Desalination 225 (2008) 1328

Adsorption of basic dye using activated carbon prepared from oil palm shell: batch and fixed bed studies
I.A.W. Tan, A.L. Ahmad, B.H. Hameed*
School of Chemical Engineering, University Science Malaysia, Engineering Campus, 14300 Nibong Tebal, Penang, Malaysia Tel. +60 (4) 599-6422; Fax +60 (4) 594-1013; email: chbassim@eng.usm.my

Received 7 December 2006; accepted 29 July 2007

Abstract The feasibility of activated carbon prepared from oil palm shell to remove methylene blue from aqueous solutions was investigated through batch and column studies. Batch experiments were carried out to study the adsorption isotherm and kinetics at 30C, with the initial concentration of 50500 mg/l and solution pH of 6.5. Equilibrium data were fitted to Langmuir, Freundlich, Temkin and Dubinin-Radushkevich isotherm models. The equilibrium data were best represented by the Langmuir isotherm model, with maximum monolayer adsorption capacity of 243.90 mg/g at 30C. The adsorption process was found to be exothermic in nature. The kinetic data were fitted to pseudo-first-order, pseudo-second-order and intraparticle diffusion models, and it was found to follow closely the pseudo-second-order model. In column experiments, the effects of the initial dye concentration, bed height and flow rate on methylene blue adsorption were studied. The highest bed capacity of 40.86 mg/g was obtained using 100 mg/l initial dye concentration, 6 cm bed height and 20 ml/min flow rate. Keywords: Adsorption; Oil palm shell; Methylene blue; Equilibrium; Kinetics

1. Introduction The discharge of dyes in the environment is worrying for both toxicological and esthetical reasons as dyes impede light penetration, damage the quality of the receiving streams and are toxic to food chain organisms [1]. Industries such as textile, leather, paper, plastics, etc., are some of the
*Corresponding author.

sources for dye effluents [2]. Methylene blue (MB), a cationic dye, is the most commonly used substance for colouring among all other dyes of its category. MB can cause eye burns, and if swallowed, it causes irritation to the gastrointestinal tract with symptoms of nausea, vomiting and diarrhea. It may also cause methemoglobinemia, cyanosis, convulsions and dyspnea if inhaled [3]. Hence, the treatment of effluents containing such

0011-9164/08/$ See front matter 2008 Elsevier B.V. All rights reserved
doi:10.1016/j.desal.2007.07.005

14

I.A.W. Tan et al. / Desalination 225 (2008) 1328

dye is of interest due to its harmful impacts on receiving waters. Adsorption onto activated carbon has been found to be superior for wastewater treatment compared to other physical and chemical techniques, such as flocculation, coagulation, precipitation and ozonation as they possess inherent limitations such as high cost, formation of hazardous by-products and intensive energy requirements [1]. However, commercially available activated carbons are still considered expensive [4]. This is due to the use of non-renewable and relatively expensive starting material such as coal, which is unjustified in pollution control applications [5]. Therefore, in recent years, this has prompted a growing research interest in the production of activated carbons from renewable and cheaper precursors which are mainly industrial and agricultural by-products, such as date pits [6], silk cotton hull and maize [7], jute fiber [3], groundnut shell [8], corncob [9], bamboo [10], rattan sawdust [11] and oil palm fibre [12]. Malaysia is the largest exporter of palm oil in the international market. One of the significant problems in the palm fruit processing is managing of the wastes generated during the processes. Palm oil mills in Malaysia produce about 4.3 million tonnes of shell annually [13]. In practice, this biomass is burned in incinerators by palm oil mills which creates environmental pollution problems in nearby localities and also offers a limited value to the industry. In order to make better use of this cheap and abundant agricultural waste, it is proposed to convert oil palm shell into activated carbon. Furthermore, very few papers have reported the dye adsorption using continuous flow conditions, which are more useful in a large-scale textile wastewater treatment. Therefore, in this study, the adsorption of MB on oil palm shell-based activated carbon was conducted in both batch and continuous conditions using a laboratory scale fixed bed column. The focus of this research was to evaluate the adsorption potential of the oil palm shell-based

activated carbon in removing MB from aqueous solutions through batch and fixed bed experiments. The equilibrium, kinetic and dynamic behaviours of the adsorption process were then evaluated to study the adsorption mechanism of MB molecules onto the prepared activated carbon. 2. Materials and methods 2.1. Methylene blue Methylene blue (MB) supplied by SigmaAldrich (M) Sdn Bhd, Malaysia was used as an adsorbate and was not purified prior to use. MB was chosen for this study because of its known strong adsorption onto solids. Deionized water was used to prepare all the solutions and reagents. Some properties of the MB dye, together with its chemical structure, are listed in Table 1. 2.2. Preparation and characterization of activated carbon Oil palm shell used for preparation of activated carbon was obtained from United Oil Palm Industries Sdn Bhd, Malaysia. The precursor was first washed to remove dirt from its surface and then dried overnight at 105C. The dried shell was crushed and sieved to desired mesh size of 1
Table 1 Some properties of the MB used
Properties Chemical structure

