Fuel 88 (2009) 1926

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

HEMP-derived activated carbon bers by chemical activation with phosphoric acid

J.M. Rosas, J. Bedia, J. Rodrguez-Mirasol *, T. Cordero
Chemical Engineering Department, School of Industrial Engineering, University of Mlaga, Campus de El Ejido s/n, 29013 Mlaga, Spain

a r t i c l e

i n f o

a b s t r a c t
Activated carbon bers were prepared by chemical activation of hemp bers with phosphoric acid at different carbonization temperatures and impregnation ratios. Surface properties of the activated carbons bers were signicantly inuenced by the activation temperature and the impregnation ratio. An increase of either of these parameters produced a high development of the porous structure of the bers. Activated carbon bers with apparent surface area of 1350 m2/g and mesopore volume of 1.25 cm3/g were obtained at 550 C with an impregnation ratio of 3. The activated carbon bers presented a high oxidation resistance, due to the presence of phosphorus compounds on the carbon surface. The oxidation resistance results suggest that COPO3 and mainly CPO3 and CP groups act as a physical barrier, blocking the active carbon sites for the oxidation reaction. 2008 Elsevier Ltd. All rights reserved.

Article history: Received 14 May 2008 Received in revised form 31 July 2008 Accepted 5 August 2008 Available online 30 August 2008 Keywords: Activated carbon bers Porosity Chemical activation Phosphoric acid Oxidation resistance

1. Introduction Hemp is one of the most productive and useful plants known. It grows quickly without any great need of pesticides in most locations and climates with only moderate water and fertilizer requirements, becoming a valuable and environmentally friendly crop. Hemp has many industrial applications such as the production of paper, textiles, building materials, food, medicine, paint, detergent, varnish, oil, ink, and fuel [1]. However, hemp transformation generates a high proportion of waste. Specically, the textile industry employs less than 50%w of the hemp, producing signicant amounts of waste [2,3]. Spain produced about 40,000 ton of hemp residue only from the textile industry (2004). Development of recycling processes of such biomass waste is generating great interest, and the production of activated carbons could be an appropriate solution. The world demand of activated carbon is steadily increasing due to its well-known extensive use as adsorbent for purication and separation in many processes. Activated carbons can be manufactured from many carbonaceous precursors, but the most commonly used raw materials are wood, coal, coconut shells and some polymers [48]. The use of several lignocellulosic by-products, such as nutshells, fruit stones, lignin, sugarcane and sawdust from a few rapid growing wood species, as renewable precursors for low cost activated carbon production, is being widely studied [923]. The use of hemp residues to produce activated carbon is very feasible and presents the advantage of the potential revalori* Corresponding author. Tel./fax: +34 952132886. E-mail address: mirasol@uma.es (J. Rodrguez-Mirasol). 0016-2361/$ - see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2008.08.004

zation of a residual material. In this respect, preparation of activated carbon bers by physical activation with steam and chemical activation with ZnCl2 of hemp bers have already been reported in the literature [2,3]. The aim of this work is the preparation of activated carbon bers by chemical activation of hemp bers with phosphoric acid, analyzing the inuence of the activation temperature and the impregnation ratio on the activated carbon porous structure and surface chemistry. The oxidation resistance of the activated carbon bers was also studied, analyzing the effect of surface phosphorus compounds, retained in the matrix structure upon activation, on the oxidation reaction. 2. Experimental 2.1. Preparation of carbon and activated carbon bers The starting materials were hemp bers supplied by Alsativa (Sociedad Cooperativa Agraria Andaluza del Camo, Prtugos, Granada). Hemp was previously cleaned from leaves and tops and kept at room temperature. The bers were cut in fractions of around 3 cm long. The precursor was impregnated by incipient wetness with 85% (w/w) H3PO4 aqueous solution at room temperature and dried for 24 h at 60 C in a vacuum dryer. The impregnation ratio, R, (H3PO4/ precursor mass ratio) varied from 1 to 3. The impregnated hemp bers were activated under continuous N2 ow (150 cm3 STP/min), in a conventional tubular furnace. The activation temperature was reached at a heating rate of 10 C/min and maintained for 2 h. Different activation temperatures within

