Luonnontieteiden ja metsätieteiden tiedekunta Faculty of Science and Forestry
Effects of thermo-mechanical modification on mechanical properties and water resistance of plantation-grown Vietnamese acacia (Acacia mangium)
and rubberwood (Hevea brasiliensis)
Rulong Cao
MASTER’S THESIS WOOD MATERIALS SCIENCE
JOENSUU 2019
Cao, R. L. 2019. Effect of thermo-mechanical modification on mechanical properties and water resistance of plantation-grown Vietnamese acacia (Acacia mangium) and rubberwood (Hevea brasiliensis). University of Eastern Finland. 57 p.
ABSTRACT
Thermo-mechanical modification of wood is a potential method to improve the utilization rate and specific properties of fast-grown plantations. The aim of this study was to investigate the effects of thermo-mechanical modification and combined thermo-mechanical modification with subsequent thermal modification on dry density, modulus of elasticity (MOE), compression strength (CS), dimensional stability, Brinell hardness (HB), and color reflectance of two Vietnamese fast-grown plantations: acacia (Acacia mangium) and rubberwood (Hevea brasiliensis). A total of 40 acacia and rubberwood boards were modified in an industrial pilot kiln where drying, compression, and subsequent thermal modification are executed in one single unit. Four combinations were applied in modification: densification of 14% and 12%
for tangentially-sawn acacia and rubberwood, respectively, either with or without subsequent thermal modification at 210°C for both species. Additionally, 14 acacia and 12 rubberwood reference boards were freshly preserved as reference groups. Properties of differently treated and reference samples were measured as following: dry density was calculated through mass and volume measurement; universal material tester was used to measure MOE, HB and CS;
dimensional changes of samples were analyzed based on water immersion test; color changes were measured with a spectrophotometer and presented in L*a*b* coordinate system. Results showed that the changes in mechanical properties after thermo-mechanical modification are more evident for rubberwood than for acacia. Thermo-mechanical modification led to higher dry density, MOE, HB, and CS of rubberwood compared to reference groups, whereas majority of these properties did not differ between the acacia samples. On the other hand, modification decreased the dimensional stability in thickness direction in both species.
Subsequent thermal modification decreased HB in both species and lowered the dry density of rubberwood. However, it improved the dimensional stability of rubberwood compared to thermo-mechanically modified samples. Regarding wood color, both species were greatly darkened and reddened after every single modification.
Keywords: thermo-mechanical modification, thermal modification, acacia, rubberwood
FOREWORD
The study materials and funding were provided by KWS Timber Tech Oy, Kuopio, and the experiment was carried out at Natural Resources Institute Finland (Luke), Joensuu and the University of Eastern Finland (UEF), Joensuu. The period of experimental work was roughly three months, from June 2018 to August 2018, followed by the data analysis and thesis writing later on until June 2019.
Here firstly I wish to thank for my main supervisors Dr. Veikko Möttönen (Luke) and Mr.
Juhani Marttila (PhD student at UEF), who granted me the opportunity to carry on this project and gave me invaluable supervisory during entire experiment and thesis writing period. They spent much time on giving me the instructions for experiments and commenting on my thesis.
Likewise, I am grateful to Mr. Pekka Ritvanen (KWS Timber Tech Oy) for his technical support for this study. Also, I would like to express my gratitude towards to Dr. Henrik Heräjärvi (Luke) and Dr. Erkki Verkasalo (Luke) for their constructive advice for designing the experiment at the beginning and also patient guidance during the experiment and thesis writing phase.
In addition, I would like to express my thanks to our associate professor Mr. Antti Haapala (UEF), and our programme coordinator Ms. Pauliina Karvinen (UEF), who have been kindly and helpfully supporting my studies during these two years. Furthermore, I wish to convey my thanks to Ramji Pandey, Sergio Martínez Suñer, and Mokbul Hossain, for their great companies along this two-year journey in Joensuu.
Last, hope this could be dedicated to my grandmother, parents, and my extended family who have been always been supporting me, encouraging me, and loving me in my entire life.
Rulong Cao Joensuu, Finland April, 2019
Table of Contents
1INTRODUCTION ... 5
1.1Wood modification ... 5
1.2 Thermo-hydro-mechanical treatments (THM) ... 6
1.3 Wood densification ... 7
1.4 Thermal modification ... 9
1.5 Wood species in this study ... 11
1.6 Studied properties of wood ... 12
1.6.1 Moisture content and density ... 12
1.6.2 Strength and mechanics of wood ... 13
1.6.3 Modulus of elasticity (MOE) ... 14
1.6.4 Compression strength (CS) ... 15
1.6.5 Hardness ... 16
1.6.6 Wood color ... 16
1.6.7 Wood swellings and shrinkages ... 17
1.7 Objective of the study ... 18
2 MATERIALS AND METHODS ... 19
2.1 Wood materials and modification processes ... 19
2.2 Methods of property determination ... 20
2.2.1 Moisture content and density ... 20
2.2.2 Modulus of elasticity (MOE) ... 22
2.2.3 Compression strength (CS) ... 23
2.2.4 Dimensional stability ... 24
2.2.5 Color reflectance ... 25
2.2.6 Brinell hardness (HB)... 27
2.3 Statistical analysis ... 28
3 RESULTS ... 29
3.1 Moisture content and density ... 29
3.2 Modulus of elasticity (MOE) ... 30
3.3 Compression strength (CS) ... 33
3.4 Dimensional stability ... 36
3.5 Color reflectance ... 39
3.6 Brinell hardness (HB) ... 41
4 DISCUSSION ... 43
4.1 Effects of treatments on density ... 43
4.2 Effects of treatments on modulus of elasticity (MOE) ... 45
4.3 Effects of treatments on compression strength (CS) ... 45
4.4 Effects of treatments on dimensional stability ... 46
4.5 Effects of treatments on color reflectance ... 48
4.6 Effects of treatments on Brinell hardness (HB) ... 49
4.7 Limitations of the study and recommendations for future research ... 49
5 CONCLUSIONS ... 50
6 REFERENCE ... 51
1 INTRODUCTION 1.1 Wood modification
As a renewable material, wood has served human being for thousands of years. However, some disadvantages of wood such as low dimensional stability and durability have restricted its applications (Shukla, 2014). Therefore, wood modification is developed and defined as a process to enhance desired properties (e.g. moisture sensitiveness, dimensional stability, hardness, wear resistance, fungi and insect resistance, UV resistance, and aesthetic value) of wood or wooden products by using chemical, mechanical, physical, or biological methods.