Chemical formula Molecular weight Type Solubility Solution pH Wave length

C16H18ClN3S.3H2O 373.9 g/mol Basic dye Soluble in water 6.5 668 nm

I.A.W. Tan et al. / Desalination 225 (2008) 1328

15

2 mm, then carbonized in a stainless steel vertical tubular reactor placed in a tube furnace under purified nitrogen (99.995%) flow of 150 cm3/min at 700C, with a heating rate of 10C/min. The sample was held at the carbonization temperature for 2 h. The char produced was soaked in a potassium hydroxide (KOH) solution with an impregnation ratio of 1:1. The mixture was then dehydrated in an oven overnight at 105C to remove moisture and then activated under the same conditions as carbonization, but to a final temperature of 850C. Once the final temperature was reached, the nitrogen gas flow was switched to carbon dioxide and activation was held for 2 h. The activated product was then cooled to room temperature under nitrogen flow and washed with hot deionized water and 0.1 molar hydrochloric acid until the pH of the washing solution reached 67. Textural characterization of the prepared activated carbon was carried out by N2 adsorption at 77 K using Autosorb I (Quantachrome Corporation, USA). The Brunauer-Emmett-Teller (BET) surface area and total pore volume of the prepared activated carbon were then determined [14,15]. Scanning electron microscopy (SEM) analysis was carried out for the precursor and the prepared activated carbon to study their surface textures and the development of porosity. In addition, the surface functional groups of the prepared activated carbon were detected by Fourier Transform Infrared (FTIR) spectroscope (FTIR-2000, Perkin Elmer). The spectra were recorded from 4000 to 400 cm1. 2.3. Batch equilibrium studies Adsorption tests were performed in a set of 43 Erlenmeyer flasks (250 ml) where 100 ml of MB solutions with initial concentrations of 50 500 mg/l were placed in these flasks. Equal mass of 0.1 g of the prepared activated carbon with the particle size of 200 m was added to each flask and kept in an isothermal shaker of 120 rpm at 30C for 30 h to reach equilibrium. The pH of the

solutions was natural (pH 6.5). Aqueous samples were taken from the solution and the concentrations were analyzed. All samples were filtered prior to the analysis in order to minimize the interference of the carbon fines with the analysis. The concentrations of MB in the supernatant solution before and after adsorption were determined using a double beam UV-visible spectrophotometer (Shimadzu, Japan) at 668 nm. Each experiment was duplicated under identical conditions. The amount of adsorption at equilibrium, qe (mg/g), was calculated by:
qe =

( C0 Ce ) V
W

(1)

where C0 and Ce (mg/l) are the liquid phase concentrations of dye at the initial and equilibrium conditions, respectively. V is the volume of the solution (l) and W is the mass of dry adsorbent used (g). 2.4. Batch kinetic studies The procedure of kinetic tests was identical to that of the equilibrium tests. The aqueous samples were taken at preset time intervals and the concentrations of MB were similarly measured. The amount of adsorption at time t, qt (mg/g), was calculated by:
qt =

( C0 Ct )V
W

(2)

where C0 and Ct (mg/l) are the liquid-phase concentrations of dye at the initial time and at any time t, respectively. V is the volume of the solution (l) and W is the mass of dry adsorbent used (g). 2.5. Column studies Continuous flow adsorption studies were conducted in a glass column made of Pyrex glass tube of 1.2 cm inner diameter and 19.5 cm height. At the bottom of the column, a stainless sieve was

16

I.A.W. Tan et al. / Desalination 225 (2008) 1328

attached followed by a layer of glass wool. A known quantity of the prepared activated carbon was packed in the column to yield the desired bed height of the adsorbent (3 and 6 cm). The column was then filled up with 5 mm size glass beads in order to provide a uniform flow of the solution through the column. Dye solution of known concentrations (50, 100 and 150 mg/l) at pH 6.5 was pumped upward through the column at a desired flow rate (20 and 30 ml/min) controlled by a peristaltic pump (Masterflex, Cole-Parmer Instrument Co.). The MB solutions at the outlet of the column were collected at regular time intervals and the concentration was measured using a double beam UV-visible spectrophotometer (Shimadzu, Japan) at 668 nm. All the experiments were carried out at room temperature (28 1C).

