3.1 Particle characterization
3.1.1 Particle thickness
The particle thickness was adjusted via the knife projection during the cutting process in the disk flaker. When analyzing the particle thicknesses, it was found that the actual average particle thickness was 0.38 mm for a set blade projection of 0.35 mm. With a set blade projection of 0.7 mm and 0.9 mm, lower particle thicknesses of 0.53 mm on average were produced in both cases. These results indicate that the particle thickness in the machining of forest-fresh paulownia wood can only be adjusted to a limited degree via the knife projection. According to these results, the particles of different thicknesses were divided into two particle variants.
3.1.2 Particle width and lenght
Analysis of particle geometry was carried out on four particle variants of different thickness (0.38 mm and 0.53 mm) and different particle fraction (B and C). Figure 1 (a) shows that the particle widths of fraction B particles are almost equally distributed. This can be seen graphically as the two graphs of the distribution function are close to each other. The same can be seen in the comparison of the particle widths of fraction C. This can be explained by the screening process in the tumbler screen, where particles are sifted based on their widths.
Based on the particle lengths, all four particle variants differ from each other (Fig. 1 (b)). Differences between the particle fractions can be seen especially by observing the particle lengths. Particles of lower thickness (0.38 mm) have been found to be longer than particles of higher thickness in both fraction B and fraction C. The slenderness ratio of thinner particles is thus positively influenced. Based on these differences, four variations of particle geometry for board production were investigated (Table 2).
3.1.3 Bulk density of paulownia particles
For further characterization of the particle variants, the bulk density of the particles was determined on three samples per variant (according to factory standard FHIS 281). The particle variants of fraction B (particle thickness 0.38 and 0.53 mm) showed higher bulk densities than the variants of fraction C (Table 2). This can be explained by the higher pore volume of fraction C particles. When the particles were scattered in the sample chamber, this pore volume formed more quickly between the coarser particles, which were not only arranged longitudinally but also edgewise in the chamber.
Table 2
Particle variants and their geometric dimensions and properties
Fraction
|
Average particle thickness [mm]
|
Average particle width [mm]
|
Average particle length
[mm]
|
Slenderness ratio [-]
|
Average bulk density [kg/m3]
|
B
|
0.38
|
0.83
|
3.89
|
10.2
|
60.9
|
B
|
0.53
|
0.79
|
3.21
|
5.8
|
64.3
|
C
|
0.38
|
1.56
|
9.58
|
25.2
|
34.4
|
C
|
0.53
|
1.48
|
8.19
|
15.4
|
42.1
|
3.2 Board production: analysis of core temperature in the particleboard
In order to investigate the potential of the high-frequency pressing technology to produce lightweight PB from paulownia, comparative boards were produced using conventional hot-pressing technology. The potential of the HF press is evident in the pressing speed of lightweight (300 kg/m3) PB (Fig. 2). Compared to the hot press, pressing times could be reduced by up to 60%. A shortening of the pressing time by using high-frequency technology was also found byMäbert et al. (2015). Heating in the core of the board cross-section up to a temperature of 100°C could already be accelerated 1.5 times using HF technology compared with HP. Whereas a heating rate of 2–4 K/min up to a target temperature of 110°C was achieved in the hot press process, the target temperature of 140°C could be guaranteed in the HF process with heating rates of 39–65 K/min.
Different heat input behavior when using the HF pressing technology was found in relation to the particle geometry. As can be seen in Fig. 3, the temperature curve as a function of pressing time for 0.53 mm thick particles can be characterized with a temperature plateau at 105°C, while for 0.38 mm thick particles this plateau can already be seen at 100°C. The temperature plateau is caused by water steam escaping from the mats to be pressed, levelling the temperature increase when the evaporation temperature is reached (Bolton et al. 1989). Accordingly, thicker particles retain moisture up to a higher temperature (105°C) in the particle mat than thinner particles.
The total pressing times for PB (300 kg/m3) made of Paulownia fraction B are the same for the two different particle thicknesses (t140 = 125 s). PB with a particle thickness of 0.38 mm of fraction C has a longer through-heating time (t140 = 130 s) for the same panel density. The highest through-heating time (t140 = 142 s) is required for PB consisting of 0.53 mm thick fraction C particles.
