UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Energy absorbing ability of wood/polyester composite laminates Haghdan, Shayesteh 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_september_haghdan_shayesteh.pdf [ 4.34MB ]
Metadata
JSON: 24-1.0166594.json
JSON-LD: 24-1.0166594-ld.json
RDF/XML (Pretty): 24-1.0166594-rdf.xml
RDF/JSON: 24-1.0166594-rdf.json
Turtle: 24-1.0166594-turtle.txt
N-Triples: 24-1.0166594-rdf-ntriples.txt
Original Record: 24-1.0166594-source.json
Full Text
24-1.0166594-fulltext.txt
Citation
24-1.0166594.ris

Full Text

ENERGY ABSORBING ABILITY OF WOOD/POLYESTER COMPOSITE LAMINATES by  SHAYESTEH HAGHDAN  B.Sc., Natural Resources Engineering, University of Tehran, 2006 M.Sc., Natural Resources Engineering, University of Tehran, 2008  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   August 2015  © Shayesteh Haghdan, 2015 ii   Abstract  Currently used energy absorbers in transportation industries are made of synthetic fiber/polymer composites as an alternative to their metal counterparts. These composites are stiff and strong but are somewhat brittle when subjected to impact load which limits their application when high energy absorbing ability is required. Wood, in contrast, has a high stiffness and strength to weight ratio and exhibits a higher deflection before failure. Despite the extensive research on the mechanical properties of synthetic fiber/polymer composites few researches are available on the effects of wood composite configuration and densification and its lamination set-up on its impact and compressive properties.    This research focused on the use of wood in the form of thin veneer to reinforce polyester and composites of them were fabricated using hand lay-up and compression molding, in different thicknesses. Various wood configurations were used to create unidirectional, cross-ply, and woven mats. The effects of each mat configuration on the impact properties of wood/polyester composites and the lamination and curing processes were investigated and discussed. The gap of knowledge on the wettability of wood to the polyester resin was informed in this dissertation using contact angle measurements and roughness tests. Energy absorbing behavior and dominant fracture mechanisms of wood/polyester laminates subjected to quasi-static compression and shear loading were examined and the results were compared with the lab-made glass fiber/polyester composites.   iii    Findings of this study demonstrated that the effect of wood configuration on the impact properties of the polyester composites was significant. Wood densification improved the impact performance of composites but this improvement was not statistically significant. It was found that wood composites had an impact energy equivalent to that of glass fiber laminates.                 iv   Preface Shayesteh Haghdan, dissertation author, was the lead investigator, responsible for the concept development and experiment design, the research and data analysis, and manuscript composition. Dr. Gregory Smith was the research supervisor who provided guidance throughout the project, commented, and edited all four manuscripts. Dr. Thomas Tannert was the co-author in two of the manuscripts and contributed to concept refinement and manuscript edits.   A version of Chapter 2 has been published: Haghdan, S., Smith, G, D., 2015, “Natural fiber reinforced polyester composites: a literature review”, Journal of Reinforced Plastics and Composites, Vol. 34, No. 14, pp. 1179-1190.   A version of Chapter 3 has been published: Haghdan, S., Tannert, T., Smith, G, D., 2015, “Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites”, Journal of Composite Materials, Vol 49, No 10, pp. 1161-1170.   A version of Chapter 4 has been published: Haghdan, S., Tannert, T., Smith, G, D., 2015, “Wettability and impact performance of wood veneer/polyester composites”, Bioresources, Vol. 10, No. 3, pp. 5633-5654.   A version of Chapter 5 has been published: Haghdan, S., Smith, G, D., 2015, “Fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading”, Composites Part A, Vol 74, pp. 114-122. v   Table of Contents  Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ...........................................................................................................................v List of Tables ................................................................................................................................ ix List of Figures .................................................................................................................................x List of Abbreviations ................................................................................................................. xiv Acknowledgements ......................................................................................................................xv Dedication ................................................................................................................................... xvi Chapter 1: Introduction ................................................................................................................1 1.1 Background ..................................................................................................................... 1 1.1.1 Impact behavior of wood/polymer composites ........................................................... 5 1.1.2 Energy absorbing ability of composites ...................................................................... 6 1.1.3 Wood densification ..................................................................................................... 8 1.2 Research problem statement ......................................................................................... 10 1.3 Research objectives ....................................................................................................... 11 1.4 Organization of the dissertation .................................................................................... 12 Chapter 2: Natural fiber reinforced polyester composites: a literature review .....................14 2.1 Synopsis ........................................................................................................................ 14 2.2 Plant fiber/polyester composites ................................................................................... 14 2.2.1 Hemp fiber ................................................................................................................ 14 2.2.1.1 Fiber loading and fiber length effects ............................................................... 15 2.2.1.2 Chemical treatments effects .............................................................................. 16 2.2.2 Jute fiber.................................................................................................................... 17 2.2.2.1 Fiber loading effects ......................................................................................... 17 2.2.2.2 Chemical treatments effects .............................................................................. 19 2.2.2.3 Long-term water absorption effects .................................................................. 20 2.2.3 Banana fiber .............................................................................................................. 21 2.2.4 Sisal fiber .................................................................................................................. 23 vi   2.2.4.1 Chemical treatments effects .............................................................................. 24 2.2.4.2 Manufacturing methods effects ......................................................................... 25 2.2.5 Coir fibers ................................................................................................................. 26 2.2.6 Kenaf fibers ............................................................................................................... 27 2.2.7 Other plant fibers ...................................................................................................... 29 2.2.7.1 Chemical treatments effects .............................................................................. 29 2.2.7.2 Fiber loading and filler size effects ................................................................... 31 2.3 Hybrid fiber/polyester composites ................................................................................ 31 2.3.1 Glass fiber-plant fiber/polyester ............................................................................... 31 2.3.1.1 Glass fiber-jute/polyester .................................................................................. 32 2.3.1.2 Glass Fiber-sisal/polyester ................................................................................ 33 2.3.1.3 Glass fiber-pineapple leaf and sisal/polyester................................................... 34 2.3.1.4 Glass fiber-oil palm/polyester ........................................................................... 35 2.3.2 Plant fiber-plant fiber/polyester ................................................................................ 36 2.3.3 Glass fiber-wood/polyester ....................................................................................... 37 2.4 Wood/polyester composites .......................................................................................... 38 2.4.1 Chemical treatments effects ...................................................................................... 39 2.4.2 Densification and configuration effects .................................................................... 40 2.5 Summary ....................................................................................................................... 41 Chapter 3: Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites ..............................................................................................43 3.1 Synopsis ........................................................................................................................ 43 3.2 Introduction ................................................................................................................... 44 3.3 Experimental investigation ........................................................................................... 45 3.3.1 Materials ................................................................................................................... 45 3.3.2 Composite manufacturing ......................................................................................... 47 3.3.3 Test methods ............................................................................................................. 49 3.4 Results ........................................................................................................................... 51 3.5 Discussion ..................................................................................................................... 54 3.6 Conclusions ................................................................................................................... 56 vii   Chapter 4: Wettability and impact performance of wood veneer/polyester composites ......67 4.1 Synopsis ........................................................................................................................ 67 4.2 Introduction ................................................................................................................... 68 4.3 Experimental investigation ........................................................................................... 71 4.3.1 Materials ................................................................................................................... 71 4.3.2 Specimen description ................................................................................................ 72 4.3.3 Composite manufacturing ......................................................................................... 73 4.3.4 Test methods ............................................................................................................. 75 4.3.4.1 Impact test ......................................................................................................... 75 4.3.4.2 Contact angle measurement .............................................................................. 76 4.3.4.3 Surface roughness measurement ....................................................................... 77 4.4 Results and Discussion ................................................................................................. 77 4.4.1 Laminating process ................................................................................................... 77 4.4.2 Microscopic observation ........................................................................................... 79 4.4.3 Fracture mechanisms ................................................................................................ 81 4.4.4 Impact energy............................................................................................................ 82 4.4.5 Wettability of wood veneers ..................................................................................... 83 4.4.6 Surface roughness of wood veneers .......................................................................... 86 4.5 Conclusions ................................................................................................................... 88 Chapter 5: Fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading .................................................................................................104 5.1 Synopsis ...................................................................................................................... 104 5.2 Introduction ................................................................................................................. 105 5.3 Experimental investigation ......................................................................................... 108 5.3.1 Materials ................................................................................................................. 108 5.3.2 Specimen description .............................................................................................. 108 5.3.3 Composite manufacturing ....................................................................................... 112 5.3.4 Test methods ........................................................................................................... 113 5.4 Results ......................................................................................................................... 114 5.4.1 Compressive properties of laminates ...................................................................... 114 viii   5.4.1.1 Effect of symmetry ......................................................................................... 115 5.4.1.2 Effect of balance ............................................................................................. 115 5.4.1.3 Effect of number of lamina ............................................................................. 116 5.4.1.4 Wood/polyester vs. glass fiber/polyester and the commercial treatments ...... 116 5.4.2 Comparison of load-displacement curves ............................................................... 117 5.4.2.1 Glass fiber/polyester (GP)............................................................................... 117 5.4.2.2 Eight-layers symmetric with face layers of 90 (8SB) ..................................... 117 5.4.2.3 Twelve-layers symmetric with face layers of 90 (12SB)................................ 118 5.4.3 Fracture mechanisms of other treatments ............................................................... 118 5.5 Discussion ................................................................................................................... 120 5.5.1 Effect of symmetry, balance, and number of lamina .............................................. 120 5.5.2 Wood/polyester vs. glass fiber/polyester and the commercial treatments .............. 121 5.5.3 Fracture mechanisms .............................................................................................. 122 5.6 Conclusions ................................................................................................................. 123 Chapter 6: Conclusion ...............................................................................................................137 6.1 Project summary ......................................................................................................... 137 6.2 Conclusions ................................................................................................................. 142 6.3 Outlook ....................................................................................................................... 143 Bibliography ...............................................................................................................................144 Appendices ..................................................................................................................................160 Appendix A Plot of failure and non-failure using staircase method ....................................... 160 Appendix B ANOVA results for the effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites ...................................................... 161 Appendix C ANOVA results for the wettability and impact performance of wood veneer/polyester composites ................................................................................................... 169 Appendix D Stress-strain curve in 8SB and 12NS samples ................................................... 190 Appendix E ANOVA results for the fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading ............................................................................ 191 Appendix F SEM micrographs of fracture mechanisms of 12SB laminates: delamination and transverse crack ....................................................................................................................... 194 ix   List of Tables Table 3.1 List of treatments, including mass and volume fractions of components; n= 20 replicates for each impact energy mean and St.Dev. Note: W denotes wood, P polyester, and GF glass fiber. ..................................................................................................................................... 58 Table 4.1 Comparison of the three wood species......................................................................... 89 Table4.2 List of treatments and mass fractions of components. .................................................. 90 Table4.3 Average contact angle (initial and after 5 seconds) and surface roughness values of different wood veneers (n= 5 replicates for each angle measurement; n=15 for each roughness measurement). ............................................................................................................................... 91 Table 5.1 List of treatments........................................................................................................ 125 Table 5.2 Mean compressive properties of laminates and the associated St. Dev. (n= 8 replicates). ................................................................................................................................... 126      x   List of Figures Figure 1.1 Relationship between Chapters 3 to 5 ........................................................................ 13 Figure 3.1 Douglas fir: (a) original sheet 0.6 mm thick; (b) woven sheet before trimming edges; (c) veneer edges deformation after densification; and (d) woven sheet of 2 mm veneer strips, circles show strips cracks and splits.............................................................................................. 59 Figure 3.2 Woven veneer sheet after edge trimming: (a) undensified 5 mm strips; (b) densified 5 mm strips; (c) undensified 12 mm strips; and (d) densified 12 mm strips. ............................... 60 Figure 3.3 Control samples: (a) ball resting on sample after impact, and (b) after removing ball........................................................................................................................................................ 61 Figure 3.4 Comparison of failure patterns: (a) densified veneer; (b) UOP; (c) WOP, top surface; (d) WOP, bottom surface; (e) circular crack in G-UDP; (f) starlike crack in GP; (g) CrossOP; and (h) G-CrossOP. ............................................................................................................................. 62 Figure 3.5 Comparison of mean impact energy of treatments (error bars represent 95% confidence intervals). .................................................................................................................... 63 Figure 3.6 Light micrographs of veneer cross section before and after densification; densification happened in tangential direction of the growth rings (vertically in the micrograph)........................................................................................................................................................ 64 Figure 3.7 SEM micrographs of resin distribution; (a) control veneer sheet – no resin applied; (b) wood-polyester composite; (c) tracheids pit in wood-polyester composite, (d) the individual glass fibers surrounded by polyester matrix in GP composites. ................................................... 65 Figure 3.8 Relationship between Chapters 3 and 4 ...................................................................... 66 xi   Figure 4.1 Lay-up of two cross-ply wood composite laminates made of four lamina; Balanced (left), and unbalanced (right); arrows indicate the fiber direction. Resin applied to the top face of the layers one to three is not shown here. ..................................................................................... 92 Figure 4.2 Composite laminates after curing: (a) glass fiber/polyester, (b) unidirectional wood/polyester, and (c) cross-ply wood/polyester. ...................................................................... 93 Figure 4.3 Light micrograph of the resinated veneer sheet in a polyester block. ........................ 94 Figure 4.4 Light microscopic images of unidirectional and cross-ply polyester laminates: (a) polyester as a white strip between layers; (b) filled cross sections of longitudinal tracheids with resin; (c) resin discoloration and area of lack of polyester; (d) neat adhesion among layers; (e) insufficient adhesion and voids; and (f) crack, void, and empty vessel elements. ....................... 95 Figure 4.5 Unidirectional and cross-ply Douglas fir laminates after impact testing: (a) unidirectional, front surface; (b) unidirectional, back surface, fracture in all layers; (c) cross-ply, front surface; and (d) cross-ply, back surface, controlled fracture in alternative layers. .............. 96 Figure 4.6 Comparison of the mean impact energy of wood/polyester and glass fiber/polyester composite laminates (error bars represent 95% confidence intervals). Note: Impact energy values of Thin GP and Thin WP were taken from the previous study (Haghdan et al. 2015a). .............. 97 Figure 4.7 Highest initial angle of polyester droplet on Oak Control (top), and lowest angle of polyester droplet after 5 seconds, DF 120 (bottom). .................................................................... 98 Figure 4.8 Contact angle of polyester droplets on wood veneers (before and after sanding) vs. time: (a) Douglas fir veneer; (b) Maple veneer; and (c) Oak veneer. ........................................... 99 Figure 4.9 Contact angles of polyester droplets on Oak veneer before and after sanding with grit size of 120 (angles at 0.5 second). .............................................................................................. 100 xii   Figure 4.10 Surface roughness of veneers as a function of sandpaper grit size; Error bars represent 95% confidence intervals. ........................................................................................... 101 Figure 4.11 Roughness profiles of the Oak veneer before and after sanding (a) Control Oak, (b) sanded with grit 120, and (c) sanded with grit 320 grit sandpaper ............................................. 102 Figure 4.12 Relationship between Chapters 4 and 5 .................................................................. 103 Figure 5.1 Schematic representation of the symmetry and balance in composite laminates (a 3D representation of stacking plies in a symmetric-balanced specimen is shown in the right side where thereby arrows indicate fiber direction). .......................................................................... 127 Figure 5.2 (a) Demolded and trimmed laminates; and (b) a typical CLC test machine. ........... 128 Figure 5.3 Trapezoid method for energy absorption calculation of laminates up to the failure point. ........................................................................................................................................... 129 Figure 5.4 Average load-displacement curve in wood/polyester and glass fiber/polyester laminates (error bars represent 95% confidence intervals). ........................................................ 130 Figure 5.5 Light micrographs of 8SA samples (a) centre delamination, and (b) multiple fractures in all layers. ................................................................................................................................. 131 Figure 5.6 Light micrographs of (a) buckling in 8SB, (b) buckling and delamination in 8SB, and (c) buckling in commercial Plywood. ......................................................................................... 132 Figure 5.7 Light micrographs of fractured 12 layers laminates: (a) delamination and transverse cracking of face layers in 12SB, (b) multiple fractures in all layers of 12NS, and (c) delamination between 4th and 5th layers of 11PW. ......................................................................................... 133 Figure 5.8 SEM micrographs of fracture mechanisms of 8SA laminate where b is a higher magnification of the area shown in the white square in a. .......................................................... 134 xiii   Figure 5.9 SEM micrographs of fracture mechanisms of 12SB laminate where b shows a higher magnification of the area shown in the white square in a. .......................................................... 135 Figure 5.10 Fiber bridging mechanism in 12SB sample. ........................................................... 136  xiv   List of Abbreviations ºC    degree Celsius  ANOVA Analysis of variance  ASTM  American Society for Testing and Materials CA  Contact Angle cm  centimeter  g   gram  h   hour J   Joule μm   micrometer m   metre  min   minute  mm   millimeter  MPa   Megapascal N   Newton Nm   Newton-meter Pa   Pascal PVC   Polyvinyl Chloride s   second  SEM   Scanning Electron Microscopy St. Dev  Standard Deviation UBC   University of British Columbia            xv   Acknowledgements  I very much appreciate the constant support and guidance of my supervisor, Dr. Gregory Smith, who gave me the invaluable opportunity to pursue my research interests and believed in my abilities and potential for the duration of my studies. His challenging questions and comments helped me to enrich the research concept and improve my critical thinking. Special thanks go to my supervisory committee members, Dr. Thomas Tannert (UBC) and Dr. Chunping Dai (FP-Innovations), for their technical advice, constructive criticism, and their extensive discussions around my work.   The financial support for this study was provided by a combination of the UBC Graduate Scholarship, and discovery grants from Natural Sciences and Engineering Research Council of Canada (NSERC). I would like to acknowledge UBC’s Centre for Advanced Wood Processing and Materials Engineering Department for making their test equipments and microscopes available for me when needed.  Many thanks go to Vincent Leung and Lawrence Günther at UBC’s Centre for Advanced Wood Processing for their technical advice and help in the use of machinery and testing samples. Last not least, I am thankful to my friends Shaghayegh Akhtari, Kamyar Gordnian, Dr. Navid Zobeiry, Dr. Alireza Forghani, Dian Xian, Solace Sam-Brew, Kunqian Zhang, Ying-Li Tsai, Jörn Dettmer, Felix Böck, and Dr. Kate Semple for their constructive comments on my project and for their unconditional support. xvi   Dedication  I gratefully dedicate this dissertation to my family.1   Chapter 1: Introduction  1.1 Background Composites overcome the limitations of a single material by combining two or more materials that have different but complementary properties; when combined intelligently, the properties of the composites can be vastly superior to that of either material alone. Composites are usually composed of two materials: a matrix phase and a reinforcing phase (Harris 1986). The matrix can be in the form of metal, ceramic, or polymer and the reinforcements can be particles or fibers. Fiber reinforced polymer (FRP) composites are commonly used in aerospace, automotive, marine, and construction industries where natural and synthetic fibers are embedded in polymer matrix.    Adding reinforcement to a polymer is a way to improve physical and mechanical properties of the resulting composite. Natural fibers offer both cost savings and a reduction in density when compared to synthetic fibers (e.g. carbon fiber and glass fiber). The most common natural fiber is wood fiber whereas glass is the most common synthetic fiber (Maldas and Kokta 1992; Thwe and Liao 2002; Jiang and Kamdem 2004; Reis et al. 2007). Wood fiber has a higher elastic modulus than many polymers and makes it a good candidate to reinforce those polymers (Clemons 2002; Bledzki and Faruk 2003). The matrix can also improve dimensional stability by surrounding the reinforcement and reduces the rate at which water is absorb into the wood phase (Wang and Morrell 2004; Tajvidi et al. 2008). Compared with wood, glass fibers provide considerably higher stiffness and strength to a polymer composite because of their high aspect 2   ratio (Gupta et al. 2001) but they are denser and more expensive compared to wood (Wambua et al. 2003).    There are several studies on the physical and mechanical properties of wood/polymer composites. Wang and Morrell (2004) investigated the water absorption behavior of wood polymer composites over long-term periods. Based on their results, there is less risk of shape change and fungal degradation for wood/polymer composites compared with wood products. Similar research shows that the physical and mechanical behavior of wood/Polyvinyl Chloride (PVC) composites is higher than either the wood or polymer separately (Chetanachan et al. 2001). Compared to wood fibers, synthetic fibers provide higher mechanical properties due to their higher aspect ratio. Long synthetic fibers facilitate stress transfer in the composite and enhance the composite’s modulus. It has been reported that the flexural strength of composites made of glass fiber and polyester resin is significantly improved by increasing the volume fraction of glass fiber (Avci et al., 2004; ).    Some work has been done on a new generation of polymer composites that are reinforced with combinations of natural and synthetic fibers in an effort to improve the mechanical properties of the final product and decrease the material costs (Short and Summerscales 1980; Jiang and Kamdem 2004). To distinguish these more complex composites from the traditional fiber composites containing only one fiber type, we will refer hereafter to a composite containing two or more reinforcement phases as a hybrid composite. Hybrid panels can be lighter and lower in cost while still possessing the required high stiffness, strength, and dimensional stability (Short and Summerscales 1980; Liang and Li 1998; Thwe and Liao 2002; Thwe and Liao 2003; 3   Jiang and Kamdem 2004; Tungjitpornkull et al. 2007; Cui and Tao 2009; Valente et al. 2011; Tewari et al. 2012).  Jiang and Kamdem (2003) and  (2004) reported that un-notched and notched impact strength of Red oak/PVC composites significantly improved by adding only 5% short glass fiber with no decrease in flexural properties. Jarukumjorn and Suppakarn (2009) examined combinations of layers of glass and sisal fibers in a polypropylene matrix. The authors reported an increase in flexural, tensile and impact strength without any significant negative changes on modulus and also improvement in thermal properties and water resistance of the polymeric composite. In similar study by Thwe and Liao (2002) a combination of glass fiber and bamboo fiber improved the stiffness and the strength of the polypropylene composite and enhanced its moisture resistant behavior. It was found that hybrid composite with hemp/glass fiber in a polypropylene matrix provided a 34% improvement in notched impact strength (Panthapulakkal et al. 2006). The positive effect of glass fiber as a second reinforcement for PVC composite has been reported by Maldas and Kokta (1992). Laminated hybrid composites with a core of hemp/polypropylene and skins of glass fiber/polypropylene were examined by Reis et al. (2007). The results indicated an improvement in fatigue performance of the composites. Based on the review on the literature of hybrid composite, it is believed that glass fiber is a good candidate for a second reinforcement in a natural fiber/polymer composite.    There are four common methods for manufacturing polymer composites: 1. Compression molding (CM) is a closed system in which the thermosetting molding material (reinforcement and resin) is pre-heated in an open heated mold cavity (Todd et al. 1994). The mold is closed and pressure applied to force the material into all areas of the mold. The heat and pressure are maintained until the material has cured; 2. Resin Transfer Molding (RTM) where the 4   thermosetting resin is injected into a closed mold in which layers of reinforcement have been previously laid. The mold remains closed until the resin cures (Strong 2008); 3. Hand lay-up where the mold is prepared with a release agent and then the liquid resin is applied on the dry reinforcement laid into the mold. The wet composite is then wiped with a squeegee to spread the resin uniformly and reduce the voids, and 4. Extrusion where a molten mixture of fiber and thermoplastic resin is pushed through a die to make a continuous profile of the desired cross section which is then cooled in a water bath (Clemons 2000; Bauser et al. 2006). In the current study a combination of hand-lay-up and compression molding were used to fabricate the composites.   Thermoplastic and thermoset polymers are commonly used in manufacturing of composite materials. There is a lot of research on combination of wood with thermoplastic polymers for example Wolcott and Englund (1999) and Chetanachan et al. (2001). But limited research is available for wood and thermoset polymers combination (Sèbe et al. 2000; Marcovich et al. 1996; Marcovich et al. 1998b; Nouwezem et al. 1994). Current industrial composites are using thermoset polymers, such as epoxy, to improve the stiffness and strength of the final product. Epoxy is recognized as a high stiffness and strength matrix for manufacturing industrial composites but its high cost limits its applications.  In contrast, general-purpose waxed polyester resin has mechanical properties slightly lower than epoxy and is also less expensive than it (Husseinsyah and Mostapha 2011).   5   1.1.1 Impact behavior of wood/polymer composites  Some research have been performed on the fracture of wood composites when they are subjected to an impact load. Razi and Raman (2000), studied the impact properties of Pine wood/High Density Polyethylene (HDPE) to analyze the effects of wood concentration and its particle size on impact damage resistance where they are subjected to drop-weight impact load. Results indicated large chips increased the energy to fracture and damage resistance significantly. The internal friction forces encountered between such large chips restricted deformation. Better impact performance of composites with higher concentration is believed to be as a result of having more obstacles for compressive deformation, crack growth, and fracture. Fracture of the impact specimen occurred by crack initiation due to debonding at the interface of wood and polymer and propagated along the weak interface. The compressive and bending strengths of a local tropical wood Kapur (Dryobalanops species.) in combination with a polymer were studied by Boey et al. (1987). Results showed a remarkable increase in these mechanical properties because of suitable impregnation of polymeric monomers into the wood cellular structure and subsequent irradiation to form a wood-polymer composite.    In drop-weight impact test a mass vertically drop on the sample (ASTM D5420 2010). Usually a tube or rails guide the falling mass. The results of this test are essentially pass/fail, because the impacted weight can stop dead on the sample (pass) or break it (fail). Both the mass and the drop height can be varied. The impact energy is the kinetic energy of the mass at impact. The energy absorbed by the test specimen is the impact energy required to just fracture or break the specimen and shows specimen’s energy absorption ability. This test can be used for thin samples like plates.  