Structure and Mechanical Behaviour of Wood-Fibre Composites

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1158

THOMAS JOFFRE

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 ISBN 978-91-554-8988-5 UPPSALA urn:nbn:se:uu:diva-229290 2014 Dissertation presented at Uppsala University to be publicly examined in Ångström 4001, Lägerhyddsvägen 1, Uppsala, Friday, 19 September 2014 at 10:00 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Stéphane Avril.

Abstract Joffre, T. 2014. Structure and Mechanical Behaviour of Wood-Fibre Composites. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1158. 34 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-8988-5.

Wood fibres have several advantages compared to man-made synthetic fibres: they have high specific stiffness, are renewable, relatively inexpensive, available in industrial quantities and biodegradable. However, to increase and diversify their utilisation, it is necessary to increase the understanding on what controls their mechanical properties. In this work, the hygroelastic behaviour of isolated wood fibres has been investigated using an analytical model and a finite element model based on three dimensional images obtained using synchrotron-based X-ray micro-computed tomography. It was thus possible to show how the cell wall responds to a mechanical load or a change in ambient relative humidity. The wood fibres were then mixed with a biopolymer aiming to produce a cost-efficient, 100% renewable composite material. The microstructure of the produced composites has been characterised using X-ray microtomography and digital image processing. It was for instance possible to measure the moisture-induced swelling of fibres embedded in a polymeric matrix. The experimental results have then been successfully compared with prediction obtained with a finite element model. The length of the fibres inside the composite has also been measured from three dimensional images, aiming to understand how each step of the processing chain is affecting the degradation of the aspect ratio of the reinforcing fibres. The presence of defects inside the composite has also been quantified using X-ray microtomography. The effects of the defects on the tensile strength have been predicted using an analytical model. The results have been compared with the measured tensile strength on each sample, showing that the size and orientation of the critical defect controls the tensile strength of the material. Finally, wood-fibre mats without any matrix material were compressed in the chamber of a microtomographic scanner. Sequential images were taken during the test. Using digital volume correlation, it was possible to calculate the local strain field inside the material. The effects of heterogeneities on the strain field have then been investigated. The applied compressive load resulted in transport of material from high to low density regions.

Thomas Joffre, Department of Engineering Sciences, Applied Mechanics, 516, Uppsala University, SE-751 20 Uppsala, Sweden.

ISSN 1651-6214 ISBN 978-91-554-8988-5 urn:nbn:se:uu:diva-229290 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-229290)

Dedicated to my wife and my child

« Les enfants seuls savent ce qu’ils cherchent. »

Antoine de Saint-Exupéry, Le petit prince

3 Abstract in Swedish

Träfibrer har många fördelar jämfört med tillverkade syntetiska fibrer. Trä- fibrer har t.ex. hög specifik styvhet, är förnyelsebara, relativt billiga, till- gängliga i industriellt gångbara kvaniteter och är biologiskt nedbrytbara. För att öka och diversifiera deras användning krävs fördjupad förståelse av vad som kontrollerar deras mekaniska egenskaper. I detta arbete har det hygroelastiska beteendet hos isolerade träfibrer stu- derats med hjälp av en analytisk modell och med finit-element-modellering utifrån tredimensionella bilder erhållna med synkrotron-Röntgen- mikrotomografi. Det var således möjligt att visa hur cellväggen påverkas av mekanisk belastning eller en förändring av omgivande relativ fuktighet. Träfibrerna blandades sedan med en biopolymer för att kostnadseffektivt kunna tillverka ett 100% förnyelsebart kompositmaterial. Mikrostrukturen hos den tillverkade kompositen har bestämts med Röntgen-mikrotomografi och digital bildbehandling. Det var t.ex. möjligt att mäta fuktsvällningen hos fibrer som är inbäddade i den polymera matrisen. De experimentella resultaten har jämförts med förutsägelser från finit-element-modellering med god överensstämmelse. Fiberlängderna i kompositmaterialet har också bestämts från de tredimensionella bilderna, med avseende att mäta hur varje steg i tillverkningskedjan bidrar till förkortning av de förstärkande fibrerna. Närvaro av defekter inuti kompositerna har också kvantifierats med hjälp av Röntgen-mikrotomografi. Defekternas inverkan på dragstryka har predik- terats med en analytisk modell. Resultaten har jämförts med uppmätt drag- styrka för varje provstav, vilket visar att storlek och orientering av den kri- tiska defekten har stor inverkan på materialets styrka. Slutligen har fibermattor utan matrismaterial belastats i tryck inne i en Röntgen-mikrotomograf. En sekvens av bilder har tagits under pålastning. Med digital volymskorrelation så kunde töjningsfältet bestämmas inuti materialet. Heterogeneiteternas inverkan på töjningsfältet har undersökts. Den pålagda lasten resutlerade i förflyttning av material från områden med hög densitet till områden med låg densitet.

4 Preface

The work presented in this thesis has been carried out at Uppsala University, in the Division of Applied Mechanics from January 2011 to September 2014, under the supervision of Professor Kristofer Gamstedt. It is the result of collaborations between many people from different universities and com- panies, and thus I would like to start with thanking everybody who helped or encouraged me in its completion. First, I express my deep and sincere thanks to Professor Kristofer Gam- stedt for giving me the opportunity to carry out my Ph.D. studies at Uppsala University as well as for the wise supervision, the encouragements and also the endless enthusiasm through the last four years. Then, I would also like to acknowledge Professor Per Isaksson for his help all along the thesis and for all the good advice. Generally, I express my gratitude to all the co-authors of the appended papers. I also take the opportunity to acknowledge all the personnel from the Di- vision of Applied Mechanics for the good working atmosphere. I enjoyed working with you! A special thanks to Alexey and my former office mate Gabriella for all the good scientific discussions and the help with the Swe- dish bureaucracy. The students I supervised are also acknowledged for their contribution to the thesis. To all the members of laboratory 3SR-LGP2 at Grenoble INP who helped me during my stay in Grenoble: merci beaucoup ! In particular I express my gratitude to Associate Professor Pierre Dumont for the wise and meticulous supervision as well as for the good courses in Grenoble which made me like mechanics. The Ph.D. students Erik L.G. Wernersson and Arttu Mietinnen, and Dr. Caroline Öhman are acknowledged for the help with microtomogaphy and digital image processing. Dr. Orlando Girlanda and Dr. Fredrik Sahlén at ABB Västerås are acknowledged for the help: it was a really good collaboration, and Orlando, sorry to have made you work during your holidays! Dr. Fredrik Bethold at Innventia is acknowledged for his grateful help with the processing of the materials.

