Adsorption behavior of estrogenic compounds on carbon nanotubes from aqueous solutions: Kinetic and thermodynamic studies

Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

Contents lists available at SciVerse ScienceDirect

Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec

Adsorption behavior of estrogenic compounds on carbon nanotubes from aqueous solutions: Kinetic and thermodynamic studies Lateefa A. Al-Khateeb, Abdualah Y. Obaid, Najwa A. Asiri, Mohamad Abdel Salam * Chemistry Department, Faculty of Science, King Abdulaziz University, PO Box 80200, Jeddah 21589, Saudi Arabia

A R T I C L E I N F O

Article history: Received 18 May 2013 Accepted 17 June 2013 Available online 25 June 2013 Keywords: Carbon nanotubes Estrogenic compounds Adsorption Kinetic Thermodynamic

A B S T R A C T

In this study different estrogenic compounds, estrone (E1), 17b-estradiol (E2), and 17a-ethinylestradiol (EE2), were removed from the model and real solution by the newly emerged multi-walled carbon nanotubes. The effects of different factors which affect the removal process were studied and optimized for efficient removal. The kinetics of E1, E2, and EE2 adsorption on MWCNTs were analyzed using different kinetic models and the results showed that the removal was mainly a pseudo-second-order process. The thermodynamic study showed the spontaneity and exothermic nature of the removal process, with negative entropy. ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

1. Introduction Recently, compounds categorized as endocrine disrupting chemicals (EDCs) have drawn attention as emerging pollutants in sources of drinking water. EDCs are a structurally diverse group of compounds that may adversely affect the health of humans, wildlife and fishes, or their progenies, by interaction with the endocrine system [1]. They include chemicals used heavily in the past, in industry and agriculture, such as polychlorinated biphenyls, bisphenol A, organochlorine pesticides, and estrogens, including estrone (E1), 17b-estradiol (E2), as well as the synthetic estrogen, 17a-ethinylestradiol (EE2). Recent research showed that natural and synthetic estrogens are effective endocrine disrupters as they cause endocrine alterations on aquatic organisms at sub ng/ L levels [2,3]. Hence effective treatments are needed for estrogen remediation. Currently, there are different traditional methods of water treatment, such as activated sludge treatment [4], adsorption by solid adsorbent [5], ozonation [6], filtration membranes [7,8], or photocatalytic oxidation [9]. Among the above mentioned methods, adsorption is regarded as a promising method for the removal of estrogens as it removed all the estrogens effectively without producing by-products, which in most cases could be as harmful as the original pollutants [10,11], and recycle of both the solid adsorbent and estrogens could be optimized. Recently, many

* Corresponding author. Tel.: +966 541886660; fax: +966 2 6952292. E-mail addresses: [email protected], [email protected] (M. Abdel Salam).

newly emerged solid adsorbents were used for the removal of estrogens from polluted water such as carbonaceous resin and high-silica zeolites [12], molecularly imprinted polymer [13], and carbon nanotubes [14,15]. Carbon nanotubes have come under intense multidisciplinary study because of their unique physical and chemical properties. These characteristics allow them to be used in a broad range of applications [16–20]. Many research studies showed the ability of CNTs for the removal of different pollutants such as heavy metals [21,22], lipase [23], resorcinol [24], tetracycline [25], aniline [26], pentachlorophenol [27] and many other pollutants [28] from aqueous solution. However, the studies on the adsorption of EDCs such as estrogens by MWCNTs are still scarce in literature [8,15,29]. Further investigations on the adsorption/removal of the most common EDCs; estrone, 17b-estradiol, and 17a-ethinylestradiol, by multi-walled carbon nanotubes as a newly emerging and promising adsorbent, are essential for aquatic environment remediation. In this manuscript, MWCNTs were used for the effective removal of targeted estrogens from aqueous solutions. The effects of different adsorption conditions such as solution pH, temperature, estrogens concentration, and adsorption time, were studied and optimized. Also, the adsorption process was studied from the kinetics and thermodynamics point of view in order to achieve a better understanding of the adsorption process. Kinetic studies are crucial to understanding the factors and means of transport for estrogens from the aqueous phase to the solid MWCNTs phase. Thermodynamics calculation of the adsorption process is required to understand the mechanism of adsorption and its spontaneity by calculating the different thermodynamic parameters.

1226-086X/$ – see front matter ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jiec.2013.06.023

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

2. Materials and methods

The adsorption of E1, E2, and EE2 on the walls of the glass flasks and the filter paper was determined by running a blank experiment without MWCNTs and was found to be negligible.

