THERMAL MODIFICATION OF SUGARCANE WASTE AND BAMBOO PARTICLES FOR THE MANUFACTURE OF PARTICLEBOARDS

ABSTRACT The thermal modification of particles of the paticleboards constituted of agroforest and industrial waste can improves the dimensional stability (thickness variation) and reduces the use of chemicals that can raise the costs of the process or be hazardous to humans and the environment. This study evaluated the effect of the thermal modification on the physical-mechanical properties and density profile of particleboards manufactured from sugarcane bagasse and bamboo (Dendrocalamus asper) (Schult f.) Backer ex Heyne). A mixture of 75% bamboo particles and 25% sugarcane bagasse was subjected to 220 °C temperature for 201 min. Urea-formaldehyde (UF)-based adhesive with three solids contents (10, 12 and 14%) based on the dry mass of the particles was used for the aggregation of the materials. Both temperature and increases in the adhesive content improved their dimensional stability, however, the thermal treatment reduced the mechanical properties. The particleboards composed of treated particles did not meet the minimum specifications established by the Brazilian norm utilized. The densitometric profiles were negatively influenced by the thermal modification and improved by the increase in adhesive content.


INTRODUCTION
The search for ecologically correct and goodquality alternative materials is one of the objectives of the sustainable development for a more rational use of natural resources (Valarelli et al., 2013).Bamboo is among such resources; it includes 34 genera and 232 species, of which 174 are endemic in Brazil, and evidences its diversity, although little explored (Arruda et al., 2011).Some species are employed as building elements due to their strength, fl exibility and versatility (Flander and Rovers, 2009), and raw material for the manufacture of particleboards (Almeida et al., 2017;Dinhane et al., 2015;Melo et al., 2015).
Agroforest and industrial waste are also used for the manufacture of particleboards, since they not only add value to the product, but also provide sustainability, reduce environmental pollution problems and help particleboard industries meet their demands (Negrão et al., 2014).Some studies have pointed to the use of sugarcane bagasse for the manufacture of particleboards (Fiorelli et al., 2013(Fiorelli et al., , 2016;;Mendes et al., 2014).However, one of the technological obstacles for its use in reconstituted particleboards is its dimensional instability in comparison to particleboards manufactured from wood particles (Freire et al., 2011).Such instability is due to the low sugarcane bagasse density, which results in a thick mattress and compression strain.On the other hand, mattresses from particles of wood or another denser material (bamboo) are less thick and can result in a low compression ratio.Therefore, the combination of materials of diff erent densities might be an option for larger availability of raw material for the industry of reconstituted particleboards.
Moreover, the thermal modifi cation of raw material might improve some technological properties of particleboards manufactured from agroforest and industrial waste, as dimensional stability, and reduce the use of chemicals that might raise the costs of the process (paraffi n) The time and temperature adopted for the thermal modifi cation process can improve the dimensional stability, strength to weather changes and biodeteriorating agents, wettability and bonding with hydrophobic adhesives, and reduce the equilibrium moisture content and swell the thickness of the particleboards (Brito, 2018).According to some studies, the thermal treatment of wood particles improved the physical properties of the particleboards (Paul et al., 2007;Mendes et al., 2013;Vital et al., 2014).This article addresses an evaluation of the thermal modifi cation eff ect on the physical-mechanical properties and density profi le of particleboards manufactured from sugarcane bagasse and bamboo (Dendrocalamus asper) (Schult f.) Backer ex Heyne).

Sugarcane waste (sugarcane bagasse) and bamboo particles
Sugarcane bagasse was obtained at a sugar mill located in Santa Bárbara D'Oeste, São Paulo state, Brazil.The collected waste showed an uniform color and no deterioration.They were transported to the Forest Sciences Department, Luiz de Queiroz College of Agriculture, University of São Paulo, Piracicaba, São Paulo state, and air-dried (18%) and dried in a greenhouse (70 ± 2 °C) for 3 h, until reaching (10%).The material (sugarcane bagasse) was classifi ed and 0.50 -0.85 mm granulometry particles were used.
Over three-year-old bamboo (Dendrocalamus asper stems showed good mechanical strength properties and availability at the Agronomics Institute of Campinas, Tatuí, São Paulo state.They were transformed into splinters, according to the procedure described by Brito et al. (2018), which were planned for the removal of the internal layer (starch-rich) and the external one (bark) and transformed into chips by a band saw.Then, they were dried and processed similarly to sugarcane bagasse.

