UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Radio frequency heating pre-treatment of sub-alpine fir to improve kiln drying Abubakari, Alhassan 2010

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

Item Metadata

Download

Media
24-ubc_2010_fall_abubakari_alhassan.pdf [ 5.19MB ]
Metadata
JSON: 24-1.0071223.json
JSON-LD: 24-1.0071223-ld.json
RDF/XML (Pretty): 24-1.0071223-rdf.xml
RDF/JSON: 24-1.0071223-rdf.json
Turtle: 24-1.0071223-turtle.txt
N-Triples: 24-1.0071223-rdf-ntriples.txt
Original Record: 24-1.0071223-source.json
Full Text
24-1.0071223-fulltext.txt
Citation
24-1.0071223.ris

Full Text

RADIO FREQUENCY HEATING PRE-TREATMENT OF SUB-ALPINE FIR TO IMPROVE KILN DRYING  by  Alhassan Abubakari  B.Sc., Kwame Nkrumah University of Science and Technology, Ghana, 2007  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE DEGREE OF  MASTER OF SCIENCE in The Faculty of Graduate Studies (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2010  © Alhassan Abubakari, 2010  ABSTRACT The objective of the study was to reveal the effect of RF heating at different power densities and time combinations as pre-conventional kiln treatment on the drying characteristics and quality of sub-alpine fir lumber. As a consequence of this objective, the study hypothesis was formulated as: “if RF heating improves the permeability of subalpine fir, then upon kiln drying, final moisture content variability between and within lumbers as well as drying defects will decrease”.  In this research, thirteen groups of one meter long, 51 x 102 mm sub-alpine fir lumbers were RF heated with two power densities (27 and 72 kW/m3) and eight time combinations (15, 30, 45, 60, 75, 90, 105 and 120 minutes), but all to the same final temperature level (100oC) before kiln drying in a two-phase experimental design. Two groups (one for each phase) served as controls. Permeability tests were also performed in the second phase of the study. The effect of RF heating on permeability, drying rates, moisture content gradient, final moisture content variability between pieces and drying defects was evaluated and analyzed. Treatment effect on total energy consumption (sum of RF heating and kiln drying energies) was also assessed to ascertain the feasibility of applying this technology in industrial setting.  Data analysis revealed in phase 1 that, not all treatments reduced moisture content variability within and between specimens and improved drying rates above and below fiber saturation point, compared to the control. Defect appearance also did not significantly reduce, and all except two treatments increased total energy consumption. In phase 2 however, permeability improved in all except one treatment but was found not to be statistically significant. Treatments also improved moisture gradient as well as drying rates above and below fiber saturation point but the moisture gradient was found not to be statistically significant. Compared to the control, not all treatments reduced moisture content variability between wood samples. Defects did not significantly improve between treatments and control, and total energy consumption was relatively higher in treatments. Results obtained within the limitations of this study led to the rejection of the hypothesis.  ii  TABLE OF CONTENTS ABSTRACT……………………………………………………………………………....ii TABLE OF CONTENTS………………………………………………………………...iii LIST OF TABLES………………………………………………………………………..v LIST OF FIGURES……………………………………………………………………..vii ACKNOWLEDGEMENTS………………………………………………………………x 1. INTRODUCTION ..........................................................................................................1 2. LITERATURE REVIEW ...............................................................................................5 2.1 Characteristics and uses of sub-alpine fir ................................................................5 2.2 Structure of wood related to drying .........................................................................7 2.3 Water in wood..........................................................................................................9 2.4 Wetwood................................................................................................................11 2.4.1 Characteristics and causes of wetwood ...........................................................11 2.4.2 Problems associated with drying wetwood......................................................13 2.5 Wood permeability and methods of improvement ................................................14 2.6 Dielectric heating of wood.....................................................................................17 2.6.1 Factors affecting dielectric properties of wood ...............................................19 2.7 Reasons for drying wood .......................................................................................21 2.8 Mechanism of wood drying ...................................................................................22 2.9 Stages of wood drying ...........................................................................................24 2.10 Drying defects......................................................................................................25 2.11 Conventional kiln drying of wood .......................................................................27 3. MATERIALS AND METHODS .................................................................................29 3.1 Specimen preparation ............................................................................................29 3.2 Radio frequency heating treatments ......................................................................30 3.3 Kiln drying experiments ........................................................................................33 3.4 Post-drying cutting and measurements ..................................................................35 3.5 Permeability experiments ......................................................................................37 3.6 Data analysis……………………………………………………………………...39 4. RESULTS AND DISCUSSION...................................................................................41 4.1 Green moisture content and basic density .............................................................41 iii  4.2 Heating rate (dT/dt) and moisture changes in samples during radio frequency heating……………………………………………………………………………45 4.3 Effect of radio frequency heating on the longitudinal air permeability.................56 4.4 Effect of radio frequency heating on drying rates .................................................60 4.5 Effect of radio frequency heating on final moisture content variability between samples after kiln drying………………………………………………………....67 4.6 Effect of radio frequency heating on moisture content variability within wood samples after kiln drying………………………………………………………...72 4.7 Wood quality after radio frequency heating and kiln drying.................................75 4.8 Energy consumption in radio frequency heating and kiln drying..........................83 4.9 Hypothesis test.......................................................................................................86 5. CONCLUSION.............................................................................................................87 6. RECOMMENDATIONS FOR FUTURE RESEARCH ..............................................89 7. REFERENCES .............................................................................................................91  iv  LIST OF TABLES Table 3.1: Radio frequency treatments in phase one and two…………………………...31 Table 3.2: Drying schedule for drying experiments……………………………………..34 Table 3.3: Analysis of variance of completely randomized experimental design……….40 Table 4.1: Moisture contents of the eighteen groups before RF heating………………...41 Table 4.2: Analysis of variance of average moisture content before RF heating………..43 Table 4.3: Basic densities of the eighteen groups before RF heating…………………...43 Table 4.4: Analysis of variance of average basic density before RF heating……………45 Table 4.5: Average heating rate of RF heating groups…………………………………..46 Table 4.6: Analysis of variance of average moisture loss during RF heating…………...49 Table 4.7: Pair-wise comparison of average moisture loss during RF heating………….49 Table 4.8: Analysis of variance of average moisture content of the main groups before kiln drying………………………………………………………………………………..51 Table 4.9: Moisture contents of the fifteen main groups before kiln drying…………....52 Table 4.10: Moisture contents of the fifteen sub-groups before kiln drying……………52 Table 4.11: Analysis of variance of average moisture content before RF heating of permeability test samples………………………………………………………………...56 Table 4.12: Longitudinal air permeability of groups in phase 2………………………...57 Table 4.13: Analysis of variance of average longitudinal air permeability of groups in phase 2…………………………………………………………………………………...58 Table 4.14: Pair-wise comparison of average longitudinal air permeability of groups in phase 2…………………………………………………………………………………...59 Table 4.15: Average drying rate of all drying runs………………………………………61 Table 4.16: Average drying rate above and below fiber saturation point for all drying runs………………………………………………………………………………………65 Table 4.17: Moisture contents of main groups after kiln drying………………………..68 Table 4.18: Analysis of variance of average final moisture contents of main groups…...69 Table 4.19: Moisture contents of sub-groups after kiln drying………………………….70 Table 4.20: Core-shell moisture content differences of main groups after kiln drying….72 Table 4.21: Core-shell moisture content differences of sub-groups after kiln drying…...74 Table 4.22: Analysis of variance of average core-shell moisture content differences…..75 Table 4.23: Pair-wise comparison of average core-shell moisture content differences of sub-groups in phase 1…………………………………………………………………….75 Table 4.24: Number of samples with surface checks in sub-groups……………………..77 Table 4.25: Average honeycomb count of sub-groups…………………………………..81 v  Table 4.26: Analysis of variance of average honeycomb count in the main groups…….82 Table 4.27: Pair-wise comparison of average honeycomb count of sub-groups in phase 2………………………………………………………………………………………….82 Table 4.28: Pair-wise comparison of average honeycomb count of main groups……….82 Table 4.29: Energy consumption in RF heating…………………………………………83 Table 4.30: Energy consumption in kiln drying…………………………………………84 Table 4.31: Total energy consumption in RF heating and kiln drying…………………..86  vi  LIST OF FIGURES Figure 1.1: Canadian exports of SPF to China from 2004 to 2008 (Statistics Canada Computer Data)....................................................................................................................2 Figure 2.1: The sub-alpine fir tree (a) and the distribution(dark section of map) of subalpine fir in British Columbia (b) (BC Forestry Innovation Investment) ………………...5 Figure 2.2: Aspirated and encrusted pit membranes on tracheid cell wall in wetwood (left) and in normal wood of sub-alpine fir (right) (Zhang et al. 2006, with permission from Springer Science and Education Media)…………………………………………….6 Figure 2.3: Principal features of a tree stem (Illustration by Marianne Markey; Reprinted with permission from Identifying wood by R. Bruce Hoadley, Published in 1990 by the Taunton Press).....................................................................................................................7 Figure 2.4: Pit cross sections: (a) Bordered pit (b) simple pit; and (c) half-bordered pit (Simpson 1991, with permission from USDA Forest Service, Forest Products Lab. Madison, WI)……………………………………………………………………………...9 Figure 2.5: The electromagnetic spectrum (Anonymous)……………………………….18 Figure 2.6: Theoretical drying rate curve for a hygroscopic solid material (Rosen 1983……………………………………………………………………………………...25 Figure 2.7: Schematic illustration of different types of warp (Simpson 1991, with permission from USDA Forest Service, Forest Products Lab. Madison, WI)…………...26 Figure 2.8: Schematic illustration of conventional dry kiln (Simpson 1991, with permission from USDA Forest Service, Forest Products Lab. Madison, WI)…………...28 Figure 3.1: Procedure for cutting drying samples from original lumber pieces………...29 Figure 3.2: Determination of wood volume by the displacement method……………….30 Figure 3.3: Arrangement of wood samples and sensor positioning during RF heating in phase 1 (left) and phase 2 (right)………………………………………………………...32 Figure 3.4: RF oscillator (left) and oven (right) with wood samples surrounded by an MDF box………………………………………………………………………………....33 Figure 3.5: The dry kiln showing the arrangement of wood samples……………………34 Figure 3.6: Post-drying cutting procedure in phase 1……………………………………35 Figure 3.7: Procedure for removing shell and core parts from wood sections for moisture content determination in phase 1………………………………………………………...36 Figure 3.8: Procedure for marking wood sections after drying in phase 2………………36 Figure 3.9: Post-drying cutting procedure in phase 2……………………………………37 Figure 3.10: Longitudinal air permeability specimen…………………………………...37 vii  Figure 3.11: Schematic illustration of falling-method apparatus………………………...38 Figure 4.1: Average moisture content of the eighteen groups before RF heating (error bars indicate ±1SD)……………………………………………………………………....42 Figure 4.2: Average basic density of the eighteen groups before RF heating (error bars indicate ±1SD)…………………………………………………………………………...44 Figure 4.3: Temperature-time curve of RF120 in phase 1……………………………….46 Figure 4.4: Average moisture loss of the main groups after RF heating (error bars indicate ±1SD……………………………………………………………………………………..48 Figure 4.5: RF heating time and moisture loss relationship……………………………..48 Figure 4.6: Relationship between heating rate in sub-groups and moisture loss in main groups…………………………………………………………………………………….50 Figure 4.7: Relationship between Experimentally determined and calculated moisture loss……………………………………………………………………………………….50 Figure 4.8: Distribution of moisture content before and after kiln drying in phase 1…..53 Figure 4.9: Distribution of moisture content before and after kiln drying in phase 1…..54 Figure 4.10: Distribution of moisture content before and after kiln drying in phase 2….55 Figure 4.11: Average moisture content of groups before heating (error bars indicate ±1SD)…………………………………………………………………………………….56 Figure 4.12: Average longitudinal permeability of groups in phase 2 (error bars indicate ±1SD)…………………………………………………………………………………….57 Figure 4.13: Drying curves of drying runs in phase 1 (left) and phase 2 (right)………...60 Figure 4.14: Normalized drying curves of drying runs in phase 1 (left) and phase 2 (right)….............................................................................................................................60 Figure 4.15: Normalized drying rate curves of drying runs in phase 1………………….62 Figure 4.16: Normalized drying rate curves of drying runs in phase 1………………….63 Figure 4.17: Normalized drying rate curves of drying runs in phase 2………………….64 Figure 4.18: Average moisture content before and after kiln drying (error bars indicate ±1SD)…………………………………………………………………………………….68 Figure 4.19: Distribution of moisture content along the length of highest, medium and lowest moisture content samples in RF60, RF75 and RF120 in phase 2………………...71 Figure 4.20: Average core-shell moisture content differences of all drying runs (error bars indicate ±1SD)…………………………………………………………………………...73 Figure 4.21: Number of samples with surface checks after RF heating…………………76 Figure 4.22: Surface checking in RF120 (left) and RF15 (right) samples in phase 1…...78 Figure 4.23: Surface checking in RF120 (left) and RF15 (right) samples in phase 2…...78 viii  Figure 4.24: Honeycomb count in RF120 (left) and RF15 (right) in phase 1…………...79 Figure 4.25: Honeycomb in RF120 (left) and Control (right) in phase 2………………..80 Figure 4.26: Average honeycomb count of all drying runs (error bars indicate ±1SD)…81 Figure 4.27: Relationship between total RF heating time and energy consumption…….84 Figure 4.28: Relationship between total kiln drying time and energy consumption…….85  ix  ACKNOWLEDGEMENTS My first gratitude and appreciation goes to my supervisor Prof. Stavros Avramidis for his immense contribution, support and guidance.  I am also grateful to Prof. Phil Evans and Dr. Luiz Oliveira, members of my supervisory committee for their advice and fruitful discussions. My appreciation goes to Prof. Tony Kozak for his advice on statistical analyses.  Special thanks also go to FP Innovations-Forintek Division for allowing me to use their laboratory and equipment for this work. Thanks to Liping Cai, Vit Mlcoch and Diego Elustondo (all of FP Innovations-Forintek Division) for their assistance.  I would also like to thank Ken Watanabe, Jianfeng Zhan, Ning Guo and Ciprian Lazarescu for their help.  x  1. INTRODUCTION Wood as a natural resource has become an indispensable raw material both industrially and domestically. Due to the high moisture content of freshly cut wood there is always the need to remove substantial amounts of water through natural or artificial drying, in order to derive maximum benefits from the use of wood. This is because wood drying increases strength, dimensional stability, reduces attack by micro-organisms and makes liquid impregnation easier. However, optimal drying is barely achieved in some wood species because the internal pathways available for moisture movement are blocked due to abnormal wood characteristics, making these species difficult to dry. One such abnormality is wetwood which is very much prevalent in sub-alpine fir.  Sub-alpine fir (Abies lasiocarpa (Hook.) Nutt) belongs to the group collectively called SPF (spruce-pine-fir) that grows in the interior British Columbia (BC) and it is harvested, processed and sold as one species group. Knudson and Mcfarling (2009), report that BC produces about 32 million cubic meters of lumber per year, and the SPF group contributes about 26 million cubic meters. Sub-alpine fir contributes about two million cubic meters (10% percent) of the SPF production per year and British Columbia Forest Service (2003) report that this is expected to increase to about 30% of the SPF production in some regions because of the gradual lodgepole pine demise from the mountain pine beetle (Dendroctonus ponderosae) epidemic.  China is one of the importers of Canadian softwood lumber, of which SPF plays a major part and export statistics show a continuous upward trend of about 100% increase per year from 2005-2008 as shown in Fig. 1.1 (COFI 2009). These statistics indicate that much of the SPF exports to China are used in utility rather than structural applications in their vast housing market. Plans are therefore far advanced to improve the grade mix of SPF exports to China with measures aimed at increasing the capacity to use structural lumber, landscaping lumber and some specialty lumber (COFI 2009). With China’s economy improving at an unprecedented rate, this trend clearly reflects a bright future of SPF lumber of which sub-alpine fir is a part. 1  Figure 1.1: Canadian exports of SPF to China from 2004 to 2008 (Statistics Canada Computer Data) Sub-alpine fir is characterized by a high proportion of wetwood and low permeability, and therefore cannot be dried as rapidly as spruce and pine. Wetwood is heartwood tissue in standing trees which has been internally infused with water (Hartley et al. 1961). Wetwood has higher extractive content and has more aspirated and encrusted pit membranes on tracheid cell walls compared to normal wood, which makes it prone to reduced permeability and problematic drying (Cai and Oliveira 2008). The long drying times for sub-alpine fir currently causes kiln capacity problems since most mills are employing accelerated drying schedules to obtain higher productivity of SPF lumber (Zhang and Cai 2008). Moreover, these schedules result in wide final moisture content disparity with significant number of kiln-dried boards either over-dried or under-dried (wets). Furthermore, substantial internal moisture profiles result in high levels of stresses that greatly affect grade recovery during planing and remanufacturing (Zhang and Cai 2008). All these problems may be addressed in one way or another by using the appropriate drying technology, optimized drying schedules and/or pre-drying strategies (i.e., sorting, steaming, and heating). Green lumber heating convectively or by dielectric  2  means as a pre-treatment has now emerged as one possible technology to alleviate this problem.  Dielectric heating is a term that encompasses both radio frequency (RF) and microwave (MW) systems. It is a method that can be used to rapidly generate heat within a piece of lumber and considerably increase internal temperature and vapor pressure (Avramidis et al. 1994). Investigations have revealed that this technology can improve the permeability and drying characteristics of hitherto, less permeable and difficult to dry wood species. Torgovnikov and Vinden (2000a, 2000b, and 2000) improved the permeability of less permeable hardwoods using microwave radiation. Hong-Hai et al. (2005) also improved the permeability of larch wood by intensive microwave irradiation.  