Effects of Industrial Heat Treatment on Some Physical and Mechanical Properties of Iroko Wood

The purpose of this study was to determine the effect of industrial heat-treatment (ThermoWood) effect on some important properties (density, anti-swelling effi ciency (ASE), electrical resistance, strength, hardness and color) of Iroko wood (Chlorophora excelsa). Samples were subjected to heat treatment at 190 °C for 120 min during the main heat treatment period (the total time of the process was 60 hours). The results showed that the applied process caused 4.6 % decrease in density. Dimensional stability was considerably improved, with ASE values of 13.2 %. The color became darker after treatment. The process caused a signifi cant (p<0.05) reduction (3.35 to 25 %) of all investigated mechanical properties at a specifi c moisture level (12 %). However, the mechanical properties of wood are closely related to its moisture content, and heat-treated wood is less hygroscopic than untreated wood. After long-term acclimatization, most heat-treated samples had almost half the equilibrium moisture content of control samples. Although heat treated samples have lower water content and higher electrical resistance at the same soaking time, the obtained data clearly showed that heat treated samples had a slightly lower electrical resistance for the same moisture content compared to control samples.

tion of wood is known to cause changes of the cell wall nanostructure, coming along with changes of accessibility and transportability of ions and an increase of ions dissociated from acetic and formic acid formed during heat treatment.Different treatment intensity levels may cause different ER characteristics.Thermally treated wood showed signifi cantly lower ER compared to untreated wood, which might be due to acetic and formic acid formed during heat treatment (Tjeerdsma et al., 1998;Weiland and Guyonnet, 2003; Wikberg and Maunu, 2004).In contrast, a decrease in density as well as the cleavage of hemicellulose side chains may have the opposite effect on electrical resistivity (Stamm, 1956;Vermaas, 1984;Du, 1991), but these parameters do not superpose the effect of dissociated acids serving as charge carriers in wood (Brischke et al., 2014).
ThermoWood (TW), the industrial-scale heat treatment process was developed in the early 90s by the Finnish Research Center (VTT) in Finland and, since then, has been expanding the market.In the process, the wood is heated in low oxygen content (under 3.5 %) with the presence of water steam.Low oxygen contents prevent wood material from burning at high temperatures.Temperatures for the actual heat treatment period range from 150 °C to 240 °C, and the time of duration changes from 0.5 h to 4 h (Viitaniemi, 2000;Homan and Jorissen, 2004;Icel et al., 2015).
The purpose of this study was to determine the effect of industrial heat-treatment effect on some important properties (density, anti-swelling effi ciency (ASE), electrical resistance, strength and hardness) of Iroko wood (Chlorophora excelsa).This wood species was chosen as study material because of its increasing demand and importance in Turkish and European market.

Testing material and specimen preparation 2.1. Ispitni materijal i priprema uzoraka
Iroko (C.excelsa) lumber, in the size used by the company for outdoor conditions (38x100x1500 mm), was obtained from an industrial plant.The lumber was stored and pre-dried at a lumber yard.The lumber was cut into two equal pieces (A: Control; B: Heat treatment).The thermal treatment was carried out in the industrial furnace of a forest product company following the industrial process (invented by Pentti Ek,Saila Jamsa, Hannu Viitanen, Pertti Viitaniemi and patented by VTT-EP0695408) with the trade name ThermoW-ood® (EU trademark number 000922765).The temperature during the heat treatment period was 190 °C and duration time at this temperature was 120 min for this species.The total process time was 60 hours.Test samples were only cut from heartwood parts according to the cutting plan as shown in Figure 1.

UVOD
Wood has been used for both indoor and outdoor applications from the beginning of human civilization.As a biological material, biodegradability and instability under changing moisture are major disadvantages of the material.Heat treatment is one of wood modification methods applied to improve dimensional stability and durability of the material (Esteves and Pereira, 2009).Since there is no chemical application during the process, heat-treated wood is mostly considered non-biocidal and an eco-friendly method, which is an alternative to classical preservation techniques in some cases (Rapp, 2001;Patzelt et al., 2002;Anonymous, 2003;Yildiz et al., 2006).