(a)

3. Results and discussion 3.1. Characterization of prepared activated carbon The BET surface area of the prepared activated carbon was found to be 596.20 m2/g, with the total pore volume of 0.34 cm3/g. Figs. 1a and 1b show the SEM images of the precursor and the derived activated carbon, respectively. Large and well-developed pores were clearly found on the surface of the activated carbon, compared to the original precursor. This might be due to the activation process used, which involved both chemical and physical activating agents of KOH and CO2. Pore development in the char during pyrolysis was also important as this would enhance the surface area and pore volume of the activated carbon by promoting the diffusion of KOH and CO2 molecules into the pores and thereby increasing the KOH-carbon and CO2-carbon reactions, which would then create more pores in the activated carbon. Stavropoulos and Zabaniotou [15] stated that KOH is dehydrated to K2O, which reacts with CO2 produced by the water-shift reaction, to give K2CO3. Intercalation of metallic potassium ap-

(b) Fig. 1. SEM micrographs: (a) raw oil palm shell (1000); (b) oil palm shell-based activated carbon (1000).

peared to be responsible for the drastic expansion of the carbon material and hence the creation of a large surface area and high pore volume. Fig. 2 displays the FTIR spectra obtained for the prepared activated carbon. The spectra displayed the following bands: 3400 cm1: OH stretching vibrations 2095 cm1: CC stretching vibrations 1559 cm1: C=C stretching vibration in aromatic rings 1086 cm1: COH stretching vibrations

I.A.W. Tan et al. / Desalination 225 (2008) 1328

6.30 6.2

17

6.0

5.8

5.6

5.4

5.2

%T 5.0

4.8

4.6

4.4

4.2

2094.51
1559.40

4.0
1086.36

3.80 4000.0 3600 3200 2800 2400 2000 1800 cm-1 1600 1400 1200 1000 800 600 450.0

Fig. 2. FTIR spectra of oil palm shell-based activated carbon.

The FTIR spectra obtained were in agreement with the results reported in the study carried out on activated carbons prepared from rice straws [16] and cherry stones [17]. 3.2. Effect of contact time and initial dye concentration on adsorption equilibrium Fig. 3 shows the adsorption capacity vs. the adsorption time at various initial MB concentrations at 30C. It indicated that the contact time needed for MB solutions with initial concentrations of 50200 mg/l to reach equilibrium was 4 5.5 h. For MB solutions with initial concentrations of 300500 mg/l, equilibrium time of 30 h was required. As can be seen from Fig. 3, the amount of MB adsorbed onto the activated carbon increased with time and, at some point in time, it reached a constant value beyond which no more MB was further removed from the solution. At

this point, the amount of the dye desorbing from the activated carbon was in a state of dynamic equilibrium with the amount of the dye being adsorbed onto the activated carbon. Similar phenomenon was also observed in adsorption of reactive dyes as well as pesticides from aqueous solutions using oil shale ash [18,19]. The amount of dye adsorbed at the equilibrium time reflects the maximum adsorption capacity of the adsorbent under those operating conditions. In this study, the adsorption capacity at equilibrium (qe) increased from 46 to 241 mg/g with an increase in the initial dye concentrations from 50 to 400 mg/l and decreased slightly to 227 mg/g for the initial concentration of 500 mg/l. Three consecutive mass transport steps are associated with the adsorption of solute from the solution by porous adsorbent [20]. First, the adsorbate migrates through the solution, i.e., film diffusion, followed by the solute movement from

18
400

I.A.W. Tan et al. / Desalination 225 (2008) 1328

300
qt (mg/g)

50 mg/l 200 mg/l 400 mg/l

100 mg/l 300 mg/l 500 mg/l

200

100

0 0 10 20 t (h) 30 40

Fig. 3. Adsorption capacity vs. adsorption time at various initial methylene blue concentrations at 30C.

the particle surface into the interior site by pore diffusion and finally the adsorbate is adsorbed into the active sites at the interior of the adsorbent particle. This phenomenon takes a relatively long contact time. A similar phenomenon was observed for the adsorption of MB from aqueous solution on jute fiber carbon and the equilibrium time was 250 min [3]. 3.3. Effect of temperature on adsorption equilibrium Fig. 4 shows the adsorption equilibrium vs. temperature at various initial MB concentrations. It was found that for low MB concentrations (50 100 mg/l), the effect of temperature on the adsorption equilibrium was not significant. However, for higher MB concentrations of 200 500 mg/l, the adsorption equilibrium decreased with the increase in temperature, indicating the exothermic nature of the adsorption reaction. This decrease in adsorption capacity with the increase in temperature is known to be due to the enhancement of the desorption step in the sorption mechanism. It is also due to the weakening of sorptive

forces between the active sites on the activated carbon and the dye species, and also between adjacent dye molecules on the sorbed phase. Similar observation was also reported for adsorption of MB using fly ash [21]. 3.4. Adsorption isotherms The adsorption isotherm indicates how the adsorption molecules are distributed between the liquid phase and the solid phase when the adsorption process reaches an equilibrium state. The analysis of the isotherm data by fitting them to different isotherm models is an important step to find a suitable model that can be used for design purposes [22]. Adsorption isotherm is basically important to describe how solutes interact with adsorbents, and is critical in optimizing the use of adsorbents. Adsorption isotherm study was carried out on four isotherm models: the Langmuir, Freundlich, Temkin and DubininRadushkevich isotherm models. The applicability of the isotherm equation to the adsorption study done was compared by judging the correlation coefficients, R2.

