- Email: [email protected]
Binary component sorption of Cu(II) and Pb(II) with activated carbon from Eucalyptus camaldulensis Dehn bark
Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
Contents lists available at ScienceDirect
Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec
Binary component sorption of Cu(II) and Pb(II) with activated carbon from Eucalyptus camaldulensis Dehn bark Apipreeya Kongsuwan a, Phussadee Patnukao b, Prasert Pavasant a,b,* a b
Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok, Thailand National Center of Excellence for Environmental and Hazardous Waste Management, Chulalongkorn University, Bangkok, Thailand
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 April 2008 Accepted 23 February 2009
The objective of this work is to illustrate the potential in the use of activated carbon in the binary component sorption of copper and lead ions. Eucalyptus bark was used as a precursor for the activated carbon which was prepared through the phosphoric acid activation process. This activated carbon was then used for the sorption of copper and lead ions. The quantity of the metal ions in the solution was measured with the Flame & Graphite Furnace Atomic Adsorption Spectrophotometer. The results indicated that the optimal pH for sorption was 5. The maximum sorption capacities for Cu(II) and Pb(II) were 0.45 and 0.53 mmol g1. Carboxylic, amine and amide groups were found to involve in the sorptions of Cu(II) and Pb(II). A major mechanism for the uptake of both heavy metals was proven not to be ion exchange but adsorption. In binary component sorptions, activated carbon still could sorb Pb(II) in a greater amount than Cu(II). However, the presence of the secondary metal ions suppressed the sorption of the primary metal ions. There seemed to have a linear inverse dependency between the sorption capacity and the concentration of the secondary metal ion. ß 2009 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
Keywords: Competitive sorption Heavy metal removal Sorption isotherm Adsorption Ion exchange
1. Introduction With a rapid increase in heavy industrial activities, pollution derived from uncontrolled escapes of heavy metals such as copper, nickel, chromium, and zinc has become serious [1,2]. Heavy metals are detected in waste streams from industries in large quantities. These heavy metals have harmful effects on human physiology and other biological systems when they exceed tolerance levels which are generally in the level of less than one part per million (ppm). Among the various treatment technologies, activated carbon sorption is commonly used because the design is simple and can be used effectively with low concentration waste streams (less than 100 ppm). Single component sorptions have extensively been investigated, however, relatively less research attention has focused on the binary component sorption. In such binary systems, sorption of heavy metal ions depended not only on specific surface properties of the adsorbent and the physicochemical parameters of the solution such as pH and initial metal ion concentration, but also on the interaction between heavy metal ion species. Competition
* Corresponding author at: Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok, Thailand. Tel.: +66 2 2186870; fax: +66 2 2186877. E-mail address: [email protected] (P. Pavasant).
among the different metal ions for the surface binding sites will certainly occur and change the sorption characteristics of the sorption of each ion species. Our recent work demonstrated a successful conversion of eucalyptus bark to activated carbon [3]. This work therefore intended to further employ this activated carbon product in the removal of heavy metal from the binary mixture solution. A synthetic mixture solution of copper and lead was arbitrarily selected as a modeled sorption system. Possible interactions between the metal ions and the surface of the adsorbent in the binary component sorption were described. 2. Experimental 2.1. Preparation of activated carbon The raw material, eucalyptus bark waste from the pulp mill, was impregnated into phosphoric acid (85% by weight) with the weight ratio of bark and phosphoric acid at 1:1. The mixture was carbonized in a muffle furnace at 500 8C for 1 h. The product was washed with hot distilled water until the pH of the leachate was 6 and dried in an oven at 105 8C for 4 h. Finally, powder activated carbon was crushed and sieved in the size ranged between mesh number 325 (0.045 mm) and 100 (0.150 mm) and stored in a desiccator.
1226-086X/$ – see front matter ß 2009 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jiec.2009.02.002
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
466
2.2. Activated carbon characteristics
Table 1 Proximate and ultimate analyses of eucalyptus bark and activated carbon.
The ultimate analysis of eucalyptus bark was performed in CHNS/O analyzer (PerkinElmer PE2400 Series II) and the proximate analysis was carried out following ASTM standards for chemical analysis of wood charcoal [4]. Characteristics of activated carbon from this work were examined. The yield of activated carbon was defined as the ratio of the weight of the activated carbon product to that of the original eucalyptus bark, both on a dry basis. Apparent (bulk) density of all samples was calculated as the ratio between weight and volume of packed dry material. The BET surface area was determined by nitrogen adsorption (196 8C) on surface area analyzer (Thermo Finnigan, Sorptomatic 1990). The cross-sectional area for nitrogen molecule was assumed to be 0.162 nm2. For sorption experiments, iodine number was determined according to the ASTM standard method [5]. Methylene blue number was determined according to the standard method [6]. The surface morphology of activated carbon was visualized via scanning electron microscopy (SEM), the corresponding SEM micrographs being analyzed with a scanning electron microscope (XL 30 ESEM FEG).
