Roughness study on homogeneous layer panels manufactured from treated wood waste

Natural resource exploration is growing, highlighting woods and joinery waste, wood industries and the like. This study presents homogeneous particleboard (PPH) roughness characterization manufactured from treated wood waste. Normative document with values of Brazilian Technical Standard Association ABNT NBR 8404 (1984), was adopted as a reference. The results show that the manufactured PPH showed roughness class N 10, with roughness values (Ra) of less than 12.5 microns.


Introduction
, Maloney (1977) and American National Standards Institute A208.1 (ANSI, 1999) report that particleboard definition depends on particles types, board structure, resin and manufacture equipment.A wide sizes and shapes variety of raw materials and particles may be used.
Particleboard is manufactured from lignocellulosic materials (usually wood) with synthetic resin or other suitable resin and glued together under heat and pressure in a hot press.Others materials may be added during manufacture in order to improve certain properties.The standard Associação Brasileira de Normas Técnicas NBR 14810 (ABNT, 2013) defines medium density particleboard, with density between 551 and 750 kg m -3 , a panel manufactured from wood particles bonded with thermosetting synthetic resin, which are consolidated under action together of heat and pressure.
The standard Associação Brasileira de Normas Técnicas NBR ISO 4287 (ABNT, 2002) defines surface roughness and geometric parameters used in determining the roughness.The obtained profile parameter was the arithmetic mean deviation of the assessed profile, defined as Ra and expressed in microns (μm), according to the standard Associação Brasileira de Normas Técnicas NBR ISO 4287 (ABNT, 2002).Ra is the main parameter of the surface roughness profile.
By the importance of the roughness analyses in particleboards, this study proposed homogeneous particleboard (PPH) manufactured evaluation from different densities and reutilizing treated wood waste to evaluate roughness after circular saw cuttings, it is important to study the roughness of the panels in the end of introduction

Material and methods
Shavings from Eucalyptus sp. and Pinus sp.treated wood were classified as shown in Figure 1.The greater shavings dimension was 2.8 mm.The shavings were mixed with two resins: urea formaldehyde and castor oil based polyurethane.Panels were made with nominal thickness of 10 mm with nominal dimensions of 400 × 400 mm and densities of 0.6, 0.8 and 1.0 g cm -3 .In mixer supplies were homogenized and, at continuation, a mold received mixture which was transferred to hot press (Figures 1 and 2).

Cutting methodology
Panels cutting methodology included as basis Brazilian standard Associação Brasileira de Normas Técnicas NBR 6162 (ABNT, 1989).Circular saws used are described in Table 1; PPHs cutting speed was 58.98 m min -1 ., and forward speed per tooth in saws were Z48 = 0.35 mm; Z96 = 0.17 mm; Z108 = 1.55 mm.From Table 1, TS is the Saw type; D is the Saw diameter, Z is the Sooth number and TD is the Tooth types and angle.Cutting number consisted of 2040 strips (Figure 3), in different densities and resin utilized.Cutting were randomly selected and divided into groups of Pinus sp., Eucalyptus sp. and resins urea formaldehyde and castor oil based polyurethane.
From selection was measured roughness with Taylor Robsow Surtronic equipment.

Roughness testing
Mean roughness is defined as absolute values arithmetic mean of removal ordered (yi), of roughness profile points in relation to mean line, within measurement pathway (lm).Specimens for this analysis presented 350 mm in length, with a width ranging from 100 to 120 mm.Roughness testings were performed at University of Sao Paulo State, Unesp, Itapeva Campus, as illustrated in Figure 4.
In Table 3 are described roughness values according to Brazilian standard Associação Brasileira de Normas Técnicas NBR 8404 (ABNT, 1984), which were used as results comparison parameters.

