ɛ-caprolactam mixture adsorption from aqueous solution onto granular activated carbon: Kinetics and equilibrium

Chemical Engineering Journal 187 (2012) 69–78

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Study of aniline/␧-caprolactam mixture adsorption from aqueous solution onto granular activated carbon: Kinetics and equilibrium Bin Tang, Yanwen Lin, Ping Yu, Yunbai Luo ∗ College of Chemistry and Molecular Sciences, Wuhan University, Wuchang District, Bayi Road, Wuhan 430072, PR China

a r t i c l e

i n f o

Article history: Received 6 December 2011 Received in revised form 19 January 2012 Accepted 19 January 2012 Keywords: Aniline Caprolactam Granular activated carbon Adsorption

a b s t r a c t Adsorptive properties of aniline (AN) and ␧-caprolactam (CPL) onto a granular activated carbon (GAC) were investigated in single and binary systems via batch adsorption experiments. As AN and CPL are toxic to environment, and they could be present in effluents from CPL manufacturing industries. The effects of pH, GAC dosage, initial adsorbate concentration and temperature on their removal were investigated. The adsorption kinetics of AN and CPL in single and binary systems were studied and was found to conform to pseudo-second order kinetic model. A competitive effect between solutes was observed since lower uptakes as well as slower adsorption kinetics of each solute were obtained in binary adsorption system. The boundary layer diffusion rate was defined as the rate limiting mechanism for AN adsorption on GAC in single and binary system, while the adsorption rate of CPL was found to be governed by particle diffusion. Adsorption isotherm data was fitted to Langmuir and Sips isotherm models for single adsorption system. For the binary solute system, the modified extended Langmuir model was developed to predict the experimental data. The reusability properties of GAC were demonstrated by adsorption–desorption cycle. © 2012 Elsevier B.V. All rights reserved.

1. Introduction ␧-Caprolactam (CPL) is an important chemical raw material which is exclusively used to manufacture Nylon-6 fiber and resins [1]. The global production of CPL amounted to 4.75 million tones in 2011 and there has been an ever increasing demand for it in the recent years. Owing to its high dissolubility in water, CPL could be present in effluents from relevant industries with large quantity. CPL has high COD and toxicity, which may have harmful effect on both public health and environmental quality [2]. Moreover, because of the complexity of chemical reactions required to transform the raw material into CPL, various impurities can be formed during the process of production. Aniline (AN) is one of the main byproducts in the process of CPL production [3]. For this reason effluents from CPL manufacturing industries may contain both CPL and AN. The latter is well known for its wide application in pharmaceutical, pesticide, dyestuff, petrochemicals and agrochemical industries as well as its high toxicity and environmental accumulation. Due to the serious environmental problems created by the pollutants containing wastewater, strict legislation for the release of these hazardous chemicals has been established.

∗ Corresponding author. Tel.: +86 27 68752511; fax: +86 27 68752511. E-mail address: [email protected] (Y. Luo). 1385-8947/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.cej.2012.01.088

Therefore, treatment of pollutants-containing wastewater is required prior to disposal. Various processes have been employed for the elimination of AN from wastewater, including photodecomposition [4–6], electrolysis [7], extraction [8], adsorption on activated carbon, resin and other adsorbent [9–11], oxidation [12,13], biodegradation [14] and other processes. While techniques available for the treatment of CPL-containing wastewater are limit to extraction [15] and biodegradation [16]. Generally, the technology of adsorption on activated carbon has been recognized as one of the most efficient, promising and widely used techniques in the separation and removal of a wide variety of organic pollutants from wastewater for its relative simplicity of design, operation and scale up, high capacity and low cost. Study of adsorptive removal of AN from aqueous solution has been reported by many researchers [17–20]. However, rather scarce works have been performed to remove CPL via activated carbon adsorption process. Usually, the industrial effluents present a mixture of pollutants, and the interactions of the solutes may be ether cooperative or competitive in the adsorption process. In those multi-component adsorption systems that have been investigated, AN and phenol for instance which have similar adsorption property were often chosen as the target adsorbates, and a synergetic effect between them was reported [19,20]. Competitive adsorption between aniline and other pollutants has been scarce in literature. However, from a practical point of view, the competition between solutes for the active adsorption sites is very

70

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

Table 1 Physico-chemical characteristics of the raw AC.

solute systems, all solutions were prepared with solution of equal concentrations of AN and CPL.

Parameters

Value

Apparent density (g/mL) Ash content (%) Moisture content (%) pHPZC Particle size (mesh) BET surface area (m2 /g) Micropore volume (cm3 /g) Total pore volume (cm3 /g) Average pore width (Å)

0.65 1.22 4.65 6.77 80–150 894.73 0.349 0.436 10.5

common in multi-component adsorption system [21–23], and the selectivity of the sorbent material for the pollutants in the solution is important in the design of adsorption systems. The aim of this work is to investigate the ability of GAC to adsorb AN and CPL from aqueous solution and the selectivity for the pollutants in the solution by the sorbent. In this work, the adsorption of AN and CPL from single and binary solute system onto a commercial available GAC was investigated. The effects of various operating parameters, such as pH, GAC dosage, initial adsorbate concentration and temperature were studied. The kinetics and isotherms for AN and CPL adsorption onto GAC were studied in single and binary systems. The Webber’s intraparticle diffusion and Boyd’s film diffusion models were attempted to study the mechanism of adsorption. The kinetic and equilibrium results were obtained from the batch adsorption experiments. 2. Materials and methods 2.1. Materials 2.1.1. Adsorbent The GAC used in this study was a coal based carbon purchased from Kermel Reagent Co., Ltd. (Tianjin, China). Prior to use, the GAC was washed thoroughly with hot deionized water to remove fines, then dried at 105 ◦ C for overnight and finally stored in a plastic container for further use. The surface properties of the GAC were characterized using a Micromeritics ASAP (Accelerated Surface Area and Porosity) 2020 adsorption apparatus with N2 as adsorbate at 77 K. The point of zero charge (pHPZC ) of GAC was determined by using a mass titration method proposed by Noh and Schwarz [24]. The main characteristics of GAC were given in Table 1. 2.1.2. Adsorbate AN (AR) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). CPL (purity > 99.9%) was kindly provided by Shijiazhuang Chemical Fiber Plant (SINOPEC, China). They were used as received without further purification and their chemical structures were shown in Fig. 1. The aqueous solutions of the adsorbates for the adsorption test were prepared by dissolving AN and CPL into deionized water without further pH adjustment. For binary