The underlying principle behind this technology is that by applying dielectric heating to wood containing substantial amount of moisture, the water molecules rapidly absorbs the electromagnetic radiation that is converted to kinetic energy and then heat. The water vaporizes quickly and this results in the generation of high steam pressure which ruptures wood cells and creates narrow voids, which results in an increase in the permeability of the wood (Hong-Hai et al. 2005). Vinden and Torgovnikov (2000) reported that microwave radiation can be used to substantially increase wood permeability to liquids and gases in both the radial and longitudinal directions. However, the drawbacks of microwave heating which include relatively lower depth of penetration and less uniform heating can reduce the efficiency and prolong the entire heating process.  The aforementioned drawbacks of microwave heating coupled with the seemingly bright future of sub-alpine fir make it imperative to explore other avenues to improve the permeability and drying characteristics of sub-alpine fir. A greater depth of penetration and more uniform heating of RF (Avramidis 1999) compared to microwave makes it the most promising technology in this regard. It therefore provides an alternative and more efficient dielectric heating technology to improve the permeability and drying characteristics of sub-alpine fir lumber. Heating and vaporizing the moisture in wet  3  pockets using RF could “open-up” the internal structure of sub-alpine fir and thus increase permeability, and at the same time improve kiln drying characteristics.  Today, no research has been conducted to investigate the potential of RF heating in improving the permeability and drying characteristics of less permeable wood species. The objective of this study therefore, is to reveal the effect of RF heating at different power densities and time combinations as pre-conventional kiln drying treatment, on the kiln drying characteristics and quality of sub-alpine fir lumber. Based on the presumption that RF heating is capable of improving the permeability and drying characteristics of sub-alpine fir, the hypothesis for this study is that if RF heating treatment methods improve the permeability of sub-alpine fir, then upon kiln drying final moisture content variability between and within lumber as well as drying defects will decrease.  4  2. LITERATURE REVIEW 2.1 Characteristics and uses of sub-alpine fir Sub-alpine fir (Abies lasiocarpa (Hook.) Nutt) belongs to the white fir group, consisting of grand fir (Abies grandis), white fir (Abies concolor), Pacific silver fir (Abies amabilis), California red fir (Abies magnifica), and noble fir (Abies procera) (Alden 1997). Subalpine fir is a mountain species found in British Columbia, Alberta, and the Yukon at altitudes of 610 to 2134 meters. It grows well on moist, well-drained loam soils, and in the interior it is associated with Douglas fir, lodgepole pine, and Engelmann spruce at lower elevations and with the limber and whitebark pines at the timber-line (Canada Department of Resources and Development, Forestry Branch 1950).  (a)  (b)  Figure 2.1: The sub-alpine fir tree (a) and the distribution (dark section of map) of sub-alpine fir in British Columbia (b) (BC Forestry Innovation Investment) Sub-alpine fir, as shown in Figure (2.1a) is a medium-sized tree usually 20 to 30 meters tall and occasionally grows to 50 meters. It has a distinctive long narrow crown of short stiff branches. The crown is grey and smooth, with resin blisters when young. However, the tree is easily infected by wood-rotting fungi between the ages of 120-140 years and dies (British Columbia Ministry of Forest and Range 2008).  5  Wood ranges from tan to brown with shades of red or pink, and the sapwood not clearly distinct from the heartwood. It has no distinctive odor or taste, and varies from very light, soft and weak to moderately heavy, hard and strong (Alden 1997). The mean basic density is reported as 329 kg/m3 (Smith 1970). Knudson et al. (2008), in their study of sub-alpine fir basic properties reported an average green moisture content of 73.8%. According to Alden (1997), sub-alpine fir wood is used for building construction, boxes, crates, planing mill products, sashes, doors, frames, food containers and as pulpwood.  Knudson and Mcfarling (2009) reported that sub-alpine fir currently contributes about two million cubic meters (10%) of the SPF production (26 million cubic meters) in BC per year. It is projected that this amount produced will increase significantly to about 30% of the SPF production within relatively short period of time in some regions. However, Zhang and Cai (2009) pointed out that sub-alpine fir wood is characterized by high proportion of wetwood (wet-pockets or wet-spots as also known) and aspirated pits, which causes difficulty in drying. Zhang et al. (2006) in their study of the anatomy of normal and wetwood of sub-alpine fir, found a higher proportion of aspirated and encrusted pit membranes on tracheid cell walls in wetwood compared to normal wood (Fig. 2.2). They reiterated that this condition has great influence on the migration of water from one cell to another, and is the cause of the problem of hard-to-dry of subalpine fir.  Figure 2.2: Aspirated and encrusted pit membranes on tracheid cell wall in wetwood (left) and in normal wood of sub-alpine fir (right) (Zhang et al. 2006 with kind permission from Springer Science and Education Media)  6  2.2 Structure of wood related to drying The cellular nature of wood influences its drying characteristics. A cross-section of a tree reveals the bark, light-colored sapwood zone next to the back, an inner and darker heartwood zone, and growth rings that show as concentric rings (Simpson 1983) as shown in Fig. 2.3. Each growth ring in softwoods consists of earlywood (springwood) and latewood (summerwood), formed at the beginning and later in the growing seasons respectively (Mackay and Oliveira 1989).  Figure 2.3: Principal features of a tree stem (Illustration by Marianne Markey; Reprinted with permission from Identifying wood by R. Bruce Hoadley, Published in 1990 by the Taunton Press) Each wood cell has a cell cavity called lumen and cell walls composed of several layers arranged in different ways. Most of the tube-like cells are oriented parallel to the long axis of the tree trunk and are termed fibers, tracheids or vessels, depending on their 7  anatomical characteristics and function (Simpson 1983). About 93% of softwood volume contains longitudinal tracheids and 1-6% is made up of longitudinal resin canals and wood rays (Panshin and de Zeeuw 1980). Running from the outermost wood towards the pith in the radial direction are narrow strands called rays (Mackay and Oliveira 1989). Most woods contain rays that function to restrain dimensional change in the radial direction, and also partially explain the fact that upon drying wood shrinks less radially than tangentially (Bowyer et al. 2007). Three major components can be found in the cellwall of wood, namely: cellulose, hemicelluloses and lignin (Koch et al. 1990). Cellulose provides most of the mechanical strength, hemicelluloses are the matrix and lignin binds the cells together (Siau 1995). Lignin is also credited with reducing dimensional change with moisture content fluctuation (Bowyer et al. 2007). The middle lamella has the greatest concentration of lignin compared to other parts of the wood (Philip et al. 2001).  According to Simpson (1983), one anatomical feature important in water flow is the pit. Kollman and Côté (1968) defined a pit as “a recess in the secondary wall of the cell, open to the lumen on one side and including the membrane closing the recess on the other side”. Pits can be simple (between parenchyma cells), bordered (between two prosenchyma cells such as two longitudinal or ray tracheids) or half-bordered (found between parenchyma and prosenchyma cells) (Fig 2.4). Bordered pits are the most important of the types of pits because they connect two conducting cells (Siau 1995).  Bordered pits may become aspirated in softwoods in some situations, and in this way the torus is displaced and covers the pit aperture, resulting in the inhibition of water flow and reduced drying rates (Simpson 1991). Pit aspiration reduces the permeability of wood and increases the difficulty of liquid impregnation (Siau 1995). Several reasons have been assigned to the causes of pit aspiration in wood. Hart and Thomas (1967) believed aspiration can be caused by drying as a result of differences in pressure that may develop on different sides of the pit membrane. Bowyer et al. (2007) however, indicated that aspiration happens as a result of liquid tensions that develop in standing trees or in processed wood that is being dried.  8  Figure 2.4: Pit cross sections: (a) Bordered pit (b) simple pit; and (c) half-bordered pit (Simpson 1991, with permission from USDA Forest Service, Forest Products Lab. Madison, WI) One characteristic feature important in wood drying is the difference between sapwood and heartwood (Mackay and Oliveira 1989). The heartwood is infiltrated with gums, resins, and other materials, making it more resistant to fluid flow than sapwood. It thus, has different drying characteristics (Simpson 1983). Heartwood varies between and within tree species and its behavior is believed to be related to growth rate, stand features and site conditions (Bamber and Fukazawa 1985). Some of the impacts of heartwood on wood drying include: reduction in moisture diffusivity and permeability, which permits slower drying if one is to avoid checking; reduced shrinkage due to bulking of the cell walls by extractives; large volumes of volatile emissions in high temperature drying that result in environmental pollution and corrosion of some kiln parts due to the release of wood acids (Keey et al. 2000). Bramhall and Wellwood (1976) indicated that heartwood dries more slowly than sapwood in most hardwoods. However, this situation is compensated for in softwoods by the lower moisture content of the heartwood, and both the sapwood and heartwood dry at almost the same rate.  2.3 Water in wood The determination of moisture content and knowledge of wood-liquid relations are very important because of the unique role these play in wood drying, impregnation, finishing 9  and bending (Kollman and Côté 1968). The green moisture content of wood varies considerably among kinds of trees, between heartwood and sapwood in the same tree, and logs cut from different heights of the tree (Skaar 1988). Moisture content is believed to be higher in sapwood of softwoods than heartwood (Kollmann and Côté 1968).  Siau (1995) defines the moisture content of wood as the mass of moisture in wood expressed as a percentage of the oven-dry mass (as shown in equation 2.1). Wood as a hygroscopic material loses or gains moisture until its moisture content is in balance with that of the atmosphere. The moisture content at this stage is called the equilibrium moisture content (MEMC), and for given exposure conditions MEMC is affected by species, heartwood and sapwood, extractive contents, previous exposure history, temperature and mechanical stress (Skaar 1988).  (2.1)  M = (w-wo)/wo *100%  where M is the moisture content (%), w is the mass of moist wood (g), and wo is the ovendry mass of wood (g).  The moisture in wood occurs in two forms: free water in the cell cavities and bound water absorbed on the cell wall structure (Kollmann and Côté 1968). The moisture content at which the cell walls are saturated with bound water but no free water exists in the cell cavities is termed the fiber saturation point (MFSP) (Tiemann 1906). The fiber saturation point falls between 25 and 32% for most species of wood. The fiber saturation point is very significant in wood drying because more energy is required to remove water from the cell wall than the cell cavities, wood cell wall will not shrink until it reaches the MFSP, and large changes in many physical and mechanical properties of wood begin to take place at the MFSP (Simpson 1991). Siau (1995) believes that bound water is hydrogen bonded to the hydroxyl groups in cellulose and hemicelluloses, and to a lesser extent to that in lignin. He went on to say that bound water content is limited by the number of sorption sites available and by the 10  number molecules of water which can be held on a sorption site. The maximum moisture content (Mmax) that can be found in a living tree is the total amount of both free and bound water (Siau 1995), and according to Keey et al. (2000) this can be calculated from the basic density ρb (kg/m3), density of water, ρw (kg/m3), and density, ρc (kg/m3) of ovendry cell tissue as shown in equation 2.2:  Mmax = (1- ρb/ ρc) (ρw/ ρb).  (2.2)  2.4 Wetwood Wetwood is a term used to describe anomalous zones in wood with very high moisture content compared to the neighboring wood (Chafe 1996, Cooper and Jeremic 1998). Two types of wetwood can be generally recognized. Sap-transition sapwood, heavy wetwood or heavy sinker heart is formed from dying sapwood that has been invaded by bacteria. The second type called wet heart or light sinker heart forms in normally developed heartwood by invasion of bacteria from nearby wetwood (Ward and Zeikus 1980). The presence of wetwood in softwood dimension lumber slows drying times and result in losses of energy and higher manufacturing costs (Ward and Pong 1980). Kabir et al. (2006) also found that wetwood has a considerable effect on the physical and mechanical properties of wood.  2.4.1 Characteristics and causes of wetwood Wetwood has characteristically different physical, biological and chemical properties from normal wood. According to Ward and Pong (1980) wetwood has a water-soaked translucent appearance in some species, while in others the wetwood has the appearance of normal heartwood or unusual dark color. Wetwood is also generally higher in moisture content compared to the neighboring normal wood. The darker coloration can be ascribed to bacteria metabolized dark polyphenolic compounds. Another prominent feature of wetwood is the distinct odor (Cooper and Jeremic 1998). Wetwood can be differentiated  11  from normal wood by their vinegar-like odor. This odor is attributed to rancid acetic and fatty acids produced by bacteria in wetwood (Kabir et al. 2006).  Wetwood is slow to dry and kiln or air dried lumber often contains wet-pockets because it has low permeability (Hartley et al. 1961). Ward and Pong (1980) however, pointed out that the absorptive capacity of wetwood can be greater than that of normal wood. Several reasons have been assigned to the low permeability of wetwood. Wetwood usually has high extractive content than sapwood and heartwood (Schroeder and Kozlik 1972), and these can cause encrustation on aspirated pits which reduces the permeability in heartwood of conifers (Bramhall and Wilson 1971, Comstock 1965, Resch and Eckland 1964). Aspiration of bordered pits in softwoods (Kozlik et al. 1972, Lin et al.1973, Ward and Kozlik 1975) and tyloses in hardwoods (Kemp 1959, Knutson 1968, Ward and Shed 1979) have all been implicated in the reduced permeability of wetwood. Bacteria slime can also block cells and contribute to reduced permeability (Ward and Pong 1980). Wetwood has a considerable effect on the physical and mechanical properties of wood (Kabir et al. 2006). Kozlik (1970) found the specific gravity of wetwood based on the green volume to be greater than normal heartwood, and Schroeder and Kozlik (1972) attributed this to the higher extractive content found in wetwood. There are conflicting reports regarding the strength properties of wetwood compared to normal wood (Ward and Pong 1980). Ward and Pong (1980) however, reported that wetwood is usually weaker than normal wood in bonding strength of the compound middle lamella between wood cells. This they attributed to the production of pectin-degrading enzymes by bacteria that degrade the pectic substance in the compound middle lamella that holds together the wood cells. Xu et al. (2001b) reported high radial and tangential shrinkage and lower tension strength perpendicular to the grain for wetwood of red oaks.  Although, Cooper and Jeremic (1998) indicated that there are no known accepted causes of wetwood, several authors have outlined probable causes of this defect in wood. According to Ward and Pong (1980) wetwood in trees can be attributed to several causes, including microbial (bacterial), non-microbial (injury) and normal age-growth. The  12  assertion that wetwood formation is caused by non-microbes is held by Bauch et al. (1975) and Coutts and Rishbeth (1977). This point was also buttressed by Hartley et al. (1961) who indicated that wetwood is associated with physical, mechanical and biological injuries. However, Hadden et al. (2004) are of the view that wetwood is attributed to bacteria infection in the sapwood and heartwood of the tree. Infection is normally associated with wounding or environmental stress on the wood. Species of bacteria in the genera Enterobacter, Klebsiella, and Pseudomonas have all been implicated in the cause of wetwood (Franc 1994). Chafe (1996) was of the opinion that the excretions of bacteria maybe hypertonic, and attract water from adjacent non-infected areas and result in the abnormally high moisture content wetwood.  2.4.2 Problems associated with drying wetwood As stated previously due to the high moisture content of freshly cut wood, there is always the need to kiln-dry wood to obtain maximum benefits from its use. The objective of a kiln operator is to achieve this feat as fast as possible with minimum defects and moisture content variation between and within lumber. This objective is barely achieved in most cases because of inherent abnormal characteristics such as wetwood in wood which retards moisture movement during drying. Wetwood increases drying time and result in energy losses as well as higher manufacturing costs (Ward and Pong 1980). Several investigators have assigned different reasons to the longer drying times of wetwood. Lihra et al. (2000) attributed this to the occluding encrustations in the bordered pits of wetwood tracheids. Passialis and Tsoumis (1984) and Wilcox and Schlink (1970) believed that slow drying is caused by aspiration of the pits. Ward and Pong (1984) further reiterated that after reaching the desired moisture content in drying, wetwood lumber still have uneven moisture where the shell is very dry but the core contains wet pockets that are still above fiber saturation point. Ward and Shedd (1981) also indicated that most wetwood lumber are above the required moisture content after drying, and usually they are separated from normal lumber in the  13  dry chain and re-dried. However, this re-drying operation can increase kiln drying costs by 40% or more.  Besides the longer time to dry wetwood, it is also highly prone to drying defects. As reported by Ward (1984), drying rates are retarded in wetwood boards and they are prone to develop a variety of defects, namely, deep surface checks, honeycombing, ring failure, collapse, and chemical brown stain. In wetwood drying, checks can develop in both radial and tangential directions and radial checks may be deep surface checks, internal ruptures called honeycomb, or bottleneck checks that develop into honeycomb (Ward and Pong 1980). Enzymes produced by bacteria degrade hemicelluloses and pectin present in the middle lamella of the cell wall which develops abnormal honeycombing and checking in wood (Kabir et al. 2006). Wetwood is also believed to have greater shrinkage than normal wood (Wilcox 1968, Ward 1984).  2.5 Wood permeability and methods of improvement Siau (1995) defined permeability as a measure of the ease with which fluids move through a porous medium under the influence of a pressure gradient. The measurement of permeability is important to the wood industry because of its relationship with drying, pulping, and the preservative treatment of wood (Bradic and Avramidis 2007). As a result of the importance of this wood parameter, several studies have been conducted over the years to measure the permeability of numerous wood species. These studies have shown wide variation in permeability results emanating from several factors inherent in wood. Rice and D’Onofrio (1996) revealed that the most obvious factors affecting the permeability are the anatomical structure and the chemical constituents of the wood. However, they indicated that the effect of the chemical constituents is only important in the comparison of permeability between sapwood and heartwood. In terms of the effect of anatomical structure on permeability Stamm (1967) indicated that softwood permeability depends on the interconnection of tracheid lumens by pit membranes. Stamm (1932) was of the opinion that the greater permeability of sapwood than heartwood might be due to far less aspiration of the pits and clogging of the membranes 14  with resin in sapwood than in heartwood. Panshin and de Zeeuw (1980) also pointed out that in conifers the openings in the margo of the pit membranes are usually the limiting structures, and the low values of longitudinal permeability in softwoods points to the small sizes of these openings.  The void volume fraction (termed as porosity) is one of the most important factors related to the internal structure of wood which affects permeability (Rice and D’Onofrio 1996). This clearly supports the claim by Siau (1995) that a solid must be porous to be permeable. He however, reiterated that not all porous bodies are permeable because permeability can only exist if the void spaces are interconnected by open pit structures between the tracheid lumens. However, if these pit openings are blocked, or if the pit membranes are in the aspirated position, the wood will have a very low permeability.  