PREPARATION AND CHARACTERIZATION OF CARBON FROM THE FRUIT OF BRAZIL NUT TREE ACTIVATED BY PHYSICAL PROCESS

Activated carbon is the name of a big group of materials that presents high degree of porosity and, consequently, an extended internal surface area, with physical and chemical adsorption properties. Innumerous raw materials can be used as precursors, especially biomass. The objective of this study was to obtain activated carbon from physical activation of the fruit of Brazil nut tree (“ouriço”) and to evaluate its physical and chemical properties in function of the diff erences between the temperatures and atmospheres of activation. The samples were carbonized at 3 diff erent temperatures and the carbons were activated under atmosphere saturated by CO 2 or steam. The results showed the infl uence of activation temperature and atmosphere on physicochemical characteristics of carbon. The carbon from woody Brazil nut seed capsule activated by CO 2 and steam at diff erent temperatures had a microporous profi le, indicating its use to adsorb organic molecules of small dimensions. Basic characteristic was observed on samples and the carbon produced was thermally stable. Best quality was attributed to carbon activated by steam at 800 °C.


INTRODUCTION
Activated carbon is the name for a big group of porous materials based on a carbonic matrix. Carbonaceous materials are carbonized after their impregnation with chemicals, it is called chemical activation; or a char is treated with oxidizing gases, a process named as physical activation. These conditions accord a high degree of porosity and, consequently, an extended internal surface area, withal physical and chemical adsorption properties (Rodriguez-Reinoso and Silvestre-Albero, 2016).
The main diff erences between activation process are: carbonization and activation are unseparated steps during chemical activation, whereas they are independent during physical activation (Prauchner and Rodríguez-Reinoso 2012). Physical activation presents several advantages and disadvantages compared to chemical activation. The main advantages are the low cost, considering there is no need of acquiring activating agents, such as acid or basic chemicals, and there is no need of an additional washing stage. Moreover, these chemicals can be corrosive (Maciá-Agulló et al., 2004).
The characteristics of activated carbons are due to the precursor material and the activation method (Bhatnagar and Sillanpää, 2010). A plenty of lignocellulosic materials were employed as precursors (Ahmad et al., 2007;Daud et al., 2000;Jaguaribe et al., 2005;Melo et al., 2015;Tsai et al., 2001). However, other materials such as animal bones and petroleum coke can also be used (Djilani et al., 2016;Kawano et al., 2008).
The woody Brazil nut (Bertholletia excelsa) seed capsule has high lignocellulose content (Rambo et al., 2015;Scussel et al., 2014) which is an important feature for obtaining a high quality carbon (Yang et al., 2007). As a byproduct, it has low cost, which, added to the lower cost of physical activation, may produce a cheaper activated carbon, to be used: in the treatment of water supply (Borges et al., 2016), in medical and environmental applications (Alkhatib and Zailaey, 2015), in natural gas storage (Gottipati et al., 2012), in the fi ltering of compounds used in agricultural activity (Melo et al., 2015), and others.
In the present study, activated carbon was obtained from physical activation of woody Brazil nut seed capsule. Its physical and chemical properties were characterized in function of the diff erences between the temperatures and atmospheres of activation.

MATERIAL AND METHODS
Samples of woody Brazil nut seed capsule were fragmented into 2 to 3 cm pieces and then washed in running water and kept in an oven with forced air circulation at 105 °C for 24 hours for drying.
Cleaned and dried material was fi tted inside a metal cylindrical container (length: 720 mm; internal diameter: 68 mm) and the fi lled container was placed in a tubular oven (FT-1200/H1Z, Fortelab, São Carlos, SP, Brazil) with heat rate of 10 °C/min. The oven was adjusted to the pyrolysis temperature and the samples were maintained at these temperatures for fi ve hours. The tested temperatures were 600 °C, 700 °C and 800 °C. Shortly after the pyrolysis of the material, the atmosphere was saturated with an oxidizing agent for 40 minutes, for the activation of the carbon. The tested agents were CO 2 (pressure of 0.58 kgf.cm -2 ) and steam at 1 kgf.cm -2 from a vertical boiler (EIT-VL, EIT, Paiçandu, PR, Brazil).
Water, volatile, ash and fi xed carbon contents were determined according to the method standardized by ASTM D 1762-84 (ASTM D 1762-84, 2007. The results were expressed as a percentage in dry basis, except for water content which was expressed as a percentage in wet basis. The pH values were measured using the standard test method ASTM D 3838-80 (ASTM D 3838-80, 1999) and bulk density by the standard test method ASTM D 2854-09 (ASTM D 2854(ASTM D -09, 2009). Activated carbon yield was determined by weighing the precursor material that fi lled the container and the material removed from the oven, after activation. The result was expressed as a percentage in wet basis.
