Selective Adsorption of Methylparaben by Submicrosized Molecularly

Oct 19, 2012 - Keke Zhi , Lulu Wang , Yagang Zhang , Yingfang Jiang , Letao Zhang ..... Minjia Meng , Yonghai Feng , Min Zhang , Yanjun Ji , Jiangdong...
0 downloads 0 Views 5MB Size
Article pubs.acs.org/IECR

Selective Adsorption of Methylparaben by Submicrosized Molecularly Imprinted Polymer: Batch and Dynamic Flow Mode Studies Minjia Meng,† Zhipeng Wang,‡ Linli Ma,‡ Min Zhang,† Juan Wang,† Xiaohui Dai,† and Yongsheng Yan*,† †

School of Chemistry and Chemical Engineering, Jiangsu University, Zhenjiang 212013, China School of The Environment, Jiangsu University, Zhenjiang 212013, China



S Supporting Information *

ABSTRACT: Highly selective submicrosized molecularly imprinted polymer (SMIPMP) for methylparaben (MP) was synthesized by molecular imprinting technique with a sol−gel process on silica submicroparticles. The prepared SMIPMP was characterized by FT-IR, SEM, TG, and N2 adsorption−desorption techniques. Compared with microsized methylparaben imprinted polymer (MMIPMP) adsorbent, SMIPMP adsorbent with small particle size and high specific surface area showed faster adsorption rate and stronger adsorption capacity for MP. The maximum static adsorption capacity for MP of SMIPMP was 32.68 mg g−1, and the adsorption equilibrium could be reached in 40 min. The SMIPMP adsorbent could be used at least 5 times without significant loss in adsorption capacity. Compared with submicrosized nonimprinted polymer (SNIP), SMIPMP indicated excellent recognition and binding affinity toward MP molecules, whose selectivity coefficients for MP relative to methyl salicylate (MS) and p-hydroxybenzoic acid (p-HB) were 5.664 and 6.129, respectively. The mechanism for static adsorption of MP onto SMIPMP was found to follow Freundlich, Redlich-Peterson isotherm, and pseudo-second-order model. Thomas’ model was applied in the quantitative description and parametrization of the dynamic adsorption of MP to SMIPMP and SNIP, which showed that the linear and nonlinear methods were both suitable to predict the breakthrough curves but the nonlinear method was better. capacity.6,7 In order to overcome these drawbacks effectively, the surface molecular imprinting technique has been developed to allow binding sites on the surface of support with more accessible sites and fast binding kinetics.8,9 Inorganic submicroparticles are new functional supports, which possess a small dimension with extremely high surface to volume ratio and have come up as effective adsorbents. Furthermore, a variety of submicro- like submicro-Al2O3, submicro-SiO2 particles, submicro-iron, and so forth have been used as adsorbents for the removal of a variety of pollutants from aqueous solutions and industrial effluents.10−12 Among the different adsorbents, the submicrosized imprinted polymers have attracted much attention as selective adsorbent for separation and enrichment of trace environmental pollutants.13,14 In the submicrostructured imprinted materials, most of the imprinted sites are situated at the surface or in the proximity of surface, which greatly improve the binding capacity and kinetics and site accessibility of imprinted materials.15,16 To our knowledge, there are few reports on selective separation MP using submicrosized imprinted polymer as an adsorbent as well as the study of isotherm models, adsorption kinetics, especially of the dynamic adsorption behavior.

1. INTRODUCTION The parabens have been widely used as preservatives in cosmetics as well as in foods and medicines for many years due to its broad antimicrobial function.1−3 However, a possible relationship between breast cancer and prolonged dermal expositions to paraben containing products has been shown in recent studies.4 Therefore, these substance have been classified as emerging environmental pollutants by the U.S. Environmental Protection Agency (U.S. EPA). Among them, methylparaben (MP) is a methyl ester of p-hydroxybenzoic acid, and it was commonly employed mainly in the formulation of some personal care products as well as some canned foods and beverages. So, it is continuously released into aquatic media through domestic wastewater and food chains. Due to the very low concentrations of target, the effective separation technology has become the main difficulty in the analysis process. In addition, the direct analysis of MP is often disturbed because of the presence of complex matrix (a homologous series of parabens including methyl, ethyl, butyl, heptyl, and benzyl parabens) in the environmental samples. Consequently, the selective separation of MP from complex samples needs much more attention. Molecular imprinting technology is now known as a promising method for the design of molecularly imprinted polymer (MIP), which allows the creation of artificial recognition sites using the target analyte (the imprint molecule).5 Traditionally, due to the thick polymeric network, most MIP was prepared by bulky polymerization with poor site accessibility to the target molecules and low rebinding © 2012 American Chemical Society

Received: Revised: Accepted: Published: 14915

July 16, 2012 October 11, 2012 October 19, 2012 October 19, 2012 dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

(MNIP) were prepared analogously except using microsilica with a particle size of ca. 2.6 μm as a support. 2.4. Characterization. The SMIPMP was characterized by SEM and FT-IR spectra. The average particle sizes of the particles were determined on the SEM images based on a weighted-averages method. The specific surface area was determined by adsorption−desorption isotherms of nitrogen at 77 K. Thermal gravimetry (TG) curves of SMIPMP were acquired by using a thermoanalysis instrument under dynamic nitrogen atmosphere at a heating rate of 10 °C min−1. 2.5. Evaluating Binding Property. For the determination of the static binding behavior of SMIPMP for MP, numbers of 9 mL of MP solutions with different concentrations ranging from 25 to 350 mg L−1 were taken into centrifuge tubes. Then 0.02 g of SMIPMP or SNIP was added into the above solutions, respectively. These mixtures were shaken on a constant temperature shaker at 25 °C and centrifuged after reaching binding equilibrium, and the residual concentrations of MP in the supernatants were determined with UV spectrophotometry at the wavelength of 256 nm, respectively. The equilibrium binding amounts (qe, mg g−1) were calculated by the following (eq 1)

