The adsorption of benzene, toluene and xylene (BTX) was investigated covering different types of commercially available activated carbons with varied surface area and surface functional groups. The physico-chemical properties were characterized by Brunauer–Emmett–Teller (BET) surface area analysis, Fourier transform infrared (FTIR) spectroscopy and the Boehm titration method. Experiments to assess the adsorption isotherms and kinetics of adsorption were performed and the results are presented. An increase in the surface acid functional groups was found to decrease the adsorption capacity, with the highest adsorption capacity corresponding to carbon with lowest acid functionality.

INTRODUCTION

Benzene, toluene and xylene, collectively abbreviated as BTX, are aromatic volatile organic compounds (VOCs) used as solvents and raw materials in the production of chemical intermediates and end-user consumer products such as paints, plastics and pharmaceuticals. Due to the wide range of applications of BTX they are in high demand and are among the most heavily produced chemicals worldwide, and are ranked among the top 50 chemicals produced in the USA (Leusch & Bartkow 2010).

However, release of BTX to the environment is a major issue due to the adverse effects it has on human health. Despite increasingly stringent government policies and laws on the discharge of toxic and hazardous materials, they still make their way into the environment at dangerous levels. The main sources of BTX release to the environment are petroleum refining plants, effluent discharges from chemical plants, spills and leaching from underground storage tanks (Everaert & Baeyens 2004; Leusch & Bartkow 2010). Groundwater, soil and the atmosphere are among the environmental media contaminated by BTX. Additionally, BTX are recognized by the US Environmental Protection Agency (EPA) as drinking and groundwater contaminants, as they account for 6% of the contaminants which affect groundwater and soil in Europe. The concentration of these contaminants in the waste stream could vary on a larger scale, depending upon the processing industries (or feeders) to the effluent stream. However, the maximum contaminant level as stipulated by the EPA should be less than 5 ppb for benzene while for toluene and xylene it should be less than 10 ppb.

There are several methods in practice for the removal of soluble organic compounds from waste water streams, including chemical oxidation, air stripping, membrane separation, biodegradation and adsorption (Ruddy & Carroll 1993; Zhang et al. 2013). Among the various techniques, adsorption is industrially widely favored in particular over separation of BTX owing to the availability of cheap adsorbents with high adsorption capacity and low concentration of BTX (Benkhedda et al. 2000). Several types of adsorbents have been reported for removal of VOCs, which include activated carbon (AC), surfactant modified zeolites, silicates, organic minerals, carbon nanotubes, polymeric resin, etc. (Cooney 1998; Ghiaci et al. 2004; Koh & Dixon 2001; Su et al. 2010; Simpson et al. 1993). Among the choice of adsorbents, AC is widely utilized in industry owing to its low cost, large surface-area and market availability.

A scan through the literature with specific reference to AC reveals a large variation in BTX adsorption capacities depending on the type of precursor, activation methods, etc. Typically a high adsorption capacity of near 800 mg/g of benzene was reported for AC prepared from coal tar with KOH being the activating agent (Asnejo et al. 2011). However, an adsorption capacity of only near 200 mg/g was reported for AC prepared from bituminous coal utilizing phosphoric acid as the activating agent (Hindarso et al. 2001). A wide variation in the surface area, pore structure and functional groups of the carbon possibly contribute to the variation in the adsorption capacity. Additionally, a trend of increase in adsorption capacity with increase in surface area and decrease in the acid functional groups was reported by Daifullah & Girgis (2003).

The present work attempts to understand the mechanism of adsorption of BTX on various physico-chemical properties of virgin activated-carbon.

EXPERIMENTAL METHODS

Materials

Various commercially available activated carbons were procured from different manufacturers. The following terminologies were utilized to identify the different carbons: C38 corresponds to AC procured from CarboTech, a low ash hard coal steam-gasified carbon, C40 corresponds to AC procured from CarboTech, a low ash coal steam-gasified carbon of a different grade, ACM40 corresponds to AC procured from CECA Chemicals using steam activation, while Norit corresponds to AC procured from NORIT carbons using steam activation. The B, T and X of analytical grade were procured from Merck (Germany) and were diluted with deionized water to the required concentrations.