2.1. Materials Multi-walled carbon nanotubes (MWCNTs) were obtained from Shenzhen Nano-Technologies, China. The MWCNTs were sonicated in acetone for 2.0 h as a pre-treatment step in order to remove any organic compounds adsorbed previously on the MWCNTs surface during the production or storage, then the MWCNTs were filtered and washed with acetone many times and dried in an oven at 110 8C, in order to evaporate the acetone. Estrone, 17b-estradiol, and 17a-ethinylestradiol, were laboratory reagent grade and obtained from Sigma–Aldrich, Canada. All other chemicals were obtained from Sigma–Aldrich (analytical grade) and all solutions were prepared using deionized water. 2.2. Characterization techniques A transmission electron microscope (TEM) (type JEOL JEM-1230 operating at 120 kV attached to a CCD camera) was used to characterize the MWCNTs for their morphological structure. The specific surface area of the different MWCNTs was determined from nitrogen adsorption/desorption isotherms measurements at 77 K using a model NOVA 3200e automated gas sorption system (Quantachrome, USA). 2.3. Analytical method Estrone, 17b-estradiol, and 17a-ethinylestradiol concentrations were determined using HPLC (Hewlett Packard 1100 series liquid chromatograph (Avondale, CA)) equipped with pump, auto injector, and diode array detector set at 200 nm and ChemStation for data collection and analysis, using different percentage of acetonitrile/water at 1 ml/min as the mobile phase. The samples (10 ml) were injected. The stationary phase was Zorbax Eclipse column (XDB-C18, 5 mm, 150 mm  4.6 mm ID). 2.4. Adsorption experiments Kinetic experiments were carried out to identify the adsorption rate and to find the effect of time and temperature on the adsorption process. The experimental procedures could be described as follows: (1) a series of 10.0 ml solutions containing 5 mg L1 of E1, E2, and EE2 was prepared in 25 ml glass bottle and kept at a certain temperature, (2) the pH of the solution was adjusted to 6.0 using universal pH buffer (Britton–Robinson buffer), (3) 100.0 mg of MWCNTs was added into the solution, (4) the solution was shaked continuously for a certain period of time, (5) after the completion of preset time intervals, the solution was taken and immediately filtered through a filter paper to collect the supernatant, (6) the residual of E1, E2, and EE2 concentrations in the aqueous solution were then determined by HPLC and the amounts adsorbed were calculated as follows: qt ¼

ðC o  C t ÞV m

2.5. Real water samples Two real samples were used to study the efficiency of MWCNTs for the removal of the estrogenic compounds under investigation. A tap water sample (TWS) was collected from our lab after allowing the tap water to flow for 10 min. The wastewater sample (MBR 6000 STP) was collected from the Membrane Bio-Reactor Technology Waste Water Treatment Plant – King Abdulaziz University (KAUWW), Jeddah City (Latitude deg. North 21.487954, Longitude deg. East 39.236748). Both samples were filtered through 0.45 mm Millipore filter paper and kept in Teflon1 bottles at 5 8C in the dark. 100 mg MWCNTs were packed inside a glass column as was explained previously [30], and 25 ml of the real sample solution was allowed to flow inside the column with a flow rate of 1.7 ml/min. 3. Results and discussion 3.1. Characterization of multi-walled carbon nanotubes Fig. 1 shows the transmission electron microscope image of the MWCNTs. The MWCNTs were long and entangled around each other with an average diameter between 30 and 60 nm and an average length between 2 and 5 mm. The TEM analysis verified the hollow structure of MWCNTs used and showed that the inner diameter was between 6 and 9 nm. The Nitrogen adsorption/ desorption isotherms were determined from N2 adsorption measured at 77 K and specific surface area was calculated using the Brunauer–Emmett–Teller equation and was found to be 84.3 m2 g1. 3.2. Optimization of the removal parameters The effects of different parameters which affected the removal of E1, E2, and EE2, by MWCNTs from an aqueous solution were investigated and optimized. The effect of the contact time on the removal of E1, E2, and EE2 by MWCNTs from solution was studied and the results are shown in Fig. 2. It is clear from the figure that the % removed was increased sharply within the first 5 min as the % removed reached 83.2, 89.2, and 91.2% for E1, E2, and EE2, respectively. Further increase in the contact time to 180 min associated with a slight increase in the % removed to 85.9, 94.4, and

(1)

where qt is the amount of E1, E2, and EE2 adsorbed by the MWCNTs (mg/g), C0 is the initial concentration (mg L1), Ct is the final concentration after a certain period of time (mg L1), V is the initial solution volume (L) and m is the MWCNT dose (g). The percentage of removed in solution was calculated using Eq. (2): % Removed ¼

Co  Ct  100 Co

917

(2)

Fig. 1. Transmittance electron microscope images for the MWCNTs.

918

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

Fig. 2. The effect of contact time on the % removed for E1, E2, and EE2. (Experimental conditions: 100 mg MWCNTs, 10 ml sample volume, pH 7.0, 298 K, and E1, E2, and EE2 concentration 5.0 mg L1.)