20

J.M. Rosas et al. / Fuel 88 (2009) 1926

400550 C range were studied. The activated samples were cooled inside the furnace maintaining the N2 ow, and then were washed with distilled water at 60 C until neutral pH and negative phosphate analysis in the eluate [24]. The resulting activated carbon bers were dried at 100 C and weighted to determine the yield of the activation process (weight of activated carbon related to weight of raw material, in dry basis). The activated carbon bers were denoted by the letter F (bers) followed by a number corresponding to the impregnation ratio and by a second number representing the activation temperature in degrees Celsius. Some carbon bers were obtained by carbonization of hemp bers under N2 ow at different temperatures without impregnation of the activation agent. The carbon bers thus obtained were denoted by F followed by the carbonization temperature in degrees Celsius. 2.2. Characterization The ultimate analysis of the samples was performed in a Leco CHNS-932 system, being the oxygen content calculated by difference. The ash content of the sample was calculated following the ASTM procedure [25]. The porous structure of the carbon bers was characterized by N2 adsorptiondesorption at 196 C, performed in an Omnisorp 100cx equipment (Coulter) and by CO2 adsorption at 0 C, carried out in an Autosorb-1 apparatus (Quantachrome). Samples were previously outgassed for at least 8 h at 150 C. From the N2 isotherm, the apparent surface area (ABET) was determined applying the BET equation [26]. The as method allows to obtain the values of the so-called external surface area (As), namely, the surface area associated to the non-microporous structure, the micropore volume (Vs), and the specic surface area (as), obtained by the high resolution method proposed by Kaneko et al. [2729], using a non-porous carbon black (Elftex-120) as standard [30]. The mesopore volume (Vmes) was determined as the difference between the adsorbed volume of N2 at a relative pressure of 0.95 and the micropore volume Vs [27]. From the CO2 adsorption data, the narrow micropore volume (VDR) and the apparent surface area (ADR) were calculated using the DubininRadushkevich equation [31]. The HorvathKawazoe [32] and the BarrettJoyner and Halenda (BJH) [33] methods applied to the N2 adsorption data were used to obtain the micropore and mesopore size distributions, respectively. The surface chemistry of the samples was analyzed by temperature-programmed desorption (TPD) and X-ray photoelectron spectroscopy (XPS). TPD proles were obtained in a custom quartz xed bed reactor placed inside an electrical furnace. The samples were heated from room temperature up to 900 C at a heating rate of 10 C/min in helium ow (200 cm3 STP/min). The amounts of CO and CO2 desorbed from the samples were monitored with nondispersive infrared (NDIR) gas analyzers (Siemens ULTRAMAT 22). Xray photoelectron spectroscopy (XPS) analyses of the samples were obtained using a 5700C model Physical Electronics apparatus with Mg Ka radiation (1253.6 eV). For the analysis of the XPS peaks, the C1s peak position was set at 284.5 eV and used as reference to locate the other peaks [3436]. The tting of the XPS peaks was done by least squares using GaussianLorentzian peak shapes. The surface texture of the samples was characterized by scanning electron microscopy (SEM). Scanning electron micrographs were obtained using a JEOL JSM-840 instrument, working at a high voltage of 2025 kV. Non-isothermal thermogravimetric analyses were carried out in a gravimetric thermobalance system, CI electronics. The thermobalance automatically measures the weight of the sample and the temperature as a function of time. Experiments were carried out in inert atmosphere (N2), and in air atmosphere, for a total ow rate of 150 cm3 (STP)/min, employing sample mass of approxi-

mately 10 mg. The sample temperature was increased from room temperature up to 900 C at a heating rate of 10 C/min. 3. Results and discussions Table 1 shows the analytical and structural properties of the hemp bers used for the preparation of activated carbon bers. The ash content of the raw material is relatively low. Hemp bers present a non-porous structure as indicated a negligible adsorption of N2 at 196 C and CO2 at 0 C. The evolution of the surface chemistry and the porous structure with the impregnation ratio of phosphoric acid and the activation temperature was studied for hemp bers. Table 2 shows the yield values and the ultimate analysis of the carbon and activated carbon bers obtained. Oxygen content was determined by difference. However, the oxygen content values shown in this table are probably overestimated as they may include the amounts of P and Si in the carbons (detected by XPS). The yield values obtained for the hemp-derived carbon bers prepared at different temperatures are similar to those reported in the literature for other lignocellulosic materials pyrolyzed at comparable temperatures [37,38] (from about 25% for a pyrolysis temperature of 450 C to about 12% for 800 C). The yields for the activated carbon bers were almost twice than those for the carbon bers at the same treatment temperature. The phosphoric acid limits the formation of tars during the carbonization treatment, increasing the yield of the remaining solid product [6]. Yield values from 41% to 50% were obtained for the activated carbon bers at different impregnation ratios and activation temperatures, similar to those obtained from other biomass natural waste [8,18] and slightly higher than those reported for hemp bers with different activation agents [2,3]. The yield values decrease slightly with the activation temperature and the impregnation ratio studied, as a consequence of a deeper dehydration of the carbonaceous structure of the precursor. The amount of oxygen

Table 1 Analytical and structural characteristics of the hemp bers used for the activated carbon preparation Fiber Ash content (% db) Ultimate analysis (% daf) C H N O Major components Cellulosea Hemicellulosea Lignina
a

3.0 43.17 5.21 0.63 50.99 67.0 16.1 3.3

Data from Ref. [1].

Table 2 Yield values and ultimate analysis of the hemp-derived carbon and activated carbon bers Yield (% db) F450 F550 F800 F1-450 F2-450 F3-400 F3-450 F3-500 F3-550 24.5 22.0 12.9 49.8 44.5 42.7 42.2 41.9 41.1 C (% daf) 81.0 84.0 96.9 64.6 78.3 83.8 80.2 77.9 76.6 H (% daf) 2.6 3.6 1.5 2.9 2.6 3.4 3.0 2.2 2.4 N (% daf) 1.7 1.8 0.9 0.4 0.2 0.3 0.2 0.4 0.8 O (% daf) 14.7 10.6 0.7 32.1 18.9 12.5 16.6 19.5 20.2