Measurement of wood permeability is believed to be independent of the fluid type (liquid or gas) if the liquid is non-swelling or a correction is made for slip flow of the gas through the small pore spaces (Comstock 1967). Milota et al. (1995) indicated that gas permeability is simpler to measure than liquid permeability, because gas permeability has no problems with air blockages or suspended particles. The product of the permeability and viscosity of the fluid gives the specific permeability which is only dependent on the porous structure of the medium (Siau 1995) (equation 2.6).  (2.6)  K= kµ  where K is the specific permeability (m3/m), k is the superficial permeability (µm3/µm) and µ is the viscosity of fluid (Pa s).  Another important wood property which can affect porosity and contribute to wide variations in permeability among different species is density. Hughes (1967) defined density as a measure of the quantity of cell wall material contained in a specific volume of wood. Zobel and Bujitenen (1980) claim that wood density is manifested by a corresponding increase in the amount of water in the cell voids of the given volume since 15  the porous space is occupied by water, which has a major role in increasing wood density. They stated that porosity of wood is therefore expected to decrease as density increases because the voids in wood are expected to be filled with moisture. Porosity and permeability values between various wood species are thus expected to vary because of differences in density.  Several studies have been conducted to find ways of improving the permeability of wood species which are known to be refractory, and create problems in their utilization. HongHai et al. (2005) improved the permeability of larch wood by intensive microwave irradiation. Torgovnikov and Vinden (2000a, 2000b) and Vinden and Torgovnikov (2000) also succeeded in improving the permeability of less permeable hardwoods using microwave radiation. Cai and Oliveira (2007) improved the permeability of sub-alpine fir by microwave and radio frequency treatments but the impact was found not to be significant. The underlying principle behind this technology is that the application of high intensity microwave and radio frequency radiation to wood with certain moisture content results in the absorption of the electromagnetic energy by the water molecules. The water vaporize quickly resulting in the generation of high steam pressure which ruptures wood cells and creates narrow voids, which results in a change in the permeability of the wood.  Zhang and Cai (2008) succeeded in improving the permeability of sub-alpine fir by an optimized high-temperature drying schedule. It is believed that the force of vaporization and/or thermal stresses due to the fast heating was able to open the aspirated pits or break the membranes in the wood cells, and therefore increased the permeability. Zhang and Cai (2009) also increased the permeability of sub-alpine fir by steaming, in which the wood was heated to generate internal steam pressure. The instantaneous release of the steam resulted in large pressure difference across the membrane of an aspirated pit pair and caused de-aspiration that led to overall improvement in permeability. Chen (1975) also succeeded in improving the permeability of black walnut by steaming to different temperatures and time lengths.  16  Stamm (1932) improved the permeability of sitka spruce by chemical treatment. Sitka spruce wood sections were treated with chlorine gas, followed in some instances by treatment with ammonia gas or ammonia water. This treatment is believed to have opened the capillary structure as a result of solvent action on the lignin of the pit membranes which is the major resistance to flow.  Johnson (1979) improved the longitudinal permeability of western hemlock, grand fir and Douglas-fir sapwood after exposure to bacteria. Scanning electron microscopy revealed that the improvement was basically due to degradation of bordered pit membranes. Efranjsah et al. (1989) improved the longitudinal permeability of spruce wood by soaking in water in the presence of bacteria. The improvement is believed to be caused by the attack of the torus by the bacteria. Schwarze et al. (2006) also improved the permeability of Norway spruce by treatment with the white rot fungus Physisporinus vitreus and attributed this to the degradation of the pit membrane by the fungus.  2.6 Dielectric heating of wood Dielectric heating is a method used to produce thermal energy within materials, which are also usually poor conductors of heat and electricity. As a result of its efficiency, dielectric heating has become an indispensable heating and drying method in various industrial processes, for the heating and drying of hygroscopic materials and materials that contain large quantities of water, such as agricultural grains, ceramics, paper products and wood.  Schiffmann (1987) indicated that, dielectric heating can be applied to all electromagnetic frequencies up to and including at least the infrared spectrum. For industrial processing however, two different frequency ranges can be differentiated; a radio frequency below 100 MHz and microwaves above 500 MHz. The location of these two bands in the electromagnetic spectrum is shown in Fig. 2.5. International agreement has given the radio frequencies 13.56, 27.12, and 40.68 MHz as well as microwaves of 915, 2450, and 5800 MHz for industrial processing (Resch 2003, Metaxas and Meredith 1983). These two dielectric heating methods can be distinguished based on their operating frequency, 17  depth of penetration, and the technology used to generate the required high frequency electric field. Microwaves have higher frequency and lower depth of penetration than radio frequencies (Strumillo and Kudra 1986). Moreover, RF heating systems use high power electrical valves, transmission lines and applicators in the form of capacitors, while microwave systems are based on magnetrons, waveguides and resonant (or nonresonant) cavities (Jones 1997, Jones and Rowley 1996).  Figure 2.5: The electromagnetic spectrum (Anonymous)  Dielectric heating is a fast and highly efficient method that depends on the inherent properties of moist wood to efficiently convert electromagnetic energy into thermal energy (Avramidis and Ruddick 1996). When wood is placed in a changing RF electromagnetic field, the electric field interacts with the whole volume of the wood causing the polar water molecules to vibrate vigorously to try and follow the rapidly changing polarity of the electromagnetic field. This results in the generation of heat through the wood due to molecular friction (Biryukov 1961). The vibration of the polar molecules is called polarization. Zaky and Hawley (1970) outlined some of the polarizations that can occur in any dielectric material; electron polarization caused by displacement of electrons with respect to positive nuclei in the atoms under the influence of an electric field; ionic polarization is caused by a shift in the relative positions of the positive and negative ions of a molecule, and may result in heat dissipation at very high frequencies. Interfacial polarization on the other hand is caused by accumulation of free ions at the interfaces of dissimilar compounds possessing different electrical conductivities. Orientation polarization is caused by permanent dipoles in polar dielectrics that re-orient under the influence of a changing electric field (Resch 2003). 18  The rate of heating of wood in a high frequency electric field is a function of the power density and the dielectric properties of the wood. The power dissipated per unit volume of a dielectric material under the influence of an external high frequency electric field is called the power density (PD, kW/m3), which can be calculated as: PD = (5.56x10-11) E2 f ε"  (2.5)  where E is the field strength (V/m), f is the frequency (Hz) and ε" is the loss factor (dimensionless) (Torgovnikov 1993). The dielectric properties of wood are very important in discussing dielectric heating. The dielectric properties are characterized by the dielectric loss tangent (tanδ) and the relative dielectric constant (ε') (Torgovnikov 1993). Resch (2003) describes tanδ as a material property reflecting the effects of applied electric field and direct current conductivity. He reiterated that tanδ stands for the high frequency electric energy dissipated in the material and transformed into thermal energy. The tanδ is numerically equal to the ratio between the active current and the reactive current in the material or to the real and reactive power ratio (Torgovnikov 1993). For practical purposes, tanδ can be regarded as an index of the fraction of the power dissipated by the dielectric material as heat (Lin 1967). The ε' of a material shows how many times the force of interaction between the electric charges in the given medium is less than that in a vacuum (Torgovnikov 1993). Peyskens et al. (1984) stated that the greater the polarization of a dielectric material, the greater will be the ε'. The ε' generally increases as the quantity of free water in a material increases, since water has a high ε' of approximately 78 at room temperature (Schiffmann 1987). The product of tanδ and ε' gives the loss factor (ε") which determines the amount of energy absorbed by the wood material (Callebaut 2007).  2.6.1 Factors affecting dielectric properties of wood The dielectric properties of wood are very important in discussing dielectric heating since they influence the amount of electric energy converted into thermal energy. However, as a result of the heterogeneous nature of wood and the dependence of its dielectric 19  properties on several factors, the dielectric properties of wood are still not fully understood and quantified (Biryukov 1961). Dielectric properties of wood are affected by parameters such as moisture content, temperature, frequency, wood species, and sapwood/heartwood percent amongst others.  Effect of temperature on dielectric properties of wood According to Zwick (1995), there is a strong relationship between temperature and the dielectric properties of wood at high moisture content. At low moisture content the temperature has little effect on the dielectric properties of wood, but the ε" of wood may increase with temperature at low moisture content. Avramidis and Zhou (1999) showed that the influence of temperature on ε" is positively linear, and more pronounced at higher moisture contents. Lin (1967) reported that there is a direct relationship between ε' and temperature, and indicated that the effect of temperature on dielectric properties is higher at higher moisture content. Siau (1995) indicated that tanδ can increase or decrease as the temperature decreases at a given frequency, and is dependent on whether the frequency is below or above the frequency corresponding to the relaxation time.  Effect of moisture content on dielectric properties of wood The most important parameter affecting the dielectric properties of wood is its moisture content; including the mineral or electrolytic solutions suspended in the water (Zwick 1995). If the moisture content of wood increases beyond the fibre-saturation point, its dielectric properties become similar to those of semiconductors having large values of tanδ and ε'. Thus, ε" increases with increasing moisture content (Biryukov 1961). Avramidis and Zhou (1999) found that ε" tends to increase as moisture content increases. This is not surprising because ε' and tanδ have a positive relationship with moisture content (Skaar 1948, Hearmon and Burcham 1954).  20  Effect of frequency on dielectric properties of wood According to Torgovnikov (1993), tanδ does not change significantly for high frequency power input between the range of 1 MHz and 1 GHz, which means that the ability of wood to absorb high frequency power is not significantly affected by frequency in the range of interest for dielectric heating. Avramidis and Dubois (1992) reported that ε" can vary greatly with frequency.  Effect of wood species and sapwood/heartwood on dielectric properties of wood According to Zwick (1995), information on the dielectric properties of wood shows inconsistent behavior for sapwood and heartwood of different wood species. He showed that, there is little difference in ε" between sapwood and heartwood of species like western hemlock. However, for species like Douglas-fir and sitka spruce, there is a higher ε" value for the sapwood zone than the heartwood zone, especially at lower frequencies and high temperatures. Avramidis and Dubois (1992) however, indicated that the variation in ε" between different species is caused by characteristics such as the sorption behaviour of wood. They studied the ε" of four different wood species at low moisture contents, and found minimal differences between them within the same grain direction.  2.7 Reasons for drying wood Freshly cut wood usually contain large amount of water which needs to be removed in order to derive optimal benefits from the use of wood in service. Wood drying has a number of distinct and important advantages:  Drying provides dimensional stability for wood in service (Simpson 1983). Wood shrinks and swells with changes in moisture content in service, however, if wood is dried to the moisture content it is expected to attain in service and placed in a stable environment, the problem of shrinkage and swelling in service can be eliminated entirely (Bramhall and Wellwood 1976). Dry lumber is less likely to experience attack by fungi and other 21  organisms, and if kept dry, it should last indefinitely (Bachrich 1980). Wood at about 22% moisture content is less susceptible to fungi which use wood or its content as food, and wood dried to less than 10% moisture content is less likely to be attacked by insects (Bramhall and Wellwood 1976). Lumber transported by rail or truck are charged based on weight (Mackay and Oliveira 1989). The weight of dry lumber is about 40% to 50% lower than that of wet wood (Denig et al. 2000), and there is therefore significant savings in shipping weight and freight reductions, with financial benefits when lumber is dried (Bachrich 1980).  According to Bachrich (1980), dry lumber exhibits significant improvement in tensile and compressive strength and therefore dry lumber is more suitable as a building material. This point was also buttressed by Denig et al. (2000) who pointed out that dry lumber is more than twice as strong and nearly twice as stiff as wet lumber. Simpson (1983) also indicated that cutting, machining, and applications of paints, varnishes and other finishes is easier after wood is dried.  2.8 Mechanism of wood drying Wood drying is an unsteady state process of heat and mass transfer in an orthotropic continuum with variable properties (Martinovic et al. 2001). Two major flow processes are involved in the drying of wood, namely, thermal and moisture. The two heat transfer mechanisms in wood drying are conduction and convection. Conduction, involves the transfer of heat from areas of high to low thermal energies, and convection involves the exchange of thermal energy between a fluid (air) and a solid surface (wood).  Movement of water in wood occurs as liquid or vapor through several passageways (Simpson 1991), which according to Panshin and de Zeeuw (1980) include cell cavities of fibers and vessels, ray cells, pit chambers and associated pit membrane openings and resin ducts. According to Simpson (1983) wood drying includes three mechanisms of moisture movement that occur in various combinations and proportions depending on species, thickness and stage of drying. These include mass flow of liquid water above 22  MFSP, diffusion of water below MFSP and the removal of moisture from the surface of the wood. Free water in wood moves in response to capillary forces created by the evaporation of water from the surface cell, and the most important parameter affecting this moisture flow is wood permeability (Bramhall and Wellwood 1976). Simpson (1983) indicated that the flow of free water requires both a continuous passageway and a driving pressure, which is the difference between pressure in the gas phase and liquid phase, given as (2.3)  Po-P1 = 2σ/r  where Po is the pressure in the gaseous phase (Pa), P1 is the pressure in the liquid phase (Pa), σ is the surface tension of liquid (N/m) and r is the radius of capillary (m). Bramhall and Wellwood (1976) pointed out that one factor affecting capillary movement is the wood temperature, because the viscosity of water is reduced at higher temperatures. Therefore, under a given capillary pull, water will move more rapidly at higher wood temperatures.  Simpson and Liu (1997) pointed out that moisture movement below MFSP is from regions of high to that of low moisture content. The driving forces of moisture diffusion in wood are moisture content, partial pressure of water vapor (Bramhall 1995), chemical potential (Skaar 1988) and water potential (Cloutier and Fortin 1993). Movement of moisture at this stage depends on the relative sizes of the orifices leading into cell lumens and the presence or absence of air bubbles in the water in the cell lumens (Siau 1984). The rate of diffusion largely depends upon the permeability of the cell walls and their thickness. Permeable species dry faster than impermeable ones, and the rate of diffusion decreases as the specific gravity increases (Simpson 1983). According to Crank (1975), the diffusion of moisture during drying obeys Fick’s second law, which is expressed mathematically as:  ࣔm/ࣔt= (ࣔ (D ࣔm/ࣔx))/ ࣔx  (2.4)  23  where m is the fractional moisture content, t is the time(s), x is the thickness direction (mm) and D is the diffusion or internal transfer coefficient (mm2/s).  2.9 Stages of wood drying Three stages of wood drying which are determined by the changes in drying rate, and result in the drying rate curve (Fig.2.6) are recognized (Rosen 1983, Kollman and Côté 1968). The first stage is called the constant rate period, and the wood at this stage is full of free and bound water. Liquid water moves by capillary forces to the surface and moisture movement across the lumber will depend on the wood permeability (Jankowsky and Dos Santos 2004). Drying rate at this stage is constant, and is controlled by the external conditions (Rosen 1983). Humidity is high and temperature is low a this stage of the drying process, to reduce the occurrence of drying defects.  The period below the fiber saturation point (MFSP) is called the falling rate period, and is divided into the first and second falling rate periods. This period is reached when the surface of the wood falls below MFSP (Rosen 1983). According to Jakowsky and Dos Santos (2004) internal moisture movement at this stage involves the liquid flow and diffusion of water vapor and hygroscopic water. Sehlstedt-Persson (2005) indicated that the drying rate at this stage is dependent on the ability of wood to transfer water vapor within the wood structure through bound water diffusion when hydrogen bonding between water molecules and hydroxyl groups in hemicelluloses and cellulose has to be overcome. According to Jankowsky and Dos Santos (2004), in the last phase (second falling rate period) there is no more liquid water in the lumber, and the drying rate is controlled only by material properties.  24  DRYINGRATE  Falling Rate Region I  Constant Rate Region  Falling Rate Region II  MOISTURE CONTENT  Figure 2.6: Theoretical drying rate curve for a hygroscopic solid material (Rosen 1983)  2.10 Drying defects The purpose of drying is to remove as much of the moisture in wood as possible and at the same time maintain material quality, to ensure efficient utilization. This is however, barely achieved because material variability and drying conditions often leads to drying defects which reduce product quality and grade recovery. Simpson (1991) defines drying defect “as any characteristic or blemish in a wood product that occurs during drying and reduces the product’s value for its intended end-use”. He went on to say that defects that occur during and after drying can be categorized into: rupture of wood tissue, warp, uneven moisture content and discoloration.  Simpson (1991) defined warping as any deviation of the face or edge of a board from flatness. Warp can be caused by differential shrinkage as a result of growth ring orientation, or the presence of cross grain, spiral grain, reaction wood or juvenile wood (Bramhall and Wellwood 1976). Warping may also be intensified by casehardening and improper piling (Kollmann and Côté 1968). According to Keey et al. (2000), the severity of warp is not determined by the presence or extent of the effect of these characteristics, but it is their gradients within and across the lumber that are so economically and physically destructive. Stamm (1964) revealed that the magnitude of warp is influenced 25  by the basic density and hemicelluloses content of the wood because the latter are responsible for adsorption of water within the cell wall and affect shrinkage on drying, and the former has a direct effect on the extent of shrinkage. The major forms of warp are bow, cup, crook, twist and diamonding (Fig. 2.7).  Figure 2.7: Schematic illustration of different types of warp (Simpson 1991, with permission from USDA Forest Services, Forest Products Lab. Madison, WI) Casehardening is a condition of stress and set in wood in which the outer fibers are under compressive stress and the inner fibers under tensile stress, after drying (Mcmillen 1958). Casehardening is caused by too rapid or uneven drying as a result of too high temperature or too low relative humidity or large fluctuation of both (Kollmann and Côté 1968). Casehardening is a normal occurrence in wood drying and can easily be relieved at the end of the drying process (Denig et al. 2000) by conditioning. According to Mackay and Oliveira (1989), the tendency towards casehardening is greater in thick than in thin lumber, in impermeable slow-to-dry species than in permeable ones, and in rapid forced drying conditions than in slower drying conditions.  