Fabrication of submicrosized imprinted spheres attached polypropylene membrane using “ two-dimensional ” molecular imprinting method for targeted separation

Abstract

The submicrosized imprinted polypropylene microfiltration membrane (SIPM) for salicylic acid (SA) was prepared by water-in-oil emulsion polymerization method and two-dimensional imprinting method based on polypropylene microfiltration membrane with salicylic acid as template molecule, 4-vinyl pyridine as functional monomer, Span 80 as emulsifier, toluene as the oil phase, ethyleneglycol dimethacrylate as cross-linking agent, and 2,2′-azobis (2-methylpropionitrile) as initiator. The structure, morphology, and surface wettability of SIPM were characterized by Raman spectra, contact angle, and scanning electron microscopy. Flux test shows that the submicrosized imprinted spheres on the membrane is beneficial for increasing the membrane flux. Static adsorption experiment indicates that SIPM has a selective rebinding for salicylic acid. Moreover, the adsorption amount of salicylic acid over SIPM increased with properly increasing the amount of imprinted spheres on the SIPM. Permeation experiment shows that transport mechanism for permeation of the salicylic acid and acetylsalicylic acid toward SIPM is accordance with the facilitated mechanism.

Keywords

Emulsion polymerization membrane separation two-dimensional molecular imprinting selective adsorption salicylic acid

Introduction

Aspirin (acetylsalicylic acid, ASA) is a well-known nonsteroidal anti-inflammatory drug (Togo et al., 2009). Recently, it has also been used for more advanced life-saving perspective related to atherosclerotic cardiovascular disease (CVD) such as myocardial infarction (MI) and occlusive stroke (Kurban and Mehmetoglu, 2010). Generally, salicylic acid (SA) is the starting material for the synthesis of ASA. Aspirin can be easily hydrolyzed to SA during storage, leading to that excess concentration of SA is often detected in aspirin, which is harmful to human being. However, SA is also a valuable and useful chemical, which can be used as intermediates in synthesis of pharmaceutical products (Carlos et al., 1997). Therefore, it is necessary to find a way for selective separation and recovery of SA from ASA.

Recently, molecularly imprinted membranes (MIM) with specific recognition sites for target molecules have been considered as an alternative membrane for improving the selective separation of individual substances (Ren et al., 2014). Thus, it is an effective method for selective separation and recovery of SA from ASA by using MIM with SA as the target molecules. Two main routes have been used to synthesize the MIM, the “traditional” in situ crosslinking polymerization and the “alternative” polymer solution phase inversion, both in the presence of templates (Chen et al., 2010; Wang et al., 2004). The crosslinking copolymerization usually can yield a MIM with a “poor mechanical stability,” obviously due to the swollen structure. The phase inversion technique is usually used for the synthesis of a number of molecularly imprinted polymer membranes (Donato et al., 2011, 2013). Donato et al. (2010) and He et al. (2015) investigated several membranes prepared by using phase inversion technique for selective separation. However, all simultaneous preparations share the same major problem that molecularly imprinted polymers (MIPs) sites and membrane morphology are formed in the same step. Therefore, surface imprinting on synthetic commercial membranes may address this limitation by requiring less imprint molecules, increasing site capacity and speed of the template molecules for the imprinting sites.

In order to prepare MIM by surface imprinting method, it is necessary to first synthesize the MIPs with specificity of target molecule in the structure (Meng et al., 2013a, 2013b; Wulff, 1995, 2002). Two imprinting methods are mainly employed for the synthesis of MIPs. The first method, “three-dimensional molecular imprinting,” involves a template molecule along with a functional monomer, a crosslinking agent, a radical initiator, and an organic solvent by bulky polymerization (Du et al., 2014). Due to the thick polymeric network, there are several shortcomings of this approach, including long preparation times, mechanical deformation of the binding sites during grinding of bulk polymers, and time-consuming sieving procedure necessary for isolation of the fraction with a narrow size distribution associated with high material loses (Huy et al., 2014; Yan et al., 2013). The second method, “two-dimensional molecular imprinting,” involves the creation of recognition cavities on the surface of supports such as gold substrates, glass, and molecularly smooth mica, which can create functional monolayers assembled on surfaces as active interfaces (Nie et al., 2011; Xu et al., 2014). As compared to the “three-dimensional molecular imprinting,” the “two-dimensional molecular imprinting” is a more promising imprinting approach for developing new MIM for selective separation.

