Inuence of different sheet compositions and commercial adhesives on the performance of multilaminated plywood

The quality of plywood panels depends on factors such as the forest species and the adhesive used in their production, and understanding the interferences of these factors in the nal properties of the plywood is of fundamental importance. The study aimed to develop multilaminate plywood panels with two forest species and two types of adhesive and to evaluate the inuences of these factors (forest species and adhesive) on the physical and mechanical properties of the plywood. The panels were produced with sheets of parica and pine with two types of adhesives, urea-formaldehyde and phenol-formaldehyde, with a weight of 150 g/cm². Then, each set was pressed for 10 minutes under a specic pressure of 0.98 MPa at a temperature of 150 °C. Three panels were produced for each type of blade and adhesive, totaling four treatments. The plywood was evaluated for physical properties (humidity, bulk density, and water absorption) and mechanical properties (parallel and perpendicular static exion and resistance to mechanical shear). The results showed that the forest species had a greater inuence on physical and mechanical properties, with the best results being observed for plywood produced with pine. The panels with sheets of parica and phenol-formaldehyde adhesive showed less moisture. The apparent density of the panels should be considered as it positively inuenced the mechanical properties and negatively impacted water absorption. It is concluded that the plywood produced can be used for internal and external applications. However, it is indicated for structural purposes as it did not meet the requirements of the NBR 31.000.001/2:2001.

with UF adhesive absorbed a greater amount of water and swelled twice as much as those produced with FF adhesive. Therefore, the type of application of the compensated panel can be determined by the type of adhesive used in its production.
Considering that both the species and the type of adhesive can in uence the nal properties of plywood, this study aimed to develop multilaminate plywood panels with two forest species and two types of adhesive and to evaluate the in uences of these factors (forest species and adhesive) on the physical and mechanical properties of the nal products.

Materials And Methods
Collection of raw material and preparation of blades Blades with dimensions of 480 x 480 mm and nominal thickness of 2 mm, obtained from two forest species, parica (Schizolobium amazonicum Herb.) and pine (Pinus oocarpa), were used to produce the compensated panels. The blades of the Amazonian species parica, 6 years old, were ceded by the Concrem Group company, located in the Dom Elizeu municipality, Pará, Brazil. The pine slats were obtained from 18-year-old trees planted at the Federal University of Lavras -UFLA, Minas Gerais, Brazil. The pine trees were laminated in a rolling mill at the Experimental Wooden Panel Production Unit (UEPAM) at the UFLA. Before the panels were produced, the veneers were properly stacked in a kiln for subsequent drying, reaching 8% humidity on a dry basis.

Production of compensated panels
Two types of adhesives were used to make the plywood panels with parica and pine blades: urea formaldehyde (UF) and phenol-formaldehyde (FF). The UF adhesive had a solid content of 51.49%, a pH of 12.86, a viscosity of 590 cp, and a Gel Time of 49 seconds. The solids content of the FF adhesive was 53.69%, at a pH of 8.32, a viscosity of 465 cp, and a Gel Time of 49 seconds. The experimental design is shown in Table 1. Three panels were produced for each type of adhesive and blade (forest species).
For each treatment, with the aid of a spatula, the adhesive was spread evenly on the faces of the blades, with a weight of 150 g/m². After the production, the panels were left (assembly) for approximately 15 minutes. This assembling period is important to ensure that the adhesive is compatible with the wood veneer before being pressed. After that, the set was pressed for 10 minutes under speci c pressure of 0.98 MPa at 150°C. The manufactured panels were conditioned in an air-conditioned chamber at a temperature of 20 ± 2°C and a relative humidity of 65%, leading to an equilibrium humidity of approximately 12%.
Physical properties of compensated panels All physical tests were performed according to ABNT 31:000.05-001/1 (2004) guidelines (Associação Brasileira de Normas 2004). The apparent density of the compensated panels was determined according to the stereometric method. Brie y, the mass of the specimens was obtained on a laboratory scale with a precision of 0.001 g, and the volume was determined with the aid of a digital pachymeter with a precision of 0.01 mm, where the length, width, and thickness were measured.
To determine the equilibrium humidity of the compensated panels, initially, the wet mass of the specimens was taken, and the samples were oven-dried until constant mass. The specimens were weighed again to determine humidity. To evaluate the water absorption of the panels, rst, the mass of the specimens was measured in a laboratory scale with a precision of 0.001 g. Subsequently, the samples were immersed in a container of water and kept for 24 hours. After the immersion period, the samples were weighed again to determine water absorption.

