europ.
Panel Products Symp., (2003) pp. 14-25 W. Wien, Eine neue Beziehung der Strahlung schwarzer Körper zum
zweiten Hauptsatz der Wärmetheorie, Berliner Berichte, 55-62 (1893)
R.W. Kessler, T. Reinhardt, W. Kessler, H. Zimmer: Adaptive processing
of wood for fibreboards by high level spectroscopic on-line control. 4th B. Wolter, U. Netzelmann, G. Dobmann, O. K. Lorenz, D. Greubel,
Europ. Panel Products Symp. (2000), pp. 227-235 Kontrastierende 1H-NMR-Messungen in Aufsatztechnik zur Bestimmung
von Feuchteverteilungen in

248 249
 Wood Panel-Based Ceramics

Chapter 7

Carbon Materials and SiC-Ceramics made from

Wood-Based Panels
Olaf Treusch

CHAPTER SUMMARY

This chapter describes the manufacture process of monolithic porous

carbon materials from specific wood-based panels. These carbon
materials can serve among other things as precursors for SiC-Ceramics
which is also illustrated here.

7.1 INTRODUCTION

A new approach to technology is to develop dense or porous monolithic

carbon materials using cellulose containing preforms (e.g. wood,
bamboo or flax). These carbon materials can be used directly in various
applications (e.g. electrodes or structural materials) or be further
converted to silicon carbide (SiC) ceramics. Combustion chambers or
heat exchangers dominate in the ordinary field of applications for SiC-
ceramics. These materials (C and SiC) can be produced from a renewable
resource.
The process is capable of converting not only solid wood but also all its
derivatives such as wood-based panels or paper and cardboard. Compared
to conventional precursors (synthetic resins) for carbon materials or
SiC-ceramics, wood-based composites are low-cost materials, since they
are basically manufactured from low-cost raw materials such as wood
chips and particles. Moreover, in contrast to natural solid wood, the
isotropy of the wood-based materials is higher and their characteristics
are easily reproducible.
Wood-based materials have been on the market for decades and there is
wide range of products (e.g. oriented strand board, particleboard and
medium density fibreboard) with certain characteristics for particular
applications. These materials are generally suitable as precursors for
carbon materials or SiC-ceramics but they have to be modified in order
to attain specific properties. This is an additional advantage of wood-

250 251
Treusch
 Wood Panel-Based Ceramics
based materials, as it is possible to adjust their properties by selecting
specific components and by varying the processing parameters. properties can be influenced specifically by selecting particle shapes and
The manufacturing process of carbon materials and SiC-ceramics can sizes as well as densities and adhesive types to suit the purpose.
be described in three stages: the production of specific wood-based In recent years various projects have been aimed at the systematic
composites, thermal degradation to carbon materials (carbonization) investigation of the potential of specifically designed wood-based
and silicon infiltration to generate the ceramics (Figure 7.1). composites as preform for biogenous carbon materials and SiC-
processes materials ceramics. The studies have shown that one of the most important
parameters for the manufacture of wood-based materials is the particle
size of the wood component (Hofenauer et al. 2004; Treusch et al 2004;
production of specific
wood-based Herzog et al. 2006)
materials
 7.2.1 Raw Materials and the Manufacturing Process

For the fabrication of wood-based materials wood fibres, particles and

carbonization biogenous carbon powders (30-120 μm) have to be mixed with liquid or powdery resin in a
materials stirring device or a charge mixer. So far different phenolic resins and
pitch have been proven to be suitable for the subsequent thermal
treatment of the boards. In the previous experiments the adhesive content
was between 5 and 50 % related to the dry mass of the wood component.
silicon infiltration biogenous SiC- Depending on the bulk density desired, a specific amount of the wood
ceramics resin mixture has to be evenly distributed into a mould for
predensification. Finally the mixture has to be pressed to achieve the
thickness required and then heated to at least 120 °C to cure the resin. To
Figure 7.1: Manufacturing process of biogenous carbon materials and
avoid a density profile in the resulting board the densification has to take
SiC-ceramics. place at room temperature.

7.2.2 Characterisation
7.2 SPECIFIC WOOD BASED MATERIALS The specific wood-based materials can be characterized by means of
thermo-gravimetric analysis, light microscopy and mechanical testing.
Apart from solid wood, wood-based materials can serve as precursors for For the observation of the material by light microscopy they have to be
biogenous carbon materials and silicon carbide ceramics. Various embedded in acrylic resin and cut into 1 μm slides. As the density has an
studies have used conventional wood-based panels for the production of important impact on the properties of the resulting materials, especially
these materials. It could be shown that these are more appropriate starting on the possibility of infiltration, the density and the density profiles have
materials for these processes than solid wood (Kercher and Nagle 2002; to be measured.
Schmidt et al 2001; Krenkel et al. 1999) due to their reproducible
properties as well as their greater homogeneity and isotropy. As the Suggested R&D Topics include:
conventional wood-based panels are designed for completely different  Preceramic polymers as adhesives (e.g. silans, siloxans)
purposes the properties of the resulting carbon or ceramic material
respectively were not satisfactory for industrial utilization. However, one  Biogenous resins
of the advantages of wood-based materials is that desired structures or

252 253
Treusch
Wood Panel-Based Ceramics

 Moulding techniques to produce near net shape wood-based

enabled. A broad variety of specifically developed wood-based materials
composites (e.g. extrusion and injection moulding with duroplastic
were carbonized to produce crack free, monolithic porous carbon
resins)
materials. The results indicated how the wood-based material parameters
 Optimization of particle geometry (size distribution) affect the properties of the resulting carbon materials (Treusch et al.
2004).
 Applying additives (e.g. carbon fibres, chemicals)
7.3.1 Cabonization
 Homogeneous density distribution
The carbonization of the wood-based materials is carried out in an inert
atmosphere (e.g. N2). To make sure the carbon materials remain crack
free and without any deformations, a slow heating rate (approx. 1 K/min)
7.3 CARBON MATERIALS has to be applied up to 500 °C. Subsequently a higher heating rate
(approx. 5-10 K/min) can be applied up to the peak temperature (900-
Most industrial carbon materials are produced from organic compounds 1500 °C). During this process, the exact replica of the former
of petrochemical origin, which were heat-treated in an inert atmosphere. macroscopic structure of the wood-based composite remains (Figure 7.2)
Conventional starting materials for the production of porous carbons are (Treusch et al. 2004; Herzog et al. 2006).
phenol formaldehyde polymers, pitch, epoxy resins and furfuryl resins
(Fitzer et al. 1969; Constant et al. 1995; Liu et al. 1996; Czosnek et al.
2002; Puziy et al. 2002).
Monolithic porous carbon materials can also be obtained by carbonising
wood in a controlled thermal decomposition process. In this process the
cellular anatomic features of wood are retained in the new carbon material
(Byrne and Nagle 1997; Moor et al. 1974; McGinnes et al. 1971). Okabe
et al. (1996a) describe a method for the production of carbon materials
based on the carbonization of wood impregnated with phenolic resin.
The phenolic resin reinforces the material and prevents the development
of cracks. As the authors define ceramics as inorganic materials with
ionic or covalent bonds they call these materials “wood ceramics”. Figure 7.2: Fibreboard before (a) and after (b) carbonization.