In this research, a highly selective submicrosized molecular imprinting polymer for methylparaben (SMIPMP) was synthesized on the support of silica submicroparticles by a sol−gel process. Aminopropyltriethoxysilane (APTES) and tetraethoxysilane (TEOS) were chosen as functional monomer and crosslinker, respectively. The SMIPMP was characterized by Fourier transform infrared spectrometer (FT-IR), transmission electron microscope (SEM), thermal gravimetry (TG), and N 2 adsorption−desorption techniques. Imprinting efficiency of SMIPMP was evaluated by static and dynamic adsorption tests. The selectivity of the obtained polymer was elucidated over competitive compounds, and the results demonstrated that the SMIPMP was capable of selective recognition of MP. Moreover, the SMIPMP was shown to be promising for regeneration.

2. MATERIALS AND METHODS 2.1. Instruments and Apparatus. Infrared spectra (4000−400 cm−1) was recorded on a Nicolet NEXUS-470 FT-IR apparatus (USA). The morphologies of the samples were obtained by scanning electron microscopy (SEM, S4800). The specific surface area of the samples was measured by NOVA2000 surface area and pore size analyzer (Quntachrome, USA). Thermal gravimetry (TG) curves of samples were acquired by using a thermoanalysis instrument (NETZSCH STA 449C). Peristaltic pump used in the experiment was supplied by Baoding Longer Precision Pump Co., Ltd. 2.2. Reagents and Materials. Tetraethoxysilane (TEOS) and aminopropyltriethoxysilane (APTES) were both purchased from Chemical Reagent (Shanghai, China). Methylparaben (MP), methyl salicylate (MS), and p-hydroxybenzoic acid (pHB) were all supplied by Sinopharm Chemical Reagent (Shanghai, China). All other chemicals used were of analytical grade and obtained commercially. Ultrapure water used throughout the experiments was obtained from laboratory purification system. 2.3. Synthesis and Characterizations of SMIPMP. 2.3.1. Preparation of Silica Submicroparticles. Silica submicroparticles were prepared by the hydrolysis of TEOS with ammonium hydroxide, according to the report by Stöber et al.17 In the typical experiment, the mixture containing total 4.45 mL of TEOS, 9 mL of H2O, 48.95 mL of NH3·H2O, and 37.6 mL of anhydrous ethanol was stirred at 25 °C for 5 h, resulting in the formation of white silica colloidal suspension. The silica submicroparticles were centrifugally separated from the suspension and washed with anhydrous ethanol four times. 2.3.2. Preparation of SMIPMP. At first, 400 mg silica submicroparticles were dispersed in 20 mL of methanol by ultrasonic vibration. Then, 0.304 g of MP (2 mmol) and APTES (1.87 mL, 8 mmol) were added into the suspension under stirring for 2 h to obtain the completely self-assembled composites with MP. Finally, 5 mL of TEOS (12.6 mmol) and 1 mL of HAc (1.0 mol L−1) were added to the mixture suspension under stirring for 18 h to obtain particles with a high cross-linking structure. The obtained polymer was then washed with 100 mL of a mixture of methanol and 6 mol L−1 HCl (1:1, V/V) for three times and neutralized with ultrapure water to remove MP. Finally, the polymer was dried at 60 °C for 12 h. The submicrosized nonimprinted polymer (SNIP) was prepared and treated in the same manner without MP. In comparison, the microsized methylparaben imprinted polymer (MMIPMP) and microsized nonimprinted polymer

qe =

(C0 − Ce)V W

(1)

where C0 (mg L−1) and Ce (mg L−1) are the initial and equilibrium concentrations of MP, respectively. V (mL) and W (g) are the volume of the solution and the weight of the polymer, respectively. For the dynamic adsorption tests, the experimental procedures are explained as follows. A certain amount (0.1 g) of SMIPMP or SNIP was packed into a piece of glass pipe of 1.5 cm i.d. and 30 cm length. The MP solution with a concentration of 100 mg L−1 was allowed to gradually flow through the packed column at the speed of 1 mL min−1. The influent to the column was pumped using a peristaltic pump, model YZ1515X. The effluents with 1 mL interval were collected, and the MP concentrations of these effluents were determined with UV spectrophotometry. The dynamic binding curve was plotted, and the Thomas model was applied to describe the experimental results of the dynamic adsorption of MP by SMIPMP or SNIP. 2.6. Evaluating Selectivities of SMIPMP or SNIP. To investigate the selectivity of SMIPMP or SNIP, both MS and pHB are selected as the contrast substances. 0.02 g of SMIPMP or SNIP was added into a 10 mL centrifuge tube, each of which contained 9 mL of solution with 25 mg L−1 of MP, MS, or pHB, respectively. After adsorption, the binding amounts of SMIPMP or SNIP for MP and the competition species were calculated as the procedure of static adsorption studies. The distribution coefficients (KD), selectivity coefficients (α) and relative selectivity coefficient (α′) of p-HB and MS with respect to MP can be obtained according to eq 2 − eq 4:

KD =

qe Ce

(2)