Parameter
2.3. Sorption 0.1 g of activated carbon was placed in a conical tube with 25 mL of 0.1 mmol L1 metal solutions. Metals in nitrate form (Cu(NO3)2 and Pb(NO3)2) were used to prepare the metal solution with initial concentration in the range from 0.1 to 10 mmol L1. The solution was mixed by a rotary shaker at 200 rpm, 25 8C for 1 h. Binary components of Cu(II) and Pb(II) were initially prepared as a mixture of equimolar concentration, e.g. binary components at 10 mmol L1 was prepared as a mixture of Cu(II) and Pb(II) both at 5 mmol L1. The range of concentration of heavy metals used to investigate sorption kinetics of the binary component systems was from 0.625 to 10 mmol L1. Loaded activated carbon was separated from solution by filter paper where the starting solutions and the filtrates were analyzed for the metal ions by Flame & Graphite Furnace Atomic Adsorption Spectrophotometer (AAS, ZEEnit 700, Analytik Jena, United Kingdom). Determination of sorption capacities [5] could be calculated from Eq. (1) q¼
VðC i C e Þ 1000W
% by weight Eucalyptus bark
Activated carbon
Proximate Moisture content Volatile matter Fixed carbon Ash content
10.50 75.05 13.10 1.35
7 22.78 65.34 4.88
Ultimate C H Others—N, O, P, S
41.36 4.67 53.97
62 4 34
carbon at 1000 magnification which confirmed amorphous and heterogeneous structures. The product contained relatively higher fixed carbon and lower volatile matter than those in the raw material as they were removed during the activation process. However, the overall yield of the process calculated from the ratio between the dry weights of the product and the raw material was only 30% (see Table 2), and the weight loss was due to the gasification of such volatile and some of the carbon in the raw materials. This level of yield was considered moderate when compared with those prepared with other raw materials such as mangrove wood (38.07%) [7], Eucalyptus camaldulensis Dehn (33.14%) [8], used tired (47.20%) [9], palm oil shell (19.66%) [10]. Table 2 also summarizes other properties of the activated carbon obtained from the technique employed in this work. The BET result illustrated that the activated carbon product had relatively high specific surface area (1240 m2 g1) when
(1)
where q is sorption capacity (mmol g1), Ci the initial metal concentration (mmol L1), Ce the concentration of metal at equilibrium (mmol L1), W the adsorbent dosage (g) and V the solution volume (L). In case of binary components systems, solution of copper–lead nitrate at 0.2 mmol L1 (equimolar concentration, Cu(II) = Pb(II) = 0.1 mmol L1) was added into the conical tube where the initial concentration of metal ions was in the range of 0.2–10 mmol L1. Note that all experiments were carried out in triplicate. 3. Results and discussion 3.1. Characteristics of eucalyptus bark raw material and activated carbon product Table 1 shows the proximate and ultimate analyses of eucalyptus bark. It was shown that the bark contained relatively high carbon and low ash, properties suitable as a raw material for the synthesis of activated carbon. After the activated process, the morphology of the carbon changed and the external surface of activated carbon obtained from scanning electron micrographs (SEMs) was shown in Fig. 1. This SEM image illustrates irregular structure with cracks and crevices on the surface of the activated
Fig. 1. SEM photograph of (a) eucalyptus bark and (b) activated carbon.
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
467
Table 2 Characteristics of activated carbon. Parameter
Value
Yield (%) Apparent (bulk) density (g cm3) BET specific surface area (m2 g1) Pore specific volume (cm3 g1) Average pore diameter (A˚)
30 0.251 1239.38 1.572 8.49
Pore size distribution Micropore (%) Mesopore (%) Iodine number (mg g1) Methylene blue (mg g1)
88.5 11.48 918 7 427 2
compared with those reported in literature such as cotton stalk (1032 m2 g1) [11], sugar cane bagasse (1132 m2 g1) [12], grain sorghum (1032 m2 g1) [13], peanut hull (1177 m2 g1) [14] Arundo donax cane (1333 m2 g1) [15], and rice husk (376 m2 g1) [16]. The difference could be due to the removal of organic components and minerals presented in the different raw materials which could react differently with phosphoric acid during the activation process [17]. This also led to a distinctive formation of pore structure in the final activated carbon product, and explained the weight loss during activation process, as phosphoric acid reacted with char and volatile matters resulting in gas products which diffused quickly out of the surface of particles and created pores [3]. The iodine and methylene blue numbers indicated the sorption capacities of the activated carbon. The iodine number indicated the sorption capacity of activated carbon in micropore (d < 200 nm) [18] and the methylene blue number represent the sorption capacity in mesopore (200 < d < 5000 nm). Table 2 illustrates that the activated carbon product from this work had a relatively high iodine number of about 918 7 mg g1, but on the other hand, a relatively low methylene blue number (427 2 mg g1), which meant that the product had more micropore than mesopore structures. The inorganic composition of activated carbon determined by microwave digester is shown in Fig. 2 and this demonstrates that the main compositions comprised alkaline and alkaline earth. These metals may be important for the sorption of heavy metals in the case that ion exchange process prevailed.
Fig. 3. Single and binary adsorption isotherms of Cu(II) and Pb(II) (shaking rate = 200 rpm, 25 8C).
significantly on the initial concentration of the metals and the isotherm results at 25 8C are displayed in Fig. 3. As a general observation, Pb(II) exhibited a higher maximum sorption capacity than Cu(II). Two mathematical expressions are commonly used to describe the isotherm of the sorption: Langmuir and Fruendlich equations [19]. The Langmuir isotherm is given by: qe ¼
qm bC 1 þ bC
where qm is the maximum sorption capacity (mg g1), b a constant related to bonding energy of sorption. The Freundlich isotherm is expressed as: qe ¼ K F C 1=n
Sorption equilibrium is established when the concentration of metal ion in the bulk solution (Ce) was in dynamic with that in the solid matrix (qe). This level of equilibrium concentration depends
(3)
where KF and n are constants. Table 3 reveals that experimental data fitted more favorably with Langmuir than Freundlich isotherms suggesting that the sorptions of Cu(II) and Pb(II) be potentially monolayer [20]. A comparison between maximum sorption capacities (qm) between the two metal ions illustrated that Cu(II) and Pb(II) had maximum sorption Table 3 Parameters of Langmuir and Freundlich isotherms. Heavy metal ion
3.2. Sorption
(2)
Cu(II) Pb(II)
Langmuir isotherm
Freundlich isotherm
qm (mmol g1)
b (L mmol1)
R
KF
1/n
R2
0.455 0.534
6.125 6.555
0.9897 0.9964
0.538 0.372
0.494 0.392
0.9041 0.7456
Fig. 2. Metal compositions in activated carbon.
2
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
468
Table 4 Comparison of reported maximum adsorption capacities. Author
Adsorbent
Heavy metal
qm (mmol g1)
This work Quek et al. [21]
Activated carbon Sago waste
Uzun et al. [22] Faur-Brasquet et al. [23]
Commercial activated carbon AC pretreated with NaCl
Chen et al. [17a]
0.450 0.200 0.530 0.230 0.174 0.147 0.040
Feng et al. [20]
Activated carbon pretreated with 1 M citric acid Rice husk ash
Cu(II) Cu(II) Pb(II) Pb(II) Cu(II) Pb(II) Cu(II)
Chen and Wu [24] Pavasant et al. [25]
GAC Dried Caulerpa lentillifera
Cu(II) Pb(II) Cu(II) Cu(II) Pb(II)
0.230 0.064 0.110 0.088 0.139
Amine N–H stretching N–H bending C–N stretching
Cu(II)
Pb(II)
* * * *
* * * *
*
*
*
*
* Indicates the potential functional group involved in the adsorption.