When wood modification is completed, no hazardous or toxic residues are supposed to be released to the environment. (Sandberg et al., 2017)
Wood modification is mainly categorized into two types: chemical treatments and thermo- hydro-mechanical treatments (THM), as it shown in the figure 1 (Navi & Sandberg, 2012).
Chemical modification is the most commonly accepted modification method due to the wide range of chemical agents, which could alter the chemical structure of wood especially cell- wall polymer hydroxyl groups and thereby changing its intrinsic properties. While in the THM treatments, changes in properties are achieved only by adjusting heat, water, and mechanical forces. (Sandberg et al., 2017) Compared to chemically treated wood products, wood products treated through THM modification are more environmentally friendly and less structure-altered (Navi & Sandberg, 2012).
Figure 1. A diagram of wood modification main processes (Navi & Sandberg, 2012)
1.2 Thermo-hydro-mechanical treatments (THM)
Different techniques of thermo-hydro (TH) and thermo-hydro-mechanical (THM) treatments are shown in the figure 2 below, based on different methods and parameters used during the modification. Compared to TH technique, mechanical forces are supplemented during THM in addition to moisture and heat (Sandberg & Kutnar, 2016). During THM modification, compressive stresses are applied against the radial or tangential direction of the solid wood (Dogu, 2010). If the temperature and moisture content are well controlled, higher values in density and stronger surface layers would be achieved so that the densified wood would be more applicable in flooring, furniture, and decoration uses compared to unmodified wood (Gao, 2016).
Figure 2. Classification of THM treatments (Sandberg & Kutnar, 2016)
Another technique defined as thermo-mechanical treatment (TM) is a sample compression process in an open system without enough control of moisture content, which is also widely applied in the industry (Kúdela, 2018). However, despite of improvements from mechanical point of view, spring back effect could lead thermo-mechanically modified wood to return back to its initial dimensions when it is exposed to a wet condition or immersed under the water (Dwainto et al., 1997; Khademi, 2011).
1.3 Wood densification
Wood densification is defined as a process where wood materials are compressed in their transverse direction without loss of lignin (Sandberg, 2013). During densification process, the void volume of lumens is largely reduced so that increased density of wood is achieved and thereby improving some mechanical properties (e.g. modulus of elasticity, surface hardness, compression strength, tensile strength, and abrasion resistance) of low-density wood species (Anshari, 2011; Sandberg, 2013).
Wood densification in its transverse direction is typically categorized into surface densification and bulk densification (also known as uniform densification or volumetric densification) (Sandberg & Kutnar, 2016). Bulk densification is commonly applied in wood
shaping and the surface densification was mainly developed to harden the surface layers and improve the resistance of abrasion of low-density solid wood (Sandberg, 2013).
According to Sandberg (2013), it is significant to soften the amorphous region of wood before densification process in order to make wood easier to be deformed and avoid mechanical damages (Figure 3). Softened state of wood is achieved by saturated stream treatment in a high temperature typically at approximately 110 °C.
Figure 3. Densification in transverse direction and different fixation methods for set recovery at cells level (Sandberg, 2013)
Densification process occurs once wood is plasticized (Navi & Sandberg, 2012). During this process, viscoelasticity of wood at glass transition temperature Tg plays an important role because at this temperature, polymeric components of wood are decomposed rapidly, and amorphous wood polymer instantly turns soft which enables the structure of amorphous molecular being rearranged without fracturing under forces. Besides, internal energy as well as properties of wood are also immediately altered at Tg. (Kúdela, 2018) On the other hand, wood shows its glassy characteristics (i.e. stiffness and brittleness) at low temperatures and once the temperature exceeds Tg, wood polymer transits to rubbery state and shows its compliance (Kutnar, 2012; Kúdela, 2018). The modulus of elasticity in the glass state shown in figure 4, for example, is 3000–4000 times higher than that of rubbery state (Navi &
Sandberg, 2012).
Figure 4. Changes in modulus of elasticity with temperature for an amorphous polymer (Navi
& Sandberg, 2012)
After densification, application of subsequent cooling process is normally required. However, the elastic energy stored in microfibrils of compressed wood would be released when atmospheric moist and heat in compressing process terminate, resulting in the thickness along compression springing back to its former shape. (Navi and Sandberg, 2012) This phenomenon is defined as “shape memory” or “set recovery”, which is concerned as the biggest problem of compressed wood (Möttönen et al., 2015). Therefore, it is important to improve the dimensional stability in the direction of compression of densified wood because springing back of thickness along compression implies the failure of densification. In general, the shape memory effect could be manipulated by adjusting the heat, compressive force, and pressing time during the modification process although it cannot be eliminated (Navi, 2000). In addition to pressure, temperature and duration of treatment, the moisture content of wood as well as wood species could also affect the quality of compressed wood (Kúdela, 2018).
Another problem of wood densification is that the surface layers are densified prior to the center during the compression process (Kúdela, 2018). In this case, the density of wood along its compressing direction may not be uniformly distributed due to the moisture and temperature gradient of wood during the process (Wang, 2000). If the internal temperature cannot meet the minimum temperature of plasticization due to some reasons e.g. insufficient preheating time, then a higher density of the surface would be caused. Therefore, an effective plasticization requires higher temperature and moisture content which are able to form more homogenous density of wood along the compression direction. (Kúdela, 2018)
1.4 Thermal modification
Thermal modification, or heat treatment, is considered as the most commercially used wood modification method nowadays (Esteves, 2009). It does not require the use of chemicals so that it could be seen as an environmental-friendly wood modification technology (Kocaefe, 2015). The temperature ranging between 160 °C and 250 °C is typically applied in most of thermal modification (Yan, 2014; Sandberg et al., 2017). During this process, wood is heated for a certain period, from a few minutes to several hours in a low oxygen and dry atmosphere or filled with non-combustible gas such as nitrogen to prevent it from carbonization and oxidation (Millet and Gerhards, 1972; Hillis, 1975; Kocaefe, 2015). Permanent degradation of cellulose and lignin would be caused once the temperature exceeds 250 °C. On the other hand, relatively low heating temperature, less than 140 °C is not able to phenomenally change the material properties. (Hill, 2006)
Some properties of wood after thermal modification are altered due to the chemical changes in the composition of cell wall material during this process (Gong, 2010). Sivonen (2002) indicated that the chemical structure changes of pine components are more significant when the temperature of heat treatment is over 200 °C. After thermal modification, the crystallinity index of cellulose increases and hemicelluloses are hydrolyzed, but lignin is only slightly decomposed (Pelaez-Samaniego et al. 2013). Reduced wood hygroscopicity achieved by these largely decomposed hemicelluloses, and the improved durability allows thermally-modified timber to be applied in conditions where weather and humidity may vary. They are applicable for both outdoor applications such as terraces, cladding, saunas, windows and garden furniture, and interiors such as flooring, decorative panels, staircase, and indoor furniture.