Conventional wood-based panels can also serve as precursors for crack- Wood-based materials shrink and lose weight during carbonization.
free, monolithic porous carbon materials. Kercher and Nagle (2002) Similar to mass loss, dimensional changes are greatest at temperatures
used commercially available medium-density fibreboards (MDF) for the between 300 and 350 °C, which is consistent with the decomposition of
production of porous carbon materials. This investigation has shown cellulose in this temperature range (Shafizadeh, 1983). At 350 °C the
that the electrical, mechanical and structural properties of carbonised material lost almost 60 % of its weight as well as 25 % of its transverse
MDF materials make them excellent candidates for lithium-ion and fuel and 10 % of its plane dimensions. Shrinkage in the plane ranges between
cell components. 30 and 40 % (Figure 7.3). Mass loss for the different types of specific
wood-based materials lies between 55 and 80 %. Because of the fibre
One of the advantages of wood-based materials used as precursors for arrangement in the direction of the plane during hot-pressing, shrinkage in
porous carbon materials is, as already mentioned, the possibility to make the plane is less than in transverse direction.
adjustments for a wide range of macro- and micro-structures by selecting
different densities, types and amounts of adhesive as well as particle
sizes. Thus the tailoring of the properties of the materials derived thereof
in terms of structure (e.g. porosity, pore size) and mechanical properties is

254 255
Treusch
 Wood Panel-Based Ceramics

mechanical strength, thermal and electrical conductivity as well as their

microscopic structure have to be matched with each other (Pierson 1993).
Porous carbon materials can be used for example as filters, adsorbents,
electrodes and catalyst supports (Rodríges-Reinoso and Linares-Solano
1982; Barton et al. 1999; Burchell 1999). More recently biogenous
carbon materials have been found as precursors for composite materials
and for reaction-formed structural ceramics (silicon or titan carbide)
(Byrne and Nagle 1997; Zhang et al. 2003).
Figure 7.3: Dimensional changes during carbonization.
Suggested R&D Topics include:
 Reducing mass loss and dimensional changes during carbonization
7.3.2 Characterisation
 Improving shaping procedures for carbon materials
The structure of carbon materials can be analyzed by scanning electron
microscopy (SEM) or by incident-light microscopy after embedding and  Increasing the mechanical properties at high porosities
polishing. To investigate dimensional changes and mass loss, dimensions
and masses have to be measured before and after carbonization in their  Upscaling
dry state. Mechanical testing of the carbon materials is carried out by
determining flexural strength by the three point method (DIN 51902) or  Finally, dispersing small catalytic particles onto such conducting
high-area rigid supports is recommended
by determining compressive strength (DIN 51910-A). The specific
surface can be measured with a gas sorption analyser, and the BET
surface area can subsequently be calculated from the isotherms by the
Brunauer-Emmett-Teller (BET) equation. The electrical conductivity can
be measured according to the 4-probe method, applying low currents so 7.4 SIC-CERAMICS
as to avoid Joule heating. The conducting properties allow these materials
to be used as porous electrodes and light electromagnetic shields. Due to Silicon carbide (SiC) is a well-known industrial ceramic material
their rather high heat-treatment temperature (900-1400 °C), the carbon characterized by outstanding properties like high mechanical strength at
materials consist of almost pure carbon. Elemental analysis leads to the temperatures up to 1300 °C, and excellent thermal shock, wear and
following elemental massic composition: C 98 %, H 0.5 %, O 1.5 % corrosion resistance. Furthermore SiC shows high hardness at a low
(1000 °C). density (3,14 g/cm³) as well as low thermal extension. These properties
make the material suitable for high temperature technologies and
Apart from the internal structure, physical and mechanical properties are chemical instruments (Gadow 1986).
mainly influenced by wood-based material parameters such as particle
size, adhesive type, adhesive content and density. Apart from the sintering technique, conventional SiC-ceramics can be
manufactured by the infiltration of a porous preform consisting of
7.3.3 Potential Applications primary silicon carbide (SiC) powder and elemental carbon (C) with
molten silicon (Si). During the infiltration process the secondary SiC,
Carbon is the basis for a multitude of materials with a broad variety of formed by the reaction (Si + C  SiC), bonds the primary SiC grains, and
industrial applications. Potential fields of application for monolithic the residual pores are filled with elemental silicon. In silicon carbide
porous carbons can be found where properties like porosity, permeability, ceramics of common quality, the SiC content is in the range of 85 to
90 vol.-% leading to bending strengths between 300 and 500 N/mm²
(Cohrt 1985). There have also been several attempts to manufacture SiC-
ceramics without using primary SiC. Polymer-derived porous carbons,
fibre-reinforced carbon materials and carbon fibres have been used for

256 257
Treusch
Wood Panel-Based Ceramics

the infiltration with liquid silicon (Singh and Behrendt 1994; Krenkel
Silicon Infiltration
2000).
Liquid Infiltration Gaseous Infiltration
In recent years carbonized wood from various species including bamboo
subjected to pyrolysis were successfully infiltrated with liquid or gaseous
silicon or silica sol resulting in silicon ceramics. In these biomorphic
ceramics the structure of the biological precursors is preserved. A
combination is thus achieved of the wood structure with the chemical and
physical characteristics of the ceramic material (Byrne and Nagle 1997;
Martinez-Fernandez et al 2000; Singh and Salem 2002; Qiao et al. 2002;
Sieber et al. 2000; Arellano-Lopez 2004; Greil 2001; Kaindl 2000). Due
to its inhomogeneity, anisotropy and the difficulties involved in adjusting Si(l) + C(s) € Si(g) + C(s) € SiC(s)
SiC(s)

its characteristics, the use of solid wood as a precursor for technical Figure 7.4: Silicon infiltration into porous carbon materials (Sieber and
ceramics is limited. Commercial wood-based panels (e.g. veneer panels, Kaindl 1998).
particleboards) have been used as precursors for SiC-ceramics as these
materials show higher isotropy and easily reproducible characteristics In a further approach to convert the carbon materials into SiC-ceramics,
(Krenkel et al. 1999; Schmidt et al. 2001). the carbon-templates formed were infiltrated with various kinds of silica
sol (SiO²). The resulting SiO²/C composite is subsequently transformed
By optimizing the wood based panels with regard to their density, internal into a SiC-ceramic via carbothermal reduction (Klingner et al. 2003;
structure and adhesive content, SiC-ceramics were formed with a Herzog et al. 2006).
structure and mechanical properties (400 N/mm²) that are promising and
show a high potential for industrial use. The mechanical strength of the
7.4.2 Characterisation
resulting ceramic depends on its structure and phase composition, which
is correlated with the structure of wood-based composite and the carbon To determine the free carbon content of the ceramized materials, they can
template respectively. The strength rises with increasing silicon carbide be heated for 1 h at 1000°C in air in an oxidation furnace with a record of
content and decreases with residual silicon and carbon content as well as the weight change (Herzog et al. 2006) or by image analysis of the cross
with higher porosity (Hofenauer et al. 2003; Gahr et al. 2004; Hofenauer section. Taking into account the bulk density of the carbon template as
et al. 2006; Herzog et al. 2006) well as the carbon content and the bulk density of the resulting ceramic,
the phase composition of the ceramic can be calculated. The phase
7.4.1 Siliconisation composition declares the relative amount of SiC and residual carbon and
silicon. To analyse the microstructure of the ceramics scanning, electron
Silicon infiltration of the carbon templates takes place in vacuum microscopy can be applied. Three or four point bending tests can be
conditions at temperatures of approx. 1650°C and can be carried out in conducted to determine the strength of the ceramic material.
liquid or gaseous form (Figure 7.4). In the case of liquid silicon
infiltration, silicon reception takes place under vacuum by capillarity of
7.4.3 Potential Applications
the existing cavity of the carbon body – one part of silicon reacts and
another remains in the cavities. During gaseous silicon infiltration Si gas Examples of the ordinary field of application for monolithic SiC-
streams into the hollow spaces, while carbon and silicon combine to pure ceramics are combustion chambers, gas turbines, heat exchangers, seal
SiC. Hereby the porosity of the original material is retained. rings, valve discs and ceramic engine parts (Hofenauer et al. 2003). Due
Siliconization in liquid or gaseous form leads to no further dimensional to the melting point of the certain amount of residual silicon, the range of
changes, so that a shaped C-template can be converted near net shape to use is limited up to approx. 1380°C. Porous SiC-ceramics provide
SiC (Gahr et al., 2003; Hofenauer et al., 2003) interesting applications in the field of filtration of hot and corrosive media
and as perform material for infiltration with light metals to fabricate metal
matrix composites (Herzog et al. 2006).