5 My friends from the computer science department (A. Pr. Alberto, Dr Al- exandra, Eirini, Kostas, Stavros, Vasilis, …), chemistry department (An- drew, Dr. Frédéric, Thibaut…) and the real doctor (Gorav) are also acknowledged: for all the good moments that I spent with you in Uppsala, Grenoble or Stockholm: Thanks guys/girls!

Finally, my thought goes to my family: my parents, my brother and my uncle and my wife for the incommensurable support during the all good and the less good moments. Merci !

6 List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Joffre, T., Wernersson, E.L.G., Miettinen, A., Luengo Hendriks, C.L., Gamstedt, E.K. (2013) Swelling of cellulose fibres in composite materials: constraint effects of the surrounding matrix. Compo- sites Science and Technology, 74:52-59

II Joffre, T., Neagu, R.C., Bardage, S.L., Gamstedt, E.K. (2014) Mod- elling of the hygroelastic behaviour of normal and compression wood tracheids. Journal of Structural Biology, 185:89-98

III Joffre, T., Miettinen, A., Wernersson, E.L.G., Isaksson, P., Gam- stedt, E.K. (2014) Effects of defects on the tensile strength of short fibre composite materials. Mechanics of Material, 75:125-134

IV Joffre, T., Miettinen, A., Berthold, F., Gamstedt, E.K. X-ray micro- computed tomography investigation of fibre length degradation during the processing steps of short-fibre composites (submitted)

V Joffre, T., Isaksson, P., Dumont, P., Rolland du Roscoat, S., Sticko, S., Orgéas, L., Gamstedt, E.K. Moisture induced swelling properties of a single wood cell (submitted)

VI Joffre, T., Girlanda, O., Forsberg, F., Sahlén F., Sjödahl. M., Gam- stedt, E.K. Effects of the microstructure on the out-of-plane deformation mechanism of cellulosic fiber based materials (manuscript)

7 Apart from the appended papers listed above, the research activity carried out during the period has led to a number of publications not covered in this thesis.

• Marais, A., Magnusson, M.S., Joffre, T., Wernersson, E.L.G., Wågberg, L. (2014) New insights into the mechanisms behind the strengthening of ligno-cellulosic fibrous networks with polyamines. (Accepted in: Cellu- lose)

• Gamstedt, E.K., Joffre, T., Miettinen, A., Berthold, F. (2013) Monitor- ing of fibre length degradation during processing of short-fibre composites by use of X-ray computed tomography, Proceedings of the 34th Risø International Symposium on Materials Science, Technical Universi- ty of Denmark

• Toungara, M., Latil, P., Dumont, P.J.J., Rolland du Roscoat, S., Orgéas, L., Joffre, T., Passas R. Observation 3D et modélisation de l'hygroexpansion de fibres lignocellulosiques (2014), Note technique du Cerig: http://cerig.efpg.inpg.fr/Note/2014/hygroexpansion-fibre- lignocellulosique.htm

8 Contents

1. Motivation ...... 10 1.1 Introduction ...... 10 1.2 Background ...... 11 2. Materials and manufacturing ...... 13 2.1 Wood fibres ...... 13 2.2 Polylactic acid ...... 14 2.3 Manufacturing ...... 15 3. Methods ...... 16 3.1 X-ray microtomography ...... 16 3.2 Modelling approaches ...... 17 3.3 In situ mechanical testing ...... 18 3. Summary of papers ...... 19 Paper I ...... 20 Paper II ...... 21 Paper III ...... 22 Paper IV ...... 23 Paper V ...... 24 Paper VI ...... 25 4. Conclusions and future work ...... 26 References ...... 29

9 1. Motivation

1.1 Introduction

During the last decade, environmental awareness has led to a considerably increased interest in developing sustainable materials made from renewable resources. Wood fibres are a renewable biomaterial already available in industrial quantity at relatively low cost. In the pulp and paper industry, wood fibres are used to produce a wide variety of products: paper for printers, newsprint or magazines, packaging materials such as board and corrugated board, tissues and fluff products for diapers. Wood fibres are also used in more unknown applications such as structural and electric insulation materials in transformers, where cellulosic materials constitute around 10% of the global weight [1, 2], or in oil filters [3]. Thanks to their decent mechanical properties, wood fibres are also suitable as reinforcement in composite materials [4, 5]. The low density of natural fibres makes their specific properties comparable to those of commonly used glass fibres [6]. One application of wood-fibre composites is interior panels in cars, where flax, hemp, kenaf or jute fibres are already used as reinforcement in synthetic resins [7]. Other promising applications for wood- fibre composites are packaging materials, furniture and non-structural building components [8, 9]. The actual production of biocomposites in Europe is given in Table 1, along with forecast for 2020. The bulk of wood-fibre composites used today is based on wood chips or wood flour rather than pulp fibres [10]. The aspect ratio of the wood flour is one order of magnitude lower than the one of a slender wood fibre [11]. For these reasons, wood flour is today considered more as an inexpensive filler rather than a reinforcing material. The use of pulp fibres instead of wood flour will lead to an improvement of the mechanical performance of the material [12-14], and thus might open up new applications for wood-fibre based materials.

10 Table 1. Use of biocomposites in Europe [15] Forecast 2020 (under fa- Estimated quantities in vourable political frame- . 2010 (tons) work) (tons) Compression moulding: 190 000 370 000 with natural fibres 40 000 120 000 with cotton fibres 100 000 100 000 with wood fibres (WPC) 50 000 150 000 Extrusion and injection moulding: 172 000 550 000 WPC 167 000 450 000 Natural Fibres Rein- 5 000 100 000 forced Plastics Biocomposites in total 362 000 (14% of all compo- 920 000 (29% of all composites market) sites market)

The focus of the present study is on the mechanical properties of wood fibres (papers II and V) and in particular on their mechanical performance in diverse materials such as biocomposites (papers I, III, IV) or high density paper, known as pressboard (paper VI).