KD (L g−1) represents the distribution coefficient, and qe (mg g−1) and Ce (mg L−1) are the equilibrium binding amount and the equilibrium concentration of the MP, respectively. The selectivity coefficient α can be obtained according to the following equation 14916

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research α=

Article

KD(MP) KDj

(3)

where KDj represents the distribution coefficient of competition species. α′ can be defined as eq 4. αM and αN are the selectivity coefficients of SMIPMP and SNIP, respectively. α α′ = M αN (4) 2.7. Reusability Experiments. The adsorption−desorption experiments were carried out to investigate the reuse property of SMIPMP. During the regeneration experiment, 0.1 g of SMIPMP was treated according to the procedure of dynamic mode. Afterward, a volume of 5 mL mixture of methanol and HCl (1:1, V/V) was found to almost elute all the MP from the cartridge. Lastly, 5 mL of distilled water was used to wash again to neutral condition for the next adsorption−desorption cycle.

3. RESULTS AND DISCUSSION 3.1. Interaction between MP and APTES. The principle of molecular imprinting lies in the preservation of the prepolymerized host−guest structure into a polymer matrix. Thus it is of obvious importance that the functional monomers strongly interact with the template through hydrogen bonding, ionic bonding, or other interaction forces and form stable host−guest complexes prior to polymerization. In the present work, APTES was directly used as functional monomer for imprinting template (MP). Here, the interaction between MP and APTES was studied by UV spectroscopic analysis, and the results were shown in Figure 1.

Figure 2. Schematic procedure of SMIPMP preparation.

APTES was chosen not only as the intermediate to link silica submicroparticles but also as the functional monomer. HAc (1.0 mol L−1) and TEOS were used as a catalyst and crosslinking agent, respectively. After the template molecules were removed, a large number of tailor-made cavities for MP on the surface of silica submicroparticles were formed. The key procedure of the preparation was neutralization in the last washing step. Only when the polymer was neutral, would the most excellent adsorbability be presented. 3.3. Characterizations of SMIPMP. To ascertain the polymeration on the surface of silica submicroparticles, FT-IR spectra (Figure 3) were obtained for silica submicroparticles, SMIPMP and SNIP, respectively. The wide and strong adsorption band around 3438 cm−1 could be ascribed to stretching vibrations of O−H. The observed features around 1103 cm−1 indicated Si−O−Si antisymmetric stretching vibrations. The bands around 1635 cm−1 resulted from H−

Figure 1. Adsorption spectra of the MP in the presence of various concentration of APTES in methanol. Concentration of MP: 0.1 mmol L−1, corresponding pure APTES solutions as blanks.

It was clearly observed that the maximum absorbance at 208 and 256 nm of the mixture solutions was declining. The change in the adsorption spectra indicated the formation of stable host−guest complexes between print molecule MP and APTES in the mixture, and the complexes was formed probably through the hydrogen bonding between N−H of APTES and −OH or −COOCH3. Moreover, the interaction was getting stronger when the concentration of APTES increased further. However, too an large amount of functional monomer may lead to their own association and increase nonselective binding sites. So the optimal molar ratio of the template and functional monomer-1:4 was chosen for the synthesis of the SMIPMP. 3.2. Preparation of the SMIPMP. The possible preparation protocol of SMIPMP was shown in Figure 2. In this procedure,

Figure 3. FT-IR spectra of silica submicroparticles (a), SMIPMP (b), and SNIP (c). 14917

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

Figure 4. SEM of the submicrosized silica (a), SMIPMP (b), SNIP (c) and microsized silica (d), MMIPMP (e), and MNIP (f).

Table 1. Particle Sizes, Specific Surface Areas, Average Particle Size, and the Maximum Static Adsorption Capacities of Adsorbents samples SMIPMP SNIP silica submicroparticle MMIPMP MNIP silica microparticle

average particle size (nm)

specific surface area (m2 g−1)

average pore diameter (nm)

static adsorption capacity (mg g−1)a

adsorption equilibrium time (min)b

475 480 425

218 195 248

0.91 1.91 1.97

32.68 20.68 11.81

40 60

2620 2630 2575

68 55 94

0.92 1.92 1.93

6.483 1.385 0.615

120 140

a

The maximum static adsorption capacities of different adsorbents were carried out in the condition as follows: adsorbent = 20 mg, 9 mL of MP water−methanol solution, concentration = 350 mg L−1, 4 h, temperature = 25 °C. bThe adsorption equilibrium time was carried out in the condition as follows: adsorbent = 20 mg, 9 mL of MP water−methanol solution, concentration = 100 mg L−1, temperature = 25 °C.

were 218, 195, and 248 m2 g−1, respectively, while the specific surface areas of MMIPMP, MNIP, and the silica microparticles were 68, 55, and 94 m2 g−1. The specific surface area of SMIPMP and SNIP was higher than that of MMIPMP and MNIP, showing that the type of supports affected the specific surface area of adsorbents. The average pore diameters of SMIPMP and MMIPMP were much smaller than that of SNIP and MNIP, revealing that the removal of template molecules formed small pores on the surface of SMIPMP. On the basis of the different thermal stability between the silica submicroparticles and polymers, the content of imprinting layer grafted on the surface of submicroparticles was measured by TG analysis. As shown in Figure 5, the silica submicroparticles (curve a) had a slight weight loss of about 10.41% from room temperature to 800 °C, attributing to the loss of absorbed water. In addition, the weight loss trend of SMIPMP was similar to that of SNIP. The amount of entire coating was estimated as 11.82% of the total mass for SMIPMP and 10.77% of the total mass for SNIP, respectively. Such distinction was due to the existence of floccules-like imprinting coating on the surface of silica submicroparticles during the imprinting process. 3.4. Batch Adsorption Kinetics. The adsorption rate is an important parameter used to image the adsorption process. As

