Brazil was the second largest producer of medium density fibreboards (MDF)/high density fibreboards (HDF) in the world in 2021, with production nearing six million cubic metres.
Out of this total, 4.9 million cubic metres were consumed in the Brazilian domestic market, resulting in a 15.7 percent growth compared to the previous year’s consumption.
The primary source of raw material for this product is wood, either in the form of logs or waste generated from other processes. MDF finds its most common application in the furniture industry.
From 1997 to 2002, Brazilian MDF production was characterised by the exclusive use of woods from the Pinus. In 2003, Eucalyptus spp woods were introduced, and they have since shown a growing trend in production participation. Currently, depending on the region of the country, one genus or the other predominates.
MDF production is concentrated in the southern and southeastern Brazil, accounting for approximately 70 percent of the total area planted with Eucalyptus spp. and Pinus spp. in the country.
Geographically, wood-based panel industries are strategically located near furniture centres or in locations with logistics that favour the efficient flow of production to customers.
In 2010, an MDF industry was established in northern Brazil, in the state of Pará, using wood from Schizolobium amazonicum Huber ex Ducke (paricá) and Eucalyptus spp plantations as raw materials.
The introduction of this industry was an innovative initiative and marked the beginning of decentralising MDF production in Brazil, as well as the utilisation of non-traditional raw materials in this industrial segment.
Among these potential raw materials are the wastes generated from the mechanical processing of wood from native Amazonian species, which are abundant in Pará.
The use of waste from processing native wood, from sawmills and veneer factories in the Amazon region, can be an interesting alternative for the particleboard industries.
However, few studies have been conducted on this topic, including research on industrial or laboratory-scale MDF production using residues from the mechanical processing of native Amazonian wood.
Paricá wood has a low density, ranging between 260 and 420 kg/cubic metre. However, most native Amazonian species used in mechanical processing consist of medium to high-density woods.
Low-density woods are preferred for the production of medium-density panels because they result in an appropriate compaction ratio (CR ≈ 1.3), which is determined by the relationship between panel density and wood density.
High-density wood cannot be adequately compressed to achieve the required compaction ratio for medium-density panels without compromising bonding. A partial solution for using high-density woods is to blend them with low-density woods, creating a mixture of raw materials that achieves the appropriate density for proper compaction.
There are numerous studies on the quality of MDF produced from the blending of different types of wood and/or other lignocellulosic materials, seeking alternative raw materials and their suitability for the production process.
In addition to the raw material, various production variables of MDF, such as density and adhesive content, as well as panel thickness (according to NBR 15316-2; ABNT 2009b), influence the product’s properties. However, there are limited studies that specifically address the effect of panel thickness.
The primary aim of the present study was to assess the properties of MDF panels produced using a combination of paricá wood and wood waste from native Amazonian species.
The specific objectives included examining the isolated effects of the blend ratio between raw materials and of the thickness of the panels on the properties of the MDF.
Experimental Design
The experimental design consisted of three independent experiments, each designed to analyse the isolated effect of a production variable on the properties of the MDF panels manufactured in an industrial production line.
The density of the panels was intentionally kept constant across treatments within each experiment because its variation could mask the effects of the variables tested in each experiment on the properties of the panels.
In Experiment I, the targeted density for the panels and panel thickness was kept constant, and the analysed variable was the fibre blend ratio (FBR) used in the panel manufacturing process, tested with three treatments. In Experiments II and III, both density and FBR were kept constant, and different treatments of panel thickness were tested. For each treatment, three panels (replicates) were used.
In this study, we used wood sourced from Schizolobium amazonicum in the form of logs (SAL) and wood veneer waste (SAV), along with a blend of wood waste resulting from the mechanical processing of various Amazonian native species (MIX).
The SAL material originated from eight-year-old plantations, SAV and MIX (lacking species identifications) materials were obtained from sawmills and veneer factories located in the municipalities of Paragominas, Dom Eliseu, and Ulianópolis.
The materials were processed into chips and subsequently into fibres within the facilities of a MDF manufacturing plant. The fibres were then used in the production of MDF panels in the same MDF plant, as detailed below.
MDF Panel Manufacturing
Except for the variables described earlier in the experimental design, including the specific density planned for each experiment, all other production parameters of the MDF panels manufactured in the plant were kept constant, as described below.