6   1.1.2 Energy absorbing ability of composites Two candidate applications for FRP composites are in the automotive and aerospace industries. The key requirements for these applications is the energy absorbing ability of the structure. For example consider floor support beams, or stanchions, in an airplane.  Stanchions are short almost vertical columns under the floor that are designed to crush in the event of a hard landing. Their role is to decelerate the passengers more gently and reduce the injuries that would otherwise occur. Stanchions are made of glass fiber/epoxy composites which, despite its high mechanical properties, are heavy and expensive (Taher et al. 2006). Composites made of natural reinforcement are currently used in appearance panels or as sandwich panels in industrial applications but there is a lack of research on their potential to be used as light-weight energy absorbers alternative to synthetic fiber composites.   An energy absorber dissipates kinetic energy either partially or totally into another form of energy. This dissipation strongly depends on the method of load application, deformation patterns, and material properties. It is either reversible as elastic strain in solids, or irreversible in plastic deformation (Alghamdi 2001). Wood is a good energy absorber and the parameters used to quantify this ability depend on the level of energy absorption required (Kretschmann 2010). Common criteria that describe the energy absorption at progressively more severe failure are: (i) work to proportional limit, (ii) work to maximum load, and (iii) work to total failure (i. e., toughness). Energy absorbing capability usually defines as Specific Energy Absorption (SEA) or crashworthiness which is the ratio of the energy absorbed by the material to its mass (Mamalis et al. 1997). SEA in composites varies depending on the fiber type, matrix type, specific geometry, processing conditions, fiber volume fraction, and testing speed (Farley 1986; Ramakrishna and 7   Hull 1993; Jacob et al. 2002). This property is measured in a compression test where the specimen compressed under axial load until failure; loading can be either static or dynamic.   Some studies have been done on crush energy absorbing ability of composite tubes. Hamada et al. (1992) studied the fracture behavior of carbon fibre/epoxy and carbon fibre/PEEK tubes made from unidirectional prepreg materials under axial compressive loads. Three fibre architectures were investigated including unidirectional fibres parallel (0°) to the axis of the tube, ±30° and ±45° and a very high specific energy absorption obtained in the 0° carbon fibre/PEEK tubes which are reported as the highest value available in the literature for any material. In research by Farley (1986), static crushing tests were conducted on square cross section composite tubes of graphite/epoxy and Kevlar/epoxy to study the influence of specimen geometry on the energy-absorption capability of these composites. The results indicated the tubes inside width-to-wall thickness (W/t) ratio significantly affect the energy-absorption capability of composite materials where it has a nonlinear relation with energy absorbing ability. An increase in energy absorbing ability of composite tubes was reported when W/t ratio decreased. Crushing behavior of corrugated composite tubes was also examined by Elgalai et al. (2004) where they tested two types of composites including carbon fibre/epoxy (in a filament form) and glass fibre/epoxy (in woven roving form). The corrugated angle in tubes was from 10° and 40°. The results showed the crushing behavior of corrugated composite tube is dependent on the change in corrugation angle and fibre type. Carbon/epoxy tubes with corrugation angle of 40° had the highest specific energy absorption ability. It was found that corrugation in tubes could significantly enhance the energy absorption ability of composite tubes in a uniform manner.   8    There are several crush mechanisms for a composite under compressive load. The failure of the structure can be catastrophic or progressive. When there is an unstable crack growth in the structure, the load increases suddenly to a peak value followed by a low post failure load. High peak load results in low energy absorption in the structure and makes the failure catastrophic. Characteristic types of progressive crushing modes are fragmentation, splaying, brittle fracture and progressive folding for synthetic fiber/polymer composites (Jacob et al. 2002). The crush fracture characteristics, however, are not identified for wood/polyester composites. Chapters 3 and 5 of this dissertation discuss the fracture mechanisms of wood composites.  1.1.3 Wood densification Wood is a cellular solid that in many ways is similar to a foam (Gibson and Ashby 1999). Under that right conditions of temperature and pressure it is possible to compress samples of low density species to much higher densities. With this increase in density comes a corresponding increase in strength and stiffness that makes these new densified woods appropriate for flooring and furniture applications for which higher density hardwoods are traditionally used (Babatunde et al. 2008; Kutnar and Šernek 2007; Diouf et al. 2011; (Fang et al. 2012).   First generation compressed woods were produced in 1930 under the name of Lignostone, Lignifol and Staypak (Hill 2007). This type of densified wood, however, was very unstable. It recovered its original shape when it came into contact with water or was re-heated. In order to improve dimensional stability after densification, the wood was heated and steam applied during densification.  Densification is usually performed at temperatures above the glass transition of hemicellulose and lignin, 150ºC, where polymer mobility and molecular 9   rearrangement is possible and facilitates the densification process. Therefore wood can be compressed above its Tg without breaking the cell walls. In addition, wood is a porous material and these pores can be compressed, with effort, until the density reaches to its cell wall density, approximately 1.5 gr/cm3 (Kellogg and Wangaard 1969; Simpson 1993; Rautkari et al. 2010). After densification, the wood is harder and darker in color. Severe darkness is expected at temperatures above 200 ͦ C due to the chemical changes in certain wood components. Above 200°C hemicellulose hydrolyzes and an increase in acid-insoluble lignin results the veneer darkness (Diouf et al. 2011).   There are several densification methods which rely on the combination of heat, steam and pressure: (i) Thermo-hygromechanical (THM) or closed system, (ii) Thermo-mechanical (TM) or open system, and (iii) Viscoelastic thermally compressed wood (VTC) (Navi and Sandberg 2012). Densification results in internal stress in the semi-crystalline region of microfibrils. Post-treatment of densified veneers at high temperatures and in saturated steam is necessary to release this stress (Cloutier et al. 2008).   Cloutier et al. (2008), densified aspen veneers by steam-injection hot press under 550 KPa steam pressure and average hydraulic pressure of 7 KPa to reach high-density veneers. Veneers were pre-heated by steam at three temperatures: 200, 220 and 240 ͦ C, and compressed from initial thickness of 3.2 mm to a final thickness of 1.6 mm.  There was a significant increase in density due to the cell wall buckling. Above 200 ͦ C, cell wall degradation occurred and the samples darkened significantly. Densification enhanced strength, stiffness and impact properties of veneers and reduced water absorption and thickness swell (Morsing 2000; Rautkari et al. 10   2010). No information, howeverm, is available on the effects of using densified veneer in fabrication of polymer composites.   1.2 Research problem statement There is a trade-off between stiffness and energy absorbing ability of a material. Depending on the candidate application, demanded properties would be different. Current energy absorbers are either relatively heavy (steel or aluminum) or expensive (synthetic fiber composites). Wood, in contrast, has a high stiffness and strength to weight ratio and exhibit a higher deflection before failure.    Despite the extensive research on various wood/polymer composites, limited information is available on the impact behavior of wood and the polyester matrix. In the current study, the polyester resin distribution around the wood component is investigated using SEM micrographs and differences in fracture patterns and impact performances between wood/polyester and glass fiber/polyester composites are examined and discussed.     Although the densification raises the modulus and strength of solid wood, there is little information available on the veneer densification of and the resulting changes in the mechanical properties of the composite.    Energy absorption is usually of significant interest when a structure is subjected to the impact or crushing. The study of fracture mechanisms and energy absorbing properties becomes more complicated when the composite structure is fabricated in the form of laminates. Energy 11   absorbing behavior of synthetic fiber/polyester laminates under quasi-static compression and shear loading is well-known as it is thoroughly studied by researchers.  The dominant fracture mechanisms and energy absorbing behavior of wood/polyester laminates, however, is not well identified in the literature.   1.3 Research objectives The principal aim of this research is to elucidate the dominant fracture mechanisms in wood/polyester composite laminates and to compare the energy absorbing ability of these composites with glass fiber/polyester counterparts. The main objectives are to:  1. Gain a better understanding of fracture behavior of wood veneer and polyester composites under the impact load, and to characterize the wood veneer densification effects on the final properties.   2. Evaluate the potential of wood veneer as an alternative raw material for manufacturing polyester laminates, investigate the effect of wood veneer lay-up configurations on the impact resistance of the composites, and to examine the roughness and wettability behavior of wood veneers to the polyester resin.  3. Investigate the effect of symmetry, balance, and number of lamina on the compressive properties of wood veneer/polyester composites, to compare the crush mechanisms of wood/polyester and glass fiber/polyester laminates, and to identify the dominant mode of fracture in each treatment. 12    1.4 Organization of the dissertation This thesis consists of an introduction chapter, a literature review chapter, three research chapters, and a conclusion. Chapter 1 introduces the problem and explains the thesis objectives. Chapter 2 is a detailed review on the improved physical and mechanical properties of composites made of wood and plant fibers in a combination with a polyester matrix. The next three chapters elaborate on the procedures and findings of the research conducted to achieve the stated objectives.    The relationship between chapters is shown in Figure 1.1. The impact energies of various polyester composites reinforced with wood and glass fibers is measured using a drop-weight impact test. The results are presented in Chapter 3 where the effects of wood reinforcement configuration and densification on impact properties of wood veneer/polyester composites are discussed. Findings of the study including fracture pattern, mat configurations, and resin distribution were used to design the next experiment. Laminates of Douglas fir, maple, and oak were fabricated using a combination of hand lay-up and compression molding. Wettability behavior of the used veneers to the polyester matrix and impact test results were presented and discussed in Chapter 4. Results of this experiment including the veneers’ wettability and effect of the mat configurations were used in Chapter 5. This chapter presents a detailed discussion on fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading and compares these mechanisms in terms of laminates symmetry, lay-up balance, and number of lamina. Chapter 6 summarizes the conclusions of Chapters 3 through 5 and provides potential research directions for future work.  13   Figure 1.1 Relationship between Chapters 3 to 5      Chapter 3  Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites • Fracture patterns  • Mat configurations • Resin distribution  Chapter 4  Wettability and impact performance of wood veneer/polyester composites Chapter 5  Fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading • Wettability and impact tests • Mat configurations  14   Chapter 2: Natural fiber reinforced polyester composites: a literature review  2.1 Synopsis Many composite products are made of thermosetting polymers reinforced with synthetic fibers. Despite the high mechanical properties associated with these fibers they are heavy and expensive compared with natural fibers. The use of natural plant fibres, combinations of natural and synthetic fibers, and wood furnish as reinforcement in polyester matrix for making low cost engineering materials has generated much interest recently. Natural fibers with good specific stiffness and strength, low density, low embodied energy, and good biodegradability have an advantage over synthetic fibers. Despite these benefits they have poor compatibility with the matrix due to their hydrophilic nature. The objective of this chapter was to review the literature on the effects of chemical treatments on fiber-matrix interfacial adhesion and the wettability of natural fibers by polyester. The efficiency of incorporating glass fiber into the natural fiber for the purpose of reducing water uptake and increasing the stiffness of composite is also discussed.   2.2 Plant fiber/polyester composites 2.2.1 Hemp fiber Hemp fiber is the most common natural fiber for reinforcing polyester composites. The composites can be manufactured using hand lay-up, compression molding, or Resin Transfer Molding (RTM) methods.  The drawback of using plant fibers as reinforcement is their susceptibility to water adsorption due to their hydrophilic nature. Surface modification of fiber, however, can limit their water uptake. The effects of hemp fiber loading, content, and chemical 15   treatments on energy absorbing ability, and on tensile, flexural, and water absorption properties of polyester composites are discussed in this section.   2.2.1.1 Fiber loading and fiber length effects The energy absorbing mechanisms of hemp fiber/polyester composites associated with the impact fracture have been investigated in detail by Sanadi et al. (1986a). Composites were made using hand lay-up method and their Izod impact toughness were measured. Results of this study showed the toughness of composites at volume fraction of 24% to be 15 times higher than the pure polyester resin. SEM micrographs of fracture surface indicated that various fracture mechanisms including fiber pull-out, plastic flow of the lignin-hemicellulose matrix, hemp fiber splitting, and crack extension at fibril-fibril interface contributed to energy absorption. In similar study by Dhakal et al. (2007) different hemp fiber volume fraction, up to 26% , were used to make non-woven hemp fiber/polyester composites. A combination of hand lay-up and compression molding methods were used to manufacture these composites. Results indicated the percentage of moisture uptake increased with fiber volume fraction. The increase was attributed to the additive amount of cellulose content at higher fiber volume fractions. Swelling of the hydrophilic hemp fibers in these composites led to micro cracking of the brittle thermosetting resin. The high cellulose content in hemp fiber further contributes to more water penetrating into the interface through these cracks. The swelling stresses led to fiber-matrix debonding and final failure of the composite.  Degradation of fiber-matrix interface significantly reduced tensile and flexural properties of the composite. It is shown in similar research that water absorption in hemp fiber composites increases with increasing fiber content. This increasing trend, however, was relatively slow after fiber saturation (Rouison et al. 2005). When the fiber content increased to 16   40%  a significant linear trend between tensile properties and fiber volume fraction was observed (Sanadi et al. 1986b).  The tensile properties of the composites increased with fiber volume fraction, as one would expect. This increasing trend was also reported by Sèbe et al. (2000) for tensile and flexural strengths. In addition to the higher cellulose content at higher fiber loading, low microfibril angle of sun-hemp fibers attributed to the linear relation.    Generally, an irregular fiber surface improves the strength of fiber-matrix bond due to its greater surface area. Nevertheless, the presence of waxes and fatty acid by-products on the fiber surface prevents a chemical bond between fiber and matrix and results in poor adhesion instead. The use of a coupling agents or fiber surface modification was recommended to improve the fiber-matrix compatibility. Dabade et al. (2006) examined the effects of fiber length, ranges from 10 to 70 mm, on tensile properties of hemp fiber/polyester composites. Random orientation of fibers in polyester matrix were used to have isotropy in the composite and have identical properties in all direction. Results indicated tensile strength of hemp fiber composites increases with fiber length up to 30 mm. Further increase of fiber length, however, decreased the tensile strength as fiber entangling limited fiber-matrix interaction.   2.2.1.2 Chemical treatments effects In a study by Rouison et al. (2005) hemp fibers were chemically treated by Styrene Maleic Anhydride (SMA), Alkylketene Dimer (AKD), Rosin Acid, and NaOH to investigate the effects of chemical treatments on the water absorption of hemp fiber/polyester composites. Samples were either immersed in water or exposed to air with a relative humidity of 94%. Among all the chemical treatments, SMA treatment improved the hydrophobicity of the fibers. The 17   improvement, however, was not statistically significant. Esterification, another chemical treatment, was applied to hemp fibers and composites were made with polyester matrix using RTM technique (Sèbe et al. 2000). The work showed that esterification treatment improved fiber-matrix bonding and resulted in an increase in flexural stiffness and strength of the composites. SEM micrographs of the fracture surface showed different textures in treated and untreated samples; the surface of untreated composite was very fibrous in appearance and had large amounts of fiber pull-out while in the treated composites it was smooth with small amount of debonding. The impact strength of the composites containing esterified fibers, however, decreased significantly. Inspection of the fibers using polarized light microscopy revealed the presence of pre-existing defects along the fibers. Authors concluded that the poor toughness performance was due to these pre-existing defects.   2.2.2 Jute fiber Compression molding and hand lay-up are the most common manufacturing methods of jute fiber/polyester composites. Both unidirectional jute fiber and woven jute fabric can be used as reinforcements in the polyester matrix to increase mechanical properties. The effects of fiber volume fraction and fiber chemical treatment on tensile, flexural, impact, shear and compression properties of jute/polyester composites and their fracture mechanisms and interface properties are discussed in this section.  2.2.2.1 Fiber loading effects Volume fraction is one of the important factors on mechanical performance of fiber reinforced polyester composites. In a study by Roe and Ansell (1985) untreated jute fiber/polyester 18   composites were made using hand lay-up method with fiber volume fraction up to 60%. Tensile stiffness and strength, Charpy impact work of fracture, and interlaminar shear strength of these composites were measured. Findings of this study showed that increasing fiber volume fraction in the composite results in an increase in tensile strain to failure. Stretching of jute fibers is a contributing factor to increase the tensile strain and also to energy absorption. Despite the good tensile performance a very poor interlaminar shear performance was reported for these composites. Having different fiber volume fraction not only affects mechanical properties but also the fracture mechanisms; for volume fraction up to 10%, fracture in tension was microscopically brittle, with fiber pull-out. Insufficient or absence of fiber-matrix interface area was usually the cause of poor interaction. By increasing fiber content the pull-out length increased and fracture paths were macroscopically longer. Due to the poor fiber-matrix adhesion and isotropy of the composites fiber pull-out was the dominant type of failure in interlaminar shear test.  Using woven jute fabric or stitch bonding of fibers were recommended to improve the shear strength of the composites.    Compressive properties of jute/polyester composites was also the point of interest. Using hand lay-up technique, untreated unidirectional jute fabric/polyester composites with fiber loading up to 45 wt%, were prepared and tested according to ASTM D3410  (Gowda et al. 1999). This standard applies a combination of shear and compression loads to the test samples in order to avoid end crushing. Fiber buckling during the test resulted a shear fracture at fiber-matrix interface and this led to ultimate failure of the samples. The final failure of composites was progressive which is beneficial for shock applications where high energy absorbing is required. Although authors did not perform any analysis on the fiber-matrix adhesion, the 19   progressive failure of the composites indicates the low stiffness value and high strain to failure in composites. In a research by O’Dell (1997) composites of jute fiber/polyester were fabricated at 10-15% fiber loading using RTM technique. Tensile, flexural, and Izod impact properties of these composites were measured and compared with glass fiber/polyester composites. Results showed that tensile and flexural strength and tensile and flexural stiffness of jute composites were half of those in glass fiber composites. SEM micrographs of the impact fracture surfaces also showed fiber pull-out failure, indicating a poor fiber-matrix interactions. It is recommended that using a coating or fiber treatment may improve the interfacial adhesion and as a result the impact strength.  2.2.2.2 Chemical treatments effects The poor interface interaction between untreated jute fiber and polyester matrix usually results in the low tensile and flexural stiffness and strength for the composite. De Albuquerquen et al. (2000), studied the effects of jute fiber chemical treatment and the presence of a wetting agent in the polyester matrix on the interface interaction and mechanical performances of unidirectionally oriented jute composites.  Different volume fractions of fiber were used up to 30%. Two types of polyester resin, with and without a wetting agent, were also used as the matrix. In high-speed manufacturing processes it is likely to get incomplete fiber wetting; this consequently results in formation of voids or air entrapment, interfacial defect, and lower adhesive bond strength. Jute/polyester composites of this study with the wetting agent in matrix, had superior mechanical properties than those without the agent due to the better fiber surface wettability. SEM analysis showed the development of a rough topography and adherence of polyester particles on the fiber 20   surface as the result of alkali treatment. This treatment significantly improved composites performance as higher wettability was achieved at the interface.   2.2.2.3 Long-term water absorption effects Seki et al. (2011) investigated the effects of water aging on the fiber-matrix interaction of woven jute fiber/polyester composites. Laminates of jute fabric and polyester were first manufactured using hand lay-up techniques and then compressed at room temperature. The composites were immersed in tanks of distilled water and synthetic seawater for up to 114 days at 34°C. During water uptake, bonding strength between fiber and matrix material was weakened owing to water molecules entering the interior structure of the jute/polyester composite. Results of short beam shear test showed that the interlaminar shear strength of the composites was significantly reduced by more than 62% in salt water and 53% in distilled water. The effect of long-term water absorption on tensile, compression, shear, impact, and flexural strength of untreated woven jute fabric and isothalic polyester composites have been studied by Ahmed and Vijayarangan (2007). Results of their finding indicated among all the measured properties, impact strength in particular, were significantly greater than those in unreinforced polyester resin. Water absorption, measured after 30 and 65 days, had detrimental effect on the tensile, flexural, and shear properties. Type of fracture varied depending on the type of test; SEM image of tensile-fractured surface of jute composite showed fracture of fiber with little pull-out. The authors concluded that interaction between fiber and matrix was poor in these composites as there was no trace of resin adhering to the fiber. Fiber microbuckling was the major failure mechanisms for the samples under compression and  matrix cracking was the dominant fracture mode for 21   interlaminar shear and flexural tests; jute composites failed under the impact load as a result of fiber splitting fracture.   2.2.3 Banana fiber Banana fiber which is obtained from the plant Musa Saptentum linn has high cellulose content and low microfibrillar angle, makes this fiber a good candidate to be used as reinforcement in polyester matrix. The effects of banana fiber length and content, and its chemical treatment on mechanical properties of banana fiber/polyester composites were the point of interest which are discussed in this section.   Pothan et al. (1997) studied the influence of banana fiber length, ranging from 10 mm to 40 mm, and content (from 20 to 50%) on the mechanical properties and aging characteristics of polyester composites reinforced with these fibers. Banana fibers of this study were also chemically treated with silane. Low fiber loading, below 10%, resulted in decrease in tensile stiffness. As mentioned, this volume fraction is not enough to impart high strength to the composites. Tensile, flexural, and impact strength had the highest value at fiber volume fraction of 40%.  Fiber chemical modification improved tensile strength of the composites by 28% and SEM micrographs of the fractured samples clearly showed the improved fiber-matrix adhesion in treated composites. As expected, water absorption was greater for the composites with higher fiber content due to the hydrophilic nature of cellulose in fibers. In a similar study by Sreekumar et al. (2008) composites of banana fiber and polyester were made using RTM technique. The authors used the same range of fiber length and content as those in Pothan et al. (1997). Increasing fiber length from 10 to 30 mm increased tensile, flexural and impact strength. Fiber 22   length above 30 mm, however, weakened the performances and was attributed to fiber curling and entanglement at longer fiber lengths resulted in a reduction in the effective length of fiber. Regarding the effect of fiber content, mechanical properties were increased as fiber loading increased up to 40%. Increasing the fiber content above this value, however, results a poor dispersion of fibers and the polyester matrix is not enough to wet the fibers. SEM micrographs of the fractured surface shows fiber pull-out, fiber breakage, and fibrillation as the dominant fracture mechanisms. Regarding the physical performance water absorption increased in banana fiber composites as the fiber content increased. In fact, the greater amount of hydrophobic cellulose at higher fiber volume fractions increased water absorption in the composites.   In a similar research by Prasad et al. (2009) mechanical properties of  banana empty fruit bunch (banana-EFB)/polyester composites including tensile, flexural and impact strength were measured as a function of fiber loading varies from 10 to 40%. Samples were made using a combination of molding and hand lay-up techniques. Results indicated the mean tensile strength of banana-EFB/polyester composites decreased up to a volume fraction of 11%, as the banana-EFB fiber acted more as filler than as reinforcement. Increasing the volume fraction up to 37%, however, increased this property due to the increased area of bonding in the interfacial region of fiber-matrix. Unlike tensile properties, the highest flexural stiffness and strength was observed at low fiber volume fraction of 11%. It was then decreased with increasing volume fraction. The authors believed poor fiber-matrix interaction and dispersion problem were the likely reasons of decreasing flexural properties with increasing fiber loading. The improved impact strength of banana composites at higher fiber volume fractions justifies these reasons. Usually poor interfacial adhesion improves toughness of the composite by introducing the pull-out fracture 23   which indeed is beneficial for energy absorption purposes. Zhu et al. (1995) used chopped banana fiber and to reinforce polyester matrix. Samples were fabricated using hand lay-up technique and at various fiber content up to 48% by weight. The optimum flexural stiffness and strength, and fracture toughness of banana fiber/polyester composites was obtained at 30% fiber weight fraction. It was surprising that the mechanical properties of the composite were lower than those of the neat polyester matrix when the fiber content was below 19%. The researchers speculated that the lower strength of the samples containing fiber content less than 19% was due to the fiber acting as a filler instead of a reinforcement.    There seems to be consensus that for banana fiber/polyester composites the highest tensile, flexural and impact strength is reached with 30 mm long fibers and at 40% fiber loading.  Chemical treatment of banana fibers is necessary to modify fiber structure and limits its hydrophilicity when the structure is exposed to a humid environment.  2.2.4 Sisal fiber Sisal fiber is usually obtain from the leaves of the Agave sisalana plant, which is largely available in tropical countries. The surface of this fiber must be chemically modified in order to make it more compatible with the polymeric matrix. Taking into consideration that fiber-matrix interface determines the mechanical properties of the composites, several experimental techniques including pull-out tests, fiber fragmentation tests, and fiber push-out have been used to characterize interface properties. The effects of fiber chemical treatment and manufacturing method on tensile, flexural, impact, hardness, and water absorption properties of sisal/polyester composites and the resulted fracture mechanisms and interface properties are discussed. 24   2.2.4.1 Chemical treatments effects In a study by Sangthong et al. (2009) Sisal fibers were treated by admicellar polymerization with a poly (methyl methacrylate) for the purpose of improving fiber-matrix interfacial adhesion in unsaturated polyester composites. Before treatment fibers were soaked in NaOH for wax removal. The concentration of MMA was varied from 0.025% to 0.1% volume per volume (v/v). Composites were prepared using a combination of hand lay-up and compression molding at different fiber volume fraction up to 40% and various fiber length up to 40 mm. Results of surface characterization of untreated and treated sisal fibers showed reduced moisture absorption of fibers in the latter. Improved fiber-matrix adhesion led to increase tensile, flexural, impact and hardness properties. This improvement showed a trend with fiber loading content as increasing fiber content resulted an increase in the mechanical properties. SEM micrographs of tensile fractured surface presented an improved interfacial bonding between sisal fiber and the polyester matrix with fiber breakage failure at the fracture point. In untreated composites, however, extensive fiber pull-out and interfacial debonding were the main reasons of composite failure. The authors recommended to manufacture polyester composites with treated sisal fiber at 30% volume fraction, with fiber length of 30 mm and, and MMA concentration of 0.075% v/v. Sydenstricker et al. (2003) investigated the effect of sisal fiber chemical treatment on its water absorption and characterized the tensile fracture surfaces of a single fiber embedded in the polyester matrix. Sisal fibers of this study were modified with two different chemicals; NaOH (0.25 to 10% concentration weight/weight (w/w)), and N-isopropyl-acrylamide (1, 2, and 3% w/w).  The degree of crystallinity of the different treated sisal fibers was also measured using X-ray diffractometer. Crystallinity index was increased in treated sisal fibers indicating the used chemical treatments extracted the amorphous portions of fiber (lignin and hemicellulose) and 25   made it more similar to a pure cellulose fiber. Density of sisal fibers decreased after chemical treatment because extraction of soluble products in NaOH and acrylamide. This lower density makes the treated sisal fiber a good candidate to be used as reinforcement in light-weight composite manufacturing. Although NaOH treated fibers showed the same water absorption as untreated fibers, acrylamide treated fibers had significantly lower uptake.  According to SEM micrographs both chemical treatments improved fiber texture by removing substances from its surface. This potentially makes the fiber more wettable to the matrix. Acrylamide treatment with highest tensile strength, lowest moisture absorption, and best performance in pull-out tests was recommended as the best treatment of this study. Despite the fact that chemical modification of sisal fiber improves the interfacial bonding with polyester matrix it also simultaneously resulted in loss of fiber strength. This loss, however, can be controlled using the proper chemicals for modification and choosing an appropriate temperature.  2.2.4.2 Manufacturing methods effects Sreekumar et al. (2007) investigated the effects of fiber length, loading, and manufacturing methods on tensile and flexural properties of sisal fiber/polyester composites. Two different techniques including RTM and compression molding were used to fabricate the composite samples. Results of tensile and flexural tests showed that both properties improved as sisal fiber length and content increased. Fiber length of 30 mm was found to be the optimum for effective reinforcement in polyester. The highest tensile and flexural properties were obtained at fiber loading between 40 to 50%.  Regarding the effect of manufacturing method, RTM composites had better performances under loads compare to the compression molded samples.  The author concluded that this was due to lower void content and better wettability (fiber-matrix interaction) 26   of the composites prepared with RTM. Higher void content in compression molded composites led to increase long-term water absorption. SEM micrographs of fractured tensile samples contained fiber breakage, fiber pull-out, and fibrillation as main fracture mechanisms. The extent of fiber pull-out was largest for compression molded samples which this is the reason of decreased mechanical properties in related composites.  2.2.5 Coir fibers Coir fiber is an important plant fiber obtained from the husk of coconut tree. This tree grows extensively in tropical countries. Low cellulose content (36–43%), high lignin content (41–45%) and high microfibrillar angle limits the use of this fiber in polyester composites. Some studies have been done on improving the interfacial fiber-matrix interaction using chemical modifications which are discussed in detail in this section.   In a study by Prasad et al. (1983) interfacial adhesion between coir fiber and a polyester matrix has been improved using alkali treatment. Two mechanical tests including tensile and fiber pull-out were performed on the treated fibers. Results indicated alkali treatment increased tensile strength of fibers by 15%, and the debonding stress of treated fibers from polyester matrix was 90% higher than that of untreated fibers. This had correlation with SEM observations showing the removal of cuticle and tyloses from the fiber surface after treating with alkali. Treated fibers, as a result, had a rough surface which led to improve their interaction with polyester matrix. Alkali treatment also prevented the flotation and segregation of fibers, led them to uniformly disperse in polyester. Due to the improved fiber-matrix adhesion treated fiber/polyester composites had properties 40% higher than those containing untreated fibers. In 27   addition to alkali treatment, coir fibers were treated by vinyl grafting and bleaching to investigate the effects of these chemical treatments on surface morphology and fiber interaction with polyester (Rout et al. 2001). Unidirectional coir fiber/polyester composites were fabricated using compression molding method. Among all modifications, bleached coir fiber reinforced polyester composite showed better performance under flexural load. The tensile strength, however, was significantly higher in alkali-treated composites. Besides improving the mechanical properties, the extent of water absorption decreased considerably after alkali treatments due to less voids and better compatibility of fiber and matrix. According to SEM micrographs of tensile fractured surface fiber pull-out was the dominant failure mode in untreated composites. In treated composites, on the other hand, reduction in fiber pull-out fracture and void contents indicates strong interfacial adhesion of fibers and polyester matrix.    