INFLUENCE OF DRYING ON THE PHYSICAL AND MECHANICAL PROPERTIES OF WOOD FROM TREES GROWN IN AN AGROFORESTRY SYSTEM

The cultivation of native and exotic species intercropped in an agroforestry system raises the interest for information on the properties of wood. Therefore, diff erent methods are being tested to improve the technological properties of this material, including drying, which causes changes in the physical and mechanical properties of the wood. The present study investigated the infl uence of drying on the physical and mechanical properties of wood from tree species grown in an agroforestry system. Parapiptadenia rigida (Benth.) Brenan, Peltophorum dubium (Spreng.) Taub., Eucalyptus grandis W. Hill × Eucalyptus urophylla S.T. Blake (hybrid), and Schizolobium parahyba (Vell.) S.F.Blake were the species selected for the study. Three 9-year-old individuals of each of the species were obtained from an agroforestry system. Thirty wood samples (2.5 × 2.5 × 41 cm) were extracted from each species. The wood samples were divided between temperature treatments; 6 samples were used for each heat treatment (control, 120, 150, 180, and 210 °C), which were then dried for two hours in an oven (with forced air circulation). Following the heat treatment, the mechanical properties of wood samples were evaluated to determine the modulus of elasticity and rupture, the tension in the proportional limit, and maximum force according to the ASTM D-143-94 (2000) standard. Finally, the physical properties of the retractability of the wood samples were evaluated according to the NBR 7190 (ABNT, 1997) standard. Specimens used to analyze this variable came from sections of the wood (sample dimensions: 2.5 × 2.5 × 5 cm) not aff ected by the static bending test. Our fi ndings indicate that, for all species investigated in this study, drying alters the physical and mechanical properties of the wood, with the most signifi cant changes occurring at temperatures between 120 and 180 °C.


1.INTRODUCTION
Wood, a product derived from the metabolism of trees, is considered a distinct raw material because of its unique properties; it is an organic material that is heterogeneous, porous, hygroscopic, and anisotropic (Almeida et al., 2016). These properties of wood make it an excellent source of quality material that can be utilized for the manufacture of wood products for industrial use (Gallio et al., 2018).
Owing to its unique characteristics, wood has been gaining popularity in diff erent scenarios, among them, in civil construction and furniture manufacturing, leading to an increase in the number of planted forests over time (Fontoura et al., 2015). In this context, agroforestry systems (SAFs) are likely a promising alternative, with the plantation of a consortium of both native and exotic tree species for wood production, while also conserving natural resources (Lenci et al., 2018). Accordingly, SAFs are responsible for a number of benefi ts when used for wood production, including an improvement in the constituent characteristics of wood, as well as ecological benefi ts such as reduced soil degradation (Martins et al., 2019), thus reducing the impacts on the surrounding environment.
In order to expand the application of wood from SAFs, it is necessary to identify the characteristics and technological behavior of diff erent types of wood (Motta et al., 2014). For this purpose, methods have been developed that aim to improve the physical and mechanical characteristics of the material, such as the process of drying (Freitas et al., 2016).
Drying is based on the application of heat; wood is exposed to diff erent temperatures (which are lower than carbonization) for varying time periods (Batista et al., 2011;Cademartori et al., 2012;Conte et al., 2014;Sahin and Güler, 2018). The exposure to heat has been found to change the chemical, physical, and mechanical properties of the wood (Sahin and Güler, 2018). Thus, an increase in temperature not only changes the color of the wood, but also results in a product with increased dimensional stability, less hygroscopicity, and greater resistance against wood pathogens (e. g., fungi) (Menezes et al., 2014;Zanuncio et al., 2014;Modes et al., 2017).
Among the technological characteristics that constitute the material, the physical and mechanical properties stand out, and the mechanics are determined by measuring the modulus of elasticity (MOE) and the modulus of rupture (MOR), which are usually obtained using static bending tests (Kol et al., 2017). These tests are very important because they identify the mechanical strength of wood with great precision and are consequently applied as classifi cation criteria (based on classifi cations recommended in national and international standards) to determine the fi nal use of the tested raw material (Gallio et al., 2016).
Among the physical properties, specifi c mass and shrinkage are emphasized, which are of signifi cant importance when considering the quality of the fi nal product. Wood does not shrink in a uniform manner due to its three-dimensional structure; the dimensions of wood can modify in varying ways depending on the equilibrium humidity of the environment to which it is exposed (when below the fi ber saturation point) (Cezaro et al., 2016).
Due to the vast occurrence of native and exotic species with characteristics that are diff erent from each other, the infl uence of methods that aim to improve the characteristics that constitute the material should be analyzed. Thus, the technological characterization of diff erent species through physical and mechanical tests is of fundamental importance, since it is possible to obtain information that helps in determining the fi nal use of the wood. Therefore, this study aims to investigate the infl uence of drying on the physical and mechanical properties of wood from species in an agroforestry system.