Checking can occur on the surface, interior or end-surfaces of lumber. Surface check is any crack in or near the surface of lumber that develops in drying (Denig et al. 2000). According to Bramhall and Welwood (1976), three factors are responsible for checking in wood, and these include differential shrinkage between surface and core, between normal and reaction wood and between radial and tangential directions. End checks occur 26  when moisture moves faster in the longitudinal direction than in the transverse direction, resulting in the ends of the board drying faster than the middle and stresses developing at the ends. End checks often indicate that the relative humidity in the kiln is too low or air velocity is too high (Denig et al. 2000). Internal checking (honeycombing) occurs when an internal tensile stress perpendicular to the grain exceeds the maximum strength of the wood (Mcmillen 1958).  Keey et al. (2000) indicated that collapse occurs at moisture contents above the fiber saturation point, and is a flattening of the fibers caused by surface tension which is generated as water evaporates from the initially water saturated lumens. Collapse is a common feature of low permeability woods, an indication of its association with high values of capillary tension as the liquid menisci retreat into cells completely filled with liquid (Siau 1995). Kollmann and Côté (1968) also believe the heartwood is more prone to collapse than the sapwood. Collapsed lumber often shows grooves on its surface, and in extreme cases it may cause distortion of the piece (Mcmillen 1958).  2.11 Conventional kiln drying of wood In conventional kiln drying, wood is dried at controlled temperature and relative humidity to attain the required moisture content without compromising the quality of the final product. Simpson (1991) reports that as the development of the modern dry kiln has progressed, changes in design have been introduced in accordance with the mechanism of heat supply, arrangement and type of fans, control of relative humidity and wet-bulb temperature. Figure 2.8 shows a typical conventional dry kiln with all the features.  Heat energy is required in the kiln to carry out the process of removing moisture from the wood. The heat of vaporization of water is about 2326 kJ/kg, and this means that considerable amount of energy is required in the kiln to effect the process of wood drying. Hansom (1988) reported that the total energy needed to remove one kilogram of water from lumber comprises of 2.4 MJ heat of vaporization, energy to raise ventilation air to kiln conditions and energy to heat kiln and wood to required temperature. The 27  transfer of heat energy into the kiln can be achieved in basically two main ways: indirect, where a hot fluid such as steam, hot water or oil is conducted into the kiln through pipes and direct in which combustion products of natural gas or other hydrocarbons are discharged directly into the kiln atmosphere (Bramhall and Wellwood 1976).  Air of controlled temperature and relative humidity must be circulated over lumber to effect drying. According to Bachrich (1980), this is required to supply sufficient heat to warm the lumber pile, and sufficient air to affect the required rate of evaporation. This is achieved in almost all kilns by the use of fans which are mostly located within the kiln and mounted overhead the lumber.  To achieve the desired result in kiln drying the amount of moisture in the kiln must be controlled. When the moisture level in the kiln gets higher than expected excess moisture must be removed from the kiln and replaced with air from the outside, and vice versa. When the humidity inside the kiln is lower than required, moisture is added to the kiln by steam spray or water spray (Simpson 1991).  Figure 2.8: Schematic illustration of conventional dry kiln (Simpson 1991, with permission from USDA Forest Services, Forest Products Lab. Madison, WI)  28  3. MATERIALS AND METHODS 3.1 Specimen preparation A total of three hundred and four (304) green sub-alpine fir lumber pieces measuring 51 x 102 mm and containing wetwood were obtained from a local sawmill in interior BC. The lumber pieces were about 3.5 m long and each was sawn to three different lengths of 0.9 m by discarding about 50 mm long pieces from each end as shown in Fig. 3.1. Wood sections of 25 mm thickness were cut from the ends of each piece for moisture content and basic density determinations (Fig.3.1). The cut sections were immediately weighed with Mettler PM 4600 Delta Range model digital balance and the displacement method (Fig. 3.2) was used to determine their volume. The wood sections were subsequently oven-dried at temperature of 103oC for 24 hours and oven-dry weights measured with Mettler PM 4600 Delta Range model digital balance. Moisture content was determined by the gravimetric method. The oven-dry weights of the wood sections, together with the green volumes were used to determine the basic density of each experimental wood piece.  Reject  Reject  0.9m  0.9m 3.5m  0.9m  Figure 3.1: Procedure for cutting drying samples from original lumber pieces Eighteen drying groups of 40 samples each were created making sure that differences in average moisture content and basic density between groups were neutralized. Eighteen groups (of five samples each) of similar moisture content were also created for 29  permeability tests. This was accomplished by using the random number generator in Excel to arbitrarily assign each sample to the different groups. Analysis of Variance (ANOVA) was later used to test any significant difference or otherwise in basic density and moisture content between the different groups. Thus, 18 groups of similar average (and standard deviation) basic density and moisture content were created for the RF heating process before drying. The experimental design used in this study was therefore completely randomized.  Figure 3.2: Determination of wood volume by the displacement method  3.2 Radio frequency heating treatments All sub-alpine fir samples (except the controls) were RF heated with different power densities and time periods, but at the same final temperature level (also called treatment temperature - Tt) before kiln drying and permeability tests. The intention here was to heat the samples and the bound/free water in both normal wood and the wet pockets, create internal steam pressure that in turn could “open-up” the internal structure by creating 30  micro-voids, possibly break aspirated pits and/or increase permeability. This situation is expected to help obtain uniform moisture content after drying and reduce defects.  The entire study was divided into two phases (Table 3.1), and the first phase involved heating eight groups (with one additional group as control) to target temperature of 100oC and maintaining this temperature for 15, 30, 45, 60, 75, 90, 105 and 120 minutes.  Table 3.1: Radio frequency treatments in phase one and two  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120  Treatment Power density temperature (oC) (kW/m3)  100  27  100  72  Samples were thus RF heated using one level of core temperature (Tc=100oC), power density (PD) of 27 kW/m3 and eight treatment time (t) lengths (15, 30, 45, 60, 75, 90, 105 and 120 minutes) after reaching Tc. The initial intention in phase one was to heat five samples (per treatment) separately for permeability tests, after heating the drying groups. However, it was impossible to mimic the same heating conditions as used for the drying groups. These made it impossible to perform permeability tests in phase 1. Based on this knowledge, permeability test samples were heated together with the drying groups in phase 2.  31  The second phase followed the same experimental design, except that samples were heated at a higher PD, namely, 72 kW/m3 (Table 3.1). This modification in experimental design was adopted to ascertain any effect of the applied PD and thus heating rate (dT/dt) on the drying quality of sub-alpine fir. To achieve this higher PD, each of the original groups was divided into two sub-groups of 20 samples each before heating. Five samples per group, of similar average moisture content and containing wet-pockets were heated along with the drying samples and used for permeability tests. In all treatment runs, upon reaching the target temperature of 100oC, it was maintained for time lengths of 15, 60, 75, 90 and 120oC.  Before RF heating, all samples were weighed on a Mettler PM 4600 Delta Range model digital balance and holes of about 25 mm deep drilled at the ends of eight wood samples with the highest moisture content within each group. Fiber optic temperature probes with quick response time were then inserted to monitor the internal temperature of the wood samples as they were heated (Fig. 3.3). The measured weights together with the estimated oven-dry ones from the initial moisture contents and weights were used to calculate the moisture content of each wood sample before RF heating.  Figure 3.3: Arrangement of wood samples and sensor positioning during RF heating in phase 1 (left) and phase 2 (right) All heating was carried out in an RF oven using a push-pull oscillator (Fig. 3.4) with a maximum power of 10 kW that operates at maximum frequency of 16 MHz and 10 kV at two amperes. The oven was an aluminum box (Fig 3.4) measuring 1.8 m long, 1.8 m 32  wide and 1.0 m high. Two vacuum tubes operating as oscillators convert direct current power into radio frequency power that was directed to the electrodes through a twin conductor balance transmission line. To reduce heat loss during RF heating, wood samples were surrounded by MDF boards (Fig. 3.4). Immediately after each RF heating treatment the weights of all wood samples were measured with Mettler PM 4600 Delta Range model digital balance to determine the moisture loss during RF heating. All wood samples were also inspected for the occurrence of surface checks, and the number of samples within each group with surface checks as well as the severity was recorded. Samples were allowed to cool down for about two hours and were subsequently wrapped in plastic sheets for onward kiln drying.  Figure 3.4: RF oscillator (left) and oven (right) with wood samples surrounded by an MDF box 3.3 Kiln drying experiments Drying experiments were carried out in a 1m3 capacity experimental kiln (Fig. 3.5), and the drying consisted of 15 runs (40 pieces per run) and overall total of 600 samples. Before drying all samples were weighed, and this together with the initial moisture content used to estimate the moisture content before drying. The arrangement of the samples in the kiln was five samples wide and eight high (Fig 3.5), with the samples arranged on aluminum stickers to allow for air flow through them only in the horizontal 33  direction. Heat for the drying experiment was supplied by two electrical heating coils (3 kW and 4 kW) and moisture for relative humidity control was provided by steam. Unidirectional fans hang at the top of the kiln, and baffles prevent air movement over the load. The kiln has a load cell unit which takes the weight change as the wood dries, and the kiln is designed to stop automatically when the desired final moisture content is attained. The kiln is also equipped with three resistance temperature devices that measure the outlet and inlet dry-bulb temperatures, as well as the wet-bulb temperature. Drying data (change in moisture content and time) were obtained automatically by a computer data acquisition system. Samples were dried to a final moisture content of 14% utilizing the same drying schedule for each run (Table 3.2).  Figure 3.5: The dry kiln showing the arrangement of wood samples  Step # 1 2 3 4 5  heat-up d1 d2 d3 d4  Table 3.2: Drying schedule for drying experiments Ramp Elapsed time Duration time M DBT WBT WBD (hr) (hr) (hr) (%) (C) (C) (C) 6 6 60 60 0 10 16 71 68 3 8 24 82 78 4 12 36 88 79 9 99 135 14 91 76 15  34  RH (%) 100 88 83 69 53  MEMC (%) 25 16 12 9 6  Air velocity (m/s) 5 5 5 5 5  3.4 Post-drying cutting and measurements After drying in phase 1, each wood sample was weighed and inspected for occurrence of any drying defects in the form of surface checking. They were subsequently marked and cut into wood sections as illustrated in Fig. 3.6. Sections marked as AM were used to determine the average final moisture content, CS were used to determine core-shell moisture content, and IC for internal checking (Fig. 3.6). Wood sections for the determination of the final average moisture content were quickly weighed with Mettler PM 4600 Delta Range model digital balance. Reject  Reject  268mm  25mm  100mm  910mm  AM  CS  IC  Figure 3.6: Post-drying cutting procedure in phase 1  Core-shell moisture content sections were marked and cut into core and shell parts as shown in Fig. 3.7, and quickly weighed with a digital balance. These together with the final average moisture content wood sections were oven-dried at 103oC for 24 hours and the moisture content determined by the gravimetric method. The number of internal checks in each wood section marked as IC was counted and the severity of checking was noted.  35  Core  Cut  Cut  Cut  Cut  Shell (10mm thick)  Figure 3.7: Procedure for removing shell and core parts from wood sections for moisture content determination in phase 1  As in phase 1, each wood sample in phase 2 was weighed after drying and inspected for occurrence of any drying defects in the form of surface checking. Slight modification was however introduced in the post-drying cutting and measurements in phase two, to allow for proper assessment of wet-pockets as well as the final moisture content distribution along the length of each wood piece. Each wood sample was marked into wood sections as shown in Fig 3.8. Wood sections measuring about 25 mm thick were cut from the ends of each piece and discarded. Sections of about 30 mm thick were subsequently marked along the length of each wood sample by leaving 21 mm thick sections between them. The 30 mm thick sections were used to determine the average final moisture content, final moisture content distribution along the length, and core-shell moisture content of each dried sample. To determine the core and shell moisture contents (Mc and Ms), Delmhorst RDM-2 pin-type moisture meter was used to measure moisture content at depths of 5 mm (for shell) and 14.5 mm (for core) for the marked wood sections. Each wood piece was subsequently cut as shown in Fig. 3.9 and the wood sections immediately weighed with a digital balance and oven-dried at 103oC for 24 hours. Moisture content was calculated based on the initial and oven-dry weights. Reject  910 mm  Reject  25 30 21 30 21 m  Figure 3.8: Procedure for marking wood sections after drying in phase 2 36  Reject  Reject  Figure 3.9: Post-drying cutting procedure in phase 2  3.5 Permeability experiments As stated previously, five wood samples of similar average moisture contents and containing wet-pockets were added to each of the drying groups during RF heating in phase two. These samples were used for longitudinal permeability tests. After RF heating, each of these wood samples were cut into permeability test samples of 35 mm long and cross-sectional area of 20 x 20 mm2, with their grain oriented in the longitudinal direction (Fig. 3.10). As many permeability test samples as possible were cut from each piece, but due to defects such as cracks, knots and stains only a few of the cut samples could be used for permeability tests. The selected test samples were subsequently conditioned at 60oC temperature and relative humidity of 67% to a final moisture content of 10%. During the conditioning, the weight of the samples was monitored until constant weights were obtained for a period of time, an indication that the required moisture content had been attained. Before the experiments, the sides of the specimens were coated with epoxy resin to prevent transverse airflow during the period of the experiments. The end surfaces of the samples were also sliced with a sharp razor blade to expose fresh surfaces, and the length and cross-sectional dimensions measured using a pair of digital calipers. 35mm 20mm  20mm Figure 3.10: Longitudinal air permeability specimen  37  Permeability experiments were carried out at the UBC-Wood Physics Laboratory, and a total of 90 samples (15 samples per treatment) were used. Experiments were conducted using the falling-water volume displacement method as described by Siau (1995) and illustrated schematically in Fig. 3.11. Vacuum  Vr b  ∆z  P changing with time  Vd a  Test specimen z  Patmospheric is constant  r  Figure 3.11: Schematic illustration of falling-method apparatus  The test specimen was tightly clamped at the point labeled as test specimen (Fig 3.11). Water is then drawn above point a by vacuum, and as the vacuum is turned off air flows through the test sample and causes the water level to drop. The time taken for the water level to drop from point b to point a, was recorded. This was repeated three times for each test sample and the average time was calculated. Permeability was then calculated using the equation as stated by Siau (1995):  kg = (Vd C L(Patm-0.074 z ))/(t A (0.074 z )(Patm- 0.03 z ))*(0.760m Hg)/1.013x105  (3.1)  where, Vd (= πr2∆z) is the volume of gas displaced by water in displacement tube, (m3), L is the length of test sample, (m), z is the average height of water over surface of reservoir  38  during the period of measurement, (m), t is time, (s), A is the cross-sectional area of test sample, (m2) and C is a correction factor.  C = 1+ (Vr (0.074 ∆z))/ Vd (Patm-0.074 z )  (3.2)  where, Vr is the gas volume remaining in the system at the end of the gas flow period, (m3).  3.6 Data analysis To test the effect of the treatments on measured variables, a completely randomized experimental design was used. Comparison of treatment means of measured variables was carried out using Analysis of Variance (ANOVA) with treatment time as the independent variable and final moisture content, core-shell moisture content, honeycomb count and longitudinal permeability as the dependent variables. The Bartlett’s and Shapiro Wilk’s tests were used to test if data met the assumptions of equal variance and normality. Where required, transformations were applied to data to meet assumptions of ANOVA i.e. normality with constant variance and where necessary, pair-wise comparison of treatment means was carried out with the Bonferroni’s and Tukey tests.  The linear model of the ANOVA is:  yij = µ + τj + εij or yij = µj + εij  (3.3)  where yij is the response variable measured on experimental unit i and treatment j, j is 1 to J treatments, µ is the grand or overall mean regardless of treatment, µj is the mean of all  measures possible for treatment j, τj is the difference between the overall mean of all measures possible from all treatments and the mean of all possible measures for treatment j, called the treatment effect, εij is the difference between a particular measure for an  experimental unit i, and the mean for the treatment j that was applied to it. The ANOVA table is shown in Table 3.3  39  Table 3.3: Analysis of variance of completely randomized experimental design Source of variation  df  SS  MS  F  Treatment  J-1  SSTR  MSTR = SSTR /(J-1) F = MSTR/MSE  Error  nT - J  SSE  MSE = SSE/(nT-J)  Total  nT - 1  SSy  where SSy is the sum of squared differences between the observed values and the overall mean, SSTR is the sum of squared differences between the treatment means and the grand mean, weighted by the number of experimental units in each treatment, SSE is the sum of squared differences between the observed values for each experimental unit and the treatment means and nT is the number of experimental units measured over all treatments. Comparison of standard deviations was done using the Bartlett’s test and all analyses were carried out with significance level of 5%.  40  4. RESULTS AND DISCUSSION 4.1 Green moisture content and basic density A total of seven hundred and fifty (750) samples were used in the entire study which were randomly divided into eighteen groups for drying and an additional six groups of five samples each for longitudinal air permeability tests. These groups were named according to the treatments to be received during RF heating. Table 4.1 and Fig. 4.1 show the average moisture content of each group before RF heating. The average initial moisture content of the eighteen groups was 86.6%, and in phase 1 the moisture content ranged from 47.2% to 154.7%. The highest average moisture content was found in the control while the lowest was in RF30 and RF120. In phase 2 however, the moisture content ranged from 26.7% to 145.3%, and the highest average was in the control whereas the lowest was in RF15. Table 4.1: Moisture contents of the eighteen groups before RF heating  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 30 45 60 75 90 105 120  Mean (%) 88.7 86.9 86.0 86.8 87.3 87.2 86.5 86.6 86.0 89.1 84.9 85.7 86.4 85.9 85.8 86.6 88.6 86.2  41  SD (%) 19.6 19.6 18.7 19.1 21.2 20.0 19.5 19.1 18.8 20.9 17.4 19.5 19.8 19.2 21.5 19.0 20.9 18.4  Min. (%) 52.3 49.6 53.8 52.6 50.3 52.6 54.0 49.4 47.2 53.5 47.9 49.0 44.1 49.9 26.7 51.1 40.5 54.0  Max. (%) 133.6 132.9 127.9 132.3 154.7 146.6 145.5 135.7 126.0 153.3 128.5 126.6 134.1 131.7 140.2 137.3 143.5 137.8  (Max) - (Min) 81.3 83.3 74.2 79.7 104.4 94.0 91.4 86.3 78.7 99.8 80.7 77.6 90.0 81.8 113.5 86.3 103.0 83.8  Knudson et al. (2008), in their study of the basic properties of sub-alpine fir reported an average green moisture content of 73.8%, which is lower than the 86.6% average recorded in this study. Skaar (1988) reported that moisture content of wood varies between logs cut from different heights in the tree and between different seasons, and all these can be possible causes of the observed differences in the average moisture content recorded in this study and that of Knudson et al. (2008).  