258 259
Treusch
Suggested R&D Topics include: Wood Panel-Based Ceramics
 Optimization of the high temperature process
7.5 REFERENCES
 Composition of the wood-based material (additives, adhesives)
 Increasing the SiC content Arellano-Lopez, A.R.; Gonzales, P.; Dominguez, C.; Fernandez-Quero,
V.; Singh, M. (2004): Biomorphic SiC: A new Engineering Ceramic
 Increasing fracture toughness Material. Int. J. Appl. Ceram. Technol., 1(1), 56-67
 Upscaling Barton. J.T.; Bull, L.M.; Klemperer, G.; Loy, D.A.; McEnany, B.;
Misono, M.; Monson, P.A.; Pez, G.; Scherer, G.W.; Vartuli, J.C.; Yaghi,
O.M. (1999): Tailored Porous Materials. Chem. Mater., 11, 2633-2656
Burchell, T.D. (Hrsg.)(1999): Carbon Materials for Advanced
Technologies. Pergamon, Amsterdam
Byrne, C.E.; Nagle, D.C. (1997): Cellulose derived Composites - A
method for materials processing. Mater. Res. Innov., 1, 137-144
Cohrt, Z. (1985): Herstellung, Eigenschaften und Anwendung von
reaktionsgebundenem, siliciuminfiltrierten Siliciumcarbid. Z.
Werkstofftech., 16, 277
Constant, K.P.; Lee, J.-R.; Chiang, Y.-M. (1996): Microstructure
development in furfuryl resin-derived microporous glassy carbons. J.
Mater. Res., 11, 2338-2345
Czosnek, C.; Ratusek, W.; Janik, J.F.; Olejniczak, Z. (2002): XRD and
29
Si MAS NMR spectroscopic studies of carbon materials obtained from
pyrolyses of a coal tar pitch modified with various silicon-bearing
additives. Fuel Process. Technol., 79, 199-206
Fitzer, E.; Schaefer, W.; Yamada, S. (1969): The formation of glasslike
carbon by pyrolysis of polyfurfuryl alcohol and phenolic resin. Carbon, 7,
643-648
Qiao, R. Ma, N. Cai, C. Zhang, Z. Jin, "Mechanical Properties and
Microstructure of Si/SiC materials derived from native Wood," Mat. Sci.
Eng., A323, 301-305
Gadow, R. (1986): Die Silizierung von Kohlenstoff. Dissertation an der
Universität Karlsruhe
Gahr M., Schmidt J., Krenkel W., Hofenauer A., Treusch O. (2003): SiC-
Keramik auf der Basis von Holzwerkstoffen. In: Hans-Peter Degischer
(Editor) Verbundwerkstoffe: 14. Symposium Verbundwerkstoffe und
Werkstoffverbunde, Wiley Europe

260 261
Treusch Processing, Microstructure and Properties. Proc. of HT-CMC 5, Seattle
Greil, P. (2001): Biomorphic Ceramics from Lignocellulosics. J. Eur.
Gahr M., Schmidt J., Krenkel W., Hofenauer A., Treusch O. (2004): Ceram. Soc. 21, 105-118
Dense SiSiC ceramics derived from different wood-based composites:
Herzog, A.; Vogt, U.; Kaczmarek, O.; Klingner, R.; Richter, K.; Wood Panel-Based Ceramics
Thoemen, H. (2006): Porous SiC Ceramics Derived from Tailored Wood-
Based Fiberboards. J. Am. Ceram. Soc. 89(5), 1499-1503
Liu, Y.; Xue, J.S.; Zheng, T.; Dahn, J.R. (1996): Mechanism of Lithium
Hofenauer A., Treusch O., Tröger F., Wegener G., Fromm J., Gahr M., insertion in hard carbons prepared by pyrolysis of epoxy resins. Carbon,
Schmidt J., Krenkel W. (2003): Dense Reaction Infiltrated Silicon/Silicon 34, 193-200
Carbide Ceramics Derived from Wood Based Composites. Advanced
Engineering Materials, 5, 794-799 M. Singh, J. A. Salem (2002):Mechanical Properties and Microstructure
of biomorphic Silicon Carbide Ceramics fabricated from Wood
Hofenauer A., Treusch O., Tröger F., Wegener, G., Fromm J. (2004): Precursors. Journal of the European Ceramic Society, 22, 2709-2717
High Strength SiSiC Ceramics derived from fine Wood Powders. Proc. of
28 th International Cocoa Beach Conference and Exposition on Advanced Martineez-Fernandez, J.; Valera-Feria, F.M.; Singh, M. (2000): High
Ceramics & Composites, Cocoa Beach 26.-30. January Temperature Compressive Mechanical Behavior of Biomorphic Silicon
Carbide Ceramics. Scripta Mater., 43, 813-818
Hofenauer A., Treusch O., Tröger F., Wegener, G., Fromm J. (2006):
Silicon infiltrated silicon carbide ceramics (SiSiC-ceramics) derived from McGinnes, E.A.; Kandeel, S.A.; Szopa, P.S. (1971): Some structural
specific wood-based composites. Holz als Roh- und Werkstoff, 64, 165 – changes observed in the transformation of wood into charcoal. Wood
166 Fiber, 3, 77-83

Kaindl, A. (2000): Zellulare SiC-Keramik aus Holz. Dissertation, Moor, G.R.; Blankenhorn, R.; Beall, F.C.; Kline, D.E. (1974): Some
Universität Erlangen-Nürnberg physical properties of birch carbonized in a nitrogen atmosphere. Wood
Fiber, 6, 193-199
Kercher, A.K.; Nagle, D.C. (2002): Evaluation of carbonized medium-
density fibreboard for electrical applications. Carbon 40, 1321-1330 Pierson, H.O. (1993): Handbook of carbon, graphite, diamond and
fullerenes. Noyes/William Andrew Publishing, Park Ridge
Klingner, R.; Sell, J.; Zimmermann, T.; Herzog, A.; Vogt, U.; Graule,
Th.; Thurner, Ph.; Beckmann, F.; Müller, B. (2003): Wood-Derived Puziy, A.M.; Poddubnaya, O.I.; Gawdzik, B.; Sobiesiak, M.; Dziadko, D.
Porous Cermaics via Infiltration of SiO2-Sol and Carbothermal (2002): Heterogeneity of synthetic carbons obtained from polyimides.
Reduction. Holzforschung, 57, 440-446 Appl. Surf. Sci., 196, 89-97

Krenkel, W.; Hall, S.; Seitz, S. (1999): Biomorphe SiC-Keramiken aus Rodríges-Reinoso, F.; Linaras-Solano, A. (1982): Microstructure of
technischen Hölzern. DGM-Tagung „Verbundwerkstoffe und Activated Carbons as Revealed by Adsorption Methods. In: Thrower,
Werkstoffverbunde“, Hamburg 5.-7. Oktober P.A. (Hrsg.) Chemistry and Physics of Carbon. Marcel Dekker Inc. New
York, 1-146
Krenkel, W. (2000): Entwicklung eines kostengünstigen Verfahrens zur
Herstellung von Bauteilen aus keramischen Verbundwerkstoffen. Schmidt, J.; Hall, S.; Seitz, S. ; Krenkel, W. (2001): Microstructure and
Forschungsbericht 2000-04, Deutsches Zentrum für Luft- und Raumfahrt properties of biomorphic SiSiC ceramics derived from pyrolysed wooden
e.V. templates. Proceedings of the 4th Int. Conference on High Temperature
Ceramic Matrix Composites (HTCMC4), München 1.-3. Oktober
Shafizadeh, F. (1983): The chemistry of Pyrolysis and Combustion. In: R.
Rowell (Hrsg.), The Chemistry of Solid Wood, American Chemical
Society, Seattle
Sieber, H.; A. Kaindl (1998): Biomimetik: Der Weg vom Holz zur
Keramik. Leaflet of FAU Erlangen
Sieber, H.; Hoffmann, C.; Kaindl, A.; Greil, P. (2000): Biomorphic
Cellular Ceramics. Adv. Eng. Mater., 2(3), 105-109

262 263
Treusch Sci. Eng., A187, 183-187
Treusch O., Hofenauer A., Tröger F., Fromm J., Wegener G. (2004): Basic
Singh, M.; Berendt, D.R. (1994): Microstructure and mechanical Properties of Specific Wood-Based Materials Carbonised in a Nitrogen
properties of reaction-formed silicon carbide (RFSC) ceramics. Mater.
Atmosphere. Wood Science and Technology 38, 323-333 Modified Wood-Based Panels
Zhang, D.; Xie, X.Q.; Fan, T.X.; Sun, B.H.; Sakata, T.; Mori, H.; Okabe,
T. (2003): Microstructure and properties of ecoceramics/metal composites Chapter 8
with interpenetrating networks. Mater. Sci. Eng,. A351, 109-116
formed wood fiber/polymer fiber composites. Forest Products Journal, Thermally and Chemically Modified Wood-Based
42(6), 42-48 Panels
Wulf Paul and Martin Ohlmeyer

