Elastic and Strength Properties of Heat-Treated Beech and Birch Wood

: This paper deals with the impact of heat treatment on the elastic and strength properties of two diffuse porous hardwoods, namely Fagus sylvatica and Betula pendula . Two degrees of the heat treatment were used at temperatures of 165 ◦ C and 210 ◦ C. The dynamic and static elasticity modulus, bending strength, impact toughness, hardness, and density were tested. It is already known that an increase in treatment temperature decreases the mechanical properties and, on the other hand, leads to a better shape and dimensional stability. Higher temperatures of the heat treatment correlated with lower elastic and strength properties. In the case of higher temperature treatments, the decline of tested properties was noticeable as a result of serious changes in the chemical composition of wood. It was conﬁrmed that at higher temperature stages of treatment, there was a more pronounced decrease in beech properties compared to those of the birch, which was the most evident in their bending strength and hardness. Our research conﬁrmed that there is no reason to consider birch wood to be of a lesser quality, although it is regarded by foresters as an inferior tree species. After the heat treatment, the wood properties are almost the same as in the case of beech wood.


Introduction
Beech (Fagus sylvatica L.) ranks among the most important European hardwoods and the most important deciduous species for Czech forestry, occupying 8.3% of total forest area [1]. It plays an important role in industry. In contrast, birch (Betula pendula Roth) is regarded as an inferior species in this region and its wood is mostly used as fuel. One of the reasons for this is its low durability and low resistance against biological agents. One of the ways to improve wood properties is thermal treatment, a natural and an environmentally friendly method of wood modification.
In many kinds of processing, wood is exposed to a treatment at elevated temperatures, e.g., drying, pulping, size stabilization, and production of particle-and fiberboard. Due to the fact that temperature influences the physical, structural, and chemical properties of wood, a number of publications are devoted to this topic , etc. The above mentioned processes are carried out at temperatures that usually do not exceed 200 • C because thermal degradation is undesirable.
Wood heating will lead to different processes that always depend on the heating mode used. It is recognized that hemicelluloses are degraded to a greater extent than other macromolecular components, but the relative stability of cellulose and lignin is much more difficult to determine. As is not the case above, when the wood is heated, heat-labile wood polymeric components (hemicelluloses) begin to decompose, resulting in the production of methanol, acid, and various volatile heterocyclic compounds

Materials
The testing material comes from tree stems from the Školní Lesní Podnik (Forest Establishment) of the Czech University of Life Sciences in Kostelec nadČernými Lesy, Czech Republic. For each species, we used wood from the basal part of three trees with diameters of about 40 cm. European beech (Fagus sylvatica L.) and European birch (Betula pendula Roth.) wood were cut into prisms with dimensions of 25 mm × 50 mm × 1000 mm (R × T × L). Six test pieces with dimensions of 20 mm × 20 mm × 300 mm were prepared from each prism to ensure the longitudinal parallelism of the testing samples with the samples prepared for two degrees of the thermal treatment.
Transversal parallelism should make the mutual comparison of two sets of tests possible (always a sample designed for the determination of density, toughness, and hardness, beneath a sample on dynamic elasticity modulus, static elasticity modulus, and bending strength). For more details, see the sampling scheme in Figure 1. Cutting diagram for testing samples preparation. Green coloring = reference, with no treatment; yellow coloring = heat treatment at 165 • C; red coloring = heat treatment at 210 • C; upper set of samples = for the determination toughness; slanted hatching = for determination density, hardness, and moisture content; bottom set of samples = for the determination dynamic elasticity modulus, static elasticity modulus, and bending strength.
In total, 360 testing samples were used (180 for beech and 180 for birch). The set of samples for each species was divided into thirds (reference, first degree of the treatment, second degree of the treatment). The following defects and irregularities were not allowed for the testing samples: knots, cracks, or reaction wood, as well as an angle of fiber declination in the bending plane larger than 5 • .
The testing samples were conditioned to reach the equilibrium moisture content (approx. 12%). We used the Climacell 707 conditioning chamber (BMT Medical Technology Ltd., Brno, Czech Republic) at 20 ± 2 • C and a relative humidity of 65 ± 5%.
One third of the testing samples were subsequently exposed to the first degree of thermal treatment (an air atmosphere at 165 • C), and the second third of the testing samples were heat-treated at 210 • C, following the Finnish technology for the wood heat treatment (Pat. EP-0759137 [25]). The lab high-temperature chamber A type KHT (Katres Ltd., Jihlava, Czech Republic) ( Figure 2, Table 1) was employed to modify both sets of the testing samples.   Figure 3 describes the process of the heat treatment. During the treatment, we used sprinkling, in contrast to the steam used in Finnish technology. We exposed the testing samples to a temperature of 20 ± 2 • C and a relative humidity of 65 ± 5% to stabilize the equilibrium moisture content ( Figure 2). For the purpose of this study, beech and birch woods were chosen deliberately. Both species represent diffuse porous hardwoods with similar densities ( Table 2). 1 Moisture content 12-15% [38]. LR = radial plane, LT = tangential plane.