(Sandberg, 2013) On the other hand, reduced mechanical and physical properties caused by cellulose depolymerization and increased crystallinity of thermally-modified wood products limit their application in wooden structures (Sandberg, 2016). Depending on the wood species and end uses, temperature and time settings should be optimized in different industrial heating processes (Sandberg, 2013).
Normally, better hydrophilicity, improved dimensional stability, greater decay resistance, and lower thermal conductivity are achieved, whereas modulus of elasticity and rupture (MOE and MOR), Brinell hardness, and impact strength are decreased after thermal modification. In addition, reduced density, improved brittleness, and darkened color are also typically caused by heat treatment. (Sandberg et al., 2017) Johansson (2006) found that 3-hour treatment at 200–°C greatly reduced the MOE of birch whereas no clear decrease was detected when birch samples were treated below 200–°C. Later study conducted by Yan (2014) also showed that
thermal treatment had no effect on its MOE within a rising temperature until 200–°C and treatment time of 6 hours, but it increased the dimensional stability of Douglas fir. However, Shi (2007) might argue that thermal treatment generally improved the MOE value although it caused decrease in MOR of some commercialized species such as pine, spruce, fir, birch and aspen. Gong (2010) investigated the effect of post-thermal modification on surface-densified aspen and pointed out that after subsequent thermal treatment, dimensional stability of sample surface was enhanced phenomenally compared to non-thermally-treated densified aspen, although the dimensional stability of surface-densified samples with thermal treatment was still lower than that of aspen without being densified. On the other hand, density, hardness and MOE of aspen surface were reduced after post heat treatment. Schneid (2013) found that long-time thermal treatment in high temperatures greatly darkened the color of eucalyptus wood samples.
According to Navi & Sandberg (2012), there are currently mainly five industrialized and commercialized wood heat treatment processes: in the Netherland, PLATO® uses water as a medium; in Germany, Oil-heat-treated wood (OHT) uses rapeseed oil as a medium; in Finland, Thermowood® uses vapor as a medium; in France, Retification uses nitrogen as a medium; and in Canada, Perdure applies drying and subsequent heating of wood in a steam atmosphere.
1.5 Wood species in this study
Acacia mangium has been successfully commercialized in Vietnam since 1980s, because of its adaptability to various conditions, short rotation period, and good quality of stems (Hai, 2015). By 2013, Acacia mangium accounted for approximately 30% (600,000 ha) of the total forest plantation area in Vietnam, which made it the most significant domestic species in Vietnam (Nambiar, 2014). Although short rotation period allows acacia to be applicable for pulp production, when it comes to saw log production, wood density should also be considered as a significant parameter indicating wood quality so that better end products will be produced. The basic density of juvenile Acacia mangium could range between 474 kg/m3 and 539 kg/m3, and that of matured acacia varies between 600-660 kg/m3, depending on the age, location of the plantation area, breeding, and hybrid situation of the species. (Dinh, 2012;
Nugroho, 2012) The higher density of the species, the better quality of the pulp and paper will be produced (Mohammed, 2011). Besides pulpwood production, other main uses of acacia
plantations with management rotation periods of 5–10 years are woodchip export and sawn timber for furniture industry (Nambiar, 2014). The average cellulose and lignin content of Acacia mangium are approximately 38% and 17%, respectively (Mohammed, 2011).
Hevea brasiliensis as a natural rubber tree species is an important industrial crop widely planted in Southeast Asia. Malaysia, Thailand and Indonesia as traditional rubber producing countries totally account for 70% of the natural rubber supply worldwide. (Nguyen, 2013) As a non-traditional rubber producing country, Vietnam contributed 550,000 ha of the rubber growing area in the year of 2007 and Vietnamese government aims to reach 700,000 ha of rubber plantation by 2020 (Fox, 2013). Hevea brasiliensis trees were traditionally used for latex and natural rubber production. However, after 25–30 years, productivity of rubberwood trees started to decrease, and those rubber trees would be felled and used as fuelwood. (Teoh, 2011) The air-dry density of rubberwood is approximately 650 kg/m3 (Balsiger, 2000).
Cellulose, hemicellulose, lignin, and extractives constitute 38%, 29%, 28%, and 4% of Hevea brasiliensis, respectively (Petchpradab, 2009). Nowadays, rubberwood as one of main raw materials for wood-based panel production (e.g. medium density fiberboard, particle board, plywood) is in high demand in southeast Asia (Jesus, 2015; Hashim, 2005). Furthermore, the homogeneous structure (i.e. similarity between heartwood and sapwood) and uniform creamy color distribution have propelled the application of rubberwood in furniture manufacturing sector (Teoh, 2011). On the other hand, the application of rubberwood is limited mainly to interiors because of its low natural durability, weak fungi resistance and poor dimensional stability (Teoh 2011; Shukla, 2018).
1.6 Studied properties of wood 1.6.1 Moisture content and density
“Moisture content in wood is expressed as a percentage by weight of the water within the wood cells, compared with the oven dry weight of the wood cells themselves” (Coulson, 2012). Properties such as weight, mechanical strength, swelling and shrinkage are all relevant to the moisture content of wood. Moisture content of wood varies between different species, different parts of wood (i.e. sapwood and heartwood), or even different boards cut from the same tree. (Wood handbook, 2000) As the moisture content of wood is strongly affected by the atmospheric moist, equilibrium moisture content (EMC) is used to define the state when the moisture content of any wood or wood product is harmonized with the atmospheric
condition (Coulson, 2012). Depending on the end uses, sawn timber or other wood products usually need to be dried until their EMC is reached, in order to prevent them from bio- degradation (Ansell, 2015).
EMC (%) End use
18 Carcassing timber
15–18 Exterior joinery
12–15 Interior joinery
12 Wood block flooring
10 Timber in permanently heated conditions 8–9 Radiator surround & timber over under-floor heating
Table 1. Expected equilibrium moisture content for different end uses (Coulson, 2012)
Density is described as the mass contained in a unit volume at some specified moisture content of wood. For instance, the oven-dried volume and mass of wood are used to determine the “dry density”, while “basic density” of wood is expressed by oven-dried mass of wood divided by the green volume. Another concept, “air-dry density” is calculated under air dry condition, that is, air-dried mass divided by air-dried volume usually at 12% moisture content conditions. (Wood handbook, 2000) Density variations could be found between species and within species due to the differences in structure, content of extractives, chemical compositions, and moisture (Tsoumis, 1991).