Sampling and evaluation
For each of the four species included in the study, three trees (approximately 9 years old) were sampled, and a log, 2 m in length, was removed from each individual in the region around the diameter at breast height (DBH). Central planks were subsequently made from the sapwood of the log samples in preparation for testing. Evaluations of wood samples were conducted at the UFSM/FW Wood Technology Laboratory.
For the assessment of static bending of wood, samples with dimensions of 2.5 × 2.5 × 41 cm were cut from the central planks; criteria such as anatomical orientation and dimensions of wood were taken into account when obtaining samples. Thirty samples per species were separated and identifi ed, resulting in 120 samples in total.
Drying of wood samples was undertaken in an oven with forced air circulation at diff erent temperatures (120, 150, 180, or 210 °C) for a period of 2 h, and one sample was retained untreated, to act as a control. Six replicates per species were exposed to each heat treatment. Following treatment, samples were subjected to static bending, which was carried out in a universal testing machine (model DL-2000), following the technical standard of the American Society for Testing and Materials (ASTM D-143-94, 2000). The values of modulus of elasticity and rupture, stress at the proportional limit, and maximum force, were obtained for wood samples tested. The retratibility of wood was then tested using samples cut from the original samples measuring 2.5 × 2.5 × 5 cm, according to the NBR 7190 (ABNT, 1997) technical standard. These were weighed on a scale (precision of 0.01 g), and their dimensions were measured with a digital caliper (precision of 0.01 mm) at points identifi ed and marked on samples. Subsequently, wood samples were immersed in water until complete saturation of wood fi bers occurred to obtain the weight and dimension values of the saturated samples. Swelling in the longitudinal, tangential, and radial planes was determined for each of the treatments.
The saturated samples were subsequently exposed to air-drying for 30 days, and then subjected to drying in an oven with forced air circulation at a temperature of 103 °C, until they reached constant weight. Weight and dimension values were collected again following drying to perform the calculations of shrinkage and swelling in the longitudinal, tangential, and radial planes. To obtain the anisotropy coeffi cient, only the tangential and radial planes were included in the measurements, and prior treatments were taken into account.

Experimental design and data analysis
A randomized design was used for the collection of data for analysis, characterized by a 4 × 5 factorial arrangement encompassing 4 wood species, 5 heat treatments, and 6 replications per treatment. The data were then analyzed using the Software "Statistical Analysis System" (SAS, 2003). Data were checked for normality using the Shapiro-Wilk test. Analysis of variance (ANOVA) and F-tests were used to determine variability between group means and the Bartlett test was used to check for homogeneity of variances. Average values were compared with the Tukey's means test at a 5% probability of error.

3.RESULTS
There was a signifi cant diff erence in static bending, shrinkage, and swelling between the four wood species studied, and among the fi ve heat treatments applied to the wood. Species × heat treatment interactions were also signifi cant for all variables studied.

Static bending
E. grandis × E. urophylla recorded the highest values of the mechanical properties tested: MOR; 122.7 MPa (120 °C), MOE; 12713.3 MPa (120 °C), tension at the proportional limit (TPL); 77.7 MPa (180 °C), and maximum force (MF); 3498.0 MPa (120 °C). The results indicated that the values tended to increase until the treatments reached 150 °C and 180 °C, showing a tendency to then decrease at 210 °C (Table 1). S. parahyba showed the lowest values among the studied species for all the mechanical properties and in all the heat treatments. For treatments measured at 210 °C, MOR, TPL, and MF recorded the lowest averages for the study (17.4, 16.5, and 507.8 MPa, respectively, Table 1).