O−H vibrations. A characteristic feature of SMIPMP and SNIP compared with silica submicroparticles was the N−H band around 1506 cm−1 and the C−H band around 2933 cm−1. It suggested that APTES were grafted onto the surface of silica submicroparticles after imprinting. The surface morphologies of the silica submicroparticles, SMIPMP, SNIP, microsized silica, MMIPMP, and MNIP, were exemplified by the SEM in Figure 4. The silica submicroparticles and microsized silica were regularly spherical particles in shape with a particle size of ca. 425 nm and ca. 2575 nm, respectively. Compared with submicro- and micro-silica particles, the surface of particles (SMIPMP, SNIP, MMIPMP, and MNIP) after polymerization became loose and rough after imprinting polymerization, which would make the binding sites exposed at the surface of silica submicroparticles. Compared with imprinted polymers, there were less differences between the morphology of imprinted and nonimprinted polymers. Therefore, the distinct adsorption properties for imprinted and nonimprinted polymers could not entirely be attributed to the morphology difference but to the imprinting effect. The specific surface area and the average pore diameters of the silica submicroparticles, SMIPMP, SNIP, microsized silica, MMIPMP, and MNIP were shown in Table 1. The specific surface areas of SMIPMP, SNIP, and the silica submicroparticles 14918

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

than the SNIP. It might be ascribed to more cavities or adsorption sites obtained on the surface of SMIPMP which made the MP easily entering into the cavities or easily binding with the recognition sites. Additionally, it can be observed from Table 2 that the adsorption capacity of MP by both SMIPMP and SNIP at equilibrium increased with the increase of initial concentration, about 12.31 mg g−1 (100 mg L−1), 24.81 mg g−1 (300 mg L−1) for SMIPMP, and 7.839 mg g−1 (100 mg L−1), 14.14 mg g−1 (300 mg L−1) for SNIP. This was expected due to a greater driving force by a higher concentration gradient pressure.23 In order to further elucidate the uptake rate of adsorbate and potential rate-controlling step, the kinetics of MP adsorption on the SMIPMP or SNIP was determined with four kinetic models: the pseudo-first-order equation, the pseudo-second-order equation, Elovich equation, and intraparticle diffusion equation.18−21 These typical kinetic models, as described in following equations, were used for fitting the experiment data obtained from batch experiments

Figure 5. TG curves of (a) silica submicroparticles, (b) SMIPMP, and (c) SNIP.

shown in Table 1, the adsorption equilibrium time of SMIPMP, SNIP, MMIPMP, and MNIP were 40, 60, 120, and 140 min, respectively. This was a clear witness for the presence of higher adsorption kinetics for submicrosized adsorbents (both the SMIPMP and SNIP) due to the existence of more affinity binding sites on the surface of submicroparticles or in the proximity of the surface. Figure 6 showed the time strong dependence of the adsorption capacities of MP on SMIPMP and SNIP adsorbents.

pseudo‐first‐order equation: qt = qe(1 − exp( −k1t )) (5)

pseudo‐second‐order equation: qt = −1

k 2qe2t 1 + k 2qet

(6)

−1

where qe (mg g ) and qt (mg g ) are the adsorption capacity at equilibrium and any time t (min), respectively. k1 (min−1) and k2 (g·(mg−1·min−1)) are pseudo-first-order and pseudosecond-order rate constant of adsorption, respectively. simple Elovich kinetic equation: qt = A + Bln t

(7)

The parameters A and B are both kinetic constants, which can be determined by regression of the experimental data intraparticle diffusion equation: qt = K idt 1/2 + C −1

(8) −1/2

where Kid is the intraparticle diffusion rate (mg g min ), and C is a constant indicating boundary layer thickness. The constants of MP adsorption kinetics were fitted with the four models above under different initial concentrations by nonlinear regression were shown in Table 2. The nonlinear regression plots of the four models were shown in Figure 6. The applicability of the kinetic models to the adsorption behaviors was studied by judging the correlation coefficien (R2). It can be seen from Table 2 that all of R2 values (R2 > 0.99) of this adsorption process on the SMIPMP or SNIP by pseudo-second-order kinetic model are higher than that by other three models. It indicated that the adsorption kinetic process was perfectly following the pseudo-second-order kinetic model. Other three kinetic equations were not well-described adsorption of MP on the SMIPMP or SNIP, especially for the intraparticle diffusion model. It was because the stage of the adsorption of MP onto SMIPMP or SNIP was a multistep process (Supporting Information Figure S1), which typically consists of three types of mechanisms as follows: the first step might be due to the diffusion of MP through the aqueous phase to the external surface of adsorbent, the second stage was intraparticle diffusion, where the MP moved through the interior solid surface, and the third portion referred to the final equilibrium stage. These results were generally in agreement with other research results that the pseudo-second-order kinetic model was able to describe properly the kinetic process of the adsorption of MIP.22

Figure 6. The nonlinear regression of kinetic models and the effect of initial concentration for SMIPMP (a) and SNIP (b) at 298 K.