Results of these two studies shows in order to get high physical and mechanical properties in coir fiber/polyester composites, the surface of coir fibers must be modified, preferably with NaOH and bleaching to get a better compatibility between fiber and matrix.   2.2.6 Kenaf fibers Kenaf (Hibiscus cannabinus) is made up an inner woody core and an outer fibrous bark surrounding the core. In a study by Ahmad et al. (2010) alkali treated kenaf fiber at various percentages (5 to 25% weight fraction) were used as reinforcement in polyester matrix. In some treatments polyester resin was mixed with a toughening agent, liquid natural rubber (LNR), to investigate its effect on impact properties of composites. Results indicated impact, flexural and fracture toughness properties all improved significantly at 20% fiber weigh fraction. Regarding 28   the effect of alkali treatment, composites contained treated kenaf fiber had superior properties than those which were not treated. The reason is alkali solution reacts with the cementing materials of the fiber, mainly hemicellulose and lignin and washed them away. The breaking down of the fiber structure into finer filaments increases the effective surface area available for wetting by the matrix. This improves the interfacial bonding between fiber and matrix and as a result the final performance of the treated composites as happened for the treated samples of this study. Rough surface of treated fiber which observed under SEM also proves the removal of hemicellulose and lignin. Addition of LNR was found to improve the impact properties and fracture toughness of composites due to the presence of rubber particles. Possibly, rubber particles acted as an impact modifier in the composites.  Flexural stiffness and strength, however, reduced in samples with LNR because of the plasticizing effect of rubber particles.    Two different kenaf fibers obtained from core and bast were used to investigate the effect of fiber size on tensile, flexural, and impact properties of polyester composites (Ishak et al. 2010). Results showed the superior performances of kenaf bast fiber composites to the other. Bast fiber with higher cellulose content (60.8% compared to 50.6% in core fiber), longer length, and smaller diameter successfully reinforced the polyester matrix. The highest tensile and flexural strength was seen at 20% weight fraction for both core and bast fibers. For impact strength, however, the optimum fiber content was 5%wt for kenaf core and 10%wt for bast fiber composites. The maximum fiber content to allow the fibers to fully get wet by polyester for both composites were subjected to these optimum fiber content. Beyond the optimal weight fraction there is poor interfacial bonding and hence lower mechanical properties.  29   2.2.7 Other plant fibers Other plant fibers including bamboo, rice, oil palm, and flax were also used to reinforce polyester matrix. Effects of chemical treatmnet, fiber loading, and length of these plant fibers on physical and mechanical properties of polyester composites are investigated and discussed in this section.   2.2.7.1 Chemical treatments effects Bamboo fiber is a cheap and fast-grown source of fiber. Despite extensive research on the effect of chemical modifications on final performance of various natural fibers/polyester composites, limited information is available for the bamboo composites. Kushwaha and Kumar (2009) investigated the effect of silane treatments on the water absorption properties of mercerized bamboo matting reinforced polyester composites. In this study after using 5% NaOH, silane coupling agent was added to the fiber. Results of short-term and long-term immersion of composites in distilled water showed that the water absorption was decreased in treated composites compared with untreated ones. Alkali treatment first removed hemicellulose, lignin, wax, and oils covering the external surface of the fiber and made the fiber more compatible with polyester. Improving water absorption properties was continued after adding coupling agent to alkali-treated fiber. The reason is inactive hydroxyl groups in bamboo fiber became active by alkali treatment which led to better silane deposition and a strong chemical interaction between the fiber and matrix was achieved. Authors recommended to use both alkali and silane treatment if wettability of bamboo fiber required. The effects of acetylation, silane or titanate coupling agents on mechanical properties of oil palm fiber/polyester composites were also investigated by Hill and Khalil (2000).  Results showed that oil palm fiber acetylation treatment increased 30   interlaminar shear strength between fiber and matrix, and tensile and flexural stiffness and strength in all acetylated composites than those in untreated. Since the surfaces of the cellulosic fibers became more hydrophobic by acetylation, the enhanced compatibility between fiber and matrix was expected. The effect of the other treatments, silane and titanate coupling agents, on mechanical properties of the composite was not significant.    In a study by Baley et al. (2006) flax fiber was modified with three different chemicals including sodium hydroxyl, acetic anhydride, and formic acid. The effect of these treatments on adhesion properties of flax fiber/microdroplet polyester resin were then investigated. Fiber surface tension was measured using micro-tension test; where an embedded single fiber in polyester matrix was pulled-out to calculate fiber-matrix adhesion. The results obtained from sodium hydroxyl plus acetic anhydride treatments and those with formic acid exhibited a general increase of the adhesion between flax fiber and polyester matrix. Sodium hydroxyl (NaOH) treatment, however, reduced the adhesion properties significantly. The poor performance of NaOH treatment alone was assumed to be due to the structural degradation of the flax fibers. It is recommended to take into consideration the severity of fiber surface chemical modifications to avoid such degradation. Although NaOH treatment, also called as alkali treatment, reduced adhesion properties of flax fiber/polyester samples but as it is mentioned earlier this treatment increased fiber surface roughness in the other plant fiber composites, led to improve their final mechanical properties. It can be concluded the bonding mechanism are remarkably depend on the nature of plant fiber itself.   31   2.2.7.2 Fiber loading and filler size effects Rice husk (RH) is an agricultural by-product from the rice milling process. Rozman et al. (2005) used RH as filler in polyester composites and investigated physical (water absorption and thickness swelling) and mechanical (tensile, flexural, and impact) properties of those composites. Three levels of fiber loading (42%, 57%, and 72%) and three levels of filler size (35-60, 60-80, and 100-140 mesh) were considered as treatments in this study. Finding of this study showed that all mechanical properties decreased as the percentage of filler was increased. This may be attributed to the decrease in the amount of matrix material or the increased irregularities in matrix. In terms of filler size, polyester composites with smaller size fillers had higher physical and mechanical properties than those made with larger filler size. In fact, the higher surface area of the RH in smaller filler size composites resulted in more interaction between hydroxyl groups of RH and the carbonyl, groups of the polyester matrix.  2.3 Hybrid fiber/polyester composites The term “hybrid composites” refers to a combination of two or more reinforcements (natural and/or synthetic) in a matrix in order to take advantages of all reinforcements and improve physical and mechanical properties of the final product.  2.3.1 Glass fiber-plant fiber/polyester Glass fibers were introduced into the plant fiber/polyester composites in order to improve their mechanical properties. In this section, the effect of this combination is discussed in detail.    32   2.3.1.1 Glass fiber-jute/polyester The effects of hybridization of glass fibers on notched sensitivity of woven jute fabric composites were evaluated using tensile test (Ahmed et al. 2007a). Four treatments of this study- one jute/polyester and other three jute–glass/polyester hybrid laminates- with different relative fiber weight fractions of jute and glass fibers were fabricated by simple hand lay-up technique in a mold at laboratory temperature. Finding of this study showed that rule of mixtures can also be applied in jute and jute-glass polyester composites for predicting the elastic properties. Composite stiffness increased by increasing glass fiber content in hybrid jute-glass composites. The high stiffness of glass fibers likely was the reason of this increase. According to the experimental values of unnotched and notched strength, jute composites exhibited the highest notch sensitivity when compared to hybrid laminates as indicated by highest reduction in their tensile strength. Observation of failed specimens showed that failure of the jute composite was sudden with little or no pull-out of jute fibers. In hybrid laminates, however, extensive fiber pull- out and breakage indicates the greater extensibility of glass fibers than jute fibers. It can be concluded that using glass fiber in combination with jute fiber improves the tensile strength and type of fracture in compare to the jute/polyester composite.    In a similar study by Ahmed et al. (2007b) the low velocity impact behavior and damage tolerance capability of jute-glass/polyester were measured through drop-weight impact test and post-impact tension test, respectively. Four drop height were used to get four different energy levels. Samples were then impacted by the weight and the load-energy-time plots were recorded. Results indicated the peak load was higher for hybrid composites than jute samples due to the greater stiffness of glass fibers. The impact energy absorption, however, was higher for 33   jute/polyester composites. Investigation of the fractured surfaces showed that matrix cracking and fiber breakage were the dominant modes of failure in jute composites whereas hybrid samples failed mostly due to delamination. Despite the higher impact energy absorbing ability of jute composites than hybrid composites they exhibited lower damage tolerance capability. Since hybrid composites presented a more gradual reduction in the normalized tensile stress, incorporation of glass fiber into the jute composite believed to improve the damage resistance and tolerance capability.    The changes in tensile and flexural properties as a function of jute and glass fibers content and as a function of ultraviolet radiation (UV) were investigated (Al-Kafi et al. 2006). Results of this research showed the best performances can be achieved at a jute to glass ratio of 1:3 and when both jute and glass fibers were UV treated. This can be explained by having the higher stiffness and elongation at break for glass fiber. Addition of glass fiber into jute/polyester composites also enhanced their impact strength due to bringing the higher extent of pull-out for the samples.  2.3.1.2 Glass Fiber-sisal/polyester Hybrid polyester composites of treated sisal fiber and coated red mud, an industrial by-product, were manufactured and their physical and mechanical properties were evaluated (Singh et al. 1995). 20% red mud, 34% sisal fibers and 8% glass fibers were incorporated in polyester matrix to fabricate the hybrid composite. Results indicated the tensile, flexural, and water resistance properties of sisal-red mud-glass/polyester composite were significantly higher than those without glass fiber. This superior performance was explained by the better inherent properties of 34   glass fiber. A mixture of chopped sisal and jute fiber of 30 mm length were used to make 5-layers hybrid glass/polyester laminate (Ramesh et al. 2013). Glass fiber layers were fixed in top, middle, and bottom of the specimen and second, and fourth layers were filled by the plant fibers. Finding of this study showed sisal-jute-glass/polyester composites exhibited superior flexural, impact, and compression properties to those in sisal-glass/polyester and jute-glass/polyester composite. This is likely due to having the advantage of reinforcing effects of both plant fibers in sisal-jute-glass/polyester composites. Authors did not report the physical properties of these hybrid composites.   2.3.1.3 Glass fiber-pineapple leaf and sisal/polyester  Hybrid composites of pineapple leaf-glass/polyester and sisal-glass/polyester were manufactured having glass fibers as skin layers and mentioned plant fibers as core (Mishra et al. 2003). The fabricated composite showed improved tensile, flexural, and impact strength compare to the pineapple leaf/polyester and sisal/polyester composites. Optimum glass fiber loadings for pineapple leaf-glass hybrid polyester and sisal-glass hybrid polyester composites were 8.6 and 5.7 % (wt), respectively. SEM micrographs of impacted samples showed improved energy dissipation mechanisms including fiber breakage, fiber fracture, and fiber pull-out. Since hydrophilic plant fibers are susceptible to high levels of water absorption, the addition of glass fiber to the composite can reduce the water absorption and thickness swelling. This superior physical performance has also reported for pineapple leaf and sisal fibers reinforced polyester composites.   35    In addition to studies on jute-glass fibers combination  Pothan et al. (2010) investigated dynamic mechanical properties of banana-glass fibers reinforced polyester composites. Banana fibers were cut at a uniform length of 30 mm and glass fibers were randomly oriented in the mat. Hybrid composites were fabricated at various glass fiber volume fraction (3 to 17%) and at different lay-up patterns, with banana as the face layer, glass as the face layer, and also as an intimate mixture of glass and banana. Results showed that the dynamic mechanical response of the hybrid composites was more affected by layering pattern than the glass fiber volume fraction.   2.3.1.4 Glass fiber-oil palm/polyester  A combination of oil palm empty fruit bunch and glass fiber were used as reinforcement in polyester composites (Khalil et al. 2007). Incorporation of both oil palm and glass fiber into the polyester improved tensile, flexural, and impact strength of the composite with the optimum properties at 35% of total fiber loading. The properties of oil palm-glass composite was found significantly higher than those with single oil palm reinforcement in polyester matrix. This improvement is due to the inherent higher stiffness and strength of glass fiber. Besides improvement in mechanical properties, incorporation of glass fiber into oil palm composite also reduced water absorption and thickness swelling as its tendency to water uptake is much lower than the natural oil palm fiber.  In similar research by Karina et al. (2008) flexural strength and physical properties of oil palm empty fruit bunch-glass/polyester composites were investigated. Results showed the addition of oil palm empty fruit bunch to glass fiber/polyester decreases flexural strength and density of the composite but increases its water absorption and thickness swelling. Incorporation of oil palm fiber up to 40%, however, resulted in similar flexural strength with the glass fiber/polyester composites. 36   2.3.2 Plant fiber-plant fiber/polyester  Athijayamani et al. (2009) investigated the effect of moisture absorption on tensile, flexural, and impact strength of roselle-sisal fiber/polyester composites under dry and wet conditions. Composites were fabricated at various weight fraction up to 40%. Roselle and sisal fibers incorporated in polyester matrix at a ratio of 1:1 and at various fiber length.  Increasing the fiber content and length at dry condition increased the tensile and flexural strength. When the samples were subjected to moisture environment a significant reduction of these properties occurred due to the degradation of fiber-matrix interface. Plant fibers of this study, specifically roselle, are well-known for their high cellulose content and their sensitivity to absorb water when exposed to wet conditions. There is one other research on combination of two plant fibers, cotton and kapok fabric, as reinforcements in polyester matrix with a cotton to kapok ratio of 2:3 (Mwaikambo and Bisanda 1999). The fabric was also treated with sodium hydroxide (NaOH) in order to improve the fiber-matrix adhesion. Some composites were subjected to weathering conditions to investigate the effect of weathering on tensile, flexural, and impact properties of the polyester composite. Inconsistency of the composites’ thickness and also in tensile properties of different area of the board was mentioned by the authors. In contrast to all previous studies which reported the trend between improving mechanical properties at higher fiber content, in this study the properties significantly decreased likely due to the poor manufacturing, insufficient fiber-matrix adhesion, and presence of voids. Treated fabric composites had lower properties than those of control composites. Extreme degradation of cellulose structure as the result of using high percentage of NaOH (5%) was believed to be the reason of the poor performance of treated composites. Authors recommended the use of NaOH must be limited up to 1%. Results of accelerated weathering at 100°C did not show the degrading effect on the hybrid composite. It 37   was likely because of the polyester itself which its degradation starts at temperatures above 200°C.   2.3.3 Glass fiber-wood/polyester Combination of wood and glass fiber in polyester matrix was first used in marine applications such as boats and kayaks. In boat manufacturing polyester resin applies as a thin layer on the wood surface to protect it against water absorption or it is used as a glue to bond wood and glass fibers. Despite the lower price of polyester compared to epoxy, using the combination of wood and polyester resin is limited in boat manufacturing due to the wood structure. Insufficient bonding to the wood causes glass fiber layer to fall off and delaminate.  If one uses an untreated wood a thick layer of polyester is required to cover the wood surface. Otherwise the inherent porous structure of wood absorbs most of the resin. The problem of having this thick layer is it takes a longer time to cure. The wettability of wood to polyester is a function of species. Therefore choosing a proper wood species should be considered when using polyester as matrix.    Incorporation of wood in the form of wood flour into glass fiber/polyester composite was done to investigate its role as filler in hybrid composite. Cerbu et al. (2010) studied the flexural behavior of a combination of wood flour and glass fiber in polyester matrix after immersion in two different environments, water and seawater. Hybrid laminates of oak-glass fiber/polyester and fir-glass fiber/polyester were fabricated using hand lay-up method. Each laminate contained six layers of E-glass woven fabric and the same weight ratio of the wood flour. A coupling agent in the form of powder was also used in order to improve the interaction between wood flour and polyester matrix. The aim of this study was comparing two different wood species, Douglas fir 38   and oak, to be introduced as fillers in polyester composites. Less moisture absorption in seawater immersed fir composites was reported compared with water immersed oak composites. Authors believed the greater resin content of fir wood flour and the presence of salt in seawater acted as barriers against the water absorption. Despite this lower water absorption, both flexural stiffness and strength decreased in fir composites after exposure to the wet environment. On the other hand, oak composites showed improved flexural stiffness indicated the specimen became more rigid after immersion in seawater while their flexural strength decreased insignificantly. The only concern of oak laminates was their tendency to develop dark stains over the surface due to their high tannin content. Although authors investigated the effect of humid environment on mechanical properties of wood flour-glass/polyetsre composite but it was necessary to provide a detailed information on lamination configuration and lay-up sequence as well as comparison of the results with the control glass fiber/polyester composites.  2.4 Wood/polyester composites Similar to the plant fiber/polyester composites various chemical treatments were used to modify the surface of wood furnish to improve interfacial adhesion. Different configurations of wood furnish including wood flour, wood fiber, veneer strips, and sawdust were used as filler to reinforce polyester. The presence of wood furnish reduces matrix shrinkage and prevent cracking of the brittle matrix during the curing process (Stamboulis et al. 2000). Wood, however, is a hygroscopic material and has tendency to absorb water in humid environment.   39   2.4.1  Chemical treatments effects Zhang et al. (2013) enhanced the interfacial adhesion between wood fiber and an unsaturated polyester matrix using chemical ingredients. Acrylic acid (AA), acrylic acid/methyl methacrylate (AAM), and acrylic acid/salinization (AAS) treatments were used to modify the fiber surface. Composites were made by mixing up the modified wood fiber, at various weight volume fraction up to 16%, and the polyester matrix. The effects of three chemical treatments on the mechanical properties, water absorption, and microstructure of the wood fiber/polyester composites were investigated. Results showed that all the three treatments improved the compatibility of wood and polyester which led to increase in the flexural properties and tensile strength. The best interfacial adhesion and optimum physical and mechanical properties was reported for AAS treated composites. Authors believed the enhanced interaction between fiber and matrix was the reason for the higher impact strength in AAS treated composites. This finding is in contrast with the findings of the study by Sèbe et al. (2000), who concluded that the slippage at interface was advantageous for energy absorption. In similar study by Marcovich et al. (1998a) who chemically modified wood flour with alkali and then maleic anhydride treatments to investigate their effect on dynamic mechanical behavior, compression, and flexural properties of the polyester composites. Alkali modification of wood flour increased the flour specific area and produced a rough surface. However, no improvement in the mechanical behavior of wood flour/polyester composites was found when particles were only alkali treated. Maleic anhydride molecules, on the other hand, were firmly attached to the wood flour and this resulted in improved water absorption behavior. Although the maleic anhydride which was the modifier in this research was well bonded to the wood particles, no improvement of final mechanical properties was reported compared to the untreated wood flour composite. This might be because 40   of the low molecular weight of the modifier which created a brittle interface. Authors reported the same results for the treated sawdust/polyester and wood fiber/polyester composites where a low interfacial area was thought to be responsible for the insignificant enhancement of the mechanical properties (Marcovich et al. 1996; Marcovich et al. 1998b).    Working on wood/polyester compatibility was continued by Nouwezem et al. (1994) where they treated pine wood meal with ozonation and esterification. Styrene maleic anhydride copolymer (SMA) was also added to the esterified pine wood in order to improve the hydrophobic nature of the composites. Results indicated ozonation and esterification of wood, together with the introduction of the SMA additive during the composite fabrication reduced water uptake in the composites. In a similar study zinc oxide and an excess of polystyrene have been tried in order to overcome the excessively slow mixing of wood fibers into the polyester dough molding compound (DMC) pastes (Freischmidt et al. 1991). Both of the used chemicals thicken the paste to maintain a cohesive dough and shortened the time needed for mixing. The tensile and flexural performance of the composites did not, however, show any significant improvement.   2.4.2 Densification and configuration effects Haghdan et al. (2015a) investigated the energy absorption and fracture morphology of wood veneer/polyester composites as a function of wood densification and configuration (woven, cross, and unwoven (unidirectional)). Glass fiber/polyester composites were also made for the purpose of comparison. This study characterized the dominant failure mode for each sample configuration. Findings of this study showed the difference of the dominant failure as splits, 41   longitudinal crack extension, zigzag pattern, and splinters in wood and starlike and circular cracks in glass fiber composites. Although veneer densification improved the impact performance of the composites without cell-wall fracture, this improvement was not statistically significant. Authors believed an improved densification method should be used to reach a higher degree of densification. In case of configuration effect, composites with woven and cross veneer strips had the best impact strength among all treatments as the number of wood strips was double of those in unidirectional composites. SEM micrographs showed that the polyester resin was well penetrated into the wood elements and distributed on its surface evenly. Chapter 3 of this dissertation discusses the results in detail.    In addition to all the described studies on the use of unsaturated polyester as matrix, the creep behavior of composites made with wood flour and a new polyester resin based on linseed oil were also investigated by Mosiewicki et al. (2011). The results of this study showed that the short-time creep deformation of composites was significantly improved with the incorporation of pine wood flour (up to 30%) into the polyester matrix. This is because of the reported excellent interfacial interaction between the ingredients as a result of good filler wetting and dispersion in the polyester matrix.  2.5  Summary The use of natural fibers in polyester composite has been a point of interest since 1980s. Several plant fibers including hemp, jute, banana, sisal, coir, kenaf, bamboo, rice, oil palm, and flax have been used to reinforce polyester matrix. These fibers, however, are highly susceptible to water absorption in a humid environment. Chemical modification of the surface of these fibers 42   improved the interfacial properties and as a result decreased water absorption of the composite. In addition to the investigations on adhesion quality, the effect of plant fiber length and content on mechanical properties of the polyester composites were also discussed. Accordingly, the use of 30 mm length plant fibers and weight fraction of 40% is recommended. Regarding the incorporation of glass fiber to natural fiber/polyester composites it generally enhanced the physical and mechanical properties of the polyester composite because of the inherently high stiffness glass fiber. Despite comprehensive research on plant fiber/polyester and hybrid fibers/polyester there is lack of research on compatibility of wood with polyester, composite lay-up configuration, curing process, and its fracture mechanisms. Chapter 4 of this dissertation includes the results of a detailed investigation on the effect of wood species on its wettability to polyester (Haghdan et al. 2015b).   43   Chapter 3: Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites  3.1 Synopsis Current energy absorbers in industrial applications are made of metals or fiber-reinforced polymers using glass and carbon fibers. These materials are stiff and strong but exhibit low energy absorption when subject to the impact load. Other issues in the use of these materials are their high cost (fiber reinforced polymers) and weight (metal). Wood reinforcements on the other hand are light-weight and economic but less stiff. In this study, thin composites of Douglas fir and a polyester matrix were fabricated using hand lay-up technique. The objectives of this chapter were (i) to investigate the impact resistance and fracture patterns of wood/polyester composites, (ii) examine polyester resin distribution and penetration into the wood material, and (3) inspect the effects of wood veneer configuration and densification on mechanical properties of polyester composites. Densified and un-densified Douglas fir veneers were used to create three different mat configurations: woven, cross and, unwoven (unidirectional) mats. A total of 350 specimens were tested under drop-weight impact load following ASTM D5420 and their impact resistance was calculated using the staircase method. Lab-made glass fiber/polyester composites were manufactured with the same technique for the purpose of comparison. Results of this study demonstrated that the effect of reinforcement configuration on the final performance of the wood-polyester composites was significant with the woven and cross configurations having remarkably higher impact energy than unidirectional composites. Densification of the wood veneer insignificantly improved the composite performance.  44   3.2 Introduction The use of composite materials, specifically fiber-reinforced-polymers, as alternatives to traditional structural materials in construction and transportation industries has dramatically increased in the past decades. Composite materials combine the advantage of light weight with the desired physical and mechanical properties of their constituents. Current industrial fiber reinforced polymer composites are made of synthetic fiber such as glass fiber and carbon fiber which provide high stiffness and strength for the composites under tensile loads. The energy absorbers currently used in cars and aircrafts are made of metal or synthetic fiber reinforced polymers. These energy absorbers provide high stiffness but are either heavy (metal) or expensive (fiber reinforced polymer). Wood fibers or veneers, in contrast, have a high stiffness and strength to weight ratio and exhibit a higher deflection before breaking (Felix and Gatenholm 1991; Chtourou et al. 1992; Li and Li 2001); if used as reinforcement, these materials may extend the time of final failure for the composite and increase its total energy absorption.   The basic physical and mechanical properties of various formulations of wood flour and thermoplastic polymers have been experimentally studied (Adhikary et al. 2008; Tajvidi et al. 2008; Tajvidi and Haghdan 2009; Xiong et al. 2009; Tasdemir et al. 2009; Ashori and Nourbakhsh 2010; Nourbakhsh et al. 2010; Bhaskar et al. 2011). In all these studies, the main goal was limited to improving the tensile and bending properties of the composite and enhancing its resistance against water absorption. In further studies (Kretsis 1987; Jiang and Kamdem 2004; Tungjitpornkull et al. 2007; Jarukumjorn and Suppakarn 2009; Pothan et al. 2010; Valente et al. 2011), a combination of wood flour with glass fiber in a polymer matrix improved tensile 45   modulus, bending strength and panels dimensional stability. The performance of these composites under static and dynamic impact loads, however, has not been investigated.    Since increasing wood density improves its mechanical properties, much research has focused on developing densification methods (Seborg et al. 1945; Navi and Girardet 2000; Blomberg and Persson 2004; Kutnar et al. 2007; Kutnar et al. 2008; Kutnar et al. 2009; Standfest et al. 2013; Macias et al. 2011). Although these compression techniques had positive effect on the general performance of solid wood, no information is available on their effect on the mechanical properties of wood reinforced polymer composites.    The objective of this research was to gain a better understanding of fracture behavior of wood veneer reinforced thermosetting composites under impact load. Three independent factors: type of reinforcement, configuration and densification, were studied to characterize their effects on impact resistance and failure modes of polyester composites.   3.3 Experimental investigation 3.3.1  Materials Douglas fir (Pseudotsuga menziesii) veneer sheets, with a nominal thickness of 0.6 mm, were cut into thin strips of 2 mm, 5 mm and 12 mm width using a veneer clipper. These strips were used to make unidirectional, woven and cross veneer mats (shown in Figures 3.1a and b) to investigate the effects of veneer configuration on the impact properties of polyester composites. To consider the effect of densification, Douglas fir veneer sheets with an initial moisture content of 7% were densified in a hot press at a temperature of 150oC and a pressure of 1.38 MPa for 9 46   minutes. Before densification, 2 gr of water was sprayed onto the each veneer sheet. The veneer sheets were then piled in stacks of 10 with aluminum foil separating the layers to prevent the veneers from sticking to each other. Following this densification, the veneer thickness decreased from 0.60 mm to 0.41 mm. Comparing the veneers initial density (0.451 g/cm3) and density after compression (0.658 g/cm3), the degree of densification was 46%. The veneer sheet darkened after densification. In addition, edges of densified veneers deformed three hours after densification as they dried out and split, shown in Figure 3.1c. Previous research found that the internal stresses in the semi-crystalline parts of microfibrils are the major cause of this deformation. These internal stresses can be relaxed by post-treatment at high temperature in saturated steam to avoid or limit veneer deformation (Rautkari et al. 2010; Navi and Heger 2004). In the study presented herein, curved edges of densified veneers did not impede the process of making woven composites. These curved edges did, however, lead to a longer mat formation time for the unidirectional samples.   Of the different strip widths, woven mats of 12 mm strips performed best in the weaving process. During manufacturing of the plain woven mat, there was a limit to how tightly one can pack the mat. Due to the stiffness of the strips, if the edges of the warp strips were in contact with each other, it was not possible for the weft strips to also be in contact. When making a transversely isotropic mat, one must have the same number of strips with the same spacing between adjacent in both directions. Weaving the 2 mm wide strips resulted in lots of splits and breakage with large gaps between them, as shown in Figure 3.1d. Because of these shortcomings, the use of 2 mm wide strips was abandoned. The process of weaving was easier using densified 5 mm strips compared to the undensified ones due to their smaller thickness. In fact, the 47   densification process made them more flexible in weaving likely the result of their lower stiffness as a result of their smaller thickness. However, there were still gaps between individual strips. This issue was solved by using wider strips. Figure 3.2 shows woven mats of densified and undensified veneer stripes of different widths.     In tight mats, gaps are smaller and the wood content per unit area is higher. Since wood has high potential to absorb impact energy, it was postulated that having a higher mass fraction of wood in polyester composite may contribute to improve the performance. The other materials used to fabricate the wood-thermosetting composites were woven E-Glass cloth (170 g per 127cm), waxed polyester resin, Methyl Ethyl Ketone Peroxide (M.E.K.P) and releasing wax which were all purchased from Coast Fiber-Tek Products Ltd.   3.3.2 Composite manufacturing In order to study the effects of wood reinforcement configuration and densification on the impact property of wood and glass fiber reinforced polyester, the following parameters were examined: densification (densified and undensified veneer), configuration (woven, cross and unwoven (unidirectional)) and type of reinforcement (wood and glass fiber). A summary of all treatments and the used labels is given in Table 3.1.Unidirectional, woven and cross wood polyester samples were made using a hand lay-up method. Before starting the lay-up process, the matrix material was prepared from the general purpose polyester resin and the catalyst MEKP in a weight ratio of 100:2 at a temperature of 15˚C. For unidirectional samples, 12 mm cut strips were resinated on both sides using a paint brush and put in the mold one after another to fully cover the mold area. A squeegee was then used to eliminate air bubbles and distribute the resin evenly 48   over the surface. It proved challenging to keep these slippery strips close to each other, especially for the densified strips as their surface was not perfectly flat after densification and consequently the mat was not flat when viewed from the side. Nevertheless, it was possible to produce flat unidirectional densified samples.    