94.4% for E1, E2, and EE2, respectively. The low % removal of E1 could be attributed to its low log Kow value; 3.43, compared with E2 and EE2; 3.94 and 4.15, respectively. The higher the Kow, the more the hydrophobic nature of the pollutants [31]. The % removed are in consistent with the hydrophobicity in the order of EE2 > E2  E1, as the hydrophobic adsorbate will adsorb preferentially on the hydrophobic MWCNTs. The effect of MWCNTs’ mass was studied at pH 7.0, 298 K, and E1, E2, and EE2 concentration of 5.0 mg L1, and the results showed that the % removed increased significantly for E1, E2, and EE2, by increasing the MWCNTs’ mass until they reach their maximum at 200 mg MWCNTs. The % removed was increased from 40.0, 57.9, 73.6% when 25 mg MWCNTs were used, to 99.1, 96.4, and 98.4%, when 200 mg MWCNTs were used for the removal of E1, E2, and EE2; respectively. This enhancement in the % removed by increasing MWCNTs’ mass was mainly due to the availability of

more binding sites on the MWCNTs’ surface for the removal and adsorption of the estrogenic compounds. For the rest of the experiments, the MWCNTs’ mass was kept at 100 mg. The effect of the solution pH on the removal of E1, E2, and EE2, by MWCNTs was studied at 120 min, and 298 K solution temperature, using 100 mg MWCNTs, and E1, E2, and EE2 concentration of 5.0 mg L1, and the results are presented in Fig. 3. At acidic pH value of 2.0, the % removed for E1, E2, and EE2, were very low; 75.6, 78.7, and 90.1, respectively, which could be attributed to the competition between the hydronium ions (H+) and the estrogenic compounds for the binding sites available at the MWCNTs’ surface. Further increase in the pH to 6.0, enhanced significantly the % removed to 85.6, 93.3, and 93.4%, for E1, E2, and EE2, respectively. Further increase in the solution pH to 9.0, did not change the % removed greatly. Further increase in the solution pH to 11.0, slightly decreased the % removed, which could be

Fig. 3. The effect of solution pH on the % removed for E1, E2, and EE2 by MWCNTs. (Experimental conditions: 60 min, 10 ml sample volume, 298 K, 100 mg MWCNTs, and E1, E2, and EE2 concentration 5.0 mg L1.)

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

919

attributed to the repulsion between the ionized and negatively charged E1, E2, and EE2, with the MWCNTs’ surface covered with the negatively charged hydroxide ions (HO), as pKa values for the estrogenic compounds under investigation is around 10.5 [32]. The effect of the E1, E2, and EE2 concentrations on the % removal by carbon nanotubes was studied and the results showed that increasing E1, E2, and EE2 concentrations at the same MWCNT mass (100 mg), was accompanied by a gradual decrease in the % removed. Increasing the concentration from 1.0 mg L1 to 20 mg L1 decreased the % removed from 94.7, 99.0, and 95.3%, to 71.1, 65.2, and 79.9%, E1, E2, and EE2, respectively, due to the lack of more active sites available for adsorption. The effect of the solution temperature on the % of E1, E2, and EE2 removed from solution was studied due to its importance for the evaluation of the solid adsorbents suitability and cost estimation. The results showed that raising the solution temperature from 281 K to 333 K, was associated with a slight decrease in the % removed from 90.8, 95.4, and 95.7%, to 79.7, 90.3, and 90.3%, for E1, E2, and EE2, respectively. This may indicates the exothermic nature of the adsorption process, as the % removed decreased by raising the solution temperature. Further discussion will be continued in the thermodynamic study. 3.3. Kinetic studies Adsorption kinetics is the study of the adsorption process rate to understand the factors that affect the adsorption process. Study of adsorption kinetics includes careful monitoring of the experimental conditions which influence the speed of adsorption, such as solution temperature, and therefore help attain equilibrium in a reasonable length of time. Kinetic studies usually use mathematical models to describe the interactions between the adsorbent and the adsorbate. Once the adsorption rates and the dependent factors are clearly identified, the development of adsorbent materials for industrial application could be evaluate and optimized. The temperature of the solution is considered one of the important factors which affect the adsorption kinetics and, accordingly, the removal process. Also, when the removal efficiency of a certain pollutant from an aqueous solution is temperature dependent, it might affect the suitability of the adsorbent, as raising the solution temperature mostly consumes time and fuel. The effect of temperature on the removal of E1, E2, and EE2 from an aqueous solution by MWCNTs was studied kinetically and the results were presented in Fig. 4. The figure showed that the removal process was enhanced by decreasing the solution temperature. The % removal of E1 was decreased from 91.1, to 85.9, then 79.5%, whereas for E2 the % removed was decreased from 96.3, to 94.4, and 92.0%, meanwhile for EE2 the % removed was decreased from 96.4, to 94.4, and 93.0, when the temperature of the solution was raised from 281 K, to 298 K, and 333 K, respectively. This inhibition in the removal of E1, E2, and EE2 with MWCNTs by raising the solution temperature may indicate the exothermic nature of the adsorption process. The figure showed also that the adsorption reached equilibrium within 5 min at all temperatures, which could be mainly due to the high hydrophobicity and low vapor pressures of E1, E2, and EE2, which prefer to transfer from the aqueous phase to the hydrophobic solid phase of the MWCNTs. The adsorption experimental data were treated kinetically using the most common and well known kinetic models in order to understand the nature of the adsorption process. The fractional power function model is a modified form of the Freundlich equation [33] and it could be written in its linearized form as the following: ln qt ¼ ln a þ b ln t

(3)

Fig. 4. The effect of solution temperature on the % removed for E1, E2, and EE2 by MWCNTs. (Experimental conditions: 10 ml sample volume, 100 mg MWCNTs, pH 7.0, and E1, E2, and EE2 concentration: 5.0 mg L1.)

where qt (mg/g) is the amount of metal ion adsorbed per unit mass of MWCNTs at any time t, while a and b are coefficients with b < 1. The function ab is the specific sorption rate when t = 1 min. Table 1 presents the kinetic parameters obtained by application of the fractional power function equation to the adsorption of E1, E2, and EE2 on MWCNTs at different solution temperatures. Although linear relationship exists between ln qt and ln t, the correlation coefficient values were low in most of the cases, as it is shown in Table 1. This may indicates that the fractional power function model was not the appropriate model to describe the adsorption of the E1, E2, and EE2 on MWCNTs.

