Wood samples were collected from 30 with 12-year-old rubber Hevea brasiliensis (Willd. ex Juss.) Muell. Arg. trees (Table 1), three of each clonal progeny, in the municipality of Selvíria, Mato Grosso do Sul State (20°20′S, 51°24′W, elevation 350 m). The trial plantation was established in 2006 at a spacing of 3 m × 3 m from seeds of 10 free-pollinated clones (IAC 40ill., IAC 41ill., IAC 326ill., IAC 311ill., IAC 301ill., IAN 873ill., GT1ill, PB 330ill., Fx 2261ill., and RRIM 725ill.). Soil in the experimental area was classified as Red Latosol with clayey texture (Santos et al. 2006). In 2018, three selected trees of clonal progenies were felled, and discs 10 cm in thickness from each tree at breast height (1,30 m from the ground) were cut from each tree. From each disc, samples close to the bark were used to determine HHV, chemical constituents, wood density, fiber dimensions, thermogravimetric parameters, and properties under Fourier-Transform Infrared Spectroscopy- FTIR.

Table 1
Improved population consisting of ten open-pollination progenies.

Ill.= illegitimate (progenies obtained from a free-pollination clone).

Wood was ground in a Wiley knife mill, of metallic steel type, manufactured by the Marconi-materials company, with the company located in the municipality of Piracicaba, in the state of São Paulo, Brazil, transformed into sawdust, and then sieved with 40-60 mesh to later characterize its chemical and energetic properties.

Samples were fragmented into smaller pieces with a hammer and chisel and milled in a micro mill. The samples, before the determination of the Higher Heating Value (HHV), were dried in greenhouses to determine the moisture content. To perform the analysis, the isoperibolic method was used with an IKA C200, calorimeter of the type coated plastic from the company Biovera-Laboratory Equipment and Technical Assistance, with its company based in Rio de Janeiro, in the state of Rio de Janeiro, Brazil. the methodology adopted for the determination of the HHV, according to ASTM D5865-98 (2004).

Prior to these analyses, the biomasses (all treatments) were dried in an oven at 100 °C for 10 minutes to get the moisture to 0 %. The determination of ash content was performed according to the standard ASTM D1102-84 (2013), and the volatile content was determined according to ASTM E872 (1982). Both tests were done in triplicate. Both standards were adopted, as all material was used for the calculation. The fixed carbon content was calculated according to Equation 1.

Where: FCC= Fixed Carbon Content (%); AC= Ash Content (%); VC= Volatile Content (%).

To determine extractives (EX) and lignin (LI) contents, TAPPI T204 (2001) and TAPPI T22 (2011) were used, respectively. The samples were fragmented into smaller pieces with a hammer and chisel and milled in a micro mill. The resulting powder was sieved through 40 mesh and 60 mesh screens, and the material retained on the last sieve was used for analysis. The analyses were sequential such that the extractives were first removed, then lignin by acid treatment, with holocellulose (HO) content finally calculated. For extractive contents, extractions were performed in solutions of toluene: alcohol (2:1v:v) and alcohol, at times exceeding 12 h in a Soxhlet extractor. For lignin, extractive-free powder was prepared in several stages with 72 % sulfuric acid to obtain insoluble and soluble lignin (Cary 100 UV-visible spectrophotometer). Finally, the two lignin values were added. The content of insoluble lignin (IL) was determined according to Equation 2.

Where: DW=dry sawdust weight, and DWlig=dry weight of insoluble lignin. Soluble lignin (SL) filtrates and the blank were read at two wavelengths (215 nm and 280 nm), respectively, using quartz cuvettes.

The soluble lignin (SL) content was determined according to Equation 3.

Where: DW=dry sawdust weight. Ex and Li were expressed as a percentage (%) of oven-dried weight of unextracted wood.

After determining the levels of extractives, soluble and insoluble lignin, the holocellulose (HO) content was calculated according to Equation 4.

Where: Ex=Extractives and Li=Lignin.

To determine each variable, a triplicate was used for each clone of Hevea brasiliensis.

Wood density was determined by the ratio between dry mass and saturated volume, according to Brazilian standard ABNT NBR-11941 (2003).

Small pieces were cut from the side of samples, and macerations were prepared according to the modified Franklin method (Berlyn and Miksche 1976). Modifications resulted from the differences caused by the higher concentration of hydrogen peroxide used in our study. Macerations were stained with alcoholic safranin and mounted in a solution of water and glycerin (1:1). Fiber measurements were performed on an Olympus CX 31 microscope equipped with a camera (Olympus Evolt E330) and a computer with image analyzer software (Image-Pro 6.3). Terminology followed the IAWA list (1989). Fiber length (FL) and fiber wall thickness (FWT) were evaluated.