To produce the woven composites, first the release wax was applied on the mold. Then the mixture of polyester and catalyst was applied on both surfaces of the woven mat and left to cure at room temperature for about 3 hours. The process for cross wood polyester composites was slightly different; in this case individual veneer strips were laid into the mold unidirectionally after being fully resinated with polyester catalyst mixture. After removing the excess amount of resin using a squeegee, a second layer of wood veneer strips was laid onto the first one in a 90° orientation. After finishing the second layer, the mat was left to cure at room temperature.    Glass fiber reinforced polyester composites were made by first cutting glass fiber cloth into the appropriate mold size. The rest of the manufacturing process was identical to the one for the wood composite samples. In the hybrid composites which contained both wood and glass fiber as reinforcements, wood strips were sandwiched between two face layers of glass fiber and polyester resin applied during the lay-up process. Hybrid composites with undensified strips led to a smoother surface compared to those with densified reinforcement.    The material composition varied for the different treatments. To examine the effect of different configuration on the impact property cross wood composites and woven wood 49   composites were made to the same mass fraction. The average mass fraction of these treatments was 11% higher than the unidirectional wood samples. The average mass and volume fractions of wood, glass fiber and polyester resin are given in Table 3.1.   3.3.3 Test methods To determine differences in impact failure energy of various glass fiber, veneer strips and polyester resin combinations, a simple impact testing apparatus was used. The apparatus consisted of a 2 m PVC pipe, with holes drilled into the side at specific heights. The pre-drilled holes were set at 250 mm, 500 mm, 750 mm, and 1000 mm from the ground. Above 1000 mm, the holes were set closer to each other in increments of 50 mm. In order to get smaller height increments for the thin composite plates of this study, the black tube was reversed. Previous research found that thin samples remained fixed in position during the impact test when they properly clamped (Morinière et al. 2013). To avoid movement of thin composite plates under the impact load, the clamping assembly, precise in opening size and dimension (ASTM D5420, 2010), was used as shown in Figure 3.3. The plates were fully clamped by eight wood screws, equally spaced along the perimeter of the clamping frame to a torque of 20 Nm each.    The up-and-down method used to determine the mean impact resistance of reinforced composite samples follows ASTM D5420 2010. In this standard, a fixed weight falls through a guide tube from various heights and hit the sample at the centre of the plates. This technique is commonly used to determine the “mean-failure energy” or the “mean impact resistance” of composite materials; the measured energy is that which causes 50% of the samples to fail. 50    For this study, failure was defined as the presence of any crack or split created by the impact of a falling ball that was visible by the naked eye under normal laboratory lighting conditions. Failure can include complete shattering of the specimen, any crack radiating out towards the edges of the specimen on either surface, any radial crack within or just outside the impact area, presence of one or more holes, brittle splitting of the top and bottom surface of the specimen.    Before starting a test, samples were visually examined to ensure they are free of cracks and other obvious imperfections. After clamping the sample, a 224 g chrome steel ball was pushed through a hole and dropped onto the sample. If the specimen failed, the drop height was decreased. In this study, five pre-test samples of each treatment were used to determine the approximate failure height. Once the approximate failure height for a given sample configuration was determined, 20 replicates were performed.    The first sample of each treatment was tested at its initial failure height. If the sample did not fail, the drop height was increased one increment (herein is 50 mm) for the second sample. Each sample was tested only once. An example of failure and non-failure plot using the staircase method is presented in Appendix A. Mean-failure height was calculated from the test data obtained, using Equation 1:             where h is the mean- failure height (mm), ho the lowest height at which an event occurred (mm), dh the increment of height (mm), N the total number of failures or non-failures (whichever  ℎ = ℎ𝑜 + 𝑑ℎ(𝐴 𝑁⁄ ± 0.5) (1) 51   is smaller), and A is the sum up of the product of i times ni, where i is a counting index starts at ho and ni is the number of events at the applicable height of failure. In calculating h, the negative sign was used when the events were failures (N=Nx). The positive sign was used when the events were non-failures (N=No). The constant of 0.5 was removed from the equation if the number of failures and non-failures were equal (No=Nx). After calculating the mean failure height, the mean-failure energy was computed using Equation 2:    where MFE is the mean-failure energy (J), h the mean-failure height (mm), w the mass (kg), and f is the conversion factor to joules (f =9.80665 × 10-3 for h measured in mm and w in kg).  3.4 Results The mean-failure energy and mean-failure height of different composite samples were calculated using equations 1 and 2. After each test, samples were examined for any sign of failures. Failure patterns differed with treatments. Figure 3.3 shows the performance of a control sheet of veneer under the impact. Veneer sheets were very brittle and all samples fail at the lowest drop height (50 mm). Based on ASTM D5420, the presence of longitudinal cracks and visible zigzag splits were taken as signs of failure for this treatment.    Failure types for unidirectional and woven samples are compared in Figure 3.4. These pictures also show the fracture of a densified veneer where it split along a single longitudinal crack (Figure 3.4a). Later samples broke into two parts and this permitted the ball to fall through the jig; it broke immediately with no splits and transverse cracks.  𝑀𝐹𝐸 = ℎ𝑤𝑓 (2) 52    UOP and UDP treatments had the same type of failure as the control samples. However, good interfacial adhesion between strips led the cracks to extend within individual strips with no samples having a crack that crossed into an adjacent strip, i.e., all samples were similar to Figure 3.4b. As described earlier, these unidirectional wood samples contained one layer of wood reinforcement.    In contrast to the unidirectional samples, woven composites were able to absorb more energy by stopping the ball right on the surface (Figure 3.4c). A number of fracture patterns were observed on the bottom surface of these samples, Figure 3.4d, including sharp splinters, zigzag patterns and strip bending. However the type of failure was completely different for the glass fiber samples. As it is shown in Figure 3.4e, circular cracks were the main type of failure in the hybrid samples reinforced with both glass fiber and wood veneers. In these samples, sandwiched wood strips between face layers of glass fibers remained relatively intact after the impact test. In the GP treatment, the ball hit to the centre of the sample and the crack extended radially around the centre, creating a starlike failure pattern, see Figure 3.4f.    Figures 3.4g and h compare the fracture patterns of cross configurations of wood in the resin matrix. Figure 3.4g shows the top surface of CrossOP samples (90° orientation of the veneer strips). These samples absorbed the impact energy by fracture along the fiber direction and zigzag splitting which was similar to what observed in woven configuration of wood. Both 0° and 90° faces had fracture parallel to direction of the fibers. The specimen shown in Figure 3.4h belongs to the hybrid cross sample, G-CrossOP, which is a combination of both wood and glass fiber reinforcements. The fracture pattern for this sample was similar to other hybrid 53   treatments and limited to the face layer of glass fiber but the type of fracture was more brittle and left sharp edges on the surface.    The average failure energies of the treatments and their respective standard deviations are listed in Table 3.1 and are compared in Figure 3.5. The hybrid composites with woven configuration had the maximum mean impact energy of 1.58 J while unidirectional wood composite samples had the lowest energy value of 0.11 J.    One-way ANOVA was used for statistical analysis to determine if the mean difference among treatments was significant (Bulmer 1979). Figure 3.5 shows the effects of reinforcement configuration on mean impact energy of polyester composites. The mean impact energy of wood composites was significantly lower than the composites which had the contribution of both wood and glass fiber reinforcements. In fact, the failure energy in wood composites reached only approximately half the magnitude. However there was no significant difference of the reinforcement configuration between densified and undensified samples. It is also found that the difference between woven and cross configuration of wood and its effect on the mean impact energy of wood composites was statistically insignificant. For the unidirectional configurations, both densified and undensified samples had significantly lower impact values. This is the same for hybrid treatments where the woven and cross composites had significantly higher mean impact energies than unidirectional samples (Appendix B).  54    Figure 3.5 further shows that densified samples in both hybrid and wood treatments and in all configurations, woven, cross and unidirectional, had higher mean impact energy than the undensified samples. Wood veneer densification improved the impact performance of composites by 7%. This increase, however, was statistically not significant.   3.5 Discussion The mode of failure and failure patterns differed between the various composites. In wood-polyester samples the presence of splits, longitudinal cracks (wood is weaker in the transverse direction, so crack extends easier along the fibers), zigzag patterns and sharp splinters were the signs of failure. For glass fiber-polyester composites, starlike and circular cracks were the dominant modes of fracture. For hybrid cross composites, G-CrossOP, the brittle failure was seen with sharp edges on the top surface. In this study, all specimens were made as thin composite plates for easier identification of dominant fracture patterns in each treatment. Increasing the wood composite thickness enhances its stiffness and brings more fracture mechanisms into the composite. This consequently results in a better impact performance for wood composites.   As mentioned earlier in the results section, densification improved the impact properties of composite samples, but was statistically insignificant. In order to investigate the effects of densification on cell wall structure, samples were examined using optical microscopy. In Figure 3.6 deformed wood cells can be seen and interestingly, there are no fractures in the cell walls. Based on this evidence one can conclude that the combination of moisture content, pressure and time was sufficient to plastically deform the wood cells. Since samples were compressed 55   tangentially to the wood growth rings, the deformation occurred primarily in the denser latewood. It is also shown in Figure 3.6 that the lumen size in both earlywood and latewood decreased significantly. Considering that the effect of densification on impact properties of polyester composites was not significant, a better densification method may assist to get a higher degree of densification taking into consideration that the wood cell walls should not break.   As can be seen in Figure 3.5, there is no difference between mean impact-resistance of control samples, unidirectional densified veneer polymer (UDP) and unidirectional undensified veneer polymer (UOP) samples. A smaller increment in impact tower might give more detailed results, but it is unlikely to make significant changes in the mean-impact resistance of the treatments.    Noted the significant effect of configuration type, woven and cross wood composites had higher mean-impact energy compare to the unidirectional samples. The average mass fraction of woven and cross wood composite was 11% higher than the unidirectional one. In fact, twice the number of strips was used to create a woven style while for unidirectional samples, half the number of strips fully covered the mold area. This higher mass fraction of wood increased the composites thickness and this contributed to its better performance under the impact load. It is also found that there was no statistical significance between woven and cross configuration. Although the woven composites were slightly stronger under impact load than the cross samples, the weaving process of its reinforcement is time consuming. Since the difference between these two treatments is not significant, the cross configuration is recommended.  56    Scanning Electron Microscopy (SEM) examination was used to study the distribution and penetration of polyester resin into the wood and glass fiber reinforcements. Figure 3.7 presents the results of SEM examination on wood and glass fiber composites demonstrating that the polyester resin completely surrounded the reinforcements and has neat distribution in the composite. Figure 3.7a shows tracheid pits in the veneer sheet before resination. After resin application, the polyester matrix fully covered the tracheid pits and flowed onto them as shown in Figure 3.7b and c. A SEM micrograph of glass fiber composite where the individual glass fibers were also entirely surrounded by the polyester is presented as Figure 3.7d.   3.6 Conclusions The impact resistance of various polyester composites reinforced with wood and glass fibers was investigated using a drop-weight impact test. Based on the results the following conclusions were reached:  1. The fracture patterns varied in wood and glass fiber reinforced polyester composites; the dominant failure modes in wood composites were splits, longitudinal crack extension, zigzag pattern and splinters while starlike and circular cracks appeared in glass fiber composites. 2. Impact resistance was higher for glass fiber composites compared to wooden samples. 3. Woven and cross polyester composites had significantly higher impact energy than unidirectional composites.  4. The densification method used in this study improved the impact properties of wood composites without cell wall fracture. However this improvement was not statistically significant. 57   5. SEM micrographs showed that the resin had penetrated into the wood elements and glass fibers and was evenly distributed over both.  The results of this study including fracture patterns of wood/polyester composites, mat configurations, and polyester resin distribution were used in chapter 4 to investigate the wettability and impact performance of wood/polyester composite laminates, Figure 3.8.    58   Table 3.1 List of treatments, including mass and volume fractions of components; n= 20 replicates for each impact energy mean and St.Dev. Note: W denotes wood, P polyester, and GF glass fiber.   Name of treatments Label Mass [%] Volume [%] Impact Energy [J] W P GF W P GF Mean St. Dev Control-veneer sheet Control 100 - - 100 - - 0.11 - * Glass-polymer GP - 56 44 - 74 26 1.14 0.097 Woven-densified-polymer WDP 64 36 - 81 19 - 0.62 0.097 Woven-undensified-polymer WOP 63 37 - 80 20 - 0.58 0.101 Cross-densified-polymer CrossDP 64 36 - 81 19 - 0.59 0.109 Cross-undensified-polymer CrossOP 63 37 - 80 20 - 0.56 0.118 Unidirectional-densified-polymer UDP 50 50 - 70 30 - 0.11 - * Unidirectional-undensified-polymer UOP 53 47 - 72 28 - 0.11 - * Glass-woven-densified-polymer G-WDP 28 46 26 53 38 9 1.58 0.097 Glass-woven-undensified-polymer G-WOP 27 47 26 52 39 9 1.55 0.094 Glass-cross-densified-polymer G-CrossDP 28 46 26 53 38 9 1.53 0.094 Glass-cross-undensified-polymer G-CrossOP 27 47 26 52 39 9 1.50 0.101 Glass-unidirectional-densified-polymer G-UDP 24 48 28 48 41 11 1.44 0.094 Glass-unidirectional-undensified-polymer G-UOP 24 49 27 48 42 10 1.35 0.101 *Since all the samples broke at the lowest height of failure, 50 mm, the standard deviation value was zero for the treatment.  59   Figure 3.1 Douglas fir: (a) original sheet 0.6 mm thick; (b) woven sheet before trimming edges; (c) veneer edges deformation after densification; and (d) woven sheet of 2 mm veneer strips, circles show strips cracks and splits.                        60   Figure 3.2 Woven veneer sheet after edge trimming: (a) undensified 5 mm strips; (b) densified 5 mm strips; (c) undensified 12 mm strips; and (d) densified 12 mm strips.  61   Figure 3.3 Control samples: (a) ball resting on sample after impact, and (b) after removing ball.  62   Figure 3.4 Comparison of failure patterns: (a) densified veneer; (b) UOP; (c) WOP, top surface; (d) WOP, bottom surface; (e) circular crack in G-UDP; (f) starlike crack in GP; (g) CrossOP; and (h) G-CrossOP.                     63   Figure 3.5 Comparison of mean impact energy of treatments (error bars represent 95% confidence intervals).             00.20.40.60.811.21.41.61.8Mean Impact Enertgy (J)Treatments64   Figure 3.6 Light micrographs of veneer cross section before and after densification; densification happened in tangential direction of the growth rings (vertically in the micrograph). 65   Figure 3.7 SEM micrographs of resin distribution; (a) control veneer sheet – no resin applied; (b) wood-polyester composite; (c) tracheids pit in wood-polyester composite, (d) the individual glass fibers surrounded by polyester matrix in GP composites.              66   Figure 3.8 Relationship between Chapters 3 and 4  Chapter 4  Wettability and impact performance of wood veneer/polyester composites Chapter 3  Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites • Fracture patterns  • Mat configurations • Resin distribution  67   Chapter 4: Wettability and impact performance of wood veneer/polyester composites  4.1 Synopsis This chapter builds on Chapter 3 in which thin composite plates of wood and polyester resin were fabricated and tested under the impact load. It reports on an experiment on the roughness and wettability behavior of three different types of wood veneers to the polyester, and discusses the effects of lay-up configurations during lamination process, and investigates the impact performance of wood/polyester composite laminates. Sheets of Douglas fir, maple, and oak veneers were assembled as unidirectional, balanced, and unbalanced cross-ply laminates using a catalyzed polyester resin. These were compared to control specimens using glass fiber as reinforcement. The impact properties of the samples, with respect to the laminate thicknesses, were characterized using a drop-weight impact tester. The wettability and surface roughness of unsanded and sanded wood veneers were also investigated. Results showed that Douglas fir cross-ply laminates had an impact energy equivalent to glass fiber laminates, making them an interesting alternative to synthetic fiber composites. Wood/polyester laminates absorbed a considerable amount of energy through a higher number of fracture modes. The balanced lay-up limited twisting of the wood/polyester composites. The lowest contact angle and highest wettability were observed in unsanded Douglas fir veneers.  68   4.2 Introduction The use of natural fiber reinforcement in polymer composites has increased over the past few decades as an alternative to conventional structural materials including concrete and metals (Li et al. 2007). Natural fibers are available in abundance and are inexpensive compared with synthetic fibers (Ashori 2008). In addition, they often have lower density than synthetic fibers, which facilitates their use in automotive and construction applications where relatively low material weight is a major advantage (Felix and Gatenholm 1991; Chtourou et al, 1992; Li and Li 2001).    The quality of the adhesion between wood and thermoplastic or thermoset polymers significantly affects the performance of such composites. Insufficient adhesion between wood and the polymer matrix results in low tensile strength and high moisture absorption (Bledzki et al. 1998). The adhesion itself depends on the wood and resin chemical structures and the manufacturing method. If one ingredient is polar and the other is non-polar (as is the case in most wood/thermoplastic composites), it is necessary to use compatibilizers, also known as a coupling agents (Sabzi et al. 2009; Gao et al. 2012; Ndiaye and Tidjani 2012; Sobczak et al. 2012). The use of a compatibilizer improves bonding but increases the cost of the final product. Unlike wood/thermoplastic composites, there is no need to add a coupling agent to wood/thermoset composites, as both wood and polyester are polar and typically bond well during the curing process (Stevens 1999).   Extensive research on the basic physical and mechanical properties of wood and thermoplastic composites is available (Adhikary et al. 2008; Tajvidi and Haghdan 2009; Xiong et al. 2009; Tasdemir et al. 2009; Ashori and Nourbakhsh 2010; Nourbakhsh et al. 2010; Bhaskar et 69   al. 2011).. The objective of most of these studies was to improve the compatibility between the composite ingredients and enhance their mechanical performances, such as tensile strength and bending stiffness. The mechanical properties of unsaturated polyester composites reinforced by various plant fibers such as jute, flax, sisal, hemp, and banana have also been investigated (Gowda et al. 1999; Pothan et al. 2003; Rodriguez et al. 2005; (Dhakal et al. 2007). These studies showed that a strong interphase between the reinforcement and matrix results in composites of high stiffness and strength but limits the number of energy absorption mechanisms. In addition to plant fibers, wood veneer strips have also been used as reinforcement in a polyester matrix to manufacture thin composite plates (Haghdan et al. 2015a). Results showed that the effects of Douglas fir veneer configuration (unidirectional, cross-ply, and woven) on the impact strength of polyester composites were significant and improved the impact strength of the cross-ply wood-polyester samples. However, they found that veneer densification, did not significantly change the impact behavior of composites.   Using a combination of two reinforcements in a polyester matrix was also previously investigated. In one example of this approach, also termed “hybrid composites,” sisal and glass fibers were combined in a polypropylene matrix, improving the tensile, flexural, and impact strengths of the composites and decreasing their water absorption (Jarukumjorn & Suppakarn 2009). In similar research on hybrid composites, the dynamic mechanical properties of samples made from a mixture of banana and glass fibers in a polyester matrix were improved compared to those of composites made with banana fiber alone (Pothan et al. 2010).  70    An important aspect when considering wood as a composite component is its wettability, which can be characterized by various methods such as the dynamic Wilhelmy method and the single-fiber Wilhelmy method (Gardner et al. 1991; Wålinder and Johansson 2001; Wålinder and Ström 2001). Wilhelmy principle-based methods measure advancing and receding contact angles after immersion of the sample into the liquid and report the force exerted upon the sample (Casilla et al. 1981). The drawback of these techniques is the difficulty of preparation since the sample must have identical front and back surfaces to achieve similar solid-liquid interactions. As an alternative, contact angle measurement using a goniometer has recently been used to determine the wettability of wood and wood-based composites. In this technique, the contact angle of an individual droplet or several droplets can be collected quickly and easily using a high-speed camera. In addition, there are no specific requirements of the cross section shape of the solid surface. Several studies have investigated the wettability of medium-density fiberboard, laminated veneer lumber, and polymer composites using the contact angle goniometer (Thwe and Liao 2002; Ayrilmis et al. 2009; Ayrilmis and Winandy 2009; Fávaro et al. 2010; Cappelletto et al. 2013).   Another parameter affecting interfacial adhesion is the veneer roughness, a measure of the fine irregularities on the veneer surface (Tabarsa et al. 2011). Several studies have reported that sanding wood veneers affects their surface wettability (Shupe et al. 2001; Aydin 2004; Kılıç et al. 2006). The surface roughness of wood and wood products also depends on structural features such as annual ring variation, wood density, cell structure, and latewood/earlywood ratio (Magoss E 2008; Kılıç et al. 2009).  71    The objectives of the present work were to evaluate the potential of wood veneer as an alternative raw material for manufacturing polyester composites; to investigate the effect of wood veneer lay-up configurations on the energy absorption of the composites compared with thin composite plates and lab-made glass fiber laminates; and to investigate the roughness and wettability behavior of wood veneers to the polyester resin.  4.3 Experimental investigation 4.3.1 Materials Improvement of the impact strength in thin composite plates of Douglas fir veneer and polyester (Haghdan et al. 2015a) was the motivation for this research. It was hypothesized that increasing the lay-up thickness would increase the number of fracture modes operating in these wood-based composites. If true, Douglas fir-reinforced polyester should have properties comparable to glass fiber-reinforced polyester. In addition to Douglas fir, two other species were also used to investigate the effects of wood species on its interaction with polyester.   Douglas fir (Pseudotsuga menziesii), sugar maple (Acer saccharum), and red oak (Quercus borealis) veneers of 0.6 mm nominal thickness were used to compare the effects of species anatomy on the manufacturing process and the resulting impact properties of the polyester composites. The texture of the veneers depends on the difference between the size of the tracheids and vessels in earlywood and latewood. Table 4.1 summarizes these properties. Douglas fir exhibits an abrupt transition from earlywood to latewood and has a medium texture. Diffuse-porous sugar maple has a fine texture as the vessels across the growth ring are similar in size. Ring-72   porous red oak has a coarse texture with larger-diameter vessels in its earlywood and smaller-diameter vessels in its latewood.   An unwaxed orthophthalic general purpose polyester resin was used as the glue in this study. This resin is characterized by average mechanical properties and lower cost than epoxy (Husseinsyah & Mostapha 2011). Unsaturated polyester solutions in styrene were used and the cross-linking of the resin and styrene during the composite manufacturing process created a three-dimensionally shaped resin (Goodman 1998). An organic peroxide, methyl ethyl ketone peroxide (MEKP), was used as a catalyst to accelerate the curing process at room temperature. The resin, the catalyst, and a releasing wax (to facilitate composite demolding) were obtained from Coast Fiber-Tek Products, Burnaby, BC, Canada. A synthetic reinforcement, woven E-Glass fiber cloth (nominally 200 g/m2) purchased from Coast Fiber-Tek Products was used to make glass fiber composites for comparison.  4.3.2 Specimen description  Three parameters were varied between treatments: 1) the type of wood species (Douglas fir, softwood and maple and oak, hardwood); 2) the veneer configuration (balanced cross-ply, unbalanced cross-ply, and unidirectional); and 3) the type of reinforcement (wood and glass fiber). A total of 13 different treatments were applied to the composites, as shown in Table 4.2. Individual sheets of Douglas fir, maple, and oak veneer were defined as the control specimens and labeled DF Control, Maple Control, and Oak Control, respectively. Unidirectional wood polyester composites and balanced and unbalanced cross-ply polyester laminates were labeled UP, BCP, and CP, respectively. The label GP was devoted to the lab-made glass fiber/polyester composites. 73    The wood-based composites were 200-mm-long, 190-mm-wide, and 2.5-mm-thick. Glass fiber/polyester composites had the same length and width but a thickness of 1.2 mm. For the impact tests, 10 replicates per treatment were produced and tested. Two different tests were applied to the veneers themselves to investigate their wettability behavior and surface roughness. To investigate the wettability of the veneers, they were sanded with two different grit sizes: 120 and 320. The treatments are listed in Table 4.3. For example, the label DF 120 denotes Douglas fir veneer sanded with grit size 120. Sanded and unsanded veneers were prepared in 100-by-100 mm squares and five replicates of each treatment were tested to determine the average contact angle. For the roughness test, 15 veneer samples 50-by-50 mm size were used to evaluate the surface roughness of each test group.  4.3.3 Composite manufacturing All samples (unidirectional, balanced, and unbalanced) were made as four-layer laminates using a simple hand lay-up technique. Before starting the lay-up process, the matrix material was prepared by mixing the general purpose polyester resin and the catalyst MEKP at a weight ratio of 100:2 under controlled laboratory conditions at a temperature of 20 °C and relative humidity of approximately 65%. Table 4.2 shows the average mass fractions of the wood and polyester in all composites. The resin content varied among treatments as each veneer absorbed different amounts of resin.   After measuring the mass of each individual veneer sheet, the lay-up process began. Using a paintbrush, the prepared mixture of resin and catalyst was applied to the veneer. A squeegee was used individually for each lamina to eliminate air bubbles and distribute the resin evenly over the 74   surface. The applied pressure was not measured but was rather low, with the resin being pushed over the surface with a hand held squeegee. The approximate thickness of resin/catalyst on each side of veneers was approximately 0.17 mm. A schematic of the balanced and unbalanced 4-ply lay-ups is shown in Figure 4.1. The first lamina was a resin-coated veneer sheet. The second lamina was added on the top of the first layer with the grain running in the opposite direction. The orientation of the third layer depended on the type of composite: in balanced cross-ply composites, the direction of the third layer was the same as the second layer, in unbalanced it was opposite. The last layer was always glued with opposite orientation to the third layer.   A balanced alternating lay-up is universally used in the production of plywood (wood composite panels made of wood veneer sheets and phenol-formaldehyde resin) and limits laminate warping and dimensional change (Dietz 1949; U.S. Forest Service 1999). In this study, however, both balanced and unbalanced configurations were manufactured to compare the effects of the lay-ups on the impact properties of the polyester composites. Samples were compressed using a flat metal block that generated a pressure of 500 Pa. The applied pressure assisted in resin distribution and made the final laminates flat. The wood/polyester mat was set aside for curing for about 24 h. The samples were then demolded and edge-trimmed.   The glass fiber-reinforced polyester composite was made by first cutting the woven glass fiber cloth into the appropriate mold size. The mold was covered with a thin layer of release agent, and then filled with the reinforcement. The glass fiber reinforcement was woven E-Glass fiber cloth. Therefore, there was no 0/90 configuration for composites made of this fiber. A squeegee was used to spread the resin over the glass fiber sheet. The rest of the manufacturing process was 75   similar to that of the wood composite samples. The glass fiber/polyester composite contained only glass fiber and polyester resin as described in Table 4.2. The goal of producing these laminates was to compare the results of impact testing with the wood/polyester samples. After making the wood composite samples, the average amount of polyester resin in each lamina was calculated. Douglas fir veneers contained 4.75 g of resin and oak veneers contained 3.28 g of resin in each lamina and had the highest and lowest resin absorption, respectively. Each maple veneer absorbed an average of 4.02 g of resin during the lamination process.  4.3.4 Test methods 4.3.4.1 Impact test The energy absorbed by the various polyester composites samples was measured according to ASTM (ASTM D5420, 2010).To limit sample movement during impact, a clamping assembly with precise dimensions was made of medium-density fiberboard (MDF). Samples were then placed between two pieces of MDFs of 254 mm by 241 mm with a 93 mm by 80 mm opening. These pieces were clamped by eight wood screws, equally spaced along the perimeter of the frame with a torque of 20 Nm each (Haghdan et al. 2015a).. Impact tests were performed by pushing a 224-g chrome steel ball through holes in a guide tube at various heights and the damage to the samples was noted. The presence of any crack or split created by the impact of a falling mass visible by the naked eye under normal laboratory lighting conditions was defined as failure. Specimen failure can also include complete shattering, radial cracking within or outside the impact area, brittle splitting of the bottom surface, and glass-type breakage. The potential failure energy was then computed using Equation 3:  76      where E is the impact energy [J], h is failure height as applicable [mm], w is the mass [kg], and f is factor for conversion to joules (9.81×10-3 when h measured in millimeters and w measured in kilograms).  4.3.4.2 Contact angle measurement The wetting behavior of the wood veneer sheets of different species was characterized by contact angle measurement. Two grit sizes (120 and 320) were used to determine the effect of sanding on veneer wettability. Angle measurements were conducted with a Contact Angle Goniometer connected to a high speed camera capable of generating 25 images per second. A drop of polyester was applied to the wood surface using a pipette. The image of the liquid drop was captured immediately after the polyester droplet was deposited on the wood surface and every 0.25 second for 15 seconds.   The images were then processed with the commercially available software ImageJ (Stalder et al. 2006; 2010). Using the sessile drop method, the contact angle of the polyester drop was determined by aligning a tangent with the sessile drop profile at the point of contact with the wood surface. Each contact angle value was taken as the average of five different measurements on different parts of the veneer surface. The results were evaluated using ANOVA (Bulmer 1979) to determine whether the differences between treatments were statistically significant.  𝐸 = ℎ𝑤𝑓 (3) 77   4.3.4.3 Surface roughness measurement A laser confocal microscope was used to measure surface roughness of the veneers before and after sanding. Before surface roughness measurements, all veneer samples were conditioned at 20°C and 65% relative humidity until constant mass was achieved. After conditioning, the average surface roughness (Ra) of the samples was determined according to ISO (ISO-4288 1996). Fifteen veneer samples 50-by-50 mm in size were used for each test group to evaluate surface roughness. Results were statistically analyzed using ANOVA.   4.4 Results and Discussion 4.4.1 Laminating process Figure 4.2 shows the unidirectional and cross-ply wood veneer/polyester and woven glass fiber/polyester composites after curing and edge trimming. All composite laminates were observed under a light microscope to inspect the cross sections and determine the presence or absence of polyester resin among the layers. Laminates of Douglas fir veneer and polyester matrix were laid flat without edge-twisting and a well-bonded composite was achieved. The manufacture of maple/polyester composites was possible using the hand lay-up technique, but complete adhesion between the four layers was not accomplished. It was expected that balanced lay-up with the grain directions of adjacent veneers perpendicular to each other would reduce shrinkage and warping in the laminates. Despite this reduced edge-twisting in the oak/polyester laminates, viable flat samples could not be manufactured.   According to ASTM D5420 (2010) for the impact testing of polymer composites, all samples should be flat with no imperfections. Hence, oak and maple laminates were not impact 78   tested. The Douglas fir laminates were visually examined to ensure they were free of cracks and other obvious imperfections before impact testing.   