1.2 Background

Even though wood fibres have suitable mechanical properties, several drawbacks currently limit the development of natural fibre-reinforced materials. An important disadvantage is the inevitable moisture uptake which occurs in moist environment. This phenomenon has been widely studied in wood [16- 19]. However, the values measured on wood cannot directly be transferred to fibres in wood-fibre composites. In wood, the fibres are attached to one another by the middle lamella which will constrain the swelling of the fibres [20, 21]. Thus, it is necessary to get a better understanding of the hygroelastic behaviour of isolated wood fibres, which will be the topic of papers II and V. The swelling of free fibres can then be related to the swelling of wood and wood-fibre composites. Another drawback limiting the interest of wood pulp fibre as reinforcement in composite materials is the difficulty to blend hydrophilic wood fibres with traditional hydrophobic polymers: a good dispersion and adhesion of the fibre is difficult to obtain [22, 23]. To overcome these problems bi- opolymers, such as Polylactic acid (PLA), which are less hydrophobic than polypropylene (PP) or polyethylene (PE), have been proven to be more technically suitable and result in better mechanical property and dispersion [12, 22]. However, regardless the matrix, the sample will contain defects such as voids, cracks or fibre agglomerates. Due to these defects, the engineering

11 properties of the material will be affected. The strength of the material is particularly sensitive to the presence of defects, since during a mechanical test the sample will typically fail and break at the weakest cross-section in the sample. To be used in a large scale application, an inexpensive way of manufacturing wood-fibre composites should be employed. Extrusion to blend fibres and polymeric matrix and injection moulding to give a shape to the com- pound have been proven to be two useful candidates for the production of wood-fibre reinforced composites [13, 24]. These two processes are well- known but unfortunately not appropriate for the processing of slender wood fibres. As a result, it can be shown that the fibres are damaged during the manufacturing processing [25, 26]. Length degradation will result in a loss of mechanical performance [27]. X-ray microtomography is a powerful tool to evaluate this degradation through the procession chain of wood-fibre composites (paper IV). All these drawbacks are current limitations for the development of biocomposites. Hopefully, none of them are an absolute limitation. The global idea of the present work is to study the mechanical behaviour of wood-fibre based materials (biocomposites, solid wood or pressboard) to first understand the deformation mechanisms, and then use the results to tailor the microstructure of the materials, in order to optimize their properties for a specific application. For instance, for outdoor applications, a dimensionally stable matrix can be used to limit the moisture sensitivity of the fibres and thus reduce the global swelling of the material (Paper I). When strength is needed, it is the presence of defects inside the material and the dispersion of the fibres which need to be addressed (paper III). The dispersion of the fibres in the matrix can for instance be improved with the use of compatibilizers [28-31]. When stiffness is needed it is important to control the aspect ratio of the fibres and the focus has to be on the manufacturing equipment to avoid damaging the fibres (Paper IV).

12 2. Materials and manufacturing

2.1 Wood fibres

The focus of the present work is on softwood fibres. Softwood fibres are generally long (2-3 mm [32, 33]), while hardwood fibres are shorter (1 mm) and less flexible [34]. The difference can also be made between the fibres growing during the spring (earlywood) and during the summer (latewood): earlywood fibres have a larger cross section and a thinner cell wall than the denser latewood fibres (see e.g. [35]). This difference creates the so-called annual rings used in dendrochronology. A wood fibre is generally presented as a hollow layered structure. The different layers are illustrated in Fig. 1. They are designated as (from the outside to the inside of the tracheid): the primary cell wall (P), the outer layer (S1), the middle layer (S2) and the inner layer (S3) of the secondary cell- wall. In the wood tissue, the tracheids are surrounded by the middle lamella, which holds the cells together.

Figure 1. Schematic illustration of the cell wall of a softwood tracheid with different orientations of cellulose microfibrils in the layers.

The secondary fibre wall (S2 layer) forms the main part of the cell-wall and thus has a large influence on the mechanical and physical properties of the fibre (see e.g. [36, 37]). The S2 layer exhibits a helical structure, where the cellulose microfibrils are wound around the axis of the fibre with an angle, known as the microfibril angle (MFA). The MFA depends principally on

13 the growing direction of the tree. When a conifer shoot is moved from its vertical position, compression wood is formed in the lower part of the shoot. Compression wood has a higher MFA than normal wood [38, 39]. The growth rate of the compression wood is faster than in the upper part, resulting in a renewed horizontal growth [40]. The cross-section of the latewood tracheid is polygonal in the green state, which would induce significant de- viations if the geometry is approximated by a cylinder [41]. Each layer is composed of several polymers: cellulose, hemicellulose and lignin. Cellulose is the most common and represents approximately 50% of the volume frac- tion [42, 43]. Due to the abundance of accessible hydroxyl groups, all the cell-wall polymers are sensitive to moisture. Due to their helical structure, the longitudi- nal and radial hygroexpansion of the fibres will be coupled with some twist [44, 45]. The cell-wall hygroexpansion, intimately linked to the fibre swelling, has been estimated from the swelling behaviour of the polymer constituents and ultrastructure [19, 46], measurements of solid wood [47], or back- calculation from measurements of wood-fibre composites [48]. In composites or paper, the fibres have been extracted from wood and will show different hygroelastic properties due to the mechanical or chemical treatment that they have undergone during the pulping process [49].

2.2 Polylactic acid

During the last decade, environmental awareness has led to a considerably increased interest in developing plastic materials made from renewable resources. The global production of renewable plastic has drastically increased in the last five years (cf. Fig. 2).

Figure 2. Global production of bioplastics [50] .

Polylactic acid (PLA) is one of the already commercially available bioplastics. It is a thermoplastic aliphatic polyester manufactured from starch-rich renewable resources such as maize, sugar beets or wheat. PLA has decent physical and mechanical properties, making it a candidate for substitution of certain petrochemical thermoplastics [51]. Thanks to the good adhesion to cellulose fibres, PLA is also suitable for use as matrix material in wood– fibre composites, where it provides stress transfer between the load-carrying fibres [4, 24, 52-54].