CHAPTER SUMMARY

The modification of wood aims to alter the absorptive behaviour of wood

and thus gives an advantage in terms of dimensional stability. The two
main methods of wood modifcation are chemical and thermal treatments;
both result in an irreversible change of the wood cell components. The
polyoses as main carrier of hydroxyl groups contribute most to moisture
uptake and thus to the risk of dimensional changes.
The principle of a chemical modification is to replace the hydroxyl groups
by a reagent; most study has undergone the substitution by acetyl groups
referred to as acetylation of wood. Any chemical modification process
results in a weight increase and keeps the wood in its swollen state. In
contrast, thermal modification is always accompanied by a weight loss.
The principle is to remove hydroxyl groups by thermal degradation thus
leaving wood in its shrunken state. Wood modification has been subject
to numerous studies and thermally modified solid wood became
commercially viable whereas chemical modification processes often lack
economical advantages due to high process and chemicals costs.
Increasingly, over recent years the focus is on applying those methods
described on wood based panels as well. But despite promising results no
commercial viabilty could be achieved yet.
This article gives an overview about different modification processes,
their advantages but also negative impacts on board properties and
highlights the limits in practical and economical terms.

8.1 INTRODUCTION

Wood modification implies all those treatments which result in a change

of wood properties. Mostly, modification processes aim to change the
hydrophilic nature of wood. Since the hygroscopicity of wood is
accompanied by anisotropic changes in dimension, swelling and shrinking

264 265
Paul, Ohlmeyer thus provoke constructional failures (Lukowsky and Böttcher 2001, Niemz
2002). Hence, especially for timber construction (e.g. window frames) high
may result in a loss of dimensional stability. Furthermore, crack dimensional stability is needed.
formation might occur as a consequence of the dimensional changes and
For wood-based panels (WBP) anisotropic changes in dimension are Modified Wood-Based Panels
minor, at least across the width and the length of a board. The limitation
of common WBP if exposed to high relative humidity or frequent
moisture changes (e.g. by weathering) is due to their tendency of here, which lead to a chemical change of the cell wall components. Since
thickness swell, i.e. a change of dimension across the thickness of a the main target of wood modification is to limit the moisture uptake, any
modification method aims to alter the hydroxyl groups as they contribute
most to the absorption of water; Figure 8.1 gives a schematic overview of
different mechanisms.
In the following, two methods will be described: The mechanism behind
thermal modification is the degradation of hydroxyl groups, while the
hydroxyl groups in the wood's polysaccharides form covalent bonds with
the chemical reagent during chemical modification.
board. In contrast to solid wood, a change in dimension of WBP consists Blocking CroSS-linking
of two elements, a reversible and an irreversible one. While the
reversible component is only due to the hygroscopic behaviour of the raw MMF −CH2−
material itself, the irreversible component can be traced back to the
production process (Adcock and Irle 1997). − OH HO− − O− -O−

During hot-pressing the particles are compressed and thus stresses are HCHO/
induced into the mat. Subsequently, if a panel is exposed to high humidity MELAMINE RESIN Swelling DMDHEU
or gets in contact with water, these stresses are released and thus lead to H2O
an irreversible increase of thickness which exceeds the reversible
component by magnitude.
Thickness swell not only has a negative impact on the appearance of the − OH HO−

panel, but also results in a decrease of its mechanical properties. If stress Grafting
build-up exceeds the strength of the adhesive bonds between the particles, Degradation
internal bond strength decreases. The stresses that develop inside the
wood due to wetting are very high (Tarkow and Turner 1958) which no −R−CH3
− OH
adhesive can permanently withstand. Therefore the adhesive bonds in a
− O− HO−
WBP in its swollen state are damaged if not broken, but stressed at the HO−
very least.
ACETYLATION HEAT
Thus, an improvement of WBP in terms of reduced thickness swell may
be facilitated by both, a change of the hygroscopic nature of wood but
also by an improved compressibility of the raw material by altering its Figure 8.1: Mechanism of different modification methods to reduce
visco-elastic properties. swelling of the wood cell wall, (adapted from Rapp et al.
2000).
For WBP principally two ways of modification exist: Either a
modification of the raw material prior to blending and hot-pressing or a
modification of the panel. In the following an overview of different 8.2 CHEMICAL MODIFICATION
modification methods is given.
As mentioned above, improving dimensional stability of WBP requires Chemical modification of wood has early been recognized as an effective
both, a reduced moisture uptake as well as a change of visco-elastic way to reduce its hygroscopicity and to improve its dimensional stability
properties. Hence, only those modification methods will be considered (Stamm 1964).

266 267
Paul, Ohlmeyer chemical configuration of the cell wall components is changed, resulting in
altered wood properties. On the one hand, chemical modification leads to
Rowell et al. (1988) defined chemical modification as any reaction improved dimensional stability and to increased resistance against fungal
between the abundant hydroxyl groups in the wood's polysaccharides attack and weathering. On the other hand there is a .reduction in tensile
with single chemical reagents, forming covalent bonds. Thus, the strength and elasticity.
The aim of any modification process is to maintain the mechanical Modified Wood-Based Panels
properties of wood while improving dimensional stability. Hence, wood
modification always is a compromise between improved properties and
reduced strength. Several studies reported in the literature deal with WBP made of
chemically modified particles (fibres, particles, strands). The modification
In accordance with Rowell (1975) suitable reagents for chemical results in decreased TS and reduced moisture uptake. Sudo (1979)
modification should contain functional groups to form covalent bonds investigated fibreboards made from acetylated fibres and found a
with hydroxyl groups. Furthermore it is necessary for the reagent to swell reduction of TS after 24 hours of water immersion. Arora et al. (1981)
the wood in order to allow access to all reactive sites. The process should used a catalyzed liquid-phase process to acetylate particles. The PF-
be carried out under mild conditions at temperatures up to 120 °C in order bonded panels made of acetylated particles showed considerably lower
to allow simple and fast reactions. The catalysts used should be mildly thickness swell at high relative humidity than boards made of non-
alkaline with a low boiling point to avoid corrosion. acetylated particles.
The mechanism of chemical modification is to replace hydroxyl groups in Acetylated flakeboards bonded with PF-resin show lower moisture
lignocellulosic materials. The reaction is accompanied with an increase of content, improved dimensional stability and improved decay resistance
weight; the degree of modification is expressed as weight percent gain (Hadi 1992).
(WPG). To enhance the degree of hydroxyl group substitution, the
reaction ideally follows a single site reaction, i.e. a reaction between the This coincides with work about acetylation of strands made of fir to
reagent and the hydroxyl group in a molar ratio of 1:1. Different reagents improve the dimensional stability of OSB (Papadopoulos and Traboulay
are used for chemical modification, e.g. anhydrides, epoxides, 2002). Ring-cut strands were acetylated at 120 °C using acetic anhydride.
isocyanates, acid chlorides, carboxylic acids, lactones, alkyl chlorides, Test panels bonded with PF-resin showed significantly lower TS and
and nitriles. Among all types of chemical modifications, the substitution water absorption. A chemical modification of wood chips using propionic
of hydroxyl groups by acetyl groups referred to as acetylation has been anhydride was performed by Papadopoulos and Gkaraveli (2003);
studied the most. A schematic overview about the reaction of acetic modification resulted in improved dimensional stability of particleboards
anhydride with hydroxyl groups of the wood is shown in Figure 8.2. manufactured with these chips. The TS values of the UF bonded boards
were more than 50 % lower than controls.
wood-OH (CH -CO) O wood-OOC-CH 3 CH 3-COOH
Beside reduced TS and water absorption, chemical modification is often
accompanied with lower strength properties. Youngquist et al. (1986) and
Rowell et al. (1987) found for particleboards made of acetylated aspen
particles and bonded with PF-resin a reduction in MOR of about 34 %.
MOE was reduced by about 11 % and IB was decreased about 36 %,
compared with panels made of non-treated flakes. Hardboards made from
acetylated hemlock fibres and bonded with 7 % phenol-formaldehyde
adhesive were tested by Youngquist et al. (1990). In static bending, MOR
was reduced by 23% and MOE by 16 % as compared to control boards of
non-acetylated fibres. Tensile strength parallel to the surface was reduced
by
the5surface
% but in
there was no change
acetylated in the tensile
boards compared strength perpendicular to
to controls.
+ 3 2 = + Rowell et al. (1991) tested fibreboards made from acetylated aspen fibres,
Acetic anhydride Acetylated wood Acetic acid
using 8 % phenol-formaldehyde resin, in static bending. They found that
Figure 8.2: Acetylation of wood. MOR increased by 15 % and MOE increased by 40 % in acetylated
fibreboards compared to controls. According to the authors, the acetylated
The chemical modification of fibres or particles prior to hot pressing boards had a more uniform density and a more consolidated surface as
rather than modification of entire panels can be regarded a relatively compared to controls.
effective method because the raw material provides a large surface area,
allowing the reagent to easily penetrate the wood.

