Molecularly imprinted polymers as tailor-made adsorbent are used to recognize target molecules. They have a memory of the size, shape, and functionalities complementary to the template molecules . Some polymerization approaches such as bulk polymerization , suspension polymerization , mini-emulsion polymerization , and precipitation polymerization  have been developed to synthesize three-dimensional network polymers.
MIPs prepared by bulk polymerization are ground and sieved to obtain a desirable size of particles which may result to destroy the cavities, irregular shape, and reduction in yield of useful size .
On the other hand, methods such as suspension polymerization and mini-emulsion polymerization may face difficulties including prolonged optimization of the experimental procedure and existing remained emulsifier or stabilizer on the adsorbent . Thus, precipitation polymerization is preferred because of resulting in the spherical and uniform shape of particles, narrow size distribution, and a surfactant or stabilizer-free polymerization.
MIPs are prepared by the co-polymerization of functional monomers with cross-linkers around template molecules. Interaction of the functional monomers with templates forms a stable complex . After the polymerization process, the template molecules are removed, leading to well-defined and highly cross-linked three-dimensional cavities. The resulting MIPs can rebind the template molecules with high selectivity. They are stable, controllable, and resistant to varying temperatures, pH, and solvent . The nucleation and growth process of MIPs can be adjusted by factors such as functional monomer, porogen, cross-linker, template, and initiator. . Duo to significant advantages of using MIPs including ease and low cost of preparation, MIPs have wide applications such as protein recognition , solid-phase extraction [12, 13], sensors [14, 15], environmental , drug delivery systems  and antibody substitutes . Bisphenol A (BPA) have extensively used in industrial chemical, as a primary raw material and as an intermediate to the production of epoxy resins, polycarbonate plastic, food and drink containers, baby formula bottles, electronic apparatus and medical facilities [19-21]. BPA has a harmful effect on the environment and endocrine systems of humans and animals [22, 23]. Therefore, it is essential to remove BPA in various samples due to its toxic influence.
Several attempts have been made to synthesize MIP by various polymerization to remove BPA from food, water, and milk solution. Hiratsuka et al. (2013) synthesized a magnetic molecularly imprinted polymers (M-MIPs) for BPA detection in river water by a multi-step swelling and polymerization method. He showed that the binding experiments and Scatchard analyses revealed two classes of binding sites . In another study, Alexiadou et al. (2008) prepared MIP for BPA by two synthetic routes: semi-covalent and noncovalent methods. They evaluated the molecular imprinting effect using the polymers in HPLC and SPE. The most critical factors of fabricated MIP were the organic content in loading–washing medium and the washing volume. Moreover, low flow rates in the elution step enhanced extraction recovery .
Despite several studies undertaken on BPA removal by MIP, there is rare information on isotherm and kinetic investigation to find out the mechanism of the process. In the present work, the precipitation polymerization is selected for MIP synthesis due to the advantages mentioned earlier. The prepared polymers are characterized by Fourier transform infrared spectroscopy (FTIR) to determine the functional groups. The morphologies and pore size of the obtaining imprinted particles are characterized by scanning electron microscopy (SEM) and Brunauer–Emmett–Teller (BET) gas adsorption measurements, respectively. The effect of molecular ratios of template: functional monomer: cross-linker on adsorbent capacity, BPA removal efficiency, polymer structure, and pore size are evaluated. Furthermore, the mechanism and binding properties of polymers are studied.
Bisphenol A (BPA), ethylene glycol dimethacrylate (EGDMA), Methacrylic acid (MAA), 2, 2´- azobisisobutyronitrile (AIBN) were obtained from Sigma Aldrich (Steinheim, Germany). AIBN was purified by recrystallization from methanol before use. Acetonitrile, acetone, acetic acid, and methanol were HPLC grade, purchased from Merck (Darmstadt, Germany) and used without further purification.
