Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A simulated study on some fundamental aspects in flakeboard manufacture Song, Dongjin 1996

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

Item Metadata

Download

Media
831-ubc_1996-0005.pdf [ 7.91MB ]
Metadata
JSON: 831-1.0075222.json
JSON-LD: 831-1.0075222-ld.json
RDF/XML (Pretty): 831-1.0075222-rdf.xml
RDF/JSON: 831-1.0075222-rdf.json
Turtle: 831-1.0075222-turtle.txt
N-Triples: 831-1.0075222-rdf-ntriples.txt
Original Record: 831-1.0075222-source.json
Full Text
831-1.0075222-fulltext.txt
Citation
831-1.0075222.ris

Full Text

A S I M U L A T E D STUDY ON S O M E F U N D A M E N T A L A S P E C T S IN F L A K E B O A R D M A N U F A C T U R E by D O N G J T N S O N G B.Eng. Beijing Forestry University, 1987 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S Department of Forestry We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A October, 1995 ©Dongj in Song, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) 11 ABSTRACT The localized properties of conventional pressed flakeboards were studied using simulated approaches. There parallel experiments were performed in this study. Firstly, Columns of trembling aspen flakes were assembled and pressed in wood particle mat in order to simulate localized variations in a flakeboard. The flake columns were constructed from 24, 32, 40 microtomed flakes. Powdered phenol formaldehyde resin was applied to all flakes at the loading level of 1%. The flake assemblies were assessed in term of bonding properties and dimensional stability. The thickness swell test were performed using an image analysis system Three variables were studied (1) pressing condition; (2) number of flakes in a flake assembly (density levels); (3) vertical layers in each flake assembly. Secondly, the compression process of the flake columns were also monitored the deformation process of the flake columns were also monitored using stress and strain relationships using a computerized cold press. The deformed wood structures were observed using a scanning electron microscope. The dimensional properties of flakes compressed at different compaction ratios were also tested. Thirdly, the flow properties of the powdered phenol formaldehyde resin was studied using a thermal mechanical analyzer (TMA). A range of pressing forces and heating rates were applied to simulate internal environmental conditions in flakeboard. The results show that density of flake assemblies influences heat transfer, and also positively affects bonding strength at the intermediate and core layers and thickness swell values the face, intermediate and core layers. Outer layers generally have greater bonding strength and thickness swell values than inner layers did. Moisture content affects bonding strength and thickness for higher density flake assemblies at the face and intermediate layers. This results from compression tested show that, at a higher CR, the thickness swelling was mainly contributed to the balloon back of the cell lumens when compressed wood was soaked in water condition. The results from (TMA) studies show that pressure I l l and heating rate affect adhesive flow and greater adhesive flow occurred at the more rapid heating rates. iv T A B L E OF CONTENT ABSTRACT ii T A B L E OF CONTENT iv LIST OF FIGURES vii LIST OF T AB L E S x ABBREVIATIONS xi A C K N O W L E D G M E N T S xii 1. INTRODUCTION 1 2. B A C K G R O U N D 3 2.1. Heat and Moisture Transfer and Density Gradient Formation 3 2.2. Adhesive Flow 6 2.2.1. Adhesive flow and cure 7 2.2.2. Flow of powdered PF resin in flakeboard 8 2.3. Flakeboard Bonding Properties 11 2.4. Dimensional Stability of Flakeboard 13 3. M E T H O D O L O G Y 16 3.1. Materials , 16 3.1.1. Flake preparation... 16 3.1.2. Flake density determination 17 3.1.3. Wood particle preparation 17 3.1.4. Adhesive synthesis 17 3.2. Adhesive Analysis 18 3.2.1. Infrared absorption spectroscopy 18 3.2.2. Gel permeation chromatography 18 3.2.3. Differential scanning calorimetry 19 3.2.4. Thermal mechanical analysis 19 V 3.3. Flake Column Cold Compression 21 3.3.1. Stress and strain determination 21 3.3.1.1. Equipment 21 3.3.1.2. Method and procedure 21 § 3.3.2. Dimensional stability 24 3.3.3. Microscopic examination 25 3.4. Flake Assembly Hot Pressing 25 3.4.1. Equipment 25 3.4.2. Method 25 3.4.3. Procedure 31 3.5. Flake Assembly Properties 31 3.5.1. Density profile determination 31 3.5.2. Tension shear testing 34 3.5.3. Dimensional stability determination 37 3.5.3.1. Instrumentation 37 3.5.3.2. Method 37 3.5.3.3. Procedure 40 3.5.4 Microscopic examination. 40 4. RESULT A N D DISCUSSION 41 4.1. Internal Conditions and Density Profiles of the Flake Assemblies in the Constituted Board 41 4.1.1. Internal conditions of the F As in the constituted mat 41 4.1.2. Density profiles of the FAs in the constituted board 49 4.2. Adhesive Analysis 53 4.2.1. Adhesive properties 53 4.2.2. Flow of powdered adhesive in flakeboard 59 4.2.2.1. Simulated studies 59 vi 4.2.2.2. Flow characterization in the flake assemblies 63 4.3. Internal Bond Properties 68 4.3.1. Bonding strength 68 4.3.2. Failure surface observations 75 4.4. Dimensional Stability 81 4.4.1. Wood deformation and "spring or swell" back 81 4.4.2. Layered thickness swell of flake assemblies 88 5. S U M M A R Y A N D CONCLUSIONS 97 5.1. Remarks 97 6. LITERATURE CITED 99 APPENDIX 1 •. 104 APPENDIX 2 106 APPENDIX 3 107 APPENDIX 4 108 APPENDIX 5 : 109 APPENDLX 6 I l l APPENDIX 7 112 V l l LIST OF FIGURES Figure 2-1. Measured temperature and vapor pressure at the face and core layers of a flakeboard mat 4 Figure 2-2. Vertical density profile of waferboard 5 Figure 2-3. A representation of the time dependent relationship between temperature and pressure at the face and core regions of a flakeboard 10 Figure 2-4. Links of a wood and adhesive bond 12 Figure 3-1. Stress and strain relationships 22 Figure 3-2. Stress-strain relationships for the flake columns pressed at different levels of magnification 23 Figure 3-3. Flake assemblies (FA) moulder 26 Figure 3-4. Schematic diagram of the mat constituted from the flake columns and wood particles 28 Figure 3-5. Pressing procedure represented by pressure and displacement curves (15 seconds press closing time, 4.5% moisture content) 29 Figure 3-6. A representation of the FAs for tension-shear test 35 Figure 3-7. Specimens and the experimental device set-up for tension-shear test 36 Figure 3-8. Schematic diagram of the image analysis system .38 Figure 3-9. A representation of the FAs for thickness swell test 39 Figure 4-1. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 15 seconds press closing time and 4.5% moisture content, and the pressure response of the constituted mat 42 Figure 4-2. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 45 seconds press closing time and 4.5 moisture content, and the pressure response of the constituted mat 43 Figure 4-3. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 15 seconds press closing time and 15% face moisture content, and the pressure response of the constituted mat 44 i viii Figure 4-4. Monitored temperature at the face, intermediate and core layers for 32 FAs at oven dry and 4.5% moisture conditions using 15 seconds press closing time 45 Figure 4-5. Monitored temperature at the face, intermediate and core layers for 32 F A and flakeboard at 4.5 moisture content and 15 seconds press closing time 46 Figure 4-6. Vertical density profiles of 24 FAs made at different pressing conditions.... 50 Figure 4-7. Vertical density profiles of 32 FAs made at different pressing conditions 51 Figure 4-8. Vertical density profiles of 40 FAs made at different pressing conditions 52 Figure 4-9. IR spectum of laboratory PF resin 55 Figure 4-10. T M A profiles of commercial and laboratory resins 56 Figure 4-11. DSC profiles of commercial and laboratory resins 57 Figure 4-12. Effect of pressure on the flow of powdered PF resin 61 Figure 4-13. Effect of heating rate on the flow of powdered PF resin 62 Figure 4-14. Effect of pressure and heating rate on the flow of powdered PF resin 63 Figure 4-15. Cross section view of 24 FAs 65 Figure 4-16. Cross section view of 32 FAs 66 Figure 4-17. Cross section view of 40 FAs 67 Figure 4-18. Tension-shear strength of40, 32, 24 FAs at the face, intermediate and core layers 72 Figure 4-19. View of the FAs' fracture surfaces of different failure types 76 Figure 4-20. Number of each failure type for the FAs at the face, intermediate and core layers , 79 Figure 4-21. Number of each failure type for 40, 32 and 24 FAs 80 Figure 4-22. Cross sections of deformed wood structure at the compaction ratio of 1.09 ...82 ix Figure 4-23. Cross sections of deformed wood structure at the compaction ratio of 1.54 82 Figure 4-24. Cross sections of deformed wood structure at the compaction ratio of 4.00 83 Figure 4-25. Collapse of cell wall in the vessel cell 84 Figure 4-26. Collapse of cell wall in the fiber cells 84 Figure 4-27. Instantaneous and delayed spring back at the different compaction ratios ..85 Figure 4-28. Thickness swell of compressed flakes at the different compaction ratios.... 87 Figure 4-29. Thickness swell after 24 hours water soaking for 40, 32, 24 FAs at the face, intermediate and core layers 92 Figure 4-30. Thickness swell after redrying for 40, 32, 24 FAs at the face, intermediate and core layers 95 LIST OF TABLES Table 3-1. Surnmary of flake assemblies from the constituted boards manufactured at different pressing conditions 30 Table 3-2. Experimental design 32 Table 3-3. Summary of property tests performed for the FAs in different constituted board 33 Table 4-1. Some peak assignments of functional groups in PF resin 54 Table 4-2. Molecular weight distribution of acetylated laboratory and commercial resin 54 Table 4-3. Comparison of average tension-shear strength at the different layers in the FAs 69 Table 4-4. Comparison of average tension-shear strength of the FAs with different density levels 70 Table 4-5. Comparison of average tension-shear strength of the FAs pressed at different conditions 71 Table 4-6. Comparison of average thickness swell at the different layers in the FAs 89 Table 4-7. Comparison of average thickness swell of the FAs with different density levels 90 Table 4-8. Comparison of average thickness swell of the FAs pressed at different conditions 91 xi ABBREVIATIONS A N O V A Analysis of variance CR Compaction ratio CUT Constant temperature and humidity DSC Differential scanning calorimetry F A Flake assembly GPC Gel permeation chromatography 1R Infrared M C Moisture content MTS Material testing system PCT Press closing time PF Phenol formaldehyde SEM Scanning electron microscopy THF Tetrahydrofuran T M A Thermal mechanical analysis UBC University of British Columbia Xll A C K N O W L E D G M E N T S I would like to thank my supervisor Dr. Simon Ellis, Faculty of Forestry, U B C , for his invaluable advise, help and patience throughout this study. Special thanks also go to Dr. Paul Steiner, Dr. Stavros Avramidis and Dr. Frank Lam, Faculty of Forestry, U B C , for their suggestions for this research and support as committee members I also wish to thank Dr. Chunping Dai, Forintek Canada Corporation, for his enjoyable and enthusiastic discussion on the studied subjects. The laboratory assistance from Bob Myronuk, Rob Johnson, Diana Hastings and Avtar Sidhu is readily acknowledged. Finally, my greatest gratitude goes to my parents, my sister and my wife for their continuous support and encouragement during my graduate studies. 1 1. INTRODUCTION Flakeboard, a special type of particleboard product, is typically manufactured by initially applying relatively small amounts of adhesive to wood flakes, then mechanically forming these constituents into a loose mat structure and lastly consolidating the mat under heat and pressure to form an integrated board. As a structural panel, the mechanical and physical properties of flakeboard have been widely studied, such as bonding strength and dimensional stability. Although the performance of the product can be assessed from various tests, the causes of such performance are still far from clear. This is mainly because of the complex nature of flakeboard structure, which makes analyses difficult to perform. The application of modeling techniques is an effective approach in the investigation of flakeboard properties, which enables us to eliminate uncertain interactions stennning from the random arrangement of flakes. A number of researchers have successfully modeled the structure of flakeboard either empirically or theoretically. The studies have provided valuable indications and predictions on various aspects of flakeboard manufacture, such as heat and mass transfer and compression behavior of the flake mat (Harless, 1987; Wolcott, 1990); horizontal density distribution (Suchsland,1962; Suchsland and Xu, 1991; Xu, 1993; Dai, 1993). However, none of the above studies have been systematically concerned with localized properties in flakeboards. The objective of the study Was to use a simulated flakeboard construction, in terms of density variation, to evaluate its localized properties. Six sub-objectives contributed to the overall goal. 1. To evaluate in situ local temperature conditions in flakeboard during the hot pressing; 2. To determine the localized adhesive flow characteristics in flakeboard; 3. To evaluate the localized bonding properties in flakeboard; 2 4. To analyze the localized dimensional stability in flakeboard; 5. To determine the effects of pressing conditions on the localized board properties; 6. To develop a technique to measure the localized thickness swell using an image analysis system 3 2. B A C K G R O U N D 2.1. Heat and Moisture Transfer and Density Gradient Formation The conventional hot pressing operation involves three simultaneous major processes in a flake-mat: (1) heat and moisture transfer; (2) adhesive flow and cure; (3) mat compression and density gradient development, which in turn deterrnine the final properties o f the flakeboard produced. Heat and moisture transfer across the mat thickness has been studied intensively by many researchers (Rackwitz, 1954; Strickler, 1959; Maku et al, 1959; Kawouras , 1977; Rauch, 1984; Kamke and Casey, 1988; Bolton and Humphrey, 1989). Time dependent temperature changes were monitored in different layers within mats. Although local moisture content was difficult to be directly measured, instead, local vapor pressure was usually measured to evaluate the moisture migration. A typical representation o f temperature and vapor pressure gradients is shown in Figure (2-1). During the pressing, heat is transferred from the face layers to the core layers by means of both conduction and convection. While wood is a poor conductor in nature, the movement o f moisture becomes a more important factor to facilitate penetration o f heat into a compressed mat. Consequently, although the initial face temperature rise results primarily from conductive heat transfer, the temperature changes in the core of mats depend to a large extent on the fluctuation o f internal vapor pressure. Since the compression properties of wood are temperature and moisture dependent (Youngs, 1957; Bodig, 1982), during the mat deformation process, the transient temperature and moisture gradients result in non-uniform compression o f wood flakes across the board thickness and, with the cure o f adhesive, result in the development of board vertical density gradients (Figure 2-2) (Strickler, 1959; Suchsland, 1962; Kelly, 1977; Smith, 1982). 