As seen in the figure, the MP adsorption of SMIPMP adsorbent increased with the time during the first 40 min and then the curve levels off as equilibrium was reached, while MP adsorption of SNIP adsorbent increased with the time during the first 60 min and the curve levels off as equilibrium was reached. The SMIPMP reached the adsorption equilibrium faster 14919

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

Table 2. Kinetic Parameters for the Adsorption of MP on SMIPMP and SNIP pseudo-first-order model adsorbents

C0 (mg L−1)

qe,exp (mg g−1)

qe,cal (mg g−1)

SMIPMP

100 300 100 300

12.31 25.10 7.839 14.14

4.595 7.972 2.859 6.444

SNIP

pseudo-second-order model

k1 (min−1)

0.02306 0.02296 0.02447 0.02327 Elovich model

R2

qe,cal (mg g−1)

0.8901 0.8752 0.8971 0.9052

12.89 25.94 8.311 14.67

k2 (g mg−1min−1)

8.471 × 10−3 5.928 × 10−3 11.88 × 10−3 8.409 × 10−3 intraparticle diffusion model

R2 0.9988 0.9991 0.9941 0.9993

adsorbents

C0 (mg L−1)

A

B

R2

Kid

C

R2

SMIPMP

100 300 100 300

2.407 4.826 1.178 2.157

2.090 4.329 1.434 2.482

0.8911 0.8770 0.7901 0.9231

0.6010 1.233 0.4003 0.7272

5.686 11.70 3.514 5.953

0.6871 0.6601 0.5614 0.7472

SNIP

3.5. Binding Isotherm. The adsorption capacity is an important factor to evaluate the imprinted polymer. As shown in Figure 7, the adsorption values increased with the increase of

where KF and 1/n are the Freundlich model constants related to the capacity and intensity of the adsorption, respectively. The Freundlich model is fit for that there are many types of sites acting simultaneously. The Freundlich equation can be linearized by taking logarithms, and constants can be determined. The Langmuir equation, which is valid for monolayer adsorption onto a surface with a finite number of identical sites, is given by eq 10 qe =

initial concentrations of MP. The maximum static adsorption capacities of adsorbents were given in Table 1. The maximum static adsorption capacities of SMIPMP, SNIP, and silica submicroparticle were 32.68, 20.48, and 11.81 mg g−1, respectively, while the maximum static adsorption capacities of MMIPMP, MNIP, and silica microparticle were 6.483, 1.385, and 0.615 mg g−1, respectively. The maximum static adsorption capacity of submicrosized adsorbents was much higher than that of microsized adsorbents. It suggested that adsorbents with small particle size and high specific surface area favored the increase in their adsorption capacity. As discussed before, more binding sites were probably exposed on the surface of SMIPMP than on silica submicroparticle, which made the adsorption capacity of SMIPMP with a relatively low specific surface was much stronger than that of submicrosized silica. In order to establish the relationship between the adsorption amount of MP and corresponding equilibrium concentration in the aqueous solution, three isotherm equations most commonly used, namely Freudlich, Langmuir, Redlich-Peterson,20,23,24 have been adopted in this work. The forms of these isotherms are presented by the following equations. The Freundlich model is an empirical equation based on adsorption on a heterogeneous surface. It is given as

qe = KFC

1 + KLCe

(10)

where KL represents the Langmuir constant associated with the free energy of adsorption. qm is the maximum adsorption capacity (mg g−1), and qe is the amount of adsorbed MP at equilibrium (mg g−1). qm and KL can be determined from the linear plot of Ce/qe versus Ce. The Redlich-Peterson model is often used to describe chemical and physical adsorption on heterogeneous surface, instead of assumption homogeneity such as equally available adsorption sites, monolayer surface coverage, and no interaction between adsorbed species. The equation for this model is

Figure 7. Comparison of Langmuir, Freundlich and Redlich-Peterson isotherm models for MP adsorption onto SMIPMP and SNIP using nonlinear regression.

1/ n

qmKLCe

qe =

KR Ce 1 + bCem

(11)

where KR and b are the Redlich constants. The value of m varies from 0 to 1. To fit the R-P isotherm equation to the experimental data set, the linear regression lines were constructed by plotting Ce/qe vs Cme of eq 12 with various values. Ce 1 b m = + Ce qe KR KR

(12)

In this study, first, the value of was calculated for each m value, and then the linear regression line associated with this specific value was plotted. In order to compare the validity of three isotherm equations, a normalized standard deviation Δqe (%) is calculated Cme

Δqe (%) =

∑ [(qe,exp − qe,cat)/qe,exp]2 N−1

(13)

where N is the number of data. Figure 8 showed the linear regression lines for the adsorption of MP on SMIPMP (a) and SNIP (b). The m values of the most

(9) 14920

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

Table 3. Adsorption Isotherm Constants for SMIPMP and SNIP adsorption isotherm models Freundlich model

Langmuir model

Redlich-Peterson model

parameters

SMIPMP

SNIP

R2 KF (mg g−1) n Δq % R2 KL (L mg−1) qm,c (mg g−1) Δq % R2 KR (L g−1) b (L mg−1) Δq %

0.9951 0.5825 1.383 4.916 0.9756 4.521 × 10−3 56.66 5.145 0.9889 0.3252 0.04379 3.143

0.9924 0.2721 1.317 5.896 0.9691 3.342 × 10−3 38.60 3.829 0.9792 0.1447 0.01596 3.177

Figure 8. Adsorption of MP on SMIPMP (a) and SNIP (b) fitted with new linear form of Redlich-Peterson isotherm equation. Figure 9. Dynamic adsorption curve of SMIPMP and SNIP contacting 100 mg L−1 MP at a rate of 1 mL min−1 at 25 °C.