I.A.W. Tan et al. / Desalination 225 (2008) 1328

500
50 mg/l 100 mg/l 400 mg/l 200 mg/l 500 mg/l

19

400 qe (mg/g) 300 200 100 0 300

300 mg/l

305

310 T (K)

315

320

325

Fig. 4. Effect of temperature on adsorption equilibrium at various initial MB concentrations.

The Langmuir isotherm assumes monolayer adsorption onto a surface containing a finite number of adsorption sites of uniform strategies of adsorption with no transmigration of adsorbate in the plane of surface [23]. The linear form of the Langmuir isotherm equation is given as:

sionless equilibrium parameter (RL) [23]. The parameter is defined by:

RL =

1 1 + bC0

(4)

Ce 1 1 = + Ce qe Qo b Qo

(3)

where Ce is the equilibrium concentration of the adsorbate (mg/l), qe is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g), Qo and b are Langmuir constants related to the adsorption capacity and the rate of adsorption, respectively. When Ce/qe was plotted against Ce, a straight line with the slope of 1/Qo was obtained (figure not shown). The correlation coefficient R2 of 0.99 indicated that the adsorption data of MB on the activated carbon was well fitted to the Langmuir isotherm. The Langmuir constants b and Qo were calculated from Eq. (3) and their values are shown in Table 2. The essential characteristics of the Langmuir isotherm can be expressed in terms of a dimen-

where b is the Langmuir constant and C0 is the highest dye concentration (mg/l). The value of RL indicates the type of the isotherm to be either unfavorable (RL > 1), linear (RL = 1), favorable (0 < RL <1) or irreversible (RL = 0). The values of RL were found to be 0.002, 0.015 and 0.009, respectively, at the solution temperatures of 30, 40 and 50C. This again confirmed that the Langmuir isotherm was favorable for adsorption of MB onto the activated carbon under the conditions used in this study. The Freundlich isotherm, on the other hand, assumes heterogeneous surface energies, in which the energy term in the Langmuir equation varies as a function of the surface coverage [23]. The well-known logarithmic form of the Freundlich isotherm is given by the following equation:
log qe = log K F + (1 n ) log Ce

(5)

20

I.A.W. Tan et al. / Desalination 225 (2008) 1328

Table 2 Langmuir, Freundlich, Temkin and Dubinin-Radushkevich isotherm model constants and correlation coefficients for adsorption of MB onto prepared activated carbon
Isotherms Langmuir (Ce/qe) = (1/Q0b) + (1/Q0) Ce Solution temperature (K) 303 313 323 303 313 323 303 313 323 303 313 323 Constants Qo (mg/g) 243.90 158.73 156.25 KF (mg/g (l/mg)1/n) 99.36 61.59 69.47 A (l/g) 98.97 65.16 121.40 qs (mg/g) 261.86 131.01 133.18 b (l/mg) 0.93 0.13 0.22 1/n 0.19 0.17 0.16 B 25.91 15.73 15.15 E 1581.14 2886.75 2672.61 R2

0.999 0.998 0.998 0.588 0.993 0.965 0.726 0.974 0.974 0.841 0.837 0.833

Freundlich log qe = log KF + (1/n) log Ce

Temkin qe = (RT/b) ln (ACe))

DubininRadushkevich qe = qs exp (B2)

where Ce is the equilibrium concentration of the adsorbate (mg/l), qe is the amount of adsorbate adsorbed per unit mass of adsorbent (mg/g), KF and n are Freundlich constants with n giving an indication of how favorable the adsorption process is. KF (mg/g (l/mg)1/n) is the adsorption capacity of the adsorbent which can be defined as the adsorption or distribution coefficient and represents the quantity of dye adsorbed onto activated carbon for a unit equilibrium concentration. The slope of 1/n ranging between 0 and 1 is a measure of the adsorption intensity or surface heterogeneity, becoming more heterogeneous as its value gets closer to zero [24]. A value for 1/n below one indicates a normal Langmuir isotherm while 1/n above one is indicative of cooperative adsorption [25]. The plot of log qe vs. log Ce (figure not shown) gave a straight line with the slope of 1/n with the value of 0.19, which showed that the adsorption of MB on the activated carbon was

favorable. Accordingly, the Freundlich constants KF and n were calculated and are listed in Table 2. Temkin and Pyzhev considered the effects of indirect adsorbate/adsorbate interactions on adsorption isotherms. The heat of adsorption of all the molecules in the layer would decrease linearly with coverage due to adsorbate/adsorbate interactions [26]. The Temkin isotherm has been used in the form as follows:
qe = ( RT / b ) ln ( ACe )

(6)

where RT/b = B. The constants A and B, together with the R2 values are as well shown in Table 1. Another popular equation for the analysis of isotherms of a high degree of rectangularity is that proposed by Dubinin and Radushkevich as follows [27]:
qe = qs exp ( B 2 )