In our present work, submicrosized imprinted polypropylene membrane (SIPM) were synthesized by a novel two-dimensional method for introducing recognition sites onto polypropylene (PP) microfiltration membrane via emulsion polymerization. This approach allowed accelerating the removal of SA because templates on the obtained thin polymeric layers could be eluted much faster due to the shortened diffusion length. Another aim of this research was to further optimize polymer attached on the surface of membrane for the optimal separation effect. Following the results, adsorption, permeation experiment, and the mechanism were discussed.

Experimental

Materials

PP microﬁltration membranes with a nominal pore size of 0.25 µm, a thickness of 125 µm, and a mass of 55 mg per piece were purchased from wiseste (Zhenjiang) Co., Ltd. SA, ASA, 4-vinylpyridine (4-VP), Span 80, oleicacid, and methylbenzene were all supplied by Sinopharm Chemical Reagent (Shanghai, China). Ethyleneglycoldimethacrylate (EGDMA) was purchased from Shanghai Xingtu Chemical, Co. Ltd. (Shanghai, China). 2,2′-Azobisisobutyronitrile (AIBN) (Shanghai No.4 Reagent & H.V., Shanghai, China) was recrystallized from methanol prior to use. Ultrapure water used throughout the experiments was obtained from laboratory puriﬁcation system.

Preparation of SIPM

Firstly, five pieces of PP membranes were immersed into acetone (10 mL) to clean the membrane and dried under nitrogen condition. Then, the membranes were transferred to Petri dishes for soaking by 3.0 mL of oleic acid.

Secondly, 1.0 mmol of 4-VP, 5.0 mmol of EGDMA, and varied amount of span 80 (0.23, 0.46, and 0.69 g) were mixed with 60 mL of methylbenzene. Then, 10 mL of water containing 0.1 mmol of SA was added into the oil phase above forming water in oil emulsion by ultrasound 15 min. Thereafter, AIBN (0.01 g) were added into the mixture.

Then, the PP membranes pre-soaked with oleicacid were added into the water-in-oil emulsion above. The mixture was sonicated, and degassed with nitrogen for 10 min to eliminate oxygen, and then was sealed for polymerization at 60℃ for 12 h. The polymeric membranes were taken out and washed with mixed solvents of methanol and acetic acid (9:1, V/V) to remove SA, emulsifier, and unreacted monomers. Finally, the SIPM was obtained by filtrating and drying. The submicrosized nonimprinted polypropylene membrane (SNPM) was prepared in the absence of the template and subjected to the same procedure. After polymerization and washing, the mass of SIPM and SNPM increased from about 55–58, 63, and 67 mg, with the span 80 amount of 0.23, 0.46, and 0.69 g, respectively. The mass of SIPM were close to that of SNPM, meaning that imprinting did not change the mass but the structure of the as-prepared membranes.

Characterization

Raman spectra (DXR) was recorded in the range of 300–3500 cm⁻¹ at ambient temperature using a WITEC Spectra Pro 2300I spectrometer equipped with an Ar-ion laser, which provided a laser beam of 532 nm wavelength. The surface morphologies and cross-sectional structures of the samples were obtained by scanning electron microscopy (SEM, S-4800).

Batch rebinding assay

In the static binding study, one piece of varied SIPM and SNPM were added into 10 mL of SA water solution with various concentrations range from 20 to 200 mg L⁻¹. After a while standing, these solutions were centrifuged after reaching binding equilibrium, and the residual concentrations of SA in the conical flask were determined with UV spectrophotometry at a wavelength of 302 nm, respectively.

The equilibrium binding amounts (q_e, mg g⁻¹) were calculated by the following equation

q_{e} = \frac{(C_{0} - C_{e}) V}{W}

(1)

where C₀ (mg L⁻¹) and C_e (mg L⁻¹) are the initial and equilibrium concentrations of SA, respectively. V (mL) and W (g) are the volume of the solution and the weight of the SIPM or SNPM, respectively.

Selective recognition experiments

To evaluate the selective properties of different SIPM for SA, SA and ASA were selected for selective adsorption experiments. One piece of the membrane (SIPM and SNPM) was added into a 50 mL conical flask, each of which containing 10 mL the coexisting compound solution with 40 mg L⁻¹ of SA and ASA. After adsorption for 4 h, the concentrations of the residual SA and ASA can be determined by UV spectrophotometry, and the corresponding equilibrium binding amounts and the selectivity coefficient (α) were calculated as the procedure of static adsorption studies.