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

920

Table 1 Different kinetic model parameters for the adsorption of estrone, 17b-estradiol, and 17a-ethinylestradiol, on MWCNTs at different temperatures. Temperature

17b-estradiol

Estrone A

Fractional power model 281 K 0.414 298 K 0.410 0.355 333 K Temperature

ab

R

A

b

ab

R

0.021 0.009 0.024

0.008 0.004 0.009

0.811 0.969 0.938

0.463 0.436 0.424

0.008 0.017 0.016

0.004 0.008 0.007

0.955 0.931 0.944

17b-estradiol qe,calc (mg/g)

Pseudo-first-order kinetic model 281 K 0.456 0.018 298 K 0.423 0.020 333 K 0.397 0.023

Pseudo-second-order kinetic model 281 K 9.25 0.456 298 K 8.42 0.423 333 K 6.31 0.397 Temperature

Temperature

Liquid film diffusion model 281 K 298 K 333 K

0.004 0.006 0.018

0.002 0.003 0.008

0.890 0.889 0.881

17a-ethinylestradiol R2

qe,exp (mg/g)

qe,calc (mg/g)

k1

R2

0.010 0.034 0.018

0.567 0.990 0.737

0.482 0.472 0.460

0.012 0.017 0.034

0.022 0.020 0.029

0.946 0.782 0.939

0.482 0.472 0.465

0.010 0.006 0.032

0.091 0.016 0.117

0.939 0.439 0.796

17a-ethinylestradiol

qe,calc (mg/g)

R2

k2 (g/mg min)

qe,exp (mg/g)

qe,calc (mg/g)

R2

k2 (g/mg min)

qe,exp (mg/g)

qe,calc (mg/g)

R2

0.456 0.430 0.399

0.999 0.999 0.999

12.91 9.08 4.45

0.482 0.472 0.460

0.482 0.473 0.460

0.999 0.999 0.999

44.9 26.8 9.41

0.482 0.472 0.465

0.482 0.472 0.465

0.999 0.999 0.999

2

17a-ethinylestradiol

b (mg/g min)

R

a (g/mg min)

b (mg/g min)

R

a (g/mg min)

b (mg/g min)

R2

0.009 0.004 0.009

0.815 0.967 0.934

4.00  1053 1.36  1026 3.55  1028

0.004 0.008 0.007

0.956 0.993 0.947

4.76  1098 2.99  1068 1.95  1025

0.002 0.003 0.008

0.890 0.889 0.881

2

17b-estradiol 1/2

)

Intra-particle diffusion model 281 K 1.38  102 298 K 3.0  103 333 K 8.3  103 Temperature

0.472 0.458 0.429

k1

Estrone kid (mg/g min

R2

qe,calc (mg/g)

17b-estradiol

Elovich kinetic model 281 K 3.96  1021 298 K 7.71  1047 333 K 3.48  1018

ab

qe,exp (mg/g)

Estrone

a (g/mg min)

b

R2

17b-estradiol qe,exp (mg/g)

A

k1

Estrone k2 (g/mg min)

17a-ethinylestradiol 2

b

Estrone qe,exp (mg/g)

Temperature

2

2

17a-ethinylestradiol

1/2

C (mg/g)

R

kid (mg/g min

0.395 0.409 0.346

0.901 0.968 0.870

4.1  103 1.2  102 1.2  102

)

2

C (mg/g)

R

kid (mg/g min1/2)

C (mg/g)

R2

0.459 0.421 0.410

0.984 0.995 0.993

2.54  103 2.46  103 1.54  102

0.470 0.456 0.407

0.831 0.961 0.999

17b-estradiol

Estrone

17a-ethinylestradiol

kfd (min1)

R2

kfd (min1)

R2

kfd (min1)

