Brazilian natural fiber (jute) as raw material for activated carbon production

Jute fiber is the second most common natural cellulose fiber worldwide, especially in recent years, due to its excellent physical, chemical and structural properties. The objective of this paper was to investigate: the thermal degradation of in natura jute fiber, and the production and characterization of the generated activated carbon. The production consisted of carbonization of the jute fiber and activation with steam. During the activation step the amorphous carbon produced in the initial carbonization step reacted with oxidizing gas, forming new pores and opening closed pores, which enhanced the adsorptive capacity of the activated carbon. N2 gas adsorption at 77K was used in order to evaluate the effect of the carbonization and activation steps. The results of the adsorption indicate the possibility of producing a porous material with a combination of microporous and mesoporous structure, depending on the parameters used in the processes, with resulting specific surface area around 470 m2.g–1. The thermal analysis indicates that above 600°C there is no significant mass loss.


INTRODUCTION
Jute fiber is by far the most common natural fiber in the world. Its use has increased significantly due to its unique combination of physical, chemical and structural properties, and also due to the growing concern with the environment (Alves et al. 2010, Zhou et al. 2013. The characteristics of a typical jute fiber are shown in Table I. There is a growing interest regarding the use of lignocellulosic materials such as jute, sisal, coconut, banana and curauá fibers as precursors for the production of activated carbon (Asadullah et al. 2013, Roy et al. 2012, Williams and Reed 2006, Phan et al. 2006, Senthilkumaar et al. 2005 as well as to strengthen polymeric matrixes of composites (Ramesh et al 2013, Ku et al. 2011. Ioannidou and Zabaniotou 2007 mention that activated carbon from natural fibers exhibit specific surface areas ranging from 80 m 2 .g -1 to 900 m 2 .g -1 , with porous structure containing mostly micropores when produced at heating rates of 2°C. min -1 , at temperatures between 200°C and 300°C during carbonization and at activation heating rates of 4°C.min -1 , with final temperature of 800°C. CARLA F.S. ROMBALDO, ANTONIO C.L. LISBOA, MANOEL O.A. MENDEZ AND APARECIDO R. COUTINHO Some studies have indicated that activated carbon from natural fibers reach adsorption rates 2 to 50 times faster the ones from coal, due to the small diameter of the fibers, around 10 μm, presenting a larger external surface area in contact with the fluid (Baudu et al. 1991). Furthermore some studies have shown its potential to withdraw water pollutants such as phenol, pesticides and dyes (Asadullah et al. 2013, Banerjee and Dastidar 2005, Hameed et al. 2007, Ahmaruzzaman and Gayatri 2010, Kumar et al. 2008).
Aiming at the use of available natural resources and the search for economical alternatives for activated carbon production, the objective of this work was to characterize the jute fiber (in natura) and to produce and characterize the resulting activated carbon.
Immediate analyses were carried out in order to determine the moisture content as well as the ash, volatile and fixed carbon contents. The analyses followed norm D1762-84 reviewed in 1985 (ASTM), which addresses such analyses for char and wood substances.
The activated carbon was characterized by gas adsorption (N 2 at 77K). All samples were degassed under vacuum at 150°C, and a gas adsorption analyzer (Quantachrome Corp. model Autosorb-1MP) was used to measure the pore distribution, pore volume and specific surface area. Data from the adsorption isotherms for nitrogen at 77K were used to determine the specific surface area according to the method proposed by Brunauer, Emmet and Teller (Rouquerol 1994, IUPAC 1985. The micropore (pore diameter narrower than 2.0nm) area was obtained by applying the t-plot method. The total pore volume was determined by converting the volume adsorbed at the saturation point P/P o ≈ 0.99 into liquid volume, whilst the micropore volume was calculated at the point of interception of the linear region of the t-plot after saturation of the micropores.