The MIPS were synthesized in two molecular ratios of 1:4:20 and 1:6:30 (template: functional monomer: cross-linker) as follows: First, for the synthesis of 1:4:20 polymer, 0.182 g (0.798 mmol) of BPA (template) and 0.35 ml (3.192 mmol) of MAA (functional monomer) were dissolved in 15 ml of the porogen acetone. The solution was stirred for 10 min. Then, 3 ml (15.96 mmol) of EGDMA as cross-linker and 50 mg (0.304 mmol) of AIBN as initiator were added to the previous solution. The pre-polymerization solution was sonicated for 10 min at room temperature and purged with nitrogen for 15 min in ice-bath to remove dissolved oxygen. The reaction was performed at 60 °C in a water bath for 24 h to achieve a solid polymer. For the synthesis of 1:6:30 polymer, the amounts of functional monomer and cross-linker were 6 and 30 folds of template amount, respectively. Non-imprinted polymers (NIPs) were synthesized exactly by the similar procedure of MIPs without bisphenol A .
Template removal from MIPs
To remove the template, the prepared polymers were transferred into a flask containing methanol/acetic acid (9:1 V/V) and the solution was continuously stirred with a magnetic stirrer during the extraction. The extraction was continued until the absorbance of the filtered solution at 278 nm reached to zero. Then, the template -free MIPs have separated from the solution by centrifuge 10000 rpm, washed with distilled water, and dried at oven at 50 °C overnight. For NIPs, Soxhlet extraction was omitted.
Preparation of BPA calibration plot
To prepare the BPA calibration plot, different concentrations of BPA were made in the acetonitrile solvent. The absorption of the samples was measured at a wavelength of 278 nm and the standard absorption plot and the corresponding equation were obtained. The achieved calibration plot is shown in Fig. 1.
Characterization of MIPs and NIPs
The scanning electron microscope (SEM) (TESCAN, VEGALL, Czech) was used for the estimation of the shape and surface morphology of the polymers. Polymeric particles were sputter-coated with gold before the SEM measurement. Nitrogen adsorption-desorption measurements were performed based on adsorption or desorption of nitrogen on or from polymer surface at 77K using BELSORP measuring instruments (Bel, Belsorp-miniII, Japan). Before measurement, the polymers were heated at 120 °C for 2 h. Standard Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) were used for calculation of speciﬁc surface area, pore-volume, and average pore diameter. Fourier transform infrared spectra (4000 – 400 cm1) of MIPs and NIPs particles were recorded on a Bruker spectrometer (Perkin Elmer, USA).
Binding afﬁnity of the imprinted and non-imprinted polymers was evaluated using a static adsorption experiment by separately mixing of 30 mg of polymer particles with various concentrations of BPA (0.1-2.5 mmol/L) in acetonitrile (CAN). The solution was shaken at 250 rpm for 3h at room temperature. After binding, the polymer particles were separated by centrifugation at 15000 rpm for 30 min. The free concentrations of BPA were determined by absorption at 278 nm. The adsorption capacity of MIPs and NIPs were determined by Equation (1) :
where C0 (mmol/L) and Cf (mmol/L) are the initial and final concentrations of BPA, (L) is the volume of solution, m(g) is the mass of the polymer, Q (µmol/g) is the amount of BPA. The removal efficiency was also calculated according to the following equation:
where, is the equilibrium concentrations of BPA (mmol/L).
In dynamic adsorption experiments to study the reaction kinetics, 30 mg of MIP particles were mixed with 15 mL of acetonitrile solution with different concentrations of BPA (0.1-2.5 mmol/L). The solution was shaken at 250 rpm. The samples were taken from the solution at an interval of 30 min and the unbound BPA was measured by a UV-visible spectrophotometer at 278 nm.
The imprinting factor which shows the diagnostic characterization of MIPs and NIPs to the template molecules was determined according to Equation (3):
Experimental data were fitted to the Langmuir,
Freundlich, and Scatchard models for the deter-mination of the isotherm parameters. Langmuir adsorption model assumes that each fixed number of homogenous sites can only adsorb on the molecule of the samples . The Langmuir model can be applied as below:
where, Ce (mmol/L) is the equilibrium concentration of the BPA, (mmol/g) is the amount BPA per unit mass of adsorbent at equilibrium concentration, (mmol/g) is the maximum adsorption capacity, b is the adsorption equilibrium constant.