Thermal modifi cation of the particles
The sugarcane bagasse and bamboo particles were dried in a greenhouse (3% humidity) and placed on hollow metallic trays (13 x 18 x 58 cm), which were deposited in rectangular metallic boxes in a greenhouse provided with a sensor that controlled the time and temperature of the thermal treatment.Nitrogen was injected into the boxes for the avoidance of ignition of the material.
The heat treatment parameters were defi ned based on the work of Mendes et al. (2013).The thermal modifi cation of the particles started at ± 30 °C and the heating rate was 3.33ºC/ min -1 for 21 min, until reaching 100 ºC.Then it was reduced to 1 °C min -1 until reaching 220 °C (141 min) and kept for 201 min.for an eff ective treatment.The greenhouse was then switched off and the material remained there overnight (± 12 h).After the treatment, the particles remained in the boxes until reaching ambient temperature (± 30 °C).

Manufacture of sugarcane bagasse and bamboo particleboards
The manufacture of the particleboards followed the recommendations of (Surdi et al. (2018) and the material showed 0.65 g.cm -3 nominal density, 15.70 mm nominal thickness and 10; 12; and 14% solids content of the urea-formaldehyde (UF)-based adhesive in relation to the mass of the particles.The percentages of solids (10, 12 and 14%) were defi ned based on those used by the industries for the production of agglomerated panels.The adhesive showed 64.16% solids content, 1.27 g.cm -3 density and 7.88 pH and received an ammonia sulphate solution (catalyst) at 5% solids proportion.The mixture was homogenized and sprinkled on the particles in a slasher of warp (12 rpm) for 5 min and a paraffi n emulsion (1.0% solids) was applied (5 min -12 rpm).
The particles were weighed and placed into a wooden hollow mold of 40 x 40 cm on an aluminum plate (50 x 50 cm).The mattress was cold preconsolidated (5 kgf.cm -2 =pressure for 5 min), removed from the box and hot-pressed (180 ºC; 35 kgf.cm -2 ; 10 min).A 1.57 cm thick delimiter was used, which were arranged on the sides of the prepressed mattress, ensuring that the applied pressure was equally distributed across the panel area.It was then arranged in a vertical position for cooling and acclimatized (22 ± 2 ºC and 65 ± 5% relative humidity -RH) prior to the removal of the samples for physicalmechanical and density profi le tests, according to Regulatory Brazilian Norm -NBR 14810 of the Brazilian Association of Technical Norms -ABNT (2013).

Density profi le
The density profi le was determined according to Surdi et al. (2014) and the eff ect of variables checked on the pressing process in the compaction of the particleboards.This study contributes to the quality control of both experimental and industrial particleboards (Wang et al., 2004).
Four samples of 15.70 x 50 x 50 mm (thickness x width x length) were removed from each particleboard, which totaled 12 per treatment, and sandpapered (# 80 sandpaper) prior to the tests.Since the density profi le analysis is non-destructive, the samples were used for internal bond (IB) strength tests.
The analysis was conducted in a QMS QDP-01X model X-ray densitometer of 10-50 KV, 1.5 mA current and initial and fi nal collimation of 180 and 90 µm beams, respectively.
The sweeping of the samples changed the X-ray beams into density values obtained every 20 µm through QMS program; a DAT-format fi le was created and read by Excel software, which enabled the construction of a graph of maximum and minimum apparent density profi le of the samples.Three reading points, i.e., two for the external layer and one for the central layer (B), were established after the obtaining of the density profi le.A 2 mm band from each end to the particleboard´s thickness direction was considered for the density measurement of points A. The average values between points A was considered for the central layer.

Experimental delineation and data analysis
The particles were subjected to 220 °C temperature and three UF-based adhesive levels were adopted (10; 12 and 14%).The particleboards were manufactured with 75% bamboo particles and 25% sugarcane bagasse were defi ned based on the best compaction ratio (relation between the specifi c masses of the panel and the raw material), among some compositions tested.Those manufactured with non-treated particles glued with 10% adhesive were taken as control.
The eff ects of both thermal modifi cation on the equilibrium moisture content of the particles and adhesive contents and thermal treatment on the properties of the particleboards were tested by variance analyses and F tests (p < 0.05).Tukey test (p < 0.05) checked the discrimination of the averages, whereas Lilliefors and Cochran tests evaluated the normality of the data and homogeneity of the variances, respectively.
The results of the physical and mechanical tests were confronted with the values established by NBR 14810 (ABNT, 2013).The humidity content of the particleboards was analyzed by descriptive statistics that considered averages and standard deviations.