The relationship between thermal conductivity and particle size of the particle variants (Wei et al. 2015) is confirmed therein. Coarser particles (particle thickness 0.53 and fraction C) lead to lower bulk densities (Table 2), resulting in a higher interparticular air void space when the particle mat is being formed manually. These voids lead to a reduction of the thermal conductivity within the particle mat and accordingly results in a higher heating time to reach the target temperature of 140°C in the center of the particle mat. The specific use of particles of certain geometries (fraction B) can therefore lead to process advantages and to a reduction of the HF time during the pressing process.
3.3 Mechanical and physical properties
3.3.1 Bending properties
Particle thickness shows a significant influence on MOR of the tested specimens at both levels of the panel’s density (300- and 400 kg/m3) (Fig. 4). Thicker particles (0.53 mm) achieve higher bending strengths (MOR) at the same raw density. An influence of particle thickness on MOE of paulownia particleboard could not be detected. Nevertheless, both properties- MOE and MOR can be influenced by the density of the PB. An increase of the board’s raw density from 300- to 400 kg/m3 leads to a doubling of these bending strengths.
The influence of thicker particles on MOR is related to the beneficial absorption of tensile and compressive stresses, which was also shown by Benthien and Ohlmeyer (2018) when varying the particle thickness. Since single-layer PB was manufactured, the distribution of particle sizes is homogeneous over the entire cross-section of the PB. However, since the bending stress particularly affects the outer layers of the PB, it can be assumed that when particles with higher thicknesses are used, they positively influence the bending properties.
The length and width of the particles (particle fraction) significantly influence both MOE and MOR of lightweight PB made from paulownia. Figure 6 shows that PB consisting of fraction C particles have significantly higher bending strengths at a board density level than PB consisting of fraction B particles. One explanation is based on the increasing slenderness ratio of the particles with increasing particle length. It has been shown that particles with a higher degree of slenderness have a positive influence on the bending properties of PB (Deppe and Ernst 2000). The same phenomenon can be seen in Fig. 7 for the consideration of MOE. Fraction C particles have a higher length compared to fraction B, which have a positive effect on the stiffness behaviour of the PB when subjected to bending stress, as they can absorb higher stresses (Benthien and Ohlmeyer 2018).
Elastic properties of paulownia single-layered PB with 8 mm thickness and densities from 300 to 1100 kg/m³ were analysed by Esteves et al. (2023). Bending strength ranged from 5 N/mm², for a density under 500 kg/m³ and 340 N/mm² for the highest density. MOE behaviour exhibited the same trend as MOR, namely 200 N/mm² and 5000 N/mm². Esteves et al. (2023) observed also that an increased density is directly correlated with a decreased particle size (0.25-4 mm), allegedly due to higher compaction when using fine-grained particles. In the case of 10 mm single-layered PB made of paulownia, bonded with UF at 200°C, (Nelis et al. 2018) reported a MOR of 4-7.5 N/mm² for a density of 350 kg/m³ and 17.5 N/mm² for PB with 500 kg/m³ density. In this study, MOE was about 1500 and 3000 N/mm² for 350 and 500 kg/m³ panel density. The requirements for P1, P2, P3 and P4 were met for densities of 500 kg/m³ and higher. Kalaycioglu et al. (2005) studied 3-layered PB with a thickness of 18 mm, bonded with UF, without hydrophobic agent. It was found out that MOR was about 14 N/mm² for the panels with a density of 550 kg/m³ and a corresponding MOE of 2396 N/mm².
All manufactured board variants in this study meet the requirements for bending strengths according to DIN CEN/TS 16368 for type LP1. For all PB with densities of 400 kg/m3, the requirements for type LP2 (use for furniture) are met.
3.3.2 Internal bond
At a board density level of 300 kg/m3, PBs consisting of thicker particles (0.53 mm) show significantly higher IB than boards made of thinner particles (0.38 mm). With higher compression of the middle layer due to the increase in raw density to 400 kg/m3, the influence of particle thickness on IB can no longer be demonstrated (Fig. 8). Here, the higher compaction leads to an increased bond between the particles. The result is an increase in bonding quality due to the minimization of interparticular voids. Possible explanations are to be found when considering the particle geometry. When using thicker particles, fewer particles distributed over the entire plate thickness are required to lie on top of each other in order to achieve the total plate thickness. A smaller number of joints between the particles can be assumed in this respect. Since the thicker particles are pressed together more strongly during the compaction process due to their higher bulk density and the smaller interparticle pore space (Table 2), it can also be assumed that the crosslinking of the adhesive between the particles is more advantageous, which could have a positive effect on IB, especially in the case of PB with low bulk densities.