2.2. Batch adsorption experiments Adsorption of AN and CPL onto GAC from single and binary solute systems was performed in batch experiments. A given mass of GAC was mixed with 100 mL of solution in glass stoppered conical flasks (250 mL). The mixtures were magnetically stirred at 150 rpm and the temperature was kept by thermostated bath. After the specified time, samples were taken from the conical flasks and the adsorbent was separated from the samples by filtering through 0.45 ␮m microporous filters. The effects of adsorption parameters on adsorption of AN and CPL onto GAC such as pH (3–11), adsorbent dosage (0.5–3.0 g/L), initial adsorbate concentration (0.5–5.0 mmol/L) and temperature (25–55 ◦ C) were investigated. The percentage removal of AN or CPL (%) was calculated according to the following equation: R=

(C0 − C) × 100% C0

(1)

The amount of AN or CPL adsorbed onto the GAC, q (mmol/g) was calculated by the following equation: q=

(C0 − C)V W

(2)

where C0 (mmol/L) is the initial concentration of AN or CPL, V (L) is the volume of the solution and W (g) is the weight of GAC and C is the residual concentration of AN or CPL at equilibrium or any time t (min), which defines qe or qt , respectively. In the effect of initial pH studies, the solution pH was adjusted by using 0.1 mol/L NaOH and 0.1 mol/L HCl solutions. The GAC dosage was fixed at 2.0 g/L and the initial concentration of AN or CPL in the solution was 1 mmol/L. In adsorption isotherm studies, the equilibrium time was set as 24 h according to the result of preliminary experiments. Concentrations of the solutes in the filtrates were determined using HPLC assembled by CBM-10A controller, LC-10AT pumps and SPD-10A dual adsorbance UV detector (Shimadzu, Japan). HPLC was performed using a Shimadzu Shim-pack VP-ODS column (4.6 mm × 150 mm, 4.5 ␮m) at a flow rate of 0.7 mL/min. The mobile phase used was acetonitrile:water = 30:70 (V/V). The adsorbency wavelength of the UV detector was fixed at 284 nm and 210 nm for the measurement of the concentration of AN and CPL, respectively. 2.3. Adsorbent regeneration experiment To investigate the reusability of GAC for AN and CPL adsorption, an adsorption–desorption procedure was performed. Firstly, a mass of 0.2 g of GAC was mixed with 100 mL solution and stirred at 25 ◦ C for 2 h. The initial concentration of AN or CPL in the solution was 1 mmol/L. The adsorbed GAC was then immersed in a 20 mL of 0.01 mol/L hydrochloric acid solution and treated with ultrasound for 30 min. The ultrasonic reactor (KSD-3000, Kexingda Co., Ltd., China.) was performed at a frequency of 80 kHz. The regenerated GAC was washed with deionized water to neutral pH and dried at 105 ◦ C for 2 h. The adsorption–desorption procedure was repeated five times. 3. Results and discussion 3.1. Adsorption study of AN and CPL

Fig. 1. Chemical structures of aniline (AN) and caprolactam (CPL).

3.1.1. Effect of pH The pH of the system plays an important role in the adsorption process for its influence on the surface charge of the adsorbent

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

Fig. 2. Effect of initial pH of the solution on AN and CPL adsorption onto GAC.

and the ionization/dissociation of the adsorbate molecule [25]. The removal of AN and CPL from single solute system as a function of pH is shown in Fig. 2. It is seen that the adsorption of both species was highest at neutral pH conditions. At an acidic pH of 3.0, which is lower than the pKa value of AN (4.6), the adsorption of AN decreased dramatically. The low adsorption can be primarily due to the electrostatic repulsion between the positively charged GAC surface and the AN cations. However, the percentage adsorption of CPL decreased slightly at lower pH of the solution, which may due to the low pKa value of CPL (−0.14). The percentage removal of both species decreased under basic conditions, which might be due to the competition between the excess OH− ions and the adsorbates molecules for the adsorption sites. Since the AN and CPL solution has a pH which is very close to that of deionized water (pH 6.5), the adsorption experiment were conducted without pH adjustment. 3.1.2. Effect of adsorbent dosage The adsorbent dosage is an important parameter due to the fact that a given mass of adsorbent can only adsorb a fixed amount of adsorbate from the solutions. The effect of GAC dosage on AN and CPL removal was determined within the GAC dosage range of 0.5–3.0 g/L for single solute system at 25 ◦ C. The results are given in Fig. 3. It was observed that the removal of AN and CPL from single solute system increased from 44.0% to 97.4% and from 23.9% to 78.7% when the dosage of GAC increased from 0.5 to 3.0 g/L. This may be explained by the fact that the increase in adsorbent

Fig. 3. Effect of GAC dosage on AN and CPL adsorption onto GAC.