According to Tsoumis (1991), Density is of great interest concerning wood production because hygroscopicity, dimensional stability and mechanical, electrical as well as thermal properties are closely related to the density of wood. In addition, it is considered as an important index for pulp and paper production as it implies the fibre yield per unit volume.
1.6.2 Strength and mechanics of wood
Mechanical properties of wood are used as general terms to determine the ability of wood to carry or to resist applied forces or loads. They are of great importance for material selection in wood construction sectors because modulus of elasticity measures the resistance to bending which directly relates to the stiffness of beams, parallel-to-grain compression strength determines the load a short post or column will carry, side hardness relates to the resistance to denting of floorings. (Bowyer et al., 2007) Unlike some other homogeneous and isotropic materials such as steel, wood behaves anisotropic and heterogeneous, which makes it respond
differently to all kinds of stresses in varying directions or degrees (Hoadley, 1995). Therefore, it is critical to better understand the mechanics of wood before application.
According to Haygreen & Bowyer (1989), stress and strain as two general terms are widely used in the study of wood mechanics. Internal stress within wood bodies is caused by external applied forces, which is expressed as unit force (i.e. the amount of force distributed on a unit of area). Strain defines the extent of deformation or distortion of wood as a result of stress, determined by unit deformation (i.e. the change in length per unit of length in the direction of stress).
1.6.3 Modulus of elasticity (MOE)
In wood materials science, modulus of elasticity or Young’s modulus, expressed as MOE, indicates the elasticity of wood, which implies that the deformation of wood occurred under low forces tends to be recovered after the load being removed (Kollmann, 1984). It is defined by the relationship between strain and applied forces only below the proportional or elastic limit of wood. Within this limit, a linearly larger deformation is caused by increasing applied forces. Better elasticity of wood suggests that less deformation will be caused under the same applied forces thereby showing superior stiffness. (Hoadley, 1995) If the load reaches the elastic limit, deformation will not be able to be fully recovered even the forces are withdrawn (Tsoumis, 1991).
Generally, MOE along longitudinal axes of wood is extensively studied whereas there is less data along the other two directions (i.e. radial and tangential axes). In addition, the modulus of elasticity along longitudinal direction of wood (varies between 2.5–17 GPa) is much higher than that of transverse direction (varies between 300–600 MPa). (Tsoumis, 1991) The differences in longitudinal modulus of elasticity could be caused by the variations in grain angle, density, moisture content, temperature, knots, and notches (Kollmann, 1984).
Figure 5. Young’s modulus (GPa) versus density (kg/m3) at 12% moisture content for selected softwood and hardwood species in longitudinal (l) and transverse (t) directions (Ansell, 2015).
1.6.4 Compression strength (CS)
The wood compression parallel to grain appears when wood is being shortened under forces along its longitudinal direction, while compression perpendicular to grain happens in wood when objects are placed on the surface of those wood products (e.g. table, floor) (Hoadley, 1995). Normally, the parallel-to-grain compression strength of wood is ranging between 25- 95 MPa, which is higher than radial or tangential compression strength (1-20 MPa). The structure of wood makes its parallel-to-grain compression strength higher than most of the constructional materials such as stone and brick, but lower than that of metals. Still, this characteristic allows wood to be applied in constructional column applications. (Tsoumis, 1991) The crash of wood in its longitudinal direction under compressive forces could be interpreted as the buckling or folding, cleavage or shearing of cells, rapture of cell walls or local fractures (Kollmann, 1968; Navi and Sandberg, 2012). Before plastic deformation caused by buckling of cell walls, wood fibre along its longitudinal direction undergoes elastic deformation. Temperature is considered as a key factor to determine the moment when buckling takes place and wood crashes, as polymers would typically be weakened in high temperatures. (Kettunen, 2006)
1.6.5 Hardness
Hardness of wood measures wood surface resistance to permanent deformation caused by the penetration or scratching of another object (Barajas, 2017). The impression on wood which is caused by external forces is closely related to the strength of wood (Kaymakci, 2014). In addition, as an important property, it reflects the flexibility and durability of products, components, and materials that are made of wood, e.g. flooring and staircase. (Kumar, 2012;
Rautkari, 2013; Gašparík, 2016). Hardness of wood varies among different species: species such as poplar, willow, basswood, and pine are relatively soft, pine, fir, and walnut have a medium hardness, while oak, ash, beech, maple, and birch are considered as hard species, depending on the density, moisture content and weight of wood itself (Tsoumis, 1991). In addition, hardness results are quite dependent on the testing methods and the shape of tools (Holmberg, 2000).
Currently, there are two main surface hardness measurements of wood: Brinell hardness measurement, which is extensively used in Europe; and Janka hardness measurement, which is widely used in South and North America (Grekin, 2013). The former test uses a specific indenter (e.g. mostly often a steel ball) and applies certain loads to the surface of wood, and then measures the size of indentation. While the latter one requires wood indentation with a specific depth to be caused by a specified tool (e.g. most often a steel ball), and the applied forces are recorded after each indentation. (Holmberg, 2000)
1.6.6 Wood color
Wood color is known as one of the most notable features of wood, and it is quite commonly used for wood identification (Navi & Sandberg, 2012). Wood color varies among species, ranging from almost white to jet black. The color differences also often exist among species or sometimes in one single tree (e.g. sapwood and heartwood color differences, earlywood and latewood color difference). (Tsoumis, 1991) Among fresh woods, dark color always indicates heartwood while light color of wood signals either heartwood or sapwood (Navi &
Sandberg, 2012). In furniture and decorative veneer industry, uniformity of wood color is considered as an extremely favorable feature in the commercialization (Moya, 2010).
Many factors such as environmental conditions and soil properties of growing trees, wood aging, long-time exposure of wood to daylight and air, water or steam, chemical or thermal treatments of wood, fungal and bacterial attacks greatly affect the color of wood (Tsoumis, 1991; Hoadley, 1995; Gierlinger, 2004; Moya, 2010; Navi & Sandberg, 2012). Consequently, the wood color is dependent on the chemical components, especially the amount and type of extractives (Gierlinger, 2004, Moya, 2012).