IB of PB made of paulownia are significantly higher at both levels of the board density when fraction B particles are used than when fraction C particles are used (Fig. 9). Since the two fractions differ significantly in terms of the length and width of the particles, it can be assumed that shorter, narrower particles (fraction B) have a positive influence on IB.
These results are consistent with the findings of (Esteves et al. 2023), where IB for paulownia PB ranged from 0.4 to 0.7 N/mm² at densities under 500 kg/m³, but at a thickness of 8 mm, that is half of the thickness of the boards in the present study. For a paulownia PB with 10 mm thickness, (Nelis et al. 2018) determined an IB of 0.8 N/mm², corresponding to a density of 350 kg/m³ and 1.25 N/mm² for 500 kg/m³ density. In the case of 3-layered paulownia PB of 18 mm studied by (Kalaycioglu et al. 2005) was determined an IB of 0.67 N/mm² for a density of 550 kg/m³.
3.3.3 Water absorption and thickness swell
At a raw density level of 300 kg/m3, the water absorption (WA) after 24 h of PB consisting of 0.38 mm thick particles is significantly lower than when using thicker particles (Fig. 10). The influence of particle thickness decreases with increasing raw density of the panels. The reduced WA at higher panel densities is consistent with the expectation supported by previous studies (Roffael and Rauch 1972). As a result of higher raw densities, the compaction of the particles in the PB increases. A decreased interparticular pore space is the result. As the pore volume decreases, also the volume into which water can penetrate decreases, which would lead to an increase in the mass of the specimen. This relationship is also observed in the WA as a function of the particle fraction used (Fig. 11). This shows an indirectly proportional relationship between decreasing WA and increasing raw density of the test specimens. It can be seen that, at both levels of panel density, particles of fraction C lead to lower WA than particles of fraction B. One explanation is based on the fact that, due to the larger particles of fraction C, there are larger interparticle voids in the PB, but the absolute space of the enclosed air volume is larger with particles of fraction B, since the particles are smaller. Furthermore, there is the assumption that, as a result of smaller particles, the individual pores become smaller and can increasingly attract and absorb water by capillary sorption effect.
At a raw density level of 300 kg/m3, PB consisting of 0.53 mm thick particles shows significantly lower swelling values than thinner particles (0.38 mm). Since thicker particles (0.53 mm) resulted in higher IB for PBs with raw densities of 300 kg/m3, it can be assumed that this counteracts the thickness swelling (TS) due to increased adhesive curing after 24 h of water storage. In addition, it can be considered that a large number of thicker particles have a smaller surface area than the same number of thinner particles for the same volume. Consequently, the same amount of adhesive used can wet a smaller area to a greater extent. Since a particle almost completely coated with adhesive has a small number of pores into which water can penetrate, it can be assumed that this particle swells less than a particle less wetted with adhesive.
With increasing raw density, TS increases after 24 h for both particle thicknesses. An influence of the particle thickness on the TS can no longer be detected at panel densities of 400 kg/m3 (Fig. 12).
The use of particles from fraction C leads to increased TS after 24 h at both levels of bulk density. Significant differences between the particle fractions are observed at the 300 kg/m3 level (Fig. 13). This is mainly due to IB of the PBs consisting of fraction B particles (Fig. 9), which counteract the swelling of the test specimens during water storage.
In addition, the particle dimensions are decisive for the TS of the test specimens. The particle length and width depend on the particle fraction. Since the particles were produced by tangentially feeding the log sections to the disk flakers knife, the width of the particle must be characterized by a tangential cut. Since particles of fraction C are wider than particles of fraction B by about 85%, a higher absolute TS of the particles in the tangential direction is to be expected. However, the effect of the particle fraction is overridden by the effect of the bulk density itself on the TS as the bulk density increases.
For 3-layered PB with 18 mm thickness, for a panel density of 550 kg/m³ the swelling in thickness reached 12.6% (Kalaycioglu et al. 2005). For 10 mm PB and a density of 350 kg/m³, the TS measured 9% and WA 120%. For panel density of 500 kg/m³, TS increases at 12.5 %, but water absorption ecreases at 75 % (Nelis et al. 2018). For a thicker singlelayered PB of 8 mm, the swelling in thickness for densities between 400 and 500 kg/m³ is from 15.5 to 25% and the WA is included in 110–145 % interval (Esteves et al. 2023).