Batch adsorption

Fifty milligrams of adsorbents and 100 ml of BTX solution were used with initial concentrations (C0) that varied from 100 to 400 mg/L for benzene and toluene, and 25–125 mg/L for xylene, respectively. The vials were sealed and mounted on a shaker, which was placed in a temperature-controlled bath and operated at 30, 40 and 50 °C at 200 rpm. The adsorption experiments were conducted for from 24 to 72 hours to ensure equilibrium. The kinetics of BTX adsorption (q, mg/g) were estimated by conducting experiments for different time durations. The amount adsorbed by the adsorbent was calculated using 
formula
where C0 and Ct are the BTX concentrations at the beginning and after a certain period of time t (mg/L); V is the initial solution volume (L); and m is the adsorbent weight (g). The concentrations of BTX were analyzed using a Varian Cary 5,000 UV–Vis–NIR spectrometer in the frequency range of 200 nm and 300 nm.

Materials characterization

The surface characteristics of the AC samples were determined by measuring the adsorption isotherms of nitrogen at 77 K using an Autosorb-6B. The AC samples were degassed at 300 °C for 24 hours in the Autosorb Degasser. Fourier transform infrared (FTIR) spectroscopy was carried out using a Nicolet Avatar 330 spectrometer. The surface functional groups containing oxygen were determined according to Boehm titration (Boehm 1994).

RESULTS AND DISCUSSION

Nitrogen adsorption–desorption isotherms of different types of adsorbents are shown in Figure 1 and their structural parameters are summarized in Table 1. These isotherms have similar features to the Type IV isotherm classification of IUPAC. The initial curvature or knee is representative of monolayer coverage and the hysteresis loop occurring at P/Po > 0.4 could indicate capillary condensation and that the type of pores is similar to a bottle neck. C38, C40 and ACM40 have high surface areas at 1,129 m2/g, 1,366 m2/g and 1,326 m2/g, respectively, while Norit-AC has lower surface area at 886 m2/g. The variation in the pore characteristics of C38, C40 and ACM40 are insignificant, while Norit-AC presents a different pore characteristic, probably owing to variation in the manufacturing process of C40, and ACM40 has the larger total pore volume at 0.86 cm3/g and 0.74 cm3/g and mesopore volumes of 0.61 cm3/g and 0.53 cm3/g, respectively. The micropore volume was observed to be significantly higher for the C38 carbon as compared to others.

Table 1

Physico-chemical properties of different types of activated carbon

Activated carbonSurface area (m2/g)Average pore diameter (nm)Total pore volume (cm3/g)Micropore volume (cm3/g)Mesopore volume (cm3/g)
C38 1,129 2.30 0.65 0.40 0.25 
C40 1,366 2.53 0.86 0.26 0.61 
ACM40 1,326 2.23 0.74 0.21 0.53 
Norit-AC 886 1.34 0.59 0.20 0.39 
Activated carbonSurface area (m2/g)Average pore diameter (nm)Total pore volume (cm3/g)Micropore volume (cm3/g)Mesopore volume (cm3/g)
C38 1,129 2.30 0.65 0.40 0.25 
C40 1,366 2.53 0.86 0.26 0.61 
ACM40 1,326 2.23 0.74 0.21 0.53 
Norit-AC 886 1.34 0.59 0.20 0.39 
Figure 1

Brunauer–Emmett–Teller (BET) surface area analysis of AC: (a) C38; (b) C40; (c) ACM40; (d) Norit-AC.

Figure 1

Brunauer–Emmett–Teller (BET) surface area analysis of AC: (a) C38; (b) C40; (c) ACM40; (d) Norit-AC.

The FTIR spectra of virgin AC samples are shown in Figure 2. The absorbance band is observed at 3,500 cm−1 which is assigned to OH from the COOH groups. These bands clearly demonstrate the presence of more OH groups on the surface. The small bands at 2,900 cm−1 represent the asymmetric and symmetric stretching vibrations of –CH3. The presence of –C ≡ C is seen by the stretching bands at 2,300 cm−1. The peaks at 1,500 cm−1 are usually assigned to –C = O stretching vibrations of ketone, aldehyde, lactone or carbonyl groups. The broad band stretching between 900 and 1,300 cm−1 is assigned to both stretching –C–O and –O–H bending of alcoholic, phenolic and carboxylic groups.