Although thermal wood modifi cation is a longknown technology, it has received increased attention in the last decade particularly in Europe, leading to an intensifi ed industrial production and commercialization (Arnold, 2010).Therefore, lots of researches have been focused on this topic recently.Modifi cation of wood by heat treatment, historical background and methods were very well reviewed by Esteves and Pereira (2009).Depending on the variation of all heat treatment processes, including industrial-scale, semi industrial scale or laboratory experiments, the properties of heat treated wood vary a lot.The extent of change in wood properties mainly depends on the heat treatment method, wood species and its characteristics, initial moisture content of the wood, the surrounding atmosphere, and treatment temperature and the time of duration (Mitchell, 1988;Rapp, 2001;Hill, 2006;Esteves and Pereira, 2009;Dubey, 2010).
Although thermally modifi ed wood tends to be more dimensionally stable than unmodifi ed wood of the same species, mechanical properties generally have a negative effect (Viitaniemi, 1997;Santos, 2000;Militz, 2002;Shi et al., 2007;Icel et al., 2015;Oliveira Araújo et al., 2016) and knowledge on electrical properties of heat treated wood is very limited.There are some recommendations from several authors for determining wood species-specifi c resistance characteristics in addition to a temperature compensation of the measurements (Du et al., 1991;Brischke et al., 2008;Meyer et al., 2012).Different parameters have an effect on the electrical conductivity/resistivity of wood and need to be considered, such as wood species, grain direction, temperature, amount and type of extractives, and the position and type of electrodes (Davidson, 1958;Brown et al., 1963;Kollmann and Cote, 1968;Du, 1991;Brischke et al., 2008).Additionally, recent developments in wood modifi cation, the impregnation of wood with preservatives as well as chemical and thermal modifi cation, have the potential to alter electrical properties of wood (Holleboom and Homan, 1998;Smith et al., 2007;Meyer et al., 2012).Electrical resistivity (ER) is affected primarily by the number or concentration of conducting ions and the mobility of the existing charges, which means the ease with which charges move in an electric fi eld (Hearle, 1953;Brown et al., 1963).Thermal modifi ca-tive humidity of 65 ±5 % to reach Equilibrium Moisture Content (EMC) in 8 weeks.At the end of the period, the dimensions of wood samples and weights were measured to the nearest 0.01 mm.The density of the samples was measured according to ISO standard (ISO 13061-2, 2014).Since wood density is a highly variable property dependent on the species of and location on the tree, the density was calculated in terms of percentage of decrease to evaluate treatment effects.The percentages of density decrease and mass loss were calculated relative to control samples.
The density data of specimens whose moisture content deviated from 12 percent were corrected (transformed to density of 12 % moisture content) using the following conversion (TS 2472, 1976) Eq. ( 1): (1) D 12 -density at 12 % moisture content (g/cm 3 ) (shown as p in the reference standard) D r -density of actual moisture content (g/cm 3 ) K -coeffi cient of volumetric shrinkage (for 1 % moisture difference) M -moisture content of the sample (shown as W in the reference standard) The dimensions in the longitudinal, tangential (width), and radial (thickness) directions were measured to an accuracy of ±0.01 mm.The volumetric swelling coeffi cients (S) were determined using Eq. ( 2) (Rowell and Young, 1981), given below, ( Where V 2 is wood volume after wetting with water and V 1 is wood volume of the oven-dried sample before wetting.As the two groups reach MC saturation at different immersion time, a fi x immersion time was not applied for heat treated and control samples.Samples were immersed in a water bath up to a constant weight at a controlled temperature (20 °C).After removing a sample from the water bath, the surface water was removed by slightly pressing it with a clean fi lter paper, and the sample was immediately weighed.