268 269
Paul, Ohlmeyer 1986, Fuwape and Oyagade 2000, Papadopoulos and Traboulay 2002,
Papadopoulos and Gkaraveli 2003). From these studies it is apparent that IB
Concerning the bonding behaviour as determined by internal bond decreases with increasing WPG. By means of electron micrographs Rowell
strength (IB), chemical modification leads to a decrease, at least if bonded et al. (1987) showed that fragmentation of the wood matrix increased with an
with formaldehyde resin systems (Chow et al. 1996, Youngquist et al. increasing level of acetylation, i.e. with higher WPG. The authors did not
attribute these defects to glue line failures since these defects were found Modified Wood-Based Panels
only in the outermost layer of surface. Besides, their studies showed that
failures during test occurred more often in the glue line for panels made of
acetylated flakes compared to panels of non- acetylated flakes. All of the procedures to acetylate wood developed over the years are
complicated reaction schemes, using either a catalyst or an organic
In considering the work of Youngquist et al. (1986), Rowell et al. (1987) cosolvent and have required long reaction times.
considered that the press pressures required for flakeboards led to glue
line failures. The reduced moisture content of the acetylated flakes made Therefore, Rowell et al. (1986a, 1989) developed a non-catalyzed liquid
them less compressible and hence higher press pressures were needed. phase acetylation process in order to achieve a modification of large
batches to facilitate industrial processing. This method was investigated
Another explanation might be the low wettability of modified wood in several studies; panels made from acetylated fibres showed reduced
particles, and therefore poor penetration of water soluble formaldehyde moisture uptake and water absorption as well as decreased TS. However,
resins into the particles (Papadopoulos and Traboulay 2002, significant improvements required more than 10 % WPG, based on dry
Papadopoulos and Gkaraveli 2003). wood mass (Chow et al. 1996, Rowell et al. 1986b–d, 1991, Tillman et al.
1987, Vick et al. 1991, Youngquist et al. 1986).
Youngquist and Rowell (1990) as well as Papadopoulos et al. (2005)
showed that using an isocyanate resin to bond the particles results in a In terms of an industrial application, this means a great amount of
much lower difference in IB between modified and non-modified chemical loading and thus a cost-intensive process as well.
particles. According to Papadopoulos et al. (2005) this might be due to
the pH independent character of isocyanate resin as well as that it Sheen (1992) reported about the production of acetylated fibres with a
completely cures during hot pressing. Another possible explanation is the relatively simple, no solvent or catalyst process at a commercial pilot
high mobility of isocyanate resins into the wood surface which causes plant. Although the acetylation results were promising, full scale
penetration to considerable depth into compressed particles and can result production was not realised because of high production costs as
in their total impregnation (Roll 1997). mentioned above. Consequently, the production costs had to be
compensated by high prices of the final panel product.
Therefore it is suggested that the isocyanate resin system is more suitable
for use in boards made from modified raw material than the formaldehyde
resin system.
Despite such research efforts, no commercial applications have yet been 8.3 THERMAL MODIFICATION
fully realised for the acetylation of wood. First attempts, one in the United
States in 1961 and one in the former USSR in 1974, came close to Another approach to alter wood properties is thermal modification. The
commercialisation but were discontinued, presumably because they were target of this procedure is the same as for chemical modification, i.e. to
not cost-effective as described by Rowell et al. (1986a). change the hygroscopic nature of wood. Contrary to chemical
modification the mechanism of action is to degrade hydroxyl groups
instead of substitute them by a chemical reagent. The chemical loading
keeps the wood in its swollen state and is accompanied with an increase
in weight. A thermal modification results in a weight loss and the wood is
kept in its shrunken stage.
Thermal modification in the typically applied temperature range of 180–
250 °C leads to a degradation of hydroxyl groups following the
mechanism of an acidic hydrolysis; Figure 8.3 gives an overview about
the basic chemical changes that the main cell wall components undergo
during thermal degradation. There are numerous studies dealing with the
mechanism behind the decomposition of wood components due to heat
treatment, e.g. Kollmann and Fengel (1965), Bourgois and Guyonnet
(1988), Garrote et al. (1999).

270 271
Paul, Ohlmeyer their structure is rather amorphous with a high portion of hydroxyl groups.
Thus, they contribute most to the sorption of water but also are sensitive to
The modification firstly affects the thermal instable polyoses. In hydrolysis. For cellulose, a thermal modification only affects the amorphous
comparison with cellulose, which is less affected by thermal degradation, regions within the otherwise crystalline structure, resulting in an increased
degree of crystallinity. Thermal degradation proceeds in three steps, Modified Wood-Based Panels
starting by deacetylation. The released acetic acid acts as a catalysing agent
which then accelerates the depolymerisation of the polysaccharides.
Alongside the depolymerisation, the heat leads to dehydration of the But also thermal modification of WBP to improve dimensional stability
monosaccharides. Depending on their ring structure (furanosidic or has been investigated for more than 40 years and still is of interest (e.g.
pyranosidic), either furfural (out of xylose) or 5- hydroxymethylfurfural Matsu and Sasaki 1956, Lehmann 1964, Heebink and Hefty 1969, Shen
(out of glucose) are formed (Bobleter and Binder 1980). Especially the 1973, Burmester 1974, Tomimura and Matsuda 1986, Hsu et al. 1988,
former substance is very reactive and leads to cross linking reactions with Subyanto et al. 1991, Sekino et al. 1997, Goroyias and Hale 2002a,
the lignin complex. At a temperature of about 200 °C the lignin itself Ohlmeyer and Lukowsky 2004, Paul and Ohlmeyer 2005, Paul et al.
undergoes a demethoxylation, leaving reactive sites which reticulate by 2006).
auto condensation forming a hydrophobic network. Since the rate of hydrolysis is accelerated by temperature and pressure,
mostly thermal modification is carried out under pressure and moist
conditions. Most investigations deal with steam pre-treatment processes
where pressures between 5 to 10 bar are applied, and temperature ranges
from 160 °C to over 200 °C. Generally, increasing treatment temperature
and time is very effective for improving dimensional stability, i.e.
reducing TS. However, Sekino et al. (1998) found a significant reduction
of the bond strength if temperatures above 200 °C are applied. This might
be explained by the partial conversion of the polyoses into furfural
polymers. This results in an increased embrittlement and a significant
reduction of shear strength (Stamm 1964).
Boonstra et al. (2006) applied a two-stage heat pre-treatment according to
Polyoses Cellulose Lignin the commercially established Plato-process for modifying solid wood
(Boonstra et al. 1998, Tjeerdsma et al. 1998). The temperature was below
Increased degree of
200 °C and the process was performed in two separate stages with an
Deacetylation Release of free molecules
crystallinity intermediate drying stage. In the first stage of the heat-treatment (hydro-
thermolysis) the chips were treated in an aqueous environment at
Depolymerisation Condensation
superatmospheric pressure with saturated steam (8–10 bar). In the second
stage, after drying, the chips were heat treated in a kiln under dry and
atmospheric conditions at 180 °C (curing-treatment). During this stage
Dehydration Cross linking superheated steam or nitrogen was used as a shielding gas to exclude
oxygen in order to reduce fire risks and preventing undesired oxidation
Formation of formaldehyde reactions. The MUF-bonded particleboards made from modified chips
and furfural show reduced thickness swell; best results were obtained if chips were
only hydro-thermolysed (first stage) without the curing step. The bonding
behaviour as determined by IB decreased.
Figure 8.3: Decomposition of the main cell wall components due to
thermal degradation (adapted from Esteves and Pereira All the treatment methods mentioned above have in common that heated
2009). steam is used and pressure has to be applied. That means in terms of an
industrial upscale extensive reconstruction of the panel production
Heat treatment of solid wood is well known for decades and published in process.
several works (e.g. Stamm et al. 1946, Seborg et al. 1953, Kollmann and
Schneider 1963, Stamm 1964, Kollmann and Fengel 1965, Noack 1969, Tomek (1966) at first discussed the possibility to using the drying process
Burmester 1973, Burmester 1975, Giebeler 1983, Hillis 1984, Bourgois to modify wood particles. The process was performed with a dryer heated
and Guyonnet 1988). by smoke gas; the oxygen content was less than 15 %. The moisture
content of the particles prior to drying was set to 40 % and treatment was

