Characterization of randomly oriented strand boards manufactured from juvenile wood of underutilized wood species

The size of the wood particles is one of the factors that significantly affect the properties of boards. In particular, the bending strength depends significantly on the size (length and thickness) of the wood strands. Since different types of wood were investigated with different densities and anatomical structures, it was important to find out whether the cutting process affected the formation of the sizes of wood strands. The dimensions of the wood strands, the resulting specific surface values and the corresponding slenderness ratios are set out in Table 1.

The lowest average thickness (0.70 mm) was recorded in beech strands and the largest (0.92 mm) was recorded in pine strands. However, the analysis of strand thickness showed that strands from all investigated wood species did not differ among themselves in this parameter. The longest length (107 mm) was recorded in alder strands and the shortest (78 mm) in pine strands. Only the alder strands differed significantly (p < 0.05) from the spruce strands in terms of length. Furthermore, the alder strands differed significantly (p < 0.05) in length from the spruce, pine, poplar and oak strands. In turn, the pine strands were significantly (p < 0.05) different in length from the willow, alder, birch and beech strands. The analysis of the width of the strands showed that, according to this parameter, the strands of all investigated wood species differed insignificantly (p ≥ 0.05) from each other, except for the birch, beech and oak strands. The width of the birch strands was the largest and differed significantly (p < 0.05) from the width of the beech and oak strands. At the same time, the widths of the birch, beech and oak strands differed insignificantly (p ≥ 0.05) from the width of the strands of other wood species. Furthermore, the strands obtained from the various investigated wood species differed insignificantly (p ≥ 0.05) in width from the spruce strands. It was also observed that as the density of the wood species decreased, the specific surface of the strands increased.

Therefore, only the alder strands were significantly (p < 0.05) different in length from the spruce strands. The strands of other wood species differed insignificantly (p ≥ 0.05) from the spruce strands in terms of thickness, length and width. Furthermore, there were no significant differences between coniferous and deciduous wood species in terms of strand sizes obtained from them. Minor differences in the strand sizes of different wood species indicate that the cutting process did not affect the strand sizes. Although it is obvious that in the process of cutting different wood species, the consumption of electricity for cutting will be different; however, this was not the purpose at this stage of the study.

The density of the input material ranged from 392 kg/m³ (spruce) to 761 kg/m³ (oak) and influenced the final board density, which ranged from 656 kg/m³ (spruce) to 799 kg/m³ (hornbeam) (Table 2). The densities of boards made from spruce, pine, poplar, larch and alder strands differed insignificantly (p ≥ 0.05), considering that the densities of these wood species differed slightly from each other. Similarly, the densities of boards made from birch, beech, oak and hornbeam strands differed insignificantly (p ≥ 0.05) from each other, considering that the densities of these wood species differed slightly from each other. The density of boards made of larch and willow strands differed insignificantly (p ≥ 0.05) from the density of boards made of beech and oak strands. In addition to the input density of wood, the compaction ratio (CR) also affected the density of the boards. Boards using such strands reached almost the same density since larch and willow strands were more compacted (56% and 28%, respectively) than beech and oak ones (4% and 11%, respectively). OSBs from low-density wood species such as alder, poplar and pine had almost the same density as OSBs from spruce strands; moreover, they also demonstrated a similar CR. It is quite natural that low-density wood species demonstrated higher CR than high-density wood species (Table 2).

Changes in the wood structure are related to the CR of the panel, higher CR caused higher changes in the deformation of the wood structure for OSB. A comparison of the non-densified wood and the final panel showed significant changes in the wood structure in the spruce OSB, shape deformation of the cell walls in the birch OSB and no changes in the beech panel compared to beech wood (Fig. 2).

The density of a board is determined by the input density of the wood species used and the pressing modes. Since in this study the pressing parameters and the planned CR of strands to form a board carpet were constant, the density of the boards was determined by the input density of the wood and the CR. In this study, the softwoods and hardwoods demonstrated different compressibility at high pressure and temperature due to the difference in the wood species (Table 2) and this difference was reflected in the vertical density profiles (VDPs) of the boards. Figure 3 illustrates the correlation between the density of boards and the density of wood based on the complete set of results obtained for all investigated wood species. The high value of the determination coefficient R² = 0.8 demonstrates a good linear correlation between the density of the wood and the density of the board. As the density of used wood increases, the density of the board increases too.

The MC of the panels after conditioning showed lower values for spruce, alder, birch and beech, and higher values for the rest of the panels.