The distribution coefﬁcients (K_d) and selectivity coefficients (α) of ASA with respect to SA can be obtained according to equations (2) and (3)

K_{d} = \frac{q_{e}}{C_{e}}

(2)

K_d (L g⁻¹) represents the distribution coefficient, q_e (mg g⁻¹) and C_e (mg L⁻¹) are the equilibrium binding amount and the equilibrium concentration of the SA, respectively. The selectivity coefficient α can be obtained according to the following equation

α = \frac{K_{d (SA)}}{K_{dj}}

(3)

where K_dj represents the distribution coefficient of competition species (ASA).

Permeation experiments

The membrane, with an effective area of 1.5 cm², was placed tightly between two permeation chambers of a permeation cell. The feed solution of SA and ASA of 50 mg L⁻¹ in mixed methanol aqueous solution (7:3, v/v) was placed in the left-hand chamber, while the same volume of methanol aqueous solution was placed in the right-hand side. Under the condition of magnetic stirring, the concentrations of substrates through the membrane were determined by UV spectrophotometry. The flux J (mg cm⁻² s⁻¹), permeability coefficient P (cm² s⁻¹) and separation factor α_ASA/SA can be calculated by the following equations

J_{i} = \frac{Δ C_{i} V}{Δ tA} i = ASA, SA

(4)

P = \frac{J_{i} d}{(C_{F 1} - C_{Ri})} i = ASA, SA

(5)

α_{ASA / SA} = \frac{P_{ASA}}{P_{SA}}

(6)

where ΔC_i/Δt is the concentration change rate of the receiving solution which can be obtained by calculating the slope value of the diffused analyte concentration–time curve, V is the volume of solution, A is the effective membrane area (cm⁻²), d is the membrane thickness, and (C_Fi − C_Ri) is the concentration difference between feeding and receiving chambers (mg L⁻¹).

Results and discussion

Characterization

Raman spectroscopy

In order to ascertain the emulsion polymerization on the surface of PP microﬁltration membrane, Raman spectra for blank PP microfiltration membrane and SIPM3 was shown in Figure 1. As shown in the spectra, the peaks between 2800 and 3000 cm⁻¹ could be ascribed to stretching vibrations of C–H of PP membrane. Compared with the blank PP membrane, the characteristic peak around 3050 cm⁻¹ represented the vibrations of Ar–H of SIPM3. In addition, the peak of 1725 cm⁻¹ could be attributed to the C = O stretching vibration. Therefore, the results revealed that the PP imprinted membranes were successfully fabricated.

Figure 1.

Raman spectrum of polypropylene membrane (PP), salicylic acid imprinted polypropylene membrane (SIPM3).

SEM analysis

To capture the surface microscopic structures of the membrane, SEM was employed after sprayed gold. The SEM images of PP, SIPM1, SIPM2, and SIPM3 were obtained with different magnifications. Among them, SIPM1, SIPM2, and SIPM3 were prepared similarly but only with different amount of span 80 of 0.23, 0.46, and 0.69 g, respectively. As seen in Figure 2(a1) and (a2), PP exhibited a smooth surface and fibrous structure. The surfaces of SIPM1, SIPM2, and SIPM3 were covered with molecularly imprinted microspheres at different densities, while the three types of membranes still maintained a high membrane porosity. Figure 2(b), (c) and (d) were the images of SIPM surface morphology with different amounts of surfactants for SIPM1, SIPM2, and SIPM3, respectively. Comparing with the increasing amount of surfactant Span 80, it can be clearly found that the molecularly imprinted microspheres on the surface of PP membrane have relatively high density, suggesting that the amount of surfactant affected the surface morphology of compound MIMs directly.

Figure 2.

Scanning electron micrograph images of polypropylene membrane (a1 × 5000, a2 × 10000); SIPM1(b1 × 5000, b2 × 10000), SIPM2 (c1 × 5000, c2 × 10000), SIPM3 (d1 × 5000, d2 × 10000).