Retratibility: shrinkage and swelling
For longitudinal shrinkage, E. grandis × E. urophylla showed the lowest values of all the species compared, recording 0.43% at 150 °C. However, at the highest temperature tested (210 °C), there was no notable diff erence between the species. The results for tangential and radial shrinkage showed optimal values for S. parahyba, recording 4.38% for tangential shrinkage, and 2.42% for radial shrinkage at 210 °C. Interestingly, the values did not diff er for P. dubium in all the heat treatments tested (Table 2).
P. rigida showed the highest shrinkage values among all the species tested. The longitudinal shrinkage Table 1 -Averages obtained by the static bending test for the mechanical properties of wood exposed to drying at diff erent temperatures. Tabela 1 -Médias obtidas pelo teste de fl exão estática para as propriedades mecânicas de madeiras expostas à secagem em diferentes temperaturas. was 1.03% in the 150 °C treatment, and the tangential and radial shrinkages were 9.96% and 5.14%, respectively, for both the control and treatment (Table 2).
For the anisotropic shrinkage coeffi cient, S. parahyba recorded the lowest value at 1.76, and P. dubium was the highest at 2.14. Both values were observed in the heat treatments at 150 °C ( Table 2).
The results showed that the longitudinal swelling was lowest for E. grandis × E. urophylla, with a value of 0.11% for the 210 °C treatment. The highest averages were obtained for P. rigida and P. dubium, with a longitudinal swelling of 0.59% at 120 °C for both species (Table 3). For the tangential and radial swelling, it was observed that the lowest values were recorded for S. parahyba, at 1.99% and 1.04%, respectively, at 210 °C. Conversely, the highest values were reported for P. rigida (8.38%) in tangential swelling, and E. grandis × E. urophylla (4.37%) in radial swelling (Table 3). Table 2 -Average values of shrinkage obtained from the retractability of wood exposed to drying at diff erent temperatures. Tabela 2 -Valores médios de contração obtidos a partir da retratibilidade de madeiras expostas à secagem em diferentes temperaturas.

Static bending
The static bending analysis demonstrated that drying signifi cantly infl uenced the mechanical properties of the wood tested, with changes identifi ed between the diff erent heat treatments, as well as between species. E. grandis × E. urophylla individuals had the highest values for all static bending variables (Table 1). No information was found in the existing literature on the eff ects of drying on the four species analyzed.
In terms of the average values of MOR, in general, it was found that there was a growth variation between the heat treatments at 120 °C and 180 °C, with a subsequent decrease until a temperature of 210 °C was Table 3 -Average swelling values obtained from the retratibility of wood exposed to drying at diff erent temperatures. Tabela 3 -Valores médios de inchamento obtidos a partir da retratibilidade de madeiras expostas à secagem em diferentes temperaturas.

Espécie
The same authors, studying the species E. grandis, reported mean values of MOR, which ranged from 77.10 to 63.87 MPa, lower than those of this study for the species of E. grandis × E. urophylla. In contrast, Ferreira et al. (2019) studied the wood of Hymenolobium petraeum Ducke, which belongs to the same family as P. rigida; Fabaceae, and reported a value of 71.38 MPa, from the material treated at 200 °C, demonstrating the similarity to the value observed in our study for the wood of P. rigida, 71.3 MPa (Table 1).
In the present study, it was also found that the properties of wood had a greater infl uence on drying when compared to the MOE, likely due to the constituents of the cell wall having specifi c functions related to these properties (Esteves and Pereira, 2009).
Notably, we found that E. grandis × E. urophylla showed the highest values for all static bending variables tested (Table 1). Analysis of the MOE variable of E. grandis × E. urophylla showed similar results to those reported by Fontoura et al. (2015) for the species Hovenia dulcis Thunb., which recorded a value of 11811.1 MPa. Notably, the change in temperature from 150 °C to 180 °C resulted in an increase in the MOE values of P. dubium and E. grandis × E. urophylla for the present study. This can be explained by the increase in lignin cross-links, degradation, and modifi cation of hemicellulose, as well as changes in the thermoplastic characteristics of heated wood (Gunduz et al., 2009).
The same variation was reported by Menezes et al. (2019), who observed that the increase occurred between the temperature ranges of 140 to 160 °C in the evaluation of the wood of two species (E. saligna and Corymbia citriodora K.D. Hill & L.A.S. Johnson). The results reported by Schneid et al. (2014), however, did not show variation in the MOE values when applied at room temperature, 160 °C, 180 °C, and 200 °C for the species Luehea divaricate Mart.. An increase in TPL was observed in relation to the increase in temperature, with the species E. grandis × E. urophylla standing out in comparison with others. This behavior can be explained by Faria et al. (2015), who reported that an increase in temperature had a direct infl uence on TPL, and resulted in an increase.
We found that, for the most part, the average values of MF for P. rigida and E. grandis × E. urophylla were the highest at temperatures of 150 °C and 120 °C, respectively (both species were higher than the control). Similarly, Cademartori et al. (2012) studied the species E. grandis and found the highest averages in the control treatment and at 180 °C, with an exposure of 4 hours.
Other research studying fl exion characteristics in the wood of the species Attalea funifera Mart., found a value of 1353.0 N, which is in accordance with S. parahyba heat-treated at 120 °C, which recorded 1220.7 N in the present study . Overall, it was found that, as the temperature increased, there was a reduction in the MF variable, demonstrating that MF is often aff ected by drying; its decrease is directly proportional to the temperature rise and exposure time of the treated wood (Korkut and Budakçi, 2009).