capacities of 0.45 and 0.53 mmol g1, respectively. The level of maximum sorption capacity of this activated carbon was considered relatively high when compared with the reported values summarized in Table 4. The functional groups related to the sorption were analyzed with FTIR where the results in Table 5 demonstrate that the activated carbon might contain more functional group binding sites for Pb(II) than for Cu(II). Moreover, Pb(II) sorption was found to have a higher b value indicating a stronger chemical and physical affinity with the activated carbon than Cu(II). The binary components isotherm models based on Langmuir assumption can be derived [25]. In case of binary component sorption, the sorption reaction can be expressed by two chemical reaction equations: K1
B þ M1 !BM 1 ; K2
B þ M2 !BM 2 ;
½BM1 ¼
K 1 ½Bt ½M1 1 þ K 1 ½M 1 þ K 2 ½M 2
(7)
½BM2 ¼
K 2 ½Bt ½M2 1 þ K 1 ½M 1 þ K 2 ½M 2
(8)
or in terms of sorption capacity as:
*
Amide N–H stretching C–O stretching
(6)
½Bt ¼ ½B þ ½BM1 þ ½BM2 : Eqs. (4)–(6) can be rearranged to:
Table 5 Possible functional groups involved with the adsorption of Cu(II) and Pb(II).
Carboxylic acid O–H stretching C5 5O stretching C–O stretching O–H bending
M1 and M2, respectively, and K1 and K2 the equilibrium constants of reactions. The total binding sites are the sum of the vacant and occupied sites:
K1 ¼
½BM1 ½B½M 1
(4)
K2 ¼
½BM2 ½B½M 2
(5)
where B are the free binding sites, M1 and M2 the first and second adsorbates in the solution, BM1 and BM2 the bound sites for metal
q1 ¼
b1 qm;1 C e;1 1 þ b1 C e;1 þ b2 C e;2
(9)
q2 ¼
b2 qm;2 C e;2 1 þ b1 C e;1 þ b2 C e;2
(10)
where qm is the maximum sorption capacity, b the affinity constant of Langmuir model, Ce the equilibrium concentration, q the sorption capacities, and subscripts 1 and 2 represent metal components 1 and 2, respectively. Fig. 3 shows the sorption isotherms of single and binary component sorptions. In this figure, the labels: Cu/Cu–Pb and Pb/ Cu–Pb: refer to the sorption isotherms of Cu(II) and Pb(II) from binary component systems, respectively: and the labels; Cu and Pb, are for the sorption isotherms of Cu(II) and Pb(II) from single component systems, respectively. Table 6 displays the equilibrium sorption capacities of single and binary component sorptions. It was also noticed that the sorption capacities of both metals for the single and binary components were similar at low initial concentration. For instance, qe for the sorption of Cu(II) at an initial concentration of 0.625 mmol L1 was almost the same as that for the binary sorption at an initial concentration of 1.25 mmol L1 (where initial Cu(II) = Pb(II) = 0.625 mmol L1). However, as the concentration became high, the sorption capacity for the binary system seemed to be lowered than that of the single component. For Cu(II) in particular, the equilibrium sorption capacity in the binary system seemed to level off at 2.16 101 mmol g1 (for systems prepared at equimolar ratio between the two metals). This could be because, at low concentration, there existed a large number of sorption sites compared with the amount of metal ion in the solution, and therefore the level of competition between the two metals was not high, and the sorption capacities for both metals remained the same as those from single component systems. However, as the concentration became high, the competition for the sorption sites became more intense and this resulted in a decrease in the sorption capacity for both metals. It should be noted that the sorption of Cu(II) in the binary system seemed to be leveled off as the initial concentration of Cu(II) reached 5 mmol L1. However, the sorption capacity of Pb(II) in such systems did not seem to reach a constant level with the range of concentration employed in this work.
Table 6 Comparison equilibrium adsorption capacity, qe (mmol g1), between single and binary adsorptions. Initial concentration (mmol L1)
Cu(II) (q0, mmol g1)
Cu(II) from binary componentsa (qmix, mmol g1)
qmix/q0 of Cu(II)
Pb(II) (q0, mmol g1)
Pb(II) from binary componentsa (qmix, mmol g1)
qmix/q0 of Pb(II)
10 5 2.5 1.25 0.625
4.59 101 3.64 101 3.00 101 2.04 101 1.27 101
2.16 101 2.16 101 2.11 101 1.24 101
0.5934 0.7200 1.0343 0.9764
5.31 101 4.57 101 3.89 101 2.95 101 1.30 101
3.79 101 3.10 101 3.02 101 1.29 101
0.8293 0.7969 1.0237 0.9923
a
Binary component is equimolar concentration, e.g. binary components at 10 mmol L1 was prepared from 5 mmol L1 of Cu(II) and Pb(II).
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
469
Langmuir isotherms as expressed in Eqs. (9) and (10) were selected to discuss experimental data to predict maximum sorption capacities. The highest achievable sorption capacities of Cu(II) and Pb(II) in binary components sorption were 0.181 and 0.357 mmol g1 in case of equimolar concentration, respectively. Table 5 indicates that Pb(II) could bind with a more variety of functional groups than Cu(II) and therefore the sorption capacity for Pb(II) was slightly higher than that of Cu(II). The maximum sorption capacities of Cu(II) and Pb(II) obtained from binary component sorption were less than those obtained from the single component system where qm,Cu = 0.45 mmol g1 and qm,Pb = 0.53 mmol g1. This could be due to the reasons given above. The effect of ionic interaction on the binary component sorption can be represented by the ratio of the sorption capacity for one metal ion in the presence of the other metal ion (binary component system), qmix, to the sorption capacity for the same metal when it was presented alone in the solution (single component system), q0 [26]. Three case scenarios could be formulated: (i) qmix/q0 > 1 the sorption was promoted by the presence of other metal ions, (ii) qmix/q0 = 1 there was no observable net interaction, and (iii) qmix/q0 < 1 sorption was suppressed by the presence of other metal ions. Table 6 shows qmix/q0 from the binary components sorption of Cu(II), all of which were almost less than unity. This indicated that Cu(II) sorption was suppressed by Pb(II) sorption. In the same manner, qmix/q0 of Pb(II) from binary component sorption were less than unity which also indicated that Pb(II) sorption was suppressed by Cu(II) sorption. To further illustrate this point, binary component sorption experiments were carried out using the mixture with nonequimolar initial concentration. In this case, the molar ratios of initial concentrations of Cu(II) and Pb(II) were varied from 2:0 to 0:2 with an increment of 0.5 mmol L1. The plot of maximum sorption capacities at various ratio of Cu(II) and Pb(II) as shown in Fig. 4 demonstrates that inhibition effect did actually occur throughout the examined concentration range. In other words, the presence of one metal species lowered the sorption capacity of the other. It is interesting to note that there existed a linear inverse dependency between the sorption capacity and the concentration of the secondary metal ion. The total sorption capacity could be well predicted (as illustrated as a top solid line in Fig. 4) provided that the ratio between the two metals was known. 3.3. Mechanism for removal of heavy metals Removal of heavy metals from wastewater could involve either adsorption or ion exchange, or both. To evaluate for the ion exchange mechanism of removal process, the measurement of the
Fig. 5. Metal ions concentration before and after adsorption.