The medium-textured anatomy of Douglas fir contributed to better resin distribution by evenly absorbing it, resulting in flat laminates of this species compared to those of maple and oak. In contrast, there were difficulties in manufacturing oak laminates even when using a balanced lay-up. Resin was easily absorbed in oak veneers during the lamination process, but since oak has large diameter-pores, the resin flowed out of the bottom face and only a small amount of resin was retained. Tangential cut maple and oak veneers and quarter cut Douglas fir veneer were used in this study. Ring-porous oak veneers with the coarse texture twisted during the curing process as they shrank more in the tangential direction. Rays in radially cut veneers usually restrain dimensional change. Softwoods with gradual earlywood-latewood transition and diffuse porous hardwoods have lower shrinkage rates species with abrupt earlywood-latewood transitions mainly because of their finer texture. The absolute shrinkage value, however, controls by cell wall thickness and density.   Veneer shrinkage was also noticed in tangential maple veneer laminates, but was controlled using the balanced lay-up. In addition to the effects of veneer texture on the lamination process, the heat generated from the exothermic curing reaction in polyester composites may have also contributed veneer deformation (Vargas et al. 2012)   79   4.4.2 Microscopic observation To determine the color of the polyester resin under a light microscope, a block of polyester was made with one wood veneer submerged inside the resin block. After curing, the block was cut, sanded, and the surface was blown off using compressed air for observation under the microscope. As shown in Figure 4.3, the polyester resin, which appears as white color, was absorbed by both earlywood and latewood tracheids, mostly however by the earlywoods. The black areas at the top and bottom of the veneer are corresponding to the polyester block material. Some air bubbles appeared at the interface.   Light micrographs of wood/polyester composite laminates are presented in Figure 4.4. The polyester matrix is visible as a white strip between the layers of Douglas fir unidirectional laminates, as shown in Figure 4.4a. The four Douglas fir veneer sheets were laid flat without any twisting. All layers were wetted by the polyester, and there were no gaps between them. The good adhesion of the Douglas fir composite is a contributing factor leading to flat laminates. This good adhesion was also seen in the cross-ply samples where all cross sections of longitudinal tracheids were filled with the polyester resin, as shown in Figure 4.4b. In contrast, unidirectional maple/polyester composites had voids and areas lacking resin. Moreover, the polyester strips between other lamina were not as flat as those found in the Douglas fir samples, as shown in Figure 4.4c. Cross-ply maple samples exhibited a neat flat laminate, as shown in Figure 4.4d, but this was not observed consistently in all samples subjected to this treatment.   Figure 4.4e clearly shows areas lacking resin in oak/polyester unidirectional composites. For cross-ply oak samples, a considerable amount of voids and empty wood vessels were observed, 80   as shown in Figure 4.4f. As explained earlier, the polyester resin was observed as a white strip between the veneer sheets. This layer in the Douglas fir samples was uniform and had approximately constant thickness (Figure 4.4a), whereas the thickness of this layer in the maple composite was much less uniform (Figure 4.4c). To achieve consistent resin thickness in all laminates, an automated manufacturing method would be better than the hand lay-up technique used.   Figure 4.4f shows the presence of an intralaminar crack between the layers of the oak laminate. A very small band-saw was used to cut the boards and prepare them for the subsequent microscopic investigation. This preparation, however, resulted in longitudinal artifactual cracks in the oak composite, as poor adhesion among the layers made them more susceptible to cracking than the Douglas fir and maple laminates.   The void areas, areas without the presence of wood or polyester, appeared black under the light microscope, as shown in Figure 4.4f. There are also resin discoloration areas where the polyester exists but has been discolored (Figure 4.4c). The thermosetting resins have low thermal conductivity. The energy generated during the exothermic crosslinking reaction may increase the internal temperature of the thick laminates and result in discoloration (Sung and Hilton 1998). However, the laminates in this study were manufactured 2.5 mm thick and it is unlikely the heat generated during the curing process caused the discoloration. The specific constituents of the general-purpose polyester resin used in this study were unknown (confidential to the manufacturer) but for applications require extreme clarity and absence of color, a clear casting polyester resin, a water-white resin, can be used to avoid discoloration. As shown in Figure 4.4f, there were empty 81   vessel and vessels covered with wood dust. This was the result of the sanding performed before laminates observed with the microscope.  4.4.3 Fracture mechanisms Figure 4.5 shows the surface of Douglas fir laminate samples after the impact test. The type of failure was different for unidirectional and cross-ply composites. While failures of both unidirectional and cross-ply composites were brittle, there were more modes of fracture operating for the cross-ply samples, including longitudinal cracks and delamination of veneer sheets. Despite the greater energy absorption of cross-ply composites, they exhibited no sign of failure on their front surface.  Figure 4.5a shows the front face of the unidirectional laminate after being hit by the impact ball. The crack initially formed on the first layer and then broke through each successive layer. Figure 4.5b shows the back face of this laminate. Longitudinal crack propagation along the fibers was accompanied by delamination of the third and fourth layers.   In contrast, the cross-ply laminates absorbed more energy than the unidirectional samples and did not exhibit any macroscopic sign of failure on their front surface, indicating that the higher impact resistance imparted by this treatment postponed crack initiation, as shown in Figure 4.5c. The cross lay-up of the veneers stopped the crack propagation generated in the first layer of the laminate. This energy absorption mechanism continued until the last layer when fracture of the veneer on the back face was visible, see Figure. 4.5d. One would expect the unidirectional samples to be more brittle than the cross-ply samples as all fibers are aligned and there is no reinforcement in the transverse direction to resist splitting. The cross-ply samples, in contrast, are reinforced in this direction and as a result resist the splitting and exhibit an increase in impact energy. The 82   bonding between the veneer sheets of the Douglas fir samples was sufficiently high that there were no significant differences between the balanced and un-balanced configurations.  4.4.4 Impact energy The failure energy of Douglas fir/polyester and Glass fiber/polyester composite laminates was calculated using Equation. 3. The average failure energies of the treatments are shown in Figure 4.6. To compare the impact energy values of different species and reinforcement configurations, one-way ANOVA was performed on the results using an α level of 0.05, showing that there was a significant difference between the Douglas fir unidirectional/polyester, Douglas fir balanced and unbalanced cross-ply/polyester, and glass fiber/polyester composites. The effect of lamination orientation (unidirectional and cross-ply) on the impact properties of the wood veneer/polyester composites was significant, with cross-ply laminates absorbing more impact energy (1.36 J) than the unidirectional laminates (0.96 J) (Appendix C). Veneers in cross-ply laminates had more cracks in comparison with the unidirectional samples. The greater fracture of the veneers indicates the creation of more surface areas resulting in greater energy absorption.   A comparison between cross-ply Douglas fir/polyester and lab-made glass fiber/polyester composites showed that, despite the greater impact energy absorbed by the glass fiber composite, the difference between these two treatments was not statistically significant. The difference between balanced and unbalanced cross-ply composites was also statistically insignificant as shown in Appendix C.  83    The impact properties of the laminates were compared with the findings of a previous study on thin composite plates (Haghdan et al. 2015a). It was found that increasing laminate thickness significantly improved the impact energy absorption of wood composites. The mean impact energies of the four tested treatments of this study and the two composite plates of the previous study are presented in Figure 4.6.  4.4.5 Wettability of wood veneers The wettability of the veneer surface was characterized by measuring the contact angle of a drop of polyester resin on the wood surface (Figure 4.7).   The resin used in this study was unwaxed acetone-free polyester resin. Acetone evaporation can significantly change the volume of the liquid; the applied resin did not have acetone but styrene with negligible evaporation. The mass of polyester resin was measured after it was poured in the aluminum foil, immediately afterwards and every 30 seconds for 5 minutes. The weight of the resin did not change over this time. Considering the duration of contact angle test was only 15 seconds, it was concluded that no resin evaporated.   The average initial contact angle and that after 5 seconds for both sanded and unsanded wood veneers are presented in Table 4.3. The lowest average contact angle, 14°, was observed for the DF 120 and DF Control after 5 seconds. The highest value, 136°, was the initial angle measured in the Oak Control. Typical droplet shapes are shown in Figure 4.7. ANOVA showed that there was no significant difference between the left and right contact angles in all samples. However, 84   significant differences in the average contact angles (p < 0.05) between all groups were found (Appendix C).   Figure 4.8 illustrates that the measured contact angles for a drop of polyester on the sanded and unsanded wood veneers decreased with time. As shown, sanding the Douglas fir, maple, and oak veneer sheets with 120-grit improved their wettability by reducing their contact angles compared to the unsanded veneers. However, this improvement was not significant for Douglas fir veneers.   Sanding with 320-grit reduced the initial contact angle in the maple and oak controls, as shown in Table 4.3, but generally reduced the wettability of all veneers. The highest wettability to polyester among the species used in this study was exhibited by Douglas fir, followed by maple and oak. In the contact angle tests, the location of the polyester drop on the surface of the veneer was chosen randomly. This location could be either earlywood, latewood, or a combination thereof. Since the test was conducted in five randomly chosen spots, the results are assumed to represent the average contact angle.   As mentioned in the Materials section, the cells in Douglas fir and maple are smaller than those of oak (tracheids in softwood and pores in hardwood). When the polyester drops were placed onto the veneers, they were absorbed by the capillary action of these elements. This absorption is slower in oak wood as the adhesive force of large pores is smaller than that of the pores of Douglas fir and maple, consequently limiting the wettability of oak.  85    The 120-grit sandpaper cleaned the surface of the veneers from dust and contamination, creating a fresh, smooth surface by blowing off the contaminants using compressed air. The surface of the sanded Douglas fir veneer was much lighter in color indicating that the old surface had been removed. It is widely recognized that a fresh surface is created by sanding with 120 grit size with a subsequent increase in veneer wettability (Moura and Hernandez 2006; Jarusombuti and Ayrilmis 2011; Cappelletto et al., 2013; Wan et al. 2014). As shown in Figure 4.9, the improved wettability was significant for maple and oak veneers when compared to the unsanded veneers.    Using 320-grit sandpaper made the veneer surface very smooth. The dust generated from sanding filled in the open pores and reduced the absorption of polyester. It should be noted that the effect of the 320-grit sandpaper was not significant on oak wettability as oak has very large vessels and the dust generated from sanding could not fill the pores. If a sanded hardwood veneer is preferred, medium grit size improves the veneer wettability.   The edge boundaries of the polyester droplet moved wider as the droplet spread; an example of that is presented in Figure 4.7. The examination of the micrographs of Figure 4.4 reveals that the resin did indeed penetrate into the veneer over long time frames. There was likely absorption of the resin into the surface of the veneer during the 15 seconds of the measurement, but it is not known how quickly this occurred.    Grinding, brushing, and sanding do not cause chemical modification of the material’s surface; rather a clean surface results, and it has a characteristic structure corresponding to the 86   composition of the material (Habenicht 2009). To chemically modify the surface, one should consider physical and chemical pre-treatment methods. Previous research findings, however, are not conclusive. For instance, Sinn et al. (2004) investigated the chemical changes of the veneers after sanding with different grit sizes. Spruce and beech veneers became slightly more acidic after sanding with 400 grit size while veneers sanded with medium grit size (grit 100), were less acidic on the outer surface. Considering that the largest grit size used in our study was 320, it is assumed unlikely to have significant chemical changes on the veneer surface.  4.4.6 Surface roughness of wood veneers The average surface roughness of the veneer sheets sanded with each grit size is given in Table 4.3. The results showed a reduction of the surface roughness of the veneers as the grit size increased. The Douglas fir veneer sanded with 320-grit had a smoother surface than that sanded with 120-grit. According to ANOVA, the effect of grit size was significant for each species (Appendix C). In all tested veneers, using higher grit size sandpaper significantly reduced the roughness. Similar statistical results were found for the effects of the type of wood species: the roughness differences among Douglas fir, maple, and oak veneers were also statistically significant.    A decrease in contact angle (Ɵ) leads to an increase in veneer wettability. The reduced roughness of the veneers sanded with the 120-grit sandpaper, increased their wettability to the polyester. The lower contact angles of these veneers as compared to those of control samples indicate this improved wettability. One exception was in the case of Douglas fir, for which there was no significant difference between the wettability before and after sanding with 120-grit paper. 87     The 120-grit sandpaper reduced roughness in the maple and oak veneers and increased their wettability to polyester by reducing their contact angles, as shown in Table 4.3. Using a higher-grit sandpaper, however, reduced the surface roughness in all used veneers, as shown in Figure 4.10. This is in agreement with the results of a similar study on the effects of sanding (Sulaiman et al. 2009). Grit size affected surface roughness as sandpaper with higher grit size contains a finer abrasive (Demirkir et al. 2014). In fact, the finest sandpapers provided smooth surfaces having considerably less surface area to be wetted in comparison with the rougher surfaces. It is believed that using grit size higher than 120 significantly reduces the surface area available for wetting.    The very smooth surface created by 320 grit absorbed less resin than those sanded with 120-grit sandpaper. Figure 4.11 compares the surface roughness profiles of oak veneer (control, sanded with 120-grit, and sanded with 320-grit). Figure 4.11a shows the rough surface of the oak veneer. This roughness, however, decreased after sanding with 120-grit sandpaper, as shown in Figure 4.11b. The results of wettability testing showed that this level of smoothness reduced the contact angle. The smooth surface of the oak veneer after sanding with 320-grit sandpaper, as shown in Figure 4.11c, provide little resistance to resin flow over that surface.  88   4.5 Conclusions  This study investigated the impact properties of wood veneer reinforced polyester composites, their fracture mechanisms, and the wettability of wood veneers to the polyester matrix. Based on the results obtained, the main conclusion points can be summarized as follows:  1. Douglas fir cross-ply laminates had an impact energy equivalent to glass fiber laminates, making them an interesting alternative to the use of synthetic fibers as reinforcement. 2. Increasing wood/polyester laminate thickness resulted in a higher number of fracture mechanisms and greater impact energy absorbed by the composites.  3. The cross-ply veneer configuration in wood/polyester composites had significantly higher impact properties than the unidirectional ones. Using a balanced lay-up also limited twisting of the wood/polyester laminates. 4. The wettability of Douglas fir veneer was greater than that of oak and maple. Sanding with medium-grit sandpaper increased the wettability of the veneers.  The results of this study may help other researchers who work on developing new composites of wood and polyester for industrial applications. Findings of this chapter, including the results of wettability and impact tests, and mat configurations were used in chapter 5 to investigate dominant fracture mechanisms of wood/polyester composite laminates under quasi-static compression and shear loading, Figure 4.12.   89   Table 4.1 Comparison of the three wood species. Species Specific Gravity* Origin Veneer Color Pore/Tracheid Size and Distribution Douglas fir 0.48 Pacific north west Red Medium diameter tracheid  (35 to 45 μm); medium-textured Maple 0.63 USA and Canada Creamy white Diffuse porous; small diameter pores  (< 50 μm); fine-textured Oak 0.69 North east USA Reddish-brown Ring porous; large diameter pores  (100 to 200 μm); coarse-textured * Property measured at 8% moisture content (ASTM-D2395 2007)  90   Table4.2 List of treatments and mass fractions of components.  Treatments Label       Mass (%) W P GF Control-Douglas fir veneer DF Control 100 - - Control-Oak veneer Oak Control 100 - - Control-Maple veneer Maple Control 100 - - Douglas fir Unidirectional Polyester DF-UP 67 33 - Douglas fir Cross-ply Polyester DF-CP 66 34 - Douglas fir Balanced Cross-ply Polyester DF-BCP 63 37 - Oak Unidirectional Polyester Oak-UP 73 27 - Oak Cross-ply Polyester Oak-CP 73 27 - Oak Balanced Cross-Ply Polyester Oak-BCP 70 30 - Maple Unidirectional Polyester Maple-UP 78 22 - Maple Cross-ply Polyester Maple-CP 79 21 - Maple Balanced Cross-ply Polyester Maple-BCP 74 26 - Glass fiber Polyester GP - 47 53 Note: W denotes wood, and P polyester. 91   Table4.3 Average contact angle (initial and after 5 seconds) and surface roughness values of different wood veneers (n= 5 replicates for each angle measurement; n=15 for each roughness measurement).  Treatment Contact angle (°) Roughness (µm) Initial After 5 s Douglas fir Control 123 (0.86) 14 (0.60) 7.87 (0.46) Douglas fir 120 123 (0.84) 14 (0.79) 6.90 (0.75) Douglas fir 320 130 (1.02) 25 (1.10) 4.86 (0.50) Maple Control 126 (1.10) 24 (0.30) 6.14 (0.73) Maple 120 118 (1.78) 17 (0.80) 5.12 (0.52) Maple 320 124 (1.59) 28 (1.52) 4.23 (0.32) Oak Control 136 (1.86) 23 (0.65) 6.97 (0.51) Oak 120 113 (2.35) 16 (1.33) 4.28 (0.64) Oak 320 127 (2.60) 25 (1.30) 2.52 (0.23) Note: Values in parenthesis are standard deviations. 92   Figure 4.1 Lay-up of two cross-ply wood composite laminates made of four lamina; Balanced (left), and unbalanced (right); arrows indicate the fiber direction. Resin applied to the top face of the layers one to three is not shown here.            [0/90/90/0]  [90/0/90/0] 1 2 3 4 93   Figure 4.2 Composite laminates after curing: (a) glass fiber/polyester, (b) unidirectional wood/polyester, and (c) cross-ply wood/polyester.94   Figure 4.3 Light micrograph of the resinated veneer sheet in a polyester block. Polyester resin in the earlywood Polyester resin in the block Air bubble 95   Figure 4.4 Light microscopic images of unidirectional and cross-ply polyester laminates: (a) polyester as a white strip between layers; (b) filled cross sections of longitudinal tracheids with resin; (c) resin discoloration and area of lack of polyester; (d) neat adhesion among layers; (e) insufficient adhesion and voids; and (f) crack, void, and empty vessel elements. 96   Figure 4.5 Unidirectional and cross-ply Douglas fir laminates after impact testing: (a) unidirectional, front surface; (b) unidirectional, back surface, fracture in all layers; (c) cross-ply, front surface; and (d) cross-ply, back surface, controlled fracture in alternative layers.97   Figure 4.6 Comparison of the mean impact energy of wood/polyester and glass fiber/polyester composite laminates (error bars represent 95% confidence intervals). Note: Impact energy values of Thin GP and Thin WP were taken from the previous study (Haghdan et al. 2015a). 1.150.600.110.961.36 1.371.4300.20.40.60.811.21.41.6Mean Impact Energy (J)Treatments98   Figure 4.7 Highest initial angle of polyester droplet on Oak Control (top), and lowest angle of polyester droplet after 5 seconds, DF 120 (bottom).  135.5º                                                                                   135.9º 13.8º                                                                                                                 14.3º 99   Figure 4.8 Contact angle of polyester droplets on wood veneers (before and after sanding) vs. time: (a) Douglas fir veneer; (b) Maple veneer; and (c) Oak veneer.            a)               b)                                                     c)  0204060801001201400 1 2 3 4 5Contact angle (degrees)Time (seconds)Douglas fir ControlDouglas fir 120Douglas fir 3200204060801001201400 1 2 3 4 5Contact angle (degrees)Time (seconds)Maple ControlMaple 120Maple 3200204060801001201400 1 2 3 4 5Contact angle (degrees)Time (seconds)Oak ControlOak 120Oak 320100   Figure 4.9 Contact angles of polyester droplets on Oak veneer before and after sanding with grit size of 120 (angles at 0.5 second).               Unsanded Sanded 59.3⁰                                                                                                                       59.4⁰  41.9⁰                                                                                                                     42.3⁰ 101   Figure 4.10 Surface roughness of veneers as a function of sandpaper grit size; Error bars represent 95% confidence intervals.2345678Control 120 320Average srface roughness (µm)Grit size of sandpaperDouglas firMapleOak102   Figure 4.11 Roughness profiles of the Oak veneer before and after sanding (a) Control Oak, (b) sanded with grit 120, and (c) sanded with grit 320 grit sandpaper                 Measured length (1000 µm) a) b) c) 103   Figure 4.12 Relationship between Chapters 4 and 5       Chapter 5  Fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading Chapter 4  Wettability and impact performance of wood veneer/polyester composites  • Wettability and impact results • Mat configurations  104   Chapter 5: Fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading  5.1 Synopsis Natural fiber/polymer composites are being used increasingly as alternatives to traditional structural materials like concrete and metals and for inorganic fibers like carbon. While the fracture mechanisms during crushing of synthetic fiber/polymer composites have been thoroughly studied, limited information is available on post-fracture investigation and identification of the dominant fracture mechanisms of wood/polyester composites, and on their energy absorbing potentials. Given the results from Chapters 3 and 4 demonstrated that wood/polyester laminates have potential to absorb a considerable amount of energy by dissipating it through a number of fracture mechanisms when subjected to impact load. In this study laminates of Douglas fir veneer were fabricated using a catalyzed polyester resin and their potential as energy absorbers were investigated and discussed. The objectives of this chapter were to investigate the effects of (i) laminates symmetry (face layers of 0º or 90º), (ii) lay-up balance (balanced and unbalanced) and (iii) number of lamina (8, 11, and 12) on compressive properties of the polyester composites. Samples were tested under quasi-static Combined Loading Compression (CLC) and their compressive performances were compared to control specimens using glass fiber as reinforcement. Results indicated that the effect of symmetry on compressive properties of wood veneer/polyester laminates was significant with laminates with face layers of 90º and core layers of 0º had the highest deflection to failure. Increasing the wood/polyester laminate thickness enhanced their energy absorbing ability by bringing more 105   fracture mechanisms into play but it noticeably reduced the laminates compressive modulus. Despite the brittle failure of glass fiber composites wood laminates exhibited a number of fracture mechanisms with shear buckling as the dominant mode of failure in symmetric samples. This progressive failure with high energy absorbing ability make wood/polyester laminates a good candidate to be used as an energy absorber structure where high deflection to failure and longer failure time are required.  5.2 Introduction Composite materials are being used extensively as they offer unique combinations of material capabilities including lighter weight, lower cost, and higher stiffness and strength values. Properties of composite materials can be optimized using different reinforcement and matrix materials or by working on the lay-up configurations in order to provide desired physical and mechanical performances for the structure (Melo and Villena 2012). Both natural and synthetic fibers have been widely used with various polymer materials (Antich et al. 2006) (Dweib et al. 2004) (Aljibori 2011). However, there is an increasing use of natural fiber/polymer composites as alternatives to traditional structural materials like concrete and metals (Ashori 2008). These alternatives with lower cost and higher strength-to-weight ratio than their counterparts have potentials to be used extensively in automotive interior panels and as building materials (Li et al. 2007; Li and Li 2001; Felix and Gatenholm 1991; Chtourou et al. 1992). Synthetic fiber polymer composites have remarkably higher stiffness when compared with natural fiber composites. These stiff composites, however, exhibit low energy absorption in impact situations. This is a result of a trade-off between stiffness and energy absorbing ability where the stiff materials exhibit a brittle failure. For energy absorber applications the structure must provide high energy 106   absorbing ability with minimum compromise in its stiffness. In addition, having low cost and light weight energy absorbers is always a priority for transportation industries.    Wood is a hygro-visco elastic material and depending on moisture content and temperature can deform as the molecules slide over each other. It surrounds with an oligopolymer matrix. The wood structure can deform significantly before fracture occurs thereby postponing the final failure.    Generally energy absorbers are systems that convert kinetic energy into other forms of energy; for example an elastic strain energy in solids is a form of potential energy and plastic deformation in deformable solids consumes energy. Wood material exhibits plastic deformation under axial compression load through closure of the cellular voids and reduced the porosity resulting an increase in density. Densification may also increase the stiffness and strength of the structure if the cell walls are not damaged. This combination of stiffness and energy absorbing ability makes wood a good candidate for impact applications (Aljawi and Alghamdi 2000; Adalian and Morlier 2001).    Several researchers aimed on improving energy absorbing ability of synthetic fiber composites. A study by Aljibori et al. (2009) showed that the energy absorbing ability of glass fiber/epoxy laminates is significantly affected by the number of layers when they are subjected to the axial compressive load. Nagel and Thambiratnam (2004) investigated the effect of specimen geometry on the energy absorption capability of composite materials and found that lay-up configuration can significantly increase the area under the load-displacement curve. Lee and 107   Soutis (2007) investigated the effect of specimen size on the axial compressive strength of carbon fiber/epoxy laminates. The results showed that failure mechanism was mainly fiber microbuckling accompanied by matrix cracking and splitting. Fiber buckling and longitudinal splitting were other fracture mechanisms reported by Davies and Hamada (2007) for carbon and silicon carbide fibers reinforced epoxy composites. Despite the extensive research on fracture mechanisms of synthetic fiber/polymer composites there is little research in the literature concerning the compressive fracture mechanisms of wood/polyester composites. In this study the potentials of using wood veneer/polyester composites as energy absorbers have been investigated under crushing load. This work focuses on the use of wood veneer instead of synthetic fibers in order to make more environmentally friendly energy absorbing structures.   Several ASTM standard are available to measure the compressive properties of the laminates ASTM D6641 (2009),  ASTM D3410 (2013), and ASTM D5467 (2010) . The Combined Loading Compression (CLC) test method has become the most common technique to measure the compressive performances. The fixture is relatively small and inexpensive, easy to use and is capable of mixed-mode loading. CLC test method promotes the testing of cross-ply laminates. Hence, samples were subjected to the combined compression load and their compressive properties including maximum compressive strength, modulus, and total energy absorbing ability have been investigated and discussed.    The objectives of this research were: i) to investigate the effect of mat lay-up, and number of lamina on the compressive properties of wood veneer/polyester composites, ii) to 108   compare the fracture mechanisms of wood/polyester and glass fiber/polyester laminates, and iii) to identify the dominant mode of fracture in each treatment.   5.3 Experimental investigation 5.3.1 Materials The motivation for the current research was based on the earlier findings of Haghdan et al. (2015a and 2015b) on the impact performance of wood-polyester composites. Woven E-Glass fiber cloth (nominally 200 g/m2) purchased from Coast Fiber-Tek Products, was used as a synthetic reinforcement for making glass fiber-polyester composites for the purpose of comparison. Sheets of Douglas fir (Pseudotsuga menziesii) veneer, with a nominal thickness of 0.6 mm, were cut into squares of 200 by 200 mm using a veneer clipper. This veneer is proven to be easily wetted by polyester when compared with Maple and Oak veneers (Haghdan et al. 2015b) with the diameter of Douglas fir longitudinal tracheids varied from 35-45 µm (Erickson and Harrison 1974). For making laminate a waxed orthophthalic general purpose polyester resin was used as a glue. This resin is characterized by its average mechanical properties and lower cost than epoxy (Husseinsyah and Mostapha 2011). An organic peroxide, methyl ethyl ketone peroxide (MEKP), was used as catalyst for the curing process at room temperature. The resin, the catalyst, and a release agent were obtained from Coast Fiber-Tek Products, Burnaby, BC, Canada.    5.3.2 Specimen description In this study there were three factors (i) veneer symmetry, (ii) lay-up balance, and (iii) number of lamina. A laminate is symmetric when the ply orientation above the centreline is a mirror image 109   of that below it. Since wood is an orthotropic material it is likely that the performance of 0º and 90º layers under compressive load will be different. Hence, two types of symmetry were used: symmetry A where both face layers are laid-up in 0º direction, and symmetry B where they are laid-up in 90º direction. The former has 90º layers around the centre line and the latter has 0º layers as it is shown in Figure 5.1. Because symmetric lay-up is balanced it improves dimensional stability of the boards (Stark et al. 2010; Drerup et al. 2012). Asymmetric treatments with no mirror image about the mid-plane of the board were introduced to investigate the effect of asymmetry on the compressive properties of the boards. Asymmetric laminates will be unbalanced as their face layers are laid-up in different directions, one in 0º, and the other in 90º directions, Figure 5.1.   Since the samples were to be tested under compressive load it was necessary that samples be of sufficient thickness to avoid Euler buckling (ASTM D6641 2009).    For this work where the samples are compressed, buckling may occur due to deformations transverse to sample length. The critical load at which Euler buckling occurs is given by the following equation:    where Pcrit is the critical or maximum axial load on the column just before it begins to buckle (N), E the Young’s modulus of elasticity (Pa), I the least second moment of area for the column’s cross sectional area (m4), and L unsupported length of the column whose ends are pinned (m).  𝑃𝐶𝑟𝑖𝑡 =𝜋2𝐸𝐼𝐿2 (4) 110     Since composites of this study were going to be tested under a combination of compression and shear loading, a modified equation for shear deformation effects was used to calculate the minimum thickness of the specimens, Equation 5.  From the equation one can see that the sufficient thickness to avoid Euler column buckling depends on not only the gage length but also to the expected ultimate compressive strength, flexural modulus, and interlaminar shear modulus (ASTM D6641 2009; ASTM D3410 2013) (Equation 5).      where t is the specimen thickness (mm), lg the length of gage section (mm), Fcu the expected ultimate compressive strength (MPa), Ef the expected flexural or compressive modulus (MPa), and Gxz is the through-the-thickness (interlaminar) shear modulus (MPa). Plywood, (wood composite panels made of wood veneer sheets and phenol-formaldehyde resin), had the most similar lay-up and composition process to the laminates of this study. Hence, data from Canadian Plywood Association (CANPLY 2014) and Handbook of Finnish Plywood (UPM 2013) were used to estimate the minimum thickness. The minimum thickness of wood laminates to preclude buckling was calculated to be greater than 3 mm. Although cross-ply wood composites such as plywood are usually made as odd number of layers to prevent warping, in modern production techniques even numbers of plies are allowed when the middle two plies are going in the same direction and acting as one ply. In order to meet the thickness requirements wood composites of this study were fabricated as 8 plies. It is thought that increasing wood  (5) 111   laminate thickness may improve the energy absorbing ability of composites by bringing more fracture mechanism for the sample. Hence, 12-layers laminates were introduced to be compared with 8 layers composites. 11 plies composites of wood and polyester was made for simulation of plywood with odd number of stacked plies, Figure 5.1.    In addition to wood/polyester samples laminates of glass fiber and polyester resin were fabricated to compare the compressive performance and fracture mechanisms of these synthetic fiber composites with the wood/polyester samples. Synthetic fiber composites were made of four layers laminates to meet the minimum thickness criteria.   A total of 8 different treatments were applied to the composites. For comparison commercial Plywood, Medium Density Fiberboard (MDF), and solid wood samples were included. Table 5.1 shows all the treatments and their associated labels.   The wood-based composites were 140 mm long and 20 mm wide. The thickness of 8, 11, and 12 layers wood/polyester laminates were 6, 8, and 9 mm, respectively. Glass fiber/polyester composites had the same length and width but a thickness of 1.2 mm. Randomly selected commercial boards were cut to the same length and width but had nominal thickness of 7 mm. All the test samples were untabbed as more than 50% of the plies were oriented in 0º direction. For the CLC tests, 8 replicates were made per treatment and subsequently tested.  112   5.3.3 Composite manufacturing All composites including glass fiber/polyester were fabricated using a simple hand lay-up technique.  Waxed orthophthalic general purpose polyester resin was mixed with the catalyst MEKP in a weight ratio of 100:2 at room temperature of 20˚C, and a relative humidity of approximately 65%. Synthetic fiber composites were made, in average, from 60% mass fraction of glass fiber and 40% polyester.  