2.3 Manufacturing

To manufacture wood-fibre reinforced composites, different approaches can be used. In Papers I and III the composites were compression moulded whereas in Paper IV injection moulding was used. In both processes the chosen temperature was around 180°C, well above the melting point of the used PLA (158°C).

15 3. Methods

3.1 X-ray microtomography

In order to acquire information about the microstructure of heterogeneous materials, X-ray micro-Computed Tomography (XµCT) has been proven to be a really powerful technique in material science [55]. It has been used in this thesis to analyse different materials: isolated wood tracheids (Paper V), fibre pressboard (Paper VI) and composites samples (Papers I, II, IV). In this section a short description of the techniques is given. More detailed information can be found in the literature (see e.g. [56-59]). The idea of XµCT is to reconstruct the three-dimensional microstructure of the sample from a number of two-dimensional projections, or more pre- cisely, the special distribution of the local X-ray attenuation coefficients. For many materials, especially non-diffractive ones, the attenuation coefficient correlates with local density of the material. As the density is different for fibres, matrix and air, the resulting images can be used to distinguish between the different constituent materials. The X-ray projections are used to computationally reconstruct the three-dimensional microstructure of the sample. The steps involved in XµCT are illustrated in Fig. 3. The two principal limitations of XµCT are today: - The size of the sample. A typical resolution for a detector such as the one of Skyscan 1171 shown in Fig. 4(a) is 4000x2000 pixels. This means that the size of the sample should not exceed 4000 times the resolution of the image. - The time necessary to perform a scan (often several hours for high resolutions), which makes such a device unsuitable for quality control in industrial production.

Figure 3. Illustration showing the different steps used in microtomography, inspired by [57]

3.2 Modelling approaches

Knowing the three-dimensional geometry, different microstructural properties can be measured from a three-dimensional image, which are of interest for composite mechanics, for instance fibre orientation distribution [60, 61], fibre length distribution [26, 62] or contacts between the fibres [63, 64]. The measured properties from the three-dimensional images are valuable input parameters for modelling. An example is the fibre orientation distribution which can be used in a laminate analogy model [48], where the composite with dispersed fibres with a certain orientation distribution is replaced by a laminate with thin unidirectional layers having different orientations from - 180 to 180º. The thickness of each unidirectional layer is linked to the prob- ability of having fibres in that direction. The stiffness of each unidirectional ply can be estimated by Hashin’s micromechanical model [65, 66]. This method has been proven reliable to predict the stiffness of biocomposites [67]. Another possibility is to use the three-dimensional image as a direct input for a finite element simulation, as in Paper V. The voxel of the three- dimensional image can be used as a hexahedron element for the finite element simulation (see e.g. [68, 69]). The other option is to adapt the mesh to the three-dimensional image aiming to reduce the number of elements. Dif- ferent routines or commercial softwares are able to mesh three-dimensional images. A comparison between the different approaches can be found in the literature [70].

17 3.3 In situ mechanical testing

It is possible to perform mechanical tests in the chamber of a microtomograph. Since XµCT is a non-destructive technique, it is possible to scan the very same sample under different loads. This is what is done in in situ mechanical tests. A typical chamber used for in situ testing is shown in Fig. 4(b). The interest of performing in situ mechanical tests is to track the deformation at a micrometre level. Then digital volume correlation can for instance be used to calculate the three-dimensional strain field inside the material [71-73].

Figure 4. (a) Microtomograph Skyscan 1172 (from the manufacturer) and (b) chamber used for in situ mechanical test.

18 3. Summary of papers

The papers are listed according to their content in table 2. Thereafter follows a brief summary of each paper based on their abstracts and conclusions.

Table 2 Paper I II III IV V VI Composite x x x x

Wood fibres x x x x x x

XµCT x x x x x

Analytical model- x x x ling

Finite element x x x modelling

Mechanical test- x x x x ing

19 Paper I

“Swelling of cellulose fibres in composite materials: Constraint effects of the surrounding matrix”

One of the main drawbacks of the use of wood fibre as reinforcement material is the swelling of these hydrophilic fibres due to moisture uptake (see Fig. 5). Since the fibres in the composite are generally embedded in a relatively hydrophobic matrix, the surrounding matrix should restrain the swelling of the fibres. In this paper, the constraint effect of the matrix on the fibre is studied and a micromechanical model is used to predict the swelling of embedded fibres. The predicted swelling is in concert with direct measurement of various wood–pulp fibre composites by means of three-dimensional X-ray microtomographic images. A small systematic difference between predicted and measured hygroexpansion coefficients could be attributed to resin-filled lumens, which are not accounted for in the model. Radial fibre swelling was ranked in decreasing order as: thermomechanical pulping, chemithermomechanical and kraft, with values from 0.27 to 0.15 strain per relative humidity change for embedded constrained fibres, and about twice as high for free unconstrained pulp fibres. It is possible to reduce by half the swelling of the fibre by using a dimensionally stable matrix.

Figure 5. Optical microscopy picture of kraft fibre (a) in a dry state (b) in a moist saturated state.

20 Paper II

“Modelling of the hygroelastic behaviour of normal and compression wood tracheids”

In this article an analytical model aiming to describe the hygroelastic behaviour of wood fibres is described. The focus is on the difference between normal and compression wood conifer tracheids. When a conifer shoot is moved from its vertical position, compression wood is formed in the under part of the shoot to control the growing direction (see Fig. 6). Compression wood shows different swelling and stiffness properties than those of usual normal wood, in order to achieve this practical function in the living plant. As a first step, a quantitative model is developed to predict the difference of moisture-induced expansion and axial stiffness between normal wood and compression wood. The model is based on a state space approach using concentric cylinders with anisotropic helical structure for each cell-wall layer, whose hygroelastic properties are in turn determined by a self-consistent concentric cylinder assemblage of the constituent wood polymers. The predicted properties compare well with experimental results found in the literature. Significant differences in both stiffness and hygroexpansion are found for normal and compression wood, primarily due to the large difference in microfibril angle and lignin content. On the basis of these numerical results, some functional arguments for the reason of high microfibril angle, high lignin content and cylindrical structure of compression wood tracheids are supported.

