The bending test was done according to Japanese Industrial Standard JIS A 5908-2003 (2003) using an Instron machine model 4204. A central point load was applied at a crosshead speed of 10 mm/min across an effective span of 150 mm. The test specimens measured 50 mm × 200 mm, and four replicates were prepared for each particleboard type. Internal bonding testing was also carried out by the Japanese Industrial Standard (JIS A 5908-2003) (2003) using the same type of Instron machine, model 4204, as the bending test machine. A sample specimen was strapped to the 50 mm x 50 mm block and then secured in the machine. The tension load was applied vertically on the face of the board with a tension loading speed of 2 mm/min. Subsequently, the maximum load (P') was recorded at the breaking point of the perpendicular tensile strength of the board. Four test samples were prepared to determine internal bonding performance.

The thickness swelling (TS) and water absorption (WA) tests of the composite board were determined based on the percentage increase in thickness and weight after water immersion, relative to their initial values (Zuo et al. 2018). These tests were conducted by soaking the composite boards in water, after which TS and WA were measured following the Japanese Industrial Standard JIS A 5908-2003. Specimens with dimensions of 50 mm × 50 mm × 10 mm were oven-dried at 105 ± 2 °C until constant weight, then immersed in distilled water at room temperature (25 ± 2 °C) for 24 h. After immersion, samples were surface-dried and re-weighed, and the thickness was re-measured. Four replicates were tested for each board.

The flammability of the composite boards was evaluated using the Limited Oxygen Index (LOI) test, following ASTM D2863 with a Fire Testing Technology LOI apparatus. Rectangular specimens (80 mm × 10 mm × 5 mm) were positioned vertically in the sample holder, and the minimum oxygen concentration required to sustain combustion was recorded. All tests were performed in a controlled atmosphere, and 15 replicate specimens of each board were prepared to ensure accuracy.

TGA was conducted with a Shimadzu TGA-50, heating the samples from room temperature to 920 °C at 20 °C/min under nitrogen flow. Weight loss profiles were obtained as a function of temperature and analyzed using the Freeman–Carroll method (Hatakeyama & Quinn, 2006).

FTIR spectra were recorded using a Perkin Elmer spectrometer at 2 cm⁻¹ resolution with 15 scans, employing the KBr pellet method at a 1:100 sample-to-KBr ratio.

The bending strength results of all samples are presented in Fig. 1. Among the OPT composite boards, sample A2 exhibited the highest bending strength, while sample B2 recorded the highest value for RW. Specifically, the mixture of OPT with 10% PVA reached 5.94 MPa, whereas RW with 10% PVA showed 5.34 MPa. Lamaming et al. (2015) reported that PVA addition enhances the mechanical performance of composite boards. The lowest bending strength values were obtained from A1 (100% OPT, 0.40 MPa) and B1 (100% RW, 0.97 MPa). As shown in Fig. 1, the bending strength decreased with the incorporation of CaCO₃ and NaCl. According to Hongzhen et al. (2017), excessive CaCO₃ (>10%) weakens intermolecular forces, destabilizes the internal structure, and reduces bending strength.

The ANOVA results for bending strength in Table 2 revealed significant differences among particleboard formulations for both OPT and RW (p < 0.001). This indicates that variations in PVA, CaCO₃, and NaCl contents had a measurable influence on the bending performance of the composites. For OPT-based boards, the between-group variance (F = 12.523) was notably higher than the within-group variance, confirming that formulation played a dominant role in determining flexural strength. Similarly, RW composites also showed significant treatment effects (F = 5.546), though the lower F value suggests that RW was less sensitive to additive changes compared with OPT. These findings highlight that the mechanical properties of the boards are strongly dependent on binder and filler composition, consistent with previous studies reporting that filler loading can either enhance or weaken stress transfer across the particle–matrix interface.

The Duncan post-hoc test shown in Table 3 further clarified which formulations differed significantly. For OPT, sample A2 clustered in a distinct subset with the highest bending strength, confirming the positive role of moderate PVA content in improving matrix cohesion. In contrast, formulations with high CaCO₃ and NaCl loadings (A3, A5, A6) fell into subsets associated with lower bending values, suggesting that excessive filler disrupted interfacial bonding. For RW, a similar pattern was observed, where B2 produced the highest bending strength, while filler-heavy formulations (B3, B4, B6) yielded lower values. These results demonstrate that PVA alone significantly enhances bending properties through hydrogen bonding interactions, whereas high mineral filler contents tend to dilute the effectiveness of the binder. Collectively, the ANOVA and post-hoc outcomes validate the experimental observations and confirm that formulation optimization is critical for achieving the desired balance between strength and functionality.

a. Uses Harmonic Mean Sample Size = 5.143

Figure 1 also presents the internal bonding (IB) performance of OPT and RW composites. The highest IB for OPT was observed in A2 (OPT + 10% PVA), while B4 (RW + 10% PVA, 20% CaCO₃, and 30% NaCl) showed the highest IB for RW composites. The lowest IB was recorded in B1 (control RW, 0.13 MPa). Overall, the addition of CaCO₃ and NaCl reduced IB in both OPT and RW, likely due to the higher CaCO₃ content. The decline in IB is attributed to changes in particle composition, where CaCO₃ partially covered the OPT particle surfaces, limiting the adhesive bonding of PVA. Similar findings were reported in CaCO₃-filled pine flake composites with phenol–formaldehyde resin (Tao et al. 2019). In general, RW particleboards exhibited higher IB than OPT particleboards, which is consistent with the findings of Lee et al. (2018).