The chips from the dosing silo, previously washed, classified, and mixed according to each fibre blend ratio (A, B, C, or D), were submitted to preheating with steam at a temperature of 90 deg C for eight minutes.
Subsequently, the chips were transferred to the digester, where they were exposed to a temperature of 190 deg C and a pressure of 0.85 MPa for five minutes. Then, the wood chips were defibrated in a double-disc refiner under a pressure of 0.95 MPa.
Upon exiting the defibrator, adhesive and paraffin emulsion were applied to the fibres. The adhesive used for fibre bonding was urea-formaldehyde (UF), with a solids content of 66 percent and a viscosity of 350 cP.
The adhesive dosage was 13 percent solids, based on dry fibre mass. The paraffin dosage was 0.7 percent of solids, calculated similarly to the resin.
After receiving the adhesive and paraffin, the fibres were dried in a tube dryer, driven by a forced air current at a temperature ranging from 140 to 150 deg C, achieving an average final moisture content of 10 percent.
The fibre mattress was formed on a mobile mat, pre-pressed, cut, and then taken to a hydraulic press equipped with 12 plates for the final pressing. The press operated at a temperature of 194 deg C, a maximum pressure of 2.1 MPa, and a total time of 189 seconds.
After pressing and cooling, one panel out of the 12 produced at a time was randomly selected, and a 600 x 600 mm sample was extracted from it, which constituted one treatment replication. This process was repeated for each new pressing and cooling cycle until three samples per treatment were collected, which were then used to obtain specimens for conducting the planned tests.
Basic Density
Samples of wood chips from the raw material were collected at the MDF plant and taken to the Mechanical Testing Laboratory of Wood and Derivatives at Universidade de São Paulo - USP, where the basic density of SAL, SAV, and MIX wood chips was determined using the maximum moisture content method.
Based on these data, the basic density of the chip mixtures that formed the fibres was estimated. The estimation procedures accounted for the percentage of each raw material’s contribution to the mixtures.
Test Specimens
The samples of panels (600 x 600 mm) produced in the MDF plant were also taken to USP for the experimental procedures. Test specimens were obtained from each panel sample using a circular saw and then conditioned at 22 ± 2 deg C and 65 ± 5 percent relative humidity.
These specimens were used to evaluate moisture content, density, water absorption (WA) and thickness swelling (TS) after two and 24 hours of immersion, internal bonding (IB), static bending (SB) to determine modulus of rupture (MOR) and modulus of elasticity (MOE), as well as resistance to screw withdrawal (RSW) on the top and face surfaces.
Density and moisture content were evaluated on the same test specimen. A similar process took place to evaluate WA and TS. The dimensions of the test specimens to evaluate SB were as follows: a length equal to 20 times the sample thickness, plus 50 mm, and a fixed width of 50 mm.
The dimensions of the test specimens to evaluate all other properties were 50 x 50 mm and the thicknesses were 12, 15 or 18 mm. For each parameter evaluated, four measurements were taken per panel, totalling 12 measurements, considering that three panels (replicates) were used per treatment.
The dimensions of specimens and testing procedures followed the NBR 15316-3 standard (ABNT 2009c), which is similar to the European standard series for MDF.
Data Analysis
Initially, the normality and homogeneity of variance of the data were checked using the Shapiro-Wilk and Bartlett tests, respectively. The requirements for normality and homogeneity were met, and there was no need to transform the data for any of the analysed variables.
The response variable in each experiment was compared among treatments using analysis of variance (ANOVA).
When the ANOVA was significant, a Tukey test was used for pairwise comparison between the treatments (α = 0.05). The average values of the assessed parameters were compared to the specifications outlined in the NBR 15316-2 standard (ABNT 2009b).
Density
The average basic density for the SAL and SAV chips did not differ statistically, as expected considering they were composed of the same source wood. The density of MIX chips was significantly higher than that of SAL and SAV.
The effective density of the panels within each experiment closely matched the planned density (target density), demonstrating that the panel production process was well-controlled.
There were no significant differences among treatments in the effective density of the panels in any of the experiments. Therefore, the fibre blend ratios (Experiment I) and panel thickness (Experiments II and III) did not have a significant influence on panel density.
Water Absorption
In Experiment I, there was no significant difference in WA and TS (both at 2 and 24 h) of the panels among treatments, indicating that the fibre blend ratio did not influence these properties.