R2

0.016 0.011 0.011

0.263 0.192 0.260

0.040 0.048 0.030

0.679 0.633 0.836

0.155 0.060 0.189

0.610 0.404 0.654

Lagergren pseudo-first-order kinetics is one of the most frequent kinetic models used to describe the adsorption of different adsorbates from an aqueous solution by solid adsorbent [34]. It could be written as the following:

given as:

lnðqe  qt Þ ¼ ln qe  k1 t

where k2 (g/(mg min)) is the pseudo-second-order rate coefficient, and qe and qt are the values of the amount adsorbed per unit mass at equilibrium and at any time t, respectively. The plot of t/qt and t of Eq. (5) must give a linear relationship from which qe and k2 could be estimated from the slope and intercept of the plot, respectively. By applying the pseudo-second-order rate equation to the adsorption of E1, E2, and EE2 experimental data converged very well with excellent regression coefficients (R2 > 0.99) and straight lines as it is presented in Fig. 5. These findings confirm the suitability of the pseudo-second-order rate equation for the description of the E1, E2, and EE2 adsorption by MWCNTs from aqueous solution. The pseudo-second-order rate equation parameters; qe, and k2, calculated from the slope and intercept of the plot of t/qt vs. t are shown in Table 1. It is clear from the table that the amount adsorbed per unit mass of MWCNTs at equilibrium (qe,calc) calculated from the slope of the pseudo-second-order plot were in good agreement with experimental values (qe,exp). Generally,

(4)

where k1 (min1) is the pseudo-first-order adsorption rate coefficient, and qe and qt are the values of the amount adsorbed per unit mass at equilibrium and at any time t, respectively. Plotting ln (qe  qt) vs. t for E1, E2, and EE2 at different temperatures did not converge well and did not give straight lines as it is clear from Table 1 with unacceptable regression coefficients. In few cases where acceptable regression coefficients were obtained, the estimated values of the amount adsorbed at equilibrium were far from the experimental values. This indicated that the pseudo-first-order Lagergren equation is not suitable for describing the adsorption of the E1, E2, and EE2 on MWCNTs. The pseudo-second-order equation has also been considered for describing the adsorption of the E1, E2, and EE2 on MWCNTs [35]. The linearized form of the pseudo-second-order rate equation is

t 1 t ¼ þ qt k2 q2e qe

(5)

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

921

adsorption behavior was exothermic, physical and reversible in nature as the adsorption capacities decreased by raising the solution temperature [36]. Also, the sorption behavior of E2, and EE2 in an activated sludge biomass was studied and the results revealed the exothermic nature of the adsorption as the portioning coefficient values were inversely related to the solution temperature [37]. In another study, a nylon membrane was used for the removal of E2 from aqueous solution and the results indicated that the adsorption was exothermic [38]. The applicability of the pseudo-second-order rate equation for the adsorption of E1, E2, and EE2 by MWCNTs agreed well with previous studies when another MWCNTs was used for the removal of EE2 [15] and the results showed that equilibrium was attained after 30 min, with a maximum adsorption capacity of 0.12 mg/g, whereas in the current study is distinguished by the low equilibration time; 5 min, and a higher adsorption capacity; 0.472 mg/g for the adsorption of EE2 by MWCNTs. The kinetic model by Elovich usually describes the adsorption of adsorbate by solid adsorbent in aqueous medium [39]. The linear form of the Elovich equation is given by the following equation: qt ¼ b lnðabÞ þ b ln t

(6)

where the a and b are the Elovich coefficients represent the initial adsorption rate (g/(mg min)) and the desorption coefficient (mg/ (g min)), respectively. The Elovich coefficients a and b were calculated from the slope and intercept of the plot of qt vs. ln t and the results are presented in Table 1. The table shows that the correlation coefficients were not satisfactory in most of the cases, which indicated that the Elovich model is not appropriate for the description of E1, E2, and EE2 adsorption by MWCNTs. Usually, the adsorption of any adsorbates from an aqueous solution to a solid adsorbent occurs in consecutive steps. This steps includes the migration of adsorbate from the bulk solution to the external surface of the solid adsorbent, followed by the diffusion through the boundary layer to the external surface of the solid adsorbent; followed by the adsorption at an active site on the solid adsorbent surface and finally the intra-particle diffusion and adsorption through the solid adsorbent pores and aggregates. Intra-particle diffusion model is another model which could be used for the description of the adsorption of E1, E2, and EE2, by MWCNTs [40]. Intra-particle diffusion model could be expressed as the following: qt ¼ kid t 1=2 þ C

Fig. 5. Pseudo-second order plots for E1, E2, and EE2 adsorbed on MWCNTs at different temperatures. (Experimental conditions: 10 ml sample volume, pH 7.0, 100 mg MWCNTs and E1, E2, and EE2 concentration: 5.0 mg L1).

raising the solution temperature was accompanied by a decrease in the amount adsorbed at equilibrium (qe) demonstrating the exothermic nature of the adsorption process of E1, E2, and EE2 adsorption by MWCNTs from aqueous solution. The same finding was observed when different adsorbents were used for the removal of different estrogenic compounds from an aqueous solution [36–38]. For example, the adsorption behaviors of E1, E2, and EE2 using activated sludge, and the results revealed that the

(7)

where qt is adsorption capacity at any time (t), kid is the intraparticle diffusion rate constant (mg/g min1/2) and C (mg/g) is a constant proportional to the thickness of the boundary layer. One of the important conditions for the validity of the intra-particle diffusion model is that the converged straight line has to pass through the origin. By applying the intra-particle diffusion model to the adsorption experimental data of E1, E2, and EE2 at different temperatures, they did not converge well and did not have straight lines that passed through the origin as is presented in Fig. 6. The last part of the adsorption data could be attributed to the intraparticle diffusion of the E1, E2, and EE2 molecules through the MWCNTs pores and aggregates. This part converges considerably well for E1, E2, and EE2 at all temperatures, and the results are tabulated in Table 1. This indicates that the intra-particle diffusion is a part of the adsorption mechanism, but it is not the rate determining step. Liquid film diffusion model is another kinetic model that assumes the flow of the adsorbate molecules through a liquid film surrounded the solid adsorbent [41]. The liquid film diffusion model could be expressed as the following: lnð1  FÞ ¼ k fd  t