A variety of terms can be used to describe the degree of dimensional stability given to wood by treatments.The anti-swelling effi ciency (ASE) was determined using Eq.(3) (Rowell and Young, 1981); (3) Where S 2 is the treated volumetric swelling coeffi cient and S 1 is the untreated volumetric swelling coeffi cient.
In this study, the samples of control and heat treated groups were immersed in a water bath at a controlled constant temperature of 20 °C to detect the moisture of samples using immersion cycles of 2, 4, 8, 12, 24, 48, 72, 96,120 and 360 hours.

Measurement of electrical resistance 2.3. Mjerenje električne otpornosti
Shigometry is a measuring technique based on electrical resistance (ER), commonly used to assess vigor of living trees and estimate degradation of wood decay.The Shigometer is a battery-operated, lightweight fi eld ohmmeter.It generates a pulsed direct electric current and registers the resistance to the current in thousands of ohms (kΩ) as it passes through wood (around and over fi ber saturation point) (Shigo and Shigo, 1974;Shortle, 1982).ER decreased with increasing T at constant MC, which has been reported earlier by Lin (1967), James (1968), and Du et al. (1991).The electrical resistance was measured with Shigometer in Figure 2 for the same moisture ingredient of control and treatment groups to determine the treatment effect.

Measurement of mechanical properties and hardness 2.4. Ispitivanje mehaničkih svojstava i mjerenje tvrdoće
Wood is a hygroscopic material and its properties are infl uenced by moisture content (MC).Mechanical properties of wood, which are very important especially for structural applications, are highly infl uencing properties (Forest Products Laboratory, 1999; Arnold, 2010).All samples were conditioned in an automatically controlled conditioning room at 20 °C (±2 °C) and relative humidity of 65 % (±5 %).Mechanical tests, such as compression strength parallel to grain (CS//), modulus of rupture (MOR) and modulus of elasticity (MOE), were applied according to TS EN 408+A1 in compliance with EN 408 (2010+A1:2012).
After the mechanical tests, the MC of each sample was measured according to ISO 13061-1(2014), and the moisture content of specimen whose moisture content deviated from 12 % was determined.To equalize the moisture contents (12 %) of the two groups, the conversion Eq. ( 4) was used (Icel et al., 2015): Where σ 12 is the strength at a 12 % moisture content (N/mm 2 ), σ m is the strength at a moisture content deviated from 12 % (N/mm 2 ), α is a constant showing the relationship between strength and moisture content (α = 0.05, 0.04, 0.02, for CS//, MOR, and MOE, respectively), and M 2 is the moisture content determined after tests (%) (Bozkurt and Göker, 1987).
The Brinell hardness was tested according to EN 1534.The diameter of the remaining indentation opened through the sphere was then measured with a microscope (Hill et al., 1989;Bektas et al., 2001).Since a moisture conversion formula could be used, which was suggested only for Janka hardness in the literature, the moisture content of the two groups (control and heat treatment) was not equalized.

Color measurement 2.5. Ispitivanje boje
The color was measured on the tangential and radial surface of the wood specimens before and after heat treatment by a Konica Minolta Chroma-Meter CR-400 (Konica Minolta Corp.; Japan) colorimeter according to ISO 7724-2-3 (1984).The sensor head was 6 mm in diameter.Measurements were made using a D65 illuminant and a 10-degree Standard observer.Percentage of refl ectance collected at 10-nm intervals over the visible spectrum (from 400 to 700 nm) was converted into the CIELAB color system, where L* describes the lightness and a* and b* describe the chromatic coordinates on the green-red and blue-yellow axes, respectively.From the L*, a*, and b* values, the difference in the lightness (DL*) and chromaticity coordinates (a* and b*) were calculated using group mean values.The quantities DL*, a*, and b* are the changes between pre-and post-treatment values.These values were used to calculate total color change (DE*) according to Eq. ( 5) (ISO 7724/3 1984): (5) All statistical calculations were based on a 95 % confi dence level.An independent sample t-test was applied to compare control and treatment groups.DTREG (Standard Version 10.6.3,Phillip H. Sherrod, USA) was used for statistical analysis.