The vertical density profile (VDP) showed a typical U-shape (Kelly 1977) for some of the low-density wood species, especially spruce, pine, poplar and alder (Fig. 4). These results are consistent with previous findings (Wang and Winistorfer 2000; Beck et al. 2009). A comparable VDP was also observed by Beck et al. (2009) for birch with a slightly higher density near the surface and almost even density through the thickness. The other high-density wood species (birch, beech, hornbeam, oak and willow) showed a nearly uniform density profile (small difference between face and core) through the thickness of the boards without higher densities at the surface. This can be explained by the easier compaction of low-density wood species strands (CR = 52–68%) under high pressure and temperature in the surface layers than in the core layer of the boards. Although high-density wood species strands are very hard to compact (CR = 4–28%), the density of such boards in the core and surface layers is almost the same (Fig. 4). Zhuang et al. (2022) also found that the difference in density between the surface and core layers of OSBs was less with an increase in species wood density. Furthermore, for the same reason, it was not possible to achieve the target thickness of the boards using high-density wood species strands. Panels made from these wood species showed a high spring break immediately after the hot press was opened – for most of them up to 15.5 mm from the target thickness of 12 mm. This will need to be taken into account in practice by selecting the appropriate pressing parameters for high-density wood species.

In general, VDP is one of the factors that most affects the physical and mechanical properties of wood-based panels. Some researchers (Chen et al. 2010) argue that a U-shaped VDP is beneficial for MOR and MOE and that a flatter VDP would result in greater dimensional stability and IB. The homogeneous density profile of an OSB provided by high-density wood species can be seen as a benefit that favors a uniform stress distribution in structural members and facilitates the use of fasteners in building construction.

The TS and WA results were unexpected. As can be seen in Table 3 and Fig. 5, lower values of TS and WA were found in boards of higher density made with high-density wood species (beech, hornbeam, oak and willow, but not birch). The TS results of this study are contrary to common experience. In this study, with an increase in the density of the boards, the TS decreased. Usually, as the density of boards increases, the TS increases and WA decreases. The reason for this can be found when we look at the density profile of the boards (Fig. 4). As mentioned above, high-density wood species showed an almost uniform density profile with practically the same density through the thickness of the boards. Usually, board samples absorb water through the edges. In the board samples made from low-density wood species, the density of the core layer is low, and water can easily penetrate through the edges of the board. However, in boards made from high-density wood species, the density of the core layer is high, practically the same as the surface layers, and it is difficult for water to penetrate inside the board. However, the same cannot be said about the effect of the density of boards on the TS. In this study, with an increase in density, TS decreases as well, which does not agree with the generally known trend. However, various research results on the influence of density on TS conducted by other authors are also contradictory. Wu and Piao (1999) demonstrated that TS and WA increased with increasing specimen density. Chen et al. (2010), investigating the influence of board density on OSB properties, found that TS and WA linearly decreased with increasing board density, similarly to our findings. In the same way, Geimer (1982), studying flakeboard TS, pointed out that TS was lower in high-density boards. Jin et al. (2009) indicated that WA was well correlated with the density of random OSB, whereas the relationship between TS and density was not straightforward. TS initially increased with an increase in density, but after reaching a peak point of around 650 kg/m³, the trend reversed. Wang et al. (2003) compared the properties of commercial aspen, pine and mixed hardwood OSBs and observed that, within a species, their data did not show consistent relationships among WA, TS and board density.

TS after 24 h showed low values for hornbeam, beech and oak of around 10% compared to spruce at 25.4% which was the highest value of TS after 24 h (Table 3). The same trend was observed in TS after 48 h. A study by Akrami et al. (2014a) of OSBs made of beech and poplar with a density of 720 kg/m³ showed TS after 24 h 12% and 23%, respectively, which is on the same trend as this study, TS after 24 h for beech 10.3% and poplar 13.9%. WA showed the lowest value for beech of 22.5% after 24 h, oak and larch showed WA 24.8% and 32.9%, respectively. The highest values for WA after 24 h were 58.3% and 58.0% for birch and spruce, respectively. The same trend was observed for WA after 48 h.