Figure 2

FTIR spectra of AC: (a) C38; (b) C40; (c) ACM40; (d) Norit-AC.

Figure 2

FTIR spectra of AC: (a) C38; (b) C40; (c) ACM40; (d) Norit-AC.

The amount of surface oxygen-containing groups on the AC was determined by the Boehm titration method and the results are presented in Table 2. The results indicate that total acidity is the sum of carboxylic, lactonic and phenolic functional groups. The lowest acidic functional groups were for C38, C40 and ACM40, while Norit-AC had the highest. The base functional groups for C38, C40 and ACM40 exhibited high values when compared with Norit-AC. Although C38, C40 and ACM40 have comparable total acidic functional groups, the proportions of carboxylic, lactonic and phenolic groups are different.

Table 2

Amount of surface functional groups from Boehm titration

Activated carbonFunctional groupAmount of functional group nCSF (mmol/g)Total acidic groups (mmol/g)Total basic groups (mmol/g)
C38 Carboxylic 0.100 1.203 1.689 
Lactonic 0.124   
Phenolic 0.979   
C40 Carboxylic 0.104 1.243 1.710 
Lactonic 0.149   
Phenolic 0.990   
ACM40 Carboxylic 0.162 1.263 1.782 
Lactonic 0.169   
Phenolic 0.932   
Norit-AC Carboxylic 0.194 1.328 1.578 
Lactonic 0.110   
Phenolic 1.024   
Activated carbonFunctional groupAmount of functional group nCSF (mmol/g)Total acidic groups (mmol/g)Total basic groups (mmol/g)
C38 Carboxylic 0.100 1.203 1.689 
Lactonic 0.124   
Phenolic 0.979   
C40 Carboxylic 0.104 1.243 1.710 
Lactonic 0.149   
Phenolic 0.990   
ACM40 Carboxylic 0.162 1.263 1.782 
Lactonic 0.169   
Phenolic 0.932   
Norit-AC Carboxylic 0.194 1.328 1.578 
Lactonic 0.110   
Phenolic 1.024   

Adsorption kinetics

The kinetics experiments were conducted using standard solutions of B, T and X (400, 400 and 125 mg/L) with 50 mg of activated carbon C38. Each of the sample vials was placed in the water bath shaker at 30 and 50 °C. Figure 3 shows the effect of time and temperature on the adsorption of BTX. The three adsorbates exhibit a very high initial uptake onto the AC while the adsorption process slows down after the initial 12 hours, eventually establishing adsorption equilibrium. The rate of adsorption was observed to increase with increasing temperature. The experimental data were tested with different well-known kinetic models such as pseudo-first order, pseudo-second order and intra-particle diffusion to identify the appropriate kinetic models for BTX adsorption. The kinetic model and correlation coefficient parameters are tabulated in Table 3. Both the pseudo-first order and intra-particle diffusion kinetic equations failed to accurately model the adsorption of BTX evidenced by the low R2 values. The pseudo-second-order kinetic equation best represents the adsorption of BTX on AC evidenced by the high R2 as well as the proximity of the predicted adsorption capacity with the experimental data.

Table 3

Kinetic models for BTX adsorption using C38

AdsorbateT (°C)Experimental qe (mg/g)Pseudo-first order
Pseudo-second order
Intra-particle diffusion
Calc. qe (mg/g)k1 (1/hr)R2Calc. qe (mg/g)k2 (g/mg.hr)R2kID (g/mg.hr0.5)C (mg/g)R2
Benzene 30 222.15 96.1 0.0670 0.930 224.225 0.00324 0.9979 17.939 100.315 0.5899 
50 189.08 39.2 0.0533 0.770 189.610 0.0103 0.9996 12.770 114.316 0.3210 
Toluene 30 278.69 88.0 0.0693 0.917 281.245 0.00432 0.9997 20.334 148.240 0.4814 
50 262.37 56.3 0.0593 0.781 263.118 0.00738 0.9998 17.931 151.998 0.3943 
Xylene 30 89.91 36.6 0.162 0.850 92.800 0.0130 0.9959 16.966 29.499 0.6231 
50 86.20 26.3 0.151 0.705 88.326 0.0199 0.9970 15.661 31.464 0.5580 
AdsorbateT (°C)Experimental qe (mg/g)Pseudo-first order
Pseudo-second order
Intra-particle diffusion
Calc. qe (mg/g)k1 (1/hr)R2Calc. qe (mg/g)k2 (g/mg.hr)R2kID (g/mg.hr0.5)C (mg/g)R2
Benzene 30 222.15 96.1 0.0670 0.930 224.225 0.00324 0.9979 17.939 100.315 0.5899 
50 189.08 39.2 0.0533 0.770 189.610 0.0103 0.9996 12.770 114.316 0.3210 
Toluene 30 278.69 88.0 0.0693 0.917 281.245 0.00432 0.9997 20.334 148.240 0.4814 
50 262.37 56.3 0.0593 0.781 263.118 0.00738 0.9998 17.931 151.998 0.3943 
Xylene 30 89.91 36.6 0.162 0.850 92.800 0.0130 0.9959 16.966 29.499 0.6231 
50 86.20 26.3 0.151 0.705 88.326 0.0199 0.9970 15.661 31.464 0.5580 
Figure 3

