Preparation of magnetic molecularly imprinted polymer beads and their recognition for baicalein

Abstract

Magnetite–β-cyclodextrin@poly(ethyleneglycol dimethacrylate-co-methacrylic acid) core–shell microspheres imprinted with baicalein (Fe₃O₄–β-CD@MIPs) were synthesized by ultrasonic assisted precipitation polymerization (UAPP) using methacrylic acid (MAA) as a functional monomer and EGDMA as a crosslinker in the presence of Fe₃O₄–β-CD nanoparticles as the template and baicalein (BAI) as the imprinting molecule. The structure, morphology and magnetic properties of the resultant Fe₃O₄–β-CD@MIPs have been systematically characterized by transmission electron microscopy (TEM), Fourier-transform infrared spectroscopy (FT-IR), vibrating sample magnetometry (VSM), and thermogravimetric analysis (TGA). The unique recognition ability of Fe₃O₄–β-CD@MIPs for BAI was evaluated by the adsorption experiments via comparing those of Fe₃O₄@MIPs and Fe₃O₄–β-CD@NIPs non-imprinted nanoparticles.

---

Introduction

Baicalein (BAI) is one of the main active ingredients of the dry root of a Chinese medicinal plant named Scutellaria baicalensis geogri (SBG).¹ BAI can be considered as one of the most promising chemo-preventive and effective agents for cancer treatment, which is mainly extracted from SBG.² Potential applications include the ability to induce HIV apoptosis,³ avoid liver poisoning⁴ and inhibit cell proliferation⁵ via in vitro studies based on its wide range of biological effects and low toxicity.

It is well-known that the pharmacologically active compounds in herbal plants have low concentrations and complex chemical constituents. It is essential to develop an effective technique for quantitative determination, recognition and purification of BAI molecules from those components in SBG. Several analytical methods, such as heat refluxing⁶ and adsorption chromatography separation,⁷ have been successfully utilized for the quantitative determination of BAI. However, these analytical procedures are always time-consuming with a high cost. A feasible way to solve these problems is via utilization of a solid adsorbent with specific recognition property, which has a great potential value for the study of biological process as well as its in-depth application. From these aspects, it is a superior choice to employ molecularly imprinted materials with specific adsorption.

Molecularly imprinted polymers (MIPs) have attracted much attention for their outstanding advantages, including unique adsorption performances, chemical resistance, mechanical stability, and relatively facile and low-cost preparation