We found that the density of wood had a strong effect on the TS and WA of the boards (Fig. 5). Both TS and WA decreased with an increase in the density of the wood. Since the dimensions of the strands were much larger than the dimensions of the structural elements of the wood tissue (cells), the strands can be considered as particles that have the properties of solid wood. Therefore, high-density wood species, which have less porosity, absorb less water and swell less than low-density wood species. Accordingly, boards with higher density absorb water more slowly, reducing the rate of TS. Jin et al. (2009) stated that the relatively poor strand-to-strand contacts in low-density boards result in weak bonds that are more likely to fail during water soaking, increasing TS. Moreover, low-density boards are highly porous because of the many voids in their structure, which facilitate the penetration and absorption of water, and thus swelling.

The CR and the specific surface of the strands also have a strong effect on the TS and WA. Boards with a higher CR and a specific surface of strands showed higher TS and WA. As shown in Table 2, a lower wood density resulted in a higher CR; consequently, according to Hsu (1987), it will increase the WA and TS values. The stress within the board is partially released when it is immersed in water.

It is known that TS and WA are exposure-time dependent. Wu and Lee (2002) demonstrated that a higher density three-layer board tended to have greater TS, whereas in single-layer uniform-density boards, density had a mixed effect on TS. Jin et al. (2009) argued that despite the greater swelling potential, higher-density boards usually do not receive full water penetration during a 24-hour water-soaking test to release deformation. Therefore, the findings of this study indicate that the relationship among TS, WA and board density appears to be fairly complex and more studies are needed to identify reliable correlations.

Spruce and alder boards had the highest MOR values (34.6 MPa and 34.3 MPa), and pine boards had the lowest MOR value (20.5 MPa). The MOR values showed that there was no statistically significant difference between the values of spruce (34.6 MPa), larch (25.9 MPa), poplar (25.2 MPa), willow (27.8 MPa), alder (34.3 MPa) and birch (27.1 MPa) (Fig. 6). Similarly, there was no significant difference in MOR values between pine, larch, poplar, willow, birch, beech, oak and hornbeam. Such a picture is unexpected since wood species of different densities were used in the experiments. Since we produced all boards under the same pressing parameters and the analysis of strand sizes showed that the strands had the same size regardless of the wood species, the MOR values practically do not differ significantly among themselves. Furthermore, no strong relationship was found between MOR and board density of the boards (R² = 0.26); a slightly better relationship was found between MOR and wood density (R² = 0.39). Moreover, a negative effect of the densities of wood and boards was found on the MOR. The relationships between MOR and the densities of board and wood are shown graphically in Fig. 7. However, as is well known, the MOR is affected not only by the density of the boards but also by many other factors, including CR, the thickness and length of the strands, slenderness ratios, specific surface, the amount of adhesive in the outer layers etc. The compaction and slenderness ratios, as well as the specific surface of the strands, had a positive effect on MOR (Fig. 7). In general, an increase in all these parameters increased MOR. The strongest relationship demonstrated the specific surface of the strands (R² = 0.70) based on the complete set of results obtained for all wood species. A higher specific surface of the strands resulted in increased MOR because strands with a higher specific surface area occupy a larger volume than those with a lower specific surface. Thus, strands with a higher specific surface at high pressure and temperature during hot pressing are more compacted and increase strand-to-strand contact because of the higher number of strands and reduced void volume. Longer and thinner strands (higher slenderness ratio) resulted in better MOR than shorter and thicker strands as a result of better alignment among longer strands and increased overlap length (Barnes 2001). Other authors (Beck et al. 2009; Chiromito et al. 2016) also showed that the length and thickness of the strand influenced MOR and MOE positively and significantly.

The results of this study show that to have comparable bending properties for spruce, pine, beech, oak and hornbeam boards, pine strands must have a higher slenderness ratio whereas beech, oak and hornbeam strands must have a higher specific surface, which can be obtained using thinner strands at a given length.

A study by Lunguleasa et al. (2021) focused on OSBs made from willow, poplar, pine and spruce and showed MOR values in the parallel direction around 44.54 MPa for willow and 27.07 MPa for spruce; in this study a decrease in the values from willow, poplar to pine was also observed. A comparison of birch, willow and poplar in a research by Dumitrascu et al. (2020) gave values for MOR in the parallel direction of 36.0 MPa, 36.6 MPa and 43.3 MPa and for MOR in the perpendicular direction of 33.0 MPa, 35.3 MPa and 18.2 MPa, respectively.