released light metal ions that was the compositions of activated carbon (Fig. 2) was carried out. This was performed in a batch system with initial concentrations of Cu(II) and Pb(II) of 10 mmol L1. All metal ions in the solution before and after sorption were analyzed by Inductively Coupled Plasma (ICP). The plots of the various metal ion concentrations were shown in Fig. 5. In this figure, the difference between the two adjacent bars is the amount of metals being adsorbed to or desorbed from the activated carbon. It was clear that there were only marginal amounts of light metals being desorbed when compared with the amount of heavy metals adsorbed. The percentage of ion exchange in the removal process is then calculated from P % ion exchange ¼
ðC 2;light C 1;light Þ 100 C 1;heavy C 2;heavy
(13)
where C1,light and C2,light are concentrations of light metal ions before and after sorption, respectively, and C1,heavy and C2,heavy the concentration of heavy metal ions before and after sorption, respectively. The calculation revealed that the percentages of ion exchange for Cu(II) and Pb(II) were 1.05% and 10.4%, respectively. Therefore ion exchange should not be a major mechanism for the uptakes of both heavy metals and the actual mechanism should be adsorption, although the uptake of Pb(II) was more contributed by ion exchange than Cu(II). It should be noted that ICP analysis was subject to some error of approximately 2%, therefore the difference in the % ion exchange could actually be due to the analytical error and not from the experiment itself. Therefore the results on % ion exchange for Cu(II) must be interpreted with extreme care. 4. Conclusions
Fig. 4. Maximum adsorption capacities for Cu(II) and Pb(II) at various molar ratios of initial metal concentrations (shaking rate = 200 rpm, 25 8C).
This work, again, illustrates the possibility in the use of activated carbon derived from Eucalyptus bark through phosphoric acid
470
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
activation method as an effective control of heavy metals in low strength wastewater. In single component sorption experiments, the activated carbon seemed to have higher affinity towards Pb(II) than Cu(II), although the difference in the sorption capacities of the two metals were not significant. This was perhaps because Pb(II) could bind with more functional groups than Cu(II). In the sorption of binary mixture, individual sorption capacity seemed to decrease when the secondary metal was presented in the system, particularly at high concentration. Within the range of the initial concentration employed in this work, a linear inverse dependency between the reduction in the sorption capacity and the initial concentration was observed. Acknowledgements The authors would like to acknowledge the financial support from the 90th Anniversary of Chulalongkorn University fund (Ratchadphiseksomphot Endowment Fund) and the Graduate School of Chulalongkorn University. References [1] S.H. Song, B.Y. Yeom, W.S. Shim, S.M. Hudson, T.S. Hwang, J. Ind. Eng. Chem. 13 (2007) 1009. [2] C. Jeon, K.H. Park, J. Ind. Eng. Chem. 13 (2007) 669. [3] P. Patnukao, P. Pavasant, Bioresour. Technol. 99 (2008) 8540.
[4] ASTM, Annual Book of ASTM Standards, 2001, D1762. [5] ASTM, Annual Book of ASTM Standards, 1999, D4607. [6] Japanese Standard Association, Japanese Industrial Standard Test Method for Activated Carbon, JIS K, 1991, D1470. [7] N. Ninlanon, Master thesis, Chulalongkorn University, Thailand, 1997. [8] V. Jindaphunphairoth, Master thesis, Chulalongkorn University, Thailand, 2000. [9] R. Isarasaenee, Master thesis, Chulalongkorn University, Thailand, 1996. [10] T. Suravattanasakul, Master thesis, Chulalongkorn University, Thailand, 1998. [11] B.S. Girgis, M.F. Isnak, Mater. Lett. 39 (1999) 107. [12] J.B. Castro, P.R. Bonelli, E.G. Cerrella, A.L. Cukierman, Ind. Eng. Chem. Res. 39 (2000) 4166. [13] Y. Diao, W.P. Walawender, L.T. Fan, Bioresour. Technol. 81 (2002) 45. [14] B.S. Girgis, S.S. Yunis, A.M. Soliman, Mater. Lett. 57 (2002) 164. [15] T. Vernersson, P.R. Bonelli, E.G. Cerrella, A.L. Cukierman, Bioresour. Technol. 83 (2002) 95. [16] A.A.M. Daifullah, B.S. Girgis, H.M.H. Gada, Colloids Surf. A: Physicochem. Eng. Aspects 235 (2004) 1. [17] (a) J.P. Chen, S. Wu, K.H. Chong, Carbon 41 (2003) 1979; (b) W. Lochananon, D. Chatsiriwech, J. Ind. Eng. Chem. 14 (2008) 84. [18] H. Jankowska, A. Swiatkowski, J. Choma, Active Carbon, Ellis Horwood, Warsaw, 1991. [19] W.W. Eckenfelder, Industrial Water Pollution Control, McGraw-Hill, Singapore, 2000. [20] Q. Feng, Q. Lin, F. Gong, S. Sugita, M. Shoya, J. Colloid Interface Sci. 278 (2004) 1. [21] S.Y. Quek, D.A.J. Wase, C.F. Forster, Water SA 24 (1998) 251. [22] I˙. Uzun, F. Gu¨zel, Turk. J. Chem. 24 (2000) 291. [23] C. Faur-Brasquet, Z. Reddad, K. Kadirvelu, P.L. Cloirec, Appl. Surf. Sci. 196 (2002) 356. [24] J.P. Chen, S. Wu, Langmuir 20 (2004) 2233. [25] P. Pavasant, R. Apiratikul, V. Sungkhum, P. Suthiparinyanont, S. Wattanachira, T. Marhaba, Bioresour. Technol. 97 (2003) 2321. [26] D. Mohan, K.P. Singh, Water Res. 36 (2002) 2304.