J.M. Rosas et al. / Fuel 88 (2009) 1926

21

3.1. Porous structure Fig. 2 shows the N2 adsorptiondesorption isotherms at 196 C for different hemp-derived carbon and activated carbon bers. Carbon ber isotherms present an almost horizontal plateau starting at very low relative pressure, corresponding to type I isotherms, characteristic of very narrow microporous solids. An increase in the carbonization temperature produces an increase of the N2 adsorbed at low relative pressures, as a consequence of a higher development of the microporous structure. The small hysteresis loop observed for the carbon bers obtained at 800 C suggests a certain development of narrow mesoporosity. The activated carbon bers show signicantly higher N2 adsorption volumes than those obtained for the carbon bers prepared at the same temperatures. This conrms the effect of the phosphoric acid activation in porosity development of hemp bers. The activated carbon bers present a modied type I isotherm, corresponding to a well-developed microporous structure with a signicant contribution of mesoporosity. The extent of the porous development of the activated carbons depends on both the activation temperature and the impregnation ratio. The increase of the activation temperature produces a progressive porosity development of the activated carbon bers, for the range of temperatures studied. The increase of the relative amount of activation agent generates a higher development of porosity, widening the porous structure of the samples. The hemp-derived activated carbon bers show a signicantly higher contribution of mesoporosity than carbon bers, as it reveals the higher N2 volume adsorbed at medium-high relative pressures and the larger hysteresis loops for the activated carbon bers, an indication of the presence of mesopores. Fig. 3 represents the as plot of the hemp-derived carbon bers obtained at 550 C and some activated carbon bers. The amount of N2 adsorbed is represented versus as, the ratio of the amount adsorbed by a non-porous standard (in this case, Elftex-120) at each relative pressure and at 0.4 relative pressure [42]. The linear multilayer region (as > 1.2, solid line) is related to the adsorption on the non-microporous surface of the carbon. The extrapolation of this linear multilayer region at as = 0 and the corresponding slope provides the total micropore volume and the external surface area, respectively. The slope of a straight line that ts as values lower than 1 and goes through the origin of coordinates (dotted line in Fig. 3) provides the specic surface area, as, as proposed by Kaneko et al. [2729]. Fig. 3 shows a higher contribution of microporosity for all the activated carbon bers. The micropore volume decreases with the impregnation ratio favoring the presence of external surface area, as shown by the higher value of the slope of the

Fig. 1. Thermogravimetric analysis in inert atmosphere for hemp bers, impregnated hemp bers (R = 1) and F1-450 activated carbon bers.

in the activated bers increases progressively with activation temperature [7,8] and decreases with the impregnation ratio [3]. Thermal decomposition of hemp ber was studied by TG analysis. Fig. 1 shows the weight-loss curve for hemp bers, hemp bers impregnated with phosphoric acid at an impregnation ratio of 1 and F1-450 activated carbon bers. Hemp bers show a small weight loss at temperatures lower than 200 C that can be attributed to release of moisture. Between 200 and 500 C a steep decrease of mass is observed, due to decomposition of the major biopolymers of the bers. At temperatures higher than 500 C, a lower weight loss rate is observed, as a consequence of a deeper degradation of the more stable volatile matter. Hemicellulose mainly decomposes from about 200300 C, cellulose pyrolysis takes place between 300 and 400 C and lignin degradation occurs under the whole range of temperatures [39,40]. As mentioned above, the main weight loss takes place from 300 to 400 C and represents 60% of the hemp ber mass, corresponding with the amount of cellulose reported in the literature [1] for this ber (Table 1). The presence of H3PO4 produces a considerable effect on the thermal decomposition of hemp ber. At lower temperatures (T < 315 C) the mass loss in hemp impregnated bers is higher than for the hemp bers. Phosphoric acid catalyzes the hydrolysis of the glycosidic linkages in hemicellulose and cellulose and it cleaves aryl ether bonds in lignin, obtaining many transformations that include dehydration, degradation and condensation. These reactions promote the release of H2O, CO, CO2 and CH4 at low temperatures [6]. The weight loss for the impregnated hemp bers at temperatures higher than 315 C is signicantly lower than for hemp bers. As it was commented previously, phosphoric acid restricts the formation of tars increasing the yield of the remaining solid product. The activation yield for sample F1-450 (shown in Table 2) is very similar to that obtained in the TG analysis for the impregnated hemp ber at 450 C, taking into account that the washing process yield represents approximately 65%. At 550 C a change in the slope of the weight-loss curve corresponding to the impregnated hemp ber can be observed, probably associated to volatilization of carbonoxygen complexes generated by the activation process, and to a lesser extent, to decomposition of phosphorous-compounds (as P2O5), produced by the reaction of phosphoric acid with the carbon matrix [41]. The weight-loss for activated carbon ber F1-450 is almost negligible at temperatures lower than 600 C, from which the decomposition of carbonoxygen groups, formed during the activation process, begins. This behavior is observed in the TPD analysis and discussed in a next section.

Fig. 2. N2 adsorptiondesorption isotherms for different hemp-derived carbon and activated carbon bers.

22

J.M. Rosas et al. / Fuel 88 (2009) 1926

Fig. 3. as curves for different hemp-derived carbon and activated carbon bers.