suitable linear regression lines are 0.66 and 0.76. The experiment data fitted with Langmuir, Freundlich, and Redlich-Peterson isotherm equations were shown in Figure 7. The adsorption capacity of MP on SMIPMP and SNIP considerably increased with the MP equilibrium concentration increasing from 0 to 286.4 and 0 to 313.1 mg L−1, respectively. But SMIPMP showed a higher adsorption capacity than that of SNIP according to the results of the MP adsorption isotherm experiments. This phenomenon may be attributed to the chemical structure and size of MP complementary to cavities in the polymeric matrix. Moreover, it was believed that the hydrophobic interactions may play an important role in the adsorption capacity of MP. As shown in Table 3, the experimental data were well fitted by the three models (Freundlich, Langmuir, and RedlichPeterson models) with R2 values: 0.9951, 0.9756, 0.9889; 0.9924, 0.9691, and 0.9792 for SMIPMP and SNIP, respectively. Clearly, the correlation coefficient is higher with the Freundlich model than with the Langmuir and Redlich-Peterson models. However, Table 3 shows that Δqe (%) values of MP adsorption on SMIPMP and SNIP are 3.143% and 3.177% for R-P isotherm equations, respectively, which are less than that of the Freundlich and Langmuir isotherm equations. Apparently, the three-parameter model provided better fitting for SMIPMP and SNIP separately. 3.6. Dynamics Binding Curves. In order to further study the binding characteristic of SMIPMP for MP, the adsorption experiments in the column method were also performed. Figure 9 displayed the dynamic adsorption curves of SMIPMP and SNIP for MP, respectively. It can be observed that the dynamic binding curve of SMIPMP for MP was obviously different from that of SNIP. The leaking volume of MP on SMIPMP was 7 mL,

which is greater than that on SNIP (about 3 mL). By calculating, the saturated adsorption amount of MP on SMIPMP was 10.45 mg g−1 higher than that on SNIP of only 6.266 mg g−1. It was further demonstrated that SMIPMP had a higher binding capacity for MP than SNIP. The trend was in accordance with batch results for MP adsorption on SMIPMP and SNIP. Additionally, the dynamic saturated adsorption amount for both SMIP MP and SNIP was lower than that (12.28 mg g−1 for SMIP MP and 7.743 mg g−1 for SNIP) of the static absorption when the initial concentration was 100 mg L−1. On the one hand, it might be due to that MP did not have enough time to contact with SMIPMP and SNIP during the dynamic adsorption process. On the other hand, the presence of the column dead volumes may influence the adsorption capacity. Generally speaking, the adsorption amount of MP adsorbed by the dynamic method was close to that in the batch method, indicating the prepared SMIPMP could be efficiently used as fixed bed adsorption material for the dynamic removal of MP in the wastewater. The Thomas model is one of the most widely used column performance theory which predicts the relationship between concentration and time.25,26 The expression by Thomas for an adsorption column is given as follows C = C0 1 + exp

(

1 KTq0M F

− KTC0t

)

(14)

where C0 is the influent MP concentration, C is the effluent concentration at time t, KT is the Thomas rate constant, q0 is 14921

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

the adsorption capacity of the adsorbent per unit mass of the adsorbent, M is the mass of adsorbent, and F is the flow rate. The linearized form of the Thomas model is as follows: ⎛C ⎞ KTq0M ln⎜ 0 − 1⎟ = − KTC0t ⎝C ⎠ F

(15)

The values of KT and q0 can be determined from a plot of ln(C0/C−1) against t at a given flow rate using linear leastsquares regressive analysis or from a plot of C/C0 against t using nonlinear regression analysis. Figure 10 showed the experimental points, linear predicted points, and nonlinear predicted points according to the Figure 11. Adsorption selectivity of MP onto the SMIPMP and SNIP adsorbents in the presence of competitive adsorbates.

adsorption capacity for them. The possible reason of SMIPMP recognizing its template molecule was due to the existence of memory cavities forming during the process of polymerization, which were perfectly complementary both in shape and functional groups with MP. Table 5 also presented the Kd and α values of SMIPMP higher than those of SNIP. The α values of SMIPMP for MP relative to Table 5. Parameters of Adsorption Selectivity of SMIPMP and SNIP SMIPMP

Figure 10. Comparison of experimental points, linear predicted points, and nonlinear predicted points with the Thomas model.

−1

Kd (L g ) MP MS p-HB

parameters of Thomas model in Table 4. The values of equilibrium uptake per gram of the adsorbent (q0, mg g−1) from experiment were also listed in Table 4. The results demonstrate that the values of the constant KT and q0 obtained by nonlinear regression and linear regression are not all consistent. In addition, the value of R2 from nonlinear regressive method was larger than that from the linear regressive method. Compared to the values of q0e and q0c, the difference of q0e from the experiment and q0c from nonlinear was smaller. From Figure 10, both linear methods and nonlinear methods are suitable for predicting the dynamic behavior of the column. Furthermore, the nonlinear regressive method was more effective in predicting the adsorption kinetics than the linear method. 3.7. Selectivity Binding. The selective binding character of SMIPMP was evaluated toward competitive substrates MP, MS, and p-HB. The adsorption experiments for each adsorbate were carried out under the same conditions, and the adsorbates were all at the same concentration of 20 mg L−1. As shown in Figure 11, SMIPMP showed a highest adsorption capacity for MP among the three competitive substrates, indicating the high adsorption selectivity for MP. Moreover, a comparison of the adsorption of the SMIPMP and SNIP adsorbents for each substrate suggested that the SMIPMP hardly show an imprinting effect for MS and p-HB as both adsorbents nearly had the same