Figure 6. Development of compression wood in softwood.

21 Paper III

“Effects of defects on the tensile strength of short-fibre composite materials”

Heterogeneous materials tend to fail at the weakest cross-section, where the presence of microstructural heterogeneities or defects controls the tensile strength. Thus the prediction of strength of composite material should generally not be based on global average properties but on the properties of the weakest cross-section of the material. In short-fibre composites, unwanted fibre agglomerates are likely to initiate tensile failure. Different samples will have different amounts of agglomerates, as illustrated in Fig.7, and thus a different tensile strength. In this study, the dimensions and orientation of fibre agglomerates have been analysed from three-dimensional images obtained by X-ray microtomography. The geometry of the specific agglomerate responsible for failure initiation has been identified and correlated with the strength. At the plane of fracture, a defect in the form of a large fibre agglomerate was almost inevitably found. These experimental findings high- light a problem of some existing strength criteria, which are principally based on a rule of mixture of the strengths of constituent phases, and not on the weakest link. Only a weak correlation was found between stress concentration induced by the critical agglomerate and the strength. A strong correlation was however found between the stress intensity and the strength, which underlines the importance of the size of the critical defect in formula- tion of improved failure criteria for short-fibre composites. The increased use of three-dimensional imaging will facilitate the quantification of dimensions of the critical flaws.

Figure 7. Agglomeration of fibres inside different composite samples. The width of each sample is 6 mm.

22 Paper IV

“X-ray micro-computed tomography investigation of fibre length degradation during the processing steps of short-fibre composites”

The mechanical properties of composites in the fibre direction are mainly attributed to the fibre slenderness, or aspect ratio. A trade-off between performance and processability is usually required, and dependent on the in- tended application. If the fibre length could be retained or not severely de- graded during various processing steps towards the injection-moulded com- ponent, a stiffer and stronger composite product could be obtained. The processing steps for injection moulded wood-fibre composites here include: Pulping, commingling, extrusion, pelletizing, and injection moulding. In Fig. 8, a three-dimensional reconstruction of the microstructure of the commingled material is found. To tune the processing parameters systematically for retained fibre length, it would be useful to investigate the degradation of the original fibre length distribution throughout the processing chain. The fibre length degradation has been monitored by X-ray micro-computed tomography through the processing steps in wood pulp-fibre reinforced polylactide. A significant fibre-length degradation was found, in particular, the extrusion step was found to result in a drastic fibre length reduction.

Figure 8. 3D visualization of the commingled of pulp and PLA fibres. The width of the sample is approximately 1 mm.

23 Paper V

“Moisture induced swelling properties of a single wood cell”

Wood cells constitute the main building block in engineered wood-based materials, whose delimiting property frequently is moisture induced swelling. The hygroexpansion properties of wood cells, technically known as fibres, are used as input in predictive micromechanical models aimed for materials design. Values presented in the literature largely depend on the microfibrillar angle, the geometry of the fibre and limiting modelling as- sumptions. Synchrotron X-ray micro-computed tomography has recently prompted means for detailed measurements of the geometry of unconstrained individual fibres undergoing moisture-induced swelling, which makes it possible to directly quantify the hygroexpansion properties of the cell wall (see Fig. 9). These values can essentially be regarded as material properties since they are not compromised by the mentioned dependencies. In addition to the well-defined three-dimensional geometry, the present approach also accounts for the large deformations and the fact that cell-wall stiffness depends on the presence of moisture. A mixed numerical- experimental approach was adopted where an finite-element updating scheme was used to simulate the swelling going from the experimental fibre geometry at 47% relative humidity to the predicted geometry of the fibre in the wet state at 80% relative humidity at equilibrium conditions. The hygroexpansion coefficients were identified by comparing the predicted and the experimental fibre geometry in the wet state. The obtained values were 0.17 strain per change in relative humidity transverse to the microfibrils in the cell wall, and 0.014 along the microfibrils.

Figure 9. A three-dimensional view of a wood fibre at ambient relative humidity. The length of the fibre is 0.8 mm.

24 Paper VI “A 3D in situ investigation of the deformation mechanisms in compressive loading in the thickness direction of cellulose fiber mats”

Fiber mat materials based on cellulose natural fibres combine a useful set of properties, including renewability, stiffness, strength, dielectric insulation, etc. The dominant in-plane fibre orientation ensures the in-plane performance, at the expense of reduced out-of-plane behaviour, which has not been studied as extensively as the in-plane behaviour. Quantitative use of X-ray micro-computed tomography and strain analyses under in situ loading open up possibilities to identify key mechanisms responsible for deformations. In the present investigation, focus is placed on the out-of-plane deformation under compressive loading of thick, high density paper, known as pressboard (see Fig. 10). The samples were compressed in the chamber of a microtomographic scanner. 3D images were captured before and after the loading the sample. From sequential 3D images, the strain field inside the material was calculated using digital volume correlation. The first principal strain compo- nent of the strain tensor showed a significant correlation with the density variation in the material. This shows that manufacturing-induced inhomoge- neities in the microstructure, related to the manufacturing process, create strain concentration zones in the sample, contributing to the overall compli- ance of the material.

Figure 10. Fibre pressboard. Dimension of the sample are 4x4x2.2 mm

25 4. Conclusions and future work

Wood-fibre based materials (e.g. biocomposites, pressboard, and cardboard) have a potential to be used as structural material in load bearing applications. However, to increase and diversify their use, the mechanical behaviour has to be better understood. Microtomography is an excellent tool to visualise the microstructure of these heterogeneous materials. The understanding of the global behaviour of material should not be limited only to the observation of phenomenon. Modelling techniques can be used to predict the deformation of some microstructures. Since microtomography is a non-destructive technique, the experimental results can be directly compared to numerical results aiming to validate a numerical approach. The main conclusions of the Ph.D. thesis are based on the conclusions of the enclosed papers on the general topic of the relation between structure and mechanical properties of wood fibres and engineered materials based on wood fibres: • The sensibility of wood fibres to moisture can be partially supressed when the fibres are embedded in a dimensionally stable polymeric matrix. This should be taken into consideration when designing composites for outdoor application • Softwood is able to control its growing direction by producing compression wood. At some point of the growing process, compression wood fibres are able to swell and due to their helical structure they will twist and consequently push the adjacent fibres to reorient the growing direction of the tree. This explains the cylindrical geometry of compression wood tracheids. By adopting a circular cross section the torsion rigidity of the section is reduced. The high microfibrillar angle of compression wood means that a higher actuation can be achieved in the axial direction as compared with normal wood. • During a tensile test, the composite material will fail at its weakest cross-section. The prediction of the strength of short-fibre composite materials should then be based on the properties of defects inside the sample and not on the average properties of the constituents of the composite. Fibre agglomerates, i.e. bundles of undispersed fibres, have been proven have a large impact on tensile strength of wood-fibre reinforced composites. Their size orientation and shape is in particular affecting the strength of the material.