71

Fig. 4. Effect of initial concentration of AN and CPL on their adsorption onto GAC.

concentration resulted in more available adsorption sites for the adsorbate. A GAC dosage of 2.0 g/L was used for the further studies. 3.1.3. Effect of initial adsorbate concentration The effect of initial concentration of AN and CPL on their removal from single solute systems is shown in Fig. 4. From the figure, it was evident that their removal decreased with the increasing initial concentration. Increasing the initial concentration from 0.5 mmol/L to 5.0 mmol/L, the AN removal decreased from 98.1% to 69.4% and the CPL removal decreased from 85.6% to 34.2%. At lower adsorbates concentrations, the ratio of number of available adsorption sites to the initial adsorbates molecule is higher than that of higher concentrations, which resulted in the decrease in their removal. 3.1.4. Effect of temperature The temperature of the solution was considered to be a critical factor affecting the adsorption process. Fig. 5 shows the effect of temperature of the solution on AN and CPL adsorption from single solute system. The adsorption efficiency of both species decreased with an increase in temperature, suggesting the exothermic nature of the adsorption process. The optimum solution temperature was selected as 25 ◦ C for the further studies. 3.2. Adsorption kinetics in single and binary systems Adsorption kinetics is an important parameter to evaluate the efficiency of adsorption. The adsorption kinetics curves in single

Fig. 5. Effect of the solution temperature on AN and CPL adsorption onto GAC.

72

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

Integration of Eq. (5) for the initial conditions t = 0 and qt = 0 gives the linear form of the pseudo-second order rate equation as: t 1 1 = + t qt qe k2 q2e

(6)

where qe , qt and k2 are the amount of AN or CPL adsorbed at equilibrium (mmol/g), the amount of AN or CPL adsorbed at time t (mmol/g) and the pseudo-second order rate constant (g/(mmol min)), respectively. The calculated kinetic models parameters for AN and CPL adsorption onto GAC in single and binary systems are given in Table 2. As shown in Fig. 7a and b, the experimental data was well fitted by the pseudo-second order model with high correlation coefficient values (R2 > 0.99), and the theoretical uptakes showed good agreement with experimental values. It is also seen that the pseudo-first order model cannot fit the experimental data accurately. Thus it can be concluded that the adsorption of AN and CPL onto GAC in both single and binary systems obeys the pseudosecond order kinetic model. It is noted that the equilibrium uptakes of CPL in binary systems were much lower than that of single system, obviously in higher initial concentrations. For instance, the removal efficiency of CPL reduced from 77.6% (in single solution of CPL) to 68% (in binary solution of AN and CPL) under the initial concentration of CPL of 0.5 mmol/L. When the initial concentration of CPL was 2 mmol/L, the CPL removal decreased from 54.7% in single solution to 37% with the coexistence of 2 mmol/L AN. The lower adsorption capacity of CPL in binary systems revealed the antagonism between the two species. However, slight decrease in AN adsorption yield was observed in binary systems in the presence of CPL, indicating that

Fig. 6. Adsorption kinetics of AN and CPL onto GAC from single and binary system.

and binary solute systems for AN and CPL are shown in Fig. 6a and b, respectively. It is seen that the adsorption rate was rapid and most of the solutes were adsorbed within the first 90 min, while the sorption kinetic of AN and CPL in binary system was slower than the individual sorption of each solute in single system. The same observation was obtained for both species at three different initial concentrations. In this work, two kinetic models namely the pseudo-first order and pseudo-second order, which are extensively used in kinetic studies, were employed to characterize adsorption kinetics. The pseudo-first order kinetics model proposed by Lagergren [26] for the solid–liquid sorption system is generally expressed as the following equation: dqt = k1 (qe − qt ) dt

(3)

where qe , is the amount of AN or CPL adsorbed at equilibrium (mmol/g), qt is the amount of AN or CPL adsorbed at time t (mmol/g), and k1 is the pseudo-first order rate constant (1/min), respectively. Integration of Eq. (3) for the initial conditions t = 0 and qt = 0 gives the linear form of the pseudo-first order rate equation as: log(qe − qt ) = log qe −

k1 t 2.303

(4)

The pseudo-second order kinetic model equation is expressed as [27,28]: dqt = k2 (qe − qt )2 dt

(5)

Fig. 7. Pseudo-second order kinetic plots for AN and CPL adsorption onto GAC.

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

73

Table 2 Parameters of the pseudo-first order and pseudo-second order kinetic models for the adsorption of AN and CPL onto GAC from single and binary system. System

Single

Sorbate

AN

CPL

Binary

AN

CPL

Co (mmol/L)

qe,exp (mmol/g)

Pseudo-first order

Pseudo-second order

qe,cal (mmol/g)

k1 (1/min)

R2

qe,cal (mmol/g)

k2 (g/(mmol min))

R2

0.5 1 2 0.5 1 2

0.245 0.483 0.911 0.194 0.335 0.547

0.191 0.372 0.711 0.147 0.241 0.388

0.018 0.017 0.013 0.004 0.003 0.003

0.976 0.972 0.965 0.964 0.945 0.954

0.255 0.500 0.922 0.192 0.317 0.525

0.168 0.083 0.036 0.118 0.085 0.054

1.000 1.000 1.000 0.996 0.997 0.996

0.5 1 2 0.5 1 2

0.242 0.472 0.877 0.170 0.275 0.370

0.193 0.366 0.679 0.121 0.197 0.257

0.018 0.015 0.011 0.004 0.004 0.004

0.986 0.971 0.953 0.923 0.936 0.939

0.255 0.478 0.881 0.158 0.253 0.340

0.157 0.083 0.035 0.133 0.074 0.057

1.000 1.000 1.000 0.999 0.997 0.996

AN was preferentially adsorbed onto the GAC in the binary solute systems and this will be further discussed in the next section. 3.3. Adsorption mechanism 3.3.1. The intra-particle diffusion model To better know the adsorption process and for design purpose, it is required to identify the mechanism that involved in the adsorption process. Although the pseudo-second order kinetic model could well fit the experimental data, it was not able to provide useful information of the mechanism of adsorption process. For a solid–liquid adsorption process, the mass transfer of the adsorbate is usually characterized by either external mass transfer (boundary layer diffusion) or intra-particle diffusion or both. It is well accepted that the adsorption process consists of three consecutive steps:

linear regions. The first region lies in the range of 0–60 min with faster adsorption rate, and it is 60–180 min and 180–360 min for the other two regions. Linear fits were applied to each part of the curves and the results were given in Table 3. It is seen that the intra-particle diffusion rate constant decreased as the contact time increased indicating the intra-particle diffusion slowed down. As mentioned above, if intra-particle diffusion controlled the overall sorption rate, the lines should pass through the origin. For AN adsorption onto GAC from single and binary solute systems, the values of I obtained from the intercept of the first linear segment are −0.019 and −0.016, which are significantly different from zero, it is concluded that the adsorption rate in this period is

(a) Transport of adsorbate molecules from bulk solution to the adsorbent exterior surface through boundary layer diffusion. (b) Intra-particle diffusion of the adsorbate molecules into the pores of the adsorbent. (c) Sorption of the adsorbate molecules on the interior surfaces of the pores and capillary spaces of the adsorbent. Of the three steps, the last step is very rapid and considered to be negligible. The overall rate of adsorption is governed by the slowest step, which would be either film diffusion or intra-particle diffusion or both. One of the most commonly used techniques for identifying the mechanism involved in the sorption process is fitting the experimental data in an intra-particle diffusion model introduced by Weber and Morris [29]. It can be expressed by the following equation: qt = kID t 1/2 + I

(7)

where qt is uptake of AN or CPL (mmol/g) at any time (t), kID is the intra-particle diffusion rate constant (mmol/(g min0.5 )) and I (mmol/g) is a constant that indicates the thickness of the boundary layer, i.e. the higher value of I, the greater the boundary layer effect. According to this model, the plot of qt versus t1/2 should produce a straight line if intra-particle diffusion is involved in the adsorption process and should pass through the origin if intra-particle diffusion is the rate controlling step. The plot of this model often has a multi-linear nature which indicates that there are two or more steps with different rate constants affecting the adsorption process, and in that case graphical analysis should be applied for each linear region [30]. In this work, the plots obtained (Fig. 8a and b) in single and binary solute systems for both species can be divided into three

Fig. 8. Intra-particle diffusion plots for AN and CPL adsorption onto GAC.

0.999 0.988 −0.060 0.001 0.983 0.979

3.3.2. The film-diffusion model In order to investigate the contribution of film resistance to the kinetics of AN and CPL adsorption, the Boyd’s film-diffusion model was employed [31]. This model has the assumption that the boundary layer surrounding the adsorbent particle is the main resistance to diffusion. The film-diffusion model is expressed by the equation as follows: F(t) = 1 −

0.402 0.103

0.999 0.995 −0.075 −0.001 0.968 0.976 0.414 0.156

R2

not controlled by intra-particle diffusion. For CPL adsorption, the intercept of the first linear region are 0.002 and −0.003, which are very close to zero, indicating that the adsorption rate in this period is likely to be governed by intra-particle diffusion.

0.003 0.006

0.002 0.006

0.957 0.995

0.965 0.967 0.222 0.086

R I2 (mmol/g) ))

−0.016 −0.003 0.048 0.020

0.016 0.008 0.987 0.983

−0.019 0.002 0.051 0.022

0.015 0.013 0.986 0.974

I1 (mmol/g)

kID2 (mmol/(g min

exp(−n2 Bt)

qt qe



0.5 2

))

R

n2

(8)

(9)

Bt = 0.4977 − ln(1 − F(t))

AN CPL Binary

a

AN CPL Single

Initial sorbate concentration: 0.1 mmol/L.

0.5

kID1 (mmol/(g min

n=1

where qt and qe are the AN or CPL uptake (mmol/g) at time t and at equilibrium, respectively. By applying the Fourier transform and then integration to Eq. (8), the following approximations were obtained by Reichenberg [32]: if the value of F(t) is higher than 0.85:

Bt =

Sorbatea

Intra-particle diffusion model

2

(10)

if the value of F(t) is lower than 0.85: 0.253 0068

kID3 (mmol/(g min

0.5 2

F(t) =

System

Table 3 Parameters of the intra-particle diffusion model and Boyd’s plot for the adsorption of AN and CPL onto GAC from single and binary system.

∞   6   1

where F(t) is the fractional attainment of equilibrium, at different times, t (min) and Bt is a function of F(t):

))

I3 (mmol/g)

R

2

Intercept

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

Boyd Plot

74

√ −



 −

2 F(t) 3

 2 (11)