In Experiment II, WA (at 2 and 24 h) and TS (at 24h) differed significantly among the treatments, indicating an effect of panel thickness (12 and 15 mm) on these variables.
In Experiment III, TS (at 2 and 24 h) differed significantly among treatments, reinforcing the effect of panel thickness (15 and 18 mm, in this case) on thickness swelling.
Static Bending, Internal Bonding & Resistance
In Experiment I, there were no significant differences in SB (MOE) and IB of the panels among treatments, indicating that the fibre blend ratio did not influence these properties. However, SB (MOR) and RSW (top and face) of the panels differed significantly among the treatments.
In Experiment II, SB (MOR and MOE) did not differ significantly among the treatments, indicating that there was no influence of panel thickness on static bending. However, IB and RSW (face) differed significantly among treatments.
In Experiment III, SB and RSW differed significantly among treatments, indicating an influence of panel thickness (15 and 18 mm), except on IB.
Discussions
The values of basic density obtained in this study for wood chips derived from veneer residues and logs of paricá agree with values reported in the literature for 14-year-old paricá wood (average 310 kg/cubic metre) (Modes et al. 2020), for five to 11-year old wood (260 to 300 kg/cubic metre), and for wood veneers (average 310 kg/cubic metre).
It is noteworthy that the basic density for the FBR used in the study (A, B, C, D) fell between those of SAL/SAV, which contained only paricá, and MIX, which contained wood waste from other species and had significantly higher density than SAL/SAV.
Thus, it is evident that the wood blending was effective in achieving suitable basic density values, which can promote a favourable compression rate (1.3 to 1.5) for good consolidation of the panels in our study.
The Brazilian standard (NBR 15316-1; ABNT 2009a) classifies panels with densities between 650 kg/cubic metre and 800 kg/cubic metre as ‘Standard MDF’.
Therefore, the panels from all treatments can be included in this class, including the panels in Experiment II even though their density slightly exceeded the ABNT upper limit, as up to five percent of variation is acceptable.
The moisture content of the panels in all experiments ranged from 7.6 percent to 8.7 percent, remaining within the range of four percent to 11 percent established by the NBR 15316-2 standard (ABNT 2009b).
Likewise, the compression ratios of the panels in the experiments ranged from 1.3 to 1.5 and were therefore above or equal to the minimum value of 1.3 indicated to ensure good fibre-to-fibre contact during mattress pressing, providing conditions for proper bonding.
In Experiments II and III, a trend of decreasing WA and TS values with increasing panel thickness was observed, agreeing with the trend observed in MDF panels produced with a density of 850 kg/cubic metre and six thicknesses from four to 14 mm and MDF panels produced with a density of 770 kg/cubic metre and three thicknesses from six to 19 mm.
The NBR 15316-2 standard (ABNT 2009b) does not specify values for WA (2 and 24 h) and TS2h, but it does establish a maximum value of 12 percent for TS24h for MDF panels with thicknesses between 12 and 19 mm.
The TS24h values of the MDF panels in our experiments ranged from 4.4 percent to 9.2 percent, below the maximum value and thus complying with the standard.
In another study of MDF panels produced by blending eucalyptus fibers (70 percent), paricá (20 percent), and sawmill residues (10 percent) bonded with UF resin, and densities ranging from 651 to 800 kg/cubic metre, panels were collected every two hours during a production shift and TS24h values ranged from 11.2 percent to 15.6 percent, and exceeding the maximum requirement stipulated by the Brazilian standard in 80 percent of the collection times.
In Experiment I, a reduction in the amount of MIX fibers led to a significant increase in MOR values. A slight increase in the compression ratio of these panels may have favoured the increase in MOR, and could benefit the mechanical strength of the panels.
The significantly higher RSW (top and face) values in T2 may be related to the IB of the panels in T2, which tended to be higher, though not significantly, from the values in the other treatments.
In Experiment II, the thinner panels had significantly higher IB and RSW (face) compared to the thicker panels, suggesting that IB and RSW (face) are positively related.
Our IB values are consistent with those obtained by Krzysik et al. for panels with thicknesses of six, 12 and 19 mm, equivalent to 1.12, 0.87 and 0.66 MPa, respectively, demonstrating that increasing the thickness led to a reduction in IB.
In Experiment III, MOR and MOE were significantly higher in the thicker panels, while RSW (top and face) was significantly higher in the thinner panels.
