Investigating sapwood and heartwood density changes of short rotation teak wood after chemical and thermal modification using X-ray computed tomography

2.1 Sample preparation

Wood was collected from trees in a 13-year-short rotation teak from plantation forest – from Bogor, West Java, Indonesia (6° 33′ 15′′ S/106° 40′ 07′′ E). It has a rainy tropical climate with an average rainfall 3,095 mm/year with an average temperature of 30 °C. The trees had an average diameter at breast height (DBH = 1.3 m) of 27 cm. Two trees were harvested and the log section in length of 1 m were taken from the bottom part starting from DBH of each tree stem. According to Figure 1, two disks (T1 and T2) with 10 cm of thickness were cut transversely from the bottom part of each stem, approximately 30 cm from soil. From each disk, a 1 cm wide strip of wood was taken in the middle and was cut from bark to bark through the pith. The increment samples, measuring 1 × 1 cm in cross-section, were prepared from bark to pith. To obtain radial density profiles, this increment sample was analyzed from pith to sapwood using XCT scanning. The same increment sample was scanned before and after treatment. There were 10 increment samples from disk T1 and 8 increment samples from disk T2. The chemical and thermal modifications implemented were 4 increment samples for furfurylation (FA), 3 increment samples for thermal treatment at 150 °C (HT150), 3 increment samples for thermal treatment at 220 °C (HT220), 4 increment samples for glycerol-maleic anhydride treatment at 150 °C (GMA150) and 4 increment samples for glycerol-maleic anhydride treatment at 220 °C (GMA220). The remaining wood from cutting the disks was converted into flat sawn timber with a thickness of 2 cm using a band saw. Meanwhile, anatomical samples were taken from short rotation teak boards taken 1 m from the disk cut. Wood boards were re-sawn into wood block samples for anatomical analysis with the dimensions 200 × 50 × 20 mm³ (L × R × T). The wood blocks were cut into smaller pieces for microscopic observation.

Figure 1:

Sampling for XCT scanning. A, furfurylation; B, thermal treatment at 150 °C; C, thermal treatment at 220 °C; D, glycerol-maleic anhydride treatment at 150 °C; and E; glycerol-maleic anhydride treatment at 220 °C.

2.2 Chemical and thermal modification

The chemical modification performed, more specifically furfurylation, was similar as in Martha et al. (2024). A furfuryl alcohol solution was prepared with 45 % of furfuryl alcohol, 5 % of tartaric acid as catalyst, and distilled water. The wood samples were impregnated with this solution using a vacuum-pressure impregnation process in an autoclave. First, a vacuum was applied at 8–10 kPa for 15 min followed by immersion in furfuryl alcohol solution under vacuum conditions 4–5 kPa for 15 min. Then, the pressure process was carried out in a pressure of 1,200 kPa for 30 min. The polymerization process was carried out after release of pressure at 120 °C for 16 h under nitrogen condition.

Thermal modification was performed at 150 and 220 °C in a nitrogen environment as described in Martha et al. (2024). The oven temperature was increased by a heating rate of 20 °C/min from room temperature to the target temperature for 4–6 h. The target temperature was maintained for 20 h.

A combination of chemical treatment using glycerol-maleic anhydride and thermal modification was conducted according to Martha et al. (2024). Glycerol-maleic anhydride adduct (GMA) was synthetized by mixing 1 mol of glycerol (92.09 g/mol) and 2 mol of maleic anhydride (98.06 g/mol) followed by heating at 80 °C for 3 h. GMA was then solubilized in water with the final concentration of 10 % w/w. The impregnation was performed at a vacuum of 8–10 kPa for 30 min followed by a pressure of 1,200 kPa for 60 min. Curing temperatures of 150 and 220 °C were applied for 20 h under nitrogen condition.

2.3 Microscopic characteristics

Light microscopy was used to analyze the microstructural changes of short rotation teak wood after chemical and thermal modification. Three kinds of treatments were investigated: untreated, FA, and GMA220. Five cube samples of 1 × 1 × 1 mm³ (L × R × T) representing sapwood (SW) and heartwood (HW) were used to measure microscopic characteristics. The wood samples were softened using boiling water and then the cross sections were cut in a sliding microtome with a thickness of 20 μm. The sections were rinsed using an ethanol series (50 %, 70 %, 90 %, 95 %, and 99 %). To maintain the chemical integrity of the sample, the section was not stained. Analysis of anatomical characteristics followed the standards as defined by the International Association of Wood Anatomists and was performed by an Olympus SC 100 microscope. Images were taken at 100× magnifications for analysis. The microscopic section of each sample was analyzed to measure vessel diameter in radial and tangential direction (μm). A minimum of 100 vessels were examined for the measurement of each parameter using Fiji Image-J software (Wheeler et al. 1989).