In case of wood laminates, the wood and resin consumption were 67% and 33%, respectively.    Before lamination the mass of each veneer sheet was measured. The mold was then covered with a thin layer of wax for demolding. The prepared mixture of resin and catalyst was applied on the both surfaces of the veneer sheet using a paint brush. The excess amount of resin and possible air bubbles were then eliminated from it using a squeegee. This step led to an even distribution of the resin over the veneer surface. The stacking sequence was an alternate lay-up of 0 and 90º. Treatment codes for each sample type are given in Table 5.1. For example the code [0/90]2s  , the subscript “s” means “symmetric”, indicates an eight-lamina composite with the stacking order of 0, 90, 0, 90, 90, 0, 90, 0º. To manufacture this laminate after setting up the fourth layer in 90º direction the fifth layer will be set up in 90º direction as well to create the mirror image and make the laminate symmetric. Samples were compressed after lamination using a flat metal block that generated a pressure of 500 Pa. The applied pressure assisted in resin distribution and makes the final laminates flat. The mat was set aside for curing for about 24 hours. The samples were demolded and edge-trimmed. Figure 5.2a shows some of these laminates.  113    To fabricate glass fiber composites the mold was first covered with the releasing agent. The cut glass fiber sheets were then resinated with the mixture of catalyst-polyester matrix. The rest of the manufacturing process was identical to the one for the wood composite samples.   5.3.4 Test methods To determine the compressive strength and stiffness properties of wood/polyester and glass fiber/polyester laminates, a combined loading compression (CLC) test method was used. The compressive test were performed according to ASTM D-6641 on the Instron universal testing machine. A typical CLC test machine is shown in Figure 5.2b. In this test the specimen is not under pure-shear or pure end-loading conditions. Instead the compressive force is introduced into the specimens by combined end- and shear-loading. Specimens were loaded in compression to failure at a nominal rate of 1.3 mm per minute. The load was continually applied to start the fracture in composite samples and then extended until their final failure. Load-strain data of specimens were collected until failure occurred. For failed specimens maximum force and strain was recorded as near as possible to the moment of failure. The mode, area, and location of the failure for each specimen was observed and analyzed according to standard failure identification codes described in the test guide. Examples of the acceptable failure modes are brooming, sample splitting in the middle, kinking, axial splitting, shear failure, buckling, fiber microbuckling and matrix cracking. Unacceptable modes include through thickness failure at the top of the specimen.    The area under load-displacement curve was measured using Trapezoid method to investigate maximum energy absorbing ability, also called work of fracture (Wf), of the 114   laminates. In trapezoid method the curve divides into series of trapezoids, each with area equals to average height multiplies the width. Total energy absorption will be then sum of the areas of the strips.  Figure 5.3 shows an example of this procedure for one of the laminates consisted of five trapezoids up to the failure point. Since tested specimens are different from each other with respect to their number of layers, specific energy absorption (SEA) was found to be a suitable parameter to compare between the laminates. The specific energy is defined as the amount of energy absorbed per unit mass of compressed material as is shown in Equation 6:    (6)    In the above equation Wf is the work of fracture and M the mass of the specimen.   5.4 Results Maximum compressive load, maximum deflection to failure, compressive strength and modulus, work of fracture up to the failure point, and specific energy absorption of all laminates were calculated from load-displacement graphs. The compressive properties are presented in Table 5.2.   5.4.1 Compressive properties of laminates From the series of experiments performed, it is found 12SB samples subjected to axial load has the maximum total energy absorption (2.10 J) as shown in Table 5.2. While 8SA samples exhibited the minimum value of total energy absorption (0.77 J). Glass fiber/polyester samples 115   had maximum specific energy absorption (0.17 J/g), followed by 12SB (0.14 J/g), 11PW (0.12 J/g), 12SA and 8SB (0.11 J/g). 12NS produced (0.09 J/g) and 8SA and 8NS came in last (0.08 J/g). Appendix D presents stress-strain curve of two samples from 8SB and 12NS treatments.     5.4.1.1 Effect of symmetry One-way ANOVA (Bulmer, 1979) was performed to determine if the difference among treatments was statistically significant. Based on the ANOVA analysis the effect of laminate symmetry on compressive properties of polyester composites was statistically significant (Appendix E). In both 8-layer and 12-layer treatments laminates with face layers of 90 had significantly higher deflection to failure than those with face layers of 0. The compressive strength of laminates and their SEA, however, had not been affected by the laminates symmetry. No significant difference of compressive modulus was reported between 12-layers wood/polyester samples.  In 8-layer composites, however, the compressive modulus was dependent to the symmetry. For example the compressive modulus of 8SB samples were significantly lower than 8SA samples.  5.4.1.2 Effect of balance Balanced and unbalanced laminates had significantly different performances under the compressive load as shown in Appendix E.  The unbalanced samples of this study were coded as 8NS and 12NS. Despite the higher deflection to failure, higher compressive modulus, and greater work of fracture values of 8NS samples in comparison with 8SAs, the differences between these two treatments were not statistically significant.  Regarding the 12-layer composites 12NS treatment had insignificantly higher deflection to the failure but less work of fracture value when 116   compared to the 12SA samples. Balanced symmetric 8-layer and 12-layer composites with the face layers in 90º had superior compressive performances in comparison with the unbalanced ones.    5.4.1.3 Effect of number of lamina Laminates thickness had a significant effect specifically on the compressive modulus and work of fracture values of wood/polyester sample. Thicker composites with 11-layers and 12-layers had remarkably higher work of fracture values than 8-layer samples. The higher number of lamina, however, reduced the compressive modulus of these composites. This trade-off was statistically significant among treatments (Appendix E).   5.4.1.4 Wood/polyester vs. glass fiber/polyester and the commercial treatments  Wood/polyester laminates of this study insignificantly tolerated higher load to failure than glass fiber/polyester composites, Table 5.2. The synthetic fiber composites had considerably higher compressive modulus but lower deflection to failure and work of fracture compared to the wood samples. The SEA of the GP laminates was not significantly higher than the wood/polyester samples. After comparison the results of compressive properties of wood/polyester samples with those of the commercial treatments it was concluded that the fabricated polyester laminates had significantly superior performances than MDF boards. Commercial Plywood and Solid wood samples, however, had the best results among all the treatments including glass fiber/polyester.  117   5.4.2 Comparison of load-displacement curves  The mode of failure and the pattern of load–displacement curves of the various composite laminates were different. Figure 5.4 shows a typical graph of an average load versus displacement in 8SB and 12SB wood/polyester composites and in glass fiber/polyester laminates. Based on the results of the ANOVA these three treatments had significantly superior properties in various compressive properties than the other samples. Therefore these three treatments were selected for comparison the load-displacement curve. Since the mode of failure and the pattern of load–displacement curve of these various composite laminates were different, each are discussed in more detail.  5.4.2.1 Glass fiber/polyester (GP) The average load-displacement curve for glass fiber-polyester composites can be divided into two regions as shown in Figure 5.4, the first region started from 0 N and raised through the elastic zone until it reached to the maximum load of 1906 N at 0.6 mm displacement. After this point, the curve dropped dramatically as a result of sudden brittle failure. The thin samples, once failed, all showed obvious kinks when under load. However, those kinks flatted out when the load was removed. The kinking was the dominant mode of fracture for all the tested samples of this treatment leading to an unstable brittle failure.   5.4.2.2 Eight-layers symmetric with face layers of 90 (8SB) 8SB treatment shows a noticeably lower slope of the average load-displacement curve indicating a lower compressive modulus than glass fiber composites, Figure 5.4. This treatment, however, had various progressive fractures up to the maximum average load of 2186 N at 1.06 mm 118   displacement. The load after this point was gradually decreased. Light micrographs of the fractured samples showed no severe plies delamination but a shear buckling mode of failure which was observed in most of the failed samples, Figure 5.6a and b. The average load-displacement curve for this graph exhibited high load carrying ability with obviously high deflection to failure under quasi-static compressive load when compared with that in GP curve.   5.4.2.3 Twelve-layers symmetric with face layers of 90 (12SB) Curve of the average load-displacement for 12SB samples has a lower slope than the GP and 8SB curves, Figure 5.4. However, samples progressively absorbed energy to the failure point and broke at the average load of 2197 N at 1.20 mm displacement. The compressive load was decreased gradually after the failure. The curve of 12SB is similar with the one for 8SB but the area under the curve is significantly larger. Regarding the fracture mechanisms12SB showed a more consistent trend, up to the failure when compared to the 8SB. This 12-layer treatment, however, had delamination as the dominant mode of fracture, Figure 5.7a. Symmetric cross-ply laminates with face layers of 90 generally underwent the largest inelastic deformation before failure. This ply configuration is capable of absorbing large amounts of energy before fracture.  5.4.3 Fracture mechanisms of other treatments Among the 8-layer treatments 8SA samples had the least maximum load to failure, 1880 N in average, and also the least deflection to failure at 0.74 mm. 6 out of 8 tested samples of this treatment had severe centre delamination as the dominant failure mode, Figure 5.5a. This severe delamination occurred at the initial steps of energy absorption process leading to laminate 119   failure. As it is shown the generated crack of delamination longitudinally extended through the sample and this significantly reduced the maximum failure load. In addition to the delamination failure mode the other two tested samples had a more consistent fracture mechanisms including plies buckling and face layers transverse crack extension, Figure 5.5b. These samples had a higher failure load and greater deflection to failure than those with the severe delamination as the main fracture mode. For 8SB samples it is worth mentioning since their fracture mechanisms were quite similar to the one in the commercial Plywood samples. Both of these treatment had shear buckling as the dominant failure mode. Unbalanced samples, 8NS and 12NS, had very steady fracture patterns including buckling, face layers delamination, and inter and intralaminar crack extensions. Figure 5.7b shows the fracture mechanisms of the 12NS sample. The 11-layer samples absorbed a considerable amount of compressive energy nearly to the peak load but then failed due to the delamination mostly between the 4th and 5th layers of the laminates, Figure 5.7c.    The damage development of wood/polyester laminates during compression was examined by scanning electron microscopy (SEM) technique. SEM micrographs, Figures 5.8 to 5.10, revealed fiber-matrix debonding, fiber pull-out, and fiber bridging as the other fracture mechanisms led to the laminates failure.  The fiber pull-out and interface fiber-matrix debonding can be an indication of poor adhesion between the fiber and the matrix observed in some of the plies.   120   5.5 Discussion 5.5.1 Effect of symmetry, balance, and number of lamina Wood is stronger parallel to the fiber direction when it is under a compressive load. Considering the fracture mechanism of 8-layer laminates those with symmetry B had strong bond in center line which avoided delamination, Figure 5.6a and b. Not surprising that the 0º layers in the centre-line of 8SB samples had higher resistance under compressive load in comparison with the 8SA laminates having 90º plies at the centre line. In addition, samples with symmetry B had shear buckling as the main fracture mechanism that indicates a good fiber-matrix interface adhesion resulted a higher deflection to failure for the composites. In contrast, 8SA laminates had severe center delamination as it is shown in Figure 5.5a. This delamination separated the plies at the first 30 seconds of load application led a low failure load and consequently a low deflection to failure, Table 5.2. Figure 5.8 shows SEM micrographs of fracture in 8SA sample. Despite the severe delamination in the centre of 8SA samples the face layer which was laid up in 0º direction showed buckling fracture which has often reported as fracture mechanism of strong fibers under axial load. Despite this good energy absorption mechanism the laminates failed due to the severe centre delamination.    Similar to the compressive behavior of 8-layesrs samples, 12SB laminates exhibited a higher deflection to failure than 12SA. The only difference is 12SB laminates did not have strong fiber-matrix adhesion as it was observed for 8SB.  In fact they failed due to the delamination fracture. Higher number of lamina in 11 and 12-layers laminates increased their deflection to failure but it significantly decreased the compressive modulus, Table5.2. Considering the trade-off between these two properties it is believed increasing laminate thickness brought unwanted 121   gaps between plies which led to a poor interface adhesion between the fiber and polyester. As it is shown in Figure 5.9 fiber-matrix debonding and fiber pull-out were fracture mechanisms for 12-layers wood/polyester laminates. These two mechanisms are known as the indicators of poor interface bonding. Depending on the final application of the composites low compressive modulus may not be an issue as the laminates had a high deflection to failure and long failure time. However automated lay-up or compression molding as other composite manufacturing techniques may improve the lamination process and the plies adhesion.   Unbalanced laminates of 8NS and 12NS had better performances than 8SA and 12SA samples but were weaker than 8SB and 12SB under the compressive load. Failure of these samples under compressive load was differed depending on the face layer position; it was buckling for 0º face layers and transverse cracking occurring in the 90º face layers. In addition, the presence of 0º in the centre-line avoided centre delamination. Accordingly, they absorbed higher energy and had more deflection to failure than the samples with symmetry A. These two unbalanced treatments, however, have had lower energy absorption than laminates with symmetry B as they had only one 0º layer in the centre-line.   5.5.2 Wood/polyester vs. glass fiber/polyester and the commercial treatments The wood/polyester laminates of this study showed a combination of fracture mechanisms which resulted a progressive failure. The maximum load to failure and deflection to failure values were then higher than those in glass fiber/polyester samples. The compressive modulus of wood/polyester laminates, however, decreased significantly as the trade-off. The curve in glass fiber/polyester composites got a very high slope but low deflection to failure because of the 122   brittle failure, Figure 5.4. This led a small work of fracture value for GP composites. Nevertheless, after eliminating the effect of weight of laminates they had the highest specific energy absorption value among wood/polyester laminates. Wood/polyester samples were stronger than MDF boards but had less energy absorption than commercial Plywood. One of the reason is the glue difference which was phenol formaldehyde in the commercial Plywood and polyester used here. These two polymers have considerably different mechanical properties which affect the final properties of the boards.  Solid wood was the strongest among all the treatments. In composite manufacturing, however, when we deal with two different ingredients it is important to enhance their compatibility and interactions.   5.5.3 Fracture mechanisms Different fracture mechanisms were recorded for different treatments.  The goal of this research was to increase the number of fracture mechanism in the wood/polyester samples in order to improve their energy absorbing ability under compression load. As it is shown in Figure 5.4 the load-displacement curve in glass fiber/polyester laminates dropped down dramatically as the result of a brittle failure. In this study fiber and ply buckling had the most energy absorption ability among other fracture mechanisms. Generally, when the applied load in the structure remains constant and it starts to warp, buckling occurs. Plastic deformation of fiber and matrix results this fracture pattern. If the laminate doesn’t have buckle formation its failure will be then catastrophic as it was seen for GP samples. Fiber kinking alone can absorb considerable amount of the energy, Table 5.2, but this fracture mechanism in high stiffness composites cannot avoid the brittle failure.   123    All wood/polyester samples in this study with lay-up A showed plies buckling as one of the fracture mechanisms. 8SA samples had severe center delamination but face layer buckling retarded the final failure. In some samples each ply was buckled under load and this resulted in global shear buckling for the laminates. As it is mentioned the commercial Plywood had a considerable work of fracture which indicates its high energy absorbing ability. Comparison of the fracture mechanisms of this treatment and laminates with symmetry B, Figure 5.6, shows that they both had buckling as the dominant mode of failure.  In 8-layer composites no fiber-matrix debonding was observed. There was one or two cases of fiber pull-out as it is shown in Figure 5.8 but it was not the dominant mechanism. In contrast, 12-layer samples had several fiber pull-out and debonding which indicates poor fiber-matrix adhesion.  Figure 5.9 shows these mechanisms. Transverse cracking, another fracture mechanisms, is presented in Appendix F. It is hard to discuss which of the fracture mechanisms came first as they were inside fractures. It is believed now the higher number of lamina in 12-layer samples added more fracture mechanism to the laminate. There were some fracture mechanisms which contributed significantly in this energy absorption. Fiber bridging, another energy absorbing mechanism retards the final failure of the composite, was seen in12SB samples, Figure 5.10. As explained, the fracture mechanisms varied among treatments according to their lay-up configurations.  5.6 Conclusions Based on the results obtained, the main conclusion points can be summarized as following: 1. The effect of symmetry on compressive properties of wood veneer/polyester laminates was statistically significant. Laminates with face layers of 90 and core layers of 0 had the highest deflection to failure. 124   2. Balanced wood veneer/polyester composites with identical face layers had higher values of compressive properties compared with the unbalanced laminates that have different face layers.  3. Increasing the wood/polyester laminate thickness enhanced their energy absorbing ability by having more fracture mechanisms available for dissipating energy. 4. There was a trade-off between compressive modulus and energy absorbing ability of the laminates; the highest area under the curve was belonged to the 12-layer samples with the lowest compressive modulus. 5. Shear buckling was the dominant mode of fracture for the symmetric balanced laminates.  6. Lab-made glass fiber/polyester composites had a higher SEA value than wood laminates. However, glass fiber/polyester laminates were very brittle and failed catastrophic only with very low deflection to failure.   7. Eight-layer samples with core layers of 0º, 8SB, had the average compressive modulus and energy absorbing ability which can be considered as the best laminate of this research.   125   Table 5.1 List of treatments. Treatments Label Lay-up code Control-Glass fiber Polyester GP - 8 layers Symmetric with 0s on faces (ǀ) 8SA [0/90]2s 8 layers Symmetric with 90s on faces (ǁ) 8SB [90/0]2s 8 layers Asymmetric 8NS [90/0]4t 12 layers Symmetric with 0s on faces (ǀ) 12SA [0/90]3s 12 layers Symmetric with 90s on faces (ǁ) 12SB [90/0]3s 12 layers Asymmetric 12NS [90/0]6t 11 layers plywood 11PW [0/90/0/90/0/ 90 ]s Commercial MDF MDF - Commercial Plywood with PF PW - Solid Wood Wood - Note: The subscripts “s” and “t” refer to the words “symmetric” and “total”, respectively.  Note: The bar over the 90 ply indicates the centerline of symmetry passes midway through this ply.      126   Table 5.2 Mean compressive properties of laminates and the associated St. Dev. (n= 8 replicates). Treatment Pmax (N) Df (mm) Fcu (MPa) Ec (MPa) Wf (J) SEA (J/g) GP 1906 (206) 0.60 (0.4) 88 (10) 1830 (664) 0.80 (0.4) 0.17 (0.08) 8SA 1880 (716) 0.74 (0.1) 18 (7) 452 (28) 0.77 (0.3) 0.08 (0.0) 8SB 2186 (145) 1.06 (0.1) 20 (3) 294 (45) 1.13 (0.3) 0.11 (0.0) 8NS 2131 (243) 0.82 (0.4) 19 (2) 592 (35) 0.85 (0.3) 0.08 (0.0) 12SA 2272 (185) 1.03 (0.5) 13 (2) 227 (53) 1.62 (0.4) 0.11 (0.1) 12SB 2197 (148) 1.20 (0.1) 14 (3) 267 (32) 2.10 (0.8) 0.14 (0.0) 12NS 2316 (139) 1.10 (0.2) 14 (2) 203 (36) 1.41 (0.3) 0.09 (0.0) 11PW 2262 (118) 1.18 (0.2) 15 (4) 231 (71) 1.78 (0.4) 0.12 (0.0) MDF 1671 (186) 0.82 (0.0) 14 (2) 295 (46) 0.67 (0.1) 0.05 (0.0) PW 4614 (302) 1.50 (0.1) 38 (5) 375 (32) 3.58 (1) 0.37 (0.1) Wood 9742 (226) 1.25 (0.1) 68 (2) 1221 (185) 5.98 (0.4) 0.44 (0.0) Note: Pmax is the maximum load and Df the displacement to failure. Note: Values in parenthesis are standard deviations. 127   Figure 5.1 Schematic representation of the symmetry and balance in composite laminates (a 3D representation of stacking plies in a symmetric-balanced specimen is shown in the right side where thereby arrows indicate fiber direction).                                    0        90  A  B    0  0  90  90  90  90  0  0  0  0  90  90  90  90  0  0  0  90  0  90  90  0  90  0  0  90  0  90  90  0  90  0  0                                                 Symmetric                    Asymmetric                                 Balanced                      Unbalanced                             8 7 6 5 4 3 2 1 128   Figure 5.2 (a) Demolded and trimmed laminates; and (b) a typical CLC test machine. 129   Figure 5.3 Trapezoid method for energy absorption calculation of laminates up to the failure point.              050010001500200025000.0000 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009Load (N)Displacement (m)8NS130   Figure 5.4 Average load-displacement curve in wood/polyester and glass fiber/polyester laminates (error bars represent 95% confidence intervals).               050010001500200025000 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2Average  load (N)Displacement (mm)GP8SB12SB131   Figure 5.5 Light micrographs of 8SA samples (a) centre delamination, and (b) multiple fractures in all layers.              132   Figure 5.6 Light micrographs of (a) buckling in 8SB, (b) buckling and delamination in 8SB, and (c) buckling in commercial Plywood.  133   Figure 5.7 Light micrographs of fractured 12 layers laminates: (a) delamination and transverse cracking of face layers in 12SB, (b) multiple fractures in all layers of 12NS, and (c) delamination between 4th and 5th layers of 11PW.  134   Figure 5.8 SEM micrographs of fracture mechanisms of 8SA laminate where b is a higher magnification of the area shown in the white square in a.  135   Figure 5.9 SEM micrographs of fracture mechanisms of 12SB laminate where b shows a higher magnification of the area shown in the white square in a.   136   Figure 5.10 Fiber bridging mechanism in 12SB sample.          137   Chapter 6: Conclusion  6.1 Project summary There is an increase in the use of fiber-reinforced polymers as alternatives to traditional structural materials such as concrete and metals. Natural reinforcements including wood and plant fibers are cheap, light-weight, and abundant when compared to the heavier and more expensive synthetic fibers.  Considering the trade-off between stiffness and energy absorbing ability of materials, stiff composites from synthetic fibers usually experience catastrophic brittle failure under impact or sudden load. In this study the potentials of using wood as an alternative material to synthetic fibers is investigated. Since the current energy absorbers in transportation industries are made of glass fiber/polyester composites, polyester resin was chosen as matrix in this research.  The focus of this study was on manufacturing laminates of wood and polyester in order to examine the wettability of wood veneers to the polyester matrix and also to characterize the impact properties and energy absorbing ability of the fabricated samples.     Chapter 2 of this dissertation reviewed the literature on the use of natural fiber in combination with polyester matrix. Plant fibers including hemp, jute, banana, sisal, coir, kenaf, bamboo, rice, oil palm, and flax are used extensively to reinforce polyester matrix. The matrix reduces the hygroscopicity of the natural fibers and improves the physical properties of the final composite. Fracture mechanisms of plant fiber/polyester composites and their performances under impact and bending loads were comprehensively reviewed and discussed in this chapter. There was a significant gap of information on the compatibility of wood with polyester, composite lamination and curing processes, and its fracture mechanisms under mechanical loads.  138   To address these issues and to fill the gap of knowledge, a detailed investigation on the use of a combination of wood with the polyester matrix was performed.    Chapter 3 experimentally studied use of wood and polyester in composites fabricated as thin plates using hand lay-up and compression molding techniques. Making thin composite plates facilitated identification of dominant fracture mechanisms for each treatment. To investigate the effect of densification on the impact properties of the plates, Douglas fir veneers were densified in150ºC heat and under pressure of 1.38 MPa. Three different mat configurations were introduced to the treatments, woven, cross and, unwoven (unidirectional) mats in order to examine the wood component configuration effect on the fracture patterns and on the absorbed impact energy. Results of this experiment showed that the densification method improved the impact behavior in the samples but this improvement was not statistically significant. The used densification technique, however, did not damage the wood cell wall. Using optical microscopy indicated the degree of densification was sufficiently high to compact the wood cells without breaking their walls. Among the various configurations, woven and cross mats led to a higher level of impact energy in comparison with the unidirectional samples. The average mass fraction of woven and cross mats were 11% higher than the unidirectional samples because twice the number of veneer strips were used to make the woven mat. This led to a thicker sample and a better performance under impact load. Comparing woven and cross configurations mats there was no significant difference in their impact outcomes. Since the woven mats preparation is a time consuming process, cross configuration is recommended as the best treatment for this experiment. There was a considerable difference in the mode of failure and failure patterns of wood/polyester samples and glass fiber/polyester composites. Since wood is weaker in the 139   transverse direction most of the cracks longitudinally propagated along the fiber direction. Other fracture patterns included splits, zigzag patterns and sharp splinters. Increased thickness in wood composites with face layers of glass fiber brought a higher number of fracture mechanisms to play and as a result improved the impact performance of the treatment. Based on the SEM micrographs before and after mat resination, the polyester resin completely surrounded the wood component and had an even distribution in the composite. Similar findings were concluded from SEM micrographs of glass fiber composites where the synthetic fibers were entirely surrounded by the matrix.    The effects of lay-up configuration was examined in Chapter 4. The finding of the previous experiment showed that increasing the thickness of the composite brings more fracture mechanisms and improves the mechanical performance of the composite. Three different wood veneers were used, Douglas fir, maple, and oak to investigate which of these veneers can make the best laminate as the anatomy of these three species is quite different. Two grit sizes (120 and 320) were used to determine the effect of sanding on veneer wettability. A laser confocal microscope was then used to measure surface roughness of the veneers before and after sanding.  The results showed that the highest wettability to polyester among the species used in this experiment was in Douglas fir, followed by maple and oak. The reason is the cells in Douglas fir and maple are smaller than those of oak (tracheids in softwood and pores in hardwood). When the polyester drops were placed onto the veneers, they were absorbed by the capillary action of these elements. This absorption is slower in oak wood as the capillary forces of large pores is smaller than that of the pores of Douglas fir and maple, consequently limiting the wettability of oak. The 120-grit sandpaper cleaned the surface of the veneers from dust and contamination, 140   creating a fresh, smooth surface, and increasing the wettability of the veneers. This increase compared to the unsanded veneers was significant for maple and oak veneers. Grit 320, in contrast, reduced wettability in all veneers as the dust generated from sanding filled in the open pores and reduced the absorption of polyester. Roughness tests showed that grit 120 reduced surface roughness in maple and oak veneers and increased their wettability to polyester by reducing their contact angles. Using grit size higher than 120 is not recommended as it reduces the wettability of wood veneers to polyester.    The texture of wood species also affected their lamination process. It was shown that Douglas fir veneer with the medium-textured anatomy had a better resin distribution by keeping the resin inside after absorption. This led to flat polyester laminates. In contrast to Douglas fir veneer sheets, maple and oak had considerable twisting and shrinkage during the lamination process and the coarse-texture of Oak with large diameter pores could not retain the polyester inside the laminate. Unidirectional, balanced, and unbalanced cross-ply laminates were compared to control specimens using glass fiber as reinforcement. It was found that cross-ply wood/polyester laminates had an impact energy equivalent to glass fiber laminates, making them an interesting alternative to the use of synthetic fibers as reinforcement. It is believed the good bonding between the veneer sheets in the Douglas fir samples was attributed to this result. Making the laminates balanced reduced veneer twisting during the lamination process and make their composites flat specifically in maple composites.   Considering the findings of experiment described in Chapter 4, Douglas fir veneer sheet was selected as the wood furnish in polyester matrix and their energy absorbing ability under 141   compressive load and their associated fracture mechanisms were investigated in Chapter 5. In the previous experiment laminates were made as four plies but in this study in order to avoid Euler buckling during compression, samples had to be fabricated in a greater thickness to meet the thickness criteria.  The effects of symmetry (face layers of 0º or 90º), lay-up balance (balanced and unbalanced), and number of lamina (8, 11, and 12) on compressive properties of wood/polyester laminates were investigated. Since wood is an orthotropic material it was hypothesized that the performance of 0º and 90º layers under compressive load will be different. Hence, two types of symmetry were used: symmetry A where both face layers are laid-up in 0º direction, and symmetry B where they are laid-up in 90º direction. The former has 90º layers around the centre line and the latter has 0º layers. Because symmetric lay-up is balanced it improves dimensional stability of the boards which is crucial in wood-based composites. In order to meet the thickness requirements wood composites of this study were fabricated as eight plies. It was hypothesized that increasing wood laminate thickness improves the energy absorbing ability of composites by bringing more fracture mechanisms into play for the sample. Hence, 12-layers laminates were introduced to be compared with 8 layers composites. 11 plies composites of wood and polyester was made for simulation of plywood with odd number of stacked plies.    Results showed that the effect of symmetry on compressive properties of wood veneer/polyester laminates was statistically significant and laminates with face layers of 90 and core layers of 0 had the highest deflection to failure. Since wood is stronger parallel to the fiber direction, the 0º layers in the centre-line of 8SB samples had higher resistance under compressive load in comparison with the 8SA laminates having 90º plies at the centre line. 8SA samples, in contrast, had severe centre delamination. Balanced wood/polyester laminates with identical face 142   layers had higher values of compressive properties compared with the unbalanced laminates that have different face layers. Similar to the compressive behavior of 8-layesrs samples, 12SB laminates exhibited a higher deflection to failure than 12SA. Higher number of lamina in 11 and 12-layers laminates increased their deflection to failure but significantly decreased the compressive modulus. It is believed that increasing laminate thickness brought unwanted gaps between plies which led to a poor interface adhesion between the fiber and polyester. Analysis of fracture mechanisms for these thick samples also showed fiber pull-out and fiber-matrix debonding which are indicators of a poor interface bonding. Despite the brittle failure of glass fiber composites wood/polyester laminates exhibited a progressive fracture mechanisms with shear buckling as the dominant mode of failure in symmetric samples. This progressive failure with high energy absorbing ability make wood/polyester laminates a good candidate to be used in applications where high deflection to failure and longer failure time are required.    6.2 Conclusions In this research, the gap of information on energy absorbing ability and fracture mechanisms of wood/polyester composite laminates was filled. The distribution of polyester resin in wood composites was examined and the wood veneers from three different species were compared in terms of their wettability to the matrix. The performances of wood veneer/polyester composites were evaluated under impact load and quasi-static combined loading compression and the associated fracture mechanisms were discussed in detail and compared with the lab-made glass fiber/polyester composites. It can be concluded that composites from wood veneer and polyester with respect to lay-up configuration and number of lamina can absorb a significant amount of 143   energy which is on par with glass fiber laminates. The progressive failure which was achieved in wood/polyester laminates of this study extended the time of final failure.  6.3 Outlook  There were, however, limitations to this research. There was limitation with available equipment to develop the densification technique, and as a result the effect of densification on impact properties of polyester composites in the current research was not significant. Using a steam-injection compression molding machine will be a better option to improve the densification process and to get a higher degree of densification, taking into consideration that the wood cell wall should not break. In addition, using an automated hand lay-up or other automatic manufacturing techniques for fabricating composite laminates will increase consistency in comparison with the manual hand lay-up.  