In the manufacturing process of the ecological particleboard panels, a fibrous reinforcement extracted from the pseudostem of Musa paradisiaca was used. The fibers were taken from unproductive banana trees existing in small plantations. Using a forage crusher consisting of steel hammers rotating at high speed, the banana fibers were transformed into particles with particle sizes ranging from 0.5 to 1.5 mm (Fig. 1). The banana fibers used in the manufacture of the panels had a moisture content of 12%.

Two types of adhesives were used to bond the constituents. The natural adhesive was made from CS harvested in small farmers’ plantations. The synthetic UF adhesive was purchased from local stores. All constituents used in the research were obtained in the Humjibre region, Ghana. The experiment was set in an entirely randomized design with five treatments and four replicates. The experimental treatments had different percentages of UF and CS adhesives to bond the particleboards: (100% UF–0% CS), (50% UF–50% CS), (30% UF–70% CS), (10% UF–90% CS) and (0% UF–100% CS).

Before the manufacturing stage of the particleboard panels, the natural adhesive obtained from CS was produced. The manufacturing followed the procedures established by Ramash et al. (2014). The CS was harvested and sent to the cleaning stage, where it was washed in running water and the husks removed. Using a forage grinder, the material was processed until a paste was formed. Then, it was diluted in distilled water at 26°C ± 2°C. The solution was filtered through a sieve with a 1 mm mesh and left to rest for 24 h at a room temperature of 30°C ± 2°C. Subsequently, the solution was decanted to yield the cassava starch. The obtained starch was used to manufacture the natural CS adhesive. The synthetic UF adhesive employed in the experiment had the following properties: 65% solids content, 1.3 g cm⁻³ specific density, 230 mPa · s viscosity, pH 7.5, and gel time of 65 s.

The proportions of the reinforcement of banana tree fibers and the CS and UF were previously set to carry out the manufacturing process of particleboard panels. Subsequently, the constituents were weighed one by one and mixed. The adhesives were placed in a polypropylene container and mixed according to the proportions established for each experimental treatment. A 2% (weight-weight) ammonium chloride-based catalyst was added to the mixture as the curing agent for the UF adhesive. Then, the banana tree pseudostem particles were brought together in the adhesive mixture. The mixture was homogenized, and the final material was introduced into an aluminum mold measuring 300 × 300 × 30 mm (Width × Length × Thickness). The mold was closed and positioned in a thermal hydraulic press. A pressure of 3.5 MPa was applied at 170°C ± 2°C for 8 min. After manufacturing, the panels were de-molded and, after that, conditioned for 6 days at 20°C ± 2°C and relative humidity of 62% ± 2%. After the conditioning stage, the panels were marked and cut to obtain samples for physical and mechanical tests. Four particleboard panels were produced for each experimental treatment. A quantity of 5 samples was taken from each panel to perform physical and mechanical tests. The research included a total of 20 repetitions per treatment. The flowchart in Fig. 2 shows all the manufacturing steps of sustainable particleboards and the proportions used in each treatment.

The apparent density of the particleboards was determined according to the methodology used by Widyorini et al. [23]. Initially, distilled water was added to the rim of a cylindrical container with a volume of 50 L. The panels were inserted individually into the container and then weighed. The density was determined from the mass ratio to the volume of each manufactured panel. The process was carried out in five replicates for each treatment evaluated in the research. Water absorption and thickness swelling tests were carried out to determine the influence of water on the properties of the particleboards and to establish the best environments for their application. The experimental procedures followed the ASTM D1037 [24] standard. Initially, the samples were weighed on an analytical balance. Width, length, and thickness were measured using a digital caliper. The samples were submerged in a polypropylene tray containing distilled water at 25°C ± 2°C. The samples were removed from the water after 2 and 24 h, and weighing and dimensional measurements were performed. The water absorption after 2 and 24 h was obtained from the mass variation of the samples before and after immersion. The thickness swelling was determined from the dimensional variations before and after 2 and 24 h immersions.

A static bending test was performed to find the particleboard’s flexural strengths. All procedures followed the procedures from the ASTM D 1037 [24] standard. The samples with dimensions of 250 mm × 50 mm × 20 mm (length × width × thickness) were individually placed in a universal testing machine (Inspekt 50-1, Hegewald & Peschke, Nossen, Germany) (Fig. 3A). A 50 kN load cell and a speed of 4 mm · min⁻¹ were used to perform the test. The tests were conducted and considered to have ended after the total rupture of the specimens (Fig. 3B). From the force and deformation data provided by the equipment, the results of the modulus of rupture and modulus of elasticity of the particleboards were determined.