Table 3 Structural characteristics of the hemp-derived carbon and activated carbon bers N2 isotherm ABET (m2/g) F450 F550 F800 F1-450 F2-450 F3-400 F3-450 F3-500 F3-550 11 187 452 723 1023 1053 1141 1160 1355 as (m2/g) 12 193 471 770 922 933 1034 1056 1257 Vs (cm3/g) 0.004 0.082 0.191 0.323 0.302 0.197 0.227 0.235 0.293 As (m2/g) 1 6 48 6 517 684 729 797 812 Vmes (cm3/g) 0.002 0.008 0.052 0.008 0.699 1.140 1.239 1.248 1.253 CO2 isotherm ADR (m2/g) 291 362 550 499 445 348 510 587 686 VDR (cm3/g) 0.111 0.138 0.209 0.190 0.169 0.133 0.194 0.223 0.261

multilayer region and the decrease of the value of the back-extrapolation of the multilayer region for higher impregnation ratio. However, the micropore volume increases with the activation temperature at the same impregnation ratio. The characteristic parameters of the porous structure of the hemp-derived carbon and activated carbon bers, obtained from the N2 and CO2 isotherms, are summarized in Table 3. The values of apparent surface area obtained by the BET and high resolution as methods are very similar. There are some differences between the micropore volumes obtained from the N2 and CO2 isotherms. N2 adsorption is carried out within the entire range of relative pressures and cooperative lling of wider micropores takes place, while CO2 adsorption at relative pressures smaller than 0.03 allows only primary lling of narrow micropores. The apparent surface area and micropore volume values obtained for the carbon bers from the N2 and CO2 adsorption isotherms, with ADR > ABET and VDR > Vs, are characteristics of solids with a narrow microporosity. This behavior can be explained by diffusional limitations of N2 in the inner of the narrower micropores during adsorption at 196 C, without reaching the equilibrium at the regular analysis times. Despite CO2 molecule presents a similar diameter to N2, the much higher adsorption temperature for CO2 (0 C) increases the diffusion rate in narrower micropores, which facilitates reaching the adsorption equilibrium in shorter times. The opposite behavior observed for activated carbon bers (ADR < ABET and VDR < Vs) reveals the presence of wider microporosity in these solids [9]. An increase of the impregnation ratio with phosphoric acid results in a higher development of porosity, mainly of larger pores. Chemical activation with phosphoric acid produces the reaction between some organic species and the acid, forming phosphate and polyphosphate bridges, which are responsible for connecting and crosslinking the biopolymer fragments. The phosphate groups

are inserted in the carbon matrix, separating the organic species. This generates an expansional process that, after removal of the acid, leaves the matrix in an expanded state, with a high pore development structure [6,14,1820]. A higher presence of phosphoric acid generates higher formation of phosphate esters, which produce a higher expansion or dilation of the porous structure [6,19]. The activated carbon bers present very high values of external area and mesopore volume, with a very wide porous distribution. The activation of hemp bers proceeds at relatively low temperature 400500 C, achieving a signicant development of micro and mesoporosity compared to the carbon bers obtained at similar temperatures. As a reference, hemp-derived activated carbons bers with ABET about 1350 m2/g and As of 800 m2/g are obtained at 550 C with an impregnation ratio of 3. Activated carbon bers obtained from different precursors as jute and coconut [21] and non-cellulosic polymers [43] at similar experimental conditions as those used in this work present higher micropore volume and signicantly lower mesopore volume than those observed for hemp-derived activated carbon bers. In a previous work [44], we have prepared activated carbon monoliths from hemp cane. The activated carbon monoliths presented higher micropore volume and lower mesopore volume than these activated carbon bers. Hemp bers present a considerable potential for the dilation, given that the content of cellulose in these bers is signicantly higher than for amorphous polymers (lignin and hemicellulose). The results seem to indicate that the presence of crystalline cellulose produces both micro and mostly mesopores [6]. Fig. 4 shows the micropore size distribution of F550 carbon ber and different activated carbon bers. F550 shows a relatively narrow micropore size distribution with a maximum at 7.5 . Activation of the hemp bers with phosphoric acid produces a widening of the micropore size distribution. An increase of both parameters treatment temperature and impregnation ratio results in a higher and broader micropore size distribution [6,19]. Fig. 5 represents the mesopore size distribution for different hemp-derived activated carbon bers. The carbon bers obtained do not present mesoporosity. Activation of hemp bers with phosphoric acid produces mainly narrow mesopores with size from about 3040 , with a low contribution of mesopores with size larger than 40 . An increase of the impregnation ratio or the treatment temperature reduces the presence of narrow mesopores in favor of those of larger size. 3.2. Surface chemistry The surface chemistry of the samples was analyzed by X-ray photoelectron spectroscopy (XPS) and temperature-programmed

Fig. 4. Micropore size distribution for different hemp-derived carbon and activated carbon bers.

J.M. Rosas et al. / Fuel 88 (2009) 1926

23

Fig. 5. Mesopore size distribution for different hemp-derived carbon and activated carbon bers.

Fig. 6. XPS P 2p spectrum of different hemp-derived carbon and activated carbon bers.

desorption (TPD). To complete the characterization of the samples a morphology study of the surfaces was carried out by scanning electron microscopy (SEM). The surface element distribution of the ber ashes and the activated carbon bers was obtained by XPS analysis. Mass surface concentrations of the samples have been obtained by numerical integration of the peaks and are reported in Table 4. The ber ashes show a very signicant amount of calcium, silicon, phosphorus and magnesium. The main elements found over the activated carbon ber surfaces were carbon and oxygen with lower amounts of phosphorus and silicon. Other inorganic components found in the ber ashes are not presented in the activated carbon bers due to their removal by the washing step after the phosphoric acid impregnation followed by the activation process. A signicant residual amount of silicon and phosphorus is observed on the surface of the activated carbon bers. Silicon, as orthosilicic acid, is a component of the hemp ber that provides stability to the ber structure. Nitrogen is also detected, but at very low concentration. The amount of phosphorus obtained by XPS slightly increases with the increase of the activation temperature from 400 to 500 C, decreasing at higher temperatures (550 C) as a consequence of volatilization of phosphorus compounds. Montan et al. [41] reported that P2O5 melts and vaporizes around this temperature (580 C). However, the content of phosphorus practically remains the same with increasing impregnation ratio [8]. Fig. 6 depicts the XPS P 2p spectra of different hemp-derived carbon and activated carbon bers. The phosphorus spectrum was deconvoluted using two doublet peaks with an area ratio of 0.5 and a separation between peaks of 0.84 eV [45]. The P 2p spectrum of the F450 carbon bers presents a signicant contribution of a peak centered at 134.2 0.2 eV corresponding to meta and polyphosphates of calcium, potassium and sodium [4548]. For activated carbon bers the bands are shifted to lower binding energies, probably due to the absence of inorganic metals such as calcium and potassium. The P 2p XPS spectrum of the hemp ber activated at 400 C with an impregnation ratio of 3, F3-400, shows a band with a main