2.4 X-ray CT scanning

2.4.1 Image acquisition and reconstruction

Figure 2 shows the workflow used to assess the density profile before and after treatment using XCT scanning. Increment samples were scanned before and after chemical and thermal modification using a flatbed scanner and the Nanowood X-ray μCT scanner, which was custom-built at the UGent Centre for X-ray Tomography (www.ugct.ugent.be) and recently refurbished in collaboration with TESCAN-XRE (www.XRE.be). This scanning system is equipped with two separate X-ray sources. This scanning system has also two different X-ray detectors with a complementary spectral sensitivity to accommodate the varying X-ray energies emitted by both sources. Increment samples were mounted in a sample holder made from dry floral foam containing 9 increment samples in batch. A reference material with a constant density value of 1.4 kg/m³ was mounted on the edge of the dry floral foam during the scanning process. All samples were scanned at an average tube voltage of 80 kV, a tube power of 8 W, a voxel size of 100  μ m and an exposure time of 500 ms per projection. The scanning process took approximately 40–60 min per sample group. The same scan settings were applied for both before (PRE) and after (POST) chemical and thermal modification samples. After image acquisition, the X-ray CT images were reconstructed using Octopus Reconstruction with beam hardening correction producing 3D reconstructed volumes. Because the scanning was done in bulk, the volumes of each sample (both PRE and POST) were cropped individually to produce individual volumes. The 3D volume datasets were then pre-aligned visually by flip and/or rotation transformation in ImageJ. After, each volume then was saved in 16 bit TIFF-stack format.

Figure 2:

Flowchart of XCT assessment for analyzing the changes in density and density profiles of wood samples before and after chemical and thermal modification.

2.5 Data analysis

The average density was calculated based on all test samples of T1 and T2 disks. The statistic tests were divided into two parts, pith-to-bark section and equalized section with balanced sapwood-to-heartwood wood ratio (Figure 2). The pith-to-bark section were calculated using all image slices from pith to sapwood without concerning the proportion of heartwood and sapwood. In this study we wanted to investigate the transition section containing heartwood and sapwood which presents significantly different wood properties. To obtain a better understanding of the homogeneity of the treatment on the heartwood and sapwood in the transition area, calculation was also performed on same proportions of sapwood and heartwood collected directly on each side of transition wood to avoid variability due to the presence of juvenile wood in inner heartwood (equalized section). The transition line, based on the color of the samples was used to determine the equal segment length. The Z-test was selected as the appropriate statistical method for comparing the mean density values of each treatment to the reference value of 0. In addition, a paired T-test was conducted to compare the mean density value between sapwood and heartwood. The Z-test and T-test were performed using R software (R Core Team 2014).

3.1 Microscopic analysis

The mass changes of chemical and thermal modification treatment in increment sample from pith to the bark are presented in Table 1. The mass change increased after furfurylation treatment by 26 %, indicating the presence of additives in the increment sample. The mass loss occurred due to thermal treatment, since high temperatures caused an increase in mass loss. The addition of GMA followed by thermal treatment caused a higher mass change compared to thermal treatment. Even if small differences were observed with previous results reported by Martha et al. (2024) probably due to the size of the samples, results indicated the same trends.