The Janka hardness was determined to evaluate the resistance of the panels to scratches and penetrations. The tests were carried out according to the parameters established by the ASTM D1037 [24] standard. To perform the test, the samples were individually positioned in a universal mechanical testing machine (Instron 4482, Instron, Norwood, MA, USA). Then, a spherical steel indenter with a diameter of 11.28 mm was placed on the surface of the specimen. The test ended when the sphere reached 5.6 mm on the test specimen surface (Fig. 3C). The data generated by the universal machine were used to determine the Janka hardness resistance property of the particleboard panels.

All experimental data obtained were subjected to the Shapiro-Wilk normality test and Levene homogeneity test to verify the existence of statistically significant differences between the properties analyzed and promote higher reliability of the results obtained. The analysis of variance (ANOVA) test was applied and followed by the Tukey test, which is intended to compare the experimental means at a 95% probability.

The apparent density had the highest value (543.0 kg cm⁻³) for the particleboards from the (0% UF–100% CS) experimental treatment. The experimental treatment (100% UF–0% CS) showed the lowest value (493.0 kg cm⁻³) for the same property. From the analysis, it is possible to verify that the increase in the percentage of CS adhesive increased the density of the particleboard panels (Fig. 4).

A panel’s density depends on factors such as particle configuration, particle distribution in the mold, pressing temperature, adhesive reactivity, and particle compressive strength [25]. Density is a fundamental property for determining the compaction rate of particleboards. It also shows the potential of the raw material for the pressing and compaction process [26]. The high value of the density property can favor the increase of the mechanical properties of the panels [27]. Other positive factors are that the use of CS in the production of materials provides a reduction in the environmental impacts generated by synthetic adhesives, a reduction in petroleum-derived materials, an incentive for the sustainable cultivation of CS, job creation, and an increase in income for rural populations [28].

The test to determine water absorption after 2 and 24 h of immersion showed that (0% UF–100% CS) treatment had the highest values, 9.9% and 23.7%. The lowest values were 7.7% and 18.2%, presented by the (100% UF–0% CS). The treatment (50% UF–50% CS) was the one that showed the lowest values among the panels composed of the two types of adhesives, 8.7% and 20.7% (Fig. 5A).

The results show that the higher the CS percentage, the higher the water absorption of the panels. For Wronka and Kowaluk [29], materials with greater porosity and empty spaces inside absorb water more easily. The large amount of hydroxyl and amino groups in the constituents can generate higher water absorption by the panels [30]. The time the panels are exposed to environmental conditions in the presence of water favors an increase in the absorption percentage [31]. The values were lower than 47.2% and 98.0% for cellulose microfiber panels bonded with cornstarch and Mimosa-tannin adhesive after 2 and 24 h [32]. Given the scarcity of raw materials from wood, Lee et al. [33] explain that using agricultural biomass materials can be an interesting alternative to produce low-cost ecological panels.

The thickness swelling test found that the 100% CS treatment presented the highest swelling percentage after 2 and 24 h of immersion in water, 3.5% and 11.5%, respectively. After 2 h of immersion in water, the lowest value was 3.1% for the (50% UF–50% CS) treatment. The lowest value after 24 h was 9.4%, presented by the (100% UF–0% CS) treatment (Fig. 5B). The analysis showed that the increasing percentages of CS adhesive caused an increase in the thickness swelling of the particleboard panels due to the higher hydrophilic character of that material. The type of adhesive and the size of the particles used to manufacture the panels determine the ability to attract water [34]. Hydrophilicity and hygroscopicity negatively affect the mechanical properties of materials reinforced with natural fibers. These composite materials tend to swell when exposed to environments with high humidity, generating stresses in the matrix and the appearance of cracks in the material [35]. Although ecological panels produced with banana tree wastes and CS adhesive have limitations for applications in environments with high humidity, they can be an interesting indoor alternative. After 2 and 24 h of water immersion, the results obtained in the research were lower than 5.2% and 14.7% of the particleboard panels of Eucalyptus badjensis and phenol-formaldehyde resin [36]. Uemura Silva et al. [37] explains that when choosing a polymer, one must consider the curing time, mechanical performance, and sustainability issues. These characteristics favor the use of ecological panels produced with CS adhesive, as demonstrated by the data found in the present work.

Performing a general analysis of the physical properties, the treatment (50% UF–50% CS) showed the best performance among the panels manufactured with the two types of adhesives. The results of water absorption and thickness swelling were better than those presented by commercial Eucalyptus sp. and UF particleboards [38] and commercial High-density fiberboard of Eucalyptus spp. and UF [39]. Fig. 6 presents a comparative analysis of the results.