peak at a binding energy around 133.9 0.2 eV, characteristic of pentavalent tetracoordinated phosphorus (PO4), as in polyphosphates and/or phosphates [6,8,4648]. These phosphates and/or polyphosphates groups are responsible of the dilation process during activation. Wu and Radovic assigned this band at binding energy of about 134.0 eV to the COPO3 groups [48]. An increase of the activation temperature up to 550 C shifted the binding energy of the main peak to a value of 133.2 0.2 eV characteristic of P atom bonded to one C atom and three O atoms, as in CPO3 groups [44,48]. Also, a higher contribution of these CPO3 surface groups is observed by increasing the impregnation ratio, given that the main XPS peak shifts to lower binding energies. A small band at around 132.1 eV is observed for the activated carbon bers obtained at high activation temperature, which can be associated to reduced phosphorus compound as CP [45]. The presence of a small amount of phosphorus as P2O5 (BE = 135.6 eV) should not be ruled out [33,47]. TPD technique is employed to characterize the carbonoxygen surface groups, whose nature and amount are related with the precursor and the activation process [4]. The groups of acidic character (carboxylic, lactonic) evolve as CO2, whereas the non-acidic (carbonyl, ether, quinone) and phenol groups give rise to CO. Anhydride surface groups evolve as both CO and CO2 [36]. Fig. 7 shows the CO evolution from TPD analysis of different carbon and activated carbon bers. F-450 shows a low amount of CO evolved mainly between 600 and 800 C, compared to that corresponding to activated carbon bers. Chemical activation with phosphoric acid produces an increase in the amount of oxygen surface groups, which evolve as CO at high temperatures (700 900 C). Based on the work of Wu and Radovic [48] and on our previous results with hemp-derived activated carbon monoliths [44], the CO evolution at high temperatures can be related to the formation P-surface groups during the phosphoric acid activation, as relatively weak metaphosphates, COPO3 and CPO3 surface groups. The activated carbon bers show most of the CO evolution at high temperatures, although a lower amount is desorbed at temperatures below 700 C, due to decomposition of anhydride, phenol

Table 4 Mass surface concentration (%) determined by XPS quantitative analysis of hemp ber ash and different hemp-derived activated carbon bers C 1s Fiber ash F1-450 F2-450 F3-400 F3-450 F3-500 F3-550 6.46 74.27 74.95 73.44 72.48 71.80 71.80 O 1s 44.89 17.62 17.42 17.36 17.74 17.83 19.80 N 1s 1.37 0.36 0.33 0.53 0.63 1.02 P 2p 8.37 5.40 5.39 5.60 5.82 5.96 3.30 Ca 2p 18.22 S 2p 2.08 Mg 2s 8.32 K 2p 2.19 Si 2p 6.19 1.34 1.87 3.27 3.63 3.79 4.08 Na 1s 1.17 Fe 2p 2.12

24

J.M. Rosas et al. / Fuel 88 (2009) 1926 Table 5 CO and CO2 evolved amount from TPD analyses of hemp-derived carbon F-450 and activated carbon bers CO (mmol/g) F-450 F1-450 F2-450 F3-400 F3-450 F3-500 F3-550 2.61 6.08 5.04 3.54 4.40 5.63 5.86 CO2 (mmol/g) 0.76 0.31 0.29 0.30 0.22 0.18 0.95

Fig. 7. CO evolution during the TPD for different hemp-derived carbon and activated carbon bers.

and ether groups. At activation temperatures as high as 550 C the formation of CPO3 groups (probably from COPO3 groups) seems to be favored, producing a signicant increase of anhydride, phenol and ether surface groups that decompose during the TPD at lower temperatures. Puziy et al. [7,8] and Fu et al. [46] also found for phosphoric acid activation of carbonaceous materials that an increase of the activation temperature generates higher amounts of phenol and ether surface groups. The amount of CO2 evolved (Fig. 8) from the carbon and activated carbon bers is signicantly lower than that corresponding to CO evolution, indicating a lower concentration of carboxyl, lactonic and anhydride surface groups. F-450 carbon displays different CO2 evolution bands between 200 and 800 C, corresponding to decomposition of the different acidic surface groups. Activated carbon bers prepared at temperatures lower than 550 C show a low CO2 evolution at temperatures below 700 C compared to that of the carbon ber. However, a signicant CO2 peak at similar temperatures where the maximum CO evolution for these activated carbon bers takes place (T > 800 C) is observed, which can be associated to the decomposition of COPO3 (or CPO3) surface groups. Also, this CO2 evolution at temperatures higher than 800 C may be caused by secondary reactions [49,50]. Diffusion of the evolved gases is rather slow in narrow micropores and a free CO (higher for the activated carbons) may react with surfacebound oxygen to produce CO2. Table 5 summarizes the amounts of CO and CO2 desorbed during the TPD. The amount of CO evolved increases with the activation temperature, but decreases with increasing the impregnation ratio. As aforementioned, little amount of acidic groups that evolve as CO2 are desorbed during the TPD, being this amount higher for the car-