4 Figure 2-1. Measured temperature and vapor pressure at the face and core layers of a flakeboard mat. ( target density: 0.72g/cm 3, pressing cycle: pressing to close at 1 minute and venting began at 6 minutes; mat moisture content: 6%:, olaten temperature: 190°C. adopted from Kamke and Casey, 1988 ). 5 Figure 2-2. Vertical density profile o f waferboard ( target density: 40pcf; press closing time: 30 seconds; mat moisture content: 5%; platen temperature: 210°C adopted from Smith, 1982 ). 6 Heat and moisture transfer and vertical density gradients development in a mat are also influenced by pressing conditions (Kelly 1977). Increasing press closing time (PCT) and mat moisture conditions are the two most effective ways to promote the rate of temperature rise in compressed mats. Since a short PCT produces a fast rate of mat densification and a rapid buildup of vapor pressure in the face region, a greater heat transfer occurs during the early stages of the press cycle. Pressing a high initial face moisture content (MC) mat in flakeboard manufacture can cause what is known as "steam shock" effect. A high moisture content in the face will again result in a rapid buildup of face vapor pressure and subsequently expedite convective heat transfer towards the core. Consequently, a different vertical density profile will be developed, where a faster PCT and a higher initial face moisture condition give a comparatively higher face density and a lower core density (Smith, 1982). The concept of horizontal density distribution of flakeboard was first introduced by Suchsland (1962), however its importance has not been given enough attention until recently. Suchsland and Xu (1989, 1991), and Steiner and Dai (1993) either empirically or theoretically modeled the mat structure of flakeboard. The potential effect of horizontal density variation on board properties was clearly demonstrated. But, to date, no attempt has been made to investigate the effect of horizontal density on heat transfer inside a flake mat. In the previous studies, temperature and moisture transfer were both assumed to be uniformly distributed in the plane of a mat. While the local mat densities may influence heat flow in different density regions within a compressed mat, and therefore affect the development of the vertical density gradient. 2.2. Adhesive Flow The formation of an adhesive bond involves several steps: wetting, adsorption, interdiffusion and hardening. The flow of an adhesive is the initial phase in the bond 7 formation, where adhesive wets, spreads and penetrates into wood particles to provide necessary resin coverage. The magnitude of flow deterrnines the potential bonding areas between wood and adhesive as well as the thickness of gluelines, and therefore affects the bonding properties of a glued product. A great deal of research has been carried out to investigate resin flow characteristics in wood composites (Nearn, 1965; Harata et al, 1968; Bryant 1977; Wilson et al., 1979; Blomquist, 1981; Ellis, 1989). It is well agreed that a moderate amount of resin flow is adequate to form a good wood-adhesive bond. Either excessive or insufficient flow could cause a starved glueline or little diffusion into the wood respectively, and subsequently result in poor bond quality. There are many variables involved in the process of resin flow. Resin molecular weight distribution, wood structure and bonding conditions such as, temperature; moisture content and pressure are recognized as the key factors which affect its flow properties. Brady and Kamke (1988) studied penetration of liquid phenol formaldehyde (PF) resin and pointed out that temperature and moisture influenced the extent of resin flow by controlling its viscosity, while pressure provided the driving force for resin hydrodynamic flow. It is also believed that the effect of resin molecular weight on resin flow is also via its viscosity. Ellis (1989, 1993) indicated that resin molecular weight distribution affects its flow through the relative proportion between high and low molecular weight species in the resin mixture. He found that resin with greater proportions of low molecular weight components exhibited larger extent of flow. In terms of resin penetration, the structure of wood and resin molecular weight appear to be the more significant factors which affect the diffusion of a resin into a wood substrate. 2.2.1. Adhesive flow and cure Adhesive flow and cure are the two major processes in the formation of bonds. Although equally important, it is well understood that the physical and mechanical changes of a resin 8 depend primarily on its polymerization progress. Hence, the flow process can be explained from the relationship between resin rheological and cure properties. Steiner and Warren (1981) were the first investigators to study the rheological behaviors of PF resin in relation to its cure. A torsional braid technique was used to monitor the resin's rigidity (oscillatory damping factor). The results demonstrated a series of progressively established resin rigidities with the development of resin cure. A major increase in resin rigidity was observed after the temperature reached 120°C. Geimer et al. (1990) further investigated the development of mechanical stiffness and chemical advancement of PF resin using a dynamic mechanical analyzer (DMA) and differential scanning calorimetry (DSC). The resin chemical-mechanical relations were quantitatively measured, the results showed that mechanical cure developed at a much higher rate than chemical cure. However, the above studies were more focused on the building up of resin mechanical properties rather than resin flow, which occurs in the early stages of the curing process. One may argue that the cure of a resin can only have a little effect on the magnitude of resin flow. This might be true for a liquid resin, where flow will occur readily as soon as a resin is applied, but, for a powdered adhesive, the resin must first melt before sufficient flow can take place. The polymerization reaction could take place during resin fusion and thus, affect the its mobility. Consequently, to understand the flow process of powdered resins, a more specific evaluation of the relationships between flow and cure for powdered PF resin is needed. 2.2.2. Flow of powdered PF resin in flakeboard Adhesives in flakeboards are exposed to changing temperature, moisture and pressure conditions during hot pressing. The effects of these conditions on the flow of resins can be interpreted both horizontally and vertically. Horizontally, because of the existence of horizontal density variation in a randomly formed mat, the localized pressure applied in a 9 plane of a mat is non-uniformly distributed. Higher density regions will receive greater pressure than do lower ones. Additionally, the rates of heat and moisture transfer among those density regions may differ to a certain extent. This subject will be discussed in the section 4.1. Vertically, as a result of heat and moisture transfer gradient, temperature and moisture conditions are different at a given time interval, and these conditions also vary with the pressing time. At the same time, pressure also changes due to the deformation, creep and stress relaxation of a mat throughout pressing. This time dependent relationship between temperature and pressure at different locations (face and core) is illustrated in Figure 2-3. As we can see, at the early stages of pressing, the higher temperature levels in the face layer are associated with high pressure, while the most rapid temperature rise in the core is accompanied by a declining pressure due to delayed heat and moisture transfer. These different temperature increase rates, together with their corresponding pressure levels across mat thickness thus result in different environmental conditions for adhesive to flow. In brief, adhesives in flakeboard are subjected to different heating rates, moisture conditions and pressure levels from both "horizontal and vertical effects". Consequently, non-uniform resin flow characteristics within a panel could result. To date, little work on the flow properties of powdered thermosetting adhesive in flakeboards has been reported. Ellis (1989) studied the adhesive flow of both commercial and laboratory resins using a thermal mechanical analysis (TMA) technique. The results showed different flow characteristics corresponding to different resin chemical structures. Nevertheless, the effect of heating rate; moisture conditions or pressure levels on adhesive flow was not studied. Furthermore, Ellis (1993) compared adhesive flow characteristics in flakeboard using powdered resins with different particle sizes. The glueline characteristics were microscopically observed. Only representative gluelines were 10 r- 3 5 I I I I I 1 1 1 1 1 o - B O 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 4 2 0 4 8 0 5 4 0 T i m e a f t e r p r e s s c l o s u r e ( s ) Figure 2-3. A representation of the time dependent relationship between temperature and pressure at the face and core regions of a flakeboard (adopted from Kawouras, 1977). 11 investigated, so, no information was provided on the variation of localized resin flow. Knowledge about the non-uniformity of resin flow in different locations within a flakeboard mat during hot pressing is limited. Yet, this goal is difficult to achieve because of the unavailability of suitable analytical techniques. However, in order to better understand resin performance in flakeboard, a reasonable and logical approach is still necessary. 2.3. Flakeboard Bonding Properties A wood/adhesive bond is comprised of three major components: wood adherent, wood/adhesive interface and adhesive (Blomquist, 1981). Adhesive bonding creates a stress transferring mechanism to from one adherent to the other. Marra (1964) introduced a chain analogy to interpret a wood/adhesive joint as a system of nine symmetrical links, representing adhesive film, adhesive boundary layer, adhesive and adherent interface, adherent boundary and adherent itself (Figure 2-4). He indicated that the performance of an adhesive bonded joint depends on the properties of each link. The failure of any one link will cause failure of such a hypothetical chain and, in the same way, failure of any one of the actual locations in a bonded assembly will result in failure of the entire assembly. Flakeboard structure is made up of a randomly fabricated flake-network connected by means of their inter-flake bonds. Unlike other glued wood products, those wood/adhesive bonds are not homogeneously distributed in the three dimensional flakeboard structure due to density variation, and heat and moisture transfer gradients (Kelly 1977; Suchsland and Xu 1991). This variation of bonding properties in the thickness direction of a panel has been investigated by a number of researchers (Shen and Carroll, 1969; Plath and Schnitzler, 1974; Brady 1989). All these studies revealed that bonding strengths were 12 Link 1: The adhesive film. Link 2 and 3: Adhesive boundary layer. Link 4 and 5: Adherent and adhesive interface. Link 6 and 7 Adherent boundary layer. Link 8 and 9: Wood adherent. Figure 2-4. Links of a wood and adhesive bond (adopted from Marra, 1964). 13 higher near the face layer and lower in the core of a board. Shen and Carroll (1969) and Plath and Schnitzler (1974) related the layer strengths to the vertical density gradient of a panel. Brady (1989) also showed that strengths of layer also depended on the heat and moisture conditions during the hot pressing. To date, the study on wood/adhesive bonding properties along the horizontal plane of a board is limited. One approach was made by Suchsland and Xu (1991), where they used a veneer strip model to simulate the structure of flakeboard in term of horizontal density variation. Internal bond tests were performed on cut elements with different numbers of veneer overlaps. The results showed a strong influence of density on the bonding properties. The strength values were in general greater in higher density regions in the constructed board, but decreased when density was above a certain limit. Nevertheless, in the above study, since internal bond tests can only measure the bonding strength at the weakest part in each element, no information upon the variability of horizontal bond strength in flakeboard was provided. To further understand inter-flake bonding properties in a flakeboard structure, such as how density, heat and moisture conditions affect wood and adhesive bonding inside flakeboard, what is required is a more specific investigation on bonding properties at different locations in term of both horizontal and vertical density variation. 2.4. Dimensional Stability of Flakeboard The dimensional behavior of a pressed flakeboard panel involves three major components (1) natural dimensional changes of wood; (2) release of compressive stresses imparted to the board in the pressing operation (Halligan, 1970; Alberta Research Council, 1987) 14 (3) non-imiformed board structure, in terms of density variation, which results in different localized dimensional properties (Suchsland, 1967; 1973; 1989). Flakeboard, like other wood products, is a hygroscopic material. It will swell or shrink when subjected to different environmental conditions that cause absorption or desorption of water below the fiber saturation point. Flakeboards generally have poorer dimensional stability than plywood or solid wood because of the magnitude of compression of wood elements in the thickness direction. This amount of compression tends to spring back either elastically and inelastically (Kunesh, 1961) or swell back when subjected to different humidity conditions. The elastic spring back is a time independent variable and obeys Hooke's law, and this amount of deformation will be recovered immediately after removal of compression pressure. While the inelastic spring back is a time dependent variable and the recovery of deformation is delayed with time. Water soaking is one of the most severe condition that causes a maximum dimensional change of a board in the thickness direction. It has been recognized by Stamm and Cohen (1956) and Rice (1976) that the thickness swell of compressed wood in the water soaking situation follows two mechanisms. One is the swelling due to the normal dimensional changes of the wood cell wall as it absorbs moisture, and the other is the balloon back of crushed wood cells when they uptake water in the cell lumen, which simultaneously also allows the release of residual compressive stresses imparted to the board during the hot pressing operation. The relationship between the non-uniform dimensional behaviors and density variation in flakeboard was first addressed by Suchsland (1967, 1973). He indicated that dimensional changes of compressed wood do not occur homogeneously within flakeboards, since higher density regions tend to spring or swell more than lower density ones do. Suchsland 15 and Xu (1989) used a veneer-strip model to simulate horizontal density variation and studied the effects of horizontal density distribution on flakeboard dimensional stability. The results further supported Suchsland's previous findings. The implications of this non-uniform local dimensional stability are that it would initially result in internal stresses between different density regions (Suchsland, 1973). Furthermore, when a panel is subjected to high humidity conditions, if the stresses caused by the swelling of compressed wood are greater than the adhesive bond strength between wood particles, bonds will be broken resulting in a separation of wood particles leading to the development of large voids in the panel (Suchsland, 1967; Kelly, 1977; Hsu et.al, 1988). Besides the normal dimensional change of wood, the other two phenomena identified in the dimensional behavior of flakeboards are not yet fully understood. However, it is generally believed that the dimensional properties of compressed wood are predominantly determined by wood deformation processes. Accordingly, a further investigation of the compression process of wood and its implication with respect to the dimensional properties of wood would be useful to improve the comprehension of the dimensional behavior of flakeboards. 