0.2800 0.04944 0.04569

SNIP αM 5.664 6.129

−1

α′

Kd (L g )

αN

4.652 4.057

0.2617 0.03378 0.02722

1.217 1.511

MS and p-HB were higher (5.664 and 6.129, respectively), the corresponding α of SNIP were much lower (1.217 and 1.511, respectively). In addition, the α′ values for MS and p-HB were 4.652 and 4.057, respectively, which was far more than 1.0, indicating the SMIPMP had higher adsorption selectivity than that of the SNIP. Therefore, it can be concluded that the SMIPMP had good adsorption selectivity in the presence of other competitive substrates. 3.8. Regeneration. After adsorption of MP onto the SMIPMP, the MP-adsorbed SMIPMP was regenerated using the CH3 OH/HCl mixture and then deionized water. The regenerated SMIPMP was used to adsorb MP in subsequent cycle. The adsorption capacity of the SMIPMP adsorbent for MP with five consecutive adsorption-regeneration cycles was shown in the Supporting Information Figure S2. It was clearly seen that SMIPMP could be effectively regenerated for further use with only about 10.70% loss of initial binding capacity after five cycles. It is reasonable to assume that the SMIPMP can be reused at least five times without decreasing their adsorption capacities significantly.

Table 4. Model Parameters by Linear Regression Analysis and Nonlinear Regression Analysis with the Thomas Model for Adsorption of MP onto SMIPMP and SNIP linear regression analysis adsorbent SMIPMP SNIP

−1

q0e (mg g ) 10.45 6.266

−1

−1

KT (L min mg ) −3

7.406 × 10 7.552 × 10−3

−1

q0c (mg g ) 9.468 7.496

nonlinear regression analysis R

2

0.9681 0.9692 14922

−1

KT (L min mg−1) −3

8.939 × 10 9.156 × 10−3

q0c (mg g−1)

R2

11.11 6.663

0.9994 0.9982

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

(4) Lokhnauth, J. K.; Snow, N. H. Determination of parabens in pharmaceutical formulations by solid-phase microextraction-ion mobility spectrometry. Anal. Chem. 2005, 77, 5938−5946. (5) Deng, D. L.; Zhang, J. Y.; Chen, C.; Hou, X. L.; Su, Y. Y.; Wu, L. Monolithic molecular imprinted polymer fiber for recognition and solid phase microextraction of ephedrine and pseudoephedrine in biological samples prior to capillary electrophoresis analysis. J. Chromatogr., A 2012, 1219, 195−200. (6) Jin, Y.; Jiang, M.; Shi, Y.; Lin, Y.; Peng, K.; Dai, B. Narrowly dispersed molecularly imprinted microspheres prepared by a modified precipitation polymerization method. Anal. Chim. Acta 2008, 612, 105−113. (7) Feng, Q. Z.; Zhao, L. X.; Yan, W.; Lin, J. M.; Zheng, Z. X. Molecularly imprinted solid-phase extraction combined with high performance liquid chromatography for analysis of phenolic compounds from environmental water samples. J. Hazard. Mater. 2009, 167, 282−288. (8) Fang, L. J.; Chen, S. J.; Guo, X. Z.; Zhang, Y.; Zhang, H. Q. Azobenzene-containing molecularly imprinted polymer microspheres with photo- and thermoresponsive template binding properties in pure aqueous media by atom transfer radical polymerization. Langmuir 2012, 28, 9767−9777. (9) Li, Y.; Ding, M. J.; Wang, S.; Wang, R. Y.; Wu, X. L.; Wen, T. T.; Yuan, L. H.; Dai, P.; Lin, Y. H.; Zhou, X. M. Preparation of imprinted polymers at surface of magnetic submicroparticles for the selective extraction of tadalafil from medicines. ACS Appl. Mater. Interfaces 2011, 3, 3308−3315. (10) Srivastava, V.; Weng, C. H.; Singh, V. K.; Sharma, Y. C. Adsorption of nickel ions from aqueous solutions by nano alumina: kinetic, mass transfer, and equilibrium studies. J. Chem. Eng. Data 2011, 56, 1414−1422. (11) Yuan, J. J.; Zhou, S. X.; You, B.; Wu, L. M. Organic pigment particles coated with colloidal nano-silica particles via layer-by-layer assembly. Chem. Mater. 2005, 17, 3587−3594. (12) Boparai, H. K.; Joseph, M.; O’Carroll, D. M. Kinetics and thermodynamics of cadmium ion removal by adsorption onto nano zerovalent iron particles. J. Hazard. Mater. 2011, 186, 458−465. (13) Shakerian, F.; Dadfarnia, S.; Shabani, A. M. H. Synthesis and application of nano-pore size ion imprinted polymer for solid phase extraction and determination of zinc in different matrices. Food Chem. 2012, 134, 488−493. (14) Alizadeh, T. Application of electrochemical impedance spectroscopy and conventional rebinding experiments for the investigation of recognition characteristic of bulky and nano-sized imprinted polymers. Mater. Chem. Phys. 2012, 1−12. (15) Connor, N. A.; Paisner, D. A.; Huryn, D.; Shea, K. J. A plasmonic photocatalyst consisting of silver submicroparticles embedded in titanium dioxide. J. Am. Chem. Soc. 2008, 130, 1680. (16) Xie, C.; Liu, B.; Wang, Z.; Gao, D.; Guan, G.; Zhang, Z. Molecular imprinting at walls of silica submicrotubes for TNT recognition. Anal. Chem. 2008, 80, 437. (17) Stöber, W.; Finker, A.; Bohn, E. J. Controlled growth of monodisperse silica spheres in the micron size range. J. Colloid Interface Sci. 1968, 26, 62−69. (18) Xue, Y. G.; Houa, H. B.; Zhua, S. J. Adsorption removal of reactive dyes from aqueous solution by modified basic oxygen furnace slag: Isotherm and kinetic study. Chem. Eng. J. 2009, 147, 272−279. (19) Fu, Q. L.; Deng, Y. L.; Li, H. S.; Liu, J.; Hu, H. Q.; Chen, S. W.; Tongmin, S. Equilibrium, kinetic and thermodynamic studies on the adsorption of the toxins of Bacillus thuringiensis subsp. kurstaki by clay minerals. Appl. Surf. Sci. 2009, 255, 4551−4557. (20) Yi, X. S.; Shi, W. X.; Yu, S. L.; Wang, Y.; Sun, N.; Jin, L. M.; Wang, S. Isotherm and kinetic behavior of adsorption of anion polyacrylamide (APAM) from aqueous solution using two kinds of PVDF UF membranes. J. Hazard. Mater. 2011, 189, 95−501. (21) Alkan, M.; Demirbas, O.; Celikcapa, S.; Dogan, M. Sorption of acid red 57 from aqueous solution onto sepiolite. J. Hazard. Mater. 2004, 116, 135−145.