The absorption bands in the infrared region of the activated carbon samples were determined in a Fourier transform (FTIR) spectrometer equipped with an attenuated total refl ectance (ATR), (Spectrum BXII, Perkin Elmer, Billerica, MA, USA) in the region between 4000 and 400 cm -1 , with resolution of 16 cm -1 .
Samples of activated carbon were submitted to BET (Brunauer, Emmett e Teller) analysis (Tristar II Kr 3020, Micromeritics, Norcross, GA, USA) to determine the specifi c surface area. The distribution for microporous size was analyzed by DFT (Density Functional Theory) method.
The thermogravimetric analysis (TGA) curves were determined in a thermogravimetric analyzer (SDT Q600, TA Instruments, New Castle, DE, USA), only for the samples that presented the smallest contents of water and volatile and the highest yield, pH and content of fi xed carbon, for both atmospheres. The explored temperature ranged from 20 °C to 1000 °C at a heat rate of 10 °C.min -1 , under atmosphere of ultrapure air (100 mL.min -1 ).
A full 2³ factorial coupled with statistical analysis of the results, by using analysis of variance (ANOVA) was carried using Action Stat Pro (ESTATCAMP and DIGUP, 2017). Five replications were used for all the physicochemical properties analyzed.
Higher results for yield were obtained from carbon activated by CO 2 at 600 °C, and from those activated by steam at 600 °C and 700 °C. Smaller contents of volatile were acquired at temperatures of 600 °C and 700 °C for the activation with CO 2 , and at 800 °C by steam activation. Steam at 800 °C allows higher ash contents. For fi xed carbon, either CO 2 and steam at 700 °C and 800 °C allowed higher results. The temperature of 800 °C, for both activation atmospheres, allowed greater basicity to samples (Table 1).
The activation by steam, regardless of the temperature, allowed smaller water contents, compared to those activated by CO 2 . The results for ash content of samples activated by steam and CO 2 , were diff erent only for the temperature of 600 °C. Steam allowed higher fi xed carbon under any temperature ( Table 1).
The specifi c surface area for samples activated by CO 2 at 600, 700 and 800 °C were 15.8, 55.0 and 395.0 m 2 g -1 , respectively. From steam activation the results were 125.0, 320.0 and 401.0 m 2 g -1 , for samples activated at 600, 700 and 800 °C, respectively.
The N 2 sorption/desorption isotherm for carbon activated by both atmospheres showed high adsorption of N 2 under low pressures (Figure 2 a and b) and higher intensity of porous of 1.16, 1.17 and 1.17 nm, for temperatures of 600, 700 and 800 °C, respectively ( Figure 3a). For the samples activated by steam at 600 °C, the porous of 1.26 nm was the most founded, followed by 1.0 and 1.7 nm for samples activated at 700 and 800 °C for the same atmosphere, respectively ( Figure 3b).
The thermogravimetric analysis (TGA) curves, for both activation methods, showed two infl ection points defi ning three ranges for mass loss. Considering   the samples activated by CO 2 , there was a mass loss of 3.08 % up to 150 °C, 0.67 % at temperatures ranging from 600 to 700 and 8.62 % above 700 °C. For the activation by steam, mass losses were 9.79 %, 0.56 % and 16.78 % for temperatures up to 150 °C, between 600 and 700 °C and above 700 °C, respectively (Figure 4).

DISCUSSION
Yields were infl uenced by the temperature but not by the atmosphere. This fact allows to infer that yield is defi ned during carbonization step, when volatile is released, and high rates of mass loss are observed (Lua and Guo, 2001;Rodríguez-Reinoso et al., 1995;Yang and Lua, 2003). So, since higher temperatures are acquired, smaller yields are observed. Similar behavior was observed for activated carbon from pistachio husk (Yang and Lua, 2003) and from palm trunk (Ahmad et al., 2007), where higher yields were getting from lower temperatures of carbonization, regardless of the activation atmosphere. Whereas the composition of biomass in general, yields close to 30 % are considered fair (Nobre et al., 2015b).
Neither the temperature nor the atmosphere of activation infl uenced water content, since water is volatilized up to 200 °C during carbonization step (Pastor-Villegas et al., 1999). However, higher temperatures of carbonization can contribute to hygroscopicity due to increasing on porous formation (Boas et al., 2010) The hygroscopic feature of the activated carbon allows chemical-physical adsorption of water inside the porous (Ahmad et al., 2007;Anisuzzaman et al., 2015). So, during the sample handling, water may be absorbed from air moisture. Lower water contents are desirable, since high water contents are related to decreasing on mechanic resistance of activated carbon, consequently, dust generation (Boas et al., 2010).