16 3. M E T H O D O L O G Y It is well agreed that the analysis of flakeboard properties involves a lot of variables from raw materials and forming processes, such as flake dimensions, flake surface quality and flake orientation, that obstruct a thorough investigation of its internal properties. In this work, in order to study the effect of internal structure in terms of density variation and adhesive flow on the internal board properties, such as localized bonding properties and dimension stability, microtomed flake assemblies (FAs) were prepared instead of a regular flakeboard and used to simulate a flakeboard's structure. The study was carried out in the following four steps: (1) cold pressing of flake columns; (2) adhesive analysis; (3) hot pressing of flake assemblies; (4) flake assembly testing techniques development and properties examination. 3.1. Materials 3.1.1. Flake preparation Green aspen (Populus tremuloides) logs were obtained from the University of British Columbia (UBC) Alex Fraser Research Forest at Williams Lake, B.C. and cut into slices of approximately 50-70 mm in thickness. In an attempt to minimize the density variation between flakes, sapwood was removed from the log slices and sawn into the desired size blocks (55x17.5x45-55 mm). After being water soaked for two weeks, the wood blocks were sliced using a sliding microtome into 55x17.5x0.69 mm flakes with the grain direction parallel to the long edge. During the microtoming, a variety of slicing conditions were tried to ensure a uniform flake surface quality. Then the microtomed flakes were conditioned in an environmental chamber for two weeks at a dry bulb temperature of 40° C; 25°C and relative humidity of 21%;73% to reach the equihbrium moisture content of 17 4.5% and 15%, respectively, and stored in polypropylene bags to maintain their equilibriiun moisture contents. 3.1.2. Flake density determination Fifty randomly selected microtomed flakes were conditioned in a constant temperature and humidity (CTH) room ( relative humidity, 50%; temperature, 20°C ) to reach an equnibrium moisture content of 9%. Then each flake was weighed and its dimensions were measured with a micrometer. The density of each flake at 9% M C was calculated based on its weight and volume. 3.1.3. Wood particle preparation Aspen flakes with about 7% moisture content, obtained from the Alberta Research Council, Edmonton, were ground to approximately lx lx l mm fine wood particles using a large Wiley mill. Then the fine wood particles were conditioned in the environmental chamber to reach the desired M C of 4.5% and 15.0%. 3.1.4. Adhesive Synthesis Chlorazol black labeled PF powder resin was synthesized in the laboratory in an attempt to make resin more detectable from wood in later microscopic analysis. The procedure for the synthesis of PF resin was based on the production of commercial resin described in the patent of Berchem et al. (1978) of Reichold Chemical Ltd. Phenol, formaldehyde and sodium hydroxide (molar ratio 1.0:2.0:0.2) were reacted in a 4L resin kettle. The phenol, 90% of formaldehyde solution, 50% of the sodium hydroxide and water to make 50% solid content was charged and held at 60°C for 30 minutes. The balance of the formaldehyde content and 25% of the sodium hydroxide was added and the mixture heated to 85°C. After a further 30 minutes, the remaining 25% of sodium hydroxide was added and the reaction mixture was then maintained until the final desired viscosity 18 (Gardner Holt H at 25°C) had been reached, at which point the reaction mixture was rapidly cooled. The liquid PF resin was stored in a resin bottle and kept in a cold storage room(5°C). Prior to converting the liquid resin to a solid phase, chlorazol dye was blended into the Uquid PF resin at a loading of 1% resin solids content. Then the chlorazol labeled PF resin was freeze-dried to the solid form The solid resin was further ground to a powder using a pestle and mortar, and sieved to pass a 200 mesh screen. 3.2. Adhesive Analysis 3.2.1. Infrared (IR) absorption spectroscopy The IR spectrum through its rich array of absorption bands provides a wealth of information about functional groups within a molecule. In this experiment, the resin samples were prepared for the powdered laboratory PF resin using a pellet forming press, the IR spectrum of the resin was recorded using a Perkin Elmer 1600 series FTXR with 64 scans in the range of 4000-400 cm"1. The sensitivity of this instrument was 8 cm 1 . 3.2.2. Gel permeation chromatography (GPC) GPC is a special type of a liquid-solid chromatography that separates molecules according to molecular size. The GPC system used in this experiment consisted of a solvent container, an isocratic Spectra Physics 8810 pump, a Rheodyne 7125 injector loop, a Micropak TSK exclusion column series (a GH8P guard column and 4 analytical columns, 1000H; 2500H; 3000H; 4000H), a Kratos Spectroflow 757 UV/VIS detector, a Spectra Physics 4290 integrator and a microcomputer with a printer. In order to make the PF resin soluble in the tetrahydrofuran (THF) solvent for the analysis, acetylated adhesive samples were initially prepared according to the techniques 19 used by Ellis and Steiner (1991). Then, each acetylated resin sample was dissolved in the THF solvent at a concentration of 0.5% w/v. During the analysis, the solution was injected into the Rheodyne 7125 injector loop: By means of the pump, the solution was forced to pass through the exclusion column series to differentiate molecular chains depending on their molecular size. At a wavelength of280 nm, this information was acquired by the detector and then transferred to the integrator for calculation. Lastly, the final results were downloaded to the microcomputer. The GPC columns had previously been calibrated using a series of commercial polystyrene standards. The analysis was made for both powdered commercial and laboratory PF resin and three replications were performed for each sample. 3.2.3. Differential scanning calorimetry (DSC) DSC is a thermal technique used to measure the differences of heat flow between a substance and a reference under a controlled heating program. In this study, DSC was used to study the cure properties of PF resin by monitoring its thermal responses. The analysis was performed on a Du Pont DSC 2910 instrument and a central computer system. An empty pan was used as the reference cell. Resin samples from both powdered commercial and laboratory adhesives were prepared. Approximately 4.5 mg of resin was carefully weighed into a DSC pan, which was then sealed with a DSC lid. The prepared resin samples were run in the DSC at a heating rate of 10°C/minute in the temperature range of 25°C to 250°C. The thermograms were automatically acquired and plotted by the central computer. 3.2.4. Thermal mechanical analysis (TMA) T M A is an analytical technique used to determine the physical responses of a material during a controlled heating program In this study, the dimensional changes of powdered PF resins during their melting and hardening were monitored by a DuPont T M A 2940 20 mermo-mechanical analyzer and a central computer system. Compact resin pellets were prepared for the test using a T M A pehet cylinder and sample encapsulating press. During the preparation, approximately 30 mg powdered resin was weighed and carefully poured into the cylinder with a diameter of approximately 4 cm and covered with the cylinder head. The cylinder was pressed in the sample encapsulating press. A uniform forming pressure was used for each sample. The samples from both commercial and laboratory PF resin were initially run in the temperature range of 25°C to 200°C at the heating rate of 10 °C/minute. The pressing and heating rates of each run were pre-programmed using the available software in the central computer system The dimensional changes were measured by the probe of the T M A attached to the resin pellet. The results were acquired by the microcomputer. In an attempt to simulate the localized environmental conditions of flakeboard during hot pressing, the flow characteristics of laboratory PF resin were examined using T M A under varied heating rates and pressing force levels. The variables studied are listed below. Three replications were performed for each test. (1) Pressing force levels at 10°C/minute: 0.05N, 0.25N, 0.45N, 0.65N (2) Heat rates at 0.05N: 10°C/minute, 30°C/minute, 50°C/minute, 70°C/minute (3) Pressing force levels and heat rates: 0.05N, 10°C/minute; 0.25N, 10°C/minute; 0.05N, 30°C/minute; 0.25N, 30°C/minute 21 3.3. Flake Column Cold Compression The objective of this experiment was to characterize the wood deformation process by defining their stress and strain relations and subsequently examine the effect of wood deformation on the dimensional stability of wood flakes. 3.3.1. Stress and strain determination 3.3.1.1. Equipment The compression tests were conducted on a Material Testing System (MTS) control testing machine. The pressing procedures were pre-programmed on the MTS controller and experimental data was acquired by a microcomputer attached to the testing system 3.3.1.2. Method and procedure Two flake moisture contents, 4.5% and 15%, were tested. In order to minimize the variability between individual flakes, microtomed flake columns formed from 15 randomly selected flakes were prepared for each pressing. The flake columns were compressed at a loading rate of 0.5 rrnn/minute at room temperature and humidity conditions. Three replications were performed for each test. This experiment was carried out in two steps. Firstly, the flake columns were compressed to the maximum strain level. The deformation of the flake columns then was characterized by a linear and non-linear stress and strain relationship (Figure 3-1). As a result of dynamic changes of wood structure, different mechanical responses were identified from the stress and strain relation curve: AB-linear elastic; BC-nonlinear; CD-curvilinear. Therefore, the physical and mechanical properties of deformed wood flakes could be studied based on the above three regions. Secondly, mid-points in the three regions, 8.5; 35; 75 percentage of maximum strain were selected as representative strain levels for further compression tests. Then, the flake columns were compressed to these pre-determined strain levels (Figure 3-2). 22 Figure 3-1. Stress/strain relationships. 23 3.0 0 0.1 0.2 0.3 0.4 Strain Strain Figure 3-2. Stress-strain relationships for the flake columns pressed at different levels of compaction ratio (a)=1.09; (b)=1.54; (c)=4.00. 24 3.3.2. Dimensional Stability The dimensional stability test involved measuring spring back of compressed flake columns and the thickness swell (TS) of individual flakes. The instantaneous spring back and delayed spring back were determined for each compressed column. Firstly, the thickness of each column prior to pressing (T0), immediately after the release of the pressure (Tj) and after being stored in sealed plastic bags for 2 weeks (T2) were measured using a micrometer. Secondly, the thickness of the deformed flake column (TJ) at the targeted strain level (S) was calculated by: Td = Tox{l-S) [1] and lastly the spring back of each column was calculated by: SB = ZLlZkxloo% [2] o where /' = 1, instantaneous spring back / = 2, delayed spring back The TS test was performed using 4.5% M C flakes. Flakes were soaked in water at room temperature for 2 hours and redryed in the C T H room. The thickness was measured at the three points of each flake before and after the water soaking, and redrying using a micrometer. Three replications were performed for each test. Thickness swell values were calculated by: TS = ^ - ^ x l 0 0 % [ 3 ] o where, TS (%): thickness swell of flake Ta (mm): original flake thickness T( (mm): thickness after / hours water soaking and redrying 2 5 3.3.3. Microscopic examination The deformed wood structure compressed at different strain level was examined using a scanning electron microscope (SEM). The flakes were randomly selected from the compressed flake column. A small wood piece of approximately 5x5 mm was removed from the central part of each flake. The observed surfaces were cut using a fresh razor blade on a plastic board. After they had been carbon coated, the samples then were examined in a SEM using a variety of magnification levels. 3.4. Flake Assembly Hot Pressing 3.4.1. Equipment The equipment used in hot pressing was a micro-computer, material testing system (MTS) controlled 300 x 300 mm electrically heated press. The pressing procedures were pre-programmed using a position and a pressure control. The experimental data of compression pressure, mat deformation and interior temperature were acquired by the micro-computer. The flake columns were formed in a flake column moulder (Figure 3-3). The moulder consists of two identical parts connected by horizontal bars with screw eyes. The two parts can move horizontally allowing to lay up flakes and grip the flake column. 3.4.2. Method FAs were made from 24, 32, 40 microtomed flakes with the targeted densities of 0.507, 0.676 and 0.844 g/cm3, respectively. In an attempt to simulate the pressing conditions in flakeboard, nine flake columns with three of each from 24, 32, 40 flake columns were placed evenly into a 450 g fine wood particle mat with a target density of 0;5 g/cm3 to 26 Figure 3-3. Flake column moulder. 