The non-homogenous and reversible adsorption of BPA on adsorbent can be described by Freundlich isotherm as Equation (5):
Where, n and (mmol/g) are Freundlich constants.
The Scatchard plot analysis is applied to obtain further knowledge on the afﬁnity of binding sites . The experimental data were analyzed using the Scatchard equation as below:
where, is the dissociation constant.
Most of the adsorption process is related to time. Kinetic models describe the adsorption rate of adsorbate and its dependency on time. Kinetic models of pseudo-first order and pseudo-second-order were used to investigate BPA adsorption onto synthesized polymers.
The pseudo-first-order equation is stated in the linear form as below:
where, qe and qt are the amount of BPA adsorbed (mg/g) on the adsorbent at the equilibrium and at time t, respectively, and K1 (1/min) is the rate of constant adsorption.
The (qe) and (K1) parameters can be calculated from the slope and intercept of the plot of Ln(qe-qt) versus time. The pseudo-second-order equation is expressed as follow:
where, k2 is the rate constant of pseudo-second-order equation (µmol/mg min).
RESULT AND DISCUSSION
Fig. 2 shows the surface morphology of MIPs and NIPs. Nanometer and spherical particles were achieved by precipitation polymerization. Since the addition of template to polymerization solution causes cavities formation in polymer network , the NIPs showed fairly regular and smooth surface rather than MIPs. Thus, the MIPs had a porous surface compared to NIPs due to the presence of BPA.
The FT-IR spectra of synthesized MIPs and NIPs are shown in Fig. 3. Similar characteristic peaks confirmed similarity in the structure of polymers. However, there are obvious differences between the IR spectra of the MIPs and NIPs. Absorption peaks at 3400 to 3500 cm-1 is related to stretching vibration of O-H. Bonds at 1730 cm-1 are linked to the stretching vibration of C=O of –COOH group of MAA. Furthermore, absorption peaks at 1380 to 1400 cm-1 and 1460 cm-1 are related to the bending vibration of CH3 and CH2 groups, respectively. The peak at 1640 cm-1 corresponded to the stretching vibration of C=C bonds.
The porosities of produced MIPs and NIPs were evaluated by the nitrogen adsorption-desorption experiment (Fig. 4). The results of the BET analysis are shown in Table 1. Pore volume, specific surface area, and average pore diameter of the MIPs are compared with the NIPs. Since the pore diameter of synthesized polymers is in the range of 2 to 50 nm, they are placed in the mesoporous category. The specific surface area of MIPs is larger than that of the corresponding NIPs, which may result due to the presence of cavities on MIPs. This may be owing to the presence template molecule, such that after its removal from the polymer, it left particles of the MIP with a higher surface area. It indicates the higher accessibility of imprinted cavities and so higher adsorption capacity of MIPs to BPA than that of the corresponding NIP due to imprinting effect. Moreover, a decrease in particle size increased their specific surface area and pore volume. In Fig. 4, at low relative pressure (P/P0), the amount of adsorption increased with a uniform gradient which is related to the adsorption of nitrogen molecules on the internal surface of the mesoporous polymer. As this ratio of (P/P0) increased, the adsorption increased rapidly due to the filling of mesoporous polymer with gas molecules and their density on the surface.
The adsorption capacity of MIPs and NIPs and isotherm models
Fig. 5 shows the effect of various initial concentrations of BPA (0.1 to 2.5 mmol/L) onto the adsorption capacity and removal efficiency of MIPs and NIPs. The adsorption capacity (amount of BPA adsorbed per unit mass of polymers) increased with an increase in the initial concentration of BPA. At low concentrations, most of the active sites remained unsaturated and the binding capacity was low. However, an increase in the initial concentration of BPA resulted in mass transfer enhancement and adsorption capacity. As shown in Fig. 5, MIPs polymer has a greater binding capacity than that of NIPs. The highest binding capacity was obtained for MIP-6 with a value of 183 µmol/g. Also, due to the reduction in active sites of adsorbent because of increasing in BPA concentration, the removal efficiency decreased.