Methods
The impact toughness (breaking power) is defined as the ability of wood to absorb the power of impact bending. The aim of this test was to determine the power consumed for the wood rupture (breaking point) under controlled conditions. Charpy's hammer (CULS, Prague, Czech Republic) was used for this determination. The hammer impact direction was tangential.
The Equation (1) was used to calculate the impact toughness: where A w is the impact toughness at the moisture content during the test time in J·cm −2 , W is the power consumed for the wood rupture in J, and b and h are the wood transversal dimensions in cm. The wood bending strength is the stress corresponding to the test sample rupture caused by the combined forces with momentum at the plane perpendicular to the cross section. For the action of a single force in the center of the supports, the bending strength was calculated according to the Equation (2): where σ pohw is the bending strength at the moisture content during the test time in MPa; F max is the force corresponding to the breaking strength in N; l 0 is the distance between supports in mm; and b and h are the width and height dimensions, respectively, in mm. The static bending tests were carried out on a Tira 50 kN testing machine (Tira GmbH, Schalkau, Germany) ( Figure 4) with support distances of 240 mm, i.e., 12-fold greater than the sample height. A theoretical basis for the determination of the bend elasticity modulus is the differential equation of the bending curve, as Equation (3) [39]: where M is the bending momentum, E is the elasticity modulus, and I is the inertia moment.
For the action of a single force in the center of the supports, the static elasticity modulus was calculated according to the Equation (4): where E ohw is the elasticity modulus at the moisture content during the test time in MPa; ∆F is the difference between the forces at maximum and minimum load limits in N; l 0 is the distance between the supports in mm; b and h are the width and height dimensions, respectively, in mm; and ∆y is the test sample deflection in the area of pure bending, equal to the difference between the bending values corresponding to maximum and minimum load limits, in mm. The dynamic elasticity modulus was calculated Equation (5) [39]: where E d is the dynamic elasticity modulus in MPa, c is the speed of sound in m·s −1 , and ρ is the wood density in kg·m −3 . We used a Fakopp Ultrasonic Timer instrument (Fakopp Enterprise Bt., Ágfalva, Hungary) ( Figure 4). Brinell hardness (BH) was calculated using a hardness tester DuraVision-30 (Struers GmbH, Willich, Germany) according to the Equation (6): where H Bw is the BH of wood (MPa), F is the maximum load force (N), D is the diameter of the carbide ball (mm), and d is the diameter of the residual indentation (mm). The force of 500 N was applied. A reading for the wood density determination was taken from each test sample after the experiment (see Figure 1). The density was calculated as Equation (7): where ρ w is the wood density at the moisture content during the testing time in g·cm −3 , m w is the wood mass at the moisture content during the testing time in g, and V w is the wood volume at the moisture content during the testing time in cm 3 .
After the tests on density were carried out, the average width of the annual rings was measured on the samples. The cross sections on samples were scanned and evaluated using image analysis software NIS Elements AR (Laboratory Imaging, Prague, Czech Republic). The average width of annual rings was measured for each individual sample in pixels, which was then recalculated to dimensions in millimeters.
After the samples were dried to zero percent moisture in a Binder FD 115 lab kiln (Binder Inc., Tuttlingen, Germany) at 103 ± 2 • C, the wood moisture content was calculated according to the following formula: where w a is the sample's moisture content in %, m w is the sample's mass at a certain moisture content in g, and m 0 is the sample's dry mass in g. A so-called specific strength [39,40] was used as another indicator of the effect of the treatment on the quality of the testing material. The specific strength represents the proportion of adequate strength and density (SI unit for specific strength is N·m/kg). This indicator is a better way of informing about the impact of the modification on the practical usability of the material.
The initial equilibrium moisture content for the testing samples of the untreated wood was 12% (standardized conditions with a relative humidity of 65 ± 5% and a temperature of 20 ± 2 • C [19,22]). The heat-treated wood exhibited a lower moisture content under these conditions depending on the degree of the heat treatment. All tests were carried out completely with the testing standards according to the Czech national standardization [41][42][43][44][45][46][47], and the determination of the dynamic elasticity modulus was based on the methodology specified in the Fakopp instrumentation manual [48].
For statistical analysis, analysis of variance ANOVA (two-factors) was used to evaluate the significance of individual factors. The Duncan's Multiple Range Test was used to compare the properties among the different treatments and species. A linear regression model was used to set the degree of correlation of selected factors. For all analyses, the same significance level of α = 0.01 (alternatively α = 0.05) was used. Tables 3 and 4 show the basic statistical characteristics of all tested properties of untreated and heat-treated beech and birch wood.