bon bers than for the activated carbon bers, obtained at similar temperature. The amount of CO2 groups desorbed decrease with the impregnation ratio. Similar amounts of CO2 desorbed during the TPD are found for other lignocellulosic precursors at similar activation conditions [24,44]. However, the amount of CO groups evolved is signicantly higher for hemp-activated carbon bers. SEM micrographs of hemp, carbon and activated carbon bers are illustrated in Fig. 9ae . Fig. 9a shows the hemp ber used as raw material in this work. The bers are in form of bundles with sizes between 10 and 60 lm. The thermal treatment at different temperatures did not modify the size of the ber bundles signicantly, as can be observed in Fig. 9b for F-550. The presence of phosphoric acid during the activation process produced an increase of the ber bundles size. This augmentation of the size of the ber bundles is more pronounced with the increase of the impregnation ratio from 25 to 100 lm, as can be observed in Fig. 9cd for an activation temperature of 450 C and impregnation ratio from 1 to 2, respectively. It is important to point out that the use of a high value of the impregnation ratio produces an agglomeration of the ber bundles, obtaining activated carbon with a monolith-like form [43], as shown in Fig. 9e for activated carbon bers obtained at 450 C and impregnation ratio of 3 (F3-450), with ber bundles size of about 330 lm (Fig. 9e). The phosphoric acid impregnated hemp bers liberates tar during the rst step of impregnation and activation at low temperatures. This tar acts as a binder, lling the inter-ber space, which is carbonized during the activation process as temperature increases, producing monolith-like activated carbon materials with mechanical rigidity and a developed porous structure [51]. No agglomeration of bers was observed in the absence of phosphoric acid, Fig 9b [43]. 3.3. Oxidation resistance Fig. 10 compares the non-isothermal oxidation resistances proles of the carbon ber F-450 and some activated carbon bers. Activated carbon bers present a higher oxidation resistance than that observed for the carbon bers, despite activated carbon bers (F1-450 and F3-450) were treated at the same temperature as that of the carbon bers (F-450) and they exhibit a more developed porous structure. The oxidation of the activated carbon bers begins to be signicant at temperatures at least 100 C higher than that of F-450 carbon. At the same impregnation ratio, an increase of the activation temperature from 400 to 450 C produces a higher oxidation resistance of the bers. However, a further increase of the activation temperature from 450 to 550 C results in a slightly increase of the oxidation reactivity at lower temperatures. This behavior can be associated to the presence of phosphorus compounds on the carbon surface, as shown by the XPS analysis (Fig. 6). Phosphorus is an important inhibitor of the oxidation reaction. The impregnation of carbons with different phosphorus compounds, such as H3PO4 [52], POCl3 [53] and organo-phosphorus compounds [54] has shown to increase the oxidation resistance. Some studies relate the enhancement of the oxidation resistance with the presence of COPO3 and CPO3 groups [48,52]. These

Fig. 8. CO2 evolution during the TPD for different hemp-derived carbon and activated carbon bers.

J.M. Rosas et al. / Fuel 88 (2009) 1926

25

Fig. 9. SEM micrographs of as received hemp bers, bar length: 50 lm (a); hemp-derived carbon F-450, bar length: 50 lm (b); hemp-derived activated carbon bers F1-450, bar length: 10 lm (c); hemp-derived activated carbon bers F2-450, bar length: 50 lm (d); hemp-derived activated carbon bers F3-450, bar length: 50 lm (e).

er contribution of CPO3 and CP surface groups (XPS bands at low binding energies, 133.2 and 132.1 eV, respectively, in Fig. 6), F3450 and F3-550, present a higher oxidation resistance than those with higher contribution of COPO3 group and P2O5 (XPS bands at high binding energies, 133.9 and 135.6 eV, respectively, in Fig. 6), F3-400 and F1-450. These results suggest that both CPO3 and CP surface groups are mainly the responsible for the high oxidation resistance of the hemp-derived activated carbon bers F3-450, although it should not be ruled out the inhibitor effect of COPO3 surface group. These P surface groups may act as a physical barrier, blocking the active carbon sites for the oxidation reaction. 4. Conclusions
Fig. 10. Non-isothermal oxidation resistances proles of hemp-derived carbon (F-450), and different activated carbon bers.

groups are suggested to act as a physical barrier, blocking the active carbon sites. Both COPO3 and CPO3 groups have been found on the hemp-derived activated carbon bers surface. However, there is a different contribution of these P surface groups in the activated carbon bers, depending on the activation conditions. It can be observed from Fig. 10 that the activated carbon with a high-

Activated carbon bers were prepared by chemical activation of hemp bers with phosphoric acid at different carbonization temperatures and impregnation ratios. Surface properties of the activated carbons bers were signicantly inuenced by the activation temperature and the impregnation ratio. The increase of either the relative amount of activation agent or the activation temperature generates a high development of the porous structure of the bers. Hemp-derived activated carbon bers with 1350 m2/g