The next step for this project will be first to investigate the interface properties of wood/polyester composite laminates and measure their bonding strength, and second to conduct dynamic crush load to the wood/polyester composites and examine the laminates failure mechanisms.  144   Bibliography Adalian, C., & Morlier, P. (2001). A model for the Behaviour of Wood Under Dynamic Multiaxial Compression. Composites Science and Technology, 61(3), 403–408. http://doi.org/10.1016/S0266-3538(00)00109-3 Adhikary, K. B., Pang, S., & Staiger, M. P. (2008). Dimensional stability and mechanical behaviour of wood–plastic composites based on recycled and virgin high-density polyethylene (HDPE). Composites Part B: Engineering, 39(5), 807–815. http://doi.org/10.1016/j.compositesb.2007.10.005 Ahmad, S. H., Rasid, R., Bonnia, N. N., Zainol, I., Mamun, a. a., Bledzki, a. K., & Beg, M. D. H. (2010). Polyester-Kenaf Composites: Effects of Alkali Fiber Treatment and Toughening of Matrix Using Liquid Natural Rubber. Journal of Composite Materials, 45(2), 203–217. http://doi.org/10.1177/0021998310373514 Ahmed, K. S., & Vijayarangan, S. (2007). Experimental Characterization of Woven Jute-Fabric-Reinforced Isothalic Polyester Composites. Journal of Applied Polymer Science, 104, 2650–2662. http://doi.org/10.1002/app Ahmed, K. S., Vijayarangan, S., & Kumar, a. (2007). Low Velocity Impact Damage Characterization of Woven Jute Glass Fabric Reinforced Isothalic Polyester Hybrid Composites. Journal of Reinforced Plastics and Composites, 26(10), 959–976. http://doi.org/10.1177/0731684407079414 Ahmed, K. S., Vijayarangan, S., & Naidu, a. C. B. (2007). Elastic properties, notched strength and fracture criterion in untreated woven jute–glass fabric reinforced polyester hybrid composites. Materials & Design, 28(8), 2287–2294. http://doi.org/10.1016/j.matdes.2006.08.002 Alghamdi, a. a a. (2001). Collapsible impact energy absorbers: An overview. Thin-Walled Structures, 39(2), 189–213. http://doi.org/10.1016/S0263-8231(00)00048-3 Aljawi, A. A. ., & Alghamdi, A. . (2000). Inversion of Frusta as Impact Energy Absorbers. Current Advances in Mechanical Design and Production, Pergamon Press, New York, II, 511–519. Aljibori, H. S. S. (2011). Energy Systems and Crushing Behavior of Fiber Reinforced Composite Materials. World Acaemy of Science, Engineering and Technology, 275–281. Aljibori, H. S. S., Ramle, R., Al-qrimli, H., & W.P, C. (2009). Quasi-Static Axial Crushing of a ± 45 ° Fiber Orientation Angle with a Different Number of Layers : Part 1 , Experimental Study. American Journal of Scientific Research, 4(4), 72–84. 145   Al-Kafi, A., Abedin, M, Z., Beg, M, D, H., Pickering, K, L., & Khan, M, A. (2006). Study on the Mechanical Properties of Jute/Glass fiber-reinforced Unsaturated Polyester Hybrid Composites: Effects of Surface Modification by Ultraviolet Radiation. Journal of Reinforced Plastics and Composites, 25(6), 575–588. Antich, P., Vázquez, A., Mondragon, I., & Bernal, C. (2006). Mechanical Behavior of High Impact Polystyrene Reinforced with Short Sisal Fibers. Composites Part A: Applied Science and Manufacturing, 37(1), 139–150. http://doi.org/10.1016/j.compositesa.2004.12.002 Ashori, A. (2008). Wood-plastic composites as promising green-composites for automotive industries! Bioresource Technology, 99(11), 4661–7. http://doi.org/10.1016/j.biortech.2007.09.043 Ashori, A., & Nourbakhsh, A. (2010). Performance properties of microcrystalline cellulose as a reinforcing agent in wood plastic composites. Composites Part B: Engineering, 41(7), 578–581. http://doi.org/10.1016/j.compositesb.2010.05.004 ASTM D3410. (2013). Standard Test Method for Compressive Properties of Polymer Matrix Composite Materials with Unsupported Gage Section by Shear, 03(Reapproved 2008), 1–16. http://doi.org/10.1520/D3410 ASTM D5420. (2010). Standard test method for impact resistance of flat , rigid plastic specimen by means of a striker impacted by a falling weight ( Gardner Impact ). http://doi.org/10.1520/D5420-10.2 ASTM D5467. (2010). Standard Test Method for Compressive Properties of Unidirectional Polymer Matrix Composite Materials Using a Sandwich Beam, 97(Reapproved 2010), 1–9. http://doi.org/10.1520/D5467 ASTM D6641. (2009). Standard Test Method for Compressive Properties of Polymer Matrix Composite Materials Using a Combined Loading Compression ( CLC ), i, 1–11. http://doi.org/10.1520/D6641 ASTM-D2395. (2007). Standard Test Methods for Specific Gravity of Wood and Wood-Based Materials, 1–9. http://doi.org/10.1520/D2395-07AE01.2 Athijayamani, a., Thiruchitrambalam, M., Natarajan, U., & Pazhanivel, B. (2009). Effect of moisture absorption on the mechanical properties of randomly oriented natural fibers/polyester hybrid composite. Materials Science and Engineering: A, 517(1-2), 344–353. http://doi.org/10.1016/j.msea.2009.04.027 Avci, a., Arikan, H., & Akdemir, a. (2004). Fracture behavior of glass fiber reinforced polymer composite. Cement and Concrete Research, 34(3), 429–434. http://doi.org/10.1016/j.cemconres.2003.08.027 146   Aydin, I. (2004). Activation of wood surfaces for glue bonds by mechanical pre-treatment and its effects on some properties of veneer surfaces and plywood panels. Applied Surface Science, 233, 268–274. http://doi.org/10.1016/j.apsusc.2004.03.230 Ayrilmis, N., Dundar, T., Candan, Z., & Akbulut, T. (2009). Wettability of fire retardant treated laminated veneer lumber (LVL) manufactured from veneers dried at different temperatures. BioResources, 4(4), 1536–1544. Ayrilmis, N., & Winandy, J. (2009). Effects of Post Heat-Treatment on Surface Characteristics and Adhesive Bonding Performance of Medium Density Fiberboard. Materials and Manufacturing Processes, 24(5), 594–599. http://doi.org/10.1080/10426910902748032 Babatunde, A., Olufemi, B., Fuwape, J., & Badejo, S. (2008). Effect of wood density on bending strength and dimensional movement of flake boards from gmelina. In 11th international inorganic-bonded fiber composites conference. Madrid, Spain. Baley, C., Busnel, F., Grohens, Y., & Sire, O. (2006). Influence of chemical treatments on surface properties and adhesion of flax fibre–polyester resin. Composites Part A: Applied Science and Manufacturing, 37(10), 1626–1637. http://doi.org/10.1016/j.compositesa.2005.10.014 Bauser, M., Sauer, G., & Siegert, K. (2006). Extrusion. ASM International. Bhaskar, J., Haq, S., & Yadaw, S. (2011). Evaluation and testing of mechanical properties of wood plastic composite. Journal of Thermoplastic Composite Materials, 25(4), 391–401. http://doi.org/10.1177/0892705711406158 Bledzki, a. K., Reihmane, S., & Gassan, J. (1998). Thermoplastics Reinforced with Wood Fillers: A Literature Review. Polymer-Plastics Technology and Engineering, 37(4), 451–468. http://doi.org/10.1080/03602559808001373 Bledzki, A. K., & Faruk, O. (2003). Wood Fibre Reinforced Polypropylene Composites: Effect of Fibre Geometry and Coupling Agent on Physico-Mechanical Properties. Appl Compos Mater. Blomberg, J., & Persson, B. (2004). Plastic deformation in small clear pieces of Scots pine (Pinus sylvestris) during densification with the CaLignum process. Journal of Wood Science, 50(4), 307–314. http://doi.org/10.1007/s10086-003-0566-2 Boey, F. Y. C., Chia, L. H. L., & Teoh, S. H. (1987). Model for the compression failure of an irradiated tropical wood-polymer composite. International Journal of Radiation Applications and Instrumentation. Part C. Radiation Physics and Chemistry, 29(5), 337–348. http://doi.org/10.1016/1359-0197(87)90003-8 Bulmer, M. (1979). Principles of statistics. 147   CANPLY. (2014). Engineered values. Cappelletto, E., Maggini, S., Girardi, F., Bochicchio, G., Tessadri, B., & Di Maggio, R. (2013). Wood surface protection with different alkoxysilanes: A hydrophobic barrier. Cellulose, 20(6), 3131–3141. http://doi.org/10.1007/s10570-013-0038-9 Casilla, R. C., Chow, S., & Steiner, P. R. (1981). An Immersion Technique for Studiyng Wood Wettability. Wood Science and Technology, 43, 31–43. Cerbu, C., Curtu, I., Ciofoaia, V., & Hanganu, L. C. (2010). Effects of the Wood Species on the Mechanical Characteristics in Case of Some E-glass Fibers / Wood Flour / Polyester Composite Materials. Journal of Material Plastice, 47(1), 109–114. Chetanachan, W., Sookkho, D., Sutthitavil, W., Chantasatrasamy, N., & Sinsermsuksakul, R. (2001). PVC wood: A new look in construction. Journal of Vinyl and Additive Technology, 7(3), 134–137. http://doi.org/10.1002/vnl.10280 Chtourou, H., Ridel, B., & Ait-Kadi, A. (1992). Reinforcement of recycled polyolefins with wood fibers. Journal of Reinforced Plastics and Composites, ll(1992), 372–394. Clemons, C. (2002). Wood-plastic composites in the United States: The interfacing of two industries. Forest Products Journal, 52(6), 10–18. Retrieved from http://www.scopus.com/inward/record.url?eid=2-s2.0-0036610248&partnerID=40&md5=4e72762072c2202faf2236b56986b125 Clemons, C. M. (2000). Woodfiber-Plastic Composites in The United States-- History and Current and Future Markets. In 3rd International Wood and Natural Fiber Composites Symposium (pp. 1–9). Kassel, Germany. Cloutier, A., Fang, C., Mariotti, N., Koubaa, A., & Blanchet, P. (2008). Densification of Wood Veneers Under the Effect of Heat , Steam and Pressure, 1–9. Cui, Y.-H., & Tao, J. (2009). Fabrication and Mechanical Properties of Glass Fiber-Reinforced Wood Plastic Hybrid Composites. Journal Of Applied Polymer Scienc, 112, 1250–1257. http://doi.org/10.1002/app Dabade, B. M., Reddy, G. R., Rajesham, S., & Kiran, C. U. (2006). Effect of Fiber Length and Fiber Weight Ratio on Tensile Properties of Sun hemp and Palmyra Fiber Reinforced Polyester Composites. Journal of Reinforced Plastics and Composites, 25(16), 1733–1738. http://doi.org/10.1177/0731684406068418 Davies, I. J., & Hamada, H. (2007). Compressive failure of unidirectional hybrid fibre-reinforced epoxy composites containing carbon and silicon carbide fibres. Composites Part A: Applied Science and Manufacturing, 38(3), 1070–1074. http://doi.org/10.1016/j.compositesa.2006.10.004 148   De Albuquerque, A. C., Joseph, K., De Carvalho Hecker, L., & Morias D’Almeida, J. R. (2000). € Effect of wettability and ageing conditions on the physical and mechanical properties of uniaxially oriented jute-roving-reinforced polyester composites, 60. DEMİRKIR, C., Aydin, İ., Çolak, S., & Çolakoğlu, G. (2014). Effects of plasma treatment and sanding process on surface roughness of wood veneers. Turkish Journal of Agriculture and Forestry, 38, 663–667. http://doi.org/10.3906/tar-1312-108 Dhakal, H. N., Zhang, Z. Y., & Richardson, M. O. W. (2007). Effect of water absorption on the mechanical properties of hemp fibre reinforced unsaturated polyester composites. Composites Science and Technology, 67(7-8), 1674–1683. http://doi.org/10.1016/j.compscitech.2006.06.019 Dietz, A. G. H. (1949). Engineering laminates. New York; London: John Wiley & Sons, Inc.; Chapman & Hall, Ltd. http://doi.org/10.1079/9781845939564.0363 Diouf, P. N., Stevanovic, T., Cloutier, A., Fang, C. H., Blanchet, P., Koubaa, A., & Mariotti, N. (2011). Effects of thermo-hygro-mechanical densification on the surface characteristics of trembling aspen and hybrid poplar wood veneers. Applied Surface Science, 257(8), 3558–3564. http://doi.org/10.1016/j.apsusc.2010.11.074 Dweib, M. a., Hu, B., O’Donnell, A., Shenton, H. W., & Wool, R. P. (2004). All Natural Composite Sandwich Beams for Structural Applications. Composite Structures, 63(2), 147–157. http://doi.org/10.1016/S0263-8223(03)00143-0 Elgalai, a. M., Mahdi, E., Hamouda, a. M. S., & Sahari, B. S. (2004). Crushing response of composite corrugated tubes to quasi-static axial loading. Composite Structures, 66(1-4), 665–671. http://doi.org/10.1016/j.compstruct.2004.06.002 Erickson, H. D., & Harrison, A. T. (1974). Douglas-Fir Wood Quality Studies Part I : Effects of Age and Stimulated Growth on Wood Density and Anatomy. Wood Science and Technology, 8, 207–226. Fang, C. H., Mariotti, N., Cloutier, A., Koubaa, A., & Blanchet, P. (2012). Densification of wood veneers by compression combined with heat and steam. European Journal of Wood and Wood Products, 70(1-3), 155–163. http://doi.org/10.1007/s00107-011-0524-4 Farley, G. L. (1986). Effect of specimen geometry on the energy absorption capability of composite materials. Journal of Composite Materials, 20, 390–400. http://doi.org/10.1177/002199838602000406 Fávaro, S. L., Ganzerli, T. a., de Carvalho Neto, a. G. V, da Silva, O. R. R. F., & Radovanovic, E. (2010). Chemical, morphological and mechanical analysis of sisal fiber-reinforced recycled high-density polyethylene composites. Express Polymer Letters, 4(8), 465–473. http://doi.org/10.3144/expresspolymlett.2010.59 149   Felix, J. M., & Gatenholm, P. (1991). The nature of adhesion in composites of modified cellulose fibers and polypropylene. Journal of Applied Polymer Science, 42(3), 609–620. http://doi.org/10.1002/app.07.v42 Freischmidt, G., Michell, A. J., Lewis, M, J., & Vanderhoek, N. (1991). Wood Fibre-Polyester Dough Moulding Compounds. Polymer International, 24, 113–117. Gao, H., Xie, Y., Ou, R., & Wang, Q. (2012). Grafting effects of polypropylene/polyethylene blends with maleic anhydride on the properties of the resulting wood-plastic composites. Composites Part A: Applied Science and Manufacturing, 43(1), 150–157. http://doi.org/10.1016/j.compositesa.2011.10.001 Gardner, D. J., Generalla, N. C., Gunnells, D. W., & Wolcott, M. P. (1991). Dynamic wettability of wood. Langmuir, 7(11), 2498–2502. http://doi.org/10.1021/la00059a017 Gibson, L. J., & Ashby, M. F. (1999). Cellular Solids: Structure and Properties (Cambridge Solid State Science Series). Cambridge solid state science series. Retrieved from http://www.amazon.co.uk/Cellular-Solids-Structure-Properties-Cambridge/dp/0521499119 Goodman, D. (1998). Advanced Low-Cost Composite Curing With High Energy Electron Beams-Phase 2. SOMERVILLE MA: CIENCE RESEARCH LAB INC. Gowda, T. M., Naidu, A. C. B., & Chhaya, R. (1999). Some mechanical properties of untreated jute fabric-reinforced polyester composites, 30, 277–284. Gupta, N., Brar, B. S., & Woldesenbet, E. (2001). Effect of filler addition on the compressive and impact properties of glass fibre reinforced epoxy, 24(2), 219–223. Habenicht, G. (2009). Applied Adhesive Bonding: A Practical Guide for Flawless Results. Wiley. Retrieved from http://books.google.com/books?hl=en&amp;lr=&amp;id=AMEwsHrAagEC&amp;oi=fnd&amp;pg=PR11&amp;dq=Applied+adhesive+bonding:+a+practical+guide+for+flawless+results&amp;ots=J7IALzB5Hs&amp;sig=PZzJTa9GUNZuJBVIV9uw07NvOZs Haghdan, S., Tannert, T., & Smith, G. D. (2015a). Effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites. Journal of Composite Materials, 49(10), 1161–1170. http://doi.org/10.1177/0021998314531308 Haghdan, S., Tannert, T., & Smith, G. D. (2015b). Effects of species anatomy and lay-up configurations on wettability and impact performance of wood veneer / polyester composites. Bioresources. Hamada, H., Coppola, J. C., Hull, D., Maekawa, Z., & Sato, H. (1992). Comparison of energy absorption of carbon/epoxy and carbon/PEEK composite tubes. Composites, 23(4), 245–252. http://doi.org/10.1016/0010-4361(92)90184-V 150   Harris, B. (1986). Engineering composite materials. Composites. London, UK.: Institue of metals. http://doi.org/10.1016/0010-4361(87)90420-4 Hill, C. (2007). Wood modification: chemical, thermal and other processes. John Wiley & Sons. Hill, C. A. S., & Khalil, H. P. S. A. (2000). Effect of Fiber Treatments on Mechanical Properties of Coir or Oil Palm Fiber Reinforced Polyester Composites, 1685–1697. Husseinsyah, S., & Mostapha, Z. (2011). The effect of filler content on properties of coconut shell field polyester composites. Malaysian Polymer, 6(1), 87–97. Ishak, M. R., Leman, Z., Sapuan, S. M., Edeerozey, a M. M., & Othman, I. S. (2010). Mechanical properties of kenaf bast and core fibre reinforced unsaturated polyester composites. IOP Conference Series: Materials Science and Engineering, 11, 012006. http://doi.org/10.1088/1757-899X/11/1/012006 ISO 4288. (1996). Geometrical Product Specifications (GPS)- Surface texture: Profile method- Rules and procedures for the assessment of surface texture. Jacob, G., Fellers, J. F., Simunovic, S., & Starbuck, M. (2002). Energy absorption in polymer composites for automotive Crashworthiness. Journal of Composite Materials, 36(7), 158. Jarukumjorn, K., & Suppakarn, N. (2009). Effect of glass fiber hybridization on properties of sisal fiber–polypropylene composites. Composites Part B: Engineering, 40(7), 623–627. http://doi.org/10.1016/j.compositesb.2009.04.007 Jarusombuti, S., & Ayrilmis, N. (2011). Surface characteristics and overlaying properties of flat-pressed wood plastic composites. European Journal of Wood and Wood Products, 69(3), 375–382. http://doi.org/10.1007/s00107-010-0440-z Jiang, H., & Kamdem, D. P. (2003). Mechanical properties of poly (vinyl chloride)/wood flour/glass fiber hybrid composites. Journal of Vinyl and …, 9(3), 138–145. Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/vnl.10075/abstract Jiang, H., & Kamdem, D. P. (2004). Development of poly(vinyl chloride)/wood composites. A literature review. Journal of Vinyl and Additive Technology, 10(2), 59–69. http://doi.org/10.1002/vnl.20009 Karina, M., Onggo, H., Dawam Abdullah, A, H., & Syampurwadi, A. (2008). Effect of oil palm empty fruit bunch fiber on the physical and mechanical properties of fiber glass reinforced polyester resin. Journl of Biological Sciences, 8(1), 101–106. Kellogg, R. M., & Wangaard, F. F. (1969). Variation in the cell-wall density of wood. Wood and Fiber. 151   Khalil, H. P. S. a., Hanida, S., Kang, C. W., & Fuaad, N. a. N. (2007). Agro-hybrid Composite: The Effects on Mechanical and Physical Properties of Oil Palm Fiber (EFB)/Glass Hybrid Reinforced Polyester Composites. Journal of Reinforced Plastics and Composites, 26(2), 203–218. http://doi.org/10.1177/0731684407070027 Kilic, M., Burdurlu, E., Aslan, S., Altun, S., & Tümerdem, Ö. (2009). The effect of surface roughness on tensile strength of the medium density fiberboard (MDF) overlaid with polyvinyl chloride (PVC). Materials and Design, 30(10), 4580–4583. http://doi.org/10.1016/j.matdes.2009.03.029 Kilic, M., Hiziroglu, S., & Burdurlu, E. (2006). Effect of machining on surface roughness of wood. Building and Environment, 41(8), 1074–1078. http://doi.org/10.1016/j.buildenv.2005.05.008 Kretschmann, D. E. (2010). Chapter 5 - Mechanical Properties of Wood. Wood Handbook - Wood as an Engineering Material, 1–46. Kretsis, G. (1987). A review of the tensile , compressive, flexural and shear properties of hybrid fibre- reinforced plastics. Composites, 18(1), 13–23. Kushwaha, P. K., & Kumar, R. (2009). Studies on Water Absorption of Bamboo-Polyester Composites: Effect of Silane Treatment of Mercerized Bamboo. Polymer-Plastics Technology and Engineering, 49(1), 45–52. http://doi.org/10.1080/03602550903283026 Kutnar, A., Kamke, F. a., & Sernek, M. (2008). The mechanical properties of densified VTC wood relevant for structural composites. Holz Als Roh- Und Werkstoff, 66(6), 439–446. http://doi.org/10.1007/s00107-008-0259-z Kutnar, A., Kamke, F., & Sernek, M. (2009). Density profile and morphology of viscoelastic thermal compressed wood. Wood Science and Technology, 43(1-2), 57–68. http://doi.org/10.1007/s00226-008-0198-1 Kutnar, A., & Šernek, M. (2007). Densi fi cation of wood, 82. Kutnar, A., Sernek, M., & Kamke, F. (2007). Viscoelastic thermal compression (VTC) of wood. In 18th International Conference on new technologies and materials in industries based on the forestry sector (pp. 73–76). Zagreb, Croatia. Lee, J., & Soutis, C. (2007). A Study on the Compressive Strength of Thick Carbon Fibre–Epoxy Laminates. Composites Science and Technology, 67(10), 2015–2026. http://doi.org/10.1016/j.compscitech.2006.12.001 Li, T. Q., & Li, R. K. Y. (2001). A fracture mechanics study of polypropylene–wood flours blends. Polymer-Plastics Technology and Engineering, 40(1), 1–21. http://doi.org/10.1081/PPT-100000116 152   Li, X., Tabil, L. G., & Panigrahi, S. (2007). Chemical Treatments of Natural Fiber for Use in Natural Fiber-Reinforced Composites: A Review. Journal of Polymers and the Environment, 15(1), 25–33. http://doi.org/10.1007/s10924-006-0042-3 Liang, J. Z., & Li, R. K. Y. (1998). Mechanical properties and morphology of glass bead-filled polypropylene composites. Polymer Composites, 19(6), 698–703. http://doi.org/10.1002/pc.10142 Macias, N., Knowles, C., Kamke, F., & Kutnar, A. (2011). Commercialization potential of viscoelastic thermal compressed wood : insights from the US forest products industry. Forest Product Journal, 61(7), 500–509. Magoss E. (2008). General Regularities of Wood Surface Roughness. Acta Silv. Lign. Hung., 4, 81–93. Maldas, D., & Kokta, B. V. (1992). Performance of Hybrid Reinforcements in PVC Composites: III. Use of Surface-Modified Glass Fiber and Wood Pulp as Reinforcements. Journal of Reinforced Plastics and Composites, 11(10), 1093–1102. http://doi.org/10.1177/073168449201101002 Mamalis, a. ., Robinson, M., Manolakos, D. E., Demosthenous, G. a., Ioannidis, M. B., & Carruthers, J. (1997). Crashworthy capability of composite material structures. Composite Structures, 37(2), 109–134. http://doi.org/10.1016/S0263-8223(97)80005-0 Marcovich, N. E., Reboredo, M. M., & Aranguren, M. I. (1996). Composites from Sawdust and Unsaturated Polyester. Journal of Applied Polymer Science, 61(7600), 119–124. Marcovich, N. E., Reboredo, M. M., & Aranguren, M. I. (1998a). Dependence of the Mechanical Properties of Woodflour – Polymer Composites on the Moisture Content. Journal of Applied Polymer Science Science, 68, 2069–2076. Marcovich, N. E., Reboredo, M. M., & Aranguren, M. I. (1998b). Mechanical properties of woodflour unsaturated polyester composites. Journal of Applied Polymer Science, 70, 2121–2131. Melo, J. D. D., & Villena, J. E. N. (2012). Effect of Fiber Volume Fraction on the Energy Absorption Capacity of Composite Materials. Journal of Reinforced Plastics and Composites, 31(3), 153–161. http://doi.org/10.1177/0731684411433061 Mishra, S., Mohanty, a. ., Drzal, L. ., Misra, M., Parija, S., Nayak, S. ., & Tripathy, S. . (2003). Studies on mechanical performance of biofibre/glass reinforced polyester hybrid composites. Composites Science and Technology, 63(10), 1377–1385. http://doi.org/10.1016/S0266-3538(03)00084-8 153   Morinière, F. D., Alderliesten, R. C., Sadighi, M., & Benedictus, R. (2013). An integrated study on the low-velocity impact response of the GLARE fibre-metal laminate. Composite Structures, 100, 89–103. http://doi.org/10.1016/j.compstruct.2012.12.016 Morsing, N. (2000). Densification of wood: The influence of hygrothermal treatment on compression of beech perpendicular to the grain. Technical University of Denmark. Mosiewicki, M. A., Marcovich, N. E., & Aranguren, M. I. (2011). Creep Behavior of Wood Flour Composites Made from Linseed Oil-Based Polyester Thermosets. Journal of Applied Polymer Science, 121, 2626–2633. http://doi.org/10.1002/app Moura, L. De, & Hernandez, R. (2006). Evaluation of varnish coating performance for three surfacing methods on sugar maple wood. Forest Products Journal, 56(11-12), 130–136. Retrieved from http://sbo.sbf.ulaval.ca/hernandezr/ArtideMoura4.pdf Mwaikambo, L., & Bisanda, T, N. E. (1999). The performance of cotton-kapok fabric-polyester composites. Journal of Polymer Testing, 18, 181–198. Nagel, G. M., & Thambiratnam, D. P. (2004). A Numerical Study on the Impact Response and Energy Absorption of Tapered Thin-Walled Tubes. International Journal of Mechanical Sciences, 46(2), 201–216. http://doi.org/10.1016/j.ijmecsci.2004.03.006 Navi, P., & Girardet, F. (2000). Effects of thermo-hydro-mechanical treatment on the structure and properties of wood. Holzforschung, 54, 287–293. Navi, P., & Heger, F. (2004). C ombined densification and mechanical processing of wood. MRS Bulletin, 29(5), 332–336. Navi, P., & Sandberg, D. (2012). Thermo-hydro mechanical wood processing. CRC Press. Ndiaye, D., & Tidjani, a. (2012). Effects of coupling agents on thermal behavior and mechanical properties of wood flour/polypropylene composites. Journal of Composite Materials, 3067. http://doi.org/10.1177/0021998311435675 Nourbakhsh, a., Ashori, a., Ziaei Tabari, H., & Rezaei, F. (2010). Mechanical and thermo-chemical properties of wood-flour/polypropylene blends. Polymer Bulletin, 65(7), 691–700. http://doi.org/10.1007/s00289-010-0288-8 Nourbakhsh, A., Ashori, A., Ziaei Tabari, H., & Rezaei, F. (2010). Mechanical and thermo-chemical properties of wood-flour/polypropylene blends. Polymer Bulletin, 65(7), 691–700. http://doi.org/10.1007/s00289-010-0288-8 Nouwezem, R., Borredon, M. E., & Gaset, A. (1994). PRODUCTION A N D PROPERTIES OF THERMOSETTING COMPOSITES BASED ON WOOD A N D UNSATURATED- POLYESTER MIXTURES. Bioresurce Technology, 48, 17–19. 154   O’Dell, J. L. O. (1997). Natural Fibers in Resin Transfer Molded Composites. Fourth International Conference on Woodfiber-Plastic Composites, Forest Product Society, Madison, WI, 280–285. Panthapulakkal, S., Zereshkian, a., & Sain, M. (2006). Preparation and characterization of wheat straw fibers for reinforcing application in injection molded thermoplastic composites. Bioresource Technology. http://doi.org/10.1016/j.biortech.2005.02.043 Pothan, L. ., Thomas, S., & Neelakantan, N. . (1997). Short banana fiber reinforced polyester composites: mechanical, failure and aging characteristics. Journal of Reinforced Plastics and Composites, 16(8), 744–765. http://doi.org/10.1177/073168449701600806 Pothan, L. a., George, C. N., John, M. J., & Thomas, S. (2010). Dynamic Mechanical and Dielectric Behavior of Banana-Glass Hybrid Fiber Reinforced Polyester Composites. Journal of Reinforced Plastics and Composites, 29(8), 1131–1145. http://doi.org/10.1177/0731684409103075 Pothan, L. A., Oommen, Z., & Thomas, S. (2003). Dynamic mechanical analysis of banana fiber reinforced polyester composites. Composites Science and Technology, 63, 283–293. Prasad, A. V. R., Rao, K. M., & Nagasrinivasulu, G. (2009). Mechanical properties of banana empty fruit bunch fibre reinforced polyester composites, 34(June), 162–167. Prasad, S. ., Pavithran, C., & Rohatgi, P. (1983). Alkali treatment of coir fibres for coir-polyester composites. Journal O Materials Science, 18, 1443–1454. Ramakrishna, S., & Hull, D. (1993). Energy absorption capability of epoxy composite tubes with knitted carbon fibre fabric reinforcement. Composites Science and Technology, 49(4), 349–356. http://doi.org/10.1016/0266-3538(93)90066-P Ramesh, M., Palanikumar, K., & Reddy, K. H. (2013). Mechanical property evaluation of sisal–jute–glass fiber reinforced polyester composites. Composites Part B: Engineering, 48, 1–9. http://doi.org/10.1016/j.compositesb.2012.12.004 Rautkari, L., Kutnar, A., Hughes, M., & Kamke, F. A. (2010). Wood surface densification using different methods. In 11th World Conference on Timber Engineering. Trentino, Italy. Razi, P. S., & Raman, A. (2000). Studies on impact fracture properties of wood-polymer composites. Journal of Composite Materials, 34(12), 980–997. Reis, P. N. B., Ferreira, J. a M., Antunes, F. V., & Costa, J. D. M. (2007). Flexural behaviour of hybrid laminated composites. Composites Part A: Applied Science and Manufacturing, 38(6), 1612–1620. http://doi.org/10.1016/j.compositesa.2006.11.010 155   Rodriguez, E., Petrucci, R., Puglia, D., Kenny, J. M., & Vazquez, A. (2005). Characterization of Composites Based on Natural and Glass Fibers Obtained by Vacuum Infusion. Journal of Composite Materials, 39(3), 265–282. http://doi.org/10.1177/0021998305046450 Roe, P. J., & Ansell, M. . (1985). Jute-reinforced polyester composites. Journal O Materials Science, 20, 4015–4020. Rouison, D., Couturier, M., Sain, M., MacMillan, B., & Balcom, B. J. (2005). Water absorption of hemp fiber/unsaturated polyester composites. Polymer Composites, 26(4), 509–525. http://doi.org/10.1002/pc.20114 Rout, J., Misra, M., Tripathy, S. S., Nayak, S. K., & Mohanty, A. K. (2001). The influence of fibre treatment on the performance of coir-polyester composites. Journal of Composite Science and Technology, 61, 1303–1310. Rozman, H. D., Musa, L., & Abubakar, a. (2005). The Mechanical and Dimensional Properties of Rice Husk-Unsaturated Polyester Composites. Polymer-Plastics Technology and Engineering, 44(3), 489–500. http://doi.org/10.1081/PTE-200048306 Sabzi, M., Mirabedini, S. M., Zohuriaan-Mehr, J., & Atai, M. (2009). Surface modification of TiO2 nano-particles with silane coupling agent and investigation of its effect on the properties of polyurethane composite coating. Progress in Organic Coatings, 65(2), 222–228. http://doi.org/10.1016/j.porgcoat.2008.11.006 Sanadi, A. ., Prasad, S. ., & Rohatgi, P. . (1986a). Sunhemp fibre-reinforced polyester. Journal of Material Science Letters, 21, 4299–4304. Sanadi, A. R., Prasad, S. V, & Rohatgi, P. K. (1986b). SEM observations on the origins of toughness of natural fibre-polyester composites. Journal of Material Science Letters, 5, 395–396. Sangthong, S., Pongprayoon, T., & Yanumet, N. (2009). Mechanical property improvement of unsaturated polyester composite reinforced with admicellar-treated sisal fibers. Composites Part A: Applied Science and Manufacturing, 40(6-7), 687–694. http://doi.org/10.1016/j.compositesa.2008.12.004 Sèbe, G., Cetin, N. S., Hill, C. A. S., & Hughes, M. (2000). RTM Hemp Fibre-Reinforced Polyester Composites. Applied Composite Materials, 7, 341–349. Seborg, R., Millet, M., & Stamm, A. (1945). Heat-stabilized compressed wood (Staypak). Journal of Mechanical Engineering, 67, 25–31. Seki, Y., Sever, K., Sarıkanat, M., Şen, İ., & Aral, A. (2011). Jute / Polyester Composites : The Effect of Water Aging on The Interlaminar Shear Strength. 6th Internationl Advanced Technologies Symposium, (May), 16–18. 156   Short, D., & Summerscales, J. (1980). Hybrids - a review: Part 2. Physical properties. Composites, 11(1), 33–38. Retrieved from http://dx.doi.org/10.1016/0010-4361(80)90019-1 Shupe, T. F., Hse, C. Y., & Wang, W. H. (2001). An investigation of selected factors that influence hardwood wettability. Holzforschung, 55(5), 541–548. http://doi.org/10.1515/HF.2001.087 Simpson, W. T. (1993). Specific gravity, moisture content, and density relationship for wood. Gen Tech Rep FPLGTR76 Madison WI US Department of Agriculture Forest Service Forest Products Laboratory (Vol. Gen. Tech.). Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.155.4926&amp;rep=rep1&amp;type=pdf Simpson, W. T. (1999). Drying and control of moisture content and dimensional changes. In Wood handbook : wood as an engineering material. (pp. 3–4). Madison, WI: USDA Forest Service, Forest Products Laboratory. Singh, B., Gupta, M., & Verma, A. (1995). mechanical behaviour of particulate hybrid composite laminates as potential building materials. Construction and Building Materials, 9(1), 39–44. Sinn, G., Gindl, M., Reiterer, a., & Stanzl-Tschegg, S. (2004). Changes in the surface properties of wood due to sanding. Holzforschung, 58(3), 246–251. http://doi.org/10.1515/HF.2004.038 Sobczak, L., Lang, R. W., & Haider, A. (2012). Polypropylene composites with natural fibers and wood - General mechanical property profiles. Composites Science and Technology, 72(5), 550–557. http://doi.org/10.1016/j.compscitech.2011.12.013 Sreekumar, P. A., Albert, P., Unnikrishnan, G., Joseph, K., & Thomas, S. (2008). Mechanical and Water Sorption Studies of Ecofriendly Banana Fiber-Reinforced Polyester Composites Fabricated by RTM. Journal of Applied Polymer Science, 109, 1547–1555. http://doi.org/10.1002/app Sreekumar, P. a., Joseph, K., Unnikrishnan, G., & Thomas, S. (2007). A comparative study on mechanical properties of sisal-leaf fibre-reinforced polyester composites prepared by resin transfer and compression moulding techniques. Composites Science and Technology, 67(3-4), 453–461. http://doi.org/10.1016/j.compscitech.2006.08.025 Stalder, A. F., Kulik, G., Sage, D., Barbieri, L., & Hoffmann, P. (2006). A snake-based approach to accurate determination of both contact points and contact angles. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 286(1-3), 92–103. http://doi.org/10.1016/j.colsurfa.2006.03.008 157   Stalder, A. F., Melchior, T., Müller, M., Sage, D., Blu, T., & Unser, M. (2010). Low-bond axisymmetric drop shape analysis for surface tension and contact angle measurements of sessile drops. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 364(1-3), 72–81. http://doi.org/10.1016/j.colsurfa.2010.04.040 Stamboulis, a., Baillie, C. a., Garkhail, S. K., Van Melick, H. G. H., & Peijs, T. (2000). Environmental durability of flax fibres and their composites based on polypropylene matrix. Applied Composite Materials, 7(5-6), 273–294. http://doi.org/10.1023/A:1026581922221 Standfest, G., Kutnar, A., Plank, B., Petutschnigg, A., Kamke, F. a., & Dunky, M. (2013). Microstructure of viscoelastic thermal compressed (VTC) wood using computed microtomography. Wood Science and Technology, 47(1), 121–139. http://doi.org/10.1007/s00226-012-0496-5 Stark, N. ., Cai, Z., & Carll, C. (2010). Wood as an Engineering Material, Chapter 11: Wood-Based Composite Materials-Panel Products- Glued-Laminated Timber, Structural Composite Lumber, and Wood-Nonwood Composite Materials, General Technical Report FPL-GTR-190. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory, 256–257. Stevens, M. J. (1999). Thoughts on the Structure of the Erythrocyte Membrane. Folia Haematologica (Leipzig, Germany : 1928), 15(8), 2773–2778. Strong, B. (2008). Fundamentals of composites manufacturing (Second). Society of Manufacturing Engineers. Sulaiman, O., Hashim, R., Subari, K., & Liang, C. K. (2009). Effect of sanding on surface roughness of rubberwood. Journal of Materials Processing Technology, 209(8), 3949–3955. http://doi.org/10.1016/j.jmatprotec.2008.09.009 Sung, Y., & Hilton, H. H. (1998). Effects of thermo-machanical properties of composites on viscosity, temperature, and degree of cure in thick thermosetting composite laminates during process. Journal of Composite Materials, 32(7), 600–622. Sydenstricker, T. H. ., Mochnaz, S., & Amico, S. C. (2003). Pull-out and other evaluations in sisal-reinforced polyester biocomposites. Polymer Testing, 22(4), 375–380. http://doi.org/10.1016/S0142-9418(02)00116-2 Tabarsa, T., Ashori, A., & Gholamzadeh, M. (2011). Evaluation of surface roughness and mechanical properties of particleboard panels made from bagasse. Composites Part B: Engineering, 42(5), 1330–1335. http://doi.org/10.1016/j.compositesb.2010.12.018 Taher, S. T., Mahdi, E., Mokhtar, a. S., Magid, D. L., Ahmadun, F. R., & Arora, P. R. (2006). A new composite energy absorbing system for aircraft and helicopter. Composite Structures, 75(1-4), 14–23. http://doi.org/10.1016/j.compstruct.2006.04.083 158   Tajvidi, M., & Haghdan, S. (2009). Effects of accelerated freeze-thaw cycling on physical and mechanical properties of wood flour/PVC composites. Journal of Reinforced Plastics and Composites, 28(15), 1841–1846. http://doi.org/10.1177/0731684408090367 Tajvidi, M., Haghdan, S., & Kazemi, N. S. (2007). Physical properties of novel layered composites of wood flour and PVC. Journal of Reinforced Plastics and Composites. http://doi.org/10.1177/0731684407081381 Tajvidi, M., Haghdan, S., & Saeed Kazemi Najafi. (2008). Physical Properties of Novel Layered Composites of Wood Flour and PVC. Journal of Reinforced Plastics and Composites, 27(16-17), 1759–1765. http://doi.org/10.1177/07316844080270160101 Tasdemir, M., Biltekin, H., & Caneba, G. T. (2009). Preparation and characterization of LDPE and PP — wood fiber composites. Journal of Applied Polymer Science, 112(5), 3095–3102. http://doi.org/10.1002/app Tewari, M., Singh, V. K., Gope, P. C., & Chaudhary, A. K. (2012). Evaluation of Mechanical Properties of Bagasse-Glass Fiber Reinforced Composite, 3(1), 171–184. Thwe, M., & Liao, K. (2002). Effects of environmental aging on the mechanical properties of bamboo–glass fiber reinforced polymer matrix hybrid composites. Composites Part A: Applied Science and …, 33, 43–52. Retrieved from http://www.sciencedirect.com/science/article/pii/S1359835X01000719 Thwe, M. M., & Liao, K. (2003). Durability of bamboo-glass fiber reinforced polymer matrix hybrid composites. Composites Science and Technology, 63(3-4), 375–387. http://doi.org/10.1016/S0266-3538(02)00225-7 Todd, R. H., Dell, K. A., & Alting, L. (1994). Manufacturing Processes Reference Guide. Industrial Press Inc. Tungjitpornkull, S., Chaochanchaikul, K., & Sombatsompop, N. (2007). Mechanical characterization of E-chopped strand glass fiber reinforced wood/PVC composites. Journal of Thermoplastic Composite Materials, 20(6), 535–550. http://doi.org/10.1177/0892705707084541 UPM. (2013). Handbook of Finnish Plywood. Valente, M., Sarasini, F., Marra, F., Tirillò, J., & Pulci, G. (2011). Hybrid recycled glass fiber/wood flour thermoplastic composites: Manufacturing and mechanical characterization. Composites Part A: Applied Science and Manufacturing, 42(6), 649–657. http://doi.org/10.1016/j.compositesa.2011.02.004 159   Vargas, M. a., Sachsenheimer, K., & Guthausen, G. (2012). In-situ investigations of the curing of a polyester resin. Polymer Testing, 31(1), 127–135. http://doi.org/10.1016/j.polymertesting.2011.10.004 Wålinder, M. E. P., & Johansson, I. (2001). Measurement of wood wettability by the Wilhelmy method. Part 1. Contamination of probe liquids by extractives. Holzforschung, 55(1), 21–32. http://doi.org/10.1515/HF.2001.005 Wålinder, M. E. P., & Ström, G. (2001). Measurement of wood wettability by the Wilhelmy method. Part 2. Determination of apparent contact angles. Holzforschung, 55(1), 33–41. http://doi.org/10.1515/HF.2001.006 Wambua, P., Ivens, J., & Verpoest, I. (2003). Natural fibres: Can they replace glass in fibre reinforced plastics? Composites Science and Technology, 63(9), 1259–1264. http://doi.org/10.1016/S0266-3538(03)00096-4 Wan, J., Smith, S. T., Asce, F., Qiao, P., & Chen, F. (2014). Experimental Investigation on FRP-to-Timber Bonded Interfaces. Journal of Composites for Construction, 18(3), 1–9. http://doi.org/10.1061/(ASCE)CC.1943-5614.0000418. Wang, W., & Morrell, J. . (2004). Water sorption characteristics of two wood-plastic composites. Forest Products Journal, 54(9736), 209–212. Wolcott, M. P., & Englund, K. (1999). A technology review of wood-plastic composites. In 33rd International Particleboard/Composite Materials Symposium (pp. 103–111). Washington, US. Xiong, C., Qi, R., & Wang, Y. (2009). Wood-thermoplastic composites from wood flour and high-density polyethylene. Journal of Applied Polymer Science, 114(2), 1160–1168. http://doi.org/10.1002/app Zhang, H., Cui, Y., & Zhang, Z. (2013). Chemical Treatment of Wood Fiber and Its Reinforced Unsaturated Polyester Composites. Journal of Vinyl and Additive Technology, 19(1), 18–24. http://doi.org/10.1002/vnl Zhu, W. H., Tobias, B, C., & Coutts, R, S, P. (1995). Banana fibre strands reinforced polyester composites. Journal of Matreial Science Letters, 14, 1–3. 