Contact angle analysis

The surface hydrophilicity was evaluated by measuring static water contact angle and Figure 3 showed the images of the water droplets in the PP membrane, SIPM1, SIPM2, and SIPM3. The value of contact angle of PP was significantly higher than 90°. However, the values of contact angles of SIPM1, SIPM2, and SIPM3 were lower than that of supporting membrane (PP). Moreover, the values of contact angles of SIPM2 and SIPM3 were clearly less than 90°, suggesting that the hydrophilicity of SIPM was higher than that of PP. It means that the existence of the imprinted microspheres on the surface of membrane changed the surface morphology of membranes and wettability.

Figure 3.

Contact angle images of polypropylene membrane (PP), SIPM1, SIPM2, and SIPM3.

Figure 4 showed the dynamic contact angle changes of PP membrane, SIPM1, SIPM2, and SIPM3. The average value of PP membrane was 115°, while the values of contact angles of SIPM1, SIPM2, and SIPM3 were obviously lower than that of PP membrane. The values of SIPM2 and SIPM3 were less than 80°, showing the fine hydrophilicity. Furthermore, the changes of contact angles of different membranes remained nearly unchanged during the entire test time, proving that the PP membrane was fixed with a layer of hydrophilic molecularly imprinted microspheres and the polymer combined with the membrane was stable.

Figure 4.

Contact angle changes of polypropylene membrane (PP), SIPM1, SIPM2, and SIPM3.

Membrane flux experiments

The relationship between operating time and the flux of methanol aqueous solution through different membranes for SA was shown in Figure 5. With the increase of operating time, the flux of blank PP membrane declined slightly. While the change of the fluxes of other imprinted membranes was stable, suggesting that the imprinted microspheres were steadily bond on the surface of PP. In addition, it can be seen that the membrane flux for SA of various membrane was in an order of PP membrane < SIPM1 < SIPM3 < SIPM2. The fluxes of SIPM1, SIPM2, and SIPM3 were higher than that of PP membrane, which maybe attribute to a lot polar group improving the hydrophilicity and facilitating transferring of SA. Moreover, different imprinted membranes also had diverse fluxes, suggesting that the amount of imprinted polymer affect membrane flux. According to the above order of flux, the increasing amount of imprinted polymer, to some extent, was beneficial to increase membrane flux. However, the excess amount of polymer might cause the partial jam of channels in original membrane which led to the decreasing porosity factor and was not conducive to increase the membrane flux.

Figure 5.

The ﬂux of SA methanol aqueous solution through polypropylene membrane (PP), SIPM1, SIPM2, and SIPM3.

Batch adsorption studies of membrane

Figure 6 presents the adsorption kinetics curves of diverse SIPM for SA. The binding amount of SIPM1, SIPM2, and SIPM3 for SA, as the figure shown, was straightly increasing in 5 min and quickly reaching to adsorption equilibrium. Compared with the corresponding nonimprinted membrane, the imprinted membrane had higher adsorption rate and binding amount for SA, which might be attributed to the higher binding sites on the surface of imprinted membrane for SA. The binding amount of SIPM1, SIPM2, and SIPM3 were 2.119, 2.780, and 2.990 mg L⁻¹, respectively. It means that the binding sites increased with the increasing amount of the imprinted polymeric microspheres. The different adsorption capacity of SIPM suggested that the amount of surfactant affected the adsorption behavior of imprinted membranes for the SA.

Figure 6.

The adsorption kinetics of different imprinted membranes at 298 K: (a) SIPM1 and SNPM1, (b) SIPM2 and SNPM2, and (c) SIPM3 and SNPM3.

As shown in Figure 7, adsorption isotherm of diverse imprinted membranes was determined in different concentrations of SA methanol solution ranging from 20 to 200 mg L⁻¹. The binding amount of SIPM1, SIPM2, and SIPM3 for SA increased with the increasing concentration of SA. In contrast, the adsorption amount of nonimprinted membrane for SA was lower than that of imprinted membrane. As discussed above, there were more binding sites on the surface of SIPM than that of SNPM. Moreover, it was obvious to find that the adsorption capacity of SIPM3 was higher than that of SIPM1 and SIPM2, showing that the increasing amount of molecularly imprinted polymeric microspheres was favor to the adsorption of SA.

Figure 7.

Comparison of the binding isotherms of different imprinted membranes at 298 K: (a) SIPM1 and SNPM1, (b) SIPM2 and SNPM2, and (c) SIPM3 and SNPM3.