272 273
Paul, Ohlmeyer
Modified Wood-Based Panels

carried out at a maximum temperature of 300 °C for four minutes.

adhesives. An example of particleboard made of thermally modified chips
Although this process had potential to modify wood particles by an
of spruce by the method described is shown in Figure 8.5.
industrial drying process it is limited to laboratory scale. Since the dryer
Tomek (1966) used was static it was necessary to spread out the wood
particles in an equal layer thickness. That was important in order to
achieve a homogenous temperature distribution within the wood particles.
In terms of an industrial upscale this requirement is hardly to fulfil.
Investigations using a rotary dryer (at laboratory scale) allow thermal
modification of larger batches and thus more potential to industrial
upscale (Paul and Ohlmeyer 2005, Paul et al. 2006). The authors applied a Figure 8.5: Particle board made of untreated and thermally modified
chips of spruce (Picea abies).
one-step treatment of wood particles (strands, chips) with an initial
moisture content of app. 5 %; Figure 8.4 displays particles of spruce
(Picea abies) before and after the process; the modified chips show the
typical discoloration following a thermal degradation. This effect is 8.4 MODIFICATION OF PANELS
linked to coloured degradation products of the polyoses due to the
hydrolytical decomposition which is similar to a Maillard reaction Steam treatments can also be effective at improving the dimensional
(Sehlstedt-Persson 2003). But also extractives are believed to contribute stability of densified WBP. Besides pre-treatment by steaming of the
to the colour change of heat-treated wood (Mcdonald et al. 1997, wood furnish before forming a mattress (see above) a number of papers
Sundqvist and Morén 2002). have been published on treatments during (simultaneous) and after hot
pressing. Two different approaches may be distinguished: steam injection
The process was carried out under atmospheric pressure in a dry and pressing in which steam not only treats the wood but affects adhesive cure
oxygen reduced atmosphere to avoid ignition. The reduction of oxygen as well (e. g. Shen 1973; Subyanto et al. 1991), and post-treatment
was achieved by temporary inducing spray of water into the treatment steaming of a panel after hot-pressing (e. g. Heebink and Hefty 1969).
chamber; besides, gaseous reaction products discharged remaining
oxygen (oxygen content in the dryer was less than 15 %). Hence, the Another approach to apply a simultaneous treatment was presented by
treatment could be performed at temperatures above 200 °C to allow short- Goroyias and Hale (2002b) using a hot press without steam injection. The
time modification. authors investigated the effect of elongated pressing times on the
dimensional stability of PF-bonded OSB. The results show that thickness
swell decreased with rising press temperature and pressing time. At the
same time internal bond strength increased. The bending strength on the
other hand decreased due to the treatment.
A post steam treatment of PF-bonded panels is an effective method for
improving dimensional stability without significant detriment to the
mechanical properties (Heebink and Hefty 1969). The principle is based
Figure 8.4: Chips of spruce (Picea abies) before (left) and after on the plastification of wood, resulting in a fixation after cooling and thus
thermal modification. preventing thickness swell. Another approach is presented by Del
The panels (bonded with formaldehyde-based resins and isocyanate, Menezzi and Tomaselli (2006). In their work they describe the effect of a
respectively) showed improved dimensional stability (thickness swell simple heating of PF-bonded panels in a hot press without applying
reduced by 50 %). Depending on the panel type (particleboard, OSB) the pressure. They found a reduction of thickness swell with increasing
decrease of internal bond strength was minor for panels bonded with heating time.
isocyanate. But the investigations have shown that pre-treatment of wood
furnish prior to blending and pressing is not restricted to specific types of A post treatment in terms of an industrial production is provided by
storing the panels in stacks after hot pressing. The stacking is used for

274 275
Paul, Ohlmeyer
conditioning the panels, mainly to achieve a unified temperature and Modified Wood-Based Panels
moisture distribution in the horizontal and vertical planes of the panel.
Another important effect is that stacking influences the properties of the
final product. By varying the temperature of the stack and the duration of Due to the improved dimensional stability it appears to be possible from a
storage panel properties are changed. With increasing temperature and general point of view to consider using modified wood-based panels for
duration of storage a reduction of thickness swell can be achieved any of those applications where moisture resistance is required, e.g.
(Ohlmeyer 2002). flooring or cladding.

Although simultaneous and post treatments improve dimensional However apart from moisture resistance and dimensional stability, boards
stability, these modification methods are restricted to adhesive systems for these fields of application also need to fulfil certain standards
which are resistant to hydrolysis, such as isocyanates and PF resins. In regarding the strength, no matter whether those standards are of
terms of an industrial upscale, simultaneous treatment is considered to be normative or industrial character.
less expedient since hot pressing is known to be the bottle neck of the The main objection to using modified wood based panels as a substitute
panel production process. That means increasing pressing time is not for common ones is the loss of mechanical strength; while the effect on
reasonable from an economical point of view. A post treatment by storing MOE is minor, the internal bond and particularly bending strength are
panels in a stack at elevated temperatures for several hours is difficult to affected more. Especially the degradation of the polyoses has major
realise. Due to the dimension of industrially produced panels it is time impact on the strength loss. Apart from being the main source for sorption
consuming and thus cost-intensive to achieve a homogenous temperature of water, the polyoses also contribute to the visco-elasticity of wood
distribution within the panel. Furthermore, post treatments might result in which provides dynamic strength.
shrinkage of the wood particles, affecting the wood-adhesive bond
network (e.g. breakages of glue lines). To a certain degree, the strength loss could be compensated by higher
resin levels; but some treatment methods may even require a different
Post treatment using chemical modification was performed by Klinga and resin system to allow for a sufficient glue joint or simply to withstand the
Tarkow (1966). The authors acetylated non-heat-treated wet-process modification itself, e.g. when applying a post treatment.
hardboard by an uncatalyzed vapour phase process. The acetylation
process resulted in board swelling and surface roughening, but it reduced Another disadvantage is the currently high cost impact on the process in
reversible and irreversible thickness swell in subsequent water soak tests. terms of thermal energy or chemicals. WBP are relatively low-value
products and so cannot support high price increases. Therefore, modified
Murphy and Turner (1989) investigated the vapour phase preservative wood-based panels would have to be used for special applications and
treatment of manufactured wood-based board materials including niche products for which the extra cost of manufacturing will pay off.
particleboards, MDF, aspen wafer boards and OSB using esterified borate
in the vapour phase to obtain complete impregnation. Chemical A lot of research has been done and still is ongoing in the field of wood
modification applications to fibre boards can reduce water absorption and modification with the focus on applying successful methods to wood
swelling of wood fibre and polymer fibre composites (Youngquist et al. based panels. The outcome is promising and further research should
1992). investigate how modification methods could be altered in order to benefit
from an improved dimensional stability but maintain the mechanical
properties which are crucial for the final application. Also, the question of
8.5 CONCLUSION economical viability needs to be addressed and ways have to be found to
integrate the modification using existing production techniques.
Modification of wood, regardless of which method is applied, is an
effective way of improving dimensional stability and durability. There is
a wide field of applications for which modified solid wood is already in
use, e.g. deckings, claddings, parquet, garden furniture, to name just a
few. But unlike for solid wood there is currently no marketable
application for modified wood-based panels.