(7)

I.A.W. Tan et al. / Desalination 225 (2008) 1328

21

where can be correlated:

1 = RT ln 1 + Ce

(8)

The constant B gives the mean free energy E of sorption per molecule of the sorbate when it is transferred to the surface of the solid from infinity in the solution and can be computed by using the relationship:
1 E= 2B

(mg/g) at equilibrium and at time t (h), respectively and k1 is the rate constant adsorption (1/h). Values of k1 were obtained from the slopes of the linear plots of ln (qe qt) vs. t in Fig. 5a. The R2 values obtained for all MB concentrations were relatively high, however the deviations of the calculated qe values from the experimental qe values were relatively large (Table 3), indicating that this model was not applicable to the kinetic data. On the other hand, the pseudo-second-order equation [29] based on equilibrium adsorption is expressed as:

(9)

where R is the gas constant (8.314 J/mol K) and T is the absolute temperature. A plot of ln qe vs. 2 (figure not shown) enables the constants qs and E to be determined. Table 2 summarizes all the constants and correlation coefficients, R2 of the four isotherm models at the three temperatures studied. The Langmuir model yielded the best fit with R2 which were higher than 0.99 at all the three temperatures. Conformation of the experimental data into the Langmuir isotherm equation indicated the homogeneous nature of oil palm shell-based activated carbon surface, i.e., each dye molecule/oil palm shell carbon adsorption had equal adsorption activation energy. The results also demonstrated the formation of monolayer coverage of dye molecules at the outer surface of the oil palm shellbased activated carbon. Similar observation was reported by the adsorption of MB onto activated carbons prepared from jute fiber [3], corncob [9], rattan sawdust [11] and oil palm fibre [12]. 3.5. Adsorption kinetics The rate constant of adsorption is determined from the pseudo-first-order equation given by Langergren and Svenska [28] as:
ln ( qe qt ) = ln qe k1t

t 1 1 = + t 2 qt k2 qe qe

(11)

where k2 (g/mg h) is the rate constant of secondorder adsorption. If the second-order kinetics is applicable, the plot of t/qt vs. t should show a linear relationship. qe and k2 can then be determined from the slope and the intercept of the plot. This procedure is more likely to predict the behavior over the whole range of adsorption. The linear plot of t/qt vs. t, as shown in Fig. 5b, shows a good agreement between the experimental and the calculated qe values (Table 3). Besides, the correlation coefficients for the second-order kinetic model were greater than 0.99 for all MB concentrations, indicating the applicability of the second-order model to describe the adsorption process of MB on the prepared activated carbon. Intraparticle diffusion model based on the theory proposed by Weber and Morris [30] was tested to identify the diffusion mechanism. According to this theory:
qt = k p t 1/ 2

(12)

(10)

where qe and qt are the amounts of MB adsorbed

where kp (mg/g h1/2), the intraparticle diffusion rate constant, is obtained from the slope of the straight line of qt vs. t1/2 (Fig. 5c). The R2 values obtained were low compared to those obtained from pseudo-first order and pseudo-second order kinetic models. As can be seen from Fig. 5c, the

22 (a)

I.A.W. Tan et al. / Desalination 225 (2008) 1328

20 15 10 5 ln (qe-q t ) 0 0 -5 -10 -15 -20 t (h) 5 10 15 20 25 30

50 mg/l 200 mg/l 400 mg/l

100 mg/l 300 mg/l 500 mg/l

(b)

1.0 0.8 t/qt (h g/mg) 0.6 0.4 0.2 0.0 0

50 mg/l 200 mg/l 400 mg/l

100 mg/l 300 mg/l 500 mg/l

10

20 t (h)

30

40

(c)

800 50 mg/l 200 mg/l 400 mg/l 100 mg/l 300 mg/l 500 mg/l

600 qt (mg/g)

400

200

0 0 1 2 t
1/2

3 (h )
1/2

Fig. 5. Kinetics for adsorption of methylene blue onto prepared activated carbon at 30C: (a) pseudo-first-order; (b) pseudo-second-order; (c) intraparticle diffusion model.

I.A.W. Tan et al. / Desalination 225 (2008) 1328

23

Table 3 Comparison of the pseudo-first-order, pseudo-second-order and intraparticle diffusion models for different initial MB concentrations at 30C
Initial MB concentration (mg/l) 50 100 200 300 400 500 qe, exp (mg/g) Pseudo-first-order kinetic model qe, cal (mg/g) 45.99 90.68 188.14 239.51 240.94 227.16 7.73 13.66 80.53 129.50 118.29 117.44 k1 (1/h) 0.27 0.32 0.26 0.12 0.12 0.13 R2 0.86 0.88 0.96 0.90 0.84 0.91 Pseudo-second-order kinetic model qe, cal (mg/g) 46.08 90.91 188.68 243.90 243.90 232.56 k2 (g/mg h) 0.277 0.202 0.017 0.005 0.005 0.005 R2 1.00 1.00 0.99 0.99 0.99 0.99 Intraparticle diffusion model qe, cal (mg/g) 52.91 104.45 215.53 264.25 267.27 250.62 kp (mg/g h1/2) 5.3925 10.518 25.508 35.322 35.022 32.215 R2 0.40 0.39 0.58 0.75 0.72 0.70