Selectivity binding

In order to study the specific adsorption performance of SIPM1, SIPM2, and SIPM3, SA and ASA were selected to perform the binary selective adsorption at the concentration of 60 mg L⁻¹. Table 1 showed the selective adsorption parameters of different membranes for SA (or ASA).

Table 1.

Parameters of batch adsorption selectivity of different membranes.

Membrane	SA			ASA
Membrane	q_SA (mg/g)	K_dSA (L/g)	α	q_ASA (mg/g)	K_dASA (L/g)
PP	0.1771	0.0035	0.9167	0.1942	0.0038
SIPM1	1.3097	0.0275	2.0896	0.6566	0.0132
SNPM1	0.5645	0.0113	1.1708	0.4879	0.0097
SIPM2	2.0133	0.0443	3.4324	0.6435	0.0129
SNPM2	0.7666	0.0156	1.2666	0.6157	0.0123
SIPM3	2.0825	0.0463	2.5977	0.8749	0.0178
SNPM3	1.1325	0.0235	1.3354	0.8681	0.0176

In Table 1, it can be seen that the value of α of PP membrane was 0.9167, suggesting that the blank membrane had almost no selectivity for SA. In addition, the values of K_d and α of SIPM were higher than that of SNPM, indicating that SIPM had higher selectivity than that of nonimprinted membrane. Moreover, the values of α of SIPM1, SIPM2, and SIPM3 were 2.0896, 3.4324, and 2.5977, showing that the imprinted membranes had high selectivity. The value of α of SIPM2 was higher than that of SIPM1 and SIPM3, indicating that SIPM2 had higher adsorption selectivity.

Permeability

The selective permeability of SA

In permeation experiments, the concentrations of binary component feed solution of SA and ASA (50 mg L⁻¹) permeated through different imprinted membranes during 3.0 h were shown in Figure 8. The figures showed the time-dependent permeation curves of SIPM1, SIPM2, and SIPM3, respectively. Compared with blank PP membrane, the concentrations of SA of other three kinds of SIPM in receiving chamber were higher than that of ASA, showing that SIPM had the abilities of specific recognition and mass transfer performance for SA. Compared with varied SIPM, SNPM, and PP membrane among the membranes, it can be found that the diffusion concentration of SA through the SNPM was close to ASA. However, the mass transfer rate was lower than that of PP membrane, showing that the functional groups on the surface of three nonimprinted membranes had nonspecific adsorption of SA (or ASA). It reduced the mass transfer rate and could not reach the target of separation. Moreover, the change of concentration (SA and ASA) in received chamber through SIPM2 was higher, suggesting that the separation of SIPM2 for SA and ASA was well. Compared with diffusion rates, it can be also found that the diffusion rate of SIPM3 for SA was lower than that of SIPM1 and SIPM2. This might attribute to the excess molecularly imprinted polymeric microspheres on the surface of SIPM3 which affected the porosity, causing lower mass transfer rate. The consequences fitted the results of SEM above.

Figure 8.

Time–permeation curves of SA and ASA through different membranes (feed concentration: 50 mg/L).

Mass transfer performance parameters

The permeability coefficients of SA and ASA through different membranes were calculated by equation (5) as shown in Table 2 in the permeation experiments. As shown in the table, the blank PP membrane has similar permeability coefficients for both SA (4.864 × 10⁻⁴ cm² s⁻¹) and ASA (5.219 × 10⁻⁵ cm² s⁻¹), showing that the mass transfer behavior for SA (or ASA) was analogous. The difference of mass transfer coefficient of nonimprinted membrane for two substances was generally low. However, the value was lower than that of corresponding blank membrane, showing that the diffusion of SA (or ASA) through nonimprinted membrane was lower than that of PP membrane. In addition, the values of P of imprinted membranes (SIPM1, SIPM2, or SIPM3) for SA were higher than that of ASA, showing that molecularly imprinted membrane was advantageous to the transfer of SA. From the SEM images, it can be seen that the molecularly imprinted polymeric microspheres were “twist” uniform distribution on the surface (or interior) of membrane with lots of recognition sites. While the molecule structure of SA is smaller than that of ASA, SA could preferentially adsorbed on the surface of the membrane and transferred continuously from one recognition site to another site of the composite membrane, resulting the faster diffusion for SA molecules. Moreover, according to the separation factors in Table 2, SIMP2 has higher α_SA/ASA value than that of SIPM1 and SIPM3, showing a better selectivity of SIMP2.