In order to identify wood color, one standardized chromaticity system, CIELab color system, has been developed as the most widely used colorimetric system, which records the reflectance spectra of wood at different wavelengths for wood color measurements (Gierlinger, 2004). It has been applied in applications such as softwood and hardwood identification;
variations among different wood species; effects of wood drying on color; and effects of thermal treatment on color (Nishino, 2000; Mononen et al., 2002; Johansson, 2006; Sotelo Montes, 2008). The color difference index is used to compare lightness, redness, and yellowness differences among or within wood species (Moya, 2012).
1.6.7 Wood swellings and shrinkages
Swellings and shrinkages describe the increased and decreased dimensions of wood caused by changing moisture content mainly in wood walls (Tsoumis, 1991). In living or freshly felled trees, the moisture content in fresh wood (approximately 150%) is generally much higher than the moisture content at their fiber saturation point (FSP) (approximately 28%). Existence of huge amount of free water in cell lumens avoids the dimensional changes of wood when the atmospheric moisture changes. (Coulson, 2012) Compared to fresh wood, the swelling and shrinking problems are more considerable in non-green wood or dried wood products (Wood handbook, 2000).
According to Shmulsky & Jones (2011), wood swells as water enters the cell wall structure while it shrinks when it loses its bound water below its FSP. When the atmospheric moisture is above the FSP of wood, wood cells start to absorb water and capture these water molecules between the hemicellulose and long-chain cellulose molecules as their bound water, which could expend the distance between chain molecules and thereby causing wood swellings.
Drying is a reverse phase of swelling. The dimensional changes are theoretically reversible, in
some small pieces of stress-free wood. However, it is usually irreversible in wood or wood panel products such as large pieces of solid wood, fiberboard or particle board.
Moisture content changes cause less expansion along wood longitudinal direction than that of transverse direction (Ansell, 2015). The main reason is that S2 layer is the main composition of wood cell wall, where molecule chains are oriented almost parallel to the long axis of the cell. Therefore, transverse dimensional changes occur when these molecule chains move closer or further to each other, whereas longitudinal swellings or shrinkages are negligible.
(Shmulsky & Jones, 2011) In addition to moisture content, factors such as density, anatomical structure of wood, extractives, chemical composition, and mechanical stresses are also affecting the dimensional changes of wood (Tsoumis, 1991).
Swellings and shrinkages need to be controlled during the production because degradation and defects caused by dimensional changes will cause problems such as opening or tightening of joints, warping, casehardening, honeycombing, change of cross-sectional shape, collapse, and loosened or raised grain (Tsoumis, 1991; Hoadley, 1995).
1.7 Objective of the study
Effects of individual thermo-mechanical modification or thermal modification on mechanical properties, dimensional stability, as well as wood color of widely used species e.g. pine, spruce, birch, aspen, beech etc. have been studied by many researchers. However, few studies were aimed to determine the combined effect of thermo-mechanical modification and subsequent thermal modification on these properties concerning the fast-grown species which might also have huge potential and help human beings to combat the scarcity of natural wood resources.
Therefore, the main objective of this study was to investigate the effects of thermo-hydro- mechanical modification (THM) and combined THM with subsequent thermal modification on selected properties of acacia (Acacia mangium) and rubberwood (Hevea brasiliensis), in order to evaluate the potential of applying an industrial modification application of these two species. More specifically, the aims were: (1) To evaluate the effects of THM and THM with subsequent thermal modification on selected mechanical properties (i.e. modulus of elasticity, parallel-to-grain compression strength, and Brinell hardness) of wood; (2) To analyze the effects of THM and THM with subsequent thermal modification on dry density and
dimensional stability in extreme conditions (immersion and 103 °C oven drying) of wood; (3) To observe the color changes of wood after THM and combined THM with subsequent thermal modification.
2 MATERIALS AND METHODS
2.1 Wood materials and modification processes
In this study, timber of two different tree species from Vietnam, acacia (Acacia mangium) and rubberwood (Hevea brasiliensis) were transported to Juankoski, Finland and then underwent thermo-mechanical modification. The pilot modification kiln patented by KWS Timber Tech Oy allows wood drying, mechanical compression, and thermal modification in one single kiln unit where different combinations of process could be achieved by adjusting parameters such as hydraulic compressive pressure, atmospheric moisture content and temperature, and duration of each modification process. The wooden boards were divided into six groups, including four experimental groups of sawn timber treated through different ways and another two reference groups dried at 50°C in the oven but without any compression or thermal modification.
Table 2. Different treatments and number of samples (A=unmodified acacia, AC=thermo- mechanically modified acacia, ACT=thermo-mechanically modified acacia with subsequent thermal modification, K=unmodified rubberwood, KC=thermo-mechanically modified rubberwood, KCT=thermal-mechanically modified rubberwood with subsequent thermal modification).
Category Treatment Number Compression
degree
Time and temperature
in thermal treatment
A Reference, 50°C oven-drying 14 – –
AC Compression 10 14% –
ACT Compression & thermal treatment 9 14% 210 °C, 3h
K Reference, 50°C oven-drying 12 – –
KC Compressed rubberwood 11 12% –
KCT Compression & thermal treatment 10 12% 210 °C, 3h
Before modification, the boards in experimental groups were placed between perforated aluminum plates where the moisture content of boards could be indirectly controlled by the air circulation rate through these plates in the kiln. Drying and hydraulic compression
proceeded simultaneously. The drying temperature was steadily increased up to 130°C for acacia samples and 120°C for rubberwood samples, and the nominal degrees of mechanical compression were set to 14% and 12% for acacia and rubberwood, respectively. After drying and mechanical compression, a subsequent 3-hour thermal modification at 210°C was applied to two groups of board samples. During this process, certain amount of steam was applied in the kiln to protect wooden boards from burning caused by high temperature. After the subsequent thermal modification process, the system was cooled down, and the pressure was released. These selected modified wooden boards were then carried to the Natural Resources Institute Finland, Joensuu, Finland and stored there for this study.
Figure 6. Acacia boards were placed between perforated aluminum plates before the compression process.
2.2 Methods of property determination 2.2.1 Moisture content and density
Samples with a size of 50 mm (thickness) × 50 mm (length) × 80 mm (width) were cut from the edge of modified boards (i.e. AC, ACT, KC, KCT) prior to the tests in order to determine the initial moisture content and density of these samples. Fresh acacia (A) and rubberwood (K) (50mm (thickness) × 60mm (length) × 80 mm (width)) were also cut through the same way and then they were put into plastic bags and stored in the refrigerator immediately after the cutting, in order to prevent the them from drying before measurements.