BTX adsorption kinetics using C38.

Figure 3

BTX adsorption kinetics using C38.

The pseudo-second-order rate constant k2 shows that the adsorption process is favored in the order xylene, toluene and benzene. This trend can be explained due to the decrease in solubility and increase in the partition coefficient. The activation energy for adsorption of BTX was calculated using the Arrhenius equation along with the rate constant (k2). The activation energy for adsorption was 47.2 kJ/mol, 21.8 kJ/mol and 17.2 kJ/mol for B, T and X, respectively. The trend of activation energies B> T + X is explained by xylene having a lower energy barrier to overcome, which confirms the greater tendency to leave the solution and adsorb on the carbon surface. The relatively low activation energies are characteristic of van der Waals interaction forces between the adsorbate and AC which describe a physical adsorption process occurring. Furthermore, physical adsorption usually has activation energy in the range of 5–40 kJ/mol, while higher energy between 40 and 800 kJ/mol suggests chemisorption (Chen & Chen 2009). Hence the adsorption of benzene can be identified as being dominated by the chemisorption process. This could be due to higher interactions between benzene active sites and surface functional groups than in the cases of toluene and xylene.

Equilibrium adsorption capacities

Equilibrium adsorption capacities of BTX were determined at 30, 40 and 50 °C by varying the initial concentration. The benzene and toluene initial concentrations were varied over the range 400–100 mg/L, and xylene over the range 125–25 mg/L, as limited by their solubility in aqueous media. The samples were analyzed after equilibrium was achieved; 72 hours for benzene and toluene, 48 hours for xylene. The adsorption isotherms were as shown in Figures 4, 5 and 6 for B, T and X, respectively. The isotherms of xylene are not smooth owing to their lower solubility and limited concentration range, which reduced the resolution of the curves, unlike for benzene and toluene.

Figure 4

Equilibrium adsorption capacities for benzene.

Figure 4

Equilibrium adsorption capacities for benzene.

Figure 5

Equilibrium adsorption capacities for toluene.

Figure 5

Equilibrium adsorption capacities for toluene.

Figure 6

Equilibrium adsorption capacities for xylene.

Figure 6

Equilibrium adsorption capacities for xylene.

The adsorbents showed clearly that an increase in temperature has a negative effect on the adsorption of BTX. The highest adsorption capacities of benzene using C38 were 222 mg/g, 210 mg/g and 189 mg/g at 30 °C, 40 °C and 50 °C, respectively. The equilibrium adsorption at 30 °C was found to be higher than at 50 °C, well in agreement with the adsorption trends in the literature (Kawasaki et al. 2004). Decrease in the equilibrium adsorption with increase in temperature could be attributed to a left shift of the adsorption equilibrium, due to the exothermic nature of adsorption. The adsorption capacities vary significantly among the various AC, which could be attributed to the variation in the properties and characteristics of the respective carbon. C40 and ACM40 have higher surface area than C38 and Norit-AC, but provide lower adsorption capacity. This could be attributed to a higher proportion of micropore volume in C38 coupled with lower acid functional groups compared to C40 and ACM40. Additionally, Norit-AC exhibited the lowest adsorption capacity, which could be attributed to higher acid functional groups coupled with low surface-area. Similar trends were also observed for toluene adsorption while the trends were in conclusive for xylene owing to low variation in the liquid phase equilibrium concentration. High acidic surface functional groups were found to reduce adsorption capacity due to an increase in hydrophilic nature as well as due to π–π interaction.