The results of the vessel area and vessel diameter in a radial and tangential direction of untreated and treated short rotation teak wood are presented in Table 2. Heartwood showed a slightly lower vessel diameter compared to sapwood. A similar result was also reported by Ren et al. (2023), measuring a larger vessel diameter in sapwood compared to heartwood of Michelia macclurei. Several studies reported that vessel diameter increases from the pith to the bark to increase hydraulic conductivity in the outer wood (Cardoso et al. 2015; Longui et al. 2014; Rodriguez-Zaccaro et al. 2019; Sperry et al. 2006). In addition, Darmawan et al. (2015) reported that short rotation teak wood has a high proportion of juvenile wood. The presence of juvenile wood in heartwood might have an impact on the size of the vessel diameter. Juvenile wood has a smaller vessel diameter size than mature wood (Bhat et al. 2001). The vessel area and the radial and tangential vessel diameter increased significantly after FA treatment. For sapwood, there was an increase of 13.8 %, and 2.2 % in the tangential and radial vessel diameter, respectively. Meanwhile for heartwood, there was an increase of 10.5 %, and 20.5 % in the tangential and radial vessel diameter, respectively. The increase in vessel diameter might be due to a combination of chemical reaction, physical pressure, and mechanical forces during the FA treatment. During the impregnation process, the introduction of furfuryl alcohol into the wood exerted pressure, therefore the vessel element might swell. The polymerization process of furfuryl alcohol that produced the formation of a polymer network within the wood structure might also cause volume expansion of the entire wood structure, including vessels. As shown in Figure 3, the brownish deposits were observed only in the cross section of both FA treated sapwood and heartwood, while no deposits were observed in the case of untreated control samples. These deposits were most probably furfuryl alcohol polymer deposits resulting from the FA treatment, even if the hypothesis of extractives deposits cannot totally exclude. On the other hand, the vessel area and vessel diameter in a radial and tangential direction decreased significantly after GMA220 treatment. There was a decrease in the tangential vessel diameter of 12.8 % and 12.6 % for sapwood and heartwood, respectively. The decrease in vessel diameter for the GMA220 treatment occurred due to the effect of thermal modification. Thermal treatment could lead to collapse of the vessel, particularly if the internal stress exceeded the mechanical strength of the cell wall. This collapse might lead to a decrease of the vessel diameter. Biziks et al. (2013) also reported that vessel area and vessel diameter of Betula pendula slightly decreased after hydrothermal treatment. Thermo-reticulation reactions resulting from depolymerization of lignin and recondensation reactions reported during wood thermal modification may also explain this decrease of vessel diameter (Gérardin 2016).

Figure 3:

Microscopic cross sections of sapwood (SW) and heartwood (HW) for untreated, furfurylation (FA), and glycerol-maleic anhydride at 220 °C (GMA220) treatment. The red arrow points to a brownish deposit.

3.2 Effect of chemical and thermal modification treatment on density

Density is an important characteristic since it determines the wood properties, especially physical and mechanical properties. To visualize the density profile of short rotation teak, the representative samples for each chemical and thermal modification treatment from pith to sapwood are presented in Figure 4. The technique used in this study made it possible to determine heartwood and sapwood. The flatbed scanning image in Figure 4A shows the transition between heartwood (HW) and sapwood (SW). In Figure 4E, the density profiles before and after treatment are plotted according to the different treatments of FA, HT150, HT220, GMA150 and GMA220. Overall, the density profile of short rotation teak wood increased from pith to sapwood, a similar result was also reported by Curvo et al. (2024). These authors noted that X-ray densitometry indicates a higher wood density from pith-to-bark direction with variations in growth rings. Darmawan et al. (2015) also reported that density of long and short rotation teak wood increased proportionally from pith to bark. This is attributed to the presence of juvenile wood near the pith which impacts the density of short rotation teak wood to be lower near the pith.

Figure 4:

Visual interpretation of the density profile. (A) Flatbed scanning image from pith to sapwood of short rotation teak before chemical and thermal modification treatments. (B) Density visualization based on an X-ray CT image of short rotation teak before treatment. (C) Density visualization based on an X-ray CT image of short rotation teak after chemical and thermal modification treatments. (D) Map of the density change before and after chemical and thermal modification treatments (E). Density profiles before and after chemical and thermal modification treatments. (F) Density profile changes before and after chemical and thermal modification treatment. FA, furfurylation; HT150, thermal treatment at 150 °C; HT220, thermal treatment at 220 °C; GMA150, glycerol-maleic anhydride treatment at 150 °C; and GMA220, glycerol-maleic anhydride treatment at 220 °C. The black vertical dotted line corresponded to the transition line.

The mean density value of short rotation teak wood before treatment was 560 kg/m³. The increase in density was observed in short rotation teak wood after FA treatment. The mean density value increased by 14.3 % after FA treatment. This result confirmed our previous study indicating that FA treatment increases the density of short rotation teak wood (Martha et al. 2021a). The increase in density is due to polymerization of furfuryl alcohol polymers within the cell wall (Baysal et al. 2004; Dong et al. 2014, 2016). As shown in Table 2, the change of vessel lumen area and vessel diameter after chemical and thermal treatment were correlated with wood density. The results in Figure 3 also confirmed the presence of furfuryl alcohol that led to the increase in density. The increase of density due to FA treatment was higher in the area near the pith (Figure 4F). However, the density profile in Figure 4E was seen to be more homogeneous from pith to the sapwood after FA treatment. This could affect the physical properties of short-rotation teak wood, in particular dimensional stability. Martha et al. (2024) reported that FA treatments can homogenize properties of sapwood and heartwood of short rotation teak wood.