Krzysik et al also reported that thickness, MOR and MOE were directly related, i.e., that MOR and MOE increased with panel thickness. According to the normative rule, in the preparation for the RSW (face) test, the screw must penetrate 15 mm into the sample.
In the 15 mm thick panels, the screw practically traversed its entire thickness. On the other hand, in the 18 mm thick panels, the screw only penetrated the outer layer of one face and reached the inner layer, where it remained anchored.
It is presumed that, in the RSW test, due to the fact that the screw passed from face to face through the 15 mm specimens, the resistance offered was greater compared to the 18 mm specimens.
The standard NBR 15316-2 (ABNT 2009b) establishes that, for MDF panels with a thickness from 12 to 19 mm intended for furniture manufacturing, the minimum values for MOR and MOE are 20 MPa and 2200 MPa, respectively.
The panels from all our experiments exhibited average MOR and MOE values above the established minimums, thus fully satisfying the normative requirements.
None of the panels reached the minimum value required for IB, stipulated at 0.55 MPa (NBR 15316-2; ABNT 2009b), intended for furniture manufacturing. On the other hand, according to the same standard, MDF panels with IB values greater than 0.30 MPa can be used as ceilings and walls. Therefore, our panels, which had IB values between 0.31 and 0.48 MPa, could be used for those purposes.
There are reports in the literature indicating that MDF panels do not always meet the regulatory requirements of European (EN 622-5; CEN 2006) and Brazilian (NBR 15316-2; ABNT 2009b) standards, of a minimum value of 0.55 MPa for IB.
MDF panels manufactured with eucalyptus and a density of 756 kg/cubic metre collected from an industrial production line in Brazil had an average IB of 0.30 MPa. The MDF panels obtained for the studies were made from pine wood from industrial production lines in four other Brazilian factories, with densities ranging from 693 to 736 kg/cubic metre, and they had an average IB ranging from 0.35 to 0.53 MPa.
In all cases evaluated, the authors found IB values below the European standard. MDF panels produced with a mixture of eucalyptus fibers (70 percent), paricá (20 percent), and sawmill residues (10 percent) had IB values ranging from 0.42 to 0.62 MPa and did not meet the minimum established by the Brazilian standard in 80 percent of the collection time.
There is no definition of minimum values for RSW (top and face) in the NBR 15316-2 standard (ABNT 2009b). MDF panels manufactured from Pinus taeda wood fibres glued with resin based on UF, with a nominal density of 700 kg/cubic metre had average values of 778 N for RSW (top) and 972 N for RSW (face).
MDF panels produced with Eucalyptus grandis fibres bonded with UF and a nominal density of 700 kg/cubic metre, had average values of 1334 N and 1206 N for RSW (top) and RSW (face), respectively.
MDF panels produced with a mixture of eucalyptus fibres (70 percent), paricá (20 percent), and sawmill residues (10 percent) had RSW (top) values ranging from 468 N to 931 N and RSW (face) values ranging from 837 N to 1343 N.
The RSW (top) and RSW (face) values found for the panels in the present study ranged from 641 N to 834 N and from 821 N to 1170 N, respectively.
Conclusions
There was a significant effect of the fibre blend ratio favouring MOR and RSW (top and face) of MDF panels produced with Schizolobium amazonicum wood and wood waste from native Amazonian species, with particular emphasis on the ratios 40SAV:60MIX and 25SAV:75MIX, respectively.
A significant effect of MDF thickness was observed for IB and RSW (face), with positive highlights for the 12 mm thick panels; and for WA (2 and 24 h) and TS24h for 15 mm thick panels, both produced with a fibre blend ratio of 15SAL:15SAV:70MIX. MDF panels produced with a fibre blend ratio of 25SAV:75MIX and a thickness of 18 mm showed better performance in TS (2 and 24 h) and SB (MOR and MOE), while the 15 mm panels had better RSW performance (top and face).
Except for IB, all other properties of the evaluated MDF panels met the specified regulatory requirements for use in furniture manufacturing.
The results indicate that mixtures of Schizolobium amazonicum (paricá) wood and wood waste from native Amazonian species, have great potential for MDF manufacturing for use in the furniture industry, but adjustments to the production process are recommended to improve IB.
To achieve this, we suggest improving the pressing cycle, increasing the adhesive (UF) content applied to the inner layer fibres of the panels, and reinforcing this adhesive with the addition of melamine (MUF), either individually or in combination.