27 form a constituted mat (Figure 3-4). In the press, the flake columns were pressed within the wood particle mat. The general pressing parameters used were: (1) Press temperature, 200°C; (2) Pressing time, 7 minutes; (3) Target thickness, 12.7 mm; (4) Maximum pressure, 0.828 MPa. Using the pressure and position control during pressing, the pressure was first applied to the constituted mat to reach its maximum value at the required press closing time (PCT) and this pressure was maintained until the targeted thickness of 12.7 mm was reached. Then this thickness was retained throughout the pressing cycle (Figure 3-5). Three different pressing conditions were used in the experiment. (1) 15 seconds PCT at 4.5% moisture content; (2) 45 seconds PCT at 4.5% moisture content; (3) 15 seconds PCT at 15% moisture content in the face layers and 4.5% moisture content in the core layers. Ehiring pressing, the internal temperature was monitored at the face (1/8), intermediate (1/4) and core (1/2) layers of flake columns by inserting thermocouples. In an attempt to compare heat and moisture transfer in the FAs within the constituted board to that of flakeboard, localized temperatures at the face, intermediate and core layers within a laboratory pressed flakeboard were also measured. Flake mats were formed using 4.5% M C flakes (0.5x25x80 mm) with a target density of 0.65 g/cm3. The mats were pressed to the thickness of 12.7 mm using a maximum pressure of 3.45 MPa and 200°C hot platen temperature. Nine replications of the constituted mat were pressed for each of the three conditions with three of them being used for interior temperature measurements. FAs that contained no resin were also made for thickness swelling tests. The total of FAs manufactured within the constituted board is summarized in Table 3-1. 28 Figure 3-4. Schematic diagram of the mat constituted from the flake columns and wood particles. 29 urc 3-5. Pressing procedure represented by pressure and displacement curves (15 seconds press closing time and 4.5% moisture content). 30 Table 3-1. Summary of the flake assemblies in the constituted boards manufactured at different pressing conditions Press Conditions No of Constituted mats Pressed No. of Flake Assemblies Manufactured or No. of Replications for Each Measurement * each mat pressed contains 9 FAs with 3 for each for 24, 32, 40 FAs 24 FAs 32 FAs 40 FAs 15 seconds PCT at 4.5% M C 6 mats (FAs were noted as: *1541, 1542, 1543, 1544, 1545, 1546) 18 18 18 3 (for temperature measurement) 3 3 3 15 seconds PCT at oven dry M C 3 (for temperature measurement) 3 45 seconds PCT at4.5%MC 6 mats (FAs were noted as: *4541, 4542, 4543, 4544, 4545, 4546) 18 18 18 3 (for temperature measurement) 3 3 3 15% MC/face at 15 seconds PCT 6 mats, (FAs were noted as: **mcl, mc2, mc3, mc4, mc5, mc6) 18 18 18 3 (for temperature measurement) 3 3 3 Note: * 15-4-1 or 45-4-1 first two numbers: press closing time (15 seconds or 45 seconds), third number: moisture content (4.5%); forth number: replications. **MC1: MC represents 15% /face moisture content, number represents replications. 31 3.4.3. Procedure Firstly, the chlorazol labeled PF resin was mixed with flakes in a plastic bag at a loading level of 1% based on oven dry weight of wood flakes for each constituted mat. Secondly, the flake columns were formed in the flake column moulder. In order to avoid losing adhesive, each flake was picked up using a tweezers and laid up into the moulder. After adding the desired number of flakes, the constructed flake columns were bonded with pre-prepared paper-tape. For the flake columns formed with 15% M C flakes on the face. Numbers 4, 5, 6 flakes were placed on the either side of 24, 32, 40 flake columns respectively. Thirdly, the constituted mats were constructed with each of three 24, 32, 40 flake columns and the fine wood particles and those mats were pressed at pre-determined conditions. After the pressing, FAs were removed from the pressed constituted mat and conditioned in the C T H room for moisture equilibration. 3.5. Flake Assembly Properties The property tests for the FAs involved density profile measurements, tension shear tests, dimensional stability tests and microscopy observations. Throughout this study, the variables studied were: three pressing conditions, three FA's density levels, three vertical layers within each FA. The experiment was arranged in a 3x3 factorial split-plot design. The variables studied are summarized in Table 3-2 and the properties tested performed for each set of 24, 32, 40 FA in a compressed constituted mat are summarized in Table 3-3. A 3.5.1' Density profile determination The density profile measurements were conducted on a R E C O M 8900/DA density analyzer at the Alberta Research Council. The main components of the density analyzer are: VAXstation 3100 computer, gamma radiation source with computer operator shutter and detection systems, computer operated sample positioning system, operator's terminal with graphic monitor and a printer. 32 Table 3-2. Experimental design Experimental factor Number of levels Levels studies Pressing conditions 3 15 PCT, 4.5% M C ; 45 PCT, 4.5% M C ; 15 PCT, 15%MC/face Number of flakes in a F A (density levels of FAs) 3 24 (0.507g/cm3) 32 (0.676g/cm3) 40 (0.844g/cm3) Layers in each F A (locations) 3 face intermediate core 33 Table 3-3. Summary of property tests performed for the FAs in different constituted mats. 24, 32, 40 FAs manufactured in the constituted mats at different pressing conditions Density profile measurement Tension shear test Thickness swell Microscopic examination 1541 4541 M C I 3 1542 4542 MC2 2 1543 4543 MC3 3 1544 4544 MC4 liillilllilllllllilillil 1 1545 4545 MC5 3 1545 4546 MC6 Replications 3 3 3 * shaded area indicates number of tests were performed for each set of 24, 32, 40 FAs in a constituted mat. 34 FAs were firstly trimmed to the dimensions of approximately 51x13 mm. Then, the samples were placed in the density analyzer and scanned along the thickness direction of each FA. Lastly, experimental data was acquired by the VAXstation 3100 computer and graphed in a microcomputer. 3.5.2. Tension shear strength Bond quality evaluations were determined on the three layers of FAs: face, intermediate and core at a distance of approximately 2 mm, 4 mm and 6 mm respectively from their surfaces. The test specimens were prepared by removing two narrow sections from each side of a F A with a table saw to expose the targeted gluelines (Figure 3-6). For the specimens tested in the face, in order to prevent failure occurring at the thin section near the surface rather than the examined gluelines, a small wood piece was glued on the surface of the F A (Figure 3-7). The tests were performed on an Instron test machine equipped with pneumatic jaws. During the preliminary tests, it was found that it was inappropriate to use the pneumatic jaws to directly grip the specimen. This was because either a lower gripping force could not hold the FAs during the tension test or a larger gripping force would damage the test sample. Therefore, a new test device was developed. Two holes were drilled in both ends of a FA at a diameter of approximately 6 mm. Then a screw was inserted in each hole. Spacers, whose diameter were just a little wider than the width of the FAs, were attached on both side the FAs by screwing matched nuts. During the test, the spacers on either side of the FA were gripped by the pneumatic jaws (Figure 3-7). In order to hold the specimen firmly, a pressure of 3.45 MPa was applied to the jaw. The specimens were pulled in tension at a loading rate of 1.27 mm/minute and failed in a shear mode. The failed specimens were retained and examined under a stereo-microscope. 35 Unit:mm Figure 3-6. A representation of the FAs for the tension-shear test. 36 Figure 3-7. Specimens and the experimental device set-up for tension-shear test. 37 3.5.3. Dimensional stability determination This experiment consisted of measuring TS of FAs and individual flakes from hot pressed flake columns. 3.5.3.1. Instrumentation The measurement of the TS for the FAs was conducted on an image analysis system, which consists of four components: (1) a Javelin JE 3462FTR color video camera; (2) a T A R G A + 16/32 frame grabber; (3) an IBM 486-compatible microcomputer with Jandel Video Analysis software; and (4) a Javelin RGB color monitor (Figure 3-8). 3.5.3.2. Method In the image analysis system, an image is comprised of a set of pixels displayed on a raster screen. Through the frame grabber, the image can be transformed to digital form Each pixel contains three pieces of information: an X location, a Y location and a gray level. The geometric measurement is based on the X and Y locations of pixels. In the measurement, an image of one side of a F A was captured from the camera and shown in the monitor. The geometric measurements were performed using a hand controlled computer mouse and the JAVA software. The experimental data and images were saved in the microcomputer for further analysis. The TS test was performed at three line segments (surface-face; face-intermediate; intermediate-core) corresponding to the three portions of a FA that is face, intermediate and core. (Figure 3-9). The FAs were soaked in water at room temperature for 24 hours and redryed. The measurements were made after 24 hours water soaking and subsequent redrying. The "layered TS": face; intermediate; core, were calculated by equation [3], Figure 3-8. Schematic diagram of the image analysis system. 39 Figure 3-9. A representation of the FA for thickness swell test. 40 3.5.3.3. Procedure In the measurement of layered TS using the image analysis system, firstly, the light condition and magnification level of F A were adjusted to produce a good image quality and they were maintained constant for all the examinations. Then, before the testing, the geometric unit in the image system was calibrated by importing an image of a standard ruler using the available software. Thirdly, using a pen, the targeted measured line segments for TS from the surface to the core were marked (Figure 3-9) at a distance of 2 mm, approximately. Fourthly, the targeted line segments were measured before water soaking, after 1, 3, 6, 12, 24 hours water soaking and redrying using the image analysis system. 3.5.4. Microscopic examination Microscopic examination was conducted on a stereo-light microscope. Samples were prepared using a modification of an embedding and pohshing technique (Ellis, 1993). Spurr's low viscosity epoxy resin was used to embed the FAs. One end of the FAs was briefly soaked in acetone and then went through a mixture of 50% acetone, 50% epoxy; 10% acetone and 90% epoxy; and finally 100% epoxy series. After having been soaked in 100% epoxy resin for 24 hours, the FAs were placed in a glass plate and heated in an oven at 70°C for 10 hours to cure the adhesive. The pohshing task was performed in a series of grinding discs and the embedded end of the FAs was ground using successively finer abrasive papers (60, 120, 180, 320, 600 grit). The polished surfaces were viewed using stereo-light microscopy. Fuji color 100 daylight film was used for color photomicroscopy. 41 4. RESULT A N D DISCUSSION 4.1. Internal Conditions and Density Profiles of the FAs in the Constituted Board It is known that adhesive performance, internal bonding strength and dimensional stability of a wood composite depend on the board density level, pressing conditions, and the resulting internal mat environments and density distributions (Kelly, 1977). The main objective of this section is to provide initial information on the internal conditions and density profiles the FAs, as a basis to further evaluate the internal properties of the FAs and flakeboard. 4.1.1. Internal Conditions of the FAs in the Constituted Mat Results of the temperature monitored at the face, intermediate, and core layers in the constructed 24, 32, 40 FAs are shown in Figure 4-1, Figure 4-2, and Figure 4-3. Recorded temperature changes at the face, intermediate and core layers of flakeboard and the 32 FA pressed at oven dry moisture content in comparison with the temperature curves of the 32 F A pressed at 4.5% moisture content are shown in Figures 4-4 and 4-5 respectively. Each curve is the average of three measurements. Figure 4-5 shows that the temperature changes observed in the different layers of the FAs are similar in trend to those observed in the flakeboard. However, the initial rate of heat transfer is somewhat slower in the FAs in the constituted mat than in the flake mat. This difference is probably because of the lack of between flake permeability in the FAs or the loss of moisture from the edge of the FAs decreased convective heat transfer within the FAs in the constituted mat. Nevertheless, Figure 4-4 demonstrates a faster rate of initial temperature rise at the intermediate layer in the FAs pressed at 4.5% M C than at oven dry MC. It is evident that moisture still played an important role for heat transfer in the FAs. 42 Figure 4-1. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 15 seconds press closing time and 4.5% moisture content, and the pressure response of the constituted mat. 43 Figure 4-2. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 45 seconds press closing time and 4.5% moisture content, and the pressure response of the constituted mat. 44 Figure 4-3. Monitored temperature at the face, intermediate and core layers for 24, 32, 40 FAs pressed at 15 seconds press closing time and 15% face moisture content, and the pressure response of the constituted mat. 45 Figure 4-4. Monitored temperature at the face, intermediate and core layers for 32 FAs at oven dry and 4.5% moisture conditions using 15 seconds press closing time. 46 Figure 4-5. Monitored temperature at the face, intermediate and core layers for 32 F A and flakeboard at 4.5% moisture content and 15 seconds press closing time (Flakeboard; flake geometry: 0.5x25x80 mm, target density: 0.65 g/cm 3, max. pressure: 3.45 MPa). 