26 • Extrusion and injection moulding are efficient ways to produce wood- fibre composites. It is then possible to utilize conventional industrial equipment to produce biocomposites. However, this equipment is not tailored for the use of natural fibres which are particularly weak in the transverse direction. As a result the fibres are drastically damaged during the manufacturing process and are losing their interesting aspect ratios. A microtomographic study has shown that the compounding stage of pelletization results in severe fibre length degradation. Microtomography can be a useful tool to identify processing steps which have a large impact microstructural parameters that control the mechanical performance. • It is now possible to characterise the three-dimensional geometry of isolated wood fibres at different ambient relative humidities in synchrotron X-ray micro-computed tomography. These images can be used as input of a finite element model. It is then possible to tune the material properties used as input of the finite element simulation to have a numerical deformation matching the experimental one. Using this technique, we were able to measure the hygroexpansion properties of the cell wall directly. • Performing a compression test in the chamber of a microtomographic scanner allowed us to track the deformation of a thick and dense paper, known as pressboard. Using digital volume correlation the three- dimensional strain field was measured inside the material. The results emphasise the effects of heterogeneities which create strain concentration within the sample.

Micromechanics has a potentially important role in the development of engineered wood materials, but it should be developed jointly together with users and stakeholders. Future close collaboration with chemists and process engineers in materials development is a necessity to make modelling work and experimental micromechanics useful and form the basis for decision making in manufacturing, choice of raw materials, and material modification. Image analysis and quantitative microtomography are potentially very useful tools in mechanics of materials. Numerical models can be validated by direct comparison with experimental results in three dimensions. Indeed, since XµCT is a non-destructive technique it is possible to acquire images under different loading conditions. The obtained results might then be directly confronted to the numerical predictions. Finite element updating techniques can be used to find the material properties of objects which lead to an agreement between simulation and experiments. It is thus possible to measure material properties on irregular heterogeneous objects. Development of 3D image analysis of XµCT in combination with finite element simulations

27 could rationalize the development of manmade materials, such as natural fibre composites.

28 References

[1] O. Girlanda, D. Tjahjanto, S. Ostlund, K. Wei, and L. Schmidt. Characterization and modelling of the mechanical properties of pressboard, in Electrical Insulation and Dielectric Phenomena (CEIDP), IEEE Conference on, 2013, 563-566. [2] S. Yamashita, M. Yoshida, S. Takayama, and T. Okuyama. Stem- righting mechanism in gymnosperm trees deduced from limitations in compression wood development, Annals of botany, vol. 99, 487- 493, 2007. [3] A. Pasila. A biological oil adsorption filter, Marine pollution bulletin, vol. 49, 1006-1012, 2004. [4] M. Huda, L. Drzal, M. Misra, and A. Mohanty. Wood fiber reinforced poly (lactic acid) composites: Evaluation of the physicomechanical and morphological properties, Journal of Applied Polymer Science, vol. 102, 4856-4869, 2006. [5] R. C. Neagu, E. K. Gamstedt, and F. Berthold. Stiffness contribution of various wood fibers to composite materials, Journal of Composite Materials, vol. 40, 663-699, 2006. [6] P. Wambua, J. Ivens, and I. Verpoest. Natural fibres: can they replace glass in fibre reinforced plastics?, Composites Science and Technology, vol. 63, 1259-1264, 2003. [7] J. Holbery and D. Houston. Natural-fiber-reinforced polymer composites in automotive applications, JOM, vol. 58, 80-86, 2006. [8] P. Clouston, S. Civjan, and L. Bathon. Experimental behavior of a continuous metal connector for a wood-concrete composite system, Forest Products Journal, vol. 54, 76-84, 2004. [9] P. M. Smith and M. P. Wilcott. Opportunities for wood/natural fiber- plastic composites in residential and industrial applications, Forest Products Journal, vol. 56, 4-11, 2006. [10] J. Morton and L. Rossi. Current and emerging applications for natural and wood fiber composites, in Presentation in the 7th International Conference of Wood Fiber-Plastic Composite, 2003, 19-20. [11] N. M. Stark and R. E. Rowlands. Effects of wood fiber characteristics on mechanical properties of wood/polypropylene composites, Wood and Fiber Science, vol. 35, 167-174, 2003. [12] H. Peltola, E. Pääkkönen, P. Jetsu, and S. Heinemann. Wood based PLA and PP composites: Effect of fibre type and matrix polymer on