On the contrary, a decrease in density was observed for thermal treatment. The increase of thermal treatment temperature led to a larger decrease in density. The mean density of short rotation teak decreased by 3.6 % and 9.1 % for HT150 and HT220, respectively. Furthermore, the trend of density change after thermal modification was uniform from pith to sapwood. The degradation of the vascular tracheal wall and the degradation of chemical compounds, especially hemicellulose, into volatile products and evaporation or degradation of extractives are the main reasons for the decrease in density (Esteves and Pereira 2009; Pfriem et al. 2012). This study showed that increasing the heating temperature led to a greater decrease in density. Shchupakivskyy et al. (2014) reported that oak wood heat treated at three different temperatures (180, 200 and 220 °C) with a duration of 2 h resulted in stronger reductions in density according to the treatment temperature.

Previous studies reported that density decreased after GMA treatment (Martha et al. 2021c; Mubarok et al. 2019). In this study, the mean density of short rotation teak increased by 1.7 % for GMA150 and decreased by 7.4 % for GMA220. The increase in the mean density of GMA150 might be due to GMA impregnation and its polymerization with wood components. The density change of GMA220 was lower than HT220 indicating GMA treatment could reduce wood degradation during the curing process. The density change tended to decrease slightly from pith to sapwood. Sapwood exhibited a higher decrease in density change compared to heartwood. This indicated that sapwood was more sensitive to thermal degradation and therefore more easily degraded.

Figure 5 shows the density changes of short rotation teak wood after chemical and thermal modification. According to the Z-test (95 % confidence level), there was a significant difference in density change after chemical and thermal treatment (Figure 5A). To evaluate how the treatment homogenized the density in the transition area, the analysis was performed with the same proportion of heartwood and sapwood (Figure 5B). The result shows that the trend of density change was similar for each treatment, indicating that there is insufficient proof that the proportion of heartwood has an impact on the magnitude of the density change.

Figure 5:

Density changes of short rotation teak wood after chemical and thermal modification for average images of pith-to-bark section (A) and average images of equalized section with balanced sapwood-to-heartwood wood ratio (B). FA, furfurylation; HT150, thermal treatment at 150 °C; HT220, thermal treatment at 220 °C; GMA150, glycerol-maleic anhydride treatment at 150 °C; and GMA220, glycerol-maleic anhydride treatment at 220 °C.

The density changes for the different treatments on sapwood and heartwood are shown in Figure 6. A T-test was applied (95 % confidence level) shows that the density change between sapwood and heartwood is statistically different (Figure 6A). A similar phenomenon also was observed with the same data number by drawing the same proportion of heartwood and sapwood (Figure 6B). The mean density values and density changes are shown in Table 3 in relation to their initial mean density values to illustrate the extent of density changes for the different treatments on sapwood and heartwood. Overall, heartwood presented a lower initial density compared to sapwood. The heartwood in short rotation teak contains juvenile wood that has thin-walled fibers with lower density and lower strength properties (Clark et al. 2006; Darmawan et al. 2015; Gryc et al. 2011). The presence of juvenile wood might have a significant contribution to the increase in density due to FA treatment. As shown in Table 3, the density change in heartwood was higher than that in sapwood. Conversely to the GMA treatment, heartwood was more resistant to thermal degradation than sapwood. The higher resistance of heartwood to thermal stress explained the lower decrease in density of heartwood compared to sapwood.

Figure 6:

Density changes of short rotation teak wood for the different treatments on sapwood (SW) and heartwood (HW): (A) Calculating based on average images from of pith-to-bark section, (B) calculating based on average images of equalized section with balanced sapwood-to-heartwood wood ratio. FA, furfurylation; HT150, thermal treatment at 150 °C; HT220, thermal treatment at 220 °C; GMA150, glycerol-maleic anhydride treatment at 150 °C; and GMA220, glycerol-maleic anhydride treatment at 220 °C.

Contrary to what could be expected at the beginning of this study, results were not in line with results generally reported in the literature on density evolution between heartwood and sapwood, especially on conventional long rotation teak wood. Heartwood generally presents a higher density compared to sapwood. However, in our study, the initial density of short rotation teak wood increased slightly from pith to sapwood. The presence of juvenile wood in the area near the pith was responsible of lower density value of heartwood. Consequently, the hypothesis suggesting that a difference in density between heartwood and sapwood could be reduced after chemical modification due to higher impregnability of sapwood was not confirmed, given the measured initial densities. The variation in density profiles of the chemical and thermal modifications appeared similar to those observed for the original wood. The results indicate that the chemical and thermal modification did not have a significant effect on the density profile, especially in the homogenization of the density profile between sapwood and heartwood after chemical modification. However, short rotation teak wood containing sapwood and heartwood does become more homogeneous, especially in leachability, color and durability, after FA and GMA220 treatment (Martha et al. 2024).