Contents lists available at ScienceDirect
Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec
Binary component sorption of Cu(II) and Pb(II) with activated carbon from Eucalyptus camaldulensis Dehn bark Apipreeya Kongsuwan a, Phussadee Patnukao b, Prasert Pavasant a,b,* a b
Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok, Thailand National Center of Excellence for Environmental and Hazardous Waste Management, Chulalongkorn University, Bangkok, Thailand
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 April 2008 Accepted 23 February 2009
The objective of this work is to illustrate the potential in the use of activated carbon in the binary component sorption of copper and lead ions. Eucalyptus bark was used as a precursor for the activated carbon which was prepared through the phosphoric acid activation process. This activated carbon was then used for the sorption of copper and lead ions. The quantity of the metal ions in the solution was measured with the Flame & Graphite Furnace Atomic Adsorption Spectrophotometer. The results indicated that the optimal pH for sorption was 5. The maximum sorption capacities for Cu(II) and Pb(II) were 0.45 and 0.53 mmol g1. Carboxylic, amine and amide groups were found to involve in the sorptions of Cu(II) and Pb(II). A major mechanism for the uptake of both heavy metals was proven not to be ion exchange but adsorption. In binary component sorptions, activated carbon still could sorb Pb(II) in a greater amount than Cu(II). However, the presence of the secondary metal ions suppressed the sorption of the primary metal ions. There seemed to have a linear inverse dependency between the sorption capacity and the concentration of the secondary metal ion. ß 2009 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
Keywords: Competitive sorption Heavy metal removal Sorption isotherm Adsorption Ion exchange
1. Introduction With a rapid increase in heavy industrial activities, pollution derived from uncontrolled escapes of heavy metals such as copper, nickel, chromium, and zinc has become serious [1,2]. Heavy metals are detected in waste streams from industries in large quantities. These heavy metals have harmful effects on human physiology and other biological systems when they exceed tolerance levels which are generally in the level of less than one part per million (ppm). Among the various treatment technologies, activated carbon sorption is commonly used because the design is simple and can be used effectively with low concentration waste streams (less than 100 ppm). Single component sorptions have extensively been investigated, however, relatively less research attention has focused on the binary component sorption. In such binary systems, sorption of heavy metal ions depended not only on specific surface properties of the adsorbent and the physicochemical parameters of the solution such as pH and initial metal ion concentration, but also on the interaction between heavy metal ion species. Competition
* Corresponding author at: Department of Chemical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok, Thailand. Tel.: +66 2 2186870; fax: +66 2 2186877. E-mail address: [email protected] (P. Pavasant).
among the different metal ions for the surface binding sites will certainly occur and change the sorption characteristics of the sorption of each ion species. Our recent work demonstrated a successful conversion of eucalyptus bark to activated carbon [3]. This work therefore intended to further employ this activated carbon product in the removal of heavy metal from the binary mixture solution. A synthetic mixture solution of copper and lead was arbitrarily selected as a modeled sorption system. Possible interactions between the metal ions and the surface of the adsorbent in the binary component sorption were described. 2. Experimental 2.1. Preparation of activated carbon The raw material, eucalyptus bark waste from the pulp mill, was impregnated into phosphoric acid (85% by weight) with the weight ratio of bark and phosphoric acid at 1:1. The mixture was carbonized in a muffle furnace at 500 8C for 1 h. The product was washed with hot distilled water until the pH of the leachate was 6 and dried in an oven at 105 8C for 4 h. Finally, powder activated carbon was crushed and sieved in the size ranged between mesh number 325 (0.045 mm) and 100 (0.150 mm) and stored in a desiccator.
1226-086X/$ – see front matter ß 2009 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jiec.2009.02.002
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
466
2.2. Activated carbon characteristics
Table 1 Proximate and ultimate analyses of eucalyptus bark and activated carbon.
The ultimate analysis of eucalyptus bark was performed in CHNS/O analyzer (PerkinElmer PE2400 Series II) and the proximate analysis was carried out following ASTM standards for chemical analysis of wood charcoal [4]. Characteristics of activated carbon from this work were examined. The yield of activated carbon was defined as the ratio of the weight of the activated carbon product to that of the original eucalyptus bark, both on a dry basis. Apparent (bulk) density of all samples was calculated as the ratio between weight and volume of packed dry material. The BET surface area was determined by nitrogen adsorption (196 8C) on surface area analyzer (Thermo Finnigan, Sorptomatic 1990). The cross-sectional area for nitrogen molecule was assumed to be 0.162 nm2. For sorption experiments, iodine number was determined according to the ASTM standard method [5]. Methylene blue number was determined according to the standard method [6]. The surface morphology of activated carbon was visualized via scanning electron microscopy (SEM), the corresponding SEM micrographs being analyzed with a scanning electron microscope (XL 30 ESEM FEG).