Releasing of volatile is due to the change to vapor phase of several molecules, by the action of the carbonization temperature (Silva and Brito, 1990). The atmosphere change during the activation step is also important for the residual volatile in the activated carbon. During this stage, the gas permeates into solid matter, contributing to desorption, distillation and removal of volatile that still existed in carbon. Also, activation gases contribute to stabilization of the radical acquired during thermal decomposition, boosting the volatile release (Ahmad et al., 2007). The formation of micro and macro pores is infl uenced by volatile release, which means that high volatile content in the activated carbon may imply a low surface area (Zhang et al., 2004). Volatile contents higher than 11 % were presented by other studies in similar conditions of carbonization and activation of the biomass (Ahmad et al., 2007;Róz et al., 2015;Yang and Lua, 2003), disclosing the high quality of the activated carbon from woody Brazil nut seed capsule.
Ash content of activated carbon is related to the composition of precursor and the characteristics of the process of carbonization and activation, since the latter may allow the contamination of the carbon with inert matter (Collet, 1956;Silva and Brito, 1990). Lower ash content may also contribute to indicate the quality of activated carbon (Jaguaribe et al., 2005). As a non-carbonic mineral additive, it may impair adsorption due to changes in interaction between carbon surface and the species to be adsorbed (Bautista-Toledo et al., 2005). Besides blocking pores in carbon and its hydrophilic characteristic, ash increases water adsorption rather than compounds of interest (Ahmedna et al., 2000;Brum et al., 2008). In general, commercial activated carbon presents ash contents ranging from 10% to 15% (Jaguaribe et al., 2005;Lopes et al., 2013), far above those found in this work.
High content of fi xed carbon confers a matrix able to generate functional groups or surfi cial complex, that serves as bond sites to adsorption of compounds of interest (Aznar, 2011;Mohan and Pittman Jr., 2006). Negative correlation with yield (Brito et al., 1987) and positive correlation with temperature (Róz et al., 2015) are observed during carbonization step. Increasing on temperature leads to losses of condensable and noncondensable compounds, as CO and CO 2 (Pinheiro et al., 2005;Róz et al., 2015;Yang et al., 2007), reducing yield and fi xing residual carbon.
pH may indicate the presence of groups on carbon surface. High values show a low index of acid groups on activated carbon (Strelko and Malik, 2002). Complexes formed between carbonic matrix and oxygen atoms can determine the acidic or basic characteristics of the activated carbon (Aznar, 2011;Mohan and Pittman Jr., 2006;Wibowo et al., 2007). The basic characteristic of the activated carbon is due to the high temperature of activation and to the activation atmosphere. Both parameters collaborate to the breakdown of bonds in the carbonic matrix and to a new arrangement with gases that compose the atmosphere (Leon y Leon et al., 1992;Mohan and Pittman Jr., 2006;Pereira et al., 2003;Wibowo et al., 2007).
The chemical surface of the activated carbon determines the capacity to retain water, catalytic characteristics, acidic or basic feature and the capacity of adsorption (Salame and Bandosz, 2001). It is related to the presence of heteroatoms (oxygen, hydrogen and nitrogen), also the atom of carbon inside the matrix (El-Sayed and Bandosz, 2004;Salame and Bandosz, 2001).
There was a broad band at 3348 cm -1 in the precursor's spectrum (woody nut seed capsule), attributed to a vibrational stretch of hydroxyl group (O-H) linked to hydrogen for alcohols and phenols, possibly due to moisture of the sample (Ramos et al., 2009;Yang and Lua, 2003). The same band was founded for other precursors used for activated carbon production (Ramos et al., 2009). Vibrations corresponding to ѵ (C-H) of alkanes (CH3 and CH2) were observed at 2924 cm -1 (Ramos et al., 2009;Yang and Lua, 2003). CH 2 and CH 3 groups are confi rmed for bands at 1446 and 1370 cm -1 , typical for angular deformation of these groups (Gomez-Serrano et al., 1996;Jagtoyen et al., 1992;Yang and Lua, 2003). Potential olefi nic ѵ (C=C) absorption in the sample allows signals at 1654 cm -1 , other two bands around 1510 and 1416 cm -1 can be due to stretch of C=C in aromatic rings (Yang and Lua, 2003). A band at 1330 cm -1 can be due to ѵ (C=O) vibrations of carboxilated groups and C-O stretch vibration is observed between 1300 and 900 cm -1 . The band at 1228 cm -1 and the shoulder at 1660 cm -1 indicate the presence of esters (R-CO-O-R'), ethers (R-O-R') or other phenolic groups on precursor. The other shoulder observed at 1107 cm -1 associated to the band at 1026 cm -1 can be associated to alcohol groups (R-OH). Finally, the band at 668 cm -1 represents γ (OH) (Yang and Lua, 2003).