Equilibrium moisture content and alteration in the color of the particles
The temperature used reduced the hygroscopicity of the particles and the equilibrium moisture contents of the control particles were larger in relation to those thermally modifi ed (Table 1).

Physical properties of the particleboards
The results of the physical properties of particleboards were calculated (Table 2).The treatments showed diff erences for all evaluated properties, except for apparent density (AD).

Mechanical properties of the manufactured particleboards
The results were obtained for the mechanical properties of the particleboards (Table 3).The treatments showed diff erences regarding modules of rupture (MOR) and modules of elasticity (MOE), surface screw withdrawal (SSW) and top screw withdrawal (TSW) and internal bond (IB).

Density profi le of the particleboards
The treatments showed diff erences regarding the densities of the external and central layers (Table 4).
The density profi les of the particleboards were obtained by the X-ray attenuation for each treatment (Figure 1).5.17 Means followed by the same letter do not diff er (Tukey; p > 0.05).
The thermal treatment modifi es the structure of the material and reduces the equilibrium moisture content through the degradation of the chemical components and formation of a lignin crossed bond, which aff ects the water adsorption (Surini et al., 2012).The decrease in the equilibrium moisture content enables the manufacture of particleboards of higher dimensional stability, however, the degradation of the OH-(hydrophilic) groupings may hamper the action of the adhesive during the consolidation of the particleboards (Vital et al., 2014).
The particles showed a darker coloration after the thermal treatment due to possible alterations in their structure and caused by the chemical modifi cation of the chromophore groups responsible for the characterization of color (Ahajji et al., 2009).

Physical properties of the particleboards
No diff erences were detected among the treatments regarding apparent specifi c density, however, it was reduced in relation to the preestablished one (0.65 g.cm -3 ), which usually occurs due to losses of materials during the manufacture stages (Iwakiri et al., 2012).
The increase in the adhesive content does not increase the specifi c mass of the particleboards (Santos et al., 2009;Colli et al., 2010;Bianche et al., 2012), which can be included in the average-density category, since they met the requirements regarding humidity content for commercialization, i.e., between 5 and 11%, according to NBR 14810 (ABNT, 2013).
Particleboards composed of control particles showed higher values for Water Absorption (WA, 2 hours and WA, 24 hours).Treatments T1 and T2 were comprised of particles glued with the same adhesive content, however, those of T2 were constituted by thermally modifi ed particles.The thermal treatment reduced the water absorption, i.e., 70.06% (WA2h) and 20.26% (WA 24 h).Mendes et al. (2013) observed a similar eff ect, i.e 25% (WA 2 h) and 36% (WA 24 h) reductions in oriented strand boards (OSB) manufactured with thermally modifi ed particles (200 and 240 °C), which proved the thermal treatment reduces the water absorption by the particles.
Particleboards composed of thermally modifi ed particles and glued with 10% adhesive showed higher water absorption in comparison to those for which 12 and 14% adhesive was used.The increase in the adhesive contents reduced 46.70% (WA 2 h) and 41% (WA 24 h) in relation to those composed of 10% adhesive content.
The higher amount of adhesive applied to the particles promotes better water proofi ng of the surfaces and reduces the water absorption rate.On the other hand, treatments T3 and T4 provided similar absorption rates and indicated 12% adhesive would be suffi cient for the improvement desired, hence, a more economical process, since the adhesive can correspond to up to 40% of the fi nal cost of the particleboard (Carneiro et al., 2007).
The increase in the adhesive contents favored the water absorption reduction for particleboards composed of Pinus spp.particles and sugarcane bagasse (Mendes et al., 2012), Eucalyptus saligna waste (Pedrazzi et al., 2006) and candeia wood associated with pinus and eucalipto (Santos et al., 2009) The absorption decrease in function of the increase in the adhesive content is due to the distribution of the same of the particles that have better bond quality (Mendes et al., 2012).
The treatments showed diff erences regarding thickness swelling (TS).The thermal modifi cation (T2) reduced 70.11%(TS 2 h) and 8.57% (TS 24 h) of the thickness in comparison to the control particleboards (T1).Mendes et al. (2013) observed 58.9% (TS 2 h) and 50.0%(TS 24 h) reductions in particleboards composed of strand particles treated at 240 °C in relation to the control particleboards.Particles subjected to thermal modifi cation showed lower swelling due to the decrease in the OH-groupings available for water adsorption in the constituents of the cell wall (Vital et al., 2014).
The increase in the adhesive contents (T3 and T4) reduced 42.15% (2 h) and 62.25% (24 h) of the TS in relation to T2 and promoted the lowest swelling.Pedrazzi et al. (2006), Colli et al. (2010) and Mendes et al. (2012) observed the same result and attributed it to the diffi cult contact between water and particles caused by a better cover of the adhesive that, consequently, delayed and reduced the thickness swelling indexes of the particleboards.
The particleboards met the 18% maximum TS after 24 h water immersion, as established by NBR 14810 (ABNT, 2013) for type 2 particleboards (internal use under dry conditions).
No diff erence was observed for Non-Recoverable Tax NRT between T1 and T2, which leads to the conclusion the thermal treatments of the particles did not infl uence such a property.T3 and T4 provided the best values due to the higher adhesive contents used in the composition of the particleboards, and 44.07 and 62.72% reductions in comparison to T1. Presumably, the higher adhesive content resulted in a better adhesion of the particles and avoided the release of higher compression strains imposed on the pressing process.The increase in the availability of adhesive per superfi cial area of the particles improves their bond, hence, properties, as TS 24 h and NRT (Mendes et al., 2003).
The increase in the adhesive content increased the values of all properties evaluated -in some cases, the values were similar to those of the control particleboards.Therefore, the increase in the adhesive content compensated for the reduction in the strength caused by the thermal modifi cation and improved the adhesion process probably due to the better distribution of the resins in the particles.
Such results confi rmed those obtained by Pedrazzi et al. (2006), Colli et al. (2010) and Ayrilmis et al. (2012), who observed improvements in the strength properties of the particleboards due to the increase in the adhesive content.The increased availability of the resin promotes higher adhesion, hence, better gluing quality (Mendes et al., 2003).