According to this model, the plot of Bt against time should be linear and pass through the origin if intra-particle diffusion controls the rate of mass transfer. If the plot is nonlinear or linear but does not pass through the origin, then it is concluded that the adsorption rate is controlled by film-diffusion or chemical reaction. Fig. 9a and b shows the Boyd plots for the adsorption of AN and CPL onto AC from single and binary solute systems in the first 60 min. It is seen that the experiment data of both species are well fitted by this model with high correlation coefficients (R2 > 0.97), and the fitting results are presented in Table 3. It is clear that the intercept of the plots of AN adsorption from single and binary solute systems are −0.075 and −0.060, which are significantly different from zero, and since the same linear segment have intercept of −0.019 and −0.016 in the intra-particle diffusion plot, it is concluded that the rate in this period of adsorption of AN from single and binary solute systems is controlled by film diffusion. The intercept of the plots of CPL adsorption from single and binary solute systems are −0.001 and 0.001, which are very close to zero, and according to the result of intro-particle diffusion plot, it is concluded that the intro-particle diffusion is the rate controlling step in CPL adsorption of this period. 3.3.3. Competitive adsorption From literature it is known that external mass transport of sorbate is usually the rate-limiting step in systems, which have poor mixing, low concentration of adsorbate, small particle size and high affinity of adsorbate for adsorbent. On the contrary, the intraparticle step controls the overall transfer for those systems that have high concentration of adsorbate, good mixing, large particle size of adsorbent, and low affinity of adsorbate for adsorbent [33]. According to the kinetics analysis above, boundary layer diffusion

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

75

3.4. Adsorption isotherms 3.4.1. Adsorption isotherms in single solute systems The equilibrium adsorption isotherm is important in the design of adsorption systems. In this study, four adsorption isotherm models namely the Langmuir isotherm, Freundlich isotherm, Redlich–Peterson isotherm and Sips isotherm models were employed to characterize the adsorption isotherms of AN and CPL in single solute systems. The Langmuir isotherm model, which is a non-linear model that based on the assumption of a monolayer uptake of the adsorbate on a homogenous surface with identical and equivalent adsorption sites [36], has been the most frequently employed model in fitting experiment data. It can be expressed as the following equation: qe =

qm KL Ce 1 + KL Ce

(12)

where qe , Ce , qm and KL are adsorbate uptake at equilibrium (mmol/g), liquid phase concentration of sorbate at equilibrium (mmol/L), maximum adsorption capacity (mmol/g) and Langmuir constant (L/mmol), respectively. A dimensionless separation factor (RL ), defined by Weber and Chakravorti [37], can be used to predict the affinity between the adsorbate and adsorbent. RL is calculated by the following equation: RL =

Fig. 9. Boyd plots for the initial period of AN and CPL adsorption onto GAC.

controlled the overall sorption rate of AN and intra-particle diffusion is the rate controlling step in CPL adsorption process. However, the adsorption kinetic studies were performed at same mixing conditions, equal initial solute concentration and fixed amount of adsorbent, hence it can be due to the affinity of AN and CPL for adsorbent that determined the adsorption rate controlling step. In the case of adsorption onto GAC with a neutral surface (pHPZC = 6.77), the main driving force can be in principle a combination of two parts: Van der Waals forces and the hydrophobic interaction between the sorbate and carbon surface [34]. AN is an aromatic compound with lower water solubility, and the higher affinity towards GAC can be attribute to the ␲–␲ dispersion interactions between the aromatic ring of AN and the graphene layers of GAC as well as the strong hydrophobic interactions between the GAC surface and the AN molecules [35]. While CPL is a typical hydrophilic compound without aromatic structure, and because of the strong amide–water interaction it is miscible in water, which determined its lower affinity towards GAC. In binary solutes adsorption systems, there was a competition for specific adsorption sites between AN and CPL molecules, in which the available adsorption sites were predominated occupied by the former owing to its higher affinity towards GAC. At lower initial concentrations of the adsorbates, there were enough available adsorption sites for both species and the competitive adsorption was weakened. While at higher initial concentrations, the competition for available adsorption sites between the two species was intensified, which resulted in a significant decrease in CPL removal from binary solute systems.

1 1 + KL C0

(13)

The value of RL indicates the adsorption nature to be either unfavorable (RL > 1), linear (RL = 1), favorable (0 < RL < 1) or irreversible (RL = 0). The Freundlich isotherm, which is the earliest relationship describing the non-idea and reversible sorption, is an empirical model assuming adsorption on heterogeneous surface and active sites with different energy [38]. It can be applied to multilayer adsorption. The Freundlich isotherm model can be represented by the following equation: 1/n

qe = KF Ce

(14)

where KF is Freundlich constant that can indicate adsorption capacity (mmol/(L1−1/n g)), and 1/n is the heterogeneity factor. The Redlich–Peterson isotherm [39] is a three parameters isotherm, which features both Langmuir and Freundlich isotherms. It can be expressed by the following equation: qe =

KR Ce ˇ

1 + aR Ce

(15)

where KR is Redlich–Peterson isotherm constant (L/g), aR is Redlich–Peterson isotherm constant (L/mmol)ˇ and ˇ is the exponent which lies between 0 and 1. Where ˇ = 1 qe =

KR Ce 1 + aR Ce

(16)

It is a Langmuir type equation. Where ˇ = 0 qe =

KR Ce 1 + aR

(17)

It becomes the Henry’s law equation. Sips isotherm [40] is another three parameters isotherm that combines the Langmuir and Freundlich isotherm. At low adsorbate concentrations, it reduces to Freundlich isotherm; while at high concentrations, it predicts a monolayer adsorption capacity

76

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78 Table 4 Values of parameters for various isotherm models for the adsorption of AN and CPL in single adsorption system. Langmuir

Isotherm models’ parameter

AN CPL

qm (mmol/g)

KL (L/mmol)

RL

R2

1.881 0.967

5.230 1.694

0.161 0.371

0.991 0.985

Freundlich

Isotherm models’ parameter

AN CPL Redlich–Peterson

KF (mmol/(L1−1/n g))

n

R2

1.565 0.562

2.869 2.673

0.974 0.966

Isotherm models’ parameter KR (L/g)

aR (L/mmol)ˇ

ˇ

R2

AN CPL

17.650 2.298

10.239 2.870

0.818 0.861

0.988 0.989

Sips

Isotherm models’ parameter

AN CPL

KS (mmol/g)

aS (L/mmol)ˇS

ˇS

R2

2.478 1.130

1.758 1.109

0.664 0.791

0.992 0.990

adsorption capacity of CPL was found in the presence of AN when compared to adsorption from single solute solution. However, a 11.9% reduction in adsorption capacity of AN in the presence of CPL was found, which indicated a higher affinity of AN for GAC than CPL.