REZULTATI I RASPRAVA
The obtained results and descriptive statistics for density and strength properties are shown in Table 1.After acclimatization period under the same conditions, while control samples reached equilibrium between 8.8-15 % moisture content, heat-treated samples reached 4.1-11.5 % MC in our experiments (Table 1 footer).Tankut et al. (2014) reported 10-12 % EMC for the control sample and 6-8 % for heat treated samples at 23 °C and 65 % relative humidity conditioning, for the same species and industrial heat treatment process at 180 °C .These results showed that heat treated iroko wood reached 4-5 % lower equilibrium moisture content than untreated ones under the same temperature and relative humidity conditions.Heat treated and control samples showed different sorption behavior and heat treated samples contained lower water content at the same water immersion time as shown in Figure 3.The availability and/or accessibility of the free hydroxyl groups of wood play an important role in water sorption (Boonstra and Tjeerdsma, 2006).The degradation of cellulose causes a reduction in available free polar adsorption sites, including free hydroxyl groups for water (Burmester, 1975;Hillis, 1984;Feist and Sell, 1987;Kartal et al., 2007).There was an increase in the relative proportion of the crystalline cellulose, where the hydroxyl groups are not easily accessible to water molecules (Pott, 2004), and the cross-linking of the lignin network (Tjeerdsma et al., 1998), which might hinder the accessibility of free hydroxyl groups to water (Pizzi et al., 1994).Therefore, the decreased equilibrium moisture content (Table 1), improved dimensional stability (Table 2), and water repellency of heat-treated wood are mainly caused by the decomposition or transformation of hemicellulose at high temperatures (Icel et al., 2015).These are the most probable causes of the lower moisture contents of heat-treated samples after acclimatization in our study.  (110.122 (1) 1.417 (1) 100.823 (3) 10.826 (3) 1.516 (3)  -3.3 *MOE, N/mm 2  51 7378.289 (1)819.689 (1) 114.779 (1) 5972.594 (3)635.252 (3) 88.953 (3)  -19 *CS//, N/mm 2 50 61.948 (2) 4.765 (2)  0.673 (2)  66.064 (4)  5.181 (4)  0.732 (4)  +6.6 SE -standard error / standardna pogreška; cv -coeffi cient of variation / koefi cijent varijacije, D 0 -oven-dry density / gustoća standardno suhog drva, D 12 -air-dry density / gustoća drva sušenog na zraku (1) 8.8 to 12.9 % MC, (2) 12 % to 15 % MC, (3) 4.1 to 6.9 % MC, (4) 7.8 to 11.5 % MC Compared with the control sample, wood density for the same moisture content (0 %) decreased significantly (p<0.05) with heat treatment.However, decrease in air-dry density (D 12 ) was lower than D 0 (Table 1) and insignifi cant (p>0.05).Mass loss of heat-treated samples was found to be 3.7 %.These results clearly show that the conversion formula (1) should be reevaluated for more realistic effects of heat treatment on air-dry density.The depolymerization reactions of wood polymers, in particular hemicelluloses, which are less stable under the effects of heat than cellulose and lignin, are the main cause of the decrease in density at high temperatures (Fengel and Wegener, 1989;Hillis, 1984).