(8)

922

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

rate and mechanism of adsorption [42–44]. The initial sharp portion might be due to the boundary layer diffusion of the estrogenic compounds through the aqueous phase to the external surface of the MWCNTs or the boundary layer diffusion of the E1, E2, and EE2 molecules, then the fast liquid film diffusion followed by the intra-particle diffusion, and finally the last step where equilibrium attained and the intra-particle diffusion starts to slow down due to extremely low E1, E2, EE2 concentrations in the aqueous phase [42–44]. From the above results and discussion, it could be concluded that the application of MWCNTs for the removal of E1, E2, and EE2 from an aqueous solution followed mainly the pseudo-secondorder kinetics as was confirmed by the converging of the experimental data, the excellent correlation coefficients (>0.99), and the good agreement between the calculated and experimental amount adsorbed per unit mass of MWCNTs. The mechanism of the adsorption of the estrogenic compounds by MWCNTs from aqueous solution was explored and the results show that it takes place in different steps including the liquid film diffusion and the intra-particle diffusion, but they were not the rate-determining steps. The adsorption mainly is due to the p–p electron-donor acceptor interaction between the E1, E2, and EE2 molecules and MWCNTs [32]. Comparing the adsorption capacities of the present study with previous research work showed suitability of the MWCNTs as a potential adsorbent for the removal of different estrogenic compounds compared with other solid adsorbent. The adsorption capacities were found to be 0.423 mg/g, 0.472 mg/g, and 0.472 mg/g for E1, E2, and EE2, respectively, whereas in the previous studies the adsorption capacities were 2533.34 ng/g, 2020.78 ng/g, 2234.09 ng/g, when activated sludge was used for the removal of E1, E2, and EE2, respectively [36], 62 ng/g for the removal of E1 by hydrophobic hollow fiber membrane [45]. 3.4. Thermodynamic studies Thermodynamic parameters; Gibbs free energy change (DG), enthalpy change (DH) and, entropy change (DS), were calculated to evaluate the thermodynamic feasibility and the spontaneous nature of the E1, E2, and EE2 adsorption by MWCNTs. Thermodynamic parameters were calculated from the variation of the thermodynamic distribution coefficient D with a change in temperature according to the equation: D¼ Fig. 6. Intra-particle diffusion model plots for E1, E2, and EE2 adsorbed on MWCNTs at different temperatures. (Experimental conditions: 10 ml sample volume, pH 7.0, 100 mg MWCNTs and E1, E2, and EE2 concentration 5.0 mg L1.)

where F is the fractional attainment of equilibrium (F = qt/qe) and kfd (min1) is the film diffusion rate coefficient. A linear plot of ln(1  F) versus t with zero intercept suggests that the kinetics of the adsorption process is controlled by diffusion through the liquid film. Application of the liquid film diffusion model to the adsorption of E1, E2, and EE2 by MWCNTs at different temperatures did not converge, and the regression coefficient values were very low as they were presented in Table 1 This indicates that the liquid film diffusion was not the rate-determining step. Based on the above results, it is clear that the mechanism of interaction between the estrogenic compounds and the MWCNTs is complex. The application of the intra-particle diffusion model, and liquid film diffusion model, on the experimental data yielded different straight lines without passing through the origin, which indicates some degree of boundary layer control. This further show that the intra-particle diffusion and liquid film diffusion are not the only rate-controlling step, but also other processes may control the

qe Ce

(9)

where qe is the amount of E1, E2, and EE2 adsorbed by the MWCNTs (mg/g) at equilibrium, and Ce is the equilibrium concentration of E1, E2, and EE2 in solution (mg L1). The DH and DS could be calculated according to the following equation [43,44]: log D ¼

DS R



DH 2:303 RT

(10)

Plotting log D vs. 1/T gives straight lines, and the DH and DS values were calculated from the slope and the intercept of the straight line, respectively. The DG values were calculated at 298 K from the relation:

DG ¼ DH  T DS

(11)

Thermodynamic parameters for the adsorption of E1, E2, and EE2 by MWCNTs were calculated using Eqs. (10) and (11). The free energy change, DG, is negative as would be expected for a product favored and spontaneous reaction. The negative values of DG were: 13.99 kJ/mole, 11.95 kJ/mole, and 9.982 kJ/mole, for the adsorption of E1, E2, and EE2 by MWCNTs from aqueous solution, respectively. DH values were all negative: 14.16 kJ/mole, 12.08 kJ/mole, and 10.11 kJ/mole, for the adsorption of E1,