linear line did not pass through the origin and this deviation from the origin or near saturation might be due to the difference in the mass transfer rate in the initial and final stages of adsorption [31]. 3.6. Validity of kinetic models The adsorption kinetics of MB onto the prepared activated carbon was verified at different initial concentrations. The validity of each model was determined by the sum of squared errors (SSE, %) given by:
SSE,% =

dicts the behavior over the whole range of studies, strongly supporting the validity and agrees with chemisorption being rate-controlling [32]. 3.7. Column studies The results of MB adsorption on the prepared activated carbon using a continuous system were presented in the form of breakthrough curves which showed the loading behaviours of MB to be adsorbed from the solution expressed in terms of normalized concentration defined as the ratio of the outlet MB concentration to the inlet MB concentration as a function of time (C/C0 vs. t). The total adsorbed MB quantity, qtotal (mg/g) in the column for a given inlet concentration and flow rate was calculated from Eq. (14) [33]:

(q

e ,exp

qe,cal )

(13)

where N is the number of data points. The lower the value of SSE indicates the better a fit is. It was found that the pseudo-secondorder kinetic model yielded the lowest SSE value of 3.10% compared to 98.45% for the pseudofirst-order kinetic model and 21.77% for the intraparticle diffusion model. This agrees with the R2 values obtained earlier and proves that the adsorption of MB onto the oil palm shell-based activated carbon could be best described by the pseudo-second-order kinetic model which is based on the equilibrium chemical adsorption that pre-

qtotal =

Q 1000

t = ttotal

t =0

Cad dt

(14)

where Cad (mg/l) is the adsorbed MB concentration (inlet concentration, C0 outlet concentration, C), Q is the volumetric flow rate (ml/min) and ttotal is the total flow time (min). Equilibrium dye uptake in the column or maximum capacity of the column (qeq) was defined by Eq. (15) as the total amount of dye adsorbed (qtotal)

24

I.A.W. Tan et al. / Desalination 225 (2008) 1328

per g of the adsorbent (X) at the end of the total flow time [33].

qeq =

qtotal X

(15)

3.7.1. Effect of initial dye concentration The effect of a variation from 50 to 150 mg/l of the initial MB concentration used with the same adsorbent bed height of 6 cm and solution flow rate of 20 ml/min is shown by the breakthrough curve in Fig. 6. At the highest MB concentration of 150 mg/l, the activated carbon bed was exhausted in the shortest time of less than 3 h leading to the earliest breakthrough. The breakpoint time decreased with increasing the inlet concentration as the binding sites became more quickly saturated in the column. This indicated that an increase in the concentration could modify the adsorption rate through the bed. A decrease in the initial MB concentration gave an extended breakthrough curve indicating that a higher volume of the solution could be treated. This was due to the fact that a lower concentration gradient caused a slower transport due to a decrease in the diffu-

sion coefficient or mass transfer coefficient. Similar trends were obtained for biosorption of phenol by immobilized activated sludge [33], removal of lead (II) ions using immobilized Pinus sylvestris sawdust [34] and biosorption of Acid Blue 15 using fresh water macroalga Azolla filiculoides [1]. The adsorption capacity was expected to increase with increasing the initial concentration because a high concentration difference provides a high driving force for the adsorption process. However, the experimental results showed that the highest adsorption capacity was obtained using MB solution of 100 mg/l initial concentration, with the adsorption capacity of 40.86 mg/g. All the adsorption capacities and the exhaustion times obtained are listed in Table 4. 3.7.2. Effect of activated carbon bed height Fig. 7 shows the breakthrough curve obtained for MB adsorption on the prepared activated carbon for two different bed heights of 3 and 6 cm, (2.0932 and 4.5055 g), at a constant flow rate of 20 ml/min and MB initial concentration of 100 mg/l. Both the breakthrough and exhaustion time increased with increasing the bed height. A higher

1.0 0.8 0.6 0.4 0.2 0.0 0 50 100 t (min) 150 200 50 ppm 100 ppm 150 ppm

Fig. 6. Breakthrough curves for MB adsorption on prepared activated carbon at different initial dye concentrations (bed height = 6 cm, flow rate = 20 ml/min, temperature = 28 1C).

