INTRODUCTION

Brazil stands out in terms of energy resources since it has all the primary energy sources, with a favorable scenario for its energy demands in long-term (^{VENTURA FILHO, 2009}). This is due to its large territorial area, almost all located in tropical and rainy regions, which offers excellent conditions for the production and energetic use of biomass in large-scale.

Most of the energy biomass in Brazil is used for the production of charcoal, which makes the country the world’s largest producer and consumer of this product (^{BARCELLOS, 2007}). In this context, Cerrado biome is one of the energy sources of Brazil due to its unique ecological aspect, as well as its peculiar physiognomies (^{SILVA; VALE; MIGUEL, 2015}).

Wood (firewood and wood products) is the oldest source of fuel used to produce energy and it is an essential component in meeting Brazil’s energy demand. This energy source will probably continue to be predominant, with most of the consumption located in the production sectors of charcoal, household, industrial, and agricultural (^{NASCIMENTO; BIAGGIONI, 2010}; ^{BRITO, 2007}).

Although the large use of wood for energy purposes, few studies describe the performance characteristics of Brazilian species (^{CINTRA, 2009}). Thus, the knowledge of some physical aspects of wood, such as its basic density, is essential for the assessment of its qualities for producing energy inputs, especially charcoal, since a high wood density implies more mass per unit volume and hence more energy (^{BARCELLOS, 2007}).

In addition, the influence of immediate analysis (ashes, volatile matter, and fixed carbon), higher calorific value, and energy density are very important in order to determine the energy potential of wood (^{ANDRADE; CARVALHO 1998}; ^{QUIRINO et al., 2005}). Thus, considering the importance of these variables in the use of wood for energy purposes, the aim of this study was to determine a model for the energy density of Cerrado species.

MATERIAL AND METHODS

RESULTS AND DISCUSSION

The Cerrado tree species assessed in this study presented values of VM between 77.4 and 85.5% (mean = 82.88%), ASH between 0 and 1% (mean = 0.27%), and FC between 14 and 23% (mean = 16.84%) (Table 1). The species presenting the highest values for VM were Byrsonima laxiflora Griseb (85.63%), Conarus perrotetti (85.62%), and Qualea parviflora (85.95%) whereas the lowest values were registered for Ferdinandusa elliptica (77.42%) and Tapirira guianensis (77.81%). The highest value of ASH was recorded for Myrcia splendens (0.83%), while the lowest value was found for Bowdichia virgiloides (0.02%). FC ranged from 13.69% (Qualea parviflora) to 22.24% (Ferdinandusa elliptica).

According to ^{Brito and Barrichelo (1978}), the immediate analysis of a fuel gives the percentage of material burning in the gaseous (volatile matter) and solid (fixed carbon) states, as well as an indication of the residual material (ashes). ^{Vale, Brasil and Leão (2002}), ^{Chaves et al. (2013}), and ^{Brito and Barrichelo (1982}) recommended for biomass, in general, volatile matter contents between 75 and 85% and fixed carbon between 15 and 25%. ^{Cintra (2009}) worked with native forest species in São Paulo State, Brazil and found values of volatile matter between 79.6 and 84.9%. The values found in our study for these variables are in accordance with those found in the literature for wood and their variation is influenced by the species. The average ASH varied according to the species, but with values considered close to those of planted forests in Brazil, indicating acceptable results for energy purposes by the analyzed material. According to Chaves et al. (2013) and ^{Pincelli (2011}), woods with a high ash content are not recommended for energy purposes for being a substance that remains in the solid form and it does not combust. ^{Vidaurre et al. (2012}) studied the energy properties of Paricá wood (Schizolobium amazonicum) and found an average ash content of 1.3%, which is considered high when compared to the ash content of Eucalypt. ^{Morais, Nascimento and Melo (2005}) found an average value of 1.26% for wood of Pinus oocarpa and related this high value to the higher content of inorganic compounds in the wood of this species when compared to others. ^{Barcellos (2007}) and ^{Moutinho et al. (2016}) found differences in ash content according to the studied species, with values ranging from 0.12 to 1.11%. Vale, Brasil and Leão (2002) reported values between 0.15 and 2.73% for 47 Cerrado species, indicating that the values found in our research were similar to those found in the literature.

According to ^{Chaves et al. (2013}), during the burning of plant biomass volatile matter rapidly volatilize, which contributes to a low energy efficiency since fuel residence time decreases due to its rapid volatilization. A strong and inversely proportional relationship is usually found between volatile matter and fixed carbon. On the other hand, fixed carbon is directly related to the calorific value. A higher fixed carbon content in the material implies a longer fuel residence time.

^{Vale, Brasil and Leão (2002}) found an average value of fixed carbon of 20.73% for Cerrado woods, ^{Almeida et al. (2015}) reported values between 17 and 18.57%, and ^{Cintra (2009}) found values between 14.4 and 19.6%, which is close to those found in our study. ^{Chaves et al. (2013}), ^{Silva (2013}), and ^{Barcellos (2007}) emphasized that fixed carbon content presents a negative relation with volatile matter content.