The mechanical properties are closely related to the moisture content of wood.Since heat-treated wood is less hygroscopic and the (maximum) amount of bound water is reduced (Boonstra et al., 2007), it can be concluded that heat treatment provides a positive contribution to the mechanical strength properties.However, the ThermoWood process caused a significant (p<0.05)reduction (up to 27.8 %) of the investigated mechanical properties (Table 1).There is a general consensus in the literature that high-temperature heat treatments have a negative effect on wood mechanical properties and this can be explained by material losses in cell lumen and hemicellulose degradation caused by high temperature (Rusche, 1973 higher MOE for the same species and similar industrial heat treatment process.The possible explanation for the increase in MOE could lie in the fact that wood is a mixture of polymers of largely amorphous polymeric hemicelluloses, lignin and of partly crystalline microfi brils (Hillis and Rozsa, 1978).It is known that, when heat treatment is carried out above a certain temperature, many amorphous polymeric materials change from their glassy states to rubbery conditions or are prone to large plastic deformation.Individual polymers are exposed to suffi cient energy to decrease their attractive forces at the glass-transition or softening temperature, so that they become rubbery or plastic to a greater degree (Yildiz et al., 2013).
Although reports on the effects of heat treatment on the anatomical structure of wood are very limited, some anatomical changes in wood structure might also contribute to reductions in mechanical properties.For example, cracks were noticed in treated species ( 2016).These cracks may contribute to abrupt fractures, which can lead to considerably different failure behavior of heat-treated wood, abrupt fracture at lower forces, as observed in bending tests presented in Figure 4.
The reason for acclimatization at specifi c conditions (i.e., 20 °C and 65 % relative humidity), before the mechanical test and the use of conversion formulas to equalize sample moistures at a specifi c level after tests, is to determine the real effects of treatment.Control and treated samples should have the same moisture content.However, changes that occur after heat treatments are irreversible.The current results and literature support the idea that heat treated wood reaches equilibrium at lower MC than untreated wood (Dos Santos et al., 2014).Moisture content of heat treated samples was found to be lower than that of control samples after 2 years weathering (Tomak et al., 2014).After acclimatization under the same conditions, the moisture contents of heat-treated samples were lower than those of untreated ones.The mechanical test results for the actual moisture of heat-treated samples  showed that the decrease (%) in strength values was lower than in samples with specifi ed moisture (12 %) content, using conversion formulas (Table 1).Therefore, realistic moisture contents should be considered to get a realistic idea about strength properties of heattreated wood in use (Icel et al., 2015).The moisture conversion formulas/coeffi cients for different strengths should be reevaluated considering MC effect on heat treated wood for more realistic results of heat treatment on strength properties.Dimensional stability is a very important criterion of wooden materials, especially under conditions of high humidity.Therefore, many studies have been conducted on this topic.Reported results of wood stability increases (%) vary from one study to another, depending on differentiations in heat treatment methods, the standards that were followed and wood species.For example, many of previous works were carried out in a laboratory-scale oven in the presence of air, while the ThermoWood heat treatment is carried out in the absence of oxygen.Oxygen acts like a catalyst for chang-es in wood components during heating, and the presence of air during heat treatment may lead to higher ASE and higher weight loss (Wang, 2014).The general consensus among published works is that heat-treated wood has a dimensional stability advantage compared to its control (Hillis, 1984;Viitaniemi, 1997 Brinell hardness values for control and heat treated samples are presented in Table 3.The highest decrease occurred in radial section (16.5 %).Additionally, the mean hardness decrease rate between tangential and radial sections for heat treated samples was approximately 4 % higher than that of the control samples.Reported results may change depending on the testing method, heat treatment conditions, anatomical structure of wood species and wood section.Thus, a 50 % hardness reduction was found by Korkut and Guller (2008) for redbud maple samples exposed to 180 °C for 6 h.Different decrease rates were reported for three  , 1968).A difference in moisture of 4.64 % (mean) and approximately 5 % in density was observed after acclimatization at the same relative humidity and temperature between the control