Langmuir, Freundlich, and Scatchard isotherm models were used to evaluate the interaction between BPA molecules and synthesized polymers. Fig. 6 and Table 2 show the Langmuir isotherm plots and model constants, respectively. The results indicated that the Langmuir model is a suitable isotherm for interpreting the adsorption data obtained for adsorption of BPA onto polymers due to high correlation coefﬁcients. The maximum adsorption capacity achieved for MIP-6 (400.1 µmol/g) was higher than that obtained for MIP-4 (245.5 µmol/g).
Freundlich isotherm plot is shown in Fig. 7. The constants and correlation coefficient (R2) are illustrated in Table 2. Results show that adsorption data were in good agreement with Freundlich isotherm with high correlation coefficients. The n values obtained for MIPs and NIPs were between 1 to 10 indicating good surface adsorption and suitability of the adsorption of BPA onto the polymers. The constant value of b and KF obtained for MIPs is smaller than that of NIPs which showed that BPA had more affinity to synthesized MIPs compared to NIPs.
The Scatchard analysis curves of MIPs and NIPs are illustrated in Fig. 8. As shown in Fig. 8, the nonlinear relationship between Qe/Ce and Qe was achieved. It indicated that the interaction sites between the template molecules and the functional monomers were not uniform during the synthesis of the MIPs for both high (left portion of the Fig.8) and low (right portion of the Fig.8)-affinity binding sites. Besides, all template molecule bonds did not participate in the polymerization reaction. The obtained result for Ka and Qmax is shown in Table 3. For NIP-6 and NIP-4, the Ka was calculated to be 357.14 and 416.66 µmol/L, respectively and the corresponding value of Qmax was 70.36 and 67.165 µmol/g, respectively.
Figs. 9 and 10 show the pseudo-first-order and pseudo-second-order plots for BPA adsorption onto polymers, respectively. Kinetic parameters are presented in Table 4. As seen in Table 4, the higher regression coefﬁcients (R2) were obtained for data fitting to the ﬁrst-order kinetic model for NIPs and MIPs. Besides, the calculated adsorption capacity (qeq,cal) achieved from the pseudo-first-order model agreed with the experimental adsorption capacity (qe,exp). Thus, the pseudo-first-order model could describe the adsorption of BPA onto synthesized NIPs and MIPs well. On the other hand, the binding kinetics obtained for MIPs are improved because of higher surface-area-to-volume ratios and more accessibility of imprinted cavities by BPA.
The imprinting factors for synthesized polymers with two molecular ratios of 1:4:20 and 1:6:30 polymers are presented in Table 5. The imprinting factor at high BPA concentrations increased in contrast to low BPA concentrations.
In this work, nano-sized MIPs were synthesized by precipitation polymerization, allowing BPA removal from water solution. The produced nanoparticles have a high surface-area-to-volume ratio; consequently, binding performance is suitable due to easier access of imprinted pores by the template. The effect of various factors such as the amount of the molecular ratio of template: functional monomer: cross-linker and initial concentrations of BPA onto adsorption capacity and removal efficiency of MIPs and NIPs were evaluated. The 1:6:30 ratio showed better removal of BPA rather than a 1:4:20 ratio. Moreover, the results indicated the nano-spherical morphology of MIPs. The adsorption isotherm and kinetic models were used to determine the mechanism and binding properties of polymers. The adsorption kinetics were in good agreement with the pseudo-first-order model. The results depicted that the synthesized MIP could be considered as an appropriate adsorbent for BPA removal.
The authors are grateful to the Islamic Azad University of Amol Branch (Amol, Iran) for its support to do this research.
CONFLICT OF INTEREST
There are no conflicts to declare.