Sieved samples consisted of remnants retained in the 30-mesh sieve. Approximately 20 mg of clonal material from each progeny was heated from 0 ºC to 700 ºC at 20 ºC/min^-1 under nitrogen atmosphere, using TGA 55 equipment. The degradation analysis was performed for each thermogravimetric event of rubber tree clones.

From TG curves, mass loss calculations were performed in the following temperature ranges: 50 ºC - 100 °C, 100 ºC - 250 °C, 250 ºC - 400 °C, 400 ºC - 600 °C and 600 ºC - 700 °C. The residual mass at 700 °C was also calculated for each clonal progeny of Hevea brasiliensis.

Sieved samples consisted of remnants retained in the 30-mesh sieve. Approximately 10 mg of the material were used to read the absorbance spectra using a Perkin Elmer model Spectrum 65 spectrometer in the region from 500 cm^-1 to 4000 cm^-1, spectral resolution of 4 cm^-1 and 32 scans.

Tests of homogeneity and analysis of variance (ANOVA) were performed, and when significant difference was detected between treatments, the Tukey test was used, at 1 % significance. The data were analyzed with the statistical software SigmaPlot version 12.

Table 2 shows the values for the energetic properties for each of 10 rubber clonal progenies. The values of HHV ranged from18357 kJ·kg^-1 (IAC-301ill.) to 19070 kJ·kg^-1 (IAC-311ill.), fixed carbon from 15,16 % (IAC-40ill.) to 15,72 % (IAC-311ill.), volatile material from 82,79 % (IAC-311ill.) to 83,92 % (IAC-301ill.), and ash content from 0,42 % (IAC-311ill.) to 1,49 % (IAC-301ill.). The values oscillated among the different rubber tree clonal progenies, presenting different values in their energetic and chemical properties.

The values of energy and chemical characteristics, wood density and wood fiber of rubber tree clonal progenies showed different behaviors. Progenies IAC 311ill., PB 330ill., IAC 41ill., IAC 301ill., and IAN 873ill. presented higher lignin and fixed carbon levels that contributed to HHV (Table 2), thus presenting superior energy performance compared to other clones. According to Trugilho and Silva (2001), these differences in energy properties can be associated with chemical composition that can affect the energy characteristics of biomass.

However, Muzel et al. (2014) reported Brazilian forest species that presented higher HHV and were, therefore, widely used for energy generation, as noted by the author above. Studying the wood of clones of Eucalyptus grandis and H. brasiliensis,Telmo and Lousada (2011) reported HHV of 17895 kJ·kg^-1 and 17897 kJ·kg^-1. Such similarity between these species highlights the wide use of Eucalyptus in Brazil for the generation of energy.

Table 2
Comparison among ten 12-year-old rubber tree clonal progenies, according to energetic properties, chemicals, wood density and fiber analysis.

ill. = illegitimate (progenies obtained from a free-pollination clone); HHV= Higher heating value; CC = carbon content; VMC = volatile matter content; AC = ash content; EC = extractives content; LC = lignin content; HC = holocellulose content; ρbas = Wood density; FL= fiber length; FWT = fiber wall thickness. Distinct letters in rows differ statistically (P<0,001) by Tukey’s test.

For decades, HHV of wood for power generation was considered 18830 kJ·kg^-1. HHV differences between clones observed in this study can be explained by the ecological group of the wood (Tan 1989). Among hardwoods, rubber tree clones have expected HHV of 19089 kJ·kg^-1. This value was observed in the present study, which reported values ranging from 18357 kJ·kg^-1 (IAC 301ill.) to 19070 kJ·kg^-1 (IAC 311ill.). Werther et al. (2000) report an HHV for rubber tree residues of 18410 kJ·kg^-1. In Brazil, values between 16500 kJ·kg^-1 and 18000 kJ·kg^-1 are approved for commercial energy uses.

Clones with higher volatile and ash contents had lower HHV, including IAC 301ill. (83,92 %), IAC 40ill. (83,75 %), RRIM 725ill. (83,69 %), IAN 873ill. (83,67 %), Fx 2261ill. (83,65 %), PB 330ill. (83,55 %), GT1ill. (83,23 %), IAC 41ill. (83,44 %), and IAC 311ill. (82,79 %), consequently reducing the energy potential of the fuel. However, fixed carbon content is directly related to calorific power, meaning that a higher fixed carbon content implies greater resistance to thermal degradation of biomass within the burning apparatus for power generation (Arantes et al. 2013).

The highest contents of fixed carbon were found for clonal progenies IAC 311ill. (15,72 %) and IAC 41ill. (15,70 %). It is said that the percentage of fixed carbon refers to the fraction of coal that burns in the solid state. Therefore, fuels with high levels of fixed carbon in both clonal progenies are preferable for steel use, owing to thermal stability and high energy power (Neves et al. 2011).