The higher calorific value was between 18,282 GJ ton^-1 for Parkia platycephala and 20,121 GJ ton^-1 for Pouteria ramiflora (mean = 19,233 GJ ton^-1). ^{Pinheiro, Rendeiro and Pinho (2005}) found values for the higher calorific value of vegetal residue ranging from 18,810 to 20,900 GJ ton^-1 regardless of particle size and density of biomass or origin wood in the case of sawdust. ^{Brito (1993}) points out that the variation in higher calorific value for wood, in general, is between 14,630 and 20,900 GJ ton^-1. The average values cited in the literature are in agreement with those found in our study, i.e. between 18,282 and 20,121 GJ ton^-1. ^{Couto et al. (2004}) observed an HCV value for wood of 18,366 GJ ton^-1. ^{Quirino et al. (2005}), studying the higher calorific value of forest species, found values between 14,003 and 21,986 GJ ton^-1, with an average value of 19,779 GJ ton^-1. In addition, ^{Vidaurre et al. (2012}) registered an average value of 18,638 GJ ton^-1 whereas ^{Vale, Brasil and Leão (2002}) observed values ranging from 18,876 to 20,854 GJ ton^-1 considering 47 Cerrado species.

Among the 34 species analyzed, Pouteria ramiflora (15,006 Mcal m^-3), Miconia cuspidata (16,177 Mcal m^-3), and Emmotum nitens (15,977 Mcal m^-3) showed the highest energy density values. On the other hand, the species with the lowest energy density were Vochysia gardineri (6,871 Mcal m^-3), Connarus perrottetti (8,380 Mcal m^-3), and Byrsonima laxiflora (9,315 Mcal m^-3). The average value of energy density for the 34 species was 12,459 Mcal m^-3.

According to ^{Vidaurre et al. (2012}) and ^{Cunha (1989}), the density has no direct relation with the amount of energy contained in the wood. Both studies, as well as the result observed in ours, mentioned no correlation between basic density and higher calorific value. However, these authors emphasize that the higher the density is, the greater the amount of energy contained per unit volume, which stimulates the interest in denser woods for burning. When the relationship between the variables of the immediate analysis (VM, ASH, and FC) and HCV was tested, a correlation was observed between VM and FC (R² = 0.9927) (Figure 1). The higher calorific value and basic density of the studied species did not present a significant relationship since the high-density species showed a low HCV (Figure 2). Table 2 and Figure 3 show the relationship between energy density and other variables.

Figure 1 Relationship between fixed carbon, ashes, and volatile matter and higher calorific value of Cerrado tree species from Tocantins State, Brazil. 

Figure 2 Relationship between higher calorific value and basic density of 34 Cerrado tree species from Tocantins State, Brazil. 

Figure 3 Relationship between energy density and calorific value, volatile material content (VM), fixed carbon content (FC), ash content (ASH) of 34 Cerrado tree species from Tocantins State, Brazil. 

In a study on the energy potential of six-year-old eucalyptus trees, ^{Lima et al. (2007}) found 9,168 Mcal m^-3 of energy density. ^{Pinheiro, Rendeiro and Pinho (2005}) studied the energy density of plant residues (fruit peels) and found values of energy density from 400 to 2,600 Mcal m^-3. When analyzing six-year-old eucalyptus trees, ^{Lima (2011}) observed a value of energy density of 11,103 Mcal m^-3 whereas ^{Brito, Barrichelo and Seixas (1983}), found values ranging from 7,843 to 12,038 Mcal m^-3 for eucalyptus species. According to ^{Lima (2011}), energy density considers the energy contained in a given wood volume. The variation of values in our study was close to the values found in the literature, ranging from 6,871 to 16,177 Mcal m^-3, with an average of 12,459 Mcal m^-3. Although the average energy density in the literature does not present so much variation, few studies with native species can be found.

^{Brito and Barrichelo (1977}) concluded that in the choice of wood to obtain charcoal with better chemical properties (higher contents of fixed carbon and lower of volatile substances and ashes), those with high lignin contents and high basic density should be sought. This provides an increase in the amount of dry matter in the oven. In a study with 108 forest species, ^{Quirino et al. (2005}) observed that the highest calorific value does not match the species with the highest basic density. In addition, ^{Quirino et al. (2012}) reported that low-density woods could be used for energy if converted into briquettes.

Fixed carbon is directly related to the calorific value, especially the charcoal, being the variation higher, ranging from 6% in a dry base of wood to up to 90% in charcoal. This may be the reason for the low relationship between the higher calorific value and fixed carbon content found in our study. Another possible explanation may be the presence of extractives. ^{Vale, Sarmento and Almeida (2005}) reported that the best use of wood for energy production as heat is linked to a higher specific mass, higher fixed carbon content, and lower volatile matter content. ^{Souza (2010}) observed that the parameter density might influence the calorific value of a material as the heat generated by combustion is related to the volume. Thus, wood and denser wood products presented higher calorific value per unit volume.

The coefficient of determination (R²) presented in Table 3 showed that 98% of the energy density variation registered can be explained by the average basic density of the tree, with mean percentage errors of less than 2.1%, as the standard error of the estimate (S_yx). The model adjusted for energy density estimation as a function of the basic density is shown in Table 3. The residual distribution did not exceed a variation of 6% either to underestimate or overestimate the energy density of the tree by the basic density of the 34 species (Figure 4).

Figure 4 Residual distribution according to the adjusted model to determine the energy density as a function of the basic density of the tree. 

The estimated energy density of the area was 1,378,541 Mcal ha^-1. The proposed model for calculating the energy density as a function of the basic density facilitates the energy surveys in areas to be explored since only the values of basic density need to be obtained, without the need of calorific value analysis. The knowledge of the energy potential of the analyzed native species allows advancing in the studies for possible improvements in the quality of wood species directed to a sustainable forest energy production.

CONCLUSION

From the adjusted model, it is possible to calculate the energy density of species as a function of the basic density.