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

923

E2, and EE2 by MWCNTs from aqueous solution, respectively. These values verify the exothermic nature of the E1, E2, and EE2 adsorption on MWCNTs, which was the explanation for the decrease in adsorption at higher temperature. The magnitude of DH suggests a weak type of bonding between the E1, E2, and EE2 and the MWCNTs such as physical adsorption and not chemical adsorption. The values obtained in this study agreed with the other values found in the literature for adsorption of estrogenic compounds, in general, on different adsorbents [36–38]. The negative values of DS, 0.56 kJ/mole K, 0.46 kJ/mole K, and 0.42 kJ/mole K, suggested the decrease in the degree of freedom at the solid–liquid interface is due to the immobilization of E1, E2, and EE2 on the MWCNTs’ surface. Generally, the negative values of DG, DH, and DS suggested that the removal process is an enthalpydriven process.

process is spontaneous as the Gibbs free energy values were negative. The negative enthalpy values for the adsorption of E1, E2, and EE2 by MWCNTs from an aqueous solution verified the exothermic nature of the removal process. The change in the entropy values were negatives indicating the decrease in randomness due to the physical adsorption of E1, E2, and EE2 molecules from the aqueous solution to the carbon nanotubes solid surface. Generally, the negative values of DG, DH, and DS suggested that the removal process is enthalpy-driven process. Finally, carbon nanotubes showed that they can be considered as a promising adsorbent for the removal of EDCs such as estrone, 17b-estradiol, and 17a-ethinylestradiol, from an aqueous solution.

3.5. Environmental applications

This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under grant number (401/247/1431). The authors, therefore, acknowledge with thanks DSR technical and financial support.

To study the applicability of MWCNTs for the removal of the estrogenic compounds, real environmental samples must be investigated. Two real environmental water samples were used for this purpose, namely, tap water sample (TWS), and wastewater collected from the King Abdulaziz University Wastewater (KAUWW). The concentrations of E1, E2, and EE2 were measured in both samples and the results showed that all pollutants were below the detection limit of the HPLC. Then, the two samples were spiked with 5.0 mg L1, and left for equilibration and then 25.0 ml of the spiked solution; at pH 6.0, were allowed to flow through the 100 mg MWCNTs packed inside the glass tube. The % adsorbed was calculated by comparing the estrogenic compounds concentrations before and after the adsorption. The % E1, E2, and EE2 removed from the real samples were calculated and were found to be 100%, 100%, and 100%, for the tap water sample, and 54.9%, 95.8%, and 94.2%, for the KAUWW samples spiked with 5.0 mg L1. These values agreed very well with those obtained in the model solution as the % adsorbed from E1, E2, and EE2 were 86.0%, 94.2% and 94.2%, respectively. This trend was expected for the TWS as the tap water, generally, contains very low concentrations of inorganic ions and organic pollutants. In the case of KAUWW sample, almost the same trend was observed, except for E1, the % adsorbed was lower than model solution and TWS. This could be attributed to the presence of other pollutants in the KAUWW sample which could compete with the estrone for the available binding sites available at the MWCNTs. This was not the case for E2 and EE2, as both of them were characterized by high binding affinities to solid adsorbent such as MWCNTs. 4. Conclusion The removal of the endocrine disrupting chemicals; estrone, 17b-estradiol, and 17a-ethinylestradiol, from an aqueous solution at different temperatures by multi-walled carbon nanotubes was studied. The results showed that most of EDC’s under investigation were removed within 5.0 min, at pH in the range between 7 and 10, and using 100 mg MWCNTs. The adsorption capacities were found to be 0.423 mg/g, 0.472 mg/g, and 0.472 mg/g for E1, E2, and EE2, respectively, which is significantly higher than other adsorbents. The % removal follow the order of EE2  E2 > E1, which agreed well with Kow values, and the hydrophobicity order. The adsorption efficiency was temperature sensitive and decreased significantly by raising the solution temperature. The adsorption was studied kinetically using different kinetic models and the results showed that the removal process could be well explained by the pseudo-second order model. The thermodynamics study showed that the adsorption