C/Co

I.A.W. Tan et al. / Desalination 225 (2008) 1328 Table 4 Column data parameters obtained at different initial MB concentrations, bed heights and flow rates (T = 28 1C)

25

Initial MB concentration (mg/l) 50 100 150 100 100

Bed height (cm) 6 6 6 3 6

Flow rate (ml/min) 20 20 20 20 30

Complete bed exhaustion time (min) 420 300 170 150 170

Bed capacity, qeq (mg/g) 38.29 40.86 24.84 18.24 33.84

1.0 0.8 0.6 0.4 0.2 0.0 0 50 100 t (min) 150 200 3 cm 6 cm

Fig. 7. Breakthrough curves for MB adsorption on prepared activated carbon at different bed heights (initial dye concentration = 100 mg/l, flow rate = 20 ml/min, temperature = 28 1C).

C/Co

MB uptake was also observed at a higher bed height due to the increase in the specific surface of the activated carbon which provided more fixation binding sites for the dye to adsorb. The increase in the adsorbent mass in a higher bed provided a greater service area which would lead to an increase in the volume of the solution treated. Taty-Costodes et al. [34] reported in their work that when the bed height was reduced, axial dispersion phenomena predominated in the mass transfer and reduced the diffusion of the solute, and therefore the solute had not enough time to diffuse into the whole of the adsorbent mass. The

total MB adsorbed together with the exhaustion times are also listed in Table 4. 3.7.3. Effect of the solution flow rate The effect of the flow rate on the adsorption of MB using the prepared activated carbon was investigated by varying the flow rate (20 and 30 ml/min) with a constant adsorbent bed height of 6 cm and the initial dye concentration of 100 mg/l, as shown by the breakthrough curve in Fig. 8. The column was found to perform better at a lower flow rate which resulted in a longer breakthrough and ex-

26
1.0 0.8 0.6 0.4 0.2 0.0

I.A.W. Tan et al. / Desalination 225 (2008) 1328

C/Co

20 ml/min 30 ml/min

50

100 t (min)

150

200

Fig. 8. Breakthrough curves for MB adsorption on prepared activated carbon at different flow rates (initial dye concentration = 100 mg/l, bed height = 6 cm, temperature = 28 1C).

haustion time. At a higher flow rate, the adsorption capacity was lower due to insufficient residence time of the solute in the column and diffusion of the solute into the pores of the adsorbent, and therefore, the solute left the column before equilibrium occurred. These results were in agreement with those reported in previous studies [1, 3436]. 3.8. Adsorption performance of prepared activated carbon Table 5 lists the comparison of the maximum monolayer adsorption capacity of MB onto various adsorbents obtained from batch studies. The activated carbon prepared in this work had a relatively large adsorption capacity of 243.90 mg/g compared to some other adsorbents reported in the literature, including commercial activated carbon, F300 which was reported to have the adsorption capacity of 240.00 mg/g on MB [15]. Therefore, the oil palm shell-based activated carbon prepared in this work could be used as an effective adsorbent for removing MB from aqueous solutions. However, the MB adsorption capacity obtained from the column experiments was lower than the values obtained from the batch experi-

ments for the same initial dye concentrations used. This might be due to the insufficient contact time between the solute and the activated carbon in the column. The difference between the batch and continuous capacity could also be attributed to the channelling of the flowing stream. Besides, the effective surface area of the activated carbons packed in the column was lower than that in the stirred batch vessels [36]. Therefore, the performance of the activated carbon bed could be enhanced by applying a higher bed height or a lower solution flow rate.

4. Conclusions The present investigation showed that activated carbon prepared from oil palm shell is a promising adsorbent for the removal of methylene blue dye from aqueous solutions over a wide range of concentrations. Methylene blue was found to adsorb strongly onto the surface of the activated carbon. Equilibrium data were fitted to the Langmuir, Freundlich, Temkin and DubininRadushkevich isotherms and the equilibrium data were best described by the Langmuir isotherm model, with the maximum monolayer adsorption

I.A.W. Tan et al. / Desalination 225 (2008) 1328 Table 5 Comparison of the maximum monolayer adsorption of MB onto various adsorbents
Adsorbents Oil palm shell-based activated carbon Bamboo dust-based activated carbon Groundnut shell-based activated carbon Jute fiber-based activated carbon Jute processing wastes Olive-seed waste residue-based activated carbon Rattan sawdust-based activated carbon Oil palm fibre *Filtrasorb F300 Maximum monolayer adsorption capacity (mg/g) 243.90 143.20 164.90 225.64 22.47 190.00263.00 294.14 277.78 240.00 References This work [8] [8] [3] [37] [15] [11] [12] [15]

27

* Commercial activated carbon

capacity of 243.90 mg/g at 30C. The adsorption process was found to be exothermic in nature. The adsorption performance of the oil palm shell-based activated carbon was comparable to the commercial activated carbon and some other adsorbents reported in earlier studies. A series of column experiments revealed that the continuous adsorption system represented by the breakthrough curves was dependent on the initial dye concentration, activated carbon bed height and the solution flow rate used. Acknowledgments The authors acknowledge the research grant provided by the Ministry of Science, Technology and Innovation (MOSTI), Malaysia under a longterm IRPA grant (Project No: 08-02-05-1021 EA001) that resulted in this article. References
[1] T.V.N. Padmesh, K. Vijayaraghavan, G. Sekaran and M. Velan, Dyes Pigments, 71 (2006) 7782. [2] K. Ravikumar, B. Deebika and K. Balu, J. Hazard. Mater., B122 (2005) 7583. [3] S. Senthilkumaar, P.R. Varadarajan, K. Porkodi and C.V. Subbhuraam, J. Colloid Interface Sci., 284 (2005) 7882.