47 As expected, at the same FA density level a greater rate of initial temperature rise inside the FAs was observed at both faster PCT (15 seconds vs. 45 seconds) and higher initial face M C (15% vs. 4.5%) (Fig. 4-1, 4-2, 4-3). These results agree qualitatively with previous studies (Strickler, 1959; Kamke and Casey, 1988). However, the significance of the effects is greatly reduced with the increase of density level from 24 F A to 40 FA, particularly in the core layer. Since the improved heat transfer from a higher press closing rate and mat moisture condition is mainly because of the increase of convective effects (Strickler, 1959), this phenomenon therefore reflects upon the influence of density on the magnitude of convective heat transfer. In the low density FAs, the inside of the F A would be more permeable for convective heat transfer to occur, while in the high density FAs, the lower void volume could obstruct the migration of moisture and diminish the convective effect. Comparing the initial temperature rise at the face, intermediate and core layers among 24, 32, 40 FAs (Fig. 4-1, 4-2, 4-3), at 15 seconds PCT, the core layer has the fastest rate of temperature rise for 24 FAs under both 4.5% and 15% moisture conditions. The intermediate and face layers present little difference in temperature rise at 4.5% M C among the 24, 32, 40 FAs. These results seem to further support the above explanation of the density effect on convective heat transfer. Furthermore, the phenomena also indicate that the lower density FAs could produce a greater rate of heat transfer inside the FAs, while the higher density FAs, due to the improvement of conductive effect, could promote the rate of temperature rise near the surface of FAs. With the increase of M C in the face layer from 4.5% to 15% aud the decrease of press closing rate from 15 seconds to 45 seconds, different results were observed. A higher M C accelerated the convective effect and resulted in a faster rate of heat flow in the intermediate layer for all FAs. On the contrary, a slower PCT reduced convective heat transfer, but simultaneously enhanced conductive heat transfer and resulted in a more rapid rate of temperature rise in the face 48 and intermediate layers for 40, 32 FAs and a similar rate of temperature rise in the core layer for 24, 32, 40 FAs. hi a comparison of the internal conditions between a constituted mat and flakeboard mat, Xu (1993) measured the horizontal density distribution of a 291.6 X 291.6 mm flakeboard and reported that the density values ranged from 0.54 g/cm3 to 0.78 g/cm3. As described in the methodology chapter, the targeted densities of 24, 32 40 FAs in the constituted board were 0.51g/cm3, 0.68g/cm3 and 0.84g/cm3 respectively, thus they are in as a similar range of horizontal density variation compared to a flakeboard. However, as a result of the two different mat structures, where the FAs in the constituted mat contained little between flake voids compared to flakeboard mat, the mode of moisture migration between the two mats may have behaved quite differently. Moisture transfer inside the FAs was most likely via the capillary structure of wood, while moisture migration through a compressed flake mat tends to follow a path around wood particles (Smith, 1982). However, Dai (1993) used a probability-based model to predict that in the cold compressed flake mat (flake geometry: 0.8x6.0x37.6 mm; mat area: 152.4 mm2, between flake voids only occupied less than 5% of the total void volume at a compaction ratio (CR) of 1.5. This result thus suggested that in some localized portions of a flake mat, such as in high density regions, where no between flake void existed, moisture flow also had to pass through the inside flake void. Accordingly, in a flakeboard mat, some of the high density regions may have a similar rate of heat transfer to that observed in the higher density level FAs in the constituted mat, while low density regions, because of the occurrence of between flake moisture transfer, could expect a much faster rate of heat propagation toward the core compared to the temperature rise in the lower density FAs. Therefore, a larger extent of heat transfer variation in different density regions may exist in flakeboard than the results exhibited from the FAs in the constituted board. 49 4.1.2. Density Profiles of the FAs in the Constituted Board The results of the determination of wood density at 9% M C for microtomed flakes are listed in Appendix 1. The average value of flake density was 0.387 g/cm3. Results of density profiles for 24, 32 and 40 FAs pressed at 15 seconds, 45 seconds and 15% MC/face are shown in Figure 4-6, Figure 4-7 and Figure 4-8. FAs constructed with a greater number of flakes show higher overall density levels. The results also indicate that pressing conditions (15, 45 PCT; 4.5%, 15% MC/face) also have a slight effect on the density gradient, as expected, a higher density value was produced on the face layer from the faster PCT. In order to compare the density profiles to the results of flake columns compressed in the cold press, the CRs corresponding to density gradients were also calculated for each F A (Fig. 4-6, 4-7, 4-8). 50 •t fl 1200 900 750 600 3.100 2.713 300 0.0 Face 3.0 4.0 Thickness (mm) 5.0 6.0 Center 1 g 1050 750 600 450 300 0.0 Face 15 seconds P C T 4.5% M C 3.0 4.0 Thickness (mm) 5.0 3.100 2.713 2.325 .2 6.0 Center 0.775 1 •f a 1200 1050 900 450 300 0.0 Face 15 seconds P C T 15% MC/face 3.0 4.0 Thickness (mm) 3.100 2.713 2.325 o 6.0 Center Figure 4-6. Vertical density profiles of 24 FAs made at different pressing conditions. 51 1 900 750 600 300 0.0 Face 1.0 45 seconds P C T 4.5% M C 3.0 4.0 Thickness (mm) 5.0 3.100 2.713 2.325 .2 6.0 Center 1.163 0.775 1200 1050 ^ 9 0 0 1 e 750 •t a £ 600 300 0.0 Face 15 seconds P C T 4.5% M C 3.0 4.0 Thickness (mm) 5.0 3.100 2.713 2.325 .2 6.0 Center 1.163 0.775 1200 300 0.0 Face 3.0 4.0 Thickness (mm) 15 seconds P C T 15% MC/face 5.0 6.0 Center 3.100 2.713 2.325 O a 1.938 -2 1" 1.550 ° 1.163 0.775 ure 4-7. Vertical density profiles of 32 F A s made at different pressing conditions. 52 o.o Face 1.0 2.0 45 seconds P C T 4.5% M C 3.0 4.0 Thickness (mm) 5.0 3.100 6.0 Center 1.550 1.163 0.775 0.0 Face 2.0 15 seconds P C T 4.5% M C 3.0 4.0 Thickness (mm) 5.0 3.100 2.713 1.163 6.0 Center o.o Face 1.0 2.0 15 seconds P C T 15% MC/face 3.0 4.0 Thickness (mm) 5.0 3.100 2.325 O 6.0 Center 0.775 ure 4-8. Vertical density profiles of 40 FAs made at different pressing conditions. 53 4.2. Adhesive Analysis 4.2.1. Adhesive properties The chemical structures of the laboratory and commercial PF resin was characterized by their functional groups and molecular weight distributions. The IR spectrum of the laboratory PF resin is shown in Figure 4-9. According to standard peak assignments (Table 4-1), the distinct peaks corresponding to the methylol groups (1010 cm - 1 and 1270 cm - 1) are clearly demonstrated. The spectrum obtained from the laboratory resin is typical to resol resins. The results from GPC are summarized in Table 4-2. The number average molecular weight ( Mn) and weight average molecular weight ( M w ) obtained from the laboratory and commercial resins indicate that both resins are in the early stage of the resinification process. Likewise, the polydispersivities ( M w / Mn) also illustrate that the resins have less low proportions of molecular weight species in their corresponding resin rnixtures. The flow and cure properties of the laboratory and commercial powdered PF resin were investigated using T M A and DSC. The results are presented in Figure 4-10 and Figure 4-11 respectively. As it was heated, powdered PF resin experienced both physical and chemical changes. Physically, the powdered resin underwent through the rheological stages of softening, melting and hardening. Chemically, the PF resin went through a polymerization process, where the resin transformed from low molecular weight prepolymers to a highly branched, cross-linked, three dimensional network. As shown in Figure 4-10, softening of the resins started at approximately 65°C, followed by the melting of the adhesive showing a dramatic drop of the probe level. After the temperature reached 130°C, the resin then became infusible and little flow was observed. From Figure 4-11, the DSC thermograms demonstrate the cure characteristics of the powdered PF resin. Two overlapping exothermic peaks were observed in the temperature range of approximately 80 to 130°C and 130 to 180°C, respectively. It is know what the resin 54 Table 4-1. Some peaks assignments of functional groups in PF resin. Wavelength (cm l ) Assignment 3300-3400 O-H stretch 2900-2800 C-H stretch 1600-1500 C — C stretch 1450 C-Hbend 1270 C H 2 O-H bend 1230 ArOH stretch 1010 C H 2 O-H stretch 690 free phenol Table 4-2. Molecular weight distribution of acetylated laboratory and commercial resin. Laboratory resin Commercial resin Average value Standard deviation Average value Standard deviation 858 13.12 861 7.64 Mw 1839 89.05 2382 147.08 Mw/Mn 2.14 2.76 55 Figure 4-9. IR spectrum of the laboratory PF resin. 56 S a m p l e : P F RESIN S i z e : 2 . 8 9 9 0 mm Method : METHOD *1 TMA F i l e : STMA.001 O p e r a t o r : SONG Run D a t e : 3 - J u n - 9 4 22: 3 9 -100H 40 80 100 120 T e m p e r a t u r e l°C) 140 160 180 200 G e n e r a l V 4 . 1 C O u P o n t 2000 S a m p l e : P F RESIN S i r e : 2 . 5 2 3 0 mm Method : METHOO #1 TMA F i l e : S T M A . 0 0 3 O p e r a t o r : SONG Run D a t e : 4 - J u n - 9 4 10: 24 5 -isoH - 2 5 0 100 120 T e m p e r a t u r e (°C) 160 180 200 G e n e r a l V 4 . 1 C DuPont 2000 Figure 4-10. T M A profiles of the (a) commercial and (b) laboratory resins. 57 Sample : COMMERCIAL RESIN j-\ Q r~\ F i l e : S 0 N G 1 . 0 0 3 S i z e : 4 . 7 3 0 0 mg U O b O p e r a t o r : SONG M e t h o d : RAMPIO'C TO 250"C Run D a t e : 1 7 - 0 c t - 9 3 23: 36 0 So" 100 150 200 250 T e m p e r a t u r e (°C) G e n e r a l V 4 . 1 C DuPont 2000 S a m p l e : SYNTHETIC RESIN S i z e : 4 . 3 3 0 0 mg Method : RAMP 10*C TO 250 "C DSC F i l e : SONG2.002 O p e r a t o r : SONG Run D a t e : i a -0c t -93 01: 31 1.0 0 . 0 -- 0 . 5 H 50 100 150 T e m p e r a t u r e (*C) 200 250 G e n e r a l V 4 . 1 C DuPont 2000 Figure 4-11. DSC profiles of the (a) commercial and (b) laboratory resins. 58 polymerization process consists of addition and condensation reactions, where methylol phenols and ether or methylene bridged methylol phenols cross-linked structures are formed. According to reaction kinetic theory, a major condensation reaction should only take off as methylol groups in a resin mixture accumulate to a certain quantity. Therefore, the above two exothermic peaks observed are more likely to be the result of addition and condensation reactions respectively (Martin, 1956). Figure 4-11 shows that the addition reaction reached its maximum rate at the temperature of 130°C. This phenomenon indicates that at this temperature, resin cure was dramatically increased. Chow and Hancock (1969) pointed out that there was a critical temperature for the propagation of resin cure. However, the temperature they observed was at 110°C. This difference is probably because different PF resins were used, which might had different chemical structures, such as concentration of methylol groups. In terms of the relationship between adhesive cure and flow, at approximately 130°C, Figure 4-10 shows a transition of resin mobility from the fusible to the infusible state. This phenomenon thus verifies that the physical properties of a resin highly depend on the development of its cure phase. Additionally, the results indicated that corresponding to the major advancement of resin cure, there was a critical temperature for a major change in resin mobility. Accordingly, this critical temperature, like the secondary exothermic peak observed between 130°C to 180°C, which indicates the temperature level requirement for adhesive cure during the pressing, is also an important parameter that provides important information on its flow properties. By comparing the properties of the commercial and laboratory resins, the ER, GPC, DSC, T M A studies showed similar results for the two resins. The results thus demonstrated that a commercial type powdered PF resin can be made in the laboratory. No significant effect from this addition of 1% chlorazol black dye to the performance of the laboratory resin 59 was observed. Results listed in Table 3-2 shows that the commercial resin has a slightly higher polydispersivity ( M w / M n ) value than does the laboratory one. This phenomenon may indicate that the latter could have a smaller proportion of high molecular weight species compared to the former one. However, no such effect on the resin flow characteristics was found in the T M A study. 4.2.2. Flow of powdered adhesives in flakeboards 4.2.2.1. Simulated studies As indicated in section 2.2, adhesives in flakeboards are exposed to different environmental conditions as a result of both "vertical and horizontal effects" during conventional hot pressing. Figure 4-12 presents the effect of heating rates on resin flow of the powdered PF resin. The results show that a greater extent of flow is associated with a faster heating rate. This phenomenon suggests that the polymerization reactions occurred during the softening and melting stages of powdered resin, and subsequently influenced resin plasticity and flow. Under different heating rates, the physical transformation and chemical advancement of powdered resin could behave differently. A lower heating rate lets the adhesive gradually to undergo its physical changes and, in the meantime, allows an adequate time for polymerization reactions to take place. As a result, the progressive increase in resin molecular weight simultaneously reduces the resin mobility. On the contrary, at a faster heating rate, melting of resin could occur before sufficient curing takes place. Therefore, resin flow capacity is less affected by the progression of its polymerization and a larger extent of flow can occur. The influence of the pressing force level on resin flow is displayed in Figure 4-13. As expected, the magnitude of resin flow was increased at higher pressure levels. In addition, the resin flow was further characterized from combined pressing force and heating rate effects. Figure 4-14 illustrates that, at a higher heating rate, flow level was significantly 60 ure 4-12. Effect of heating rate on the flow of powdered PF resin. 61 Figure 4-13. Effect of pressing force on the flow of powdered PF resin. 6 2 Figure 4-14. Effect of pressing force and heating rate on the flow of powdered PF resin. 63 increased with the increase of pressure, while the increase of pressing force could only slightly improve the extent of resin flow. Since it has been found that different heating rates will yield different resin plasticity throughout the heating process, it is concluded that the effect of pressure on resin flow largely depends on the mobility of the resin. Blomquist (1981) indicated that the principal function of pressure in bonding is to force an adhesive to wet and flow to the bonding areas. At the higher heating rate, with the rapid increase of resin plasticity, pressure was able to make the resin flow to a greater extent. While at the lower heating rate, because of the low resin mobility, the extent of flow was improved only slightly with the increase of the pressure level. Considering the vertical flow of powdered adhesive inside a flakeboard, the face layers experience a higher heating rate compared to the core, a greater extent of resin flow would be induced in the face layers. In addition, these different heating rates in the face and core regions are associated with the same pressure changes (Figure 2-3). This means that, within the face layers, a higher pressure is applied to the adhesive associated with a higher temperature level. Whereas in the core, the highest pressure is firstly applied to the adhesive at a low temperature level. When the temperature in the core reaches high enough levels for the adhesive to melt, the pressure has already declined. According to the above, these conditions can further result in an even greater difference in the flow characteristics of the adhesive between the face and core than that caused by heating rates. Horizontally, higher density regions receive greater pressures but lower heating rates, while lower density regions receive lower pressures but exhibit higher heating rates, so no positive conclusion could be drawn from this simulated study. 