130 Phase 1 Phase 2  120  Moisture content (%)  110 100 90 80 70 60 50 40 30 20 10 Control RF15  RF30  RF45  RF60  RF75  RF90 RF105 RF120  RF heating groups  Figure 4.1: Average moisture content of the eighteen groups before RF heating (error bars indicate ±1SD)  Pre-RF heating moisture content variability (standard deviation) of the 18 groups ranged from 18.7% to 21.2% and 17.4% to 21.5% in phases 1 and 2, respectively, with an overall standard deviation of 19.5%. The use of a one-way analysis of variance (Table 4.2) revealed no significant differences in the average moisture content between groups before RF heating. Comparison of standard deviations with the Bartlett’s test also showed no significant differences within phase 1 (α = 0.05 and p = 0.99) and phase 2 (α = 0.05 and p = 0.96).  42  Table 4.2: Analysis of variance of average moisture content before RF heating  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 92.32 133429.89 133522.22 333.77 133621.82 133955.59  df 8 351 359 8 351 359  MS 11.54 380.14  F 0.03  p-value 1.00  41.72 380.69  0.11  0.99  Table 4.3 and Fig. 4.2 present the basic density of the eighteen groups. In phase 1, it ranged from 297 kg/m3 in RF120 to 510 kg/m3 in RF75. The highest average basic density of the groups on the other hand was 364 kg/m3 whiles the lowest was 366 kg/m3. In phase 2, however, this ranged from 269 kg/m3 in RF15 to 534 kg/m3 in the control, with the highest and lowest average of the groups being 366 kg/m3 and 363 kg/m3, respectively.  Phase 1  Phase 2  Table 4.3: Basic densities of the eighteen groups before RF heating SD Min. Max. RF heating RF treatment Mean 3 3 3 groups time (mins) (kg/m ) (kg/m ) (kg/m ) (kg/m3) (Max) - (Min) Control 0 364 34.1 303 452 149 RF15 15 365 36.1 302 462 161 RF30 30 365 34.6 305 456 150 RF45 45 365 34.9 303 458 156 RF60 60 364 34.1 302 448 146 RF75 75 366 38.7 304 510 206 RF90 90 366 37.4 306 494 188 RF105 105 365 34.7 303 464 162 RF120 120 364 35.2 297 453 156 Control 0 366 41.3 305 534 229 RF15 15 363 36.6 269 456 188 RF30 30 364 34.8 297 454 157 RF45 45 365 38.1 296 492 196 RF60 60 365 35.8 300 466 167 RF75 75 365 36.0 296 473 177 RF90 90 364 35.9 299 473 174 RF105 105 365 36.7 295 477 182 RF120 120 364 35.0 305 459 154  43  The average basic density of the eighteen groups was 365 kg/m3 and higher than the value of 329 kg/m3 reported by Smith (1970), a disparity that might be due to differences in wood structure, amount of extractive constituents in the wood, the proportion of juvenile and mature wood contained in the wood and the part of the stem (butt or upper portion) where lumber was obtained from. As stated by Kollmann and Côté (1968), all these aforementioned factors can bring about variations in density.  500 Phase 1 Phase 2  450  Basic density (kg/m 3)  400 350 300 250 200 150 100 50 Control RF15  RF30  RF45  RF60  RF75  RF90  RF105 RF120  RF treatments  Figure 4.2: Average basic density of the eighteen groups before RF heating (error bars indicate ±1SD) The variability (standard deviations) of basic density in phase 1 ranged from 34.1 kg/m3 to 38.7 kg/m3 whereas in phase 2 it ranged from 35 kg/m3 to 41.3 kg/m3. The overall variability of basic density of the eighteen groups was 36 kg/m3. One-way analysis of variance (Table 4.4) however, revealed no significant differences in average basic density between the groups. The Bartlett’s test also showed no significant differences in standard deviations in phase 1 (α = 0.05 and p = 0.99) and phase 2 (α = 0.05 and p = 0.99).  44  Table 4.4: Analysis of variance of average basic density before RF heating Source of Variation SS df MS F p-value Between Groups 76.13 8 9.52 0.01 1.0 Phase 1 Within Groups 444427.72 351 1266.18 Total 444503.85 359 Between Groups 210.40 8 26.30 0.02 1.0 Phase 2 Within Groups 473814.37 351 1349.90 Total 474024.76 359 Although eighteen groups were initially created for RF heating and kiln drying, improper functioning of RF machine and dry kiln led to the loss of three groups in phase 2. Therefore, only fifteen groups were used in the entire study, and the following presentation of results and discussion will now be based on these groups.  4.2 Heating rate (dT/dt) and moisture changes in samples during RF heating As stated previously, temperature probes were inserted into eight samples with the highest moisture content within each group and heated to same final core temperature. This was then maintained at that level for different lengths of time in both phases of this study by regulating the applied power. Thus, two groups were created; the group in which temperature probes were inserted (sub-group) and the entire group (main group) consisting of 40 samples per group. In phase 1, the average moisture content of the subgroups before RF heating ranged from 102.6% in RF30 to 119.0% in RF60. In phase 2, the highest average of 100.8% was in RF75 whereas the lowest of 96.0% was in RF15. The variability in the moisture content (measured by the standard deviations) in phase 1 was from 8.1% in RF120 to 21.5% in RF75. In phase 2, however, this ranged from 15.5% in RF120 to 20.6% in RF90. Analysis of variance revealed no significant differences in the average moisture content in phase 1 (α = 0.05, p = 0.36) and phase 2 (α = 0.05, p = 0.95). The Bartlett’s test also showed no significant differences in the standard deviations in phase 1 (α = 0.05, p = 0.07) and phase 2 (α = 0.05, p = 0.88).  Figure 4.3 shows the characteristic temperature-time curve of the fifteen groups during RF heating, where T1-T8 are the temperatures recorded by temperature probes 1-8, 45  respectively. Based on the temperature-time data obtained, the average dT/dt for each sub-group in both phases was calculated and tabulated (Table 4.5).  110  90  T1  o  Temperature ( C )  100  T2  80  T3  70 60  T6  T5  T4  T7  50  T8  40 30 20 0  50  100  150  200  250  Time (min)  Figure 4.3: Temperature-time curve of RF120 in phase 1  Phase 1  Phase 2  Table 4.5: Average heating rate of RF heating groups RF heating RF treatment time Average dT/dt groups (mins) (oC/min) RF15 15 0.84 RF30 30 0.93 RF45 45 0.72 RF60 60 0.89 RF75 75 0.67 RF90 90 0.76 RF105 105 0.68 RF120 120 0.71 RF15 15 1.56 RF60 60 1.33 RF75 75 1.37 RF90 90 1.46 RF120 120 1.41  The highest dT/dt of 0.93oC/min in phase 1 was recorded in RF30 whereas the lowest of 0.67oC/min was in RF75. The highest and lowest dT/dt in phase 2 was 1.56oC/min and 46  1.33oC/min in RF15 and RF60, respectively. In both phases however, there was no clearcut trend in dT/dt between the groups, and this could be due to differences in the moisture content of the samples from which temperature-time data were obtained as well as the moisture content of the section of the wood samples where temperature probes were inserted. This is because the amount of energy required to increase the temperature of wood is dictated by the specific heat capacity, which according to Kollmann and Côté (1968) varies considerably with moisture content. Sehlstedt-Persson (2005) also reported that the higher the moisture content, the higher the specific heat capacity of wood. Therefore, differences in moisture content of wood can result in differences in specific heat capacity, heating time and consequently dT/dt. Generally, dT/dt was higher for all groups in phase 2 than in phase 1 because of reduced heating time in the former as a result of applying higher power. This is consistent with the findings of Avramidis (2008) who reported of improvement in RF heating time with power density increase. Because the free water (and in some cases below MFSP the bound water) in the wood vaporizes to generate the needed internal pressure, varying degrees of moisture loss were recorded in the groups that were a function of heating temperature and time. Fig. 4.4 shows the average moisture loss by each of the main groups in phases 1 and 2. It is evident from Figs. 4.4 and 4.5 that the moisture loss increased with heating time. The moisture loss followed the same trend in the sub-groups in which temperature probes were inserted. The highest moisture loss was in RF120 while the lowest was in RF15. Such a trend is not surprising because of the longer period of water vaporization during the 120 minutes of heating compared to that of 15 minutes. This longer period resulted in greater moisture evaporation and subsequent moisture loss. However, comparison of moisture loss between phase 1 and 2 reveals relatively greater amount in phase 2, although heating time was longer in the former than the latter. This trend is possibly due to the higher power density applied in phase 2. This suggestion is consistent with that of Torgovnikov and Vinden (2009) in their study of the modification of wood to increase permeability using microwave. They reported a relatively higher moisture loss of 16 to 64% with the application of higher power compared to lower power that resulted to a loss of moisture of 4 to 13%.  47  85  Phase 1 Phase 2  75  Moisture loss (%)  65 55 45 35 25 15 5 RF15  RF30  RF45  RF60  RF75  RF90  RF105  RF120  RF heating groups  Figure 4.4: Average moisture loss of the main groups after RF heating (error bars indicate ±1SD)  70 2  R = 0.9467  Moisture loss (%)  60 50  2  R = 0.963  Phase 2 40 30  Phase 1  20 10 0 1.0  1.5  2.0  2.5  3.0  3.5  4.0  Total RF heating time (hrs)  Figure 4.5: RF heating time and moisture loss relationship 48  4.5  Comparison of moisture loss during RF heating in both phases using one-way analysis of variance (Table 4.6) showed significant differences between the main groups. Significant differences were also observed between the sub-groups in phase 1 (α = 0.05, p = 0.0001) and phase 2 (α = 0.05, p = 0.0001). The subsequent pair-wise comparison to show groups with significantly different means was done and the results are shown in Table 4.7. In both phases, treatments with the same letters are not significantly different and vice versa. Table 4.6: Analysis of variance of average moisture loss during RF heating  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 40.61 15.70 319.00 954.68 391.33 1346.01  df 7 312 56 4 195 199  MS 5.80 0.05  F 115.28  p-value <0.0001  238.67 2.01  118.93  <0.0001  Table 4.7: Pair-wise comparison of average moisture loss during RF heating Main group Sub-group RF heating RF treatment time groups (min) Mean Grouping Mean Grouping RF15 15 7.5 E 6.6 D RF30 30 10.8 E 12.4 C D RF45 45 26.0 C 26.6 B Phase 1 RF60 60 19.1 D 23.6 B C RF75 75 33.0 B 28.1 B RF90 90 35.0 B 28.6 B RF105 105 44.6 A 33.2 B RF120 120 48.0 A 56.7 A RF15 15 18.3 D 20.3 C RF60 60 43.6 C 41.8 B Phase 2 RF75 75 51.4 B 52.1 A B RF90 90 56.4 A 61.1 A RF120 120 59.0 A 60.1 A The dT/dt of the sub-group was related to the average moisture loss of the main group and two regression models capable of predicting the moisture loss from dT/dt were developed for low and high PDs (Fig. 4.6), and tested with experimental data. The 49  experimentally determined moisture loss as well as the calculated loss was fairly in agreement, with reasonably high R2 values (Fig. 4.7). 60 y = -1834.1x2 + 5197.3x - 3633.1 R2 = 0.9377  Average moisture loss (%)  50 40 30  Phase 1 y = 152.01x2 - 364.37x + 217.78 R2 = 0.6673  20  Phase 2  10 0 0.2  0.4  0.6  0.8  1  1.2  1.4  1.6  1.8  Average heating rate (oC/min)  Figure 4.6: Relationship between heating rate in sub-groups and moisture loss in main groups  60  Calculated moisture loss (%)  2  R = 0.9377  50  Phase 2 40 2  R = 0.6673  30  Phase 1 20 10 0 0  10  20  30  40  50  Experimental moisture loss (%)  Figure 4.7: Relationship between experimentally determined and calculated moisture loss 50  60  Due to RF heating moisture losses, large variations in moisture content between groups were recorded before kiln drying. As shown in Table 4.9 and Fig. 4.18 the moisture content of samples in the main groups in phase 1 ranged from 12.9% to 131.9%, with the highest average of 87.4% in the control while the lowest of 35.9% was measured in RF120. In phase 2, however, moisture contents ranged from 10.1% to 145%, and the highest average of 87.8% was found in the control while the lowest of 26.7% was in RF120. As listed in Table 4.10, sub-groups in phase 1 on the other hand had the highest average moisture content before drying of 115.4% in the control while the lowest of 48.9% was in RF120. In phase 2, the highest and lowest averages of 103.6% and 32.43 were recorded in the control and RF90, respectively. Analysis of variance (Table 4.8) revealed significant differences in moisture content between the main groups in phases 1 and 2 before kiln drying. Average moisture content of the sub-groups was also found to be significantly different in phases 1 (α = 0.05, p = 0.0001) and 2 (α = 0.05, p = 0.0001). Subsequent pair-wise comparison of the main groups showed significant differences between the controls and RF heated groups in both phases. In the sub-groups, however, all RF heated groups in phase 1 except RF15, and all RF heated groups in phase 2 were found to be significantly different from the controls. Table 4.8: Analysis of variance of average moisture content of the main groups before kiln drying  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 118726.86 112784.07 231510.94 1094.40 1049.61 2144.01  51  df 8 351 359 5 234 239  MS F 14840.86 46.19 321.32 218.88 4.49  48.80  p-value < 0.0001  <0.0001  Table 4.9: Moisture contents of the fifteen main groups before kiln drying  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120  Mean (%) 87.4 70.9 69.8 56.6 60.9 49.7 47.5 39.3 35.9 87.8 60.2 38.3 31.8 28.9 26.7  SD Min. Max. (%) (%) (%) 19.4 52.3 131.9 18.8 36.7 109.1 16.9 39.4 108.8 17.6 27.9 101.7 18.1 19.6 100.1 17.6 20.1 107.6 18.2 15.9 103.0 18.8 12.9 94.8 13.9 16.7 63.1 19.5 54.9 145.3 16.7 22.2 102.8 18.3 10.8 72.7 13.9 11.9 62.6 15.5 10.2 65.9 14.3 10.1 66.2  (Max) - (Min) 79.7 72.4 69.4 73.8 80.5 87.5 87.0 81.9 46.4 90.4 80.6 61.9 50.7 55.7 56.1  Table 4.10: Moisture contents of the fifteen sub-groups before kiln drying  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120  Mean (%) 115.4 97.9 83.1 79.2 85.2 69.0 73.1 61.8 48.8 103.6 67.0 47.5 39.8 32.4 36.8  52  SD (%) 10.4 7.0 19.4 11.8 10.1 18.3 15.8 20.8 11.3 16.7 16.8 16.0 12.4 18.9 12.4  Min. (%) 103.0 88.3 59.1 66.0 65.7 50.5 48.6 23.8 23.5 84.4 37.4 21.7 15.4 3.2 17.3  Max. (%) 131.9 109.1 108.8 101.7 100.1 107.6 103.0 94.8 63.1 145.3 102.8 72.7 62.6 65.9 66.2  (Max) - (Min) 28.7 20.9 49.7 35.7 34.4 57.2 54.4 70.9 39.6 60.9 65.4 50.9 47.2 62.7 48.9  Moisture content variability (standard deviation) between samples in the main groups (Table 4.9) in phase 1 ranged from 13.9% in RF120 to 19.4% in the control. In phase 2, however, moisture content variability ranged from 14.3% in the control to 19.5% in RF120. Generally, the standard deviations were relatively lower in the RF heated groups than the controls in both phases, and this can be attributable to the moisture loss during heating which tended to narrow the moisture content variability. The standard deviation of the sub-groups in phase 1 (Table 4.10) ranged from 7.0% in RF15 to 20.8% in RF105. In phase 2, this ranged from 12.4% in RF75 and RF120 to 18.9% in RF90. However, the Bartlett’s test showed no significant differences in standard deviations between the main groups in phase 1 (α = 0.05 and p = 0.58) and phase 2 (α = 0.05 and p = 0.21). No significant differences in standard deviations were also observed between the sub-groups in phase 1 (α = 0.05 and p = 0.11) and phase 2 (α = 0.05 and p = 0.51). The distribution of moisture content of the main groups before drying in both phases is illustrated in Figs. 4.8, 4.9 and 4.10.  RF15 30  25  25  20  20  Initial moisture content Final moisture content  15  Frequency  Frequency  Control 30  15  10  10  5  5  0  Initial moisture content Final moisture content  0 0  10  20  30  40  50  60  70  80  90 100 110 120 130 140  0  10  20  30  40  Moisture content (%)  60 70 80 90 100 110 120 130 140 Moisture content (%)  RF45  RF30 30  30  25  25 20 Initial moisture content Final moisture content  15  Frequency  20 Frequency  50  Initial moisture content Final moisture content  15  10  10  5  5 0  0 0  0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%)  10  20  30  40  50 60 70 80 90 100 110 120 130 140 Moisture content (%)  Figure 4.8: Distribution of moisture content before and after kiln drying in phase 1 53  RF75  RF60 30  30  25  25 20  F req u en c y  Initial moisture content Final moisture content  15  Frequency  20  Initial moisture content Final moisture content  15  10  10  5  5  0  0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%)  0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%) RF105  RF90 30  30  25  25 20 Initial moisture content Final moisture content  15  Frequency  F req uen cy  20  Initial moisture content Final moisture content  15  10  10  5  5  0  0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%)  0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%)  RF120 30 25  F req u en cy  20 Initial moisture content Final moisture content  15 10 5 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Moisture content (%)  Figure 4.9: Distribution of moisture content before and after kiln drying in phase 1  54  RF15  Control 30  30  25  25  Initial moisture content Final moisture content  Frequency  Frequency  20 15 10  Final moisture content Initial moisture content  20 15 10  5  5  0  0 0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  RF75  RF60 30  30  25  25 Initial moisture content Final moisture content  Initial moisture content Final moisture content  20 Frequency  Frequency  20 15  15  10  10  5  5  0  0 0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  RF90  RF120  30  30  25  25 Initial moisture content Final moisture content  Initial moisture content Final moisture content  20 Frequency  Frequency  20 15  15  10  10  5  5  0  0 0  10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Moisture content (%)  Figure 4.10: Distribution of moisture content before and after kiln drying in phase 2  55  4.3 Effect of radio frequency heating on the longitudinal air permeability. To ascertain the ability of RF heating to improve the vapor flow as well as the drying characteristics of sub-alpine fir, longitudinal air permeability tests were carried out in phase 2 of this study. Figure 4.11 shows the average moisture content of all groups before RF heating, and the lowest average of 77.2% was in RF120 whiles the highest of 82.7% was in the control. Analysis of variance (Table 4.11) showed no significant differences in moisture content before RF heating.  110 100  Moisture content (%)  90 80 70 60 50 40 30 20 10 Control  RF15  RF60  RF75  RF90  RF120  RF heating groups  Figure 4.11: Average moisture content of groups before heating (error bars indicate ±1SD)  Table 4.11: Analysis of variance of average moisture content before RF heating of permeability test samples Source of Variation SS df MS F p-value Between Groups 128.37 5 25.68 0.06 0.99 Within Groups 9570.98 24 398.79 Total 9699.35 29  56  Figure 4.12 and Table 4.12 show the longitudinal permeability values for all groups. The lowest value of 0.001 µm/µm3, among all samples was recorded in RF15 while the highest value of 3.378 µm/µm3 was recorded in RF60. The lowest average permeability of 0.014 µm/µm3 was also recorded in RF15 whereas the highest of 0.330 µm/µm3 was recorded in RF60. The standard deviation was highest in RF60 and lowest in RF15, and all groups except RF15 had relatively higher permeability than the control.  