Acknowledgements

References [1] B. Gillesby, T. Zacharewski, Environmental Toxicology and Chemistry 17 (1998) 3. [2] S. Rodriguez-Mozaz, M. Lopez de Alda, D. Barcelo, Journal of Chromatography A 1045 (2004) 85. [3] F.C. Robert, P.S.S. Rominder, H.X. Fu, Water Environment Research 79 (2007) 969. [4] A. Johnson, J. Sumpter, Environmental Science & Technology 35 (2001) 4697. [5] J.L. Zhou, R. Liu, A. Wilding, A. Hibberd, Environmental Science & Technology 41 (2007) 206. [6] M.G. Maniero, D.M. Bila, M. Dezotti, Science of the Total Environment 407 (2008) 105. [7] S. Lee, J. Lee, S. Kim, P. Park, J. Kim, C. Lee, Journal of Membrane Science 326 (2009) 84. [8] J. Heo, J.R.V. Flora, N. Her, Y. Park, J. Cho, A. Son, Y. Yoon, Separation and Purification Technology 90 (2012) 39. [9] M. Bistan, T. Tisˇler, A. Pintar, Catalysis Communications 22 (2012) 74. [10] R. Pereira, M. de Alda, J. Joglar, L.A. Daniel, D. Barcelo´, Chemosphere 84 (2011) 1535. [11] R. Pereira, C. Postigo, M. de Alda, L. Daniel, D. Barcelo´, Chemosphere 82 (2011) 789. [12] A. Rossner, S.A. Snyder, D.R.U. Knappe, Water Research 43 (2009) 3787. [13] I˙. Koc¸, G. Baydemir, E. Bayram, H. Yavuz, A. Denizli, Journal of Hazardous Materials 192 (2011) 1819. [14] L. Joseph, J. Heo, Y. Park, J.R.V. Flora, Y. Yoon, Desalination 281 (2011) 68. [15] A.K. Kumar, S.V. Mohan, Separation and Purification Technology 87 (2012) 22. [16] S. Sirivisoot, T.J. Webster, Comprehensive Biomaterials 5 (2011) 303. [17] R. Orinˇa´kova´, A. Orinˇa´k, Fuel 90 (2011) 3123. [18] D.A. Gomez-Gualdro´n, J.C. Burgos, J. Yu, P.B. Balbuena, Progress in Molecular Biology and Translational Science 104 (2011) 175. [19] C.H. Ahn, Y. Baek, C. Lee, S.O. Kim, S. Kim, S. Lee, S.-H. Kim, S.S. Bae, J. Park, J. Yoon, Journal of Industrial and Engineering Chemistry 18 (2012) 1551. [20] L. Pan, X. Pei, R. He, Q. Wan, J. Wang, Colloids and Surfaces B 93 (2012) 226. [21] S¸.S. Bayazit, I˙. I˙nci, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.03.023. [22] M. Abdel Salam, S.A. Kosa, G. Al-Zhrani, Comptes Rendus Chimie 15 (2012) 398. [23] N.Zˇ. Prlainovic´, D.I. Bezbradica, Z.D. Knezˇevic´-Jugovic´, S.I. Stevanovic´, M.L. Avramov Ivic´, P.S. Uskokovic´, MijinF Zˇ.D., Journal of Industrial and Engineering Chemistry 19 (2013) 279. [24] Q. Liao, J. Sun, L. Gao, Colloids and Surfaces A 312 (2008) 160. [25] L. Zhang, X. Song, X. Liu, L. Yang, F. Pan, J. Lv, Chemical Engineering Journal 178 (2011) 26. [26] H. Al-Johani, M. Abdel Salam, Journal of Colloid and Interface Science 360 (2011) 760. [27] M. Abdel Salam, R. Burk, Water, Air, & Soil Pollution 210 (2010) 101. [28] X. Ren, C. Chen, M. Nagatsu, X. Wang, Chemical Engineering Journal 170 (2011) 395. [29] R. Gao, X. Su, X. He, Y. Zhang, Talanta 83 (2011) 757. [30] M. Abdel Salam, M.S.I. Makki, M.Y. Abdelaal, Journal of Alloys and Compounds 509 (2011) 2582. [31] G. Ying, R.S. Kookana, Y. Ru, Environment International 28 (2002) 545. [32] B. Pan, D.H. Lin, H. Mashayekhi, B.S. Xing, Environmental Science & Technology 42 (2008) 5480. [33] R.C. Dalal, Communications in Soil Science and Plant Analysis 5 (1974) 531. [34] S. Lagergren, Zur theorie der sogenannten adsorption gel.oster Stoffe – About the theory of so-called adsorption of soluble substances. Kungliga Svenska Vetenskapsakademiens. Handlingar, Band 24, 1898, p. 1. [35] W. Rudzinski, W. Plazinski, Adsorption 15 (2009) 181.

924

L.A. Al-Khateeb et al. / Journal of Industrial and Engineering Chemistry 20 (2014) 916–924

[36] Y. Ren, K. Nakano, M. Nomura, N. Chiba, O. Nishimura, Water Research 41 (2007) 2341. [37] A.O. Ifelebuegu, S.C. Theophilus, M.J. Bateman, International journal of Environmental Science and Technology 7 (2010) 617. [38] M. Liu, E.P.C. Lai, Y. Yang, International Journal of Research and Reviews in Applied Sciences 11 (2012) 67. [39] S.H. Chien, W.R. Clayton, Soil Science Society of America 44 (1980) 265. [40] W.J. Weber, J.C. Morris, Journal of Sanitary Engineering Division, American Society of Civil Engineers 89 (1963) 31.

[41] G.E. Boyd, A.W. Adamson, L.S. Myers, Journal of the American Chemical Society 69 (1947) 2836. [42] E. Bulut, M. Ozacar, I.A. Sengil, Microporous and Mesoporous Materials 115 (2008) 234. [43] X. Yang, B. Al-Duri, Journal of Colloid and Interface Science 287 (2005) 25. [44] A. Kiran Kumar, S. Venkata Mohan, P.N. Sarma, Journal of Hazardous Materials 164 (2009) 820. [45] S. Chang, T.D. Waite, A.I. Scha¨fer, A.G. Fane, Desalination 146 (2002) 381.