4.2.2.2. Flow characterization in the flake assemblies The cross section of FAs pressed at different pressing conditions were microscopically examined. The results of glueline characteristics of 24, 32, 40 FAs are shown in Figure 4-64 i i 15, Figure 4-16, and Figure 4-17 respectively. At the same density level, the photographs show that adhesive flow occurred to a greater extent near the face than in the core. These results agreed with former T M A studies. By comparing the flow characteristics at the relevant positions at different density levels (40, 32, 24 FAs), higher density FAs generally appeared to have a thinner glueline than the lower density ones did. This is believed to be caused by the higher pressure that 40 or 32 FAs received. There is no clear evidence to show a difference in adhesive flow between FAs pressed under different pressing conditions. However, a slightly greater flow was observed for higher moisture content FAs presumably due to the moisture effects. The embedding and poh'shing technique used in this study was found to be a useful method to detect the adhesive flow characteristics in flakeboards. It is know that a flakeboard contains a considerable amount of deformed wood structure and these amounts of deformation tend to spring back when subjected to a high moisture condition. Since this technique does not require water soaking in the preparation of microscopy specimens, the unaltered wood and adhesive structure can be obtained and observed. However, because embedded specimen can only be examined by light microscopy under reflected light conditions at low magnification level detailed information of resin flow, such as resin penetration, could not be obtained. In this study, a variety of microscopic techniques was tried. In an attempt to quantitatively assess flow, the adhesive flow was examined under fluorescence microscopy in association with an image analysis system The observed sections of 2 pim in thickness were sliced from small pieces of samples embedded in epoxy resin using a glass knife on a microtome. It was found that the amount of fluorescence obtained was not sufficient to be detected by the image analysis system due to the short "exposure time" that the frame grabber used (1/60 second). A scanning electron microscopy (SEM) technique was also tried. Samples were briefly water soaked and surfaces for observation were cut with a sledge microtome. Although adhesive Figure 4-15. Cross section view of 24 FAs. 66 Figure 4-16. Cross section view of 32 FAs. 67 Figure 4-17. Cross section view of 40 FAs 68 was tagged with the black dye, poor contrast between wood adhesive was still encountered. Ellis (1989) used a bromine labeling technique to determine the resin penetration. The adhesive penetration was successfully measured. This chemical label technique appears to be a more suitable method to precisely determine adhesive flow in wood. 4.3. Internal Strength Properties 4.3.1. Bonding Strength The tension-shear strengths of 40, 32, 24 FAs at the face, intermediate and core layers are summarized in Appendix 2 and the value of each test cell is graphed in Figure 4-18. Analysis of variance (ANOVA) was performed based on split-plot design for this 33 factorial experiment. Three-way and two-way interactions were firstly examined in the sub-plot. It was concluded that the two-way interactions, density and layer position, were sigriificant at the 95% confidence level. Therefore, further A N O V A were carried out on individual layers i.e., face, intermediate and core. The results are listed in Appendix 3 and Appendix 4, respectively. Additionally, in order to test the significance of the difference of bonding strength for the FAs at different layers, density levels and pressing conditions, average strength values were analyzed and ranked using the Tukey multiple comparison test. The results are summarized in Tables 4-3, 4-4 and 4-5. Comparisons of the tension-shear strengths among the face, intermediate and core layers of the FAs pressed within the same density level and pressing conditions as shown in Figure (4-18). The results show that strength values increase from the core, through the intermediate layer to the face layer. Between the intermediate and face layers, a significant improvement in bonding strength (a=0.05) was also observed. Brady (1989) indicated that layered bonding strengths were closely related to layered density and pressing 69 Table 4-3. Comparison of average tension-shear strength at the different layers in the FAs. Average tension-shear strength (MPa) Assembly type (Density) Layer Pressing condition 15PCT 45PCT 15%MC 40FA Face 4.65 a 4.69 a 5.19 a Intermediate 3.87 b 3.90 b 4.39 b Core 3.55 c 3 . 5 9 ° 3.58 c 32FA Face 4.50 a 4.37 a 4.79 a Intermediate 3.38 b 3.20 b 3.26 b Core 3.05 b 2.99 b 3.14 b 24FA Face 4.43 a 4.34 a 4.55 a Intermediate 2.63 b 2.84 b 2.66 b Core 2.49 b 2.67 b 2.59 b Within a cell, means designated by the same letter are not significantly different at the 95% confidence level. 70 Table 4-4. Comparison of average tension-shear strength of the FAs with different density levels. Average tension-shear strength (MPa) Pressing Condition Assembly type (Density) Face Intermediate Core 15PCT 40FA 4.65 a 3.87 a 3.55 a 32FA 4.50 a 3.38 b 3.05 b 24FA 4.43 a 2.63 c 2.49 c 45PCT 40FA 4.69 a 3.90 a 3.59 a 32FA 4.37 a 3.20 b 2.99 b 24FA 4.34 a 2.84 c 2.67 b 15%MC 40FA 5.19 a 4.39 a 3.58 a 32FA 4.79 a b 3.26 b 3.14 a 24FA 4.55 b 2.66 c 2.59 b Within a cell, means designated by the same letter are not significantly different at the 95% confidence level. 71 Table 4-5. Comparison of average tension-shear strength of FAs pressed at different conditions. Average tension-shear strength (MPa) Assembly type (Density) Pressing Condition Face Intermediate Core 40FA 15%MC 5.19 a 4.39 a 3.58 a 15 PCT 4.65 b 3.87 b 3.55 a 4 5 PCT 4.69 b 3.90 b 3.59 a 32FA 15%MC 4.79 a 3.26 a 3.14 a 15 PCT 4.50 a 3.38 a 3.05 a 45PCT 4.37 a 3.20 a 2.99 a 24FA 15%MC 4.55 a 2.66 a 2.59 a 15 PCT 4.43 a 2.63 a 2.49 a 45PCT 4.34 a 2.84 a 2.67 a Within a cell means designated by the same letter are not significantly different at the 95% confidence level. 72 Figure 4-18. Tension-shear strength of 40, 32, 24 FAs at the face, intermediate and core layers. 73 temperature conditions. Figures 4-1, 4-2, 4-3 demonstrate different temperature conditions, and temperature induced differences in compaction ratio (CR) (Figures 4-6, 4-7, 4-8) between the intermediate and face layers. Likewise, the higher strength values observed in the face layer can be attributed to the heat transfer gradient and layered density levels. Additionally, for intermediate and core layers, 24 FAs and 32 FAs only provided a shghtly superior strength value in the intermediate layer than that in the core layer (Figure 4-18, Table 4-3), even through the intermediate layer experienced much earlier temperature rise than that in the core layer. However, when the density of FAs increased from 32 F A to 40 FA, a greater difference in CR was observed between the intermediate and the core layers (Figure 4-8). Consequently, a significant increase (a =0.05) in bonding strength was also observed in the intermediate layer (Table 4-3). This result verifies that wood and adhesive bonds were not only controlled by the pressing temperature conditions, but were also dependent on the degree of C K With an increase in the density of the FAs from 24, 32 to 40 FAs, at 4.5% flake moisture content, it was observed that wood and adhesive bonds at different layers responded differently. For the intermediate and core layers, Figure 4-18 shows that a greater FAs' density always produced stronger bonds than did lower density FAs, even though the temperature transfer was shghtly faster in the lower density FAs (Figure 4-1). In addition, Table 4-4 shows that the increases were significant at 95% confidence level except for the core layer at 45 second PCT. Accordingly, those results indicate that density had a considerable effect on the quality of wood and adhesive bonds for the internal layers in the FAs. In the case of face layers, higher density FAs only produced a shghtly superior strength value and no significant difference was observed (Figure 4-18, Table 4-4). As mentioned in the previous section, a great extent of resin flow could take place in the face layer. The above result suggests that the flow of resin in the face layers was capable of providing adequate wood/adhesive contact to form a stable wood/adhesive bond, and 74 further compression of wood element could not significantly improve bonding strength. Nevertheless, for the intermediate and core layers, since the flow of adhesive occurred to a lesser extent due to the delayed temperature rise and the associated declined pressure, a higher CR could increase the magnitude of wood/adhesive bonding area and subsequently enhance the bonding strength. Furthermore, these phenomena also imply that the extent of wood/adhesive bonding area is a critical parameter in the determination of bonding quality, particularly when low resin loading levels are applied. Table 4-5 and Figure 4-18 also illustrate the effect of pressing conditions on the layered bonding strength in the FAs. Press closing time showed little influence on wood/adhesive bond for the FAs. However, moisture content exhibited a positive effect on bonding properties at the face layers. At a higher moisture content, a upward shift of strength values was observed in the face layer for 40 and 32 FAs. Ellis (1989) indicated that a higher moisture condition can promote the extent of adhesive flow in wood. As a consequence, the above results may be attributed to a higher degree of adhesive flow. However, since this phenomenon was only found in the face layer and for higher density FAs, where damage of wood structure could occur (Geimer et al., 1985), the results may also indicate that high moisture content may slow down the strength reduction of wood during the pressing. Considering the wood/adhesive bonding in flakeboards, the vertical bonding strengths at the different layers in the FAs show similar increasing trends as in the case of flakeboards (Kelly, 1977). Bonding strength generally increased from the core to the face regions. Horizontally, FAs provided similar strength values in the face layer, but varied strength values in the intermediate and core layers depending upon different density regions. These results may suggest that, in flakeboard, bonding strength may show less variation in the face regions, while a larger variation in the middle regions of a board may occurs. 75 However, since flakeboards have a very different flake fabrication pattern from that in the FAs, the extent of variation on bonding strength in flakeboard may not be generated from the results in the FAs. From the foregoing, it could be readily concluded that pressing temperature and panel density positively influenced the inter-flake bonding. Temperature effects appeared to be closely related to their associated pressure levels, from which controlled the adhesive flow and subsequently detennined bonding quality. While density effects were to provide additional bonding area for wood and adhesive when adhesive flow had occurred insufficiently. 4.3.2. Failure Surface Observations After microscopically examining all the failure surfaces for the FAs, four basic failure types could be differentiated (Appendix 2): A: Cross fiber failure; B: Wood and adhesive failure near the interface; C: Wood/adhesive failure at the interface; D: Adhesive failure The typical types of failure surface are shown in Figure 4-19. Cross fiber failure was characterized by a large percentage of wood failure and fractures usually cross wood fiber and gluelines. Adhesive failures were observed where failure surfaces were covered by adhesive exposing little wood. Interface failures were recognized in the FAs' fracture surface as failure occurring both in wood and adhesive at then boundary layer. The failure surfaces with deeper areas of wood failure were designated as failure type B, while the fracture surfaces exhibiting shallower wood failure were designated as failure type C. (A) Cross fiber failure (B) Wood and adhesive failure near the interface Figure 4-19. View of the F A s ' fracture surfaces of different failure types (C) Wood/adhesive failure at the interface (D) Adhesive failure Figure 4-19. View of the FAs' fracture surfaces of different failure types (continue). 78 The purpose of distinguishing the failure type for the FAs was an attempt to identify the failure zones in wood and adhesive bonds that formed under different conditions. Figure 4-20 and Figure 4-21 graphed the number of failure types according to layered position and density (24FA, 32FA, 40FA) of the FAs respectively. The results showed little presence of adhesive failure. Therefore, they indicate good wood/adhesive bonds. Figure 4-20 shows the difference in failure mode among the face, intermediate and core layers. No significant effect from the layered position was observed. The face layer presented a large number of type A failure and the intermediate and core layers showed more proportion of interface failure. A trend of failure types occurred in 24, 32, 40 FAs is clearly demonstrated in Figure 4-21. Failure type A dominated the failure of 40 FAs, while failure type C principally occurred in 24 FAs. These phenomena indicated that higher density FAs have a relatively lower cohesion strength in wood, and lower density FAs have a smaller adhesion strength between wood and adhesive. Geimer (1985) reported that a flakeboard with high density could cause some damage of wood structure and result in a strength reduction. Accordingly, the higher proportion type A failure occurred in 40 FAs may also because the high CR damaged the structure of wood. The above result may also indicate that failure modes in the FAs are somewhat density dependent. On the other hand, this result would suggest that, in order to improve the wood and adhesive bond strength, it may only be required to reinforce strength property in specific bonded area according to a certain CR level. 70 Figure 4-20. Frequency of failure type occurring for FAs at the face, intermediate and core layers. Note: A : Cross fiber failure B: Wood and adhesive failure near the interface C: Wood/adhesive failure at the interface D : Adhesive failure so Figure 4-21. Frequency of failure type occurring for 40, 32 and 24 FAs. Note: A : Cross fiber failure B. Wood and adhesive failure near the interface C: Wood/adhesive failure at the interface D: Adhesive failure 81 4.4. Dimensional Stability 4.4.1. Wood deformation and "spring or swell" back The deformation process perpendicular to the grain direction of wood flakes was characterized using a stress and strain relationship (Figure 3-1). The linear and non-linear stress and strain curve illustrated three distinct regions: AB-linear; BC-nonlinear; CD-curvilinear. The typical deformed wood structures at the compaction ratio of 1.09, 1.54 and 4.00 corresponding to the three representative strain levels of 8.5, 35, 75 are illustrated in Figure 4-22. The results show that, in the AB-elastic range, there was little change in the structure of wood flakes, while in the BC range, it was observed that wood cells started to buckle and damage of cell walls upon vessel cells occurred (Figures 4-23, 4-25). In the CD range, most of wood cells were collapsed and the crashing of the cell wall of the fiber cells was apparently observed (Figures 4-24, 4-26). The study indicates that the changes in wood structures are associated with different mechanical responses during compression process. Gibson and Ashby (1988) studied the compression of balsa wood in the tangential direction. Their results also showed a similar stress-stain relation. The microscopic observations made at different stages of deformation also illustrated a series of compressed wood structures. In order to characterize dimensional properties of compressed wood according to different levels of wood structure deformation, the elastic and non-elastic spring back and water soaking thickness swelling were studied at the typical CRs of 1.09, 1.54 and 4.00. The elastic and non-elastic properties of compressed wood are illustrated in Figure 4-27. It was found that both instantaneous and delayed spring back were increased with the increase in CR, indicating the increase of wood dimensional instability at a higher compression level. However, Figure 4-27 also shows that the increase in the instantaneous spring back was less going a CR of 1.09 to 1.54 and it was going from a CR 1.54 to 4.00. This phenomenon is probably because the fact that at a higher CR level, a Figure 4-23. Cross sections of deformed wood structure at the compaction ratio o f 1.54 (x 500). 