Parameter
2.3. Sorption 0.1 g of activated carbon was placed in a conical tube with 25 mL of 0.1 mmol L1 metal solutions. Metals in nitrate form (Cu(NO3)2 and Pb(NO3)2) were used to prepare the metal solution with initial concentration in the range from 0.1 to 10 mmol L1. The solution was mixed by a rotary shaker at 200 rpm, 25 8C for 1 h. Binary components of Cu(II) and Pb(II) were initially prepared as a mixture of equimolar concentration, e.g. binary components at 10 mmol L1 was prepared as a mixture of Cu(II) and Pb(II) both at 5 mmol L1. The range of concentration of heavy metals used to investigate sorption kinetics of the binary component systems was from 0.625 to 10 mmol L1. Loaded activated carbon was separated from solution by filter paper where the starting solutions and the filtrates were analyzed for the metal ions by Flame & Graphite Furnace Atomic Adsorption Spectrophotometer (AAS, ZEEnit 700, Analytik Jena, United Kingdom). Determination of sorption capacities [5] could be calculated from Eq. (1) q¼
VðC i C e Þ 1000W
% by weight Eucalyptus bark
Activated carbon
Proximate Moisture content Volatile matter Fixed carbon Ash content
10.50 75.05 13.10 1.35
7 22.78 65.34 4.88
Ultimate C H Others—N, O, P, S
41.36 4.67 53.97
62 4 34
carbon at 1000 magnification which confirmed amorphous and heterogeneous structures. The product contained relatively higher fixed carbon and lower volatile matter than those in the raw material as they were removed during the activation process. However, the overall yield of the process calculated from the ratio between the dry weights of the product and the raw material was only 30% (see Table 2), and the weight loss was due to the gasification of such volatile and some of the carbon in the raw materials. This level of yield was considered moderate when compared with those prepared with other raw materials such as mangrove wood (38.07%) [7], Eucalyptus camaldulensis Dehn (33.14%) [8], used tired (47.20%) [9], palm oil shell (19.66%) [10]. Table 2 also summarizes other properties of the activated carbon obtained from the technique employed in this work. The BET result illustrated that the activated carbon product had relatively high specific surface area (1240 m2 g1) when
(1)
where q is sorption capacity (mmol g1), Ci the initial metal concentration (mmol L1), Ce the concentration of metal at equilibrium (mmol L1), W the adsorbent dosage (g) and V the solution volume (L). In case of binary components systems, solution of copper–lead nitrate at 0.2 mmol L1 (equimolar concentration, Cu(II) = Pb(II) = 0.1 mmol L1) was added into the conical tube where the initial concentration of metal ions was in the range of 0.2–10 mmol L1. Note that all experiments were carried out in triplicate. 3. Results and discussion 3.1. Characteristics of eucalyptus bark raw material and activated carbon product Table 1 shows the proximate and ultimate analyses of eucalyptus bark. It was shown that the bark contained relatively high carbon and low ash, properties suitable as a raw material for the synthesis of activated carbon. After the activated process, the morphology of the carbon changed and the external surface of activated carbon obtained from scanning electron micrographs (SEMs) was shown in Fig. 1. This SEM image illustrates irregular structure with cracks and crevices on the surface of the activated
Fig. 1. SEM photograph of (a) eucalyptus bark and (b) activated carbon.
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
467
Table 2 Characteristics of activated carbon. Parameter
Value
Yield (%) Apparent (bulk) density (g cm3) BET specific surface area (m2 g1) Pore specific volume (cm3 g1) Average pore diameter (A˚)
30 0.251 1239.38 1.572 8.49
Pore size distribution Micropore (%) Mesopore (%) Iodine number (mg g1) Methylene blue (mg g1)
88.5 11.48 918 7 427 2
compared with those reported in literature such as cotton stalk (1032 m2 g1) [11], sugar cane bagasse (1132 m2 g1) [12], grain sorghum (1032 m2 g1) [13], peanut hull (1177 m2 g1) [14] Arundo donax cane (1333 m2 g1) [15], and rice husk (376 m2 g1) [16]. The difference could be due to the removal of organic components and minerals presented in the different raw materials which could react differently with phosphoric acid during the activation process [17]. This also led to a distinctive formation of pore structure in the final activated carbon product, and explained the weight loss during activation process, as phosphoric acid reacted with char and volatile matters resulting in gas products which diffused quickly out of the surface of particles and created pores [3]. The iodine and methylene blue numbers indicated the sorption capacities of the activated carbon. The iodine number indicated the sorption capacity of activated carbon in micropore (d < 200 nm) [18] and the methylene blue number represent the sorption capacity in mesopore (200 < d < 5000 nm). Table 2 illustrates that the activated carbon product from this work had a relatively high iodine number of about 918 7 mg g1, but on the other hand, a relatively low methylene blue number (427 2 mg g1), which meant that the product had more micropore than mesopore structures. The inorganic composition of activated carbon determined by microwave digester is shown in Fig. 2 and this demonstrates that the main compositions comprised alkaline and alkaline earth. These metals may be important for the sorption of heavy metals in the case that ion exchange process prevailed.
Fig. 3. Single and binary adsorption isotherms of Cu(II) and Pb(II) (shaking rate = 200 rpm, 25 8C).
significantly on the initial concentration of the metals and the isotherm results at 25 8C are displayed in Fig. 3. As a general observation, Pb(II) exhibited a higher maximum sorption capacity than Cu(II). Two mathematical expressions are commonly used to describe the isotherm of the sorption: Langmuir and Fruendlich equations [19]. The Langmuir isotherm is given by: qe ¼
qm bC 1 þ bC
where qm is the maximum sorption capacity (mg g1), b a constant related to bonding energy of sorption. The Freundlich isotherm is expressed as: qe ¼ K F C 1=n
Sorption equilibrium is established when the concentration of metal ion in the bulk solution (Ce) was in dynamic with that in the solid matrix (qe). This level of equilibrium concentration depends
(3)
where KF and n are constants. Table 3 reveals that experimental data fitted more favorably with Langmuir than Freundlich isotherms suggesting that the sorptions of Cu(II) and Pb(II) be potentially monolayer [20]. A comparison between maximum sorption capacities (qm) between the two metal ions illustrated that Cu(II) and Pb(II) had maximum sorption Table 3 Parameters of Langmuir and Freundlich isotherms. Heavy metal ion
3.2. Sorption
(2)
Cu(II) Pb(II)
Langmuir isotherm
Freundlich isotherm
qm (mmol g1)
b (L mmol1)
R
KF
1/n
R2
0.455 0.534
6.125 6.555
0.9897 0.9964
0.538 0.372
0.494 0.392
0.9041 0.7456
Fig. 2. Metal compositions in activated carbon.
2
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
468
Table 4 Comparison of reported maximum adsorption capacities. Author
Adsorbent
Heavy metal
qm (mmol g1)
This work Quek et al. [21]
Activated carbon Sago waste
Uzun et al. [22] Faur-Brasquet et al. [23]
Commercial activated carbon AC pretreated with NaCl
Chen et al. [17a]
0.450 0.200 0.530 0.230 0.174 0.147 0.040
Feng et al. [20]
Activated carbon pretreated with 1 M citric acid Rice husk ash
Cu(II) Cu(II) Pb(II) Pb(II) Cu(II) Pb(II) Cu(II)
Chen and Wu [24] Pavasant et al. [25]
GAC Dried Caulerpa lentillifera
Cu(II) Pb(II) Cu(II) Cu(II) Pb(II)
0.230 0.064 0.110 0.088 0.139
Amine N–H stretching N–H bending C–N stretching
Cu(II)
Pb(II)
* * * *
* * * *
*
*
*
*
* Indicates the potential functional group involved in the adsorption.