Table 2.

Experimental permeability parameters for SA and ASA through the various imprinted membranes.

Membrane	substrate	J × 10⁻⁴ (mg/(cm²·s))	P × 10⁻⁵ (cm²/s)	α _SA/ASA	Mass transfer model
Membrane	substrate	J × 10⁻⁴ (mg/(cm²·s))	P × 10⁻⁵ (cm²/s)	α _SA/ASA	K_OV × 10⁻³ (cm/s)	R ²
PP	ASA	1.305	5.219	0.9320	5.846	0.9914
	SA	1.216	4.864		4.934	0.9960
SIPM1	ASA	0.7283	2.913	1.748	1.998	0.9933
	SA	1.273	5.092		5.521	0.9960
SNPM1	ASA	0.9183	3.673	1.090	2.782	0.9968
	SA	1.001	4.003		3.253	0.9885
SIPM2	ASA	0.4891	1.956	2.473	1.171	0.9872
	SA	1.210	4.839		4.720	0.9861
SNPM2	ASA	0.7093	2.837	1.143	1.907	0.9926
	SA	0.8107	3.243		2.312	0.9917
SIPM3	ASA	0.4722	1.889	1.891	1.115	0.9934
	SA	0.8930	3.572		2.612	0.9837
SNPM3	ASA	0.6270	2.508	1.040	1.590	0.9844
	SA	0.6523	2.609		1.719	0.9858

The following equation was used to compute the overall mass transfer coefﬁcient (K_OV) using concentration as the driving force (Dingemans et al., 2008)

(\frac{K_{OV} A}{V_{F}}) t = \frac{1}{1 + φ} ln (\frac{C_{F, 0}}{(1 + φ) C_{F} - φ C_{F, 0}})

(7)

where

φ = \frac{V_{F}}{V_{R}}

The right hand side of equation is plotted against time to estimate K_OV. This plot yields a linear curve with the slope (K_OVA/V_F) containing K_OV. The values of K_OV can be determined from a plot of C_F,0 against t using nonlinear regression analysis.

The value of overall mass transfer coefﬁcient (K_OV) obtained by equation (7) was shown in Table 2. High R² values (R²> 0.96) showing the fine fitting of the experimental permeation data for both imprinted and nonimprinted membrane. The nonlinear model provided fine fitting of SA and ASA for different membranes in Figure 9. From Table 2, the value of K_OV for SA is close to that of ASA toward blank PP membranes. However, the value of K_OV of SA was all much higher than that of ASA for SIPM1, SIPM2, and SIPM3. Moreover, the value of K_OV of SA for SIPM2 was higher than that for SIPM1 and SIPM3, indicating better separation effect of SIPM2 among the SIPMs. It also revealed that the amount of sphere on the membrane may play an important role in the permeation selectivity of SIPM. The inverse of the mass transfer coefficient is the mass transfer resistance. Compared to the SIPM1 and SIPM3, SIPM2 has the biggest mass transfer coefficient for SA, suggesting the SIPM2 has minimum resistance for transferring SA. Such transport for permeation of SA toward the SIPM was accordance with the “facilitated permeation” mechanism mentioned above (Fan et al., 2014; Meng et al., 2013 b).

Figure 9.

Nonlinear regression of mass transfer model for SA and ASA through different membranes.

Conclusions

This work highlighted the combination of membrane separation and “two-dimensional” molecular imprinting technology to prepare SIPM. The SIPM were proved to effectively separate SA from ASA. The separation effect greatly depended on the amount of imprinted spheres on the surface of SIPM. The imprinted spheres on the surfaces of membrane were beneficial to the increase in the membrane flux. This work provided an effective and novel way to remove SA and recover it from ASA, and provide a general strategy for the fabrication of highly selective membrane for puriﬁcation of trace analytes in the complex matrix.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by the National Natural Science Foundation of China (Nos. 21207051 and 21406085), PhD Programs Foundation of Ministry of Education of China (No. 20123227120015), Programs of Senior Talent Foundation of Jiangsu University (No. 15JDG024), Natural Science Fund for Colleges and Universities in Jiangsu Province (No. 15KJB550003), China Postdoctoral Science Foundation funded project (2015M581743), Postdoctoral Fund of Jiangsu Province (No. 1501104B), and Natural Science Foundation of Jiangsu Province (No. BK20150483).