The results can be analyzed based on the π–π interaction mechanism (Coughlin & Ezra 1968; Coughlin et al. 1968). More acidic carbons have a greater amount of surface oxygen groups, which are electron-withdrawing groups. The oxygen groups withdraw π-electrons from the carbon basal planes and therefore the density of π-delocalized π-electrons on the carbon basal planes decreases. This causes a reduction in π–π interactions between π-electrons of the aromatic rings of the BTX and the ACs' basal planes. Toluene and xylene are more electron-rich (Newcombe & Dixon 2006), due to the presence of the methyl groups, which act as electron-donating groups. Additionally, an increase in surface oxygen functional groups would render carbon more polar, reducing the hydrophobicity of the AC, resulting in water molecules being adsorbed, eventually blocking the active pore sites and hindering adsorption.

CONCLUSIONS

The work assesses the ability of porous activated-carbon to be an effective adsorbent for the removal of BTX. Equilibrium adsorption capacity of the chosen adsorbents was found to increase in the order T > B > X. The highest adsorption capacity corresponds to the adsorbent C38, which had the lowest acidic functional groups. The decrease in adsorption for the acid carbons is attributed to reduced π–π interactions and blockage of active pore sites by water molecules. The kinetics of adsorption were found to follow pseudo-second-order kinetics for BTX. In conclusion, from the kinetics and equilibrium adsorption capacity, the active carbon with high surface-area and low acid functional groups could potentially be a good candidate for removal of aromatic compounds from waste streams.

ACKNOWLEDGEMENT

The authors would like to thank GRC, The Petroleum Institute for financial support.

REFERENCES

REFERENCES
Asnejo
N. G.
Alvarez
P.
Granda
M.
Blanco
C.
Santamaría
R.
Menéndez
R.
2011
High performance activated carbon for benzene/toluene adsorption from industrial wastewater
.
J. Hazard. Mater.
192
,
1525
1532
.
Benkhedda
J.
Jaubert
J.-N.
Barth
D.
Perrin
L.
2000
Experimental and modeled results describing the adsorption of toluene onto activated carbons
.
J. Chem. Eng. Data
45
,
650
653
.
Cooney
D. D.
1998
Adsorption Design for Wastewater Treatment
.
Lewis
,
New York
.
Coughlin
R. W.
Ezra
F. S.
Tan
R. N.
1968
Influence of chemisorbed oxygen in adsorption onto carbon from aqueous solution
.
J. Colloid. Interf. Sci.
28
,
386
396
.
Daifullah
A. A. M.
Girgis
B. S.
2003
Impact of surface characteristics of activated carbon on adsorption of BTEX
.
J. Coll. Surf. A: Physicochem. Eng. Aspects
214
(
1–3
),
181
193
.
Everaert
K.
Baeyens
J.
2004
Catalytic combustion of volatile organic compounds
.
J. Hazard. Mater.
109
,
113
139
.
Hindarso
H.
Ismadji
S.
Wicaksana
F.
Mudjijati
Indraswati
N.
2001
Adsorption of benzene and toluene from aqueous solution onto granular activated carbon
.
J. Chem. Eng. Data
46
,
788
791
.
Kawasaki
N.
Kinoshita
H.
Oue
T.
Nakamura
T.
Tanada
S.
2004
Study on adsorption kinetic of aromatic hydrocarbons onto activated carbon in gaseous flow method
.
J. Colloid. Interf. Sci.
275
,
40
43
.
Leusch
F.
Bartkow
M.
2010
A Short Primer on Benzene, Toluene, Ethylbenzene and Xylenes (BTEX) in the Environment and in Hydraulic Fracturing Fluid
.
Prepared for the Queensland Department of Environment and Resource Management. Smart Water Research Center, Griffith University, Queensland, Australia
.
Newcombe
G.
Dixon
D.
2006
Interface Science in Drinking Water Treatment: Theory and Applications
.
Academic Press, Amsterdam, The Netherlands
.
Ruddy
E. N.
Carroll
L. A.
1993
Select the best VOC control strategy
.
Chem. Eng. Prog.
89
,
28
35
.