capacities of 0.45 and 0.53 mmol g1, respectively. The level of maximum sorption capacity of this activated carbon was considered relatively high when compared with the reported values summarized in Table 4. The functional groups related to the sorption were analyzed with FTIR where the results in Table 5 demonstrate that the activated carbon might contain more functional group binding sites for Pb(II) than for Cu(II). Moreover, Pb(II) sorption was found to have a higher b value indicating a stronger chemical and physical affinity with the activated carbon than Cu(II). The binary components isotherm models based on Langmuir assumption can be derived [25]. In case of binary component sorption, the sorption reaction can be expressed by two chemical reaction equations: K1
B þ M1 !BM 1 ; K2
B þ M2 !BM 2 ;
½BM1 ¼
K 1 ½Bt ½M1 1 þ K 1 ½M 1 þ K 2 ½M 2
(7)
½BM2 ¼
K 2 ½Bt ½M2 1 þ K 1 ½M 1 þ K 2 ½M 2
(8)
or in terms of sorption capacity as:
*
Amide N–H stretching C–O stretching
(6)
½Bt ¼ ½B þ ½BM1 þ ½BM2 : Eqs. (4)–(6) can be rearranged to:
Table 5 Possible functional groups involved with the adsorption of Cu(II) and Pb(II).
Carboxylic acid O–H stretching C5 5O stretching C–O stretching O–H bending
M1 and M2, respectively, and K1 and K2 the equilibrium constants of reactions. The total binding sites are the sum of the vacant and occupied sites:
K1 ¼
½BM1 ½B½M 1
(4)
K2 ¼
½BM2 ½B½M 2
(5)
where B are the free binding sites, M1 and M2 the first and second adsorbates in the solution, BM1 and BM2 the bound sites for metal
q1 ¼
b1 qm;1 C e;1 1 þ b1 C e;1 þ b2 C e;2
(9)
q2 ¼
b2 qm;2 C e;2 1 þ b1 C e;1 þ b2 C e;2
(10)
where qm is the maximum sorption capacity, b the affinity constant of Langmuir model, Ce the equilibrium concentration, q the sorption capacities, and subscripts 1 and 2 represent metal components 1 and 2, respectively. Fig. 3 shows the sorption isotherms of single and binary component sorptions. In this figure, the labels: Cu/Cu–Pb and Pb/ Cu–Pb: refer to the sorption isotherms of Cu(II) and Pb(II) from binary component systems, respectively: and the labels; Cu and Pb, are for the sorption isotherms of Cu(II) and Pb(II) from single component systems, respectively. Table 6 displays the equilibrium sorption capacities of single and binary component sorptions. It was also noticed that the sorption capacities of both metals for the single and binary components were similar at low initial concentration. For instance, qe for the sorption of Cu(II) at an initial concentration of 0.625 mmol L1 was almost the same as that for the binary sorption at an initial concentration of 1.25 mmol L1 (where initial Cu(II) = Pb(II) = 0.625 mmol L1). However, as the concentration became high, the sorption capacity for the binary system seemed to be lowered than that of the single component. For Cu(II) in particular, the equilibrium sorption capacity in the binary system seemed to level off at 2.16 101 mmol g1 (for systems prepared at equimolar ratio between the two metals). This could be because, at low concentration, there existed a large number of sorption sites compared with the amount of metal ion in the solution, and therefore the level of competition between the two metals was not high, and the sorption capacities for both metals remained the same as those from single component systems. However, as the concentration became high, the competition for the sorption sites became more intense and this resulted in a decrease in the sorption capacity for both metals. It should be noted that the sorption of Cu(II) in the binary system seemed to be leveled off as the initial concentration of Cu(II) reached 5 mmol L1. However, the sorption capacity of Pb(II) in such systems did not seem to reach a constant level with the range of concentration employed in this work.
Table 6 Comparison equilibrium adsorption capacity, qe (mmol g1), between single and binary adsorptions. Initial concentration (mmol L1)
Cu(II) (q0, mmol g1)
Cu(II) from binary componentsa (qmix, mmol g1)
qmix/q0 of Cu(II)
Pb(II) (q0, mmol g1)
Pb(II) from binary componentsa (qmix, mmol g1)
qmix/q0 of Pb(II)
10 5 2.5 1.25 0.625
4.59 101 3.64 101 3.00 101 2.04 101 1.27 101
2.16 101 2.16 101 2.11 101 1.24 101
0.5934 0.7200 1.0343 0.9764
5.31 101 4.57 101 3.89 101 2.95 101 1.30 101
3.79 101 3.10 101 3.02 101 1.29 101
0.8293 0.7969 1.0237 0.9923
a
Binary component is equimolar concentration, e.g. binary components at 10 mmol L1 was prepared from 5 mmol L1 of Cu(II) and Pb(II).
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
469
Langmuir isotherms as expressed in Eqs. (9) and (10) were selected to discuss experimental data to predict maximum sorption capacities. The highest achievable sorption capacities of Cu(II) and Pb(II) in binary components sorption were 0.181 and 0.357 mmol g1 in case of equimolar concentration, respectively. Table 5 indicates that Pb(II) could bind with a more variety of functional groups than Cu(II) and therefore the sorption capacity for Pb(II) was slightly higher than that of Cu(II). The maximum sorption capacities of Cu(II) and Pb(II) obtained from binary component sorption were less than those obtained from the single component system where qm,Cu = 0.45 mmol g1 and qm,Pb = 0.53 mmol g1. This could be due to the reasons given above. The effect of ionic interaction on the binary component sorption can be represented by the ratio of the sorption capacity for one metal ion in the presence of the other metal ion (binary component system), qmix, to the sorption capacity for the same metal when it was presented alone in the solution (single component system), q0 [26]. Three case scenarios could be formulated: (i) qmix/q0 > 1 the sorption was promoted by the presence of other metal ions, (ii) qmix/q0 = 1 there was no observable net interaction, and (iii) qmix/q0 < 1 sorption was suppressed by the presence of other metal ions. Table 6 shows qmix/q0 from the binary components sorption of Cu(II), all of which were almost less than unity. This indicated that Cu(II) sorption was suppressed by Pb(II) sorption. In the same manner, qmix/q0 of Pb(II) from binary component sorption were less than unity which also indicated that Pb(II) sorption was suppressed by Cu(II) sorption. To further illustrate this point, binary component sorption experiments were carried out using the mixture with nonequimolar initial concentration. In this case, the molar ratios of initial concentrations of Cu(II) and Pb(II) were varied from 2:0 to 0:2 with an increment of 0.5 mmol L1. The plot of maximum sorption capacities at various ratio of Cu(II) and Pb(II) as shown in Fig. 4 demonstrates that inhibition effect did actually occur throughout the examined concentration range. In other words, the presence of one metal species lowered the sorption capacity of the other. It is interesting to note that there existed a linear inverse dependency between the sorption capacity and the concentration of the secondary metal ion. The total sorption capacity could be well predicted (as illustrated as a top solid line in Fig. 4) provided that the ratio between the two metals was known. 3.3. Mechanism for removal of heavy metals Removal of heavy metals from wastewater could involve either adsorption or ion exchange, or both. To evaluate for the ion exchange mechanism of removal process, the measurement of the
Fig. 5. Metal ions concentration before and after adsorption.