3  Average longitudinal Permeability (µm /µm)  RF heating groups Control RF15 RF60 RF75 RF90 RF120  Table 4.12: Longitudinal air permeability of groups in phase 2 RF treatment Mean SD Max. Min. time (mins (µm/µm3) (µm/µm3) (µm/µm3) (µm/µm3) (Max) - (Min) 0 0.031 0.021 0.068 0.004 0.064 15 0.014 0.01 0.038 0.001 0.037 60 0.33 0.874 3.378 0.005 3.373 75 0.08 0.149 0.582 0.004 0.578 90 0.039 0.077 0.258 0.003 0.255 120 0.038 0.027 0.08 0.002 0.078  1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Control  RF15  RF60  RF75  RF90  RF120  RF heating groups  Figure 4.12: Average longitudinal permeability of groups in phase 2 (error bars indicate ±1SD)  57  The observed increases in permeability could possibly be due to the creation of microvoids and possible de-aspiration of pits in the wood samples by steam pressure, which were generated as a result of heating and vaporizing the free water in the wood. This is because Hong-Hai et al. (2005) improved the permeability of larch wood by microwave irradiation and attributed this to the creation of micro-cracks in the wood caused by steam pressure. Zhang and Cai (2008) also improved the permeability of sub-alpine fir by fast and slow heating, and by microscopic analysis detected micro-voids as well as broken aspirated pits in the wood. They then attributed this to the force of vaporization generated by steam pressure as a result of moisture evaporation in the wood. The lowest longitudinal permeability observed in RF15 can possibly be due to relatively lower steam pressure generated which could not “open-up” the internal structure of the wood well or higher number of wet-pockets contained in the samples for that group. Analysis of variance (Table 4.13) indicated significant differences in longitudinal permeability between the groups. The subsequent pair-wise comparison with the Bonferroni’s test (Table 4.14) showed no significant differences between the control and all groups (all groups with same letters are not significantly different and vice versa). The only difference was between RF60 and RF15, and this could be due to large variation among samples. As seen in Fig. 4.12, in some cases, shorter duration of RF heating resulted in relatively higher permeability values than longer ones, although the opposite was expected because of the possible higher vapor pressure to be generated when treatment time increases. This trend therefore goes on to support the claim of probable variation between samples in each group. Table 4.13: Analysis of variance of average longitudinal air permeability of groups in phase 2 Source of Variation SS df MS F p-value Between Groups 24.5616 5 4.9123 3.03 0.0146 Within Groups 136.2053 84 1.6215 Total 160.7669 89  58  Table 4.14: Pair-wise comparison of average longitudinal air permeability of groups in phase 2 RF heating groups Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 60 75 90 120  Mean 0.031 0.014 0.330 0.080 0.039 0.038  Grouping A B A B A B A B A B  In kiln drying of timber, moisture movement above MFSP is mostly bulk flow controlled by permeability in the transverse direction, which is the ideal permeability to investigate in this study and relate any improvement in drying characteristics to. Torgovnikov and Vinden (2009) reported the ability of microwave heating to substantially increase the permeability of hitherto less permeable wood species in the longitudinal-radial direction, due to the creation of micro-voids in these directions, caused by the generation of internal steam pressure. Longitudinal permeability was therefore chosen in this study to ascertain the capability of RF heating to improve the permeability of sub-alpine fir in this direction and relate this to the transverse permeability. If the claim by Torgovnikov and Vinden (2009) is anything to go by, then improvement in longitudinal permeability (as recorded by some RF heated groups in phase 2) also points to a high possibility of improvement in the transverse direction. This claim is also supported by the improvement in dT/dt above MFSP (section 4.4) of all RF heated groups in phase 2, which is an indication of  permeability increase in the transverse direction.  However, Siau (1995) reported that due to the smaller number of pit openings traversed in series and larger number in parallel in the longitudinal direction, the ratio of longitudinal to transverse permeability can vary from 10,000 to 40,000 depending on the degree of overlap of the tapered ends of the tracheids. Therefore, any improvement in transverse permeability might not be of the same magnitude as that in the longitudinal direction reported in this study.  59  4.4 Effect of radio frequency heating on drying rates The drying curves for all kiln drying runs in both phases are shown in Fig 4.13. However, for comparison purposes there was a need to normalize them so they begin from the same point. This was achieved by dividing the momentary moisture content of each drying run by the initial moisture content (M/Mi). Drying curves portraying this are shown in Fig. 4.14. It is apparent that in both phases all drying runs exhibited similar drying characteristics with almost no differences.  100  100  90  90 80  Control  M o istu re co n te n t (% )  Moisture content (%)  80 70  RF15  60  RF30  50  RF45 RF60  40  RF75 RF105  30  RF90  Control  70 60  RF15  50 40 RF60  30  RF75  RF120  20  20  RF120 RF90  10 0  10  20  30  40  50  60  70  80  10  90 100 110 120  0  10  20  30  40  Drying time (hrs)  50  60  70  80  90  100 110 120  Drying time (hrs)  Figure 4.13: Drying curves of drying runs in phase 1 (left) and phase 2 (right)  1.0  Control  Normalized moisture content  0.9 N orm alized m oisture content  1.0  Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120  0.8  0.7  0.6  RF15 RF60  0.9  RF75 RF90 RF120  0.8  0.7  0.6  0.5  0.5 0  5  10  15  20  25  30  35  40  0  45  5  10  15  20  25  30  35  40  45  Drying time (hrs)  Drying time (hrs)  Figure 4.14: Normalized drying curves of drying runs in phase 1 (left) and phase 2 (right)  60  From the drying data obtained, overall average dM/dt based on different moisture contents was calculated for each drying run in both phases. This was accomplished by calculating dM/dt at five-hour intervals and taking the average of the resultant dM/dt values for each group. Table 4.15 shows the dM/dt of all groups in both phases, and it can be seen that this was relatively higher in the controls than all RF heated groups. This could be due to the higher moisture content of the controls compared to the other groups. However, a closer look at Table 4.15 revealed that some groups with relatively lower moisture contents before drying recorded higher dM/dt than those with higher moisture contents. This trend is possibly due to differences in the number of wet-pockets per sample in each group as well as differences in the improvement of permeability as a result of RF heating. Table 4.15: Average drying rate of all drying runs RF heating RF treatment Average dM/dt groups time (mins) (%/hr) Control 0 0.85 RF15 15 0.78 RF30 30 0.78 RF45 45 0.57 Phase 1 RF60 60 0.65 RF75 75 0.53 RF90 90 0.59 RF105 105 0.44 RF120 120 0.42 Control 0 0.70 RF15 15 0.63 Phase 2 RF60 60 0.44 RF75 75 0.44 RF90 90 0.36 RF120 120 0.35 Due to differences in Mi between drying runs which could affect dM/dt values, there was a need to have common moisture content for all drying runs to compare dM/dt in both phases. The Mi of all drying runs in phase 1 was therefore normalized to 36% (the least Mi among the group) and that of phase 2 was normalized to 32% (for all groups in phase  2 with Mi above MFSP) and 26% (for groups with Mi less than MFSP). The normalization was done by assuming the aforementioned moisture contents as the initial for each group 61  during drying. From these normalized data, drying curves were plotted for each kiln drying run. A third degree polynomial was then fitted to the drying curves and the regression equation used to recalculate the moisture content. From this recalculated moisture content, dM/dt values were also recalculated. However, due to slight changes in the recalculated moisture contents, a second normalization was done for drying runs in both phases. In phase 1, Mi and Mf were normalized to 33% and 17%, respectively whereas those in phase 2 were normalized to 32% and 17% (for all runs with Mi above MFSP). Groups in phase 2 with Mi below MFSP were normalized to 26% and 17% Mi and Mf, respectively. The normalization was done with the assumption that the  aforementioned moisture contents were the initial and final of the groups during drying. Drying rate curves showing this are plotted in Figs 4.15, 4.16 and 4.17  RF15 1.2  1.0  1.0  0.8  0.8  dM/dt (%/hr)  dM/dt (%/hr)  Control 1.2  0.6 0.4 0.2  0.6 0.4 0.2  0.0  0.0 10  15  20 25 Moisture content (%)  30  35  10  15  20  25  30  35  Moisture content (%)  RF45 1.2  1.0  1.0  0.8  0.8  dM/dt (%/hr)  dM /dt (% /hr)  RF30 1.2  0.6  0.6  0.4  0.4  0.2  0.2 0.0  0.0 10  15  20  25  30  10  35  15  20 25 Moisture content (%)  30  Moisture content (%)  Figure 4.15: Normalized drying rate curves of drying runs in phase 1  62  35  RF60  RF75 1.2  1.0  1.0  0.8  0.8  d M /d t (% /h r)  dM /dt (% /hr)  1.2  0.6  0.6 0.4  0.4  0.2  0.2  0.0  0.0 10  15  20 25 Moisture content (%)  30  35  10  15  30  1.2  1.0  1.0  0.8  0.8  dM /dt (% /hr)  1.2  0.6 0.4 0.2  0.6 0.4 0.2 0.0  0.0 10  15  20 25 Moisture content (%)  30  35  30  35  10  15  20 25 Moisture content %)  30  RF120 1.2 1.0  dM /dt (% /hr)  35  RF105  RF90  dM /dt (% /hr)  20 25 Moisture content (%)  0.8 0.6 0.4 0.2 0.0 10  15  20 25 Moisture content (%)  Figure 4.16: Normalized drying rate curves of drying runs in phase 1  63  35  RF15 0.8  0.7  0.7  0.6  0.6 dM /dt (% /hr)  dM/dt (%/hr)  Control 0.8  0.5 0.4 0.3  0.5 0.4 0.3  0.2  0.2  0.1  0.1  0.0  0.0 10  15  20  25  30  10  35  15  20  Moisture content (%)  25  0.8  0.8  0.7  0.7  0.6  0.6 d M /d t (% /hr)  dM/dt (%/hr)  35  RF75  RF60  0.5  0.5  0.4  0.4  0.3  0.3  0.2  0.2  0.1  0.1  0.0  0.0 10  15  20 25 Moisture content (%)  30  10  35  15  RF90  20 25 Moisture content (%)  30  35  30  35  RF120  0.8  0.8  0.7  0.7  0.6  0.6 dM /dt (% /hr)  dM/dt (%/hr)  30  Moisture content (%)  0.5  0.5  0.4  0.4  0.3  0.3  0.2  0.2  0.1  0.1  0.0  0.0 10  15  20 25 Moisture content (%)  30  35  10  15  20  25  Moisture content (%)  Figure 4.17: Normalized drying rate curves of drying runs in phase 2  As mentioned before, the purpose of the study was to use RF heating to “open-up” the internal structure of sub-alpine fir, i.e., to improve air permeability and hopefully it’s drying characteristics. In timber kiln drying, improved permeability corresponds to 64  increased dM/dt, especially above MFSP where moisture movement is via bulk flow controlled by the permeability of the wood. To ascertain the effect of the RF heating on dM/dt, average dM/dt values above and below MFSP were calculated for drying runs in  phases 1 and 2. This was accomplished using the dM/dt data obtained by fitting a third degree polynomial to the drying curves. Average dM/dt values were subsequently calculated from 33-30% and 30-17% for phase 1 drying runs, whereas that of phase 2 was from 32-30% and 26-17%. These values are shown in Table 4.16. Table 4.16: Average drying rate above and below fiber saturation point for all drying runs RF heating groups Control RF15 RF30 RF45 Phase 1 RF60 RF75 RF90 RF105 RF120 Control RF15 Phase 2 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120  dM/dt above MFSP (%/hr) 0.74 0.92 0.95 0.85 0.90 0.75 0.88 0.58 0.18 0.56 0.68 0.66 0.73  dM/dt below MFSP (%/hr) 0.43 0.59 0.64 0.44 0.58 0.55 0.68 0.67 0.40 0.29 0.35 0.43 0.44 0.45 0.37  In phase 1, all groups except RF105 and RF120 had relatively higher dM/dt above MFSP than the control. The average dM/dt of the groups above MFSP ranged from 0.18%/hr in RF120 to 0.95%/hr in RF30. The percent increase in dM/dt compared to the control was from 1% in RF75 to 28% in RF30. The percent decrease however, ranged from 22% in RF105 to 76% in RF120. All groups in phase 2 had relatively higher dM/dt above MFSP than the control. The average dM/dt of the groups above MFSP ranged from 0.56%/hr in the control to 0.73 %/hr in RF75. The percent increase in dM/dt compared to the control was in the range of 18% in RF60 to 30% in RF75.  65  The movement of water during drying above MFSP is by capillary forces (Kollmann and Côté 1968) and according to Bramhall and Wellwood (1976) permeability is the most  important factor affecting capillary flow. Therefore, the observed higher dM/dt in some RF heated groups than the control in both phases can possibly be due to improved permeability caused by the creation of micro-cracks and voids in the wood samples as a result of RF heating. The expectation was however that the longer the RF heating the higher dM/dt should be because of the higher vapor pressure associated with longer periods of moisture evaporation in wood and the likelihood of more micro-voids been created.  A closer look at the dM/dt values above MFSP in Table 4.16 indicates that some groups with relatively shorter period of RF heating recorded higher dM/dt than those with longer periods. The likely reasons for this result is the possibility of differences in permeability after RF heating, caused by differences in the number of wet-pockets contained in the samples of the various groups. Another reason which can also be assigned to the higher dM/dt of the control than RF105 and RF120 is the differences in drying conditions  (temperature and relative humidity) caused by differences in Mi. Temperature can have some level of influence on dM/dt because moisture movement during drying above MFSP is controlled to some extent by the wood temperature. At higher wood temperatures the viscosity of water is reduced, so under given capillary pull water will flow more rapidly (Bramhall and Welwood 1976) resulting in rapid drying and higher dM/dt.  Average dM/dt below MFSP followed a different trend altogether from that above MFSP. In phase 1 all groups except RF120 had relatively higher dM/dt than the control. The lowest dM/dt of 0.40%/hr was recorded in RF120 whiles the highest of 0.68%/hr was in RF90.  Compared to the control increases ranging from 2% in RF45 to 58% in RF90 were observed. A decrease in dM/dt of 7% was also recorded in RF120. All groups in phase 2 recorded relatively higher dM/dt than the control. The lowest average dM/dt of 0.29%/hr was observed in the control whereas the highest of 0.45%/hr was in RF90. Increases ranging from 21% in RF15 to 55% in RF120 were recorded.  66  Drying below MFSP is largely by diffusion, and according to Kollmann and Côté (1968) this is due to moisture gradient set up across the cell wall. Simpson (1991) also indicated that the rate of diffusion depends to a large extent upon the permeability of the cell wall (permeable species dry faster than impermeable ones). Therefore, the higher dM/dt recorded by some groups in both phases compared to control could be due to a steeper moisture gradient created within the samples as a result of moisture loss during RF heating, subsequent drying to reach MFSP, and possible improvement in permeability from the RF heating. However, since stickers were not used, the RF heating was not like regular drying. Moreover, vacuum was not used so it cannot be likened to RF vacuum drying where moisture moves mostly in the longitudinal direction, so moisture probably dried from the side surfaces and created internal moisture gradient. The lower dM/dt recorded by RF120 in phase 1 than the control might be due to wet-pockets which restricted moisture diffusion during drying.  It can also be observed from Table 4.16 that some treatments with shorter period of RF heating recorded higher dM/dt than those with longer periods, although the opposite was expected. This trend could be due to differences in the number of wet-pockets which also resulted in differences in moisture loss, moisture gradient as well as improvement in permeability.  4.5 Effect of radio frequency heating on final moisture content variability between samples after kiln drying Table 4.17 and Fig. 4.18 show the variation in Mf of all drying runs in the main group. The Mf of all samples in phase 1 ranged from 3.1% to 51.6%, and the lowest and highest averages of the groups were 15.7% and 20.8%, respectively. The highest Mf was found in RF90, while the lowest was in RF120. In phase 2, Mf of all samples ranged from 5.2% to 30.6%, with the average of the groups ranging from 11.4% to 14.5%. The extreme Mf of 51.6% and 30.6% observed in phases 1 and 2, respectively, might probably be due to relatively higher number of wet-pockets contained in these samples, which were all not eliminated by the RF heating to allow for the free movement of water during drying. 67  Analysis of variance (Table 4.18) revealed significant differences in the final moisture content between groups in phase 1, but no differences in phase 2. However, comparison between RF heated groups and control in phase 1 showed no significant differences. Table 4.17: Moisture contents of main groups after kiln drying RF heating RF treatment Mean SD Min. Max. groups time (mins) (%) (%) (%) (%) (Max) - (Min) Control 0 17.5 7.7 7.1 42.3 35.2 RF15 15 20.1 7.4 9.2 36.5 27.3 RF30 30 17.2 6.0 9.6 35.1 25.5 RF45 45 16.4 6.8 8.0 33.3 25.3 Phase 1 RF60 60 19.5 8.7 8.0 42.3 34.4 RF75 75 17.0 9.2 7.0 50.9 43.9 RF90 90 20.8 10.5 3.1 51.6 48.5 RF105 105 17.1 10.9 6.9 46.9 40.1 RF120 120 15.7 7.3 8.5 42.1 33.6 Control 0 14.5 5.0 9.8 30.6 20.8 RF15 15 11.4 4.7 6.0 27.2 21.3 RF60 60 12.9 6.2 5.2 29.7 24.6 RF75 75 11.5 5.4 5.8 25.8 20.0 Phase 2 RF90 90 11.7 4.9 5.2 24.8 19.6 RF120 120 11.4 5.6 6.1 24.8 18.7  110 Phase 1 Phase 1 Phase 2 Phase 2  100  Moisture content (%)  90  Initial moisture content Final moisture content Initial moisture content Final moisture content  80 70 60 50 40 30 20 10 Control RF15  RF30  RF45 RF60 RF75 RF heating groups  RF90  RF105 RF120  Figure 4.18: Average moisture content before and after kiln drying (error bars indicate ±1SD) 68  Table 4.18: Analysis of variance of average final moisture content of main groups  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 2.22 48.24 50.47 317.06 6632.46 6949.52  df 8 351 359 5 234 239  MS 0.28 0.14  F 2.02  p-value 0.043  63.41 28.34  2.24  0.051  The final moisture contents of the sub-groups are shown in Table 4.19, and in phase 1 the moisture content ranged from 9.2% in RF120 to 51.6% in RF90. The highest average moisture content was in RF90 whereas the lowest was in RF120. In phase 2 however, this ranged from 6.3% in RF15 to 30.6% in the control. The highest and lowest moisture content were recorded in the control and RF15 respectively. The relatively higher moisture contents recorded by some samples in both phases could possibly be due to wetpockets which the RF heating could not eliminate to allow for free flow of moisture during drying. Analysis of variance showed significant differences in the average moisture content between the groups in phase 1 (α = 0.05 and p = 0.002) but no difference in phase 2 (α = 0.05 and p = 0.224). The subsequent pair-wise comparison of the groups in phase 1 revealed no significant differences between the RF heated ones and the control.  The Mf distributions of all drying runs in the main group are also shown in Figs. 4.8, 4.9 and 4.10. It is evident that most samples reached Mf values in the 10-20% moisture content range. Table 4.17 shows the variability (standard deviations) of Mf between pieces of lumber in the main groups of both phases. In phase 1, the lowest and highest variability of 6% and 10.9% were recorded in RF30 and RF105, respectively. Compared to the control, all groups except RF60, RF75, RF90 and RF105 recorded reduced moisture content variability between pieces. This reduction ranged from 4% in RF15 to 22% in RF30. However, RF60, RF75, RF90 and RF105 respectively recorded 14%, 21%, 36% and 43% increases in moisture content variability when compared to the control. In phase 2, the variability of Mf ranged from 4.7% in RF15 to 6.2% in RF60. All except RF15 and RF90 had relatively higher standard deviations than the control. Reduction in 69  standard deviation of 2 to 6%, and increases of 8 to 23% were observed in groups compared to control.  Phase 1  Phase 2  Table 4.19: Moisture contents of sub-groups after kiln drying RF heating RF heating Mean SD Min. Max. groups time (mins) (%) (%) (%) (%) (Max) - (Min) Control 0 25.1 10.4 15.4 42.3 26.9 RF15 15 22.6 8.9 9.6 36.5 26.9 RF30 30 17.4 6.1 9.6 27.6 18.0 RF45 45 20.0 8.2 10.3 33.3 23.0 RF60 60 27.5 10.9 13.9 42.3 28.5 RF75 75 30.0 11.4 15.4 50.9 35.5 RF90 90 33.9 10.6 20.6 51.6 31.0 RF105 105 32.9 13.9 10.9 46.9 36.0 RF120 120 16.3 10.8 9.2 42.1 32.9 Control 0 18.1 6.0 10.6 30.6 19.9 RF15 15 13.5 6.3 6.3 27.2 20.9 RF60 60 17.1 6.8 7.5 29.7 22.2 RF75 75 14.8 5.2 6.7 25.6 18.8 RF90 90 14.3 5.4 7.3 24.8 17.5 RF120 120 15.1 6.6 7.0 24.8 17.8  The standard deviations of the final moisture content of sub-groups in phase 1 ranged from 6.1% in RF30 to 13.9% in RF105 (Table 4.19). All except RF15, RF30 and RF45 had relatively higher standard deviations than the control, and increase in standard deviations ranging from 1-33% as well as reductions of 14-41% were observed between control and RF heated groups. In phase 2 however, the standard deviations ranged from 5.2% in RF75 to 6.8% in RF60, and all except RF75 and RF90 had higher standard deviations than the control. Compared to control, standard deviations increased by 4-13% and also reduced by 10-14%. The Bartlett’s test however, showed no significant differences in standard deviations between groups in phase 1 (α = 0.05 and p = 0.745) and phase 2 (α = 0.05 and p = 0.897).  The relatively higher standard deviations of Mf in some of the groups compared to the control is possibly due to the fact that RF heating had a relatively greater effect on some of the wood samples than others, possibly because several were dry with no wet-pockets while others still had wet-pockets. This is possible because of likely differences in the 70  number and size of wet-pockets contained in each sample. Relatively smaller-sized wetpockets are completely eliminated by the RF heating whereas larger ones are slightly altered. This claim is supported by the moisture distribution along the length of samples with lowest, medium and highest Mf in each of the groups with higher standard deviations than the control in phase 2 (Fig. 4.19). It is evident from these plots that whereas some of the samples had no or less wet-pockets along their length, others still had pockets of wetwood with relatively high moisture content, a situation which can result in wide moisture content distribution and consequently higher standard deviations.  