Fig. 10. Adsorption isotherms of AN and CPL onto GAC in single system.

characteristic of the Langmuir isotherm. It can be expressed by the following equation: ˇ

qe =

KS aS Ce S ˇ

1 + aS Ce S

(18)

where KS is Sips isotherm constant (mmol/g), aS is Sips isotherm constant (L/mmol)ˇs and ˇS is the Sips isotherm constant dimensionless. As shown in Fig. 10a and b, non-linear fittings were applied to the isotherm data of AN and CPL adsorption onto AC from single solute system using origin 7.5 by the four adsorption isotherm models. The calculated parameters for each adsorption isotherm model were given in Table 4. It is seen that the Sips isotherm provides the best correlation for both species with highest correlation coefficient (R2 > 0.99). It is also seen that the Langmuir isotherm gives better fitting than the Freundlich isotherm for AN and CPL with higher correlation coefficient (R2 > 0.98), indicating monolayer adsorption of the adsorbate. The calculated maximum adsorption capacities of AN and CPL in single system were 1.881 and 0.967 mmol/g, respectively. The RL values in single system were 0.161 and 0.371 for AN and CPL, indicating favorable adsorption of AN and CPL onto GAC. 3.4.2. Adsorption isotherms in binary solute system Fig. 11a and b illustrates adsorption isotherms for AN and CPL in binary adsorption systems. Compared with the adsorption isotherms of AN in binary adsorption system, the adsorption isotherms of CPL has a less regular shape. From the figure, it is clear that the uptakes of CPL decreased significantly with the presence of AN in binary adsorption system. A 52.2% reduction in

Fig. 11. Adsorption isotherms of AN and CPL onto GAC in binary system.

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

77

Table 5 Values of parameters for isotherm models for AN–CPL binary system.

AN CPL

Extended Langmuir

Modified Langmuir

Modified extended Langmuir

SSE

SSE

a

b

SSE

0.720 0.034

0.734 0.034

−0.241 −0.126

−0.332 0.245

0.022 0.014

Additionally, the maximum uptakes of CPL and AN that determined in single adsorption system were 0.855 and 1.735 mmol/g under the initial concentration of 5 mmol/L. While the combined uptakes of the two species at an initial concentration in total of 5 mmol/L was 1.348 mmol/g (0.992 and 0.356 for AN and CPL, respectively), which is lower than the maximum uptakes of AN in single solute system. This can be attributed in the main to the inhibitory effect of the loaded solute molecules on the adsorption of the other. There are several multi-component isotherm models derived from single component adsorption system can be applied to describe the multi-component adsorption system, among which the extended Langmuir model was the most frequently employed one [41]. The extended Langmuir model could be expressed as follows: qe,j =

qm,J KL,j Ce,j 1+



KL,j Ce,j

qe,CPL =

(19)

(qm,AN − qm,CPL )KL,AN Ce,AN qm,AN KL,AN Ce,AN + 1 + KL,AN Ce,AN 1 + KL,AN Ce,AN + KL,CPL Ce,CPL (20)

qm,CPL KL,CPL Ce,CPL 1 + KL,AN Ce,AN + KL,CPL Ce,CPL

qm,CPL KL,CPL Ce,CPL (1 − CPL ) 1 + KL,AN Ce,AN + KL,CPL Ce,CPL

and it is assumed that competitive efficiency () varies linearly with the amount of another sorbate adsorbed at equilibrium (qe ): AN (qe,CPL ) = (aAN qe,CPL + bAN )

(24)

CPL (qe,AN ) = (aCPL qe,AN + bCPL )

(25)

where a and b are the constant parameter of competitive efficiency. The sum of the squares of the error (SSE) was used to evaluate the accuracy of each model. SSE =

n 

(qe,cal − qe,exp )2i

qm,AN KL,AN Ce,AN (1 − AN ) 1 + KL,AN Ce,AN + KL,CPL Ce,CPL

(26)

Evaluation of each isotherm model was done by comparing the SSE value. The isotherm parameters of AN and CPL in binary adsorption system were estimated and listed in Table 5. The modified extended Langmuir isotherm model provided the best prediction with lowest SSE values for AN and CPL adsorption in binary system, indicating the inhibitory effects of the loaded AN and CPL molecules on the adsorption of each other. As can be seen in Fig. 11a and b, the extended Langmuir model under-predicted the binary adsorption data of AN and over-predicted that of CPL, which may due to the great variation in the values of qm for AN and CPL as explained previously. Additionally, the negative values of a and b for AN suggested the slight inhibitory effects caused by the loaded CPL molecules on the carbon surface. While the positive value of b for CPL indicated that the adsorption of CPL was greatly restrained by the loaded AN molecules. 3.5. Reusability Reusability was an important factor for the practical application of GAC in adsorption industry. Therefore, the regeneration properties of GAC were studied. In effect of pH study, it is known that the adsorption of AN onto GAC was greatly inhibited at low pH, which indicated the AN adsorbed GAC can be regenerated at

(21)