The initial weight (recorded as M1) of all samples was measured by the scale, and the initial volume (recorded as V1) of fresh acacia (A) and rubberwood (K) was measured by using gravimetric method which could be further used for basic density calculations. After the initial weight and volume measurements, all the samples were oven-dried at the temperature of 103 °C and their mass (recorded as M2) and volume (recorded as V2) were recorded immediately after they were taken out from the oven, in order to calculate the moisture content and dry density of all the samples.
Their moisture content was defined as followed:
MC =M1−M2
M2 × 100% (1) Where
MC = moisture content of samples (%) M1 = initial mass of samples (g)
M2 = oven-dried mass of samples (g)
The basic density of A and K was calculated as below:
ρb =M2
V1 × 1,000 (2) Where
ρb = basic density of A and K (kg/m3) M2 = oven-dried mass of samples (g) V1 = green volume of samples (cm3)
The dry density of all samples was calculated as below:
ρb′ =M2
V2 × 1,000 (3) ρb’= dry density (kg/m3)
M2= oven-dried mass of samples (g) V2= oven-dried volume of samples (cm3)
Figure 7. Volume measurement of fresh rubberwood specimen by using the gravimetric method.
2.2.2 Modulus of elasticity (MOE)
According the standard ISO 13061-4, static three-point bending test was used to measure the modulus of elasticity (MOE) of all the sample boards from different modification processes.
Matertest model FMT-MEC 100kN was the equipment for testing the MOE value. Before the bending test, the width and thickness of all boards were measured by the caliper and recorded into the computer system for further calculations. Forces were always applied to the direction of thickness with a loading speed of 0.02 mm/s, regardless the sawing pattern and two spans of 750 mm (modified samples) and 840 mm (non-modified samples) were used in the test in accordance with the standard (thickness: span = 1:14). The bending was stopped before each board reaching its proportional limit to prevent it from irreversible bending. The deflection and applied forces in the mid-span of each board were recorded in the computer system, and then the MOE was calculated automatically according to these given defection and force values.
Prior to the test, fresh samples of acacia (A) and rubberwood boards (K) were dried in an oven at a temperature of 50 °C roughly 2 weeks, and thermo-mechanically modified acacia and rubberwood (AC and ACT), thermo-mechanically modified acacia and rubberwood with subsequent thermal modification (ACT and KCT) were stored indoors for two weeks.
After the bending test, small pieces were sliced from the edge of boards in order to determine the moisture content of samples. The weight of small pieces was measured twice: once right
after the bending test (marked as M1) and once after being oven-dried at the temperature of 103°C (marked as M2). The moisture content was defined as the same way as it shown previously in the formula (1).
The modulus of elasticity (MOEw) of A and K was adjusted to the level at 12% moisture content by using the following formula, which is valid for moisture contents of (12% ± 5%):
MOE12 = MOEw
1−α(W−12) (4) Where
α is the correction factor for the moisture content, equal to 0.02;
W is the moisture content of the wood, determined according to ISO 13061-1.
Figure 8. Setup of three-point bending test for MOE measurement
2.2.3 Compression strength (CS)
Prior to the compression test, samples with length of 60 mm were cut from all 6 groups of boards by circular and then further cut to smaller pieces by band saw, with a cross-section of 20 mm × 20 mm and a height of 60 mm from the center and surface. Then they were stored in the conditioning chamber at 20 °C, 65% relative humidity for roughly one week until their weight became to be stable.
Universal physical testing machine Zwick was used to perform the experiment and the compression strength along parallel direction of the grain was measured. According to standard ISO 13061-12:2017 (E), the ultimate stresses σ (MPa) in compression parallel to
grain is determined by a gradually increasing load in direction parallel to grain of a test piece until failure, presented as:
σ =Fmax
a∗b (5) Where
Fmax is the maximum load force, N;
a and b are the cross-sectional dimensions of the test piece, mm.
The compression strength parallel to grain of A and K was adjusted to its level at 12%
moisture content using the following formula (6), which is valid for moisture contents of (12%± 5%):
σ12= σ ∗ (1 + α(W − 12)) (6) Where
α is the correction factor for the moisture content, equal to 0.02;
W is the moisture content of the wood, determined according to ISO 13061-1.
Figure 9. Setup of parallel-to-grain compression strength measurement
2.2.4 Dimensional stability
Defect–free specimens with length of 50mm were cut from each board and marked as it shown in the figure 10 below. After that all the specimens were stabilized in the conditioning chamber at 20 °C, 65% relative humidity until their equilibrium moisture content (EMC) was reached. Roughly after one week, samples were taken out from the conditioning chamber and
the initial length, width and thickness of specimens were measured by caliper. After that, these specimens were soaked in buckets filled with water and stored in room temperature for 14 days. The dimensions were measured once again after 14 days. Immediately after that, all the specimens were put into an oven at the temperature of 103 °C till they were fully dried.
The last step was to measure the oven-dry dimensions of them in order to calculate dimensional shrinkage.
The dimensional changes of swelling and shrinkage were determined as formulas shown below:
α =L1−L0
L0 × 100% (8) α′= L2−L1
L1 × 100% (9) Where
α= swelling rate in length, width, and thickness (%) α’= shrinkage rate in length, width, and thickness (%) L0= initial dimensions (length, width, and thickness) (mm)
L1= dimensions (length, width, and thickness) after 14 days’ immersion (mm) L2= dimensions (length, width, and thickness) after oven drying (mm)
Figure 10. Markings of each specimen in extreme condition test. Thickness was measured from 8 points (a, b, c, d, e, f, g, h), length measured from 3 points (a-g, b-f, c-e), and width measure from 1 point (h-d).
2.2.5 Color reflectance
A length of 300 mm was cut from each board after MOE measurements, and surface with 0.5 mm depth was removed from boards by using planner. Then 5 different steps were machined with spindle molder by exposing their surfaces at 0, 3, 6, 9, and 12 mm depths along the thickness profile correspondingly. Konica Minolta CM-2600d portable spectrometer was used to carry out color reflectance measurements of all samples. Each board was measured five times at different steps by portable spectrometer, and the visible wavelength ranged between 360 mm and 740 mm was collected as original data set to be further converted to L*a*b*
color coordinates. The total color difference ΔEab* represents the color difference between two measurements, as it shown in the formula (11):
Δ𝐸𝑎𝑏∗ = √(ΔL∗)2+ (Δa∗)2+ (Δb∗)2 (11) where
ΔEab*: the color difference between two measurements (i.e. modified samples and unmodified samples),
ΔL*: difference between two measurements in lightness; 0 corresponds to black and 100 corresponds to white,
Δa*: difference between two measurements in redness; negative values stand for greenness and positive value stand for redness,
Δb*: difference between two measurements in yellowness: negative and positive values stand for blueness and yellowness, respectively.