Density profi le of the particleboards
T2 decreased 17.65% and 7.17% the apparent density of such layers in relation to T1, respectively, which shows negative infl uences of the thermal modifi cation on the density profi le of the particleboards.
The increase in the adhesive content increased the specifi c density of the external layer.Since the results of 12 and 14% content were similar, the use of 12% is recommended, due to economic issues (the adhesives may correspond to up to 50% of the process cost).
From the profi les of the analyzed particleboards it is observed that T1 showed more pronounced peaks of apparent specifi c density on the external layer in comparison to T2, whereas the peaks of particleboards composed of 12% (T3) and 14% (T4), were more pronounced in relation to T2.
T4 particleboards showed a denser region on the central layer, of characteristics similar to those of T1, which indicates a higher uniformity of the readings in relation to T2 and T3.Therefore, the increase in the adhesive content promoted a better consolidation in the external and internal layers that improved the performance of the particleboards regarding the mechanical properties evaluated.

CONCLUSIONS
Both thermal modifi cation and increase in the adhesive content improved the WA and TS properties, but reduced the mechanical strength of the particleboards.
Although the adhesive content improved NRT and all mechanical properties evaluated, the particleboards did not meet the requirements of the Brazilian Norm.Similarly, the densitometric profi les were negatively infl uenced by the thermal treatment, which resulted in a reduction in the apparent density.However, they were improved by increase in adhesive content which provided more homogeneous profi les, thus confi rming the improvements in the evaluated mechanical properties.
Equilibruim moisture content and alteration in the color of the particles It was observed that the thermal treatment reduced the equilibrium moisture content of the particles and 35.86 and 12.38% were obtained for sugarcane bagasse and bamboo, respectively.Vital et al. (2014) obtained T1 -Control (10%); T2 (Heat treatment 10%); T3 (Heat treatment 12%); T3 (Heat treatment 14%).*Datas contained inBrito (2018), used as comparison in this article.Means followed by the same letter do not diff er (Tukey; p > 0.05).CV: coeffi cient of variation.

Table 1 -
Equilibrium moisture content of sugarcane residue and bamboo particles.Tabela 1 -Teor de umidade de equilíbrio das partículas de resíduo de cana e bambu.
Datas contained in Brito (2018), used as comparison in this article.Means followed by the same letter do not diff er (Tukey; p > 0.05).CV: coeffi cient of variation. *

Table 4 -
Mean values of the apparent density of the outer layer and central layer of the particleboards according to the treatments and adhesive contents used.Tabela 4 -Valores médios da densidade aparente das camadas externa e central dos painéis de acordo com os tratamentos e teores de adesivo empregados.