26

J.M. Rosas et al. / Fuel 88 (2009) 1926 [21] Phan NH, Rio S, Faur C, Le Coq L, Le Cloirec P, Nguyen TH. Production of brous activated carbons from natural cellulose (jute, coconut) bers for water treatment applications. Carbon 2006;44:256977. [22] Diao Y, Walawender WP, Fan LT. Activated carbons prepared by phosphoric acid activation by grain sorghum. Bioresour Technol 2002;81:4552. [23] Toles CA, Marshall WE, Wartelle LH, Johns MM. Acid activated carbons from almond shells: physical, chemical and adsorptive properties and estimated cost of production. Bioresour Technol 2000;71:8792. [24] Gonzlez-Serrano E, Cordero T, Rodrguez-Mirasol J, Cotoruelo L, Rodrguez JJ. Removal of water pollutants with activated carbons prepared from H3PO4 activation of lignin from kraft black liquors. Water Res 2004;38:304350. [25] ASTM. Standard test method for total ash content of activated carbon. D286694; 2004. [26] Brunauer S, Emmett PH, Teller E. Adsorption of gases in multimolecular layers. J Am Chem Soc 1938;60:30919. [27] Kaneko K, Ishii C. Superhigh area determination of microporous solids. Colloids Surf 1992;67:20312. [28] Kaneko K, Ishii C, Ruike M, Kuwabara H. Origin of superhigh surface area and microcrystalline graphitic structures of activated carbons. Carbon 1992;30: 10751088. [29] Kaneko K. Determination of pore size and pore size distribution. I. Adsorbents and catalysts. J Membr Sci 1994;96:5989. [30] Gregg SI, Sing KSW. Adsorption, surface area and porosity. London: Academic Press; 1982. [31] Dubinin MM, Zaverina ED, Radushkevich LV. Sorption and structure of active carbons. I. Adsorption of organic vapors. J Phys Chem [URSS] 1947;21: 13511362. [32] Horvath G, Kawazoe K. Method for the calculation of effective pore size distribution in molecular sieve carbon. J Chem Eng Jpn 1983;16:4705. [33] Barret EP, Joyner LG, Halenda PP. The determination of pore volume and distribution in porous substances. I. Computations from nitrogen isotherms. J Am Chem Soc 1951;73:37380. [34] Biniak S, Szymanski G, Siedlewski J, Swiatkowski A. The characterization of activated carbons with oxygen and nitrogen surface groups. Carbon 1997;35: 17991810. [35] Desimoni E, Casella GI, Morone A, Salvi AM. XPS determination of oxygen containing functional groups on carbon-bre surfaces and the cleaning of these surfaces. Surf Interf Anal 1990;15:62734. [36] Figuereido JL, Pereira MFR, Freitas MMA, rfo JJM. Modication of the surface chemistry of activated carbons. Carbon 1999;37:137989. [37] Raveendran K, Ganesh A. Adsorption characteristics and pore-development of biomass-pyrolysis char. Fuel 1998;7:76981. [38] Nik-Azar M, Hajaligol MR, Sohrabi M, Dabir M. Mineral matter effects in rapid pyrolysis of beech wood. Fuel Process Technol 1997;51:717. [39] Yang H, Yan R, Chen H, Lee DH, Zheng C. Characteristics of hemicellulose, cellulose and lignin pyrolysis. Fuel 2007;86:17818. [40] Cordero T, Rodrguez-Maroto JM, Rodrguez-Mirasol J, Rodrguez JJ. On the kinetics of thermal decomposition of wood and wood components. Thermochim Acta 1990;164:13544. [41] Montan D, Torn-Fernndez V, Fierro V. Activated carbons from lignin: kinetic modelling of the pyrolysis of Kraft lignin activated with phosphoric acid. Chem Eng J 2005;106:112. [42] Carrott PJM, Roberts RA, Sing KSW. Adsorption of nitrogen by porous and nonporous carbons. Carbon 1987;25:5968. [43] Surez-Garca F, Martnez-Alonso A, MD Tascn J. Activated carbon bers from Nomex by chemical activation with phosphoric acid. Carbon 2004;42: 14191426. [44] Rosas JM, Bedia J, Rodrguez-Mirasol J, Cordero T. Preparation of hemp-derived activated carbon monoliths. Adsorption of water vapor. Ind Eng Chem Res 2008;47:128896. [45] Moulder JF, Stickle WF, Sobol PE, Bomben KD. In: Chastain J, King Jr RC, editors. Handbook of X-ray photoelectron spectroscopy. Eden Prairie, MN: Physical Electronics Inc.; 1995. p. 48725. [46] Fu R, Liu L, Huang W, Sun P. Studies on the structure of activated carbon bers activated by phosphoric acid. J Appl Polymer Sci 2002;87:225364. [47] Puziy AM, Poddubnaya OI, Ziatdinov AM. On the chemical structure of phosphorus compounds in phosphoric acid-activated carbon. Appl Surf Sci 2006;252:80368. [48] Wu X, Radovic LR. Inhibition of catalytic oxidation of carbon/carbon composites by phosphorus. Carbon 2006;44:14151. [49] Hall PJ, Calo JM. Secondary interactions upon thermal desorption of surface oxides from coal chars. Energ Fuel 1989;3:3706. [50] Boehm HP. Surface oxides on carbon and their analysis: a critical assessment. Carbon 2002;40:1459. [51] Nakagawa Y, Molina-Sabio M, Rodrguez-Reinoso F. Modication of the porous structure along the preparation of activated carbon monoliths with H3PO4 and ZnCl2. Micropor Mesopor Mater 2007;103:2934. [52] Labruqure S, Pailler R, Naslain R, Desbat B. Oxidation inhibition of carbon ber performs and C/C composites by H3PO4. J Eur Ceram Soc 1998;18: 19531960. [53] Dhami TL, Bahl OP, Awasthy BR. Oxidation-resistant carboncarbon composites up to 1700 C. Carbon 1995;33:47990. [54] Wu X, Pantano CGM, Radovic LR. Inhibition of catalytic oxidation by boron and phosphorus. In: Int carbon conference, Lexington, USA; 2001. p. 2.33.