Ash content should be in the range of 1 % - 3 % for good energy performance (Schoninger and Zinelli 2012). Thus, the values for this energetic characteristic are within the expected pattern for good energy yield since ash values ranged from 0,42 % (IAC 301ill.) to 1,49 % (IAC 311ill.).

In Brazil, species of Eucalyptus are widely used for energy production because of their rapid growth and energetic properties (Jesus et al. 2017). However, recent studies show similarity in energetic properties between rubber tree clones. Bersh et al. (2018b) characterized Eucalyptus and reported HHV values between 13970 kJ·kg^-1 and 14250 kJ·kg^-1, volatile materials of 83,17 % - 86,16 %, ash content of 0,57 % - 0,60 %, and 13,27 % - 14,25 % of fixed carbon content. These values show that rubber tree clones share similar and superior energetic properties, highlighting their use for bioenergy production when compared to Eucalyptus wood.

According to Paula (2003), wood from clones has higher fiber wall thickness and length, as well as higher lignin and extractive content. Consequently, this wood is increasingly recommended for energy production for its biomass production, resulting in a greater amount of mass per unit of volume and, consequently, greater energy release capacity. Wood with high lignin content contributes to high gravimetric yield and gives greater resistance to charcoal thermal degradation since this wood has more condensed structures that degrade at higher temperatures (Castro et al. 2013, Oliveira et al. 2010).

In this context, clones of higher basic densities present larger dimensions in their anatomical constitution. This characteristic is not true across the clonal spectrum as clone IAC 301ill. presented lower density and smaller fiber dimensions. The increase in wood density is followed by an increase of fiber wall thickness and length (Oliveira et al. 2010), but this was not observed in the present study. IAN 873ill. was taller and wider, but still had lower density than other clones.

As lignin and extractive content increase, density increases proportionally. Thus, we find an increase in the energy yield of fixed carbon content and HHV (Dias-Junior et al. 2015). This premise was not observed for rubber tree clones IAN 873ill. and IAC 40ill. in that they presented lower wood density, even while their HHV and fixed carbon content were higher than those values in other clones. The lower wood density and HHV value of IAC 301ill. may be related to the age of the tree and anatomical characteristics of the species itself (Protásio et al. 2014).

Basic density is an important property of wood and should be considered for energy use of a given biomass since that characteristic is causally related to energy production. That is, the higher the density, the greater the amount of energy stocked per unit volume (Queiroz et al. 2004). It is recommended that wood used as an energy source should present values above 0,40 kg/cm³ of wood density (Alzate et al. 2005), which is confirmed for all clones listed in Table 2.

The differences found in energetic properties of wood among H. brasiliensis clones result from the chemical composition values, fiber dimensions and wood density of the biomass. Studies on other species report that HHV is mainly dependent on lignin and extractive content. High levels of lignin favor the energetic properties of fixed carbon and HHV, which can be attributed to carbon-carbon bonds between monomeric phenyl-propane units present in lignin, which hinder their decomposition, and to the higher content of carbon of this molecular component of biomass (Bufalino et al. 2012, Dietenberger and Hausburg 2016).

Holocellulose is considered the most abundant component of the cell wall. However, it is amorphous and offers no resistance to high temperatures of thermal degradation, hampering its use in energy production because of its negative interference in biomass energy, HHV and physics properties (Tan and Largervist 2011). Clones with higher holocellulose levels present lower HHV, specifically IAC 326ill., Fx 2261ill., IAC 41ill., and GT1ill.

Energy from biomass is usually obtained through thermochemical technologies, especially combustion. High ash content is disadvantageous because it decreases the transfer of heat in fuel and increases corrosion of the equipment used in the process. It also decreases the HHV of biomass (Protásio et al. 2014, Soares et al. 2014).

Only a few studies have reported on the characteristics of wood for energy purposes. For example, Menucelli et al. (2019) studied 10 clonal progenies of rubber tree and reported characteristics superior to clones evaluated in the present study, such as fiber length between 1189 μm - 1097 μm, fiber wall thickness of 4,55 μm - 5,13 μm, extractive contents between 12,42 % - 16,34 %, lignin content between 27,51 % - 22,47 %, and wood density from 0,57 kg/cm^-3 to 0,66 kg/cm^-3. These wood characteristics promoted higher HHV content between 18592 kJ·kg^-1 and 19757 kJ·kg^-1, values that could be attributed to differences in genetic material (Soares et al. 2014).

