Mechanical Properties

The mean values of modulus of rupture (MOR) ranged from 1.28 MPa to 7.35 MPa. The adhesive that presented the lowest values was TPF 80:20 (Table 4). According to CS 236-66 (Commercial Standard ^{[CS], 1968}), the minimum required value is 5.6 MPa; only three adhesives types reached this value which were UF and PF at 120 °C and UF at 140 °C. This property, overall, showed best results at the temperature of 120 °C.

In the same table are the values for the modulus of elasticity (MOE), which ranged from 186.21 MPa to 931.71 MPa. Again, the adhesive that presented the lowest value was the TPF 80:20. None of the adhesives tested for this property met the minimum value required by CS 236-66 (1968), which is 1050 MPa. The best value observed in this study for MOE was 100% UF adhesive at 120 °C (931.71 MPa), but still 11.27% below the established by the standard.

The temperature influenced the results on all adhesives used. The addition of tannin to urea-formaldehyde showed no statistical difference in the properties (MOR and MOE) at both temperatures; different behavior was observed for the addition of tannin to phenol-formaldehyde.

^{Gonçalves, Lelis and Oliveira (2008)}, when studying agglomerated panels of Eucalyptus urophylla at different levels of Acacia mearnsii tannin with urea-formaldehyde adhesive, found mean values for MOR of 13.53 MPa (100% UF) and 12.94 MPa (90% UF + 10% tannin). They used a pressure of 2.95 MPa and a temperature of 170 °C with particle moisture of 16%. For the present study the MOR for the 100% UF adhesive was 7.35 MPa (120 °C), this value being 54.32% smaller than that found by the cited authors. For the adhesive 90% UF + 10% tannin, the same authors obtained a result 35.09% higher than the present study.

For both properties (MOR and MOE), the 100% UF adhesive and 100% PF at 120 °C showed the best results. This can be explained by the fact that the particles used in the production of agglomerated panels presented acid pH (4.86), facilitating the curing of the urea-formaldehyde resin that occurs in the same pH range. In the panels with addition of tannin (20%) the properties were impaired, at both temperatures.

The uncontrolled presence of bark in the particles contributed to the static bending values not reaching CS 236-66 (1968), being respectively 10.98 MPa and 12.35 MPa for MOR (UF and FF) and 2,401 MPa and 1,715 MPa for MOE (UF and FF). ^{Nemli, Gezer, Yildiz, Temiz and Aydin (2006)}, when evaluating the mechanical properties of agglomerated panels produced with Alnus glutinosa wood impregnated with Pinus brutia bark extract, concluded that larger bark quantities promoted reduced modulus of rupture and elasticity; however, for those panels manufactured with 1% of bark of Pinus brutia the minimum requirements of standard CS 236-66 (^{CS, 1968}) for MOR were met.

The influence of particle size distribution on the production of agglomerated panels of Pinus taeda was studied by ^{Peixoto and Brito (2000)}, who observed that the best results for MOR and MOE were obtained with adhesive based on urea-formaldehyde and with particles of smaller dimensions. These results are similar to those found in the present study that showed better averages in the urea-formaldehyde adhesive for both MOR and MOE. The size and shape of the particles in agglomerated panels influence the physical and mechanical properties (^{Li, Cai, Winandy, & Basta, 2010}).

The mean values of internal bonding ranged from 0.26 MPa to 2.11 MPa, highlighting the 100% UF adhesive at 120 °C (Table 5). All the tested conditions complied with the minimum required - 0.42 MPa (^{CS, 1968}), except for the use of 20% of tannin with PF at both temperatures. ^{Teodoro (2008)} found internal bonding values higher than the standard even when tannin was added in the PF adhesive composition. However, in this work, the increase of tannin in the commercial adhesives provided a reduction of this property. ^{Melo et al. (2010b)} also found lower values than this study.

The use of processed edgings with conventional particles in panels with 50% conventional particle mixed with 50% of edgings glued with 6% of urea-formaldehyde adhesive showed better results - 0.459 MPa, than those found in a study by ^{Brito, Sampaio, Oliveira and Batista (2006)}. ^{Hillig, Haselein and Santini (2002)} found results similar to those for wood panels of the same species and density, with mean values between 0.185 MPa and 0.286 MPa. Similar results were found in the present study.

The mechanical strength of the panels can be compromised if the adhesive distribution is not adequate, which in turn is also associated with the shape of the particles. Therefore, since the particles used were not produced for this purpose, the quality of the panels was impaired. In this sense, by using a composition with virgin particles and a particle size control (mainly thickness), a better mechanical performance of the panels could have been achieved.

Mean values of screw withdrawal at the surface ranged from 283.83 N to 795.33 N (Table 6). The value required by the ANSI A-208.1 (American National Standards Institute [ANSI], 1999) for this property is 400 N; only the 80:20 TPF adhesive at 120 °C did not reach the established. In general, this result is very important for the manufacture of furniture, since screws are used to join the pieces.