83 Figure 4-24. Cross sections of deformed wood structure at the compaction ratio of 4.00 (x 600). Figure 4-26. Collapse of cell wall in the fiber cells (x 3000). 85 Figure 4-27. Instantaneous and delayed springback at the different compaction ratios. 86 greater damage was induced in wood and it simultaneously, the plasticity of the wood was increased. The thickness swelling of compressed wood has reversible and irreversible components. The reversible thickness swelling is due to dimensional changes of the cell walls, while the irreversible thickness swelling is a result of the swell back of the cell lumens (Stamm and Cohen, 1956). Figure 4-28 summaries the thickness swelling of deformed wood and the proportion between reversible and irreversible components. The results show that the thickness swelling values are considerably increased for wood pressed to higher CRs. Similarly, the irreversible thickness swelling also increased proportionally with the increase in total thickness swelling. This result shows that, at a higher CR, the thickness swelling was largely composed to the balloon back of the cell lumens. However, the degree of increase for the irreversible thickness swelling was reduced from CRs of 1.09-1.54 to CRs of 1.54-4.00. The above results provide initial information on the relationships between deformed structures of wood and their associated dimensional properties. The relationships differ when wood is compressed under different heat and moisture conditions. Kunesh (1961) and Gerard (1966) investigated the effects of temperature and moisture conditions on spring back and thickness swelling of wood respectively. They reported that higher pressing temperature levels and moisture contents could reduce the spring and swell back for compressed woods. Since flakeboards are pressed in the presence of heat and moisture, a further study on the relationships between structure and dimensional stability of wood compressed at different environmental conditions is recommended. 87 Figure 4-28. Thickness swell of compressed flakes at the different compaction ratios. 88 4.4.2. Layered thickness swell of flake assemblies The thickness swell after 24 hours water soaking and subsequent redrying of 40, 32, 24 FAs at the face, intermediate and core layers are summarized in Appendix 5. Analysis of variance (ANOVA) was performed for the thickness swell values based on split-plot design. Three-way and two-way interactions were firstly examined in the sub-plot. It was observed that the two-way interaction between layered position and density, was significant at the 95% confidence level for 24 hour thickness swell. All the two-way interaction between layered position, density and pressing conditions were significant at the 95% confidence level for redryed thickness swell. The results are listed in Appendix 6 and Appendix 7 respectively. Additionally, in order to test the significance of the difference of 24-hour thickness swell values for the FAs at different layers, density levels and pressing conditions respectively, average thickness swell values were analyzed and ranked using Tukey multiple comparison test. The results were summarized in Table 4-6, 4-7 and 4-8. Figure 4-29 showed the average thickness swell values for the face, intermediate and core layers after 24 hours water soaking. Comparing thickness swell values between the face, intermediate and core layers of the FAs pressed within the same density level and pressing condition, the results show that, except for 40 FAs pressed at 15% moisture condition, thickness swell values were usually higher for the outer layers of the FAs. As Figures 4-6, 4-7 and 4-8 show the outer layers generally had a greater CR than the inner ones did due to the early temperature rise. The higher thickness swell values observed for the face and intermediate layers are thus believed to result from their greater degree of compression. However, for the 40 FAs and 24 FAs, particularly for the 40 FAs pressed at 15.0% moisture condition, a smaller incremental increase in thickness swell from the intermediate to face layers or from the core to the intermediate layers was noticed. According to the 89 Table 4-6. Comparison of average thickness swell at the different layers in the FAs. Average thickness swell values (%) Assembly type (Density) Layer Pressing condition 15PCT 45PCT 15%MC 40FA Face r~ 81.98 a 80.01 a 58.94 a Intermediate 73.10 a 74.44 a 55.64 a b Core 49.89 b 53.88 b 44.34 b 32FA Face 67.36 a 65.73 a 54.65 a Intermediate 38.93 b 36.00 b 29.84 b . Core 23.05 c 24.48 b 20.67 b 24FA Face 34.28 a 36.76 a 34.64 a Intermediate 8.22 b 7.33 b 7.61 b Core 6.41 b 5.73 b 5.50 b Within a cell, means designated by the same letter are not significantly different at the 95% confidence level. 90 Table 4-7. Comparison of average thickness swell of the FAs with different density levels. Average thickness swell values (%) Pressing Condition Assembly type (Density) Face Intermediate Core 15PCT 40FA 81.98 a 73.10 a 49.89 a 32FA 67.36 b 38.93 b 23.05 b 24FA 34.28 c 8.22 c 6.41 c 45PCT 40FA 80.01 a 74.44 a 53.88 a 32FA 65.73 a 36.00 b 24.48 b 24FA 36.76 b 7.33 c 5.73 c 15%MC 40FA 58.94 a 55.64 a 44.34 a 32FA 54.65 a 29.84 b 20.67 b 24FA 34.64 b 7.61 c 5.50 c Within a cell, means designated by the same letter are not significantly different at the 95% confidence level. 91 Table 4-8. Comparison of average thickness swell of FAs pressed at different conditions. Average thickness swell values (%) Assembly type (Density) Pressing Condition Face Intermediate Core 40FA 15%MC 58.94 a 55.65 a 44.34 a 15PCT 81.98 b 74.44 b 53.88 a 45PCT 80.01 b 73.10b 49.89 a 32FA 15%MC 54.65 a 29.84 a 20.67 a 15PCT 67.36 a 38.93 a 23.05 a 45PCT 65.73 a 36.00 a 24.48 a 24FA 15%MC 34.64 a 7.61 a 5.50 a 15 PCT 34.28 a 8.22 a 6.41 a 4 5 PCT 36.76 a 7.33 a 5.73 a Within a cell, means designated by the same letter are not significantly different at the 95% confidence level. 92 Figure 4-29. Thickness swell of 24 hours water soaking for 40, 32, 24 FAs at the face, intermediate and core layers. 93 statistical analysis, Table 4-6 also shows that no significant difference (a=0.05) was produced between the intermediate and face layers for 40 FAs or between the core and intermediate layers for 24 FAs. Figures 4-6, 4-7 and 4-8 show that a greater degree of difference in CR between the intermediate and face layers for 24 FAs and 32 FAs was produced than that between the intermediate and face layers for 40 FAs or between the core and intermediate layers for 24 FAs. The cause of the above phenomenon thus might also be due to the difference in CR found among different layers. Furthermore, for the 40 FAs pressed at 15.0% moisture condition, thickness swell values were only slightly increased from the intermediate layer to the face layer. This phenomenon thus might imply that, at a higher compression level, higher moisture contents could improve the dimensional stability of deformed wood. The positive effect of moisture content on dimensional stability of wood was also reported by Gerard (1965). He studied the thickness swell of cross-plied veneer assembly pressed at different moisture conditions and found that greater swelling occurred in the outer plies of assemblies pressed at 4% moisture content than those pressed at 12 percent. Back (1987) discussed the bonding mechanism in hardboard and indicated that moisture, functioning as a plasticizer, could reduce the glass transition temperature for Ugnin, cellulose and hemicellulose and subsequently affect polymer bonds such as hermceUulose-hgnin bonds, in the wood fibers,. Hse (1989) also contributed the dimensional properties of steam injected flakeboards to the plastic flow of the lignin and the partial hydrolysis of hemicellulose. Considering thickness swell among 24, 32 to 40 FAs at the face, intermediate and core layers, higher density FAs generally resulted in greater values (Figure 4-17). The results from the statistical analysis also showed significant increases in thickness swell from lower density FAs to higher density FAs, except for the face layer produced at 45 second PCT and 15% moisture content conditions between 32 FAs and 40 FAs (Table 4-7). Those results reflect the positive influence of density on thickness swell. Additionally, it was 94 also found that, FAs pressed at the face layer, 15% moisture content, resulted in thickness swell for the 40 FAs was only shghtly higher than that in 32 FAs (Figure 4-17). This phenomenon suggests that, at a higher temperature and moisture condition, an increase in compression level would not significantly affect the thickness swell properties. Meyer (1984) studied the manufacturing methods for heat-stabilized compressed wood (staypak), a wood product used to make tool handles and table legs. He reported that at the pressing conditions of 160°C, 1500-2500 psi, and about 12% moisture content, the compressed wood could gain a relatively stable state. The press closing time showed little influence on thickness swell values for the various FAs. However, moisture condition, again, illustrates its significant effect on the dimensional stability of compressed wood on the face layer with higher density FAs (32 F A and 40 FA) (Figure 4-27, Table 4-8). This observation further supports the above discussions. Figure 4-30 summarizes the average thickness swell values after redrying for the face, intermediate and core layers. Thickness swell after subsequent redrying generally followed the same variation trends to those observed in the water soaking conditions, where outer layers or higher density FAs general had a greater thickness swell value. In addition, it is interesting to note that moisture again exhibited a strong influence on thickness swell values for the 40 FAs and 32 FAs at the face layer (Table 4-8). This result also supports the idea that the flow of lignin could be occurring in the face layer of high density FAs during the pressing thus resulting in a rather stable deformed wood structure. From the above study, it was found that both density and layered position in FAs iivfluence thickness swell and that the moisture effect tended to be significant under high temperature and pressure condition, such as in the face layer of 40 FAs. Suchsland (1989) 95 Face Intermediate Core Layer - 15-40 - 15-32 - 15-24 o - 45-40 o - 45-32 A 45-24 - M C - 4 0 - M C - 3 2 - M C - 2 4 Figure 4-30. Thickness swell of redrying for 40, 32, 24 FAs at the face, intermediate and core layers. 96 indicated that dimensional stability of flakeboards is controlled by higher density regions. Since this study showed that a high moisture condition could improve dimensional properties at the outer layers for higher density FAs, it is logical to conclude that a higher flake initial moisture content may be able to improve the dimensional stability of flakeboards. 97 5. SUMMARY A N D CONCLUSIONS This work has, for the first time, systematically evaluated the localized properties for flakeboards. The FAs used as a structural model to simulate pressing conditions in flakeboards were shown to be a reasonable approach in the analysis of their internal properties. Consequently, this study provides initial information on internal temperature conditions; localized adhesive flow; layered bonding properties and layered dimensional stability for flakeboards and leads to following findings: 1. Density affects heat transfer rate, where lower density FAs resulted in faster rates of heat flow in the core layers, while higher density FAs tended to have an more rapid temperature rise near the surface. 2. Pressure and heating rate affect adhesive flow and greater adhesive flow occurs at a more rapid heating rate. 3. Bond strength values were higher for the outer layers in FAs. Density positively affected bonding strength at the intermediate and core layers. However, at the face layer, increase of compaction ratio did not significantly improve the bonding strength values. 4. Density of FAs affects their wood and adhesive failure modes. 5. The balloon back of the cell lumens for compressed wood is the main reason for its large amount of "water soaking" thickness swelling. 6. At the same layer in the FAs, thickness swell values were higher for greater density FAs, and within a FA, outer layers generally had greater thickness swell values. 7. Moisture facilitated heat transfer, particularly for in the lower density FAs. A higher face moisture content improved the bonding strength at the face layer for 40 FAs and greatly reduced the thickness swell at the face and intermediate layers for 40 FAs. 5.1. Remarks The objectives of this work were not merely to study the localized properties for a flakeboard, but also to study the fundamental processes involved in the flakeboard 98 manufacture and address their implications to the panel physical and mechanical properties. In flakeboard manufacture, wood elements are compressed under pressure; heat and moisture. This compression of wood is required in order to produce an intimate wood-adhesive-wood contact for qualified wood/adhesive bonding (Back, 1987). But, on the other hand, it is also found that the compression of wood simultaneously brings some "inherent defects" to flakeboard properties. Dimensionally, flakeboards are not stable in the thickness direction due to the tendency of releasing compression stress introduced during the press (Halligan, 1970; Alberta Research Council, 1987). Although a large number of research has been done to study the bonding properties and dimensional stability of flakeboards, in order to improve the board performance, there is a still lack of fundamental knowledge on those two important panel properties. The results from compression studies showed that water soaking thickness swell of compressed wood was mainly contributed to the balloon back of the cell lumens. This finding may indicate that for an improvement of dimensional properties there needs to be a decrease in the compression level of wood. However, a certain amount CR is needed, between 1.6-1.8 (Back, 1987), to produce good wood/adhesive bonding. Additionally, it is also found that at the intermediate and core layers of FAs, bonding strength and thickness swell values were increased with higher density levels of FAs, while at the face layer, the increase of CR did not improve the bonding strength, and a high moisture content improved bonding strength and reduced thickness swell values for 40FAs. These results also demonstrated the bonding and dimensional properties deformed wood at different pressure, temperature and moisture conditions. Therefore, the above findings might be helpful to improve flakeboard properties. For further research, to optimize properties of composites products, the key studies should also be directed towards the fundamental properties of wood and adhesive under different pressure, heat, and moisture conditions. 