The performance variations can be attributed not only to the filler content but also to interfacial interactions between PVA and the lignocellulosic particles. PVA, rich in hydroxyl (–OH) groups, forms hydrogen bonds with hydroxyl groups present on cellulose and hemicellulose surfaces of both OPT and RW. This molecular interaction improves adhesion and stress transfer across the matrix, explaining the enhanced internal bonding observed at moderate PVA levels, consistent with recent reports of hydrogen bonding and physical entanglement effects in PVA–wood systems (Liu et al. 2024; Huang et al. 2024; Chen et al. 2024). However, excessive filler loading (NaCl or CaCO₃) disrupts this interaction by occupying surface sites and reducing effective contact between PVA and wood particles, thereby lowering mechanical strength.

The ANOVA results for internal bonding in Table 4 demonstrated statistically significant differences among formulations for both OPT and RW composites (p < 0.001). The F values 10.319 for OPT and 6.680 for RW indicate that binder and filler ratios had a strong influence on tensile strength perpendicular to the board surface. Similar to the bending test, OPT appeared more sensitive to additive variation than RW, suggesting that the lower density and higher porosity of OPT fibers amplified the effects of PVA and mineral fillers.

The Duncan post-hoc analysis shown in Table 5 revealed that A2 had the highest internal bonding strength, confirming that moderate PVA addition optimizes interfacial adhesion. In contrast, formulations with excessive fillers for A3 and A6 exhibited the lowest IB, reflecting poor particle binder contact. For RW composites, B2 and B5 formed distinct subsets with the highest IB values, whereas B1 showed the weakest bonding. These results emphasize the role of PVA as a hydrogen-bonding binder while demonstrating that excessive mineral loading diminishes adhesion by masking hydroxyl-rich wood surfaces.

a. Uses Harmonic Mean Sample Size = 5.143

Dimensional stability, typically evaluated through thickness swelling (TS) and water absorption (WA), is a crucial property influencing the service life of particleboards under environmental exposure. As shown in Fig. 2, both WA and TS of OPT and RW particleboards decreased with the incorporation of PVA, CaCO₃, and NaCl. Srivabut et al. (2018) reported that such reductions occur because chemical fillers act as barriers that limit water penetration into the boards. This modification therefore enhances the durability of both OPT and RW based composites. Swelling occurs when water molecules diffuse into the cell wall, and the extent of swelling is directly related to the quantity of water absorbed (Younis 2017). Figure 2 further indicates that RW particleboards exhibit better dimensional stability than OPT boards, consistent with findings by Lee et al. (2018). Wahab et al. (2018) also explained that the porous structure of OPT particles promotes water uptake, leading to higher WA, which in turn results in swelling and greater TS.

The ANOVA results for TS and WA confirmed highly significant differences among board types for both OPT and RW (p < 0.001). For TS in Table 6, OPT had a higher F value (11.280) compared with RW (7.516), indicating stronger treatment effects in OPT-based boards. Similarly, for WA in Table 8, the influence of formulation was more pronounced in RW boards (F = 56.757) than in OPT (F = 13.265), highlighting that dimensional stability improvements depend not only on additive type but also on intrinsic wood characteristics.

Duncan post-hoc tests in Tables 7 and 9 showed that boards containing fillers and PVA for A5 and B5 exhibited significantly lower TS and WA values compared with control boards for A1 and B1. Particularly, B5 recorded the lowest water uptake, reflecting the effectiveness of mineral fillers in reducing porosity and limiting water penetration. Conversely, controls and low-PVA boards absorbed more water and swelled extensively, demonstrating the crucial role of both binder and filler in enhancing dimensional stability. These results validate that filler incorporation, when optimized, can act as a hydrophobic barrier, thereby prolonging the service life of the boards under humid conditions.

a. Uses Harmonic Mean Sample Size = 5.143

a. Uses Harmonic Mean Sample Size = 5.143

The Limited Oxygen Index (LOI) test is commonly used to evaluate the flammability and flame retardancy of materials. It is regarded as one of the key screening and quality control techniques in flame-retardant research and development (Birck et al. 2014). A higher LOI value indicates that a material is more resistant to ignition and combustion (Zhang et al. 2018). Figure 3 presents the LOI results for a comparative study of OPT and RW particleboards.