Mechanical properties
The static bending mechanical tests with ber direction parallel and perpendicular to the test body length (of the cover sheets) and glue line resistance to shear stress in dry condition were also performed according to the recommendations of ABNT 31:000.05-001/1 (2004) (Associação Brasileira de Normas 2004). The mechanical tests were conducted on a universal testing machine (EMIC DL-30000).

Statistical analysis
The design used was entirely random, arranged in a factorial design composed of two factors: forest species and adhesive, with three repetitions, totaling 12 panels. For the statistical analysis, we considered each type of panel produced (treatments) and the species x adhesive interactions. The results of physical and mechanical properties were submitted to analysis of variance (ANOVA) and those detected as signi cant by the F test were evaluated by the Scott-Knott test at 5% signi cance, using the statistical program SISVAR -System of Analysis of Variance. Pearson's simple linear correlation analysis was performed among all properties evaluated for plywood, considering all treatments. For Pearson's correlation, the signi cance levels of p < 0.01 and p < 0.05 were adopted; the analysis was performed in the R software.

Results And Discussion
Physical properties of compensated panels The apparent density and equilibrium humidity of the compensated panels produced from parica sheets, UF and FF pine, and adhesive are presented in Table 2. The use of different adhesives (UF and FF) did not signi cantly affect (p ≤ 0.05) the apparent density for the same lamina composition, that is, the same forest species. On the other hand, independently of the type of adhesive, the apparent density was statistically higher for the compensated panels produced from the pine slats. Knowing that the production conditions were the same for the treatments, it is inferred that the difference observed in this physical property of the compensated panels is mainly related to the basic density of the different forest species used.
According to Silva et al. (2016), the Amazonian species parica presents a basic density between 0.31 and 0.35 g/cm³. The forest species Pinus oocarpa has a basic density close to 0.45 g/cm³ . Thus, the basic density of pine is at least 1.28 times higher than that of parica, and therefore, this characteristic in uenced the apparent density of the compensated panels. The apparent density values of the panels are close to those found in the literature for Pinus oocarpa panels (0.54 g/cm³) (Lisboa et al. 2016) and for parica panels (0.40 g/cm³) (Machado et al. 2018).
In relation to the equilibrium humidity, only the panels produced with parica sheets and FF adhesive (PAFF) differed statistically from the other treatments. However, all treatments were below the 12% limit suggested by the Brazilian Association of the Processed Wood Industry -ABIMCI (Associação Brasileira da Indústria de Madeira Processada Mecanicamente -ABIMCI 2007). As plywood is commonly used as a building material for outdoor applications, understanding its behavior in relation to moisture is of extreme importance (Windt et al. 2018) because the moisture absorbed by the plywood can compromise the mechanical and physical properties and the durability of the material.
The average values for water absorption did not differ and ranged between 62.43 and 82.84%. Regarding the results presented in the literature, for water absorption of plywood, Lisboa et al. (2016) worked with native species of the Brazilian Cerrado and with the species Pinus oocarpa and reported values for water absorption, ranging from 27.43 to 71.58%, being the highest value found for pine. Costa and Menezzi (2017) reported water absorption values of 59.58% for plywood produced with parica blades, whereas Machado et al. Machado et al. (2018) found values of 86.30%.
The results of the breakdown of the species x adhesive interaction for each physical characteristic of the panels are presented in Table 3. For moisture, apparent density, and water absorption, the most in uential parameter was the wood species. For group I (PAFF and PIFF), both humidity and apparent density were higher for the panels produced with FF (PIFF) glued pine blade. The apparent density of the panels produced with pine blade, but with UF adhesive (PIUF, group II), was also statistically higher than that for the PAUF panels. In contrast, the PAUF treatment absorbed more water after 24 hours of testing compared to the PIUF panels (group II).
The interaction of UF adhesive with the different wood species (parica and pine) showed that the density and anatomical structure of the species affected the water absorption of the panels (group II). The number of voids in the glue line increases with the increase of the pot size of the forest species used, causing greater water absorption. Bekhta et al. (2020) showed that water absorption is related to panel density, where higher density results in a lower number of pores and, consequently, lower water absorption. This is in accordance with the results observed for the composite PIUF, which presented a density of 0.55 g/cm³ and lower water absorption (62.43%) compared to the panel PAUF, which has a density of 11.58 g/cm³ and higher water absorption (82.84%).
The species of wood used in the production of plywood has a signi cant impact on the dynamics of moisture, since the wood contains a peculiar anatomical structure and several substances that may have a hydrophobic character and that, in turn, will in uence the wetting capacity (Aydin et al. 2006;Brischke et al. 2013). Therefore, to enhance the properties of the plywood panel in addition to the choice of a waterresistant adhesive, it is also essential to choose forest species that have low absorption and high-water desorption.
The results presented for the analysis of the adhesive unfolding indicated that the type of adhesive in uenced only the humidity of the panels (group III), indicating that the use of UF adhesive on the composites produced with parica blade was higher (11.58%) in relation to the PAFF panels (10.10). Therefore, possibly, the glue barrier formed by the FF adhesive affected the humidity dynamics of the parica compensated panels; however, this effect was not evidenced in the water absorption test, in which the results were statistically equal. The FF adhesive is commonly used in panels destined for external applications due to its excellent resistance to humidity. No statistical difference was observed in the physical properties of the pine plywood (group IV).