29 fibre morphology, dispersion and composite properties, Composites Part A, vol. 61, 13-22, 2014. [13] P. Nygård, B. Tanem, T. Karlsen, P. Brachet, and B. Leinsvang. Extrusion-based wood fibre–PP composites: Wood powder and pelletized wood fibres–a comparative study, Composites Science and Technology, vol. 68, 3418-3424, 2008. [14] S. Butylina, O. Martikka, and T. Kärki. Properties of wood fibre- polypropylene composites: effect of wood fibre source, Applied Composite Materials, vol. 18, 101-111, 2011. [15] M. Carus. A first quantitative assessment of biomass use in the EU industry, Nova-Institut Bio-based Economy in the EU-27, 2012. [16] D. Derome, M. Griffa, M. Koebel, and J. Carmeliet. Hysteretic swelling of wood at cellular scale probed by phase-contrast X-ray tomography, Journal of Structural Biology, vol. 173, 180-190, 2011. [17] A. Patera, D. Derome, M. Griffa, and J. Carmeliet. Hysteresis in swelling and in sorption of wood tissue, Journal of Structural Biology, vol. 182, 226-234, 2013. [18] E. Badel and P. Perré. Using a digital X-ray imaging device to measure the swelling coefficients of a group of wood cells, NDT & E International, vol. 34, 345-353, 2001. [19] K. Persson. Micromechanical modelling of wood and fibre properties, Ph.D thesis, Division of Structural Mechanics, Lund Institute of Technology, 2000. [20] R. C. Neagu and E. K. Gamstedt. Modelling of effects of ultrastructural morphology on the hygroelastic properties of wood fibres, Journal of Materials Science, vol. 42, 10254-10274, 2007. [21] I. Burgert, M. Eder, N. Gierlinger, and P. Fratzl. Tensile and compressive stresses in tracheids are induced by swelling based on geometrical constraints of the wood cell, Planta, vol. 226, 981-987, 2007. [22] A. Bledzki, M. Letman, A. Viksne, and L. Rence. A comparison of compounding processes and wood type for wood fibre-PP composites, Composites Part A, vol. 36, 789-797, 2005. [23] G. Chinga-Carrasco, A. Miettinen, C. L. L. Hendriks, E. K. Gamstedt, and M. Kataja. Structural Characterisation of Kraft Pulp Fibres and Their Nanofibrillated Materials for Biodegradable Composite Applications, InTech, 2012. [24] V. L. Finkenstadt and B. Tisserat. Poly (lactic acid) and Osage Orange wood fiber composites for agricultural mulch films, Industrial Crops and Products, vol. 31, 316-320, 2010. [25] B. Nyström, R. Joffe, and R. Långström. Microstructure and strength of injection molded natural fiber composites, Journal of Reinforced Plastics and Composites, vol. 26, 579-599, 2007. [26] A. Miettinen, C. L. L. Hendriks, G. Chinga-Carrasco, E. K. Gamstedt, and M. Kataja. A non-destructive X-ray

30 microtomography approach for measuring fibre length in short-fibre composites, Composites Science and Technology, vol. 72, 1901- 1908, 2012. [27] S. Migneault, A. Koubaa, F. Erchiqui, A. Chaala, K. Englund, C. Krause, et al.. Effect of fiber length on processing and properties of extruded wood‐fiber/HDPE composites, Journal of Applied Polymer Science, vol. 110, 1085-1092, 2008. [28] A. Marais, J. J. Kochumalayil, C. Nilsson, L. Fogelström, and E. K. Gamstedt. Toward an alternative compatibilizer for PLA/cellulose composites: Grafting of xyloglucan with PLA, Carbohydrate Polymers, vol. 89, 1038-1043, 2012. [29] M. Takatani, K. Ikeda, K. Sakamoto, and T. Okamoto. Cellulose esters as compatibilizers in wood/poly (lactic acid) composite, Journal of Wood Science, vol. 54, 54-61, 2008. [30] A. Karmarkar, S. Chauhan, J. M. Modak, and M. Chanda. Mechanical properties of wood–fiber reinforced polypropylene composites: effect of a novel compatibilizer with isocyanate functional group, Composites Part A, vol. 38, 227-233, 2007. [31] D. Plackett. Maleated polylactide as an interfacial compatibilizer in biocomposites, Journal of Polymers and the Environment, vol. 12, 131-138, 2004. [32] S. L. Bardage. Three-dimensional modeling and visualization of whole Norway spruce latewood tracheids, Wood and Fiber Science, vol. 33, 627-638, 2001. [33] J. Brandström, S. L. Bardage, G. Daniel, and T. Nilsson. The structural organisation of the S1 cell wall layer of Norway spruce tracheids, IAWA Journal, vol. 24, 27-40, 2003. [34] T. Tabarsa and Y. H. Chui. Characterizing microscopic behavior of wood under transverse compression. Part II. Effect of species and loading direction, Wood and Fiber Science, vol. 33, 223-232, 2001. [35] C. Plomion, G. Leprovost, and A. Stokes. Wood formation in trees, Plant physiology, vol. 127, 1513-1523, 2001. [36] J. Barnett and V. A. Bonham. Cellulose microfibril angle in the cell wall of wood fibres, Biological Reviews, vol. 79, 461-472, 2004. [37] P. Fratzl, R. Elbaum, and I. Burgert. Cellulose fibrils direct plant organ movements, Faraday Discuss., vol. 139, 275-282, 2008. [38] S. Andersson, R. Serimaa, M. Torkkeli, T. Paakkari, P. Saranpää, and E. Pesonen. Microfibril angle of Norway spruce [Picea abies (L.) Karst.] compression wood: comparison of measuring techniques, Journal of Wood Science, vol. 46, 343-349, 2000. [39] T. E. Timell. Compression wood in gymnosperms, vol. 3: Springer- Verlag: Berlin, 1986. [40] H. Yamamoto, F. Sassus, M. Ninomiya, and J. Gril. A model of anisotropic swelling and shrinking process of wood, Wood Science and Technology, vol. 35, 167-181, 2001.