References

Carlos

Hooshang

Christian

(1997) Separation of phenols from eucalyptus wood tar. Biomass and Bioenergy 13: 25–37.

Chen

Qin

Jia

(2010) Novel surface-modiﬁed molecularly imprinted membrane prepared with iniferter for permselective separation of lysozyme. Journal of Membrane Science 363: 212–220.

Donato

Greco

Drioli

(2011) Preparation of molecularly imprinted membranes and evaluation of their performance in the selective recognition of dimethoate. Desalination and Water Treatment 30: 171–177.

Donato

Tasselli

Drioli

(2010) Molecularly imprinted membranes with affinity properties for folic acid. Separation Science and Technology 45: 2273–2279.

Donato

Tasselli

Luca De

(2013) Novel hybrid molecularly imprinted membranes for targeted 4,4′-methylendianiline. Separation and Purification Technology 116: 184–191.

Dingemans

Dewulf

Braeckman

(2008) Mass transfer characteristics for VOC permeation through flat sheet porous and composite membranes: the impact of the different membrane layers on the overall membrane resistance. Journal of Membrane Science 332: 234–242.

Lei

Zhang

(2014) Determination of clenbuterol from pork samples using surface molecularly imprinted polymers as the selective sorbents for micro extraction in packed syringe. Journal of Pharmaceutical and Biomedical Analysis 91: 160–168.

Fan

Tian

(2014) A novel free-standing flexible molecularly imprinted membrane for selective separation of synephrine in methanol-water media. Journal of Membrane Science 467: 13–22.

Meng

Yan

(2015) Fabrication of new cellulose acetate blend imprinted membrane assisted with ionic liquid ([BMIM]Cl) for selective adsorption of salicylic acid from industrial wastewater. Separation and Purification Technology 145: 63–74.

10.

Huy

Seo

Zhang

(2014) Selective optosensing of clenbuterol and melamine using molecularly imprinted polymer-capped CdTe quantum dots. Biosensors and Bioelectronics 57: 310–316.

11.

Kurban

Mehmetoglu

(2010) Effects of acetylsalicylic acid on serum paraoxonase activity, Ox-LDL, coenzyme Q (10) and other oxidative stress markers in healthy volunteers. Clinical Chemistry 43: 287–290.

12.

Meng

Feng

Zhang

(2013a) Highly efficient adsorption of salicylic acid from aqueous solution by wollastonite-based imprinted adsorbent: A ﬁxed-bed column study. Chemical Engineering Journal 225: 331–339.

13.

Meng

Feng

Zhang

(2013b) Optimization of surface imprinted layer attached poly(vinylidene ﬂuoride) membrane for selective separation of salicylic acid from acetylsalicylic acid using central composite design. Chemical Engineering Journal 231: 132–145.

14.

Nie

Jiang

Zhang

(2011) Two-dimensional molecular imprinting approach for the electrochemical detection of trinitrotoluene. Sensors and Actuators B: Chemical 156: 43–49.

15.

Ren

Yang

(2014) The selective binding character of a molecular imprinted particle for bisphenol A from water. Water Research 50: 90–100.

16.

Togo

Suzuki

Yoshimaru

(2009) Aspirin and salicylates modulate IgE-mediated leukotriene secretion in mast cells through a dihydropyridine receptor-mediated Ca²⁺ influx. Clinical Immunology 131: 145–156.

17.

Wang

Xia

Sun

(2004) Molecularly imprinted copolymer membranes functionalized by phase inversion imprinting for uracil recognition and permselective binding. Journal of Chromatography B 804: 127–134.

18.

Wulff

(1995) Molecular imprinting in cross-linked materials with the aid of molecular templates-a way towards artificial antibodies. Angewandte Chemie International Edition in English 34: 1812–1832.

19.

Wulff

(2002) Enzyme-like catalysis by molecularly imprinted polymers. Chemical Reviews 102: 1–27.

20.

Deng

Tang

(2014) Preparation of 2D molecularly imprinted materials based on mesoporous silicas via click reaction. Journal of Materials Chemistry B 2: 8418–8426.

21.

Yan

Liu

Gao

(2013) Ionic liquids modified dummy molecularly imprinted microspheres as solid phase extraction materials for the determination of clenbuterol and clorprenaline in urine. Journal of Chromatography A 1294: 10–16.