99 ) 6. LITERATURE CITED Alberta Research Council. 1987. Dimensional stabilization state of the art review FP 2.4.1. Alberta Research Council. Industrial Technologies Department. Forest Products Program Edmonton, Alberta. Back, E.L. 1987. The bonding mechanism in hardboard manufacture. Holzforschung 41 (4):247-258. Blomquist, R. F. 1981. Adhesive~an overview. In adhesive bonding of wood and other structure materials. Vol. III. Madison, Wisconsin. 436pp. Bodig, J and B.A. Jayne. 1982. Mechanics of wood and wood composites. Van Nor strand Reinhold company. New York. 712pp Bolton, A.J. and P.E. Humphrey. 1988. The hot pressing of dry-formed wood-based composites. Part I. A review of the literature, identifying the primary physical process and the nature of their interaction. Holz. 42(4): 403-406. Bolton, A.J. and P.E. Humphrey. 1989. The hot pressing of dry-formed wood-based composites. Part IQ. Predicted vapor pressure and temperature variation with time, compared with experimental data for laboratory boards. Holz. 43(4):265-274. Brady, D.E. 1987. The effect of hot-pressing parameters on resin penetration and flakeboard layer properties. M.Sc. thesis. Forestry and Forest Products. Virginia Polytech. Inst, and State University. 122pp. Brady, D.E. and F.A. Kamke. 1988. Effects of hot-pressing parameters on resin penetration. For. Prod. J. 38(11/12): 63-68. Bryant, B.S. 1977. Wood adhesion: hi handbook of adhesives. 2nd edition; Edited by Irving Skeist. Von Nostand Reichold Co., Inc., N.Y. Chow, S.Z. and W.V. Hancock. 1969. Method for detennining degree of cure of phenolic resin. For. Prod. J. 19(4): 21-29. 100 Dai, C P . 1993. Modeling structure and processing characteristics of a randomly-formed wood-flake composite mat. Ph.D. dissertation. Department of Wood Science. University of British Columbia. Ellis, S. 1989. Some factors affecting the flow and penetration of powdered phenolic resin into Wood. Ph.D. Dissertation. Department of Wood Science. University of British Columbia. 213pp. Ellis S.C. and P.R. Steiner. 1991. Charaterization of chemical properties and flow parameters of powdered phenol-formaldehyde resins. Wood and Fiber Sci. 23(1): 85-97. Ellis, S. 1993. The performance of waferboard bonded with powdered phenol-formaldehyde resins with selected molecular weight distributions. For. Prod. J. 43(2): 66-68. Ellis, S. 1993. The efficiency of powdered phenol-formaldehyde waferboard adhesfves. Wood and Fiber Sci. 25(3): 214-219. Geimer, R.L., R.A. Follensbee, A.W. Christiansen, J.A. Koutsky and G.E. Myers. 1990. Resin characterization. Prepared for publication in 24th International Washington State University Particleboard/composites Materials Symposium. 320pp. Geimer, R.L., R.J. Mahoney, S.P. Loehnertz and R.W. Meyer. 1985. Influence of processing-induced damage on strength of flakes and flakeboards. Res. Pap. FPL 463. Madison, WI: US. 14pp. Gerard, J.C. 1966. Dimensional behavior of particles in simulated particleboard constructions. For. Prod. J. 16(6): 40-48. Gibson, L.J. and M.F. Ashby. 1988. Cellular solid: structure and properties. Pergamon Press. N. Y. 357 pp. Halligan, A.F. 1970. A review of thickness swelling of particleboard. Wood Sci. Tech. 4: 301-312. Harada, H. , G.W. Davies and K.F. Plomley. 1968. Preliminary macroscopic studies of wood structure and adhesion in plywood. For. Prod. J. 18(2): 240-244. 101 Harless, T.E.; Wagner, F.G.; Seale, R.D.; Mitchell, P.H. and D.S. Ladd. 1987. A model to predict the density profile of particleboard. Wood Fiber Sci. 19(1): 81-92. Hsu, W.E., W. Schwald, J. Schwald and J.A. Shield. 1988. Chemical and physical changes required for producing dimensionally stable wood-based composites. Part 1. Steam pretreatment. Wood Sci. Technol. 22: 281-289. Kawouras, P.K. 1977. Fundamental process variables in particleboard manufacture. Ph.D. dissertation. University of Wales. U.K. Kamke, F.A. and L.J. Casey. 1988. Fundamentals of flakeboard manufacture: internal mat conditions. For Prod. J. 38(6):38-44. Kelly, M.W. 1977. Critical literature review of relationships between processing parameters and physical properties of particleboard. USDA Forest Ser. Gen. Tech. Report. FPL-10. Forest Prod. Lab., Madison, WI. 64pp. Kunesh, R.H. 1961. The inelastic behavior of wood: a new concept for improved panel forming processes. For. Prod. J. 11(9): 395-406. Maku, T.R. Hamada and H. Sasaki. 1959. Studies on the particleboard. Rept. 4. Temperature and moisture distribution in particleboard during hot-pressing. Wood Research, Kyoto Univ. 21:34-46. Marra, A.A. 1964. Proceedings of the conference on the theory of wood adhesion. University of Michigan. Ann Arbor. Mich. Martin, R.W. 1956. The chemistry of phenolic resins. John Wiley and Sons Inc., New York. 288pp. Meyer, J.A. 1984. Wood-polymer materials. The Chemistry of Solid Wood. American Chemistry Society, Washington, D.C. 614pp. Nearn, W.T. 1965. Wood adhesive interface relations. J. Paint Technol. 37:720-733. 102 Plath, L. and E. Scknitzler. 1974. The density profile, a criterion in evaluating particleboard. Holz Roh-Werkst. 32(11): 443-449. Rackwitz, G. 1954. Ein Beitrag zur Kenntnis der Vorgaei der Verleimung von Holzspan zu Holzspanplatten in beheizten hydrauhschen Pressen. Dissertation, Technische Hochschule Braunschweig, 121 S. Rauch, W. 1984. Temperatur-und Dampfdruckverlauf bei der Herstellung von Spanplatten und ihr Einfluss auf die technologischen Eigenschaften. HolzRoh. Werk. 42: 281-286. Parardy, R.D., B.A. Haataja, S.M. Shaler, A.D. Williams and T.L. Laufenberg. 1989. Pressing of wood composite panels at moderate temperature and high moisture content. For. Prod. J. 39(4): 27-32. Shen, K.C. and M.N. Carroll. 1969. Measurement of layer-strength distribution in particleboard. For. Prod. J. 20(6): 53-55. Smith, D.C. 1982. Waferboard press closing strategies. For. Prod. J. 32(3): 40-45. Stamm, A.J. and W E . Cohen. 1956. Swelling and dimensional control of paper. Australian Pulp and Paper Industry Technical Association Proc. No. 10. Steiner, P .K and C P . Dai. 1993. Spatial structure of wood composites in relation to processing and performance characteristics. Part 1. Rational for model development. Wood Sci. Technol. 28: 45-51. Steiner, P.R. and S.R. Warren. 1981. Rheology of wood-adhesive cure by torsional braid analysis. Holzforschung. 35: 273-278. Strickler, M.D. 1959. Effect of press cycles and moisture content on properties of Douglas-fir flakeboard. For. Prod. J. 7:203-215. Suchsland, O. 1962. The density distribution in flake board. Michigan Quarterly Bulletin. 45(1): 104-121. 103 Suchsland, O. 1967. Behavior of a particleboard mat during the press cycle. For. Prod. J. 17(2): 51-57. Suchsland, O. 1973. Hygroscopic thickness swelling and related properties of selected commercial particleboard. For. Prod. J. 23(7): 26-30. Suchsland, O. and H. Xu. 1989. A simulation of the horizontal density distribution in a flakeboard. For. Prod. J. 39(5): 29-33. Suchsland, O. andH. Xu. 1991. Model analysis of flakeboard variables. For. Prod. J. 41(11/12): 55-60. Wilson, J.B. and R.L. Krahmer. 1976. Particleboard: microscopic observations of resin distribution and board fracture. Forest Prod. J. 26(11): 42-45. Wolcott, M.P. 1990. Modeling viscoelasitic cellular materials for the pressing of wood composites. Ph.D. dissertation. Department of Wood Science and Forest Products, Virginia Polytechnic Institute and State University. Xu, W. 1993. Horizontal density distribution of particleboard: origin and implications. Ph.D. dissertation. Department of Wood Science. University of British Columbia. 166pp. Youngs, R.L. 1957. The perpendicular-to-grain mechanical properties of red oak as related to temperature, moisture content, and time. U.S. FPL. Report No. 2079. 104 Appendix 1. Flakes Density Measurment Weight(g) Thickness(min) Density(g/cm3) 0.232 0.682 0.360 0.256 0.691 0.392 0.269 0.702 0.406 0.271 0.703 0.408 0.263 0.702 0.396 0.256 0.707 0.384 0.228 0.684 0.353 0.230 0.691 0.353 0.258 0.692 0.395 0.276 0.710 0.411 0.261 0.689 0.400 0.247 0.683 0.382 0.252 0.708 0.377 0.274 0.712 0.408 0.259 0.683 0.402 0.260 0.715 0.385 0.254 0.673 0.400 0.281 0.712 0.418 0.258 0.683 0.399 0.234 0.682 0.363 0.259 0.708 0.387 0.261 0.706 0.392 0.230 0.690 0.352 0.260 0.691 0.398 0.258 0.716 0.381 0.230 0.688 0.353 0.256 0.682 0.396 0.251 0.685 0.387 0.255 0.693 0.389 105 30 0.261 0.695 0.397 31 0.232 0.684 0.359 32 0.252 0.680 0.392 33 0.273 0.700 0.413 34 0.255 0.684 0.395 35 0.251 0.694 0.382 36 0.244 0.678 0.381 37 0.231 0.692 0.354 38 0.245 0.692 0.374 39 0.264 0.694 0.403 40 0.257 0.693 0.392 41 0.260 0.690 0.399 42 0.255 0.688 0.392 43 0.257 0.685 0.397 44 0.245 0.700 0.371 45 0.258 0.695 0.393 46 0.255 0.685 0.393 47 0.234 0.688 0.360 48 0.265 0.702 0.399 49 0.256 0.690 0.392 50 0.248 0.695 0.377 Average 0.254 0.693 0.387 Standard 0.013 0.010 0.017 deviation 106 Appendix 2. Shear Strength and Failure Mode for the Flake Assemblies Press Flake Specimen Face Intermedate Core Conditions No. 15PCR 40FA 1541 1543 1545 Shear- Failure Shear-Strength Mode Strength (MPa) (MPa) 4.52 A 3.94 4.65 A 3.81 4.79 A 3.87 Failure Shear- Failure Mode Strength Mode (MPa) A 3.68 A B 3.51 B A 3.47 A 32FA 1541 4.68 A 3.55 D 3.33 D 1543 4.72 B 3.38 B 2.96 B 1545 4.10 B 3.21 B 2.86 B 24FA 1541 4.55 B 2.75 C 2.44 C 1543 4.49 D 2.47 B 2.70 B 1545 4.25 C 2.68 C 2.34 C 45PCR 40FA 4541 4.76 A 3.93 A 3.57 A 4543 4.43 B 4.05 A 3.43 A 4545 4.88 A 3.71 A 3.76 A 32FA 4541 4.17 B 3.26 A 2.98 A 4543 4.46 A 3.15 A 2.92 A 4545 4.49 A 3.18 B 3.05 B 24FA 4541 4.49 B 2.99 B 2.99 B 4543 4.37 B 2.79 C 2.48 C 4545 4.15 B 2.74 B 2.53 B M C 40FA mcl mc3 mc5 32FA mcl mc3 mc5 24FA mcl mc3 mc5 5.19 A 4.34 5.24 B 4.60 5.14 A 4.2.5 4.67 A 3.28 5.11 A 3.08 4.60 B 3.43 4.42 B 2.48 4.93 A 2.79 4.29 B 2.70 A 3.38 A A 3.61 A A 3.75 A B 3.62 B B 2.97 B B 2.85 B C 2.72 C A 2.44 A B 2.61 B 1 0 7 Appendix 3. Analysis of Variance for Tension-shear Strength Source Sum of Squares DF Mean Square F Ratio Main-plot COND$ 0.565 2 0.283 6.902 n.s. COND$*R$ 0.165 4 0.041 DENS$ 11.365 2 5.683 145.718 * DENS$*R$ 0.155 4 0.039 COND$*DENS$ 0.420 4 0.105 2.442 n.s. COND$*DENS$*R$ 0.342 8 0.043 Sub-plot POSTS 36.423 2 18.211 337.241 * POST$*R$ 0.214 4 0.054 COND$*POST$ 0.306 4 0.077 2.200 n.s. COND$*POST$*R$ 0.277 8 0.035 DENS$*POST$ 2.046 4 0.512 9.309 * DENS$*POST$*R$ 0.439 8 0.055 POST$*COND$*DENS$ 0.333 8 0.042 1.448 n.s. ERROR 0.462 16 0.029 Note: "*" means significant at 95% confidence level "n.s." means not significant at 95% confidence level 1 0 8 Appendix 4. Analysis of Variance for Tension-shear Strength at Individual Layers Source Sum of Squares DF Mean Square F Ratio Face COND$ 0.736 2 0.368 7.006 * DENSS 0.788 2 0.394 7.506 * COND$*DENS$ 0.149 4 0.037 0.708 n ERROR 0.945 18 0.052 Intermediate COND$ 0.111 2 0.056 2.581 n DENSS 8.211 2 4.105 190.324 * COND$*DENS$ 0.542 4 0.135 6.277 * ERROR 0.388 18 0.022 Core CONDS 0.025 2 0.012 0.251 n DENSS 4.412 2 2.206 44.580 * COND$*DENS$ 0.063 4 0.016 0.317n ERROR 0.891 18 0.049 Note: "*" means significant at 95% confidence level "n.s." means not significant at 95% confidence level 109 Appendix 5. Thickness Swell After 24 hours Water Soaking and Subsequent Redrying for the Flake Assemblies Press Flake Specimen Thickness Swell (%) Condition No. Face Intermediate Core 24hrs Water After 24hrs Water After 24hrs Water After Soaking Redrying Soaking Redrying Soaking Redryin 15PCT 24FA 1542 32.87 21.38 8.72 4.51 6.39 1.51 1542 33.18 18.18 6.28 3.89 4.28 2.28 1544 30.89 18.89 9.18 4.42 7.87 4.25 1546 40.19 18.11 8.70 3.13 7.09 3.61 32FA 1542 68.35 45.66 43.22 21.57 22.79 9.09 1542 74.03 52.34 33.04 23.54 22.70 13.62 1544 61.79 42.76 36.36 22.79 27.27 13.70 1546 65.28 43.01 43.10 25.43 19.45 11.46 40FA 1542 82.87 47.94 62.44 45.77 49.95 36.39 1542 87.39 54.55 83.52 50.64 53.91 40.92 1544 80.08 49.73 78.18 51.34 43.51 34.94 1546 77.73 47.25 68.25 49.55 52.21 34.94 45PCT 24FA 1542 34.76 21.66 9.00 4.53 4.28 2.13 1542 34.63 21.66 5.53 4.53 5.69 3.33 1544 37.56 20.77 7.38 4.41 4.10 3.23 1546 40.07 23.36 7.41 5.54 8.83 4.42 32FA 1542 76.05 48.68 32.17 23.48 21.95 13.12 1543 73.68 47.63 45.34 31.26 33.45 16.07 1544 52.49 42.96 32.27 19.70 17.53 13.25 1546 60.70 43.04 34.23 22.67 25.00 16.67 40FA 1542 90.59 52.01 73.80 52.01 46.32 31.07 1543 83.33 50.00 63.16 45.45 50.00 33.33 1544 72.49 45.26 74.03 47.92 56.62 39.22 1546 73.64 47.75 86.77 47.73 62.56 48.21 15%MC 24FA 1542 37.39 16.65 4.23 2.23 4.23 2.22 1543 30.35 17.25 8.83 4.42 4.25 2.23 1544 40.45 .22.79 8.33 4.45 9.09 4.09 1546 30.35 17.38 9.03 4.41 4.42 2.42 32FA 1542 67.65 25.67 30.25 25.98 21.66 8.69 1543 46.61 29.54 33.43 26.73 22.79 13.70 1544 53.41 28.85 26.67 18.46 17.38 12.97 1546 50.94 31.12 29.00 20.90 20.82 12.56 40FA 1542 52.14 37.25 52.34 36.10 44.55 30.98 1543 50.07 35.48 59.56 42.14 49.61 38.12 1544 60.83 34.35 54.23 41.67 43.22 36.07 1546 72.73 45.45 56.42 39.76 39.98 34.30 I l l Appendix 6. Analysis of Variance for 24-hours Thickness Swell Source Sum of Squares DF Mean Square F Ratio Main-plot COND$ 1535.42 2 767.71 42.14* COND$*R$ 109.33 6 18.22 DENS$ 40283.70 2 20141.80 556.05 * DENS$*R$ 339.43 6 56.57 COND$*DENS$ 961.90 4 240.47 5.74 * COND$*DENS$*R$ 502.95 12 41.91 Sub-plot POSTS 18022.90 2 9011.48 305.47 * POST$*R$ 177.00 6 29.50 COND$*POST$ 272.39 4 68.10 1.72 n.s. COND$*POST$*R$ 475.49 12 39.628 DENS$*POST$ 2312.78 4 578.19 54.96 * DENS$*POST$*R$ 126.21 12 10.52 . POST$*COND$*DENS$ 201.68 8 25.21 0.67 n.s. ERROR 1021.37 27 37.83 Note: "*" means significant at 95% confidence level "n.s." means not significant at 95% confidence level 112 Appendix 7. Analysis of Variance for Redryed Thickness Swell Source Sum of Squares DF Mean Square F Ratio Main-plot CONDS 5717.5 COND$*R$ 56.19 DENS$ 20381.5 DEN.S$*R$ 86.3 COND$*DENS$ 1585.2 COND$*DENS$*R$ 51.1 2 6 2 6 4 12 2858.7 9.3 10190.7 14.2 396.3 4.2 305.1 * 716.6 * 93.0 *. Sub-plot POSTS 596.6 POST$*R$ 53.7 COND$*POST$ 331.9 COND$*POST$*R$ 141.6 DENS$*POST$ 207.3 DENS$*POST$*R$ 75.4 POST$*COND$*DENS$ 251.7 ERROR 282.1 2 6 4 12 4 12 8 27 298.3 8.9 82.9 11.8 51.8 6.2 31.4 10.4 33.3 * 7.0 * 8.2 * 3.0 * Note: "*" means significant at 95% confidence level "n.s." means not significant at 95% confidence level 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075222/manifest

Comment

Related Items