of apparent surface area and 1.25 cm3/g of mesopore volume were obtained. At high impregnation ratios, phosphoric acid impregnated hemp bers liberate tar during the rst step of impregnation and activation at low temperatures. This tar acts as a binder, lling the inter-ber space, which is carbonized during the activation process as temperature increases, producing monolith-like activated carbon materials, with mechanical rigidity and a developed porous structure. The activated carbon bers presented a high oxidation resistance, due to the presence of phosphorus compounds on the carbon surface. The oxidation resistance results suggest that COPO3 and mainly CPO3 and CP groups act as a physical barrier, blocking the active carbon sites for the oxidation reaction. Acknowledgements The authors thank Ministry of Science and Education of Spain (DGICYT, Projects CTQ2006/11322 and PPQ2003-07160). JB acknowledges the assistance of the Ministry of Science and Education of Spain for the award of a FPI grant. References
[1] Batra SK. In: Lewin M, Pearce EM, editors. Handbook of bre chemistry. New York: Marcel Dekker Inc.; 1998. [2] Williams PT, Reed AR. Pre-formed activated carbon matting derived from the pyrolysis of biomass natural ber textile waste. J Anal Appl Pyrolysis 2003;70:56377. [3] Williams PT, Reed AR. High grade activated carbon matting derived from the chemical activation and pyrolysis of natural bre textile waste. J Anal Appl Pyrolysis 2004;71:97186. [4] Bansal RC, Donnet JB, Stoeckli F. Active carbon. New York: Marcel Dekker; 1988. [5] Rodrguez-Reinoso F. Activated carbon: structure, characterization, preparation and applications. In: Marsh H, Heintz EA, Rodrguez-Reinoso F, editors. Introduction to carbon technologies. Universidad de Alicante; Secretariado de Publicaciones; 1997. [6] Jagtoyen M, Derbyshire F. Activated carbons from yellow poplar and white oak by H3PO4 activation. Carbon 1998;36:108597. [7] Puziy AM, Poddubnaya OI, Martnez-Alonso A, Surez-Garca F, Tascn JMD. Synthetic carbons activated with phosphoric acid. I. Surface chemistry and ion binding properties. Carbon 2002;40:1493505. [8] Puziy AM, Poddubnaya OI, Martnez-Alonso A, Surez-Garca F, Tascn JMD. Surface chemistry of phosphorous-containing carbons of lignocellulosic origin. Carbon 2005;43:285768. [9] Rodrguez-Mirasol J, Bedia J, Cordero T, Rodrguez JJ. Inuence of water vapor on the adsorption of VOCs on lignin-based activated carbons. Sep Sci Technol 2005;40:311335. [10] Rodrguez-Mirasol J, Cordero T, Rodrguez JJ. Activated carbons from CO2 partial gasication of eucalyptus Kraft lignin. Energ Fuel 1993;7:1338. [11] Rodrguez-Mirasol J, Cordero T, Rodrguez JJ. High-temperature carbons from kraft lignin. Carbon 1996;1:4352. [12] Rodrguez-Mirasol J, Cordero T, Rodrguez JJ. Preparation and characterization of activated carbons from eucalyptus kraft lignin. Carbon 1993;1:8795. [13] Guo Y, Rockstraw DA. Physicochemical properties of carbons prepared from pecan shell by phosphoric acid activation. Bioresour Technol 2007;98:151321. [14] Blanco Castro J, Bonelli PR, Cerella EG, Cukierman AL. Phosphoric acid activation of agricultural residues and bagasse from sugar cane: inuence of the experimental conditions on adsorption characteristics of activated carbons. Ind Eng Chem Res 2000;39:416672. [15] Tancredi N, Cordero T, Rodrguez-Mirasol J, Rodrguez JJ. Activated carbons from Uruguayan eucalyptus wood. Fuel 1996;15:17016. [16] Marquez-Montesinos F, Cordero T, Rodrguez-Mirasol J, Rodrguez JJ. CO2 and steam gasication of a grapefruit skin char. Fuel 2002;81:4239. [17] Molina-Sabio M, Rodrguez-Reinoso F, Caturla F, Sells MJ. Porosity in granular carbons activated with phosphoric acid. Carbon 1995;33:110513. [18] Vernersson T, Bonelli PR, Cerrella EG, Cukierman AL. Arundo donax cane as a precursor for activated carbons preparation by phosphoric acid activation. Bioresour Technol 2002;83:95104. [19] Solum MS, Pugmire RJ, Jagtoyen M, Derbyshire F. Evolution of carbon structure in chemically activated wood. Carbon 1995;33:124754. [20] Molina-Sabio M, Rodrguez-Reinoso F. Role of chemical activation in the development of carbon porosity. Colloid Surf A: Physicochem Eng Aspects 2004;241:1525.