Figure 11. Setup of color reflectance measurement
2.2.6 Brinell hardness (HB)
After measuring the color reflectance of these 6 groups of samples, they were immediately stabilized in the conditioning chamber at 20 °C, 65% relative humidity until they reached equilibrium moisture content (roughly one week). After that samples were taken out from the chamber piece by piece to the testing machine Lloyd L6000R in order to carry out Brinell hardness measurements. There were three measuring points at different steps with depths of 0 mm (surface), 6 mm, and 12 mm from the surface, and each point was measured once during the experiment.
According to standard EN 1534, a nominal value of 1000 N was reached in 12–18 s and maintained for 20–30 s on the surface of the samples before the force was withdrawn.
Dimeters of the residual indentation along the grain (d1) and across the grain (d2) were measured by caliper after indentation recovering for approximately 3 minutes.
Brinell hardness was calculated according to the formula (10) in the same standard:
HB = 2F
g×π×D×[D−√D2−d2] (10) where
F is the maximum load applied force, F=1,000N;
HB is the Brinell hardness, kg/mm2; g is the acceleration of gravity, g=9.8 m/s2; π is the “pi” factor, π=3.14;
D is the diameters of the indenter, D=10 mm;
d is the average value of the diameter of the residual indentation (𝑑 =d1+d2
2 ), mm.
Figure 12. Setup of Brinell hardness measurement
2.3 Statistical analysis
IBM SPSS Statistics 25 software was used for statistical analysis. Analysis of variance (one- way ANOVA) tests were performed to analyze the effects of THM and combined THM with subsequent thermal modification on dry density, modulus of elasticity, and parallel-to-grain compression strength of acacia and rubberwood samples at 0.05 significance level. If significance is detected, further Tukey HSD test would be used to evaluate the statistically pairwise differences between treated groups and their control. Significance level was at 0.05 level.
Regarding the test for Brinell hardness and color reflectance, firstly, one-way ANOVA was used to determine the significant HB or color reflectance differences between different steps of acacia and rubberwood. If significance exists, one-way ANOVA and further pairwise Tukey comparisons would be carried out to evaluate the statistical HB or color reflectance value differences at different steps between treated groups and their control at 0.05 level. If there is no sign of significance, the average HB or color reflectance value from different steps would be calculated and the same procedure as dry density would be followed to determine the statistically significant differences between treated groups and their controls.
In the case of dimensional stability, one-way ANOVA was firstly performed to observe whether there were statistically significant dimensional changes in different directions of the
same sample affected by immersion or 103 °C oven-drying. If significance is found, the test would be performed again to determine the effects of THM and combined THM with subsequent thermal modification on dimensional changes of acacia and rubberwood samples at 0.05 significance level. If significance is again detected, pairwise Tukey comparisons would be performed to determine the statistical differences between treated groups and their control at 0.05 level.
3 RESULTS
3.1 Moisture content and density
As it shown in the table 3, after the thermo-mechanical modification, the moisture content decreased significantly compared to that of 50-degree oven-dried samples in both acacia and rubberwood. Moreover, the moisture content in thermo-mechanically modified samples with subsequent thermal modification was even lower than that of thermo-mechanically modified samples.
Dry density of fresh and modified samples is also presented in table 3. According to statistical analysis in table 4, 5, and 6, there was significant elevation of rubberwood dry density from K to KC, suggesting that the thermo-mechanical modification may have a positive effect on dry density of rubberwood. On the other hand, subsequent thermal modification decreased the dry density of KC significantly, as it indicated in the table 6. In acacia samples, although dry density changes were detected between differently treated samples, no statistical significance was found between any two groups, which meant that there was no increase or decrease in dry density after the thermo-mechanical modification or thermo-mechanical modification with subsequent thermal modification.
Table 3. Moisture content, basic density, and dry density of differently treated acacia and rubberwood (A= 50-degree oven-dried acacia, AC=thermo-mechanically modified acacia, ACT=thermo-mechanically modified acacia with subsequent thermal modification; K= 50- degree oven-dried rubberwood, KC=thermo-mechanically modified rubberwood, KCT=thermo-mechanically modified rubberwood with subsequent thermal modification). The figures in the table represent the mean value of moisture content in percentage and basic density ± standard deviation of corresponding specimens
Moisture content
%
Basic density kg/m3
Dry density kg/m3 A 22.56± 11.57 530.33± 63.55 570.70± 65.14
AC 7.12± 2.23 - 590.19± 78.69
ACT 3.14± 0.37 - 552.81± 70.47
K 7.87± 1.28 569.56± 35.35 620.03± 40.09
KC 6.01± 0.31 - 665.04± 39.65
KCT 4.20± 0.16 - 615.37± 38.10
Table 4. Dry density (kg/m3) of acacia between modification processes, and significance of difference according to analysis of variance.
ANOVA Sum of Squares df Mean Square F Sig.
Between Groups 6645.456 2 3322.728 .662 .523
Within Groups 150621.910 30 5020.730
Total 157267.366 32
Table 5. Dry density (kg/m3) of rubberwood between modification processes, and significance of difference according to analysis of variance.
ANOVA Sum of Squares df Mean Square F Sig.
Between Groups 16401.024 2 8200.512 5.294 .011*
Within Groups 46471.357 30 1549.045
Total 62872.381 32
Table 6. Multiple comparison of dry density (kg/m3) of rubberwood between modification processes which are significantly different, according to Tukey HSD test.
(I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Lower Bound Upper Bound
K KC -45.00462* 16.42892 .027* -85.5064 -4.5029
KCT 4.65583 16.85205 .959 -36.8890 46.2007
KC KCT 49.66045* 17.19671 .019* 7.2659 92.0550
*. The mean difference is significant at the 0.05 level.
3.2 Modulus of elasticity (MOE)
The moisture contents of 50-degree oven-dried acacia boards (A) and rubberwood boards (K) were defined as (22.56± 11.57%) and (7.87± 1.28%) respectively, which would be used for MOE calibration before the experiment. The moisture contents of thermo-mechanically
modified acacia (AC), thermo-mechanically modified acacia with subsequent thermal modification (ACT), thermo-mechanically modified rubberwood (KC), and thermal- mechanically modified rubberwood with subsequent thermal modification (KCT) were (7.12±
2.23%), (3.14± 0.37%), (6.01± 0.31%), (4.20± 0.16%) correspondingly. The MOEs of these modified samples were not adjusted.