The bands at: 2928 cm -1 for carbon activated by steam at 600 and 800 °C; 2891 cm -1 also for activation by steam at 700°C; and 2884 cm -1 for activation by CO 2 at all temperatures, can be attributed to ether group (-O-CH3) and/or to C-H bonds, symmetric or asymmetric, of methyl and methylene groups (Ahmad et al., 2007). Bands at 1584 cm -1 for carbons activated by CO2 at temperatures of 600 and 700 °C and at 1134 cm -1 for temperature of 800 °C represent the stretch of carbonyl group in quinone, also the structure of γ-pyrone with strong vibrations of the C=O and C=C combination (Ahmad et al., 2007;Tsai et al., 2001). At 1127 cm -1 , the band found in the samples activated by steam at 800 °C, and at 1134 cm -1 for carbon activated by CO2 at 700 °C, represent ketones, alcohols, pyrones and aromatic deformations in C-H plan (Ahmad et al., 2007). The bands at 984 cm -1 (carbon activated by CO2 at 800 °C), 953 cm -1 (carbon activated by steam at 700 and 800 °C) and 920 cm -1 (carbon activated by steam at 600 °C) can be due to vibration of ethers (-C-O-C-) (Ahmad et al., 2007). The bands at 847 and 771 cm -1 found in the samples activated by steam at 600 °C can be related to C-H out of the plan in aldehydes -CHO, compounds of pyranose and others benzene derivatives (Ahmad et al., 2007).
The basic characteristic of the carbon confers greater capacity of adsorption of acidic substances (Leon y Leon et al., 1992;Mohan and Pittman Jr., 2006;Pereira et al., 2003;Wibowo et al., 2007). The formation of structures of pyrone, ether and groups carbonyl contributes to the basic characteristic of the produced carbon.
The results for specifi c surface area, both for the carbon activated by CO 2 and by steam, are considered fair for activated carbons from wood pyrolysis (Herzog et al., 2006). That fact indicates the quality of woody Brazil nut seed capsule as a precursor for activated carbon production, besides being a raw material of low cost, renewable and abundantly found.
Independent of the atmosphere of activation, the samples presented a microporous profi le, confi rmed by the higher adsorption volume of N 2 under low pressure (Sun and Webley, 2010) and by the identifi cation of peaks at diameter smaller than 2 nm (20 Å) (Ahmad et al., 2007). Activated carbons with microporous profi le provide high capacity of adsorption of organic molecules of small dimensions, with potential to retain gases and common solvents (Nobre et al., 2015a).
From BET analysis, the microporous profi le was confi rmed for both activation process, since the isotherms were type I, typical of microporous materials with external surfaces relatively small (Sing et al., 1985). Type I isotherms confer characteristic of chemisorption for activated carbon (Shaji and Zachariah, 2017), on that case, the adsorption is limited by the volume of accessible micropores instead of internal surface area (Sing et al., 1985). This type of isotherm is typical for carbon produced from biomass as a precursor (Oliveira et al., 2016;Yang and Lua, 2003). Pore distribution was not aff ected by activation atmosphere, indicating that woody Brazil nut seed capsule can be used as a precursor independent of the atmosphere used.
From thermogravimetric curves, both carbons activated by CO 2 and activated by steam, at 800 °C, presented similar thermal stability. The initial mass loss observed at temperatures up to 200 °C can be explained by water evaporation (Pastor-Villegas et al., 1999), which could have been adsorbed during the storage or handling of the samples (Oliveira et al., 2016). The second mass loss, occurred between 600 and 700 °C, could be due to the decomposition of groups formed on carbon surface during activation process (carbonyl groups, ethers, structures of pyrone) and/or also by the decomposition of the carbon skeleton (Oliveira et al., 2016). The small loss of mass over a wide temperature range provides the desired stability to activated carbon.

CONCLUSIONS
Temperature and atmosphere of activation had signifi cant eff ects on physico-chemical characteristics but did not present interaction between them. Samples presented a microporous profi le for both atmospheres, indicating their high capacity to adsorb organic molecules of small dimensions. Pore distribution was not aff ected by activation atmosphere and the results for specifi c surface area confi rmed the quality of woody Brazil nut seed capsule as precursor for activated carbon. Considering the results, activation by steam at 800 °C led an activated carbon with better quality.