A similar trend, with values on the MOR, was also observed on the MOE values (Figs. 8 and 9). The highest values of MOE were 5,185 MPa for spruce and 4,472 MPa for alder, without significant differences. The MOE values for larch (4,110 MPa), poplar (3,815 MPa) and willow (3,939 MPa) did not show a significant difference from alder. The results of this study are in good agreement with those obtained by other authors (Chen et al. 2010; Zhuang et al. 2022). Zhuang et al. (2022) showed a similar tendency for MOE values to decrease with increasing species wood density. Chen et al. (2010) found that the relationships between the bending properties and density were best described with quadratic regression curves. Both the MOR and MOE values increased with increasing board density, but the increase slowed when the density reached approximately 690 kg/m³ (for MOR) and 660 kg/m³ (for MOE). The density of boards in this study was in the range of 650–800 kg/m³.

The study by Lunguleasa et al. (2021) showed MOE values for OSBs made from willow of around 4,700 MPa and for poplar around 4,600 MPa in the parallel direction; for pine, MOE values in the parallel direction were around 3,200 MPa and in the perpendicular direction around 3,000 MPa, which is comparable to the results from our study. Poplar OSBs in an earlier study by Dumitrascu et al. (2020) showed a lower value of MOE in the parallel direction of 3,147 MPa than the values of this study; willow (5,665 MPa) and birch (4,636 MPa) showed higher MOEs in the parallel direction than we did in this study focused on unoriented OSBs.

The IB strength reached the highest average value for birch, 0.9 MPa, but there was no significant difference between spruce and other wood species (Fig. 10), which confirms that, for IB as one of the crucial parameters, it is possible to make OSBs with comparable properties from all our investigated underutilized wood species. The reasons for the lack of difference in IB strength values may be the same as those for MOR. The degree of correlation between IB and board density in this study was very low, and this may be the result of differences in VDP and strand geometry. The effects of investigated variables on IB are very complicated. In general, increasing the density of wood resulted in a decrease of IB, whereas increasing the CR and the specific surface of the strands increased IB. The relationships between IB and these factors are shown graphically in Fig. 11. Beck et al. (2009) also found that strand geometry and wood species had no significant effect on IB. Xu and Winistorfer (1995) found a positive correlation between IB and OSB, although the coefficient of determination (R²) was only 0.20–0.25.

OSB reached higher IB values in the study by Dumitrascu et al. (2020) of 1.05 MPa, 1.33 MPa and 1.28 MPa for birch, willow and poplar, respectively. The IB values from the study by Beck et al. (2009), with a comparison of the different strand dimensions, showed the average values for aspen of 0.69 MPa, and for birch of 0.77 MPa, which are comparable with the results of this study. A study by Akrami et al. (2014a) for OSBs made from beech and poplar with a density of 720 kg/m³ showed IB of 0.7 MPa, and 0.6 MPa, which are contrary to this study. IB for beech was 0.55 MPa and for poplar 0.67 MPa.

It is well known that high-density woods are difficult to bond because of their thicker cell walls and smaller diameter lumens (Frihart and Hunt 2010). In this case, the adhesives do not easily penetrate into the wood, preventing good mechanical interlocking. Furthermore, high-density wood species are characterized by a higher concentration of extractives that can interfere with the cure of adhesives (Frihart and Hunt 2010). Therefore, much greater pressure is required to compress stronger, stiffer, high-density wood to bring the adhesive into contact with the wood surfaces.

The overall results obtained for OSB properties met the EN 300 (2006) standard of minimum requirements for OSB/2 (MOR > 20 MPa, MOE > 3,500 MPa (except boards made with pure pine, birch, beech or oak), IB > 0.32 MPa, TS 24 h < 20% (except boards made with pure spruce, pine, alder or birch) for board thickness range of 10 mm to 18 mm) and OSB/3 (MOR > 20 MPa, MOE > 3,500 MPa (except boards made with pure pine, birch, beech or oak), IB > 0.32 MPa, TS 24 h < 15% (except boards made with pure spruce, pine, alder or birch) for board thickness range of 10 mm to 18 mm) boards. Furthermore, IB was found to reach values 1.7–2.8 times higher than the levels specified in the EN 300 (2006) standard. TS after 24 h for boards of larch, poplar, willow, beech, oak and hornbeam strands was 9.1–13.9%, which is around 54.5–30.5% and 39.3–7.3% less than the EN 300 (2006) standard allows for OSB/2 and OSB/3, respectively.

The results of this study correspond to the results described by other authors (Chen et al. 2010) who found that the effects of board density on MOR, MOE and IB can be approximated with convex quadratic functions, implying that in the lower density region, these board properties improve more rapidly with increasing density than for higher density. When the density reached a certain elevated level (660–690 kg/m³), little benefit was achieved with further increases in density.