Static bending
The results obtained through the static bending test in the parallel direction of the bers showed that the rupture modulus (MOR) was statistically similar among the treatments (Fig. 1). On the other hand, the modulus of elasticity (MOE) was higher for the compensated panels produced with pine sheets. The PIUF treatment presented a percentage increase, for the MOE, of approximately 105.44%, in relation to the PAUF panels. The panels made with pine and FF adhesive (PIFF) had a perceptual increase of 169.95%, in comparison to the parica and FF adhesive (PAFF) panels.
Regarding the results presented in the literature, the mean values of MOR in the parallel direction obtained in this study, for the parica panels, were similar to the values reported by Iwakiri et al. (2011), which ranged from 21.20 to 33.20 MPa for parica plywood glued with UF and FF adhesive, respectively. In contrast, the MOE values are below the average values in the range of 3.44 to 5.28 found by Iwakiri et al. for parica composites with UF and FF adhesive, respectively. In the literature, results close to those found The results of the MOR and MOE show that the plywood panels tested in the parallel direction the bers were considerably larger than those of the samples tested perpendicular to the bers. This behavior was consistent with the ndings of Auriga et al. 2020). Although the values of MOR and MOE in the parallel direction were higher than those of MOR and MOE in the perpendicular direction, the results did not meet the requirements of the NBR 31.000.001/2:2001 standard, which establishes values of MOR for concreteshaped plywood (FOR) of at least 45 MPa. For the parallel MOE, the standard establishes a minimum value of 5.00 GPa, and therefore, only the multilaminated PIFF plywood met the requirements.
The results of the modulus of rupture (MOR) and the modulus of elasticity (MOE), determined by means of the static bending test in the perpendicular direction of the bers, are presented in Figure 2. The resistance to perpendicular exion (MOR) was signi cantly higher for the plywood produced with pine sheets, irrespective off the type of adhesive. The PIUF treatment presented a percentage increase, for the MOR, of approximately 225.97% in relation to the PAUF panels. The plywood produced with pine and FF adhesive (PIFF) had a percentage increase of 45.56% in comparison to the plywood with parica sheets and FF adhesive (PAFF).
According to Kretschmann (2010), the bending resistance and the elasticity modulus of the plywood panel are affected by several factors, such as forest species, wood quality and humidity, density, number of layers, veneer thickness, and binder adhesive. Taking into account that the panels were produced with the same glue grammage, the same layer number and thickness of sheets, and under the same conditions of pressure and temperature, it can be inferred that the density in uenced the results of perpendicular bending. This relation was expressly linear for the perpendicular elasticity modulus, in which the PIFF panel presented higher density (0.58 g/cm³) and higher MOE (4.24 MPa) values. The PIFF treatment presented a percentage increase of 731.37, 311.65, and 85.15% in relation to the PAUF, PAFF, and PIFF composites, respectively.
Regarding the values presented in the literature for the perpendicular MOR, the results found by Machado et al. (2018) were higher for parica panels, with 24.70 MPa. The perpendicular MOE results obtained for parica plywood were also lower compared to parica panels with FF and UF adhesive produced by Iwakiri et al. (2011), whose minimum average values were 1.18 and 1.23 MPa, respectively. The result obtained for the PIFF treatment for the perpendicular MOE was well above those found for Pinus taeda plywood, i.e., 2.24 MPa (Mendes et al. 2013). Reis et al. (2019) studied the mechanical properties of plywood produced with Pinus oocarpa and found a value higher than that observed in this study for perpendicular MOR, namely 31.5 MPa.
Although the values of MOR were better for the plywood produced with pine blades, with the exception of the PAUF, all other treatments exceeded the values required by the NBR 31.000.001/2 (2001) standard, which establishes a minimum value of 15.59 MPa. However, only the PIFF plywood reached the required values for the perpendicular MOE, at least 2.50 GPa.
The results presented for the breakdown analysis of the species and adhesive indicated that only the species in uenced mechanical properties (Table 4). Therefore, the type of adhesive (UF and FF) did not in uence the results obtained for the mechanical properties of the panels produced with the same composition of blades (group III and IV). Analyzing the panels produced with parica blades (PAFF) and pine blades (PIFF) and glued with FF adhesive, we observed that the parallel and perpendicular MOE was higher for the PIFF treatment (group I). On the other hand, there was no statistical difference for the MOE.
For group II (PAUF and PIUF), only the perpendicular MOR showed statistical difference, so that the PIUF treatment obtained a higher rupture module (26.11 MPa). These differences can be attributed to the lower apparent density of parica composites (0.39 to 0.41 g/cm³) in relation to Pinus oocarpa plywood (0.55 to 0.58 g/cm³).