276 277
Paul, Ohlmeyer grain. In: Proceedings of The First European Panel Products Symposium,
Llandudno, 9-10 October, 1997
8.6 REFERENCES Arora, M.; Rajawat, M.; Gupta, R. (1981): Effect of acetylation on properties
of particle boards prepared from acetylated and normal particles of wood.
Adcock, T. and Irle, M.A. (1997): The effect of compaction ratio on the Holzforschung und Holzverwertung, 33(1), 8–10.
dimensional recovery of wood particles pressed perpendicular to the
Bobleter, O.; Binder, H. (1980): Dynamischer hydrothermaler Abbau von
Holz. Holzforschung 34: 48-51 Modified Wood-Based Panels
Boonstra, M.J.; Tjeerdsma, B.F.; Groeneveld, H.A.C. (1998): Thermal
modification of non-durable wood species. In: Part 1. The Plato Esteves, B.M.: Pereira, H.M. (2009): Heat treatment of wood. Bio
technology: thermal modification of wood. International Research Group Resources, 4(1), 370–404
on Wood Preservation, Document no. IRG/WP 98-40123
Fuwape, J.A.; Oyagade, A.O. (2000): Strength and dimensional sta-bility
Boonstra, M.J.; Pizzi, A.; Zomers, F.; Ohlmeyer, M; Paul, W. (2006): The of acetylated Gmelina and Sitka spruce particleboard. J. Tropical Forest
effects of a two stage heat treatment process on the properties of Prod., 6(2), 184–189
particleboard. Holz als Roh- und Werkstoff, 64(2), 157-164
Garrote, G.; Dominguez, H.; Parajo, J.C. (1999): Hydrothermal
Bourgois J, Guyonnet R (1988) Characterization and analysis of torrified processing of lignocellulosic materials. Holz als Roh- und Werkstoff 57:
wood. Wood Sci Tech, 22, 143–155 191-202
Burmester, A. (1973) Einfluss einer Wärme-Druck-Behandlung Giebeler, E. (1983): Dimensionsstabilisierung von Holz durch eine
halbtrockenen Holzes auf seine Formbeständigkeit. Holz Roh- Werkst, Feuchte/Wärme/Druck-Behandlung. Holz Roh- Werkst, 41, 87–94
31, 237–243
Goroyias, G.J.; Hale, M.D. (2002a): Heat treatment of wood strands for
Burmester, A. (1974): Erfolgreiche Quellungsvergütung mit einfachen OSB production: Effect on the mechanical properties, water absorption
Mitteln, part 1 and 2. Holz- und Kunststoffverarbeitung, 8 and 9, 534- and dimensional stability. International Research Group on Wood
538 and 610-617 Preservation, Document no. IRG/WP 02-40238
Burmester A (1975): Zur Dimensionsstabilisierung von Holz. Holz Roh- Goroyias, G.J.; Hale, M.D. (2002b): The Effect of High Temperature and
Werkst, 33:333–335 Long Pressing Time on the Dimensional Stability and Decay Resistance
of OSB. Paper prepared for the 33rd Annual Meeting of the International
Bourgois, J.; Guyonnet, R. (1988): Characterization and analysis of Research Group on Wood Preservation in Cardiff, Wales
torrified wood. Wood Science and Technology 22: 143-155
Hadi, Y.S. (1992): Acetylated flakeboard properties. In Chemical
Chow, P.; Bao, Z.; Youngquist, J.A.; Rowell, R.M.; Muehl, J.H.; Krzysik, modification of lignocellulosics, Rotorua, New Zealand, 7-8 November
A. (1996): Properties of hardboards made from acetylated aspen and 1992 (compiled by Plackett D.V. & Dunningham E.A.). Forestry
southern pine. Wood Fiber Sci., 28(2), 252–258 Research Institute, Ministry of Forestry FRI Bulletin, 176, 9-15
Del Menezzi, C.H.S.; Tomaselli, I. (2006): Contact thermal post- Heebink, B.G.; Hefty, F.V. (1969): Treatment to reduce thickness
treatment of oriented strandboard to improve dimensional stability: A swelling of phenolic bonded particleboard. Forest Products Journal
preliminary study. Holz als Roh- und Werkstoff, 64(3), 212-217 19(11), 17-29
Hillis, W.E. (1984): High temperature and chemical effects on wood
stability. Wood Science and Technology, 18, 281-293
Hsu, W.; Schwald, W.; Shields, J.A. (1988): Chemical and physical
changes required for producing dimensionally stable wood-based
composites. Part 1. Steam pre-treatment. Wood Sci Tech 22, 281–289
Klinga, L.; Tarkow, H. (1966): Dimensional stabilization of hardboard by
acetylation. Tappi, 49(1), 23–27
Kollmann F, Schneider A (1963): Über das Sorptionsverhalten
wärmebehandelter Hölzer. Holz Roh- Werkst, 41, 87–94

278 279
Paul, Ohlmeyer Lehmann, W.F. (1964): Retarding dimensional changes in particleboards.
Information Circular 20 Oregon State University, Corvallis
Kollmann, F.; Fengel, D. (1965): Änderungen der chemischen Lukowsky, D.; Böttcher, P. (2001): Beschichtung von hitzevergütetem Holz.
Zusammensetzung von Holz durch thermische Behandlung. Holz Roh- (Technischer Bericht 2323/2001. In: Burk, R. (2002): Evaluierung
Werkst, 21(3), 77–85 technischer Eigenschaften von hitzevergütetem Holz unter besonderer
Berücksichtigung der Entwicklung hochwärmegedämmter Holzfenster,
Abschlussbericht des FuE-Vorhabens Kap. 0810, Titel 682 01-04. FH Modified Wood-Based Panels
Kaiserslautern/ Pirmasens). 31 Seiten
Maku, T.; Sasaki, H. (1956): Some factors which affect swelling Paul, W.; Ohlmeyer, M.; Leithoff, H.; Boonstra, M.J.; Pizzi, A. (2006):
properties of chipboard. Composite Wood, 3, 135-139 Optimising the properties of OSB by a one-step heat pre-treatment
process. Holz als Roh- und Werkstoff, 64(3), 227-234
McDonald, A.; Fernandez, M.; Kreber, B. (1997): Chemical and UV-VIS
spectroscopic study on kiln brown stain formation in Radiata pine. In: 9 th Papadopoulos, A.N.; Traboulay, E. (2002): Dimensional stability of OSB
International Symposium of Wood and Pulping Chemistry, Montreal, made from acetylated fir strands. Holz Roh- Werkstoff, 60, 84–87
Canada 70, 1-5
Papadopoulos, A.N.; Gkaraveli, A. (2003): Dimensional stabilization and
Murphy, R.J.; Turner, P. (1989): A vapour phase preservative treatment strength of particleboard by chemical modification with propionic anhy-
of manufactured wood based board materials. Wood Science Technology, dride. Holz Roh- Werkst., 61(2), 142–144
23(3), 273-279
Rapp, A.O.; Sailer, M.; Peek, R.-D. (2000): Innovative Holzvergütung zur
Niemz, P. (2002): Thermisch vergütetes Fichtenholz. Holzforschung Erhöhung der Dauerhaftigkeit. Mitteilungen der Bundesforschungsanstalt
Schweiz 1: 28-29 fuer Forst- und Holzwirtschaft Nr. 200, 27-34
Noack, D. (1969): Über die Heisswasserbehandlung von Rotbuchenholz Roll, H. (1997): Distribution of EMDI on particles and particleboard.
im Temperaturbereich von 100 bis 180°C. Holzforsch Holzverwert, 21(5), Proceedings of the 1st European Panel Products Symposium. Llandudno,
118–124 Wales, UK, p 235
Ohlmeyer, M. (2002): Untersuchung über die Eigenschaftsentwicklung Rowell, R.M. (1975): Chemical modification of wood: Advantages and
von Holzwerkstoffen nach dem Heißpressen. Dissertation Universität Disadvantages. In: Proceedings of the 71st Annual Meeting of the
Hamburg, 203 p. AWPA, San Francisco, 71, 41-51
Ohlmeyer, M.; Lukowsky, D. (2004) Wood-based panels produced from Rowell, R.; Tillman, A.; Simonson, R. (1986a): A simplified procedure
thermal treated wood – Properties and perspectives. In: Wood-frame for the acetylation of hardwood and softwood flakes for flakeboard
housing durability and disaster issues conference. October 4–6 2004, Las production. Wood Chemistry and Technology. 6(3), 427–448
Vegas, USA
Rowell, R.; Simonson, R.; Tillman, A. (1986b): Dimensional stability of
Paul, W.; Ohlmeyer, M. (2005): Optimisation of wood based panel particleboard made from vapor phase acetylated pine wood chips. Nordic
properties by heat pre-treatment. In: Proceedings of the Ninth European Pulp & Paper Research Journal, 2(1),11–17
Panel Product Symposium, October 5-7 2005, Llandudno, UK
Rowell, R.; Tillman, A.; Liu, Z. (1986c): Dimensional stabilization of
flakeboard by chemical modification. Wood Science and Technology,
20(1), 83–95
Rowell, R.; Wang, H.; Hyatt, J. (1986d): Flakeboards made from aspen
and southern pine wood flakes reacted with gaseous ketene. Wood
Chemistry and Technology. 6(3), 449–471
Rowell, R; Youngquist JA; Sachs, I.B. (1987): Adhesive bonding of
acetylated aspen flakes. I: Surface changes, hydrophobicity, adhesive
penetration and strength. International Journal of Adhesion and
Adhesives, 7(4), 183-188
Rowell, R.; Youngquist, J.; Imamura, Y. (1988): Strength tests on
acetylated aspen flakeboards exposed to a brown-rot fungus. Wood and
Fiber Science, 20(2), 266–271