31 [41] T. Innes. Stress model of a wood fibre in relation to collapse, Wood Science and Technology, vol. 29, 363-376, 1995. [42] J. Dinwoodie. Timber: its structure, properties and utilisation, MacMillan Press Ltd, London, 1981. [43] D. Fengel. The ultrastructure of cellulose from wood, Wood Science and Technology, vol. 3, 203-217, 1969. [44] I. Burgert, K. Frühmann, J. Keckes, P. Fratzl, and S. Stanzl- Tschegg. Properties of chemically and mechanically isolated fibres of spruce (Picea abies [L.] Karst.). Part 2: Twisting phenomena, Holzforschung, vol. 59, 247-251, 2005. [45] M. Johansson, M. Perstorper, R. Kliger, and G. Johansson. Distortion of Norway spruce timber Part 2. Modelling twist, Holz als Roh-und Werkstoff, vol. 59, 155-162, 2001. [46] I. Cave. A theory of the shrinkage of wood, Wood Science and Technology, vol. 6, 284-292, 1972. [47] L. Wallström and K. Lindberg. Measurement of cell wall penetration in wood of water-based chemicals using SEM/EDS and STEM/EDS technique. Wood Science and Technology, vol. 33, 111-122, 1999. [48] K. M. Almgren, E. K. Gamstedt, and J. Varna. Contribution of wood fiber hygroexpansion to moisture induced thickness swelling of composite plates, Polymer Composites, vol. 31, 762-771, 2010. [49] R. Evans, R. H. Newman, U. C. Roick, I. D. Suckling, and A. F. Wallis. Changes in cellulose crystallinity during kraft pulping. Comparison of infrared, X-ray diffraction and solid state NMR results, Holzforschung-International Journal of the Biology, Chemistry, Physics and Technology of Wood, vol. 49, 498-504, 1995. [50] E. Bioplastics. Life Cycle Assessment of Bioplastics, http://en.european-bioplastics.org/, 2008. [51] O. Wolf, M. Crank, M. Patel, F. Marscheider-Weidemann, J. Schleich, B. Hüsing, et al.. Techno-economic feasibility of large- scale production of bio-based polymers in Europe, Technical Report series, European commission, 2005. [52] M. S. Huda, L. T. Drzal, A. K. Mohanty, and M. Misra. Chopped glass and recycled newspaper as reinforcement fibers in injection molded poly (lactic acid)(PLA) composites: a comparative study, Composites Science and Technology, vol. 66, 1813-1824, 2006. [53] K. M. Almgren, M. Åkerholm, E. K. Gamstedt, L. Salmen, and M. Lindström. Effects of Moisture on Dynamic Mechanical Properties of Wood Fiber Composites Studied by Dynamic FT-IR Spectroscopy, Journal of Reinforced Plastics and Composites, vol. 27, 1709-1721, 2008. [54] A. Gregorova, M. Hrabalova, R. Kovalcik, and R. Wimmer. Surface modification of spruce wood flour and effects on the dynamic

32 fragility of PLA/wood composites, Polymer Engineering & Science, vol. 51, 143-150, 2011. [55] L. Salvo, P. Cloetens, E. Maire, S. Zabler, J. Blandin, J.-Y. Buffière, et al.. X-ray micro-tomography an attractive characterisation technique in materials science, Nuclear instruments and methods in physics research section B: Beam interactions with materials and atoms, vol. 200, 273-286, 2003. [56] A. Kak and M. Slaney. Principles of computerized tomographic imaging, IEEE Press, New York, 1988. [57] F. Forsberg. X-ray microtomography and digital volume correlation for internal deformation and strain analysis, Ph.D thesis, Luleå University of Technology, 2008. [58] J. Elliott and S. Dover, X‐ray microtomography. Journal of Microscopy, vol. 126, 211-213, 1982. [59] E. N. Landis and D. T. Keane. X-ray microtomography, Materials characterization, vol. 61, 1305-1316, 2010. [60] M. V. Ginkel. Image Analysis using Orientation Space based on Steerable Filters, Ph.D thesis, Deft Univerisity of technology, The Netherlands, 2002. [61] J. Tan, J. Elliott, and T. Clyne. Analysis of tomography images of bonded fibre networks to measure distributions of fibre segment length and fibre orientation, Advanced Engineering Materials, vol. 8, 495-500, 2006. [62] C. L. Luengo Hendriks. Constrained and dimensionality- independent path openings, Image Processing, IEEE Transactions on, vol. 19, 1587-1595, 2010. [63] J. Viguié, P. Latil, L. Orgéas, P. Dumont, S. Rolland du Roscoat, J.- F. Bloch, et al.. Finding fibres and their contacts within 3D images of disordered fibrous media, Composites Science and Technology, vol. 89, 202-210, 2013. [64] K. M. Almgren, E. K. Gamstedt, P. Nygård, F. Malmberg, J. Lindblad, and M. Lindström. Role of fibre-fibre and fibre-matrix adhesion in stress transfer in composites made from resin- impregnated paper sheets, International Journal of Adhesion and Adhesives, vol. 29, 551-557, 2009. [65] Z. Hashin. Analysis of Properties of Fiber Composites With Anisotropic Constituents, Journal of Applied Mechanics, vol. 46, 543, 1979. [66] Z. Hashin. Analysis of composite materials-a survey, Journal of Applied Mechanics, vol. 50, 481, 1983. [67] G. Josefsson, F. Berthold, and E. K. Gamstedt. Stiffness contribution of cellulose nanofibrils to composite materials, International Journal of Solids and Structures, vol. 51, 945-953, 2014. [68] S. J. Shefelbine, U. Simon, L. Claes, A. Gold, Y. Gabet, I. Bab, et al.. Prediction of fracture callus mechanical properties using micro-

33 CT images and voxel-based finite element analysis, Bone, vol. 36, 480-488, 2005. [69] M. Lengsfeld, J. Schmitt, P. Alter, J. Kaminsky, and R. Leppek. Comparison of geometry-based and CT voxel-based finite element modelling and experimental validation, Medical engineering & physics, vol. 20, 515-522, 1998. [70] D. Ulrich, B. Van Rietbergen, H. Weinans, and P. Rüegsegger. Finite element analysis of trabecular bone structure: a comparison of image-based meshing techniques, Journal of Biomechanics, vol. 31, 1187-1192, 1998. [71] B. K. Bay, T. S. Smith, D. P. Fyhrie, and M. Saad. Digital volume correlation: three-dimensional strain mapping using X-ray tomography, Experimental Mechanics, vol. 39, 217-226, 1999. [72] F. Forsberg, M. Sjödahl, R. Mooser, E. Hack, and P. Wyss. Full Three‐Dimensional Strain Measurements on Wood Exposed to Three‐Point Bending: Analysis by Use of Digital Volume Correlation Applied to Synchrotron Radiation Micro‐Computed Tomography Image Data, Strain, vol. 46, 47-60, 2010. [73] H. Leclerc, J.-N. Périé, S. Roux, and F. Hild. Voxel-scale digital volume correlation, Experimental Mechanics, vol. 51, 479-490, 2011. [74] J. Andersons and R. Joffe. Estimation of the tensile strength of an oriented flax fiber-reinforced polymer composite, Composites Part A, vol. 42, 1229-1235, 2011.

Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1158 Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-229290 2014