Glue line resistance to shear stress
The resistance of the glue line of the panels produced with the different species and adhesives was evaluated through shear tests in dry, wet, and post-boiling conditions. As shown in Table 5, the highest values of shear resistance to dry, wet, and post boiling conditions were obtained for the plywood panels produced with pine blade, independently of the type of adhesive. This can be explained by the higher basic density values of the Pinus oocarpa species compared to parica wood. Demirkir et al. (2013) proved that shear strength is higher in species with higher densities.
The values found in this study for the resistance of the glue line of the pine panels produced with UF adhesive were higher than those found in the study of Iwakiri et al. (2001), who worked with different species of pine and found values ranging from 1.19 to 1.74 MPa for the dry condition and values between 0.59 to 1.29 MPa for the wet condition. However, for shear resistance under dry conditions, the panels with FF adhesive presented lower values in relation to the results mentioned in the work of Iwakiri et al. (2001), who report values from 2.29 to 3.50 MPa. The resistance of the glue line after boiling, in this study, was higher than that reported by Iwakiri et al. (2009), who worked with ve species of tropical pine and found values ranging from 0.88 to 1.42 MPa.
Although the results were better for the plywood made with the pine slats, all samples met the minimum requirements of the European Standard EN 314-2 (CEN, 1993) (1993) for outdoor panels, with reference values from 0.6 to 1.0 MPa. Therefore, multilaminated plywood produced with a pine and parica blade can be indicated for both internal and external use.
The shear strength values of plywood panels depend more on the type of species than on the type of adhesive applied (Table 6). The breakdown analysis of species and adhesive showed that the type of adhesive (UF and FF) did not in uence the results obtained for the shear in dry, wet, and post-boiling conditions, for the same blade composition (groups III and IV). On the other hand, the type of forest species in uenced the shear strength in dry, wet, and post-boiling conditions. For group I (PAFF and PIFF), only wet shear resistance differed statistically, in which the PIFF treatment showed higher resistance (2.11 MPa). Observing the panels produced with parica sheets (PAUF) and pine sheets (PIUF) and glued with UF adhesive, shear strength in dry, wet, and post-boiling conditions was higher for the PIUF treatment (group III).