4. CONCLUSIONS In this work, a highly selective submicrosized imprinted polymer for methylparaben (SMIPMP) was synthesized by a surface imprinting technique. Compared with MMIP MP adsorbent, SMIPMP exhibited higher adsorption capacity and offered a faster kinetics for the rebinding of MP. The equilibrium adsorption isotherms of MP on MIP SMIPMP and SNIP can be well fitted by the Freundlich and RedlichPeterson model. In addition, the SMIPMP adsorbent had excellent selectivity toward structural analogues of MP. The pseudo-second-order rate equation characterized the kinetic curves well and intraparticle diffusion plots showed that more than one step affecting the adsorption process. During the dynamic experiment, the Thomas model was found suitable for the mathematical description of MP adsorption in the fixed-bed column, and the nonlinear regression was more effective. Moreover, the SMIPMP could be used at least five times without weakening the binding capacity significantly. All the abovementioned indicates the potential application of the MIP adsorbents for selective removal of MP in water or wastewater treatment. Further, it can be predicted to be exploited in concentration and purification of trace analytes in the complex matrix.



ASSOCIATED CONTENT

S Supporting Information *

Figures: The kinetic data were treated with the intraparticle diffusion model. (Figure S1) and the regeneration of SMIPMP (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the National Natural Science Foundation of China (No. 21077046, No. 21107037, No. 21176107, No. 21174057, No. 21004031), National key basic research development program (973 Program, No. 2012CBB21500), Ph.D. Programs Foundation of Ministry of Education of China (No. 20093227110015), and Natural Science Foundation of Jiangsu Province (BK2011461, SBK2011459, BK2011514).



REFERENCES

(1) Handa, O.; Kokura, S.; Adachi, S.; Takagi, T.; Naito, Y.; Tanigawa, V.; Yoshida, V.; Yoshikawa, V. Methylparaben potentiates UV-induced damage of skin keratinocytes. Toxicology 2006, 227, 62− 72. (2) Benedict, M. Q.; Hood-Nowotny, R. C.; Howell, P. I.; Wilkins, E. E. Methylparaben in Anopheles gambiae s.l. sugar meals increases longevity and malaria oocyst abundance but is not a preferred diet. J. Insect. Physiol. 2009, 55, 197−204. (3) Šatínský, D.; Huclová, J.; Ferreira, R. L. C.; Montenegro, M. C. B. S. M.; Solich, P. Determination of ambroxol hydrochloride, methylparaben and benzoic acid in pharmaceutical preparations based on sequential injection technique coupled with monolithic column. J. Pharm. Biomed. Anal. 2006, 40, 287−293. 14923

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924

Industrial & Engineering Chemistry Research

Article

(22) Pan, J. M.; Zou, X. H.; Wang, X.; Guan, W.; Yan, Y. S.; Han, J. Selective recognition of 2,4-dichlorophenol from aqueous solution by uniformly sized molecularly imprinted microspheres with β-cyclodextrin/ attapulgite composites as support. Chem. Eng. J. 2010, 162, 910−918. (23) Xue, Y. J.; Hou, H. B.; Zhu, S. J. Adsorption removal of reactive dyes from aqueous solution by modified basic oxygen furnace slag: Isotherm and kinetic study. Chem. Eng. J. 2009, 147, 272−279. (24) Albadarin, A. B.; Mangwandi, C.; Al-Muhtaseb, A. H.; Walke, G. M.; Allen, S. J.; Ahmad, M. N. M. Kinetic and thermodynamics of chromium ions adsorption onto low-cost dolomite adsorbent. Chem. Eng. J. 2012, 179, 193−202. (25) Thomas, H. C. Heterogeneous ion exchange in flowing system. J. Am. Chem. Soc. 1944, 66, 1664−1666. (26) Sotelo, J. L.; Ovejero, G.; Rodríguez, A.; Á lvarez, S.; García, J. Removal of atenolol and isoproturon in aqueous solutions by adsorption in a fixed-bed column. Ind. Eng. Chem. Res. 2012, 51, 5045−5055.

14924

dx.doi.org/10.1021/ie301890b | Ind. Eng. Chem. Res. 2012, 51, 14915−14924