RF75 45  40  40 Moisture content (%)  Moisture content (%)  RF60 45  35 High MC  30 25 Medium MC  20 15  High MC  30 25 20  Medium MC  15 Low MC  10  Low MC  10  35  5  5 0  100  200  300  400  500  600  700  800  0  900  100  200  300  400  500  600  700  800  900  Length of wood (mm)  Length of wood (mm)  RF120 45  Moisture content (%)  40 35  High MC  30 25 Medium MC  20 15 10  Low MC  5 0  100  200  300  400  500  600  700  800  900  Length of wood (mm)  Figure 4.19: Distribution of moisture content along the length of highest, medium and lowest moisture content samples in RF60, RF75 and RF120 in phase 2  71  4.6 Effect of radio frequency heating on moisture content variability within wood samples after kiln drying The moisture content variability within samples was determined by the difference between the core (Mc) and shell (Ms) moisture contents. The average Mc-Ms difference of the drying runs (main groups) is given in Table 4.20 and Fig. 4.20. In phase 1, the lowest and highest Mc-Ms difference of 7.9% and 16.4% were observed in RF120 and RF90, respectively. With the exception of RF15, RF60, RF75 and RF90, all groups had lower Mc-Ms difference than the control. This reduction ranged from 9% in RF105 to 39% in  RF120. The RF15, RF60, RF75 and RF90 groups on the other hand, respectively recorded 16%, 22%, 8% and 26% higher Mc-Ms difference than the control. This higher Mc-Ms difference in some of the groups than the control might be due to relatively larger-  sized wet-pockets in some samples which were not completely eliminated during RF heating and were sent to the kiln for drying. However, moisture movement restrictions caused by these wet-pockets during drying possibly led to these samples remaining wet after drying. Table 4.20: Core-shell moisture content differences of main groups after kiln drying  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120  Mean (%) 13.0 15.0 10.1 10.0 15.9 14.0 16.4 11.8 7.9 11.8 11.4 9.5 10.8 6.6 10.3  72  SD (%) 11.5 10.9 11.6 10.0 13.6 16.8 15.3 15.6 10.6 9.7 8.9 10.2 9.2 7.1 11.6  Min. (%) 0.4 0.4 22.4 1.9 1.1 4.8 0.2 1.0 9.2 1.5 0.0 2.4 0.2 1.3 1.7  Max. (%) 48.7 37.0 41.9 41.8 50.6 79.1 63.2 56.8 46.4 40.8 37.1 35.1 37.3 25.7 39.5  (Max) - (Min) 48.3 36.6 19.5 39.9 49.5 74.3 63.0 55.8 37.3 39.3 37.1 32.7 37.1 24.4 37.8  Core-shell moisture content difference (%)  37  Phase 1 Phase 2  32 27 22 17 12 7 2 Control RF15  RF30  RF45  RF60  RF75  RF90  RF105 RF120  RF heating groups  Figure 4.20: Average core-shell moisture content differences of all drying runs (error bars indicate ±1SD) In phase 2, the lowest average Mc-Ms difference of 6.6% was recorded in RF90 while the highest of 11.8% was in the control. All groups had relatively lower average Mc-Ms difference than the control, and this reduction ranged from 3% in RF15 to 44% in RF90. This trend is possibly due to improved permeability which has the effect of improving moisture movement during drying and in turn, can lead to narrow Mc-Ms difference. The Mc-Ms differences of the sub-groups are presented in Table 4.21, where it can be seen that the highest average in phase 1 was recorded by RF90 whereas the lowest was in RF120. All groups except RF15, RF30, RF45 and RF120 had higher Mc-Ms difference than the control. This trend could possibly be due to large number of wet-pockets contained in the samples of these groups which were not completely eliminated during RF heating to allow for free flow of moisture during drying. The Mc-Ms difference in phase 2 however, followed a different trend altogether. The highest average of 18.8% was in the control whereas the lowest of 9.1% was in RF90, and all RF heated groups had relatively lower average Mc-Ms difference than the control. This could be due to  73  improved permeability from RF heating, which in turn allowed for free flow of moisture during drying. Table 4.21: Core-shell moisture content differences of sub-groups after kiln drying RF heating RF treatment Mean SD Min. Max. groups time (mins) (%) (%) (%) (%) (Max) - (Min) Control 0 24.2 13.1 11.3 48.7 37.4 RF15 15 20.9 12.6 2.1 37.0 34.9 RF30 30 12.4 10.2 1.7 31.9 30.2 RF45 45 15.5 14.3 1.3 41.8 40.5 Phase 1 RF60 60 29.1 17.1 5.9 50.6 44.7 RF75 75 35.6 25.4 6.0 79.1 73.2 RF90 90 36.8 17.4 12.6 63.2 50.6 RF105 105 35.4 19.3 4.1 56.8 52.7 RF120 120 10.1 15.1 0.3 46.4 46.1 Control 0 18.8 11.1 6.9 40.8 33.9 RF15 15 14.7 12.0 0.0 37.1 37.1 Phase 2 RF60 60 17.3 10.1 2.3 35.1 32.8 RF75 75 16.6 8.5 0.1 37.3 37.2 RF90 90 9.1 7.2 0.5 25.7 25.2 RF120 120 18.3 14.0 1.1 39.5 38.4 Due to significant differences in the moisture content between the different groups before drying, comparing Mc-Ms difference should best be done with analysis of covariance which makes an adjustment for the initial differences in moisture content (covariate). This adjustment assumes that the groups have similar regression coefficients. When the coefficients are not similar (like the Mc-Ms difference data) the effect of the adjustment will be different for dissimilar values of the covariate. Since the assumption of similar regression coefficients could not be met, although Mi was found to have a significant effect on the average Mc-Ms difference, analysis of covariance could not be used to analyze these data. The only option left was the use of analysis of variance (Table 4.22) which showed significant differences in Mc-Ms difference between the main groups in phase 1, but no difference in phase 2. Comparison of the groups to the control in phase 1 revealed no significant differences.  74  Table 4.22: Analysis of variance of average core-shell moisture content differences  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 113.73 2302.39 2416.12 0.05 9.97 10.46  df 8 351 359 5 234 239  MS 14.22 6.56  F 2.17  p-value 0.029  0.09 0.04  2.28  0.05  The ANOVA of the sub-groups also showed significant differences between the groups in phase 1 (α = 0.05 and p = 0.0016) but no difference in phase 2 (α = 0.05 and p = 0.964). The pair-wise comparison was done for the groups in phase 1 (Table 4.23), and it revealed no significant differences between the RF heated groups and control.  Table 4.23: Pair-wise comparison of average core-shell moisture content differences of sub-groups in phase 1 RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120  Mean 24.2 20.9 12.4 15.5 29.1 35.6 36.8 35.4 10.1  Grouping A B A B A B A B A B A A A B  4.7 Wood quality after radio frequency heating and kiln drying All wood samples were evaluated before and after RF heating and kiln drying for surface checking. No surface checks were observed on side and end surfaces of the samples before RF heating. However, after heating, some samples developed surface checking and no change in the already existing surface checks was observed after drying. The initial intention was to evaluate these checks based on the length and possibly thickness, but because some of them were very small and close to each other, it was impossible to 75  evaluate the checks based on the length and thickness. Therefore, samples within each group with surface checks were counted and the severity of the checks noted. Figure 4.21 shows the number of samples with surface checks within each main group in phases 1 and 2. It is clear that no surface checking was found in the controls in both phases, which indicates that surface checking was caused primarily by RF heating. In phase 1, all samples in RF105 had one or more surface checks whereas only 18% of samples in RF15 had surface checks. In phase 2 however, 98% of samples in RF120 had one or more surface checks whereas 93% of the samples in both RF15 and RF60 had surface checks. The number of samples with surface checks in the sub-groups is presented in Table 4.24.  Number of samples with surface checks  50 Phase 1 Phase 2  45 40 35 30 25 20 15 10 5 0 Control RF15  RF30  RF45  RF60  RF75  RF90  RF105 RF120  RF heating groups  Figure 4.21: Number of samples with surface checks after RF heating  As shown in the sub-groups (Table 4.24), no surface checks were found in the controls in both phases and RF15 in phase 1. In phase 1, all samples in RF105 and RF120 had one or more surface checks while only 13% of samples in RF30 and RF45 had surface checks. In phase 2 however, all samples in RF120 had one or more surface checks compared to 81% in RF15. 76  Table 4.24: Number of samples with surface checks in sub-groups RF Number of Number of samples RF heating treatment samples with with surface checks groups time (mins) surface checks (%) Control 0 0 0 RF15 15 0 0 RF30 30 1 13 RF45 45 1 13 Phase 1 RF60 60 5 63 RF75 75 7 88 RF90 90 6 75 RF105 105 8 100 RF120 120 8 100 Control 0 0 0 RF15 15 13 81 Phase 2 RF60 60 15 94 RF75 75 16 100 RF90 90 15 94 RF120 120 16 100  According to Bramhall and Wellwood (1976) surface checking can be caused by differential shrinkage between the shell and core of wood which result in shrinkage stresses. They went on to say that when these shrinkage stresses exceed the strength of the wood across the grain, surface checking along the wood rays occur. Therefore, this relatively greater occurrence of surface checking in the RF heated groups could be due to shrinkage stresses which exceeded the strength of the wood across the grain and resulted in checking along the wood rays. This is probably because the surface of wood lost much moisture in the early stages of the RF heating and began to shrink while the core was still wet, resulting in differential shrinkage and consequently, shrinkage stresses.  There was no clear-cut relationship between the number of samples with surface checking and the duration of RF heating in both the main and sub-groups. In the main groups in phase 1 for instance, RF heating duration of 75 minutes led to 10% more samples having surface checks than a heating duration of 90 minutes. The same can also be said of phase 2 in which a heating duration of 15 minutes led to the same number of samples with surface checks as the 60 minutes treatment. This could be due to differences 77  in shrinkage stresses during RF heating. It must, however, be emphasized that the severity of the surface checking tended to increase with the RF heating time (Figs. 4.22 and 4.23). Thus, the less severe surface checking was found in RF15 whereas the most severe was in RF120. Most checks observed were however, superficial and could easily be eliminated in planing without compromising the quality of the final product. As stated by Simpson (1991) superficial surface checks that can be removed during machining are not a problem.  Figure 4.22: Surface checking in RF120 (left) and RF15 (right) samples in phase 1  Figure 4.23: Surface checking in RF120 (left) and RF15 (right) samples in phase 2 All samples were also evaluated for internal checking (honeycomb) after kiln drying. In the main and sub-groups in phase 1, no honeycomb was found in the control, but it did appear in the RF heated samples. This points to the fact that honeycomb was caused largely by the RF heating. This result is not surprising because when the wood’s internal temperature in RF heating reaches 100oC there is rapid vaporization of water. This results in the generation of internal steam pressure that when it exceed the strength of the cellwall, micro-checking that propagates to full level honeycomb might result. In phase 2 78  (both main and sub-groups) however, samples in the control as well as the RF heated groups had honeycomb. Simpson (1991) reported that honeycomb is caused by tension failure along the grain of wood, as a result of internal tension stresses in the core when Mc is too high and drying temperatures are maintained at high levels for a long time.  Therefore, the honeycomb observed in the control of phase 2, could be due to tension stresses which developed in the core of the wood because it was still wet and the drying temperature (probably during the last step of the schedule) was maintained at a higher level for a long time.  The percentage of samples with honeycomb within each main group in phase 1 ranged from zero in the control to 93% in RF120, whereas that in the sub-groups ranged from zero in the control to 100% in RF105. In the main groups in phase 2, 5% of samples in the control and 95% in RF90 had honeycomb whereas in the sub-groups, 6% of samples in the control and 100% in RF120 had honeycomb. In phase 1, the honeycomb was most severe in RF120 and least severe in RF15 (Figs. 4.24). In phase 2, the least severe honeycomb was in the control whiles the most severe was in RF120 (Fig. 4.25). This could be due to differences in the magnitude of internal pressure generated within each group as a result of RF heating to vaporize the water in the wood. Steam pressure was expected to be greater in RF120 than RF15 because of the longer period of heating in the former.  Figure 4.24: Honeycomb in RF120 (left) and RF15 (right) in phase 1.  79  Figure 4.25: Honeycomb in RF120 (left) and control (right) in phase 2  The initial intention was to evaluate these checks based on their number, length and width. However, some appeared as very small dot-like cracks and it was impossible to measure their length, let alone their width. Honeycomb was therefore evaluated based on the counts found in each sample as well as the severity. The evaluation of checks based on counts has been previously used by Harris et al. (2008). The average honeycomb count for each drying run in the main and sub-groups are shown in Fig. 4.26 and Table 4.25 respectively. In phase 1 (main groups), the lowest average honeycomb count of zero was found in the control whereas the highest of eight was in RF45 and RF60. In the subgroups however, the control had zero counts while RF120 recorded 17 counts. In the main group in phase 2, the lowest average of 5 was in RF15 whereas the highest of 10 was found in the control and RF90. However, in the sub-groups in phase 2, the lowest average of one was in the control whereas the highest of 10 was in RF90.  Analysis of variance (Table 4.26) revealed significant differences in honeycomb counts between the main groups in both phases. Analysis of variance for the sub-groups showed no significant differences between the groups in phase 1 (α = 0.05 and p = 0.304) but did show significant differences between the groups in phase 2 (α = 0.05 and p = 0.0001). In the main groups the subsequent pair-wise comparison with the Tukey-test (Table 4.28) indicated significant differences between the control, RF45 and RF60 in phase 1. In phase 2 however, no significant differences were observed between the control and all other groups. Pair-wise comparisons of the sub-groups in phase 2 are shown in Table 80  4.27. Figures in this table show significant differences between RF90, RF120 and the control.  25 Phase 1 Phase 2  Honeycomb count  20  15  10  5  0 Control  RF15  RF30  RF45  RF60  RF75  RF90  RF105 RF120  RF heating groups  Figure 4.26: Average honeycomb count of all drying runs (error bars indicate ±1SD)  Table 4.25: Average honeycomb count of sub-groups  Phase 1  Phase 2  RF heating groups Control RF15 RF30 RF45 RF60 RF75 RF90 RF105 RF120 Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 30 45 60 75 90 105 120 0 15 60 75 90 120 81  Mean 0 4 4 16 10 9 10 5 17 1 2 3 5 10 8  SD 0.0 8.3 4.8 21.3 8.5 14.3 16.3 1.4 26.7 3.0 3.7 3.5 4.4 6.6 6.7  Table 4.26: Analysis of variance of average honeycomb count in the main groups  Phase 1  Phase 2  Source of Variation Between Groups Within Groups Total Between Groups Within Groups Total  SS 1955.35 23595.025 25550.375 9.23456 104.87305 114.1076  df 8 351 359 5 157 162  MS 244.419 67.222  F 3.636  p-value 0.0004  1.84691 0.66798  2.7649  0.0201  Table 4.27: Pair-wise comparison of average honeycomb count of sub-groups in phase 2 RF heating groups Control RF15 RF60 RF75 RF90 RF120  RF treatment time (mins) 0 15 60 75 90 120  Mean Grouping 1 C 2 C 3 B C 5 A B C 10 A 8 A B  Table 4.28: Pair-wise comparison of average honeycomb count of main groups RF heating RF treatment groups time (mins) Mean Grouping Control 0 0 B RF15 15 4 A B RF30 30 3 A B RF45 45 8 A Phase 1 RF60 60 8 A RF75 75 4 A B RF90 90 4 A B RF105 105 5 A B RF120 120 5 A B Control 0 10 A B RF15 15 5 B RF60 60 8 A B Phase 2 RF75 75 7 A B RF90 90 10 A RF120 120 9 A B  82  4.8 Energy consumption in radio frequency heating and kiln drying Energy consumption in both RF heating and kiln drying was calculated for each group to ascertain the feasibility of applying this technology in an industrial setting. As shown in Table 4.29, RF energy consumption in phase 1 ranged from 8 kW-hr in RF15 and RF30 to 16 kW-hr in RF120, whereas that in phase 2 was from 14 kW-hr in RF15 to 30 kW-hr in RF120. In both phases, the energy consumption was doubled as the heating time increased from 15 minutes to 120 minutes. This is however, not surprising because the longer the heating time, the more energy is consumed. This point is buttressed by Fig. 4.27 which clearly reveals a strong linear relationship between RF heating time and energy consumption in both phases. Thus, at the same PD, energy consumption will tend to increase as the heating time increases. Comparing the RF energy consumption in phases 1 and 2, it is evident that this was relatively higher in phase 2 than in phase 1 although heating time in the former was relatively shorter. This can be explained by the higher PD (72 kW/m3) used in phase 2 compared to that of phase 1 (27 kW/m3).  Phase 1  Phase 2  Table 4.29: Energy consumption in RF heating RF heating RF treatment Total RF heating Energy consumption in RF groups time (mins) time (hrs) heating (kW-hr) RF15 15 1.8 8 RF30 30 1.9 8 RF45 45 2.6 11 RF60 60 2.5 10 RF75 75 2.8 11 RF90 90 3.3 13 RF105 105 3.7 15 RF120 120 3.9 16 RF15 15 1.1 14 RF60 60 2.0 25 RF75 75 2.2 25 RF90 90 2.4 27 RF120 120 2.9 30  83  Energy consumption (kW-hr)  35 R2 = 0.9595  30 25  Phase 2 20 R2 = 1 15 Phase 1 10 5 1  2  3  4  5  Total RF heating time (hrs)  Figure 4.27: Relationship between total RF heating time and energy consumption  Phase 1  Phase 2  Table 4.30: Energy consumption in kiln drying RF heating RF treatment Average moisture Energy consumption groups time (mins) content before drying (%) in kiln drying (kW-hr) Control 0 87.4 28 RF15 15 70.9 24 RF30 30 69.8 24 RF45 45 56.6 20 RF60 60 60.9 21 RF75 75 49.7 17 RF90 90 47.5 17 RF105 105 39.3 13 RF120 120 35.9 12 Control 0 87.8 28 RF15 15 60.2 21 RF60 60 38.3 13 RF75 75 31.8 10 RF90 90 28.9 7 RF120 120 26.7 7  Energy consumption in kiln drying was calculated based on the amount of water removed from each drying group to attain the required Mf. The energy consumption in phase 1 ranged from 12 kW-hr to 28 kW-hr in RF120 and control respectively (Table 4.30). 84  Phase 2 followed the same trend; with energy consumption ranging from 7 kW-hr in RF120 to 28 kW-hr in the control. This trend can be attributable to the higher moisture content of the controls relative to RF120. Much moisture therefore needed to be taking out, and this prolonged the drying time. Longer drying times result in higher energy consumption (Fig. 4.28).  30  Energy consumption (kW-hr)  R2 = 0.8452 R2 = 0.9797  25  Phase 1  20  Phase 2 15  10  5 20  40  60  80  100  120  140  Total drying time (hrs)  Figure 4.28: Relationship between total kiln drying time and energy consumption The total energy consumption in both RF heating and kiln drying was also calculated for each drying run (Table 4.31). In phase 1 the lowest energy consumption of 28 kW-hr was recorded in the control, RF105 and RF120 whiles the highest of 32 kW-hr was in RF15 and RF30. However, in phase 2, the lowest of 28 kW-hr was in the control whereas the highest of 37 kW-hr was in RF60 and RF120. All groups in phase 1 except RF105 and RF120 consumed much energy than the control while in phase 2 the energy consumption in all the groups was higher than the control. This can be explained by the longer RF heating time in phase 1 and the relatively higher power applied in phase 2.  85  Table 4.31: Total energy consumption in RF heating and kiln drying RF heating RF treatment Total energy groups time (mins) consumption (kW-hr) Control 0 28 RF15 15 32 RF30 30 32 RF45 45 30 Phase 1 RF60 60 31 RF75 75 29 RF90 90 30 RF105 105 28 RF120 120 28 Control 0 28 RF15 15 34 Phase 2 RF60 60 37 RF75 75 34 RF90 90 34 RF120 120 37  4.9 Hypothesis test The hypothesis for this study was that if RF heating methods improve the permeability of sub-alpine fir, then after kiln drying Mf variability between and within lumbers as well as drying defects will decrease. With permeability tests performed only in phase 2, hypothesis could only be applicable to experiments performed in this phase of the study. However, it is evident from the discussion that permeability did not significantly improve between RF heated groups and control for the range of inputs and sample size tested, so did moisture content variability within and between lumbers as well as drying defects. The hypothesis was therefore rejected within the experimental constraints of this study.  86  5. CONCLUSION The objective of the study was to reveal the effect of RF heating as pre-treatment on the kiln drying characteristics and quality of sub-alpine fir 51 x 102 mm lumber. In the light of this investigation, the following conclusions can be made for the two phases:  Phase 1 1. Increase in dM/dt of 1-28% above MFSP was observed between treatments and control. Treatments also resulted in 22-76% decrease in dM/dt above MFSP. 2. Increase in dM/dt of 2-58% as well as decrease of 7% below MFSP was observed between treatments and control. 3. Treatments resulted in reduced final moisture content variability of 4-22%, as well as increases of 14-43% in the main groups. Treatments also reduced final moisture content variability of the sub-groups by 14-21% and also increased by 133%. 4. Reduction in moisture gradient ranging from 9-39% and increases ranging from 16-26% were observed between treatments and control in the main groups, but was found not to be statistically significant. Treatments reduced moisture gradient by 14-58% and also increased by 20-52% in the sub-groups, but was found not to be statistically significant. 