280 281
Paul, Ohlmeyer Rowell, R.; Youngquist, J.; Rowell, J.; Hyatt, J. (1991): Dimensional
stability of aspen fiberboard made from acetylated fiber. Wood and Fiber
Rowell, R.; Tillman, A.; Simonson, R.. (1989): Acetylation of Science, 3(4), 558–566
lignocellulosic materials. U.S. Patent 4,804,384.
Seborg RM, Tarkow H, Stamm AJ (1953): Effect of heat upon the Modified Wood-Based Panels
dimensional stabilisation of wood. J For Prod Res Soc, 3(9), 59–67
Sekino, N.; Inoue, M.; Irle, M.A. (1997): Thickness swelling and internal Tillman, A.; Simonson, R.; Rowell, R. (1987): Dimensional stability and
bond strength of particles made of steam-pretreated particles. Mokuzai resistance to biological degradation of wood products by a simplified
Gak-kaishi 43(12), 1009–1015 acetylation procedure. In: Proceedings, 4th International symposium of
wood and pulping chemistry, April 27–30; Paris, France. Wood and
Sekino, N.; Inoue, M.; Irle, M. (1998) The bond quality of steam pre- Pulping Chemistry Association.
treated particles. In: The 2nd European Panel Products Symposium, 30–
38 Tarkow, H.; Turner, H.D. (1958): The Swelling Pressure of Wood. Forest
Products Journal 8(7):193-197.
Sehlstedt-Persson, M. (2003) Colour responses to heat treatment of
extractives and sap from pine and spruce. 8 th International IUFRO Wodd Tjeerdsma, B.F.; Boonstra, M. ;Militz, H. (1998): Thermal modification
Drying Conference, Brasov, Romania, 459-464 of non-durable wood species. Part 2. Improved wood properties of
thermally treated wood. International Research Group on Wood
Sheen, A.D. (1992): The preparation of acetylated wood fibre on a Preservation, Document No. IRG/WP 98-40124
commercial scale. In: Chemical modification of lignocellulosics, Rotorua,
New Zealand, 7-8 November 1992 (compiled by Plackett D.V. & Tomek, A. (1966): Die Heißvergütung von Holzspänen, ein neues
Dunningham E.A.). Forestry Research Institute, Ministry of Forestry FRI Verfahren zum Hydrophobieren von Spanplatten. Holztechnologie, 7(3),
Bulletin, 176, 1-8 157-160
Shen, K.C. (1973): Steam-press process for curing phenolic-bonded Tomimura, Y.; Matsuda, T. (1986): Particleboard made of steamed flakes.
particleboard. Forest Prod J, 23(3), 21–29 Mokuzai Gakkaishi 32(3), 170–175
Stamm, A.J.; Burr, H.K.; Kline, A.A. (1946): Heat Stabilized Wood Vick, C.; Krzysik, A.; Wood, J. (1991): Acetylated isocyanate-bonded
(Staybwood). Forest Products Laboratory Madison (USA), Rep. No. R. flakeboards after accelerated aging. Holz als Roh- und Werkstoff, 49, 221–
1621 228.
Stamm, A.J. (1964): Wood and Cellulose Science. The Ronald Press Youngquist, J.A.; Rowell, R.M.; Krzysik, A. (1986): Mechanical
Company, USA. Chapter 19, pp 312–342 properties and dimensional stability of acetylated aspen flake-boards.
Holz als Roh- und Werkstoff, 44, 453–457
Sudo, K. (1979): Character of hardboards from acetylated Asplund pulp.
Japan Wood Research Society Journal, 25(3), 203–208 Youngquist, J.A.; Rowell, R.M. (1990): Adhesive bonding of acetylated
aspen flakes. Part III. Adhesion with isocyanates. International Journal of
Subyanto, B.; Takino, S.; Kawai, S.; Sasaki, H. (1991) Production of Adhesion and Adhesives, 10(4), 273-276
thick low-density particleboard with a semi-continuous steam-injection
press. Mokuzai Gakkaishi 37(1):24–30 Youngquist, J.A.; Rowell, R.M.; Ross, N.; Krzysik, A. M; Chow, P.
(1990): In: Proceedings of the 1990 Joint International Forest Products
Sundqvist, B.; Morén, T. (2002) The influence of wood polymers and Conference, Taipei, Republic of China, 159-162
extractives on wood colour induced by hydrothermal treatment. Holz als
Roh- und Werkstoff, 60, 375-376 Youngquist J.A., Krzysik A.M., Muehl J.H. et al (1992): Mechanical and
physical properties of air formed wood fiber/polymer fiber composites.
Forest Products Journal, 42(6), 42-48

282 283
 COST

COST- the acronym for European Cooperation in Science and

Technology- is the oldest and widest European intergovernmental
network for cooperation in research. Established by the Ministerial
Conference in November 1971, COST is presently used by the scientific
communities of 36 European countries to cooperate in common research
projects supported by national funds.
The funds provided by COST - less than 1% of the total value of the
projects - support the COST cooperation networks (COST Actions)
through which, with EUR 30 million per year, more than 30 000
European scientists are involved in research having a total value which
exceeds EUR 2 billion per year. This is the financial worth of the
European added value which COST achieves.
A "bottom up approach" (the initiative of launching a COST Action
comes from the European scientists themselves), "à la carte participation"
(only countries interested in the Action participate), "equality of access"
(participation is open also to the scientific communities of countries not
belonging to the European Union) and "flexible structure" (easy
implementation and light management of the research initiatives) are the
main characteristics of COST.
As precursor of advanced multidisciplinary research COST has a very
important role for the realisation of the European Research Area (ERA)
anticipating and complementing the activities of the Framework
Programmes, constituting a "bridge" towards the scientific communities
of emerging countries, increasing the mobility of researchers across
Europe and fostering the establishment of "Networks of Excellence" in
many key scientific domains such as: Biomedicine and Molecular
Biosciences; Food and Agriculture; Forests, their Products and Services;
Materials, Physical and Nanosciences; Chemistry and Molecular Sciences
and Technologies; Earth System Science and Environmental
Management; Information and Communication Technologies; Transport
and Urban Development; Individuals, Societies, Cultures and Health. It
covers basic and more applied research and also addresses issues of
prenormative nature or of societal importance.
Web:http://www.cost.esf.org
Vi
ew
pu
bli
cat
ion
sta
ts

Formal publisher: Brunel University Press

Book title: Wood-Based Panels: An Introduction for Specialists
Year of publication: 2010
ISBN: 978-1-902316-82-6

ESF provides the COST COST is supported by the EU RTD

Office through an EC Framework programme
contract

ISBN 978-1-902316-82-6