Pearson´s correlation for physical and mechanical properties
Pearson's correlation analysis was applied to correlate the mechanical properties with each other and with the apparent density of composites (Fig. 3). The results indicate that density correlated signi cantly with all mechanical properties. The correlations found between the apparent density and the modulus of elasticity in the parallel and perpendicular directions were strong and positive, while the other mechanical properties presented moderate positive correlations. This indicates that mechanical properties tend to increase with increasing apparent density. Density is commonly considered one of the most important characteristics of the material, as it is strongly correlated with the mechanical resistance of wood (Kretschmann 2010).
On the other hand, the correlation between apparent density and water absorption (-0.50) was weakly negative, i.e., the increase in density caused a reduction in water absorption. This fact was con rmed by the Scott-Knott test among the PIUF compensated panels, which presented higher density (0.55 g/cm³) and lower water absorption (62.43%) compared to the PAUF panel with a density of 11.58 g/cm³ and obtained higher water absorption (82.84%). Lisboa et al. (2016) reported a relationship between the apparent density of plywood and water absorption, and in this sense, the increase in density promotes the reduction of empty spaces, impeding/limiting water entry into the wood structure.

Conclusions
The forest species in uenced the apparent density of plywood and the dynamics of water absorption. In general, among the treatments, the pine panels presented higher apparent density and the breakdown analysis of the species showed that the PIUF treatment (group I) absorbed less water compared to the PAUF treatment (group II). These differences were mainly attributed to the basic density of the forest species. The unfolding analysis for the adhesive showed that the parica panels glued with FF adhesive absorbed less moisture compared to those glued with UF adhesive.
No statistical difference was observed for the MOR in the parallel direction, and the values did not meet the minimum requirements stated by NBR 31.000.001/2:2001. The pine panels obtained statistically higher values for the parallel MOE and for the perpendicular MOR. However, in relation to the minimum values required by the NBR 31.000.001/2:2001 standard, only the PIFF treatment met the minimum values of parallel MOE, whereas the minimum requirements for perpendicular MOE, with the exception of the PAUF treatment, were met by all other treatments. The PIFF treatment obtained a higher perpendicular MOE and was the only one that met the requirements of the NBR 31.000.001/2:2001.
Analysis of the breakdown of mechanical properties showed that only the forest species interfered signi cantly with the parallel MOE and the perpendicular MOE and MOR; the values were statistically higher for the pine plywood.
The basic density of the forest species statistically in uenced the mechanical resistance to shear; the composites produced with pine showed higher values. However, all treatments met the minimum requirements stated in the EN 314-2 standard. Unfolding analysis con rmed that the shear strength of composites depended more on the type of forest species than on the adhesive used.
Pearson's correlation analysis con rmed that the apparent density of the panels is a relevant characteristic as it positively in uences the mechanical properties and negatively impacts the physical properties, such as water absorption.   Averages followed by different upper-or lower-case letters within the same group differ statistically from each other by Scott-Knott's (1974) test, at 5% signi cance. Ns: no signi cant.