5. All except RF45 and RF60 in the main groups resulted in honeycomb counts not significantly different from the control. Honeycomb count in sub-groups was not significantly different. 6. Radio frequency heating of more than 90 minutes does not increase total energy consumption (sum of RF heating and kiln drying energies).  Phase 2 1. All except RF15 improved longitudinal permeability of sub-alpine fir, but was found not to be significantly different from the control. 2. Treatments resulted in increased dM/dt of 18-30% above MFSP. 3. All except RF15 and RF90 in the main groups resulted in higher final moisture content variability than control, but was found not to be statistically significant. 87  Final moisture content variability of sub-groups increased in all treatments except RF75 and RF90 but was found not to be statistically significant. 4. Treatments reduced moisture content gradient by 3-44% in the main groups, but was found not to be statistically significant. Moisture gradient in the sub-groups also reduced by 3-52%, but was not statistically significant. 5. Honeycomb count in all main groups was not statistically different from the control. All except RF90 and RF120 in the sub-groups had honeycomb count not significantly different from control. 6. Total energy consumption in RF heated groups is relatively higher than the control.  Both phases 1. There is a direct relationship between total RF heating time and moisture loss. 2. Drying rate is more pronounced below MFSP. 3. There is a direct relationship between severity of both surface and internal checking and total RF heating time. 4. There is a direct relationship between both RF heating and kiln drying times and energy consumption. Although the range of inputs and the sample size used in this study led to the rejection of the hypothesis, the results have revealed the potential for pre-drying RF heating to improve the permeability and possibly drying characteristics of sub-alpine fir ( and other less permeable wood species) because of the increase in longitudinal permeability in phase 2. Because all possible options were not exhausted, the results of this study can serve as a useful guide for future modifications of the experimental design to optimize this potential pre-drying treatment scheme.  88  6. RECOMMENDATIONS FOR FUTURE RESEARCH The range of inputs and sample size of this study led to the rejection of the hypothesis. However, since all possibilities were not exhausted the following recommendations based on observations from this study are made for future research.  This research aimed at using RF heating as a treatment before kiln drying to eliminate wet-pockets and reduce final moisture content variability between and within lumber after drying. Results obtained however, indicated higher variability between pieces in some RF heated groups than the control. This is believed to be due to the higher standard deviation of moisture content of kiln loads before RF heating. This resulted in the heating having relatively greater effect on the lower moisture content samples and less effect on higher ones. This high initial standard deviation of moisture content also restricted the use of relatively high power in RF heating due to the likelihood of severe occurrence of surface and internal checking in the samples. Future research should therefore, focus on sorting samples into groups of similar moisture content before RF heating, and applying the appropriate power (lower power for lower moisture content groups and vice-versa) during the heating treatment. This way, the effect of the RF heating will be similar on all samples and potentially help reduce final moisture content variability after drying as well as the occurrence of defects.  This study also used RF heating to “open-up” the internal structure of sub-alpine fir to improve permeability and also allow free movement of water during kiln drying. This was also expected to help to significantly reduce moisture gradient after kiln drying. Results obtained in this study however, proved otherwise, because the treatment temperature as well as the power densities used in this study was not strong enough to significantly improve the permeability of sub-alpine fir. Future research should look into using higher power densities (say, 90, 110 and 120 kW/m3) and temperatures (say, 110 and 120oC) in RF heating.  89  Probably a good understanding of moisture movement in wet-pockets could help improve this pre-treatment scheme. Additional fundamental work should therefore be carried out in order to understand moisture movement mechanism from wet-pockets and how RF could interact with that mechanism to produce the desirable results of faster drying rates and moisture content uniformity.  90  7. REFERENCES Alden, Harry A. 1997. Softwoods of North America. Gen. Tech. Rep. FPL–GTR–102. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. 151 pp. Avramidis, S. 1999. Radio frequency vacuum drying of wood. In: Proc. of the COST E15 Wood Drying Conference, Edinburgh, UK, Oct. 13-14, 1999. 125-138. Avramidis, S. 2008. Pasteurization of Lumber Using Dielectric Heating: A Preliminary Study. Final project Report for the Canada Wood Group. 27 pp. Avramidis, S. and Dubois, J. 1992. The study of dielectric properties of spruce, hemlock, western red cedar and Douglas-fir at varying moisture content, temperature, grain direction and radio frequency. Report No. 93 (SA-3), Science Council of British Columbia, BC, Canada. 50 pp. Avramidis, S. and Ruddick, J. N. R. 1996. CCA accelerated fixation by dielectric heating. Forest Prod. J. 46(7-8): 52-55. Avramidis, S. and Zhou, B. 1999. On the loss factor of wood during radio frequency heating. Wood Sci. Technol. 33(4): 299-310. Avramidis, S., Liu, F. and Neilson, B. J. 1994. Radio-frequency/vacuum drying of softwoods: Drying of thick western red cedar with constant electrode voltage. Forest Prod. J. 44(1): 41-45. Bachrich, J. L. 1980. Dry Kiln Handbook, Vancouver: H.A. Simons (International) Ltd., 374 pp. Bamber, R. K. and Fukazawa, K. 1985. Sapwood and heartwood: a review. For. Abstr. 46(9): 567-580. Bauch J., Hőll, W. and Endeward, R. 1975. Some aspects of wetwood formation in fir. Holzforschung 29(6): 198-205. British Columbia Forest Service, Ministry of Forests. 2003. Timber supply and the mountain pine beetle infestation in British Columbia. British Columbia Ministry of Forests and Range. 2008. Tree Book: Learning to recognize trees of British Columbia. 130 pp. Biryukov, V. A. 1961. Dielectric heating and drying of wood. Israel Program for Scientific Translation Ltd. 116 pp. Bowyer, J. L., Shmulsky, R. and Haygreen, J. G. 2007. Forest Products and Wood Science, an Introduction. Blackwell Publishing. 558 pp. 91  Bradic, S. and Avramidis, S. 2007. Longitudinal air permeability of pinewood with beetle transmitted blue-stain. Holz Roh Werkst 65(3): 183-185. Bramhall, G. 1995. Diffusion and the drying of wood. Wood Sci. Technol. 29(3): 209215. Bramhall, G. and Wellwood, R. W. 1976. Kiln drying of western Canadian lumber. Information Report, VP-X-159. 112 pp. Bramhall, G. and Wilson, J.W. 1971. Axial gas permeability of Douglas fir microsections dried by various techniques. Wood Sci. Technol. 3(4): 223-230. Cai, L., and Oliveira, L. C. 2007. Gas permeability of wetwood and normal wood of subalpine fir in relation to drying. Drying Technol. 25 (5): 501-505. Cai, L., and Oliveira, L. C. 2008. A simulation of wet pocket lumber drying. Drying Technol. 26(5): 525-529. Callebaut, J. 2007. Dielectric heating. Power Quality and Utilization Guide. 9 pp. Chafe, S.C. 1996. Drying wetwood-infected hoop pine. Holzforschung. 50(1): 55-61. Chen, P. Y. S. 1975. The effect of steaming and temperature on the longitudinal permeability of black walnut. Wood and Fiber. 7 (3): 222-225. Cloutier, A. and Fortin, Y. 1993. A model of moisture movement in wood based/water potential and the determination of the effective water conductivity. Wood Sci. Technol. 27(2): 95-114. Comstock, G. L. 1965. Longitudinal permeability of green eastern hemlock. Forest Prod. J. 15(10): 441-449. Comstock, G. L. 1967. Longitudinal permeability of wood to gases and non-swelling liquids. Forest Prod. J. 17(10): 41-46 Cooper, P. and Jeremic, D. 1998. Overview of the causes and occurrence of wet pockets. In: Proceedings of the Quality Lumber Drying in the Pacific North-West, Penticton, BC, Canada. Forest Products Society Publication No. 7271, USA. 45-53. Council of Forest Industries. 2009. COFI Market Access Newsletter. International Market Development Activities Update. March 2009 Edition. 17 pp. Coutts, M. P. and Rishbeth, J. 1977. The formation of wetwood in grand fir. Eur. J. For. Pathol. 7(1): 13-22. Crank, J. 1975. The mathematics of diffusion. Second edition. Oxford Univ. Press, London. 414 pp. 92  Denig, J., Wengert, E. M. and Simpson, W. T. 2000. Drying hardwood lumber. Gen. Tech. Rep. FPL–GTR–118. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. 138 pp Department of Resources and Development, Forestry Branch, Canada. 1950. Native Trees of Canada. 293 pp. Efransjah, F., Kilbertus, G. and Bucur, V. 1989. Impact of water storage on mechanical properties of spruce as detected by ultrasonics. Wood Sci. Technol. 23 (1). 35-42. Franc, G. D. 1994. Bacterial Wetwood Slime Flux. University of Wyoming Cooperative Extension Service. 4 pp. Hadden, C. H., Windham, A. S., Hopper, G. M. and Clatterbuck, W. K. 2004. Bacterial Wetwood Disease of Trees. UT Extension. 2 pp. Hansom, O. P. 1988. Contemporary Timber Drying. TBL 60, Timber Research and Development Association (TRADA), Buckinghamshire. UK. 84 pp. Harris, G. A., Torgovnikov, P., Vinden, P., Brodie, G. I. and Shaginov, A. 2008. Microwave pretreatment of backsawn messmate boards to improve drying quality: Part 1. Drying Technol. 26 (5): 579-584. Hart, C. A. and Thomas, R. J. 1967. Mechanism of bordered pit aspiration as caused by capillarity. Forest Prod. J. 17(11): 61-67.  Hartley, C., Davidson, R. W. and Crandall, B. S. 1961. Wetwood, bacteria, and increased pH in trees. USDA Forest Products Laboratory Report 2215. 35 pp. Hearmon, R. F. S. and Burcham, J. N. 1954. The dielectric properties of wood. Department of Science and Industry Research, London. Forestry Products Research Special Report No. 8. Hoadley, R. B. 1990. Identifying wood: Accurate Results with Simple Tools. The Taunton Press. 223 pp. Hong-Hai, L., Qing-Wen, W., Lin, Y., Tao, J. and Yin-Chun, C. 2005. Modification of larch wood by intensive microwave irradiation. J. For. Res. 16(3): 237-240. Hughes, J. F. 1967. Density as an index of wood quality with special reference to the development of rapid and efficient methods of estimating density. In: Proceedings of the Tropical Forestry Meeting by Commonwealth Forestry Institution. Oxford. 1-4. James, W. L. 1975. Dielectric properties of wood and hardboard: variation with temperature, frequency, moisture content and grain orientation. USDA, Forest Service Research Paper, FPL 245. 32 pp. 93  Jankowsky, I. P. and Dos Santos, G. R. 2004. Drying behavior and permeability of Eucalyptus grandis lumber. Proceedings of the 14th International Drying Symposium (IDS 2004), Sao Paulo, Brazil, 22-25 August 2004, Vol B, 1385-1389. Johnson, B. R. 1979. Permeability changes induced in three western conifers by selective bacterial inoculation. Wood and Fiber. 11 (1). 11-21 Jones, P. L. 1997. RF Heating, an Old Technology with a Future. In Microwaves: Theory and Applications in Materials Processing, IV; Clark, D. E.; Sutton, W. H., Lewis, D. A., (Eds.). The American Ceramic Society: Westerville, Ohio, 1997. 261–268. Jones, P. L. and Rowley, A. T. 1996. Dielectric drying. Drying Technol. 14(5): 10631098. Kabir, M. F., Leininger, T. D., Araman, P. A. and Winn, M. F. 2006. Detection of wetwood by ultrasonics. Forest Prod. J. 56(3): 70-74 Keey, R. B., Langrish, T. A. G. and Walker, J. C. F. 2000. Kiln-Drying of Lumber. Berlin Heidelger: Springer-Verlag. 326 pp. Kemp, A. K. 1959. Factors associated with the development of collapse in aspen during kiln drying. Forest Prod. J. 9(3): 124-130. Knudson, R. M. and Mcfarling, S. M. 2009. Extracting greater value from sub-alpine fir: Veneer products. Forest Prod. J. 59(3): 17-23. Knudson, R. M., Chow, S., Lau, K. K., Lougheed, D. E. and Zaturecky, I. 2008. Extracting greater value from sub-alpine fir: some basic properties. Forest Prod. J. 58(1/2): 31-39. Knutson, D. M. 1968. Wetwood in trembling aspen (Populus tremuloides Michx.). Ph.D. thesis. Univ. Minn., St. Paul. 152 pp. Koch, P., Côté Jr., W. A., Schlieter, J. and Day, A. C. 1990. Incidence of compression wood and stem eccentricity in lodgepole pine of North America. USDA Forest Service, Intermountain Research Station, Research Paper INT-420. Ogden, USA. 42 pp. Kollmann, F. F. P. and Côté, W. A. 1968. Principles of Wood Science and Technology, Vol. 1, Solid Wood, Berlin Heidelberg: Springer-Verlag, 592 pp. Kozlik, C. J. 1970. Problems of drying western hemlock heartwood to uniform final moisture content. Proc. 21st Annual Meeting, Western Dry Kiln Clubs, WashingtonIdaho-Montana Seasoning Club, Missoula, Montana. Kozlik, C. J., Krahmer, R. L. and Lin, R. T. 1972. Drying and other related properties of western hemlock sinker heartwood. Wood and Fiber 4(2): 99-111.  94  Lihra, T., Cloutier, A. and Zhang, S. Y. 2000. Longitudinal and transverse permeability of balsam fir wetwood and normal heartwood. Wood and Fiber 32(2): 164-178. Lin, R. T. 1967. Review of the dielectric properties of wood and cellulose. Forest Prod. J. 17(7): 61-66. Lin, R. T., Lancaster, E. K. and Krahmer, R. L. 1973. Longitudinal water permeability of Western hemlock.1. Steady State Permeability. Wood and Fiber 4(4): 278-289. Mackay, J. F. G. and Oliveira, L. C. 1989. Kiln Operator’s handbook for Western Canada. Forintek Canada Corp. Special Publication No. SP-5. 53 pp. Martinovic, D., Horman, I. and Demirdic, I. 2001. Numerical and experimental analysis of wood drying process. Wood Sci. Technol. 35(1/2): 143-156. Mcmillen, J. M. 1958. Stresses in Wood during Drying. Forest Products Laboratory, Forest Service, United States Department of Agriculture. 33 pp. Metaxas, A. C. and Meredith, R. J. 1983. Industrial microwave heating. IEE Power Engineering Series No.4, Peter Peregrinus Ltd., London. 357 pp. Milota, M. R., Tscheritz, J. L., Verrill, S. P. and Mianowski, T. 1995. Wood and Fiber 27(1): 34-40. Panshin, A. J. and C. de Zeeuw. 1980. Textbook of Wood Technology: Structure, Identification, Properties and Uses of the Commercial Woods of the United States and Canada. McGraw-Hill Book Company. 722 pp. Passialis, C. and Tsoumis, G. 1984. Characteristics of discolored and wet wood in fir. IAWA Bull. New Ser. 5(2): 111–120. Philip, L. R., Kretschmann, D. E., Clark, A. and Isebrands, J. G. 2001. Formation and properties of juvenile wood in southern pines: a synopsis. Gen. Tech. Rep. FPL-GTR129. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. 42 pp. Peyskens, E., de Pourcq M., Stevens, M. and Schalck, J. 1984. Dielectric properties of softwood species at microwave frequencies. Wood Sci. Technol. 18(4): 267-280. Resch, H. 2003. High-Frequency Heating Combined with Vacuum Drying of Wood. In: Proceedings of 8th International IUFRO Wood Drying Conference. Brasov, Romania, Aug. 24-29, 2003. 127-132 Resch, H., and Eckland, B. A. 1964. Accelerating the drying of redwood lumber. Calif. Agric. Exp. Stn. Bull 803. 27 pp.  95  Rice, W. R. and D’Onofrio, M. 1996. Longitudinal gas permeability measurements from eastern white pine, red spruce and balsam fir. Wood and Fiber 28(3): 301-308. Rosen, H. N. 1983. Recent advances in the theory of drying lumber. In: Proceedings of the IUFRO Division V Conference, Southern Illinois University, USDA, Madison. 3262. Schiffmann, R. F. 1987. Microwave and dielectric drying. Handbook of Industrial Drying, 327-356. Schroeder, H. A. and Kozlik, C. J. 1972. The characterization of wetwood in western hemlock. Wood Sci. Technol. 6(2): 85-94. Schwarze, F. W. R., Landmesser, H., Zgraggen, B. and Heeb, M. 2006. Permeability changes in heartwood of Picea abies and Abies alba induced by incubation with Physisporinus vitreus. Holzforschung. 60 (4): 450-454 Sehlstedt-Persson, M. 2005. Properties of Solid Wood: Responses to Drying and Heat Treatment. Licentiate Thesis. 52 pp. Siau, J. F. 1971. Flow in Wood, Syracuse, N.Y. Syracuse University Press, 131 pp. Siau, J. F. 1984. Transport Processes in Wood, Berlin Heidelberg: Springer-Verlag, 245 pp. Siau, J. F. 1995. Wood: Influence of moisture on physical properties. Department of Wood Science and Forest Products, Virginia Polytechnic Institute and State University. 227 pp. Simpson, W. T. 1983. Drying Wood: A Review. USDA Forest Service, Forest Products Laboratory. 34 pp. Simpson, W. T. (Editor). 1991. Dry Kiln Operator’s Manual. Forest Products Laboratory, Madison, Wisconsin. 274 pp.  Simpson, W. T. and Liu, J. Y. 1997. An optimization technique to determine red oak surface and internal moisture transfer coefficients during drying. Wood and Fiber 29(4): 312-318. Skaar, C. 1948. The dielectric properties of wood at several radio frequencies. New York State College of Forestry Technical Publication No. 69, Syracuse, New York. 36 pp. Skaar, C. 1988. Wood-water Relations. Springer-Verlag. 279 pp. Smith, W. J. 1970. Wood density survey in western Canada. Canadian Forestry Service, Western Forest Products Laboratory, Information Report VP-X-66, Vancouver, B.C. 29 pp. 96  Stamm, A. J. 1932. Effect of chemical treatment on wood permeability. Industrial and Eng. Chem. 24(1): 51-53. Stamm, A. J. 1964. Wood and Cellulose Science. The Ronald Press Company. 549 pp. Stamm, A. J. 1967. Movement of fluids in wood. Part 1: Flow of fluids in wood. Wood Sci. Technol. 1(2): 122-141. Strumillo, C. and Kudra, T. 1986. Drying: Principles, Applications and Design. Gordon and Breach Science Publishers, New York. 448 pp. Tiemann, H. D. 1906. Effect of moisture upon the strength and stiffness of wood. U.S. Dep. Agric. For. Serv. Bull. 70. 144 pp.  Torgovnikov, G. I. 1993. Dielectric Properties of Wood and Wood-Based Materials. Springer-Verlag. 194 pp. Torgovnikov, G. and Vinden, P. 2000a. New wood-based materials “Torgvin” and “Vintorg”. Proceedings of 5th Pacific Rim Bio-based Composites Symposium, December 10-13, Canberra, Australia. 756-764. Torgovnikov, G. and Vinden, P. 2000b. Microwave modification of wood properties: improvements in wood permeability. International Research Group on Wood Preservation, 1RG31, IRG/WG-40181, May 14-19, Kona, Hawaii, U.S.A. Torgovnikov, G. and Vinden, P. 2009. High-intensity microwave wood modification for increasing permeability. Forest Prod. J. 59(4): 84-92. Vermaas, H. F., Pound, J. and Borgin, K. B. 1974. The loss tangent of wood and its importance to dielectric heating. South Afri. For. J. 89: 5-8. Vinden, P. and Torgovnikov, G. 1998. A method for increasing the permeability of wood. Patent Appl. Intern. Publication NoWO 99/64213. Vinden, P. and Torgovnikov G. 2000. The physical manipulation of wood properties using microwave: Proceedings, International Conference of IUFRO. The future of Eucalypts for Wood Production, March 19-24, Tasmania, Australia. 240-247. Ward, J. C. 1984. Influence of wetwood on pulse-current resistances in lumber before and during kiln drying. Wood and Fiber 16(4): 598-617. Ward J. C. and Kozlik, C. J. 1975. Kiln drying sinker heartwood from young-growth western hemlock. Preliminary evaluation. 26th Ann. Mtg. Western Kiln Clubs. Oregon State Univ., Corvallis, OR. 44-63.  97  Ward J. C. and Pong, W.Y. 1980. Wetwood in trees: A timber resource Problem. USDA For. Serv. Gen. Tech. Rep. PNW 112, Pacific North West For. Range Stn. Portland, OR. 56 pp. Ward, J. C. and Shedd, D. 1981. Characteristics for presorting white fir lumber with wetwood. Proc. 32nd Annual Meeting, Western Dry Kiln Clubs, Oregon State Univ. Corvallis, OR. 32-47. Ward, J. C. and Shedd, D. 1979. California black oak drying problems and the bacterial factor. USDA For. Serv. Res. Pap. FPL-344. 14 pp. Ward, J. C. and Zeikus, J. G. 1980. Bacteriological, chemical and physical properties of wetwood in living trees. In: Bauch, J. (ed.). Natural variations of wood properties. Mitt. Bundesforschungsanstalt f. Forst-und Holzwirtschaft, Nr. 131. Max Wiedehusen, Hamburg. 133-166 Wilcox, W. W. 1968. Some physical and mechanical properties of wetwood in white fir. Forest Prod. J. 18(12): 27–31. Wilcox, W. W. and Schlink, C. G. R. 1970. Absorptivity and pit structure as related to wetwood in white fir. Wood and Fiber 2(4): 373–379. Xu, Z., Leininger, A., Lee, W. C. and Tainter, F. H. 2001b. Physical, mechanical, and drying properties associated with bacterial wetwood in red oaks. Forest Prod. J. 51(3):7984. Zaky, A. A. and Hawley, R. 1970. Dielectric solids. Routledge and Kegan Paul Ltd., London. 106 pp. Zhang, Y., and Cai, L. 2008. Impact of heating speed on permeability of sub-alpine fir. Wood Sci. Technol. 42(3): 241-250. Zhang, Y., and Cai, L. 2009. Mechanism for de-aspirating pits in sub-alpine fir by steam explosion prior to drying. Drying Technol. 27(1): 84-88. Zhang, Y., Cai, L. and Xu, Y. 2006. Aspirated pit in wetwood and micro-morphology of microbial degradation in sub-alpine fir. Front. For China 1(4): 449-452. Zobel, B. J. and van Bujitenen, J. P. 1980. Wood Variation: Its causes and Control. Springer-Verlag. Berlin. 363 pp. Zwick, R. L. (Principal investigator). 1995. The COFI radio frequency/vacuum kiln project. The Council of Forest Industries, Vancouver. 121 pp.  98  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0071223/manifest

Comment

Related Items