CARBONIZATION AND ACTIVATION OF JUTE FIBER
The raw material was submitted to carbo nization followed immediately by activation in the same run. Samples 190mm long were placed within a quartz reactor capable of accommodating 50g.
The experimental set-up is shown in Figure  1. Heat was supplied by an electrically heated 630mm vertical furnace (1) to a 700mm long 60mm diameter quartz reactor (2). Temperature and heating rates were controlled (3). A flask (9) was heated by resistances (8) connected to a power source (11). Temperatures were measured by thermocouples inside (4) and outside (5) the reactor as well as inside the flask (10). The whole set-up was placed under an exhausting system (6) to vent off gases. The N 2 flow rates to carry steam were previously tested and defined for all experiments: 200mL.min -1 of N 2 carried 0.17mL.min -1 of water (activation steam), 300mL.min -1 of N 2 carried 0.33mL.min -1 of water and 400mL.min -1 of N 2 carried 0.50mL.min -1 of water. Figure 2 shows the temperature profile selected for the production of steam activated jute fiber.
Some preliminary runs were carried out to determine the experimental parameter values. Temperatures below 600°C (T2) produced activated carbons with too low specific surface area. Above 800°C the material was totally or partially degraded. Within these limits the following conditions were specified.
All analyses employed samples of 5mg and final temperature of 950°C. The TG curves of in natura jute fibers with inert atmosphere, synthetic air and CO 2 are presented in Figures 3A, 3C and 3E, respectively. Their respective first derivative thermogravimetry (DTG) curves are presented in Figures 3B, 3D and 3F. Figures 3A, 3C and 3D show that the heating rate did not affect the mass losses, indicating they depend only on temperature. Jute fiber degradation starts at about 260°C and proceeds fast with temperature rise. This behavior was also noticed by Cao et al. 2013, Monteiro et al. 2012and by Banerjee and Dastidar 2005 Comparing the DTG curve using inert gas ( Figure 3B) with two other gases ( Figures 3D and  3F), one observes that the first peak (9.5% mass loss) around 80°C is common to both situations, which is due to the loss of adsorbed water in fiber (intra and intermolecular dehydration reactions); the second peak is also common, accounting for the major organic material loss (cellulose and hemicellulose), between 330°C and 380°C. However the third peak which appears when using O 2 , indicating the loss of material due to an oxidizing reaction involving the lignin. For all situations, i.e., for all types of gases, all material loss occurs below 600°C.

CHARACTERIZATION OF ACTIVATED CARBON FROM JUTE FIBER
Jute fiber (in natura) presented moisture content of 9%, volatile content of 84%, ash content of 9% and fixed carbon content of 7%. In order to grant total volatile elimination, the heating rate must be low during the carbonization period, which also boosts pore formation. Table III shows the variables investigated and the answer for yield (ɳ) and specific surface area (SSA) for the jute fiber carbonization and activation with steam. The data in Table III indicate a mass based average yield of 18% and average specific surface area of 470 m 2 .g -1 . The highest the variable levels, the lowest the yield, and the highest the SSA. The low yield can be explained by the elimination of the large amount of volatile matter present in the raw material (carbonization) as well as by the reaction between fixed carbon and steam (activation). The yields obtained varied inversely with temperature. These results are in agreement with the ones from Asadullah et al. 2010.
The specific surface values vary proportionally with temperature, with maximum value (730 m 2 .g -1 ) at 700°C, when starting the destruction pore wall. Figure 4 shows the N 2 (77K) adsorption/desorption isotherms of activated carbon from jute fiber. 2nm (2.95nm). All other samples had average pore diameter narrower than 2nm (according to IUPAC micropores). It is possible to conclude that the activated carbon obtained from the jute fiber is essentially microporous.
The increase of SSA is mostly related to wall destruction, turning micropores into mesopores, and not to the creation of new pores. This phenomenon has been reported in previous investigations (Phan et al. 2006, Asadullah et al. 2014. Figure 5A shows pore size distributions of two samples (JFA1 and JFA2) and Figure 5B shows pore size distributions of JFA6 and JFA8, all obtained according to the Non-Local Density Functional Theory (NLDFT). Sample JFA1 exhibits just one peak, which is inside the micropore region (less than 2nm), confirming the data from the N 2 isotherms, coherent with the 98% of micropores in the sample. Sample JFA2 also contains a large presence of micropores (diameter narrower than 5nm). Samples JFA6 and JFA8 have similar behavior, with large incidence of micropores and two other smaller peaks in the mesopore region, in agreement with the data from Table IV, with area composed of 93% of micropores and only 7% pores with diameter wider than 5nm. The percentage of volume in the micropore region of these samples is 85%.

CONCLUSIONS
The thermogravimetric analysis heating rate did not affect the mass losses, indicating they depend only on temperature. The major portion of the volatile matter was removed within the range 260°C -430°C. Above 600°C there is no significant mass loss. For the inert and carbonic gas atmospheres the thermogramimetric analysis indicates the presence of two peaks of mass loss while for the synthetic air atmosphere it indicates the presence of three peaks of mass loss.
A mass based average yield of 18% and an average specific surface area of 470 m 2 .g -1 was observed. The activated carbon from jute fiber structure characterization indicates a microporous The isotherms of Figure 4 show similar trend except JFA8 that presents an increase in the adsorption capacity at higher relative pressures. There was no hysteresis, which indicates no capillary condensation, typical of macro and meso pores. Sample JFA1 exhibited the lowest adsorption capacity, as also obtained by Asadullah et al. 2014. Table IV shows that activated carbon from jute fiber is widely microporous. Around 95% of SSA belongs to micropores and 90% of the pore volume belongs to micropores. Only sample JF8 showed an average pore diameters larger than material with minor presence of mesopores, although this may be changed by activation conditions. Mild temperatures and mild steam flow rates produce low specific surface areas. The porous structure is made up of micropores, totalling 92% of total volume. It was also observed that extreme activation times and gas flow rates result in higher surface areas, but still with high concentration of micropores. Therefore the jute fibers proved to be a good and versatile precursor of activated carbon.