Retratibility: shrinkage and Swelling
In terms of the values of longitudinal shrinkage in the present study, we found that there was no tendency for growth or reduction between thermal treatments. With regard to tangential shrinkage, the same behavior was observed for P. rigida and E. grandis × E. urophylla, with a reduction in values between the range of 150 to 210 °C, whereas for P. dubium and S. parahyba, the results decreased from 120 to 210 °C.
Regarding radial shrinkage, our results indicated that E. grandis × E. urophylla was the most notable; E. grandis × E. urophylla showed a decrease in radial shrinkage with treatments applied after heat treatment at 120 °C. Another study investigating E. grandis × E. urophylla reported 4.5% shrinkage in the radial direction (Eleotério et al., 2015), which is in accordance with the results of this study. Further, Ferreira et al. (2019) investigated radial shrinkage in H. petraeum and found that wood treated at 180 °C for two hours recorded a radial alteration of 4.71%, in line with our results for S. parahyba, when heat treated at 210 °C.
Results for the anisotropic coeffi cient showed a reduction in values with increased temperature, which are according to the results from other research. This is expected to be caused by the degradation of accessible hydroxy groups, which are directly associated with the absorption of water in the cell walls of wood, thus resulting in a reduction in the anisotropic coeffi cient (Poubel et al., 2013).
We also found that the swelling variable in the longitudinal direction, for E. grandis × E. urophylla, showed a reduction behavior with increased temperature. Notably, the values found for the control samples of our study are in accordance with those found by Huller et al. (2017); E. grandis recorded a swelling of 0.54% under similar testing conditions. Interestingly, the tangential and radial senses of the species P. dubium, when heat-treated at 120 °C, were found to be comparative to another woody species, Tectona grandis L.f., without the application of heat treatment, highlighting the similarity between the two species (Gil et al., 2018).
As observed for the anisotropic coeffi cient of shrinkage, there was a decrease in the results of anisotropic swelling coeffi cient with an increase in temperature. A study by Huller et al. (2017) evaluating Eucalyptus cloeziana F. Muell. found that wood samples not exposed to drying recorded a value of 1.66%, similar to our fi ndings for E. grandis × E. urophylla, when treated at 150 °C.
This behavior can be explained by the high percentage of shrinkage and swelling in the tangential plane, compared to the radial plane of wood, which is defi ned by the orientations and dimensions of the wood being tested, thus allowing for contrasting dimensional variations of the anatomical planes (Oliveira et al., 2010).
In this sense, the smaller the diff erence between the tangential and radial planes, the more dimensionally stable the wood (Acosta et al., 2020). Therefore, as a general rule, anisotropic coeffi cients below 1.5 establish wood as 'optimal,' whereas values between 1.5 and 2.0 characterize wood as 'normal' (Logsdon et al., 2008). Accordingly, the higher the coeffi cient, the greater the probability of cracking and warping in the wood, thus making it more dimensionally unstable (Müller et al., 2014).
E. grandis × E. urophylla recorded the highest values for all static bending variables.
Additionally, we found that an increase in temperature directly infl uenced the values of shrinkage and swelling (retratibility of wood) in the diff erent dimensional planes.
The anisotropic coeffi cient values of wood decreased with increasing temperature.
Drying alters the physical and mechanical properties of the four wood species studied.