Statistical analysis
Factors and experimental levels investigated in obtaining cutting roughness values [Ra (μm)] of homogeneous particleboard manufactured were Wood Species [Specie], varying between Pinus (Pinus) and Eucalyptus (Euca.), and adhesive type [Adhes.],consisting of castor oil based polyurethane resin (PU) and urea formaldehyde (UF).Two levels of two factors combination resulted in a full factorial design with four treatments (Tr), clarified in Table 4.
For each of four experimental treatments delineated were manufactured 8 homogeneous particleboards, being removed two samples from each panel, in a total of 64 specimens tested in roughness cutting values (Ra) obtaining.
In order to investigate isolated factors effect [Specie; Adhes.] and interaction between them [Specie × Adhes] on cutting roughness values was used variance analysis (ANOVA), considered at 5% significance level (α).Equivalency between treatment means was considered as null hypothesis (H 0 ), and no equivalence between means (at least two treatments) as alternative hypothesis (H 1 ).By ANOVA formulation, evaluated by Minitab ® software version 14, P-value lower than significance level (P-value < 0.05) implies to reject null hypothesis (at least one treatments mean differs from the others), accepting it in otherwise (the treatments means are equivalent).To validate Anova were tested normality in cutting roughness values distribution, with help of Anderson-Darling test (AD), variance homogeneity with Bartlett (Bt) and Levene (Le) tests and waste independence by means of waste versus order graph.
Accused significance of isolated factors on cutting roughness values, in sequence was used multiple comparison Tukey test (contrast test between means) enabling to group factor significant levels.From Tukey test, A indicates highest mean group, B is the second highest mean group and so on, same letters implied treatments with equivalent means.In case of significant interaction between factors, interaction between factors graphic was used as an auxiliary way to interpret interaction effects.

Results and discussion
Table 5 shows mean values, variation coefficients (CV) along with smaller (Min.) and with bigger   Rolleri and Roffael (2010) studied the surface roughness of particleboards produced from wood particles and urea formaldehyde resin and obtained Ra roughness values in the range of 5.2 to 11.2 μm.Their roughness results are similar to this study.
Tabarsa, Arshorie, and Gholamzadeh (2011) evaluated the roughness of particleboard made with sugarcane bagasse and urea formaldehyde resin and found roughness Ra of between 10 and 25 μm.The roughness results reported by these authors are superiors to those found in this study.
Varanda, Alves, Gonçalves, and Santiago (2010) evaluated the roughness of Eucalyptus grandis wood pieces submitted to sanding and found an average roughness value of 7.4 μm, after sanding.The results of Varanda et al. (2010) are similar to the results of this study roughness.
Statistical analysis positive results obtained were demonstrated.Figure 5 shows cutting roughness mean values graph from four experimental treatments investigated.Table 6 shows Anova results on panel cutting roughnesses values, where DF represents freedom degrees, Seq SS squares sum, Adj SS the set squares sum, Adj MS set mean square and F Fisher's statistics.As observed in Table 6, the two individual factors were significant on panels cutting roughness values, same did not occur with the interaction between them, considered not significant by Anova.

Conclusion
Results obtained of this research allow concluding that: -individual factors Specie and Adhesive were significant on roughness values of particleboards.The interaction between factors was considered not significant by Anova; -particleboards produced with Pinus sp.particles presented roughness values (5.67 μm) lower than particleboards produced with Eucalyptus sp.particles (6.33 μm); -particleboards produced with castor oil base polyurethane resin presented roughness values (5.59 μm) lower than particleboards produced with ureaformaldehyde resin (6.42 μm); -Particleboards produced with urea-formaldehyde resin showed worse surface finish (Ra roughness greater value), still showed environmental problems in the use of this formaldehyde-based resin; -roughness values of panels manufactured according to four experimental treatments delineated showed roughness class N 10, with roughness values (Ra) of less than 12.5 microns, according to Associação Brasileira de Normas Técnicas NBR 8404 (ABNT, 1984).

Figure 6 presents
Figure 6 presents Anova tests validation results on cutting roughness values.By means of P-values (> 0.05) found for both tests, cutting roughness values have normal distribution, variances homogeneity and waste independence, validating Anova model.

Figure 6 .
Figure 6.Anova validation test results of cutting roughness values: normality test (a), variance homogeneity test (b) and waste independence (c).

Figure 7 .
Figure 7. Graphs of Main effects on roughness values: wood species (a) and adhesive type (b).

Table 1 :
Circular saws for wear testing.

Table 4 .
Experimental treatments delineated in homogeneous particleboard manufacturing.

Table 5 .
Results obtained from cutting roughness from fabricated panels according four delineated experimental treatments.

Table 6 .
Anova results on the cutting roughnesses values.
Table 7 presents tukey's test results of two factors considered significant, where x is the mean and Ag.groups formed.

Table 7 :
Tukey's test results on panels cutting roughness values.