released light metal ions that was the compositions of activated carbon (Fig. 2) was carried out. This was performed in a batch system with initial concentrations of Cu(II) and Pb(II) of 10 mmol L1. All metal ions in the solution before and after sorption were analyzed by Inductively Coupled Plasma (ICP). The plots of the various metal ion concentrations were shown in Fig. 5. In this figure, the difference between the two adjacent bars is the amount of metals being adsorbed to or desorbed from the activated carbon. It was clear that there were only marginal amounts of light metals being desorbed when compared with the amount of heavy metals adsorbed. The percentage of ion exchange in the removal process is then calculated from P % ion exchange ¼
ðC 2;light C 1;light Þ 100 C 1;heavy C 2;heavy
(13)
where C1,light and C2,light are concentrations of light metal ions before and after sorption, respectively, and C1,heavy and C2,heavy the concentration of heavy metal ions before and after sorption, respectively. The calculation revealed that the percentages of ion exchange for Cu(II) and Pb(II) were 1.05% and 10.4%, respectively. Therefore ion exchange should not be a major mechanism for the uptakes of both heavy metals and the actual mechanism should be adsorption, although the uptake of Pb(II) was more contributed by ion exchange than Cu(II). It should be noted that ICP analysis was subject to some error of approximately 2%, therefore the difference in the % ion exchange could actually be due to the analytical error and not from the experiment itself. Therefore the results on % ion exchange for Cu(II) must be interpreted with extreme care. 4. Conclusions
Fig. 4. Maximum adsorption capacities for Cu(II) and Pb(II) at various molar ratios of initial metal concentrations (shaking rate = 200 rpm, 25 8C).
This work, again, illustrates the possibility in the use of activated carbon derived from Eucalyptus bark through phosphoric acid
470
A. Kongsuwan et al. / Journal of Industrial and Engineering Chemistry 15 (2009) 465–470
activation method as an effective control of heavy metals in low strength wastewater. In single component sorption experiments, the activated carbon seemed to have higher affinity towards Pb(II) than Cu(II), although the difference in the sorption capacities of the two metals were not significant. This was perhaps because Pb(II) could bind with more functional groups than Cu(II). In the sorption of binary mixture, individual sorption capacity seemed to decrease when the secondary metal was presented in the system, particularly at high concentration. Within the range of the initial concentration employed in this work, a linear inverse dependency between the reduction in the sorption capacity and the initial concentration was observed. Acknowledgements The authors would like to acknowledge the financial support from the 90th Anniversary of Chulalongkorn University fund (Ratchadphiseksomphot Endowment Fund) and the Graduate School of Chulalongkorn University. References [1] S.H. Song, B.Y. Yeom, W.S. Shim, S.M. Hudson, T.S. Hwang, J. Ind. Eng. Chem. 13 (2007) 1009. [2] C. Jeon, K.H. Park, J. Ind. Eng. Chem. 13 (2007) 669. [3] P. Patnukao, P. Pavasant, Bioresour. Technol. 99 (2008) 8540.
[4] ASTM, Annual Book of ASTM Standards, 2001, D1762. [5] ASTM, Annual Book of ASTM Standards, 1999, D4607. [6] Japanese Standard Association, Japanese Industrial Standard Test Method for Activated Carbon, JIS K, 1991, D1470. [7] N. Ninlanon, Master thesis, Chulalongkorn University, Thailand, 1997. [8] V. Jindaphunphairoth, Master thesis, Chulalongkorn University, Thailand, 2000. [9] R. Isarasaenee, Master thesis, Chulalongkorn University, Thailand, 1996. [10] T. Suravattanasakul, Master thesis, Chulalongkorn University, Thailand, 1998. [11] B.S. Girgis, M.F. Isnak, Mater. Lett. 39 (1999) 107. [12] J.B. Castro, P.R. Bonelli, E.G. Cerrella, A.L. Cukierman, Ind. Eng. Chem. Res. 39 (2000) 4166. [13] Y. Diao, W.P. Walawender, L.T. Fan, Bioresour. Technol. 81 (2002) 45. [14] B.S. Girgis, S.S. Yunis, A.M. Soliman, Mater. Lett. 57 (2002) 164. [15] T. Vernersson, P.R. Bonelli, E.G. Cerrella, A.L. Cukierman, Bioresour. Technol. 83 (2002) 95. [16] A.A.M. Daifullah, B.S. Girgis, H.M.H. Gada, Colloids Surf. A: Physicochem. Eng. Aspects 235 (2004) 1. [17] (a) J.P. Chen, S. Wu, K.H. Chong, Carbon 41 (2003) 1979; (b) W. Lochananon, D. Chatsiriwech, J. Ind. Eng. Chem. 14 (2008) 84. [18] H. Jankowska, A. Swiatkowski, J. Choma, Active Carbon, Ellis Horwood, Warsaw, 1991. [19] W.W. Eckenfelder, Industrial Water Pollution Control, McGraw-Hill, Singapore, 2000. [20] Q. Feng, Q. Lin, F. Gong, S. Sugita, M. Shoya, J. Colloid Interface Sci. 278 (2004) 1. [21] S.Y. Quek, D.A.J. Wase, C.F. Forster, Water SA 24 (1998) 251. [22] I˙. Uzun, F. Gu¨zel, Turk. J. Chem. 24 (2000) 291. [23] C. Faur-Brasquet, Z. Reddad, K. Kadirvelu, P.L. Cloirec, Appl. Surf. Sci. 196 (2002) 356. [24] J.P. Chen, S. Wu, Langmuir 20 (2004) 2233. [25] P. Pavasant, R. Apiratikul, V. Sungkhum, P. Suthiparinyanont, S. Wattanachira, T. Marhaba, Bioresour. Technol. 97 (2003) 2321. [26] D. Mohan, K.P. Singh, Water Res. 36 (2002) 2304.