[4] C. Sourja, D. Sirshendu, D. Sunando and K.B. Jayanta, Chemosphere, 58 (2005) 10791086. [5] M.J. Martin, A. Artola, M.D. Balaguer and M. Rigola, Chem. Eng. J., 94 (2003) 231239. [6] B.S. Girgis and A.A. El-Hendawy, Micropor. Mesopor. Mater., 52 (2002) 105117. [7] K. Kadirvelu, M. Kavipriya, C. Karthika, M. Radhika, N. Vennilamani and S. Pattabhi, Bioresour. Technol., 87 (2003) 129132. [8] N. Kannan and M.M. Sundaram, Dyes Pigments, 51 (2001) 2540. [9] R.L. Tseng, S.K. Tseng and F.C. Wu, Colloids Surf. A: Physicochem. Eng. Aspects, 279 (2006) 6978. [10] B.H. Hameed, A.T.M. Din and A.L. Ahmad, J. Hazard. Mater., 141 (2007) 819825. [11] B.H. Hameed, A.L. Ahmad and K.N.A. Latiff, Dyes Pigments, 75 (2007) 143149. [12] I.A.W. Tan, B.H. Hameed and A.L. Ahmad, Chem. Eng. J., 127 (2007) 111119. [13] Z. Husain, Z. A. Zainal and M. Z. Abdullah, Biomass Bioenergy, 24 (2003) 117124. [14] S. Wang, Z.H. Zhu, A. Coomes, F. Haghseresht and G.Q. Lu, J. Colloid Interface Sci., 284 (2005) 440 446. [15] G.G. Stavropoulos and A.A. Zabaniotou, Micropor. Mesopor. Mater., 82 (2005) 7985. [16] G.H. Oh, C.H. Yun and C.R. Park, Carbon Sci., 4(4) (2003) 180184. [17] M. Olivares-Marn, C. Fernndez-Gonzlez, A. Macas-Garca and V. Gmez-Serrano, Appl. Surf. Sci., 252 (2006) 59805983.

28

I.A.W. Tan et al. / Desalination 225 (2008) 1328 Handlingar, 24(4) (1898) 139. [29] Y.S. Ho and G. McKay, Chem. Eng. J., 70 (1998) 115124. [30] W.J. Weber and J.C. Morris, Proc. Int. Water Pollution Symposium, Pergamon, Oxford, 2 (1962) 231 266. [31] K. Mohanty, D. Das and M.N. Biswas, Chem. Eng. J., 115 (2005) 121131. [32] R.L. Tseng and S.K. Tseng, J. Colloid Interface Sci., 287 (2005) 428437. [33] Z. Aksu and F. Gnen, Process Biochem., 39 (2004) 599613. [34] V.C. Taty-Costodes, H. Fauduet, C. Porte and Y.S. Ho, J. Hazard. Mater., B123 (2005) 135144. [35] E. Malkoc, Y. Nuhoglu and Y. Abali, Chem. Eng. J., 119 (2006) 6168. [36] Z. Al-Qodah and W. Lafi, J. Water Supply: Res. Technol., 52(3) (2003) 189198. [37] S. Banerjee and M.G. Dastidar, Bioresour. Technol., 96 (2005) 19191928.

[18] Z. Al-Qodah, Water Res., 34(17) (2000) 42954303. [19] Z. Al-Qodah, A.T. Shawaqfeh and W.K. Lafi, Desalination, 208 (2007) 294305. [20] D.S. Faust and M.O. Aly, Chemistry of Wastewater Treatment. Butterworths, Boston, 1983. [21] V.V. Basava Rao and S. Ram Mohan Rao, Chem. Eng. J., 116 (2006) 7784. [22] M. El-Guendi, Adsorpt. Sci. Technol., 8(2) (1991) 217225. [23] T.W. Weber and R.K. Chakkravorti, AIChE J., 20 (1974) 228. [24] F. Haghseresht and G. Lu, Energy Fuels, 12 (1998) 11001107. [25] K. Fytianos, E. Voudrias and E. Kokkalis, Chemosphere, 40 (2000) 36. [26] M. Hosseini, S.F.L. Mertens, M. Ghorbani and M.R. Arshadi, Mater. Chem. Phys., 78 (2003) 800. [27] S. Rengaraj, Y. Kim, C.K. Joo, K. Choi and J. Yi, Korean J. Chem. Eng., 21(1) (2004) 187194. [28] S. Langergren and B.K. Svenska, Veternskapsakad