As shown in Fig. 3, the LOI of OPT particleboards increased with the addition of CaCO₃ and NaCl fillers. The highest value was obtained for sample A6, which consisted of OPT with 20% PVA, 10% CaCO₃, and 30% NaCl. Wang et al. (2017) reported that CaCO₃ enhances flame retardancy by delaying ignition, reducing the heat of combustion, and improving the overall fire resistance of particleboards. While OPT boards generally exhibited higher LOI improvements with additive incorporation, RW-based boards also demonstrated notable flame-retardant enhancement, achieving values above 27%, comparable to OPT. The slightly lower values in RW composites may reflect differences in density and inherent lignin content, which influence char formation during combustion. Compared with conventional flame-retardant systems such as ammonium polyphosphate, borates, and halogen-free mineral fillers, the NaCl/CaCO₃ approach demonstrates competitive LOI values above 26% while maintaining lower toxicity and cost (Rajaei et al. 2017; Schirp and Su 2016). While APP- or phosphorus-based systems often yield higher LOI values, they may involve higher cost or environmental concerns, highlighting the novelty of adopting NaCl as a low-cost, sustainable alternative in wood composites.

The ANOVA results for LOI in Table 10 revealed a clear distinction between OPT and RW composites. OPT boards showed significant differences among formulations (F = 42.882, p < 0.001), indicating that NaCl and CaCO₃ loadings had a strong influence on flame resistance. By contrast, RW composites exhibited no significant differences (p = 0.206), suggesting a more uniform flame-retardant performance across formulations. This discrepancy highlights the stronger responsiveness of OPT to additive incorporation, likely due to its more porous structure and higher moisture affinity.

The Duncan post-hoc test in Table 11 identified A6 as having the highest LOI, indicating that it falls into a distinct subset with superior flame-retardant performance. In contrast, A1 showed the lowest value, underscoring the necessity of filler incorporation for fire safety. For RW, although differences were not statistically significant, a gradual increase in LOI with higher filler content was observed, with B5 and B6 showing slightly better fire resistance than B1. Overall, the statistical analysis validates that mineral fillers, particularly in OPT boards, substantially enhance flame retardancy, positioning NaCl/CaCO₃ systems as viable, eco-friendly alternatives to conventional halogenated flame retardants.

b. Uses Harmonic Mean Sample Size = 5.143

Thermogravimetric analysis (TGA) is a widely applied technique for assessing the thermal stability of materials and identifying the decomposition behavior of polymers within composite systems, such as OPT and RW-based particleboards (Younis 2017). Figures 4a and 4b display the thermogravimetric (TG) weight-loss curves and derivative thermogravimetric (DTG) profiles, respectively, for comparative particleboards made from OPT and RW. In this evaluation, the specimens were subjected to a controlled increase in temperature to assess their heat resistance and decomposition patterns.

As shown in Fig. 4a, the TGA curves of both OPT and RW composites revealed two distinct stages of weight loss occurring between 50–150 °C and 200–250 °C. The first stage corresponds to the evaporation of moisture and volatile compounds (Lamaming et al. 2015), while the second stage reflects polymer decomposition leading to carbonaceous char formation (Zuo et al. 2018). For both OPT and RW particleboards bonded with PVA, weight loss began around 200 °C, with complete decomposition observed at approximately 543 °C, consistent with Schirp and Su (2016). This behavior is expected, as PVA undergoes dehydration at the onset of thermal degradation, producing unsaturated compounds (Younis 2017).

Moreover, particleboards containing 10% PVA, 10% CaCO₃, and 30% NaCl (samples A5 and B5) demonstrated greater thermal resistance. This improvement can be attributed to the decomposition of CaCO₃ into carbon dioxide at elevated temperatures, which contributes to thermal stabilization (Schirp and Su 2016). Each additive displayed a distinct degradation pathway (Jiang et al. 2017), and Lamaming et al. (2017) further noted that the incorporation of inorganic materials enhances crystallinity, thereby improving thermal stability. The DTG results presented in Fig. 4b provide a detailed view of mass loss across distinct temperature ranges (Amini et al. 2013). In both OPT and RW composites, the addition of CaCO₃ and NaCl shifted decomposition onset to higher temperatures. However, RW composites consistently retained slightly higher residual mass at 600 °C compared to OPT, suggesting greater thermal stability, likely due to their denser microstructure and lower porosity.

Figure 5 shows the FTIR spectra of OPT and RW particleboards. A broad band near 3415 cm⁻¹ corresponds to O–H stretching vibrations, characteristic of hydroxyl groups in cellulose and hemicellulose. Absorption bands between 2800–3000 cm⁻¹ reflect C–H stretching vibrations associated with polysaccharides and lignin, while peaks around 1104–1110 cm⁻¹ and 1155–1159 cm⁻¹ indicate cellulose contributions from both crystalline and amorphous domains (Lee et al. 2018). The pronounced O–H band also supports hydrogen bonding between PVA and lignocellulosic surfaces, which underpins the enhanced dimensional stability and bonding strength observed at moderate PVA contents. In CaCO₃-filled boards, carbonate-related bands further suggest its contribution to flame retardancy, as thermal decomposition releases CO₂ that dilutes flammable volatiles and delays ignition.