and heat treated groups.Therefore, hardness values presented in Table 3 were not only affected by heat treatment but also by a combined effect of moisture and density differences caused by heat treatment.The color values showed clear effects of high temperature on color changes (Table 4).Positive and negative values show a tendency of wood surface to become reddish and greenish for a*, and yellowish and bluish for b*, respectively.Lower L* value of heat treated samples indicated that color became darker after heat treatment.In this study, L* and b* values in color change tests decreased after heat treatment; on the other hand, a* value generally increased.There was no signifi cant difference (p>0.05) between the color values of radial and tangential sections.Surfaces of iroko wood turned reddish and bluish after heat treatment.Higher darkening (DL*) and total color change (DE*) were observed in this study than in a laboratory experiment carried out by Korkut (2012).Electrical conductivity of wood increases or, in other words, electrical resistivity decreases with the increase of water content at constant temperature (Kollmann and Cote, 1968;Gora and Yanoviak, 2015;Zelinka et al., 2016).Higher ER values of heat treated samples for the same immersion time were measured, but it was also determined that MC of two groups was different at the same immersion time (Figure 3).Therefore, ER data of the samples with the same MC (different immersion time) were compared.Heat treated samples reached 28 % MC, which is the maximum for ER experiment, after 15-day water soaking, while the control samples reached the same MC in 3 days.ER data for the same moisture content clearly showed that heat treated samples had lower ER compared to untreated wood (Figure 5), which might be due to acetic and formic acid formed during heat treatment (Tjeerdsma et al., 1998;Weiland and Guyonnet, 2003;Wikberg and Maunu, 2004) and/or mass loss (Brischke et al., 2014).As a result, heat treatment seems to cause a slight reduction in ER under the investigated conditions.was observed in the modulus of elasticity.However, the mechanical properties of wood are closely related to moisture content, and heat-treated wood is less hygroscopic than untreated wood.After acclimatization under the same conditions, moisture contents of heat-treated samples were lower than those of untreated ones and the mechanical test results for the real moisture of heat-treated samples showed that the decrease in strength values was lower than in samples with the specifi ed (12 %) moisture content.Exceptionally, compression strength parallel to grain was found higher in heat treated samples at their real moisture content.2. Heat treatment caused a slight reduction in ER, detectable when the control and heat treated samples had the same moisture content.Mass loss of heattreated samples was 3.7 %, which may be the possible cause of this small reduction in ER.On the other hand, because of the lower hygroscopicity of heattreated wood, treated samples attained lower equilibrium than untreated samples under the same relative humidity and temperature.These results showed that the equilibrium moisture content of heat treated iroko wood was 4-5 % lower than that of untreated samples under the same conditions.MC has a higher effect on ER than heat treatment, and under the same humidity conditions, heat treated wood showed lower ER. 3. Different failure behavior (abrupt fracture at lower forces) of heat-treated wood was observed in bending tests.4. Timber color became uniformly darker after treatment for the clear parts of the lumber.There was no signifi cant (p>0.05)difference between the color values of radial and tangential sections.Iroko wood surfaces turned reddish after heat treatment.5.The biggest decrease in hardness occurred in radial section (16.5 %).Additionally, the mean decrease in hardness between tangential and radial sections was approximately 4 % higher in heat treated samples than in control samples.Moisture content and density were different for control and heat treated groups.Density decrease and lower equilibrium are the result of heat treatment.Therefore, at least moisture content of the sample should be equalized by changing acclimatization conditions for the two groups to determine the effect of heat treatment.Psychrometric table/chart for heat treated woods requires further research.Additionally, the conversion formula (1) should be reevaluated for more realistic results of the effect of heat treatment on air-dry density.

ACKNOWLEDGMENTS
We would like to thank the Suleyman Demirel University OYP department for their partial fi nancial support (Project No: 06088-YL-14).We also would like to thank all administrative and technical staff of the NOVA Forest Products Company for their invaluable help and support.

Table 4
Heat treatment effects on color values of ıroko wood on radial and tangential section Tablica 4. Utjecaj toplinske obrade na parametre boje za iroko drvo na radijalnome i tangencijalnom presjeku