Results
The influence of wood species on a specific property (quantity), at a particular level of treatment (REF = reference, with no treatment, 165 = heat treatment at 165 • C, 210 = heat treatment at 210 • C), is almost always statistically significant (p < 0.01) (see Tables A1-A9). This is not only the logical reason for moisture content in untreated wood (hereinafter REF) and also in treated at a temperature of 165 • C (hereinafter 165). For wood treated at a temperature of 210 • C (hereinafter 210), this difference is significant. Furthermore, there is a statistically insignificant influence of wood on the bending strength "210" and the impact bending strength "REF" and "210".    Clearly, the resilience of both woods was proved to be particularly evident against dynamic strain (impact bending strength), even at a lower level of heat treatment (decrease towards "REF" by 28% or beech and 56% for birch) and at a higher level by 81% for beech and by 86% for birch (see Table 5). On the other hand, the statistical method of this type of load (MOR = bending strength) at a lower level of heat treatment showed an increase of 6% for beech and up to 26% for birch at a lower heat treatment, and there was only a significant decrease at a higher degree of heat treatment, namely 59% for beech and 47% for birch. Regarding the elastic properties (MOE) and hardness, there was only a slight decrease for the beech and an increase for the birch (see Table 5 or Figure 5c,d). Respectively, the above mentioned practically means that the deformation potential in the plastic zone is significantly limited, particularly at a higher degree of heat treatment (above 200 • C).  Expected correlation (as stated Dinwoodie [49]) between statistical and dynamical MOEs has proved to be not very significant for possible predictions, especially for heat-treated wood (see Figure 6e,f). The explanation has already been described in detail in research by Borůvka [37], i.e., the different influence of moisture content during the measurement of dynamic and static moduli, as well as the existence of shear stress during the static three-point bending test. Interestingly, there is a considerable difference in the width of the annual rings between the two trees, with the birch having more than three times the width of the beech, but the density of both trees was more or less standard, within the limits indicated in the literature (Table 2). Specific strength is a part of the results only for a simple material comparison, i.e., the quasi-removal of the effects caused by the density. This characteristic is better for comparison on a different basis, such as wood with metals. In our case, the development trend of the compared properties with the temperature increase proved to be logically similar, without taking into account the density.