The figure 13 below shows that the MOE value of A (10605± 2289 MPa) was slightly higher than that of ACT (10557± 1047 MPa). The MOE value of AC (9761± 1226 MPa) stayed at the lowest among these three groups. However, no significance of difference was found between any two groups according to analysis of variance shown in table 7, which indicated that neither thermo-mechanical modification nor thermo-mechanical modification with subsequent thermal modification affected the MOE value of acacia samples. In rubberwood samples, KC has the highest MOE value at 11342± 1385 MPa, closely followed by KCT at 11050± 790 MPa. The lowest value was detected in K, with MOE of 8130± 750 MPa. In addition, table 8 and 9 indicated that the differences between K and KC, and K and KCT were statistically significant (P<0.05), which revealed that there was MOE elevation of rubberwood through both thermo-mechanical modification and thermo-mechanical modification with subsequent thermal modification. However, it could not be concluded that the subsequent thermal modification degraded the MOE value of KC, according to statistical analysis in table 9.
Figure 13. MOE (MPa) of acacia and rubberwood with different modification processes (A=50-degree oven-dried acacia, AC=thermo-mechanically modified acacia, ACT= thermo- mechanically modified acacia with subsequent thermal modification, K=50-degree oven-dried rubberwood, KC= thermo-mechanically modified rubberwood, KCT=thermo-mechanically modified rubberwood with subsequent thermal modification). Each bar represents the average value of MOE (MPa) ± its standard deviation.
Table 7. MOE (MPa) of acacia between modification processes, and significance of difference according to analysis of variance.
ANOVA Sum of Squares df Mean Square F Sig.
Between Groups 4760292.229 2 2380146.115 .789 .463
Within Groups 90450721.422 30 3015024.047
Total 95211013.652 32
Table 8. MOE (MPa) of rubberwood between modification processes, and significance of difference according to analysis of variance.
ANOVA Sum of Squares df Mean Square F Sig.
Between Groups 67777827.449 2 33888913.725 31.821 .000*
Within Groups 29820023.409 28 1065000.836
Total 97597850.859 30
Table 9. Multiple comparison of MOE (MPa) of rubberwood between modification processes which are significantly different, according to Tukey HSD test.
(I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Lower Bound Upper Bound K KC -3212.01636* 440.04149 .000* -4300.8331 -2123.1996
KCT -2920.37283* 463.84446 .000* -4068.0865 -1772.6592
KC KCT 291.64354 463.84446 .806 -856.0701 1439.3572
*. The mean difference is significant at the 0.05 level.
0 2000 4000 6000 8000 10000 12000 14000
A AC ACT K KC KCT
MOE (MPa)
Treatments for acacia and rubberwood
3.3 Compression strength (CS)
After being conditioned in the chamber for one week, moisture contents of acacia samples (A) and rubberwood samples (K) were stabilized at (16.51± 6.90%) and (10.53± 0.74%) respectively, which were used as original data in the formula for compression strength calibration.
Figure 14. Compression strength (MPa) of acacia with different modification processes in different parts (A=conditioned acacia, AC-C=core part from thermo-mechanically modified acacia, AC-S= surface part from thermo-mechanically modified acacia, ACT-C=core part from thermo-mechanically modified acacia with subsequent thermal modification, ACT- S=surface part from thermo-mechanically modified acacia with subsequent thermal modification). Each bar represents the average value of CS (MPa) ± its standard deviation.
As it shown in the figure 14, the compression strength of AC and ACT was higher than A (49.59± 6.54 MPa), both from the surface (AC-S 61.17± 5.99 MPa, ACT-S 60.98± 3.08 MPa) and core (AC-C 56.52± 8.76 MPa, ACT-C 58.81± 4.19 MPa). However, statistical analysis in table 10 and 11 showed that only compression strength of AC-S and ACT-S was significantly higher than A whereas there was no significant difference between AC-C and A, or ACT-C and A, which meant that both thermo-mechanical modification and thermo-mechanical modification with subsequent thermal modification elevated the compression strength from the surface of acacia but neither of them improved compression strength from the core.
Besides, there was no difference between AC-C and AC-S or ACT-C and ACT-S.
0 10 20 30 40 50 60 70 80
A AC-C AC-S ACT-C ACT-S
Compression strength (MPa)
Treatments for acacia
Table 10. Compression strength (MPa) of acacia between modification processes, and significance of difference according to analysis of variance.
ANOVA Sum of Squares df Mean Square F Sig.
Between Groups 888.703 4 222.176 4.576 .004*
Within Groups 1796.458 37 48.553
Total 2685.161 41
Table 11. Multiple comparison of compression strength (MPa) of acacia between modification processes which are significantly different, according to Tukey HSD test.
(I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Lower Bound Upper Bound
A
AC-C -6.92906 3.04454 .176 -15.6573 1.7991
AC-S -11.58339* 3.04454 .004* -20.3116 -2.8552
ACT-C -9.21978 3.75826 .124 -19.9941 1.5545
ACT-S -11.39089* 3.53639 .021* -21.5291 -1.2526
AC-C
AC-S -4.65433 3.11618 .573 -13.5879 4.2793
ACT-C -2.29072 3.81653 .974 -13.2321 8.6506
ACT-S -4.46182 3.59826 .728 -14.7774 5.8538
AC-S ACT-C 2.36361 3.81653 .971 -8.5778 13.3050
ACT-S .19250 3.59826 1.000 -10.1231 10.5081
ACT-C ACT-S -2.17110 4.21933 .985 -14.2672 9.9250
*. The mean difference is significant at the 0.05 level.
Similar results were found in rubberwood as it shown in figure 15. The compression strength of KC-C, KC-S, KCT-C, and KCT-S was (59.44± 6.25 MPa), (59.76± 7.92 MPa), (60.84±
5.51 MPa), and (63.46± 3.88 MPa) respectively, of which were all significantly higher than that of K (49.72± 3.85 MPa), according to table 12 and 13. It suggested that both thermo- mechanical modification and thermo-mechanical modification with subsequent thermal modification penetrated through the whole thickness profile and elevated the compression strength of rubberwood. On the other hand, no difference was detected between KC-C and KC-S, ACT-C and ACT-S, KC-C and KCT-C, or KC-S and KCT-S, which meant that there was no compression strength difference between the core and surface of thermo-mechanically modified rubberwood and thermo-mechanically modified rubberwood with subsequent thermal modification.