^{Tinti, Gonçalves, Batista, and Vidaurre (2012)}, who performed screw withdrawal on the Oriented Strand Board (OSB) panel surface at the edge and center, found 1857.10 N and 1676.95 N, respectively. These values are higher than the ones found in the present study, corresponding to an OSB panel classified as medium density, of structural character with the particles following a specific index of slenderness, unlike this work.

For screw withdrawal at the top, mean values ranged from 225.83 N to 1034.8 N. According to NBR 14810-2 (^{ABNT, 2006}), the minimum value for this property is 800 N. Some panels did not produce a satisfactory result in accordance with this standard, except UF 100% (120 °C); 100% PF (120 °C and 140 °C); TUF 90:10 (140 °C) and TPF 90:10 (140 °C).

Some possible explanations for the screw removal at the top not being satisfactory, could be due to its particle size, since these do not have an ideal slenderness index, i.e., close to 200 (^{Maloney, 1993}). Also, the uncontrolled presence of bark in the particles, contributed negatively to the adhesion between the substrate and the adhesive impeding it to occur satisfactorily, making the panel with weak bonds between the particles in this sense.

The slenderness index is an important parameter and exerts influence on the characteristics of the panel because it is directly related to the contact area between wood particles and adhesive, influencing the mechanical properties of the adhesive (^{Iwakiri, Keinert Júnior, & Mendes, 2005}; ^{Saldanha, 2004}). Therefore, greater particle size control may provide improvements in mechanical performance on the panels.

Regarding the degradation by Postia placenta fungus, all the panels were classified as highly resistant (Fig. 1), except the 100% TF adhesive which presented moderate resistance, indicating that the TF adhesive has low fungal-resistant chemical properties and requires more studies.

*Means followed by lowercase letter, at the same temperature, and upper case in the same color pattern, do not differ statistically (p ≤ 0.05) by the Tukey test. The bar is the standard deviation.

Figure 1 Mean values of mass loss of the agglomerated panels due to Postia placenta fungus in the tested adhesives and temperatures. 

Regarding the fungus Trametes versicolor the 90:10 TFF adhesive was considered highly resistant; the 100% FF and 100% TF adhesives were resistant and the UF 100% adhesives; TUF 90:10; TUF 80:20 and TFF 80:20 were considered moderately resistant (Fig. 2), according to ASTM D-2017 (2005b).

* Means followed by the same letter do not differ statistically (p ≤ 0.05) by the Tukey test. # Mean mass loss value for each adhesive.

Figure 2 Mean values of mass loss of agglomerated panels due to Trametes versicolor fungus for each tested adhesive. 

Resistance to white rot fungus (Trametes versicolor) in agglomerated panels of Eucalyptus grandis wood particles bonded with urea-formaldehyde and tannin- formaldehyde showed no significant difference, but on average the panels produced with tannin-based adhesive were the least resistant (^{Melo et al., 2010a}). This behavior is similar to that found in the present work, since the adhesives that lost more mass were the combinations of urea-formaldehyde with tannin (in proportions of 10% and 20%).

In agglomerated panels produced with acacia wood glued with urea-formaldehyde-based adhesive and proportions of tannins (10, 20 and 30%) in the presence of all adhesives, the classification was mentioned as resistant to white rot fungus (Trametes versicolor) (^{Gonçalves, Brocco, Paes, Loiola, & Lelis, 2014}). The same authors also observed that there is a tendency of improvement in the resistance of the panels with the increase of the compaction ratio, but unrelated to the increase of the proportion of tannins in the urea-formaldehyde.

Resistance to Cryptotermes brevis termites

After 45 days, a significant mortality rate of termites was observed in all treatments. Moi (1980), ^{Kartal and Green III (2003)} state that this may be associated with the presence of extractives, resins and ashes (silicates), mainly in the bark. In wood with higher ash contents, ^{Paes, Medeiros Neto, Lima, Freitas, and Diniz (2013)} found faster death of dry wood termites.

The adhesive that lost more mass was 80:20 TPF with 1.51% loss, and the one that lost less mass was the 100% PF with 0.16% (Fig. 3). Thus, the amount of tannin present in the adhesive contributed to a decrease in the resistance of the panel to the dry wood termite in the phenol-formaldehyde adhesive.

* Means followed by the same letter do not differ statistically (p > 0.05) by the Tukey test. # Average mass loss value for each adhesive.

Figure 3 Mean values of mass loss of the agglomerated panels to the dry wood termite. 

For the UF, PF and TF (100%) adhesives, with 20 days of testing, mortality reached 100%; despite this, the 100% TF adhesive was the second most attacked, with 1.34% mass loss.

^{Ribeiro et al. (2014)} evaluated commercial agglomerates of eucalyptus, eucalyptus + pinus and sugarcane bagasse on dry wood termite attack (Cryptotermes brevis), and found a greater loss of mass in the eucalyptus panels (2.78%) and lower loss in the panel of sugarcane bagasse (0.19%). The latter was less attacked due to chemical components present in sugarcane bagasse, since this residue may present mineral components, such as silica in large quantities, as observed by ^{Belini, Tomazello Filho, Mendes, Leite and Lima (2012)}.