The first term of Eq. (20) is the expression of the Langmuir isotherm for the molecule number of AN that adsorbed without a competition. The second term represents the molecule number of AN that adsorbed on the sites with the competition of CPL based on the extend Langmuir adsorption model. The molecule number of CPL that adsorbed on the sites with the competition of AN can be calculated by Eq. (21). For the AN and CPL binary adsorption system onto GAC, the competitive effect should be taken into account due to the inhibitory effect of the loaded solute molecules on the adsorption of the other. In this paper the modified extended Langmuir model was developed to describe the AN and CPL binary adsorption system by introducing competitive efficiencies () of adsorbates into Eq. (19), which can expressed by the following equations: qe,AN =

(23)

i=1

where qm,j and KL,j are the maximum adsorption capacity and Langmuir adsorption constant of solute j estimated from corresponding single adsorption system. It should be noted that the extended Langmuir isotherm provides reasonable estimates for multi-component system as long as the values of qm of the solutes were close to each other. From the single solute adsorption study, it is known that AN shows a higher affinities towards GAC than CPL. Therefore, the extended Langmuir isotherm model cannot give a satisfactory description of the target system. Derived from the original Langmuir isotherm model and extend Langmuir isotherm model, a modified Langmuir model proposed by Jain and Snoeyink [42] based on the assumption that a part of adsorption occurs without competition can be used to describe the binary adsorption system where the values of qm of the solutes were far from each other. For the AN and CPL binary adsorption system, the model representation is given by the following equations: qe,AN =

qe,CPL =

(22) Fig. 12. Adsorption–desorption cycle of GAC.

78

B. Tang et al. / Chemical Engineering Journal 187 (2012) 69–78

an acidic condition [10]. However, the adsorption capacity of CPL varied slightly at different pH of the solution. Since ultrasound has provided an efficient way for the regeneration of adsorbent [43], a combined regeneration process was applied in this study. Adsorption–desorption cycle of GAC is shown in Fig. 12. From the figure, it is clear that the GAC could be used repeatedly without a significant decrease in adsorption amount. 4. Conclusions Adsorption of AN and CPL onto a granular activated carbon from single and binary solute aqueous solutions was evaluated by adsorption kinetics and equilibrium study. The adsorption efficiency of AN and CPL from single solute system was optimum at a pH of 6.5 and at the solution temperature of 25 ◦ C. The sorption kinetics of both species in single and binary systems obeyed the pseudo-second order model. For AN adsorption from aqueous solution, the adsorption rate in the first 60 min was determined to be controlled by boundary layer diffusion, while intra-particle diffusion governed the overall adsorption rate of CPL in this period. The adsorption isotherms of both species in single system were found to confirm to the Langmuir isotherm and Sips isotherm models. The monolayer adsorption capacities of AN and CPL were 1.881 and 0.967 mmol/g, respectively. A great inhibitory effect on CPL adsorption from binary solute system by AN was observed, resulting lower adsorption capacity of CPL. The experiment data indicated the competitive adsorption of AN and CPL for the available sites of GAC. The competitive adsorption isotherms data of both species in binary system was well predicted by the modified extended Langmuir model. The experimental results obtained from adsorption–desorption cycle suggested that GAC could be used repeatedly for AN and CPL adsorption from single and binary systems. Acknowledgements The authors thank Dr Junwei Ge and Dr Jianquan Zhao for their efforts in characterization of granular activated carbon. Additionally, we wish to appreciate Prof Yunbai Luo for his valuable comments during the paper preparation. References [1] F. Ullmann, Ullmann’s Encyclopedia of Industrial Chemistry, Interscience, New York, 2004. [2] R.P. Pohanish, Sittig’s Handbook of Toxic and Hazardous Chemicals and Carcinogens, 2012. [3] Y.Y. Zhao, Z.Z. Jing, H.P. Li, H.S. Zhang, The determination of impurities in caprolactam by capillary gas chromatography–mass spectrometry, Microchem. J. 69 (2001) 213–217. [4] C. Karunakaran, S. Senthilvelan, Solar photocatalysis: oxidation of aniline on CdS, Sol. Energy 79 (2005) 505–512. [5] A. Kumar, N. Mathur, Photocatalytic degradation of aniline at the interface of TiO2 suspensions containing carbonate ions, J. Colloid Interf. Sci. 300 (2006) 244–252. [6] W. Chu, W.K. Choy, T.Y. So, The effect of solution pH and peroxide in the TiO2 induced photocatalysis of chlorinated aniline, J. Hazard. Mater. 141 (2007) 86–91. [7] Y. Han, X. Quan, S. Chen, H. Zhao, C. Cui, Y. Zhao, Electrochemically enhanced adsorption of aniline on activated carbon fibers, Sep. Purif. Technol. 50 (2006) 365–372. [8] C.H. Li, Recovery of aniline from wastewater by nitrobenzene extraction enhanced with salting out effect, Biomed. Environ. Sci. 23 (2010) 208–212. [9] K. Laszlo, Adsorption from aqueous phenol and aniline solutions on activated carbons with different surface chemistry, Colloid Surf. A 265 (2005) 32–39. [10] F.Q. An, X.Q. Feng, B.J. Gao, Adsorption of aniline from aqueous solution using novel adsorbent PAM/SiO2 , Chem. Eng. J. 151 (2009) 183–187.