Discussion
The aim of this study was to compare beech and birch wood and to better explain the effect of thermal treatment on their physical and mechanical properties. The related objective of this paper was primarily to verify the negativity of the higher level of treatment of deciduous woods ("210") against conifers (see [37]). This hypothesis has been completely confirmed and it is clear that for the mentioned species, the maximum temperature is about 200 • C. Above this temperature, there are already significant changes in the chemical structure, especially the hemicellulose components (see more information in Introduction). Higher values of some properties (e.g., MOE) at the lower level of treatment, i.e., "165", are more or less related to the fact that the changes in the wood structure are negligible and only the positive effect of the lower moisture is manifested, which at the higher stage "210" is important due to significant changes in the chemical structure, especially the hemicellulose components of the polysaccharide complex [9][10][11]13]. The general trend corresponds with the results specified for the example in the handbook of the International Thermowood Association [50], etc. [28,30,51].
Pentosans prevail among hardwood hemicelluloses, whereas hexosans are predominant in coniferous species. However, hemicelluloses are mainly copolymers of different carbohydrates. Xylans predominate among hemicelluloses in all deciduous species, in particular, glucuronoxylans. Mannane fractions are more skeleton than filling material of wood, having good links with the cellulose (see the representation of selected types of wood in the Table 6), mainly galactoglucomannans. Following on from this, the results of a comparison of the values of the selected properties of deciduous woods compared to representative of coniferous woods, which are much more resistant, are emerging. This is the same for static and dynamic mechanical loading (see Table 7). The decrease of properties in hardwoods is higher for beech and birch, which have a higher value of reference properties than the alder, but the values at "210", especially in toughness, are almost identical, which means that the decrease is ultimately stronger. It is necessary to realize that the variability of most properties [52,53] is not eliminated by the heat treatment, often on the contrary.
From the above mentioned, at the finale, eventually limiting conditions are developed so as to guarantee the appropriate and safe utilization of the relevant type and degree of treatment of the modified wood in terms of its required utility properties. From the achieved results, it is clearly seen that usage of wood, treated with high temperatures, especially wood of deciduous species (i.e., beech and birch), is not suitable for construction purposes, because of the significant decrease of bending strength and toughness.

Conclusions
There is a partial increase in the values of most properties at a lower treatment temperature, eventually leading to the preservation of values at the level of untreated wood, for example, for birch, the modulus of rupture increased by 26%, the modulus of elasticity by 24%, and the hardness in the radial plane by 34%. This is related to the fact that chemical changes are not yet significant, and they only case the restriction of the wood's ability to absorb bound water.
With higher treatment temperatures, there is a decrease in the elastic and especially the strength properties of the heat-treated wood. At higher treatment temperatures, more markedly right above 200 • C, the significant reduction of equilibrium moisture has no such effect as the consequence of more significant changes in the chemical structure of wood and the decrease in properties is significant.
Apparently, wood with a higher hemicellulose content, i.e., a lower overall resistance, exhibits a lower density, static bending strength, and toughness. Therefore, a more significant decrease was observed for the beech and birch woods than for the softwoods at higher treatment temperatures. The decrease of toughness by about 81% for beech wood with treated temperatures of 210 • C was observed in comparison to untreated wood (respectively 86% at birch). Static bending strength at heat treated birch wood decreased by 47% (respectively 59% at beech).
The higher strength resistance (respectively mainly hardness) of birch wood compared to beech in relation to heat treatment has been demonstrated, which is probably due to the higher content of mannan fractions of hemicelluloses.
The existing correlation between static and dynamic modules of elasticity was confirmed, but it was not statistically significant in all cases.
Our research confirmed that although untreated birch wood is not equal to beech wood from the view of wood properties, the heat treatment provides wood of similar properties. The impact of the heat treatment on the wood properties is less pronounced in the case of birch than beech, and the birch is thus more suitable for thermal modification. This simple and environmentally friendly method provides one of the ways to increase the utilization of birch wood in the industry for more valuable products than fuelwood.

Conflicts of Interest:
The authors declare no conflict of interest.