Regarding bending strength tests, the treatment (100% UF–0% CS) presented the highest value for the modulus of rupture (MOR), 16.5 MPa. The lowest values for the MOR were 12.9 MPa of the (10% UF–90% CS) and (0% UF–100% CS) experimental treatments (Fig. 7A). Among the particleboards composed of the 50% UF and 50% CS was the one that presented the highest values for the rupture and elasticity modules, being 15.0 and 2302.0 MPa (Fig. 7).

Therefore, when exposed to static bending tests, it was discovered that the panels’ resistance decreased and the proportion of CS adhesive increased. The bending properties depends directly on the microstructure of the constituents and the bond at the interface between the reinforcement and the matrix. The low resistance can also be linked to a decrease in the total surface area for the reinforcement-matrix interaction, causing the transfer of negative mechanical loads to the interface region of the material [40]. The presence of voids causes a decrease in the strength of particleboards [41]. Although the CS adhesive caused a reduction in the bending strength, the values were higher than 11.1 MPa for cocoa wood waste particleboards (Theobroma cacao) bonded with a CS adhesive [42]. On the other hand, the values reached 6.2–9.8 MPa for ground Pinus sylvestris particleboards glued with a UF adhesive [29]. All bending strength values in the present experiment met the 8 MPa minimum limit established for type-8 panels described in the JIS A 5908 standard [43]. Therefore, ecological panels produced with particles from banana tree pseudostems can be an acceptable alternative for mechanical bending applications that do not exceed resistance limits.

Concerning the modulus of elasticity, the highest observed value was 2413.0 MPa, indicated by the treatment (100% UF–0% CS), and the lowest was 2173.0 MPa, presented by the treatment (0% UF–100% CS). The (50% UF–50% CS) and (30% UF–70% CS) had intermediate values, not differing statistically from each other (Fig. 7B). The analyses showed a pattern similar to the results obtained for the modulus of rupture. The increase in the percentage of the CS adhesive also caused a reduction in the modulus of elasticity of the panels. The fibers’ orientation and the adhesive type must be considered in manufacturing the panels, as they can influence the properties. Mensah et al. [42] also found a decrease in the modulus of elasticity after adding 50% CS adhesive to panels particleboard with Theobroma cacao wood waste. Zakaria et al. [44] verified a reduction in the modulus of elasticity after adding 25% CS adhesive to agglomerated panels reinforced with palm oil. However, the results presented in the research were superior to cassava starch panels with percentages of 10%, 60%, 70%, and 80% of banana leaf fibers in studies made by Sivakumar et al. [45]. Another positive point is that all the panels manufactured in the research meet the minimum limits of 10.0 and 1550.0 MPa for the modulus of rupture and elasticity established by ANSI A208.1 [46] and 11.5 and 1600.0 MPa for general-purpose particleboards panels established by EN 312-2 [47].

Performing a general analysis of the mechanical property of flexural strength, it was found that among the panels manufactured with the two types of adhesives, the treatment (50% UF–50% CS) was the one that showed the best results for the modulus of rupture and elasticity. The values were higher than those presented by commercial Eucalyptus sp. particleboards with UF [38] and commercial high-density fiberboard [38]. Fig. 8 presents a comparative analysis of the results.

The highest resistance value for the Janka hardness test was 8.8 MPa. However, the value did not differ statistically from the 8.7 MPa indicated by the treatment (50% UF–50% CS). The treatment’s lowest value was 4.4 MPa (30% UF–70% CS) (Fig. 9).

From the results obtained, a significant decrease in the Janka hardness was verified after adding 70% CS adhesive. The reduction in the hardness of the panels may have occurred due to a higher amount of waste, reducing the resistance of the material’s surface [48]. For de Oliveira Paula et al. [49], fibrous layers on the surface provide a soft, light surface area with less resistance to piercing loads. Although the panels produced here showed a decrease in the Janka hardness compared to the 100% UF treatment, the material still presented a hardness higher than 7.1 MPa found for particleboard panels produced by Mensah et al. [42] with Theobroma cacao wood waste and UF adhesive and 5.7 MPa of particleboards with the same waste and CS adhesive. The experimental results demonstrated that the panels manufactured with banana tree waste CS adhesive have a noteworthy potential to produce ecological particleboard panels. Uemura Silva et al. [37] highlights that the use of waste in particleboard manufacturing can be an acceptable alternative to improve the life cycle of these materials, reduce the steps related to processing, and contribute to not increasing the environmental burden generated by the waste processing steps.