[11] X.Y. Lin, J.P. Zhang, X.G. Luo, C. Zhang, Y. Zhou, Removal of aniline using lignin grafted acrylic acid from aqueous solution, Chem. Eng. J. 172 (2011) 856–863. [12] N. Jagtap, V. Ramaswamy, Oxidation of aniline over titania pillared montmorillonite clays, Appl. Clay Sci. 33 (2006) 89–98. [13] H.T. Gomes, P. Selvam, S.E. Dapurkar, J.L. Figueiredo, J.L. Faria, Transition metal (Cu, Cr, and V) modified MCM-41 for the catalytic wet air oxidation of aniline, Micropor. Mesopor. Mater. 86 (2005) 287–294. [14] L. Wang, S. Barrington, J. Kim, Biodegradation of pentyl amine and aniline from petrochemical wastewater, J. Environ. Manage. 83 (2007) 191–197. [15] X.C. Gong, Y.C. Lu, G.S. Luo, Caprolactam recovery by a column packed with polysulfone microcapsules containing 1-octanol, Sep. Purif. Technol. 69 (2009) 71–77. [16] C.C. Wang, C.M. Lee, Isolation of the ␧-caprolactam denitrifying bacteria from a wastewater treatment system manufactured with acrylonitrile–butadiene–styrene resin, J. Hazard. Mater. 145 (2007) 136–141. [17] H.A. Johani, M.A. Salam, Kinetics and thermodynamic study of aniline adsorption by multi-walled carbon nanotubes from aqueous solution, J. Colloid Interf. Sci. 360 (2011) 760–767. [18] X.X. Yang, Q.X. Guan, W. Li, Effect of template in MCM-41 on the adsorption of aniline from aqueous solution, J. Environ. Manage. 92 (2011) 2939–2943. [19] C. Valderrama, J.I. Barios, M. Caetano, A. Farran, J.L. Cortina, Kinetic evaluation of phenol/aniline mixtures adsorption from aqueous solutions onto activated carbon and hypercrosslinked polymeric resin (MN200), React. Funct. Polym. 70 (2010) 142–150. [20] W.M. Zhang, J.L. Chen, B.C. Pan, Q. Chen, Synergistic effect on phenol/aniline mixture adsorption on nonpolar resin adsorbents from aqueous solution, React. Funct. Polym. 66 (2006) 395–401. [21] M. Chiban, A. Soudani, F. Sinan, M. Persin, Single, binary and multi-component adsorption of some anions and heavy metals on environmentally friendly Carpobrotus edulis plant, Colloids Surf. B: Biointerf. 82 (2011) 267–276. [22] D. Mohan, K.P. Singh, Single- and multi-component adsorption of cadmium and zinc using activated carbon derived from bagasse—an agricultural waste, Water Res. 36 (2002) 2304–2318. [23] L.J. Li, F.Q. Liu, X.S. Jing, P.P. Ling, A.M. Li, Displacement mechanism of binary competitive adsorption for aqueous divalent metal ions onto a novel IDA-chelating resin: isotherm and kinetic modeling, Water Res. 45 (2011) 1177–1188. [24] J.S. Noh, J.A. Schwarz, Estimation of the point of zero charge of simple oxides by mass titration, J. Colloid Interf. Sci. 130 (1989) 157–164. [25] C. Moreno-Castilla, Adsorption of organic molecules from aqueous solutions on carbon materials, Carbon 42 (2004) 83–94. [26] S. Lagergren, Zur theorie der sogenannten adsorption geloster stoffe, K. Sven. Vetenskapsakad. Handl. 24 (1898) 1–39. [27] Y.S. Ho, G. McKay, Pseudo-second order model for sorption process, Process Biochem. 34 (1999) 451–465. [28] Y.S. Ho, Review of second order model for adsorption systems, J. Hazard. Mater. 136 (2006) 681–689. [29] W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution, J. Sanit. Eng. Div. Proc. Am. Soc. Civil Eng. 89 (1963) 31–59. [30] G.F. Malash, M.I. El-Khaiary, Piecewise linear regression: a statistical method for the analysis of experimental adsorption data by the intraparticle-diffusion models, Chem. Eng. J. 163 (2010) 256–263. [31] G.E. Boyd, A.W. Adamson, L.S. Myers Jr., The exchange adsorption of ions from aqueous solutions by organic zeolites. II. Kinetics, J. Am. Chem. Soc. 69 (1947) 2836–2848. [32] D. Reichenberg, Properties of ion exchange resins in relation to their structure. III. Kinetics of exchange, J. Am. Chem. Soc. 75 (1953) 589–598. [33] K.V. Kumar, A. Kumaran, Removal of methylene blue by mango seed kernel powder, Biochem. Eng. J. 27 (2005) 83–93. [34] J. Lyklema, Fundamentals of Interface and Colloid Science, Vol. I. Fundamentals, Academic Press, New York, 1993. [35] R. Coughlin, F.S. Ezra, Role of surface acidity in the adsorption of organic pollutants on the surface of carbon, Environ. Sci. Technol. 2 (1968) 291–297. [36] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum, J. Am. Chem. Soc. 40 (1918) 1361–1403. [37] T.W. Weber, R.K. Chakravorti, Pore and solid diffusion models for fixed bed adsorbents, J. Am. Inst. Chem. Eng. 20 (1974) 228–238. [38] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906) 385–471. [39] O. Redlich, D.L. Peterson, A useful adsorption isotherm, J. Phys. Chem. 63 (1959) 1024–1026. [40] R. Sips, Combined form of Langmuir and Freundlich equations, J. Chem. Phys. 16 (1948) 490–495. [41] K.K.H. Choy, J.F. Porter, G. McKay, Langmuir isotherm models applied to multicomponent sorption of acid dyes from effluent onto activated carbon, J. Chem. Eng. Data 45 (2000) 575–584. [42] J.S. Jain, V.L. Snoeyink, Adsorption from biosolute systems on active carbon, J. Water Pollut. Control Fed. 45 (1973) 2463–2479. [43] O. Hamdaoui, E. Naffrechoux, J. Suptil, C. Fachinger, Ultrasonic desorption of p-chlorophenol from granular activated carbon, Chem. Eng. J. 106 (2005) 153–161.