160   Appendices Appendix A  Plot of failure and non-failure using staircase method  Treatment: Woven-undensified-polymer (WOP)  ho =250; N=Nx=9; d=50 h= ho + dh (A/N-0.5)   =250+50(7/9-0.5)   =263.88 mm  Total mass height(mm) Outcome of test (X=failure, O=non-failure) nx no i ni ini 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20      300     x  x  x    x  x  x  x  7 0 1 7 7 250  x  o  o  o  x  o  o  o  o  o 2 8 0 2 0 200 o  o        o          0 3     9 (Nx) 11 (No)  9 (N) 7 (A) MFE= hwf        = (263.88) (0.224) (9.80665×10-3)        = 0.58 J 161   Appendix B  ANOVA results for the effects of reinforcement configuration and densification on impact strength of wood veneer/polyester composites  I. Mean impact energy among treatments  SUMMARY        Groups Count Sum Average Variance   Control 20 2.19669 0.109834 7.81E-34   UDP 20 2.19669 0.109834 8.11E-34   UOP 20 2.19669 0.109834 8.11E-34   G-UOP 20 27.01928 1.350964 0.010286   G-UDP 20 28.77663 1.438832 0.008762   WDP 20 12.30146 0.615073 0.009397   WOP 20 11.64245 0.582123 0.010286   G-WDP 20 31.63233 1.581617 0.009397   G-WOP 20 30.97332 1.548666 0.008762   GP 20 22.84557 1.142279 0.009397   CrossDP 20 11.86212 0.593106 0.011937   CrossOP 20 11.20312 0.560156 0.013841   G-CrossDP 20 30.53399 1.526699 0.008762   G-CrossOP 20 30.09465 1.504732 0.010286   P 20 2.19669 0.109834 8.11E-34   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 101.205 14 7.228931 975.8958 2E-231 1.726605 Within Groups 2.111132 285 0.007407    Total 103.3162 299        162   II. Effects of configurations                       SUMMARY       Groups Count Sum Average Variance   G-UOP 20 27.019282 1.350964 0.010286   G-UDP 20 28.776634 1.438832 0.008762   G-WDP 20 31.63233 1.581617 0.009397   G-WOP 20 30.973323 1.548666 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.667721 3 0.222574 23.92833 5.25E-11 2.724944 Within Groups 0.706928 76 0.009302    Total 1.374649 79     SUMMARY       Groups Count Sum Average Variance   G-UOP 20 27.019282 1.350964 0.010286   G-UDP 20 28.776634 1.438832 0.008762   G-CrossOP 20 30.094648 1.504732 0.010286   G-CrossDP 20 30.533985 1.526699 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.373972 3 0.124657 13.08889 5.61E-07 2.724944 Within Groups 0.723817 76 0.009524    Total 1.097789 79     SUMMARY      Groups Count Sum Average Variance   G-CrossOP 20 30.09465 1.504732 0.010286   G-CrossDP 20 30.53399 1.526699 0.008762   G-WOP 20 30.97332 1.548666 0.008762   G-WDP 20 31.63233 1.581617 0.009397   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.06454033 3 0.021513 2.312856 0.082718 2.724944 Within Groups 0.70692772 76 0.009302    Total 0.77146805 79     163                  SUMMARY       Groups Count Sum Average Variance   CrossOP 20 11.20312 0.560156 0.013841   CrossDP 20 11.86212 0.593106 0.011937   WOP 20 11.64245 0.582123 0.010286   WDP 20 12.30146 0.615073 0.009397   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.031365 3 0.010455 0.919926 0.4354 2.724944 Within Groups 0.863755 76 0.011365    Total 0.89512 79       SUMMARY       Groups Count Sum Average Variance   UOP 20 2.1966896 0.109834 8.11E-34   UDP 20 2.1966896 0.109834 8.11E-34   WDP 20 12.301462 0.615073 0.009397   WOP 20 11.642455 0.582123 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 4.788651 3 1.596217 324.3892 3.23E-43 2.724944 Within Groups 0.373972 76 0.004921    Total 5.162623 79     SUMMARY      Groups Count Sum Average Variance   UOP 20 2.19669 0.109834 8.11E-34   UDP 20 2.19669 0.109834 8.11E-34   CrossOP 20 11.20312 0.560156 0.013841   CrossDP 20 11.86212 0.593106 0.011937   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 4.36883745 3 1.456279 225.9721 9.05E-38 2.724944 Within Groups 0.48978269 76 0.006445    Total 4.85862013 79     164   SUMMARY      Groups Count Sum Average Variance   WDP 20 12.30146 0.615073 0.009397   UDP 20 2.19669 0.109834 8.11E-34   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.552661 1 2.552661 543.2973 4.13E-24 4.098172 Within Groups 0.178541 38 0.004698    Total 2.731202 39            SUMMARY      Groups Count Sum Average Variance   UDP 20 2.19669 0.109834 8.11E-34   CrossDP 20 11.86212 0.593106 0.011937   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.335515 1 2.335515 391.3191 1.32E-21 4.098172 Within Groups 0.226796 38 0.005968    Total 2.562311 39      SUMMARY      Groups Count Sum Average Variance   WOP 20 11.64245 0.582123 0.010286   UOP 20 2.19669 0.109834 8.11E-34   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.230562 1 2.230562 433.716 2.2E-22 4.098172 Within Groups 0.195431 38 0.005143    Total 2.425993 39      SUMMARY      Groups Count Sum Average Variance   WDP 20 12.30146 0.615073 0.009397   CrossDP 20 11.86212 0.593106 0.011937   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.004825 1 0.004825 0.452381 0.505276 4.098172 Within Groups 0.405337 38 0.010667    Total 0.410163 39     165         SUMMARY      Groups Count Sum Average Variance   UOP 20 2.19669 0.109834 8.11E-34   CrossOP 20 11.20312 0.560156 0.013841   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.027893 1 2.027893 293.0183 1.87E-19 4.098172 Within Groups 0.262987 38 0.006921    Total 2.29088 39                   SUMMARY      Groups Count Sum Average Variance   WOP 20 11.64245 0.582123 0.010286   CrossOP 20 11.20312 0.560156 0.013841   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.004825 1 0.004825 0.4 0.530877 4.098172 Within Groups 0.458417 38 0.012064    Total 0.463243 39     SUMMARY      Groups Count Sum Average Variance   G-WOP 20 30.97332 1.548666 0.008762   G-UOP 20 27.01928 1.350964 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.390861 1 0.390861 41.04 1.58E-07 4.098172 Within Groups 0.361908 38 0.009524    Total 0.752769 39     SUMMARY      Groups Count Sum Average Variance   G-UOP 20 27.01928 1.350964 0.010286   G-CrossOP 20 30.09465 1.504732 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.236447 1 0.236447 22.98765 2.52E-05 4.098172 Within Groups 0.390861 38 0.010286    Total       0.627308 39     166                     SUMMARY      Groups Count Sum Average Variance   G-WDP 20 31.63233 1.581617 0.009397   G-CrossDP 20 30.53399 1.526699 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.030159 1 0.030159 3.321678 0.076248 4.098172 Within Groups 0.345019 38 0.009079    Total 0.375178 39     SUMMARY      Groups Count Sum Average Variance   G-WOP 20 30.97332 1.548666 0.008762   G-CrossOP 20 30.09465 1.504732 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.019302 1 0.019302 2.026667 0.162719 4.098172 Within Groups 0.361908 38 0.009524    Total 0.38121 39     SUMMARY      Groups Count Sum Average Variance   G-WDP 20 31.63233 1.581617 0.009397   G-UDP 20 28.77663 1.438832 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.203875 1 0.203875 22.45455 2.99E-05 4.098172 Within Groups 0.345019 38 0.009079    Total 0.548894 39         SUMMARY      Groups Count Sum Average Variance   G-UDP 20 28.77663 1.438832 0.008762   G-CrossDP 20 30.53399 1.526699 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.077207 1 0.077207 8.811594 0.005159 4.098172 Within Groups 0.332956 38 0.008762    Total 0.410163 39     167   III. Effects of densification            SUMMARY      Groups Count Sum Average Variance   G-UDP 20 28.77663 1.438832 0.008762   G-UOP 20 27.01928 1.350964 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.077207 1 0.077207 8.106667 0.007076 4.098172 Within Groups 0.361908 38 0.009524    Total 0.439116 39     SUMMARY       Groups Count Sum Average Variance   UDP 20 2.1966896 0.10983448 8.11E-34   UOP 20 2.1966896 0.10983448 8.11E-34   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0 1 0 0 1 4.098172 Within Groups 3.08149E-32 38 8.10918E-34    Total 3.08149E-32 39     SUMMARY       Groups Count Sum Average Variance   WDP 20 12.3014618 0.615073088 0.009397   WOP 20 11.6424549 0.582122744 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.010857252 1 0.010857252 1.103226 0.300193 4.098172 Within Groups 0.373972003 38 0.009841368    Total 0.384829255 39     SUMMARY       Groups Count Sum Average Variance   CrossDP 20 11.86212 0.593106 0.011937   CrossOP 20 11.20312 0.560156 0.013841   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.010857 1 0.010857 0.842365 0.36451 4.098172 Within Groups 0.489783 38 0.012889    Total 0.50064 39     168   SUMMARY      Groups Count Sum Average Variance   G-WDP 20 31.63233 1.581617 0.009397   G-WOP 20 30.97332 1.548666 0.008762   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.010857 1 0.010857 1.195804 0.281046 4.098172 Within Groups 0.345019 38 0.009079    Total 0.355877 39             IV. Effects of reinforcement type SUMMARY      Groups Count Sum Average Variance   Wood composites 20 2.19669 0.109834 7.81E-34   20 2.19669 0.109834 8.11E-34   20 2.19669 0.109834 8.11E-34   20 12.30146 0.615073 0.009397   20 11.64245 0.582123 0.010286   20 11.86212 0.593106 0.011937   20 11.20312 0.560156 0.013841   Glassfiber composites  20 27.01928 1.350964 0.010286   20 28.77663 1.438832 0.008762   20 31.63233 1.581617 0.009397   20 30.97332 1.548666 0.008762   20 22.84557 1.142279 0.009397   20 30.53399 1.526699 0.008762   20 30.09465 1.504732 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 89.18129 13 6.860099 864.3639 4.4E-209 1.757085 Within Groups 2.111132 266 0.007937    Total 91.29243 279     SUMMARY       Groups Count Sum Average Variance   G-CrossDP 20 30.53399 1.526699 0.008762   G-CrossOP 20 30.09465 1.504732 0.010286   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.004825 1 0.004825 0.506667 0.480933 4.098172 Within Groups 0.361908 38 0.009524    Total 0.366734 39     169   Appendix C  ANOVA results for the wettability and impact performance of wood veneer/polyester composites  I. Impact energy among treatments   SUMMARY       Groups Count Sum Average Variance   DF-UP 8 7.688414 0.961052 0.006032   DF-CP 8 10.87361 1.359202 0.003231   DF-BCP 8 10.98345 1.372931 0.003447   GP 8 11.42279 1.427848 0.006893   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.108 3 0.369322 75.35897 1.6E-13 2.946685 Within Groups 0.1372 28 0.004901    Total 1.2452 31      SUMMARY      Groups Count Sum Average Variance   DF-UP 8 7.688414 0.961052 0.006032   DF-CP 8 10.87361 1.359202 0.003231   DF-BCP 8 10.98345 1.372931 0.003447   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.875617 2 0.437809 103.339 1.35E-11 3.4668 Within Groups 0.088969 21 0.004237    Total 0.964586 23      SUMMARY      Groups Count Sum Average Variance   DF-UP 8 7.688414 0.961052 0.006032   DF-CP 8 10.87361 1.359202 0.003231   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.634094 1 0.634094 136.907 1.29E-08 4.60011 Within Groups 0.064842 14 0.004632    Total 0.698936 15      170   SUMMARY      Groups Count Sum Average Variance   DF-UP 8 7.688414 0.961052 0.006032   DF-BCP 8 10.98345 1.372931 0.003447   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.678578 1 0.678578 143.1818 9.7E-09 4.60011 Within Groups 0.06635 14 0.004739    Total 0.744928 15      SUMMARY      Groups Count Sum Average Variance   DF-CP 8 10.87361 1.359202 0.003231   DF-BCP 8 10.98345 1.372931 0.003447   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.000754 1 0.000754 0.225806 0.641979 4.60011 Within Groups 0.046747 14 0.003339    Total 0.0475 15      SUMMARY      Groups Count Sum Average Variance   DF-CP 8 10.87361 1.359202 0.003231   DF-BCP 8 10.98345 1.372931 0.003447   GP 8 11.42279 1.427848 0.006893   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.021111 2 0.010556 2.333333 0.121597 3.4668 Within Groups 0.095001 21 0.004524    Total 0.116112 23       II. Wettability of wood veneers  - DF Control 171         SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 4 121.283 30.32075 0.161035   Right CA 4 118.789 29.69725 0.564038   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.777504 1 0.777504 2.144624 0.193411 5.987378 Within Groups 2.17522 6 0.362537    Total 2.952724 7      SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 4 94.903 23.72575 2.974732   Right CA 4 95.655 23.91375 4.201622   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.070688 1 0.070688 0.0197 0.892971 5.987378 Within Groups 21.52906 6 3.588177    Total 21.59975 7      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 4 79.57 19.8925 0.466998   Right CA 4 80.406 20.1015 0.260196   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.087362 1 0.087362 0.240271 0.641417 5.987378 Within Groups 2.181584 6 0.363597    Total 2.268946 7       SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 4 492.825 123.2063 0.148967   Right CA 4 491.792 122.948 1.972689   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.133386 1 0.133386 0.125738 0.735023 5.987378 Within Groups 6.364967 6 1.060828    Total 6.498353 7     172   SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 4 71.769 17.94225 0.119901   Right CA 4 73.46 18.365 0.042279   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.357435 1 0.357435 4.407893 0.080534 5.987378 Within Groups 0.486539 6 0.08109    Total 0.843974 7               - DF 120  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 5 613.9 122.78 1.020171   Right CA 5 615.067 123.0134 1.500902   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.136189 1 0.136189 0.10804 0.750825 5.317655 Within Groups 10.08429 8 1.260536    Total 10.22048 9      SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 5 154.018 30.8036 1.286497   Right CA 5 155.19 31.038 9.411082   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.137358 1 0.137358 0.02568 0.876656 5.317655 Within Groups 42.79031 8 5.348789    Total 42.92767 9     SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 4 57.091 14.27275 0.129181   Right CA 4 55.295 13.82375 0.311094   ANOVA      F crit Source of Variation SS df MS F P-value 5.987378 Between Groups 0.403202 1 0.403202 1.831591 0.224701  Within Groups 1.320826 6 0.220138    Total 1.724028 7     173   SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 5 117.4 23.48 2.533014   Right CA 5 118.903 23.7806 3.239861   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.225901 1 0.225901 0.078263 0.786762 5.317655 Within Groups 23.0915 8 2.886437    Total 23.3174 9      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 5 98.076 19.6152 0.734725   Right CA 5 98.681 19.7362 0.862368   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.036602 1 0.036602 0.045836 0.83583 5.317655 Within Groups 6.388372 8 0.798546    Total 6.424974 9       SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 5 88.105 17.621 0.605934   Right CA 5 90.109 18.0218 0.62064   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.401602 1 0.401602 0.654835 0.44179 5.317655 Within Groups 4.906295 8 0.613287    Total 5.307896 9        SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 5 71.506 14.3012 0.100933   Right CA 5 68.56 13.712 0.295761   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.867892 1 0.867892 4.375626 0.069822 5.317655 Within Groups 1.586775 8 0.198347    Total 2.454666 9     174   - DF 320 SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 3 390.038 130.0127 6.641854   Right CA 3 387.973 129.3243 29.60182   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.710704 1 0.710704 0.039218 0.852674 7.708647 Within Groups 72.48734 4 18.12184    Total 73.19805 5      SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 3 153.288 51.096 19.781   Right CA 3 154.231 51.41033 27.59029   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.148208 1 0.148208 0.006257 0.94075 7.708647 Within Groups 94.74258 4 23.68564    Total 94.89079 5      SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 3 117.783 39.261 21.93588   Right CA 3 121.652 40.55067 22.19317   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.49486 1 2.49486 0.113071 0.753574 7.708647 Within Groups 88.2581 4 22.06453    Total 90.75296 5      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 2 56.754 28.377 0.023328   Right CA 2 57.72 28.86 0.025992   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.233289 1 0.233289 9.460219 0.091439 18.51282 Within Groups 0.04932 2 0.02466    Total 0.282609 3     175   SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 3 85.294 28.43133 0.509525   Right CA 3 86.447 28.81567 0.668858   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.221568 1 0.221568 0.376054 0.572877 7.708647 Within Groups 2.356767 4 0.589192    Total 2.578336 5      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 3 74.322 24.774 6.198427   Right CA 3 75.816 25.272 4.599244   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.372006 1 0.372006 0.068905 0.805903 7.708647 Within Groups 21.59534 4 5.398836    Total 21.96735 5      - Maple Control  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 2 255.742 127.871 2.411208   Right CA 2 249.67 124.835 0.426888   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 9.217296 1 9.217296 6.495408 0.125598 18.51282 Within Groups 2.838096 2 1.419048    Total 12.05539 3      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 2 56.754 28.377 0.023328   Right CA 2 57.72 28.86 0.025992   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.233289 1 0.233289 9.460219 0.091439 18.51282 Within Groups 0.04932 2 0.02466    Total 0.282609 3     176   SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 2 54.425 27.2125 0.489061   Right CA 2 55.004 27.502 0.276768   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.08381 1 0.08381 0.218875 0.685926 18.51282 Within Groups 0.765829 2 0.382914    Total 0.849639 3      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 2 49.583 24.7915 0.041761   Right CA 2 48.063 24.0315 0.160745   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.5776 1 0.5776 5.704551 0.139527 18.51282 Within Groups 0.202505 2 0.101253    Total 0.780105 3       - Maple 120  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 5 596.958 119.3916 2.604213   Right CA 5 583.513 116.7026 5.157319   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 18.0768 1 18.0768 4.65805 0.062961 5.317655 Within Groups 31.04613 8 3.880766    Total 49.12293 9       SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 5 180.927 36.1854 4.212465   Right CA 5 183.316 36.6632 3.442601   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.570732 1 0.570732 0.149112 0.709452 5.317655 Within Groups 30.62026 8 3.827533    Total 31.191 9     177   SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 5 134.005 26.801 3.070429   Right CA 5 136.604 27.3208 2.427305   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.67548 1 0.67548 0.24573 0.633427 5.317655 Within Groups 21.99093 8 2.748867    Total 22.66641 9      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 5 118.089 23.6178 0.315009   Right CA 5 119.533 23.9066 0.273326   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.208514 1 0.208514 0.708825 0.424285 5.317655 Within Groups 2.353342 8 0.294168    Total 2.561856 9       SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 5 109.39 21.878 1.334559   Right CA 5 110.971 22.1942 0.84416   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.249956 1 0.249956 0.229452 0.644749 5.317655 Within Groups 8.714873 8 1.089359    Total 8.964829 9      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 5 86.104 17.2208 1.102471   Right CA 5 83.266 16.6532 0.705995   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.805424 1 0.805424 0.890727 0.372906 5.317655 Within Groups 7.233864 8 0.904233    Total 8.039288 9     178   - Maple 320  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 5 620.446 124.0892 2.619067   Right CA 5 616.979 123.3958 6.394954   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.202009 1 1.202009 0.266698 0.619524 5.317655 Within Groups 36.05608 8 4.50701    Total 37.25809 9      SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 5 221.214 44.2428 31.75066   Right CA 5 225.593 45.1186 26.74683   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.917564 1 1.917564 0.065561 0.804375 5.317655 Within Groups 233.99 8 29.24874    Total 235.9075 9      SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 4 151.513 37.87825 9.603138   Right CA 4 153.815 38.45375 6.717136   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.6624 1 0.6624 0.081175 0.78529 5.987378 Within Groups 48.96082 6 8.160137    Total 49.62322 7      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 4 134.089 33.52225 16.01381   Right CA 4 137.871 34.46775 12.06551   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.787941 1 1.787941 0.127349 0.733411 5.987378 Within Groups 84.23796 6 14.03966    Total 86.0259 7     179   SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 5 155.678 31.1356 12.72038   Right CA 5 155.128 31.0256 10.88662   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.03025 1 0.03025 0.002563 0.960866 5.317655 Within Groups 94.42798 8 11.8035    Total 94.45823 9      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 5 142.45 28.49 7.756394   Right CA 5 141.378 28.2756 5.132751   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.114918 1 0.114918 0.017832 0.897068 5.317655 Within Groups 51.55658 8 6.444572    Total 51.6715 9      - Oak Control  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 4 537.604 134.401 32.68161   Right CA 4 546.621 136.6553 11.7909   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 10.16329 1 10.16329 0.457059 0.52419 5.987378 Within Groups 133.4175 6 22.23626    Total 143.5808 7      SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 4 157.027 39.25675 3.141635   Right CA 4 160.288 40.072 1.898677   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.329265 1 1.329265 0.527454 0.495023 5.987378 Within Groups 15.12093 6 2.520156    Total 16.4502 7     180   SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 4 128.955 32.23875 24.80613   Right CA 4 128.487 32.12175 16.99515   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.027378 1 0.027378 0.00131 0.972303 5.987378 Within Groups 125.4038 6 20.90064    Total 125.4312 7      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 4 111.417 27.85425 6.853511   Right CA 4 111.508 27.877 3.43206   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.001035 1 0.001035 0.000201 0.989141 5.987378 Within Groups 30.85671 6 5.142785    Total 30.85775 7      SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 4 104.534 26.1335 2.124187   Right CA 4 104.935 26.23375 0.785962   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.0201 1 0.0201 0.013814 0.910274 5.987378 Within Groups 8.730446 6 1.455074    Total 8.750546 7      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 4 92.342 23.0855 1.041708   Right CA 4 91.528 22.882 0.166303   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.082825 1 0.082825 0.137125 0.723876 5.987378 Within Groups 3.624035 6 0.604006    Total 3.70686 7      181   - Oak 120  SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 3 336.58 112.1933 101.7375   Right CA 3 342.239 114.0797 55.39699   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 5.33738 1 5.33738 0.067934 0.807237 7.708647 Within Groups 314.269 4 78.56725    Total 319.6064 5       SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 4 130.469 32.61725 6.204348   Right CA 4 137.12 34.28 3.029798   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 5.529475 1 5.529475 1.197615 0.315774 5.987378 Within Groups 27.70244 6 4.617073    Total 33.23191 7      SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 4 101.466 25.3665 12.80858   Right CA 4 103.388 25.847 5.800205   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.461761 1 0.461761 0.049628 0.831102 5.987378 Within Groups 55.82635 6 9.304392    Total 56.28811 7      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 4 87.248 21.812 15.30497   Right CA 4 90.137 22.53425 13.93977   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.04329 1 1.04329 0.071349 0.798327 5.987378 Within Groups 87.73421 6 14.62237    Total 88.7775 7     182   SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 3 58.496 19.49867 10.85149   Right CA 3 59.584 19.86133 10.44879   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.197291 1 0.197291 0.018525 0.898313 7.708647 Within Groups 42.60056 4 10.65014    Total 42.79785 5      SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 4 64.944 16.236 3.138615   Right CA 4 64.958 16.2395 0.872235   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2.45E-05 1 2.45E-05 1.22E-05 0.997325 5.987378 Within Groups 12.03255 6 2.005425    Total 12.03258 7      - Oak 320 SUMMARY(Initial CA)      Groups Count Sum Average Variance   Left CA 4 503.743 125.9358 27.61999   Right CA 4 513.083 128.2708 20.86483   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 10.90445 1 10.90445 0.449809 0.527384 5.987378 Within Groups 145.4545 6 24.24241    Total 156.3589 7      SUMMARY(1s)      Groups Count Sum Average Variance   Left CA 4 166.224 41.556 49.83167   Right CA 4 181.734 45.4335 30.59592   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 30.07001 1 30.07001 0.747754 0.420411 5.987378 Within Groups 241.2828 6 40.2138    Total 271.3528 7     183   SUMMARY(2s)      Groups Count Sum Average Variance   Left CA 4 125.599 31.39975 34.14899   Right CA 4 144.483 36.12075 15.34795   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 44.57568 1 44.57568 1.801149 0.228127 5.987378 Within Groups 148.4908 6 24.74847    Total 193.0665 7      SUMMARY(3s)      Groups Count Sum Average Variance   Left CA 4 108.832 27.208 16.18413   Right CA 4 124.63 31.1575 10.8064   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 31.1971 1 31.1971 2.311708 0.179217 5.987378 Within Groups 80.97158 6 13.49526    Total 112.1687 7      SUMMARY(4s)      Groups Count Sum Average Variance   Left CA 4 104.256 26.064 10.67982   Right CA 4 116.29 29.0725 12.85806   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 18.10214 1 18.10214 1.538129 0.261195 5.987378 Within Groups 70.61364 6 11.76894    Total 88.71579 7       SUMMARY(5s)      Groups Count Sum Average Variance   Left CA 4 93.66 23.415 7.003858   Right CA 4 105.175 26.29375 17.30388   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 16.5744 1 16.5744 1.363714 0.287196 5.987378 Within Groups 72.92323 6 12.15387    Total 89.49763 7      184   - Differences of the CA among treatments  SUMMARY (Initial CA)     Groups Count Sum Average Variance   DF Control 5 614.5715 122.9143 0.685678   Maple Control 5 629.54 125.908 2.780347   Oak Control 5 675.5245 135.1049 12.02682   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 403.5932 2 201.7966 39.07545 5.56E-06 3.885294 Within Groups 61.97138 12 5.164282    Total 465.5646 14       SUMMARY(Initial CA)      Groups Count Sum Average Variance    DF 120 5 614.4835 122.8967 0.715884    Maple 120 5 590.2355 118.0471 3.149081    Oak 120 5 564.8055 112.9611 37.95947    ANOVA        Source of Variation SS df MS F P-value  F crit Between Groups 246.8369 2 123.4185 8.85261 0.004346  3.885294 Within Groups 167.2977 12 13.94148     Total 414.1347 14                  SUMMARY(Initial CA)      Groups Count Sum Average Variance   DF 320 5 643.6735 128.7347 8.207047   Maple 320 5 618.7125 123.7425 2.5227   Oak 320 5 631.659 126.3318 18.89165   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 62.33411 2 31.16705 3.156542 0.079162 3.885294 Within Groups 118.4856 12 9.873797    Total 180.8197 14     185    SUMMARY(After 5s)      Groups Count Sum Average Variance   DF Control 5 69.652 13.9304 0.097787   Maple Control 5 121.977 24.3954 0.050743   Oak Control 5 114.677 22.9354 0.328118   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 321.2298 2 160.6149 1010.901 4.22E-14 3.885294 Within Groups 1.906595 12 0.158883    Total 323.1363 14      SUMMARY(After 5s)      Groups Count Sum Average Variance   DF 120 5 70.033 14.0066 0.037018   Maple 120 5 84.685 16.937 0.646961   Oak 120 5 80.596 16.1192 1.399343   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 22.8652 2 11.4326 16.46304 0.000363 3.885294 Within Groups 8.333287 12 0.694441    Total 31.19849 14      SUMMARY(After 5s)      Groups Count Sum Average Variance   DF 320 5 124.179 24.8358 3.264037   Maple 320 5 141.914 28.3828 6.361006   Oak 320 5 124.0745 24.8149 4.002638   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 42.18593 2 21.09296 4.643408 0.032094 3.885294 Within Groups 54.51073 12 4.542561    Total 96.69665 14         186   - Effect of grit size  SUMMARY(Grit size)      Groups Count Sum Average Variance   DF Control 5 614.5715 122.9143 0.685678   DF 120 5 614.4835 122.8967 0.715884   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.000774 1 0.000774 0.001105 0.974296 5.317655 Within Groups 5.606249 8 0.700781    Total 5.607023 9      SUMMARY(Grit size)      Groups Count Sum Average Variance   Maple Control 5 629.54 125.908 2.780347   Maple 120 5 590.2355 118.0471 3.149081   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 154.4844 1 154.4844 52.10768 9.08E-05 5.317655 Within Groups 23.71771 8 2.964714    Total 178.2021 9      SUMMARY(Grit size)      Groups Count Sum Average Variance   Oak Control 4 542.1125 135.5281 14.84163   Oak 120 4 449.6635 112.4159 48.63082   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1068.352 1 1068.352 33.66349 0.001149 5.987378 Within Groups 190.4174 6 31.73623    Total 1258.77 7          187   SUMMARY(Grit size)      Groups Count Sum Average Variance   DF 120 5 614.4835 122.8967 0.715884   DF 320 5 643.6735 128.7347 8.207047   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 85.20561 1 85.20561 19.09812 0.00238 5.317655 Within Groups 35.69172 8 4.461466    Total 120.8973 9       SUMMARY(Grit size)      Groups Count Sum Average Variance   Maple 120 5 590.2355 118.0471 3.149081   Maple 320 5 618.7125 123.7425 2.5227   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 81.09395 1 81.09395 28.59558 0.000688 5.317655 Within Groups 22.68713 8 2.835891    Total 103.7811 9        SUMMARY(Grit size)      Groups Count Sum Average Variance   Oak 120 4 449.6635 112.4159 48.63082   Oak 320 4 508.413 127.1033 21.22129   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 431.438 1 431.438 12.3529 0.0126 5.987378 Within Groups 209.5563 6 34.92606    Total 640.9943 7          188   III. Surface roughness of wood veneers  SUMMARY(Roughness)      Groups Count Sum Average Variance   DF Control 8 62.987 7.873375 0.213389   DF 120 8 55.188 6.8985 0.558598   DF 320 8 38.843 4.855375 0.245361   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 37.95484 2 18.97742 55.96143 3.85E-09 3.4668 Within Groups 7.121438 21 0.339116    Total 45.07628 23      SUMMARY(Roughness)      Groups Count Sum Average Variance   Maple Control 8 49.092 6.1365 0.528936   Maple 120 8 40.975 5.121875 0.271338   Maple 320 8 33.846 4.23075 0.099254   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 14.54787 2 7.273934 24.25917 3.48E-06 3.4668 Within Groups 6.296696 21 0.299843    Total 20.84456 23          SUMMARY(Roughness)      Groups Count Sum Average Variance   Oak Control 8 55.76 6.97 0.263347   Oak 120 8 34.264 4.283 0.407576   Oak 320 8 20.125 2.515625 0.055199   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 80.49344 2 40.24672 166.2809 1.33E-13 3.4668 Within Groups 5.082852 21 0.242041    Total 85.57629 23       189   SUMMARY(Roughness)      Groups Count Sum Average Variance   DF Control 8 62.987 7.873375 0.213389   Maple Control 8 49.092 6.1365 0.528936   Oak Control 8 55.76 6.97 0.263347   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 12.07345 2 6.036725 18.00802 2.79E-05 3.4668 Within Groups 7.03971 21 0.335224    Total 19.11316 23      SUMMARY(Roughness)      Groups Count Sum Average Variance   DF 120 8 55.188 6.8985 0.558598   Maple 120 8 40.975 5.121875 0.271338   Oak 120 8 34.264 4.283 0.407576   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 28.53586 2 14.26793 34.5886 2.26E-07 3.4668 Within Groups 8.662581 21 0.412504    Total 37.19844 23      SUMMARY(Roughness)      Groups Count Sum Average Variance   DF 320 8 38.843 4.855375 0.245361   Maple 320 8 33.846 4.23075 0.099254   Oak 320 8 20.125 2.515625 0.055199   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 23.48331 2 11.74165 88.10346 6.12E-11 3.4668 Within Groups 2.798695 21 0.133271    Total 26.282 23           190   Appendix D  Stress-strain curve in 8SB and 12NS samples                       024681012141618200 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2Stress (MPa)Strain (mm/mm)8SB212NS1191   Appendix E  ANOVA results for the fracture mechanisms of wood/polyester laminates under quasi-static compression and shear loading  SUMMARY(Ec)      Groups Count Sum Average Variance   GP 8 14647.78 1830.973 296187.5   8SA 8 3615.342 451.9177 25564.46   8SB 8 2352.124 294.0155 1212.849   8NS 8 4740.717 592.5896 31617.23   12SA 8 1817 227.125 2831.839   12SB 8 2132 266.5 3108.286   12NS 8 1623 202.875 647.5536   11PW 8 1845 230.625 3550.554   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 16907726 7 2415389 52.98064 2.07E-22 2.178156 Within Groups 2553042 56 45590.04    Total 19460768 63      SUMMARY(Ec)      Groups Count Sum Average Variance   8SA 8 3615.342 451.9177 25564.46   8SB 8 2352.124 294.0155 1212.849   8NS 8 4740.717 592.5896 31617.23   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 356981.8 2 178490.9 9.169912 0.001373 3.4668 Within Groups 408761.7 21 19464.84    Total 765743.6 23      SUMMARY(Ec)      Groups Count Sum Average Variance   8SA 8 3615.342 451.9177 25564.46   8SB 8 2352.124 294.0155 1212.849   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 99732.46 1 99732.46 7.449029 0.016294 4.60011 Within Groups 187441.1 14 13388.65    Total 287173.6 15     192   SUMMARY(Wf)      Groups Count Sum Average Variance   8SA 8 6.129 0.766125 0.075183   8SB 8 8.975 1.121875 0.074978   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.506232 1 0.506232 6.74256 0.021115 4.60011 Within Groups 1.051122 14 0.07508    Total 1.557354 15      SUMMARY(Wf)      Groups Count Sum Average Variance   12SA 8 12.937 1.617125 0.015152   12SB 8 16.887 2.110875 0.021906   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 0.975156 1 0.975156 52.62911 4.19E-06 4.60011 Within Groups 0.259404 14 0.018529    Total 1.23456 15      SUMMARY(Ec)      Groups Count Sum Average Variance   12SA 8 1817 227.125 2831.839   12NS 8 1623 202.875 647.5536   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 2352.25 1 2352.25 1.352104 0.264349 4.60011 Within Groups 24355.75 14 1739.696    Total 26708 15      SUMMARY(Ec)      Groups Count Sum Average Variance   8SA 8 3615.342 451.9177 25564.46   12SA 8 1817 227.125 2831.839   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 202127 1 202127 14.23616 0.002057 4.60011 Within Groups 198774.1 14 14198.15    Total 400901.1 15      193   SUMMARY(Df)      Groups Count Sum Average Variance   8SA 8 5.92868 0.741085 0.011071   8SB 8 8.48464 1.06058 0.016872   8NS 8 6.46771 0.808464 0.046522   12SA 8 8.19937 1.024921 0.096922   12SB 8 9.79706 1.224633 0.00815   12NS 8 8.87723 1.109654 0.008442   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1.362729 5 0.272546 8.69927 1.01E-05 2.437693 Within Groups 1.315849 42 0.03133    Total 2.678578 47      SUMMARY(Wf)      Groups Count Sum Average Variance   8SB 8 8.975 1.121875 0.074978   12SB 8 16.887 2.110875 0.021906   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 3.912484 1 3.912484 80.76663 3.45E-07 4.60011 Within Groups 0.678186 14 0.048442    Total 4.59067 15      SUMMARY(Pmax)      Groups Count Sum Average Variance   GP 8 15255.99 1906.999 34104.91   8SA 8 15040.48 1880.06 440386.6   8SB 8 17489.44 2186.18 82844.07   8NS 8 17048.49 2131.062 37934.76   12SA 8 18187.18 2273.397 138183.2   12SB 8 17559.47 2194.934 372217.2   12NS 8 18509.56 2313.695 100489.7   11PW 8 18098.69 2262.337 424734.9   ANOVA       Source of Variation SS df MS F P-value F crit Between Groups 1519336 7 217047.9 1.064681 0.39813 2.178156 Within Groups 11416267 56 203861.9    Total 12935603 63      194   Appendix F  SEM micrographs of fracture mechanisms of 12SB laminates: delamination and transverse crack  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166594/manifest

Comment

Related Items