Table 3 summarizes the main chemical and functional groups found in the wood of H. brasiliensis clones according to Tyutkova et al. (2019), being identified by the number of waves of the carboxylic groups: = C-H (902 cm^-1), C-O-O (1114 cm^-1), - C-H (1464 cm^-1), C = O (1743 cm ^-1), C-H (2924 cm ^-1) and O-H (3370 cm ^-1).

Table 3
Characteristic bands of infrared spectra (FTIR) for Hevea brasiliensis clones and their respective functional groups.

One can observe that all clones have a strong absorption band at 1114 cm^-1 where the lignin present in wood is located (Schuerch 1989). However, clones IAC 40, GT1, PB 330, IAC 41, RRIM 725, and IAN 873 had higher wavelengths in this spectrum. This could be explained by the chemical structure of wood because these clones have higher levels of lignin (Table 2) and, hence, higher resistance to thermal degradation of the material, as characterized by the stretch mode of the O-H combination.

According to Figure 1, the group found at 1743 cm^-1 can be attributed to C=O, indicating the presence of an acetyl or carboxylic acid group derived from lignin. The high proportion of these chemical groups indicates an increase in HHV (Table 1). The same behavior was reported by Schuerch (1989).

It was possible to notice a strong striation between bands 1743 cm^-1 and 2924 cm^-1, larger than the other spectral bands reported. The bands are characterized by strong bonding between hemicellulose and lignin, which is characterized mainly by C-H, N-H and O-H that are desirable for energy production (Popescu et al. 2011).

At band 3370 cm^-1, the phenolic groups bonded to hydrogen formed O-H chemical groups (Pandey and Pitman 2003). This group is responsible for high degradation resistance of the wood, which is dependent on lignin content, as was possible to observe in this study. Clones with higher lignin levels present higher absorption spectra, namely IAC 40 and GT1, in the present study (Figure 1).

Figure 1
Main spectra obtained for Hevea brasiliensis wood clone.

Table 4 shows averages of mass loss of different clonal progenies of Hevea brasiliensis, as a function of temperature ranges from 50 ºC to 700 ºC.

Table 4
Mean values of mass loss (%) of different clonal progenies of Hevea brasiliensis as a function of temperature ranges.

* Residual mass, considering the mass of wood absolutely dry (s).

The first two temperature ranges, 50 ºC - 100 ºC and 100 ºC - 250 ºC, correspond to drying of wood (Vieira 2019), making an average total of 4,68 % of initial average total loss for clonal progenies of Hevea brasiliensis.

Between 250 ºC and 400 ºC, an average loss in biomass of 65,11 % was verified. It can be inferred that most of this lost mass results from degradation of hemicellulose, cellulose, and volatile emissions, as well as the start of partial lignin decomposition (López-González et al. 2013). The maximum rate observed among clonal progenies of Hevea brasiliensis in this range of degradation refers to the maximum degradation of cellulose, as this constituent corresponds to 40 % - 45 % of the wood (Pereira et al. 2013). According to Figure 2, it is possible to observe the mass losses of Hevea brasiliensis clones at different temperatures.

Figure 2
TG and DTG curves of mass losses for Hevea brasiliensis clones.

Between 400 and 600 ºC, the average mass loss was 5,94 %. According to Shen et al. (2010) for this temperature range, wood thermal degradation of 5 % to 10 % can be expected, as was observed in Hevea brasiliensis.

The general average mass loss observed in the 600 ºC to 700 ºC range was 1,47 %, showing that hemicelluloses and celluloses were totally degraded and that the observed mass loss refers to lignin in small proportion since it is a more thermally stable wood component. The absence of a peak of degradation related to lignin likely results from the fact that its thermal decomposition occurs over a wide temperature range, emphasizing that only a fraction decomposes at temperatures below 450 ºC, as mentioned by Huang et al. (2009).

The residual mass refers to total wood that is converted into charcoal at the end of the carbonization process. Lower residual masses were observed for clones IAC 41ill. (18,45 %) and IAN 873ill. (21,95 %), but higher residual masses for IAC 326ill. (24,99 %) and IAC 301ill. (24,49 %). The values for other clonal progenies were close. These differences between residual masses can be explained by the clonal effect of the species (Vieira 2019).

In Brazil, Eucalyptus clones are widely cultivated for energy production (Ferreira et al. 2017). However, it was found that clonal progenies of Hevea brasiliensis are more thermally resistant than 12 Eucalyptus clones studied by Vieira (2019), which obtained an average residual mass of 22,03 %. On the other hand, they are less thermally resistant to Eucalyptus clones studied by Pereira et al. (2013), who report residual masses at 7,5 years with values of 25,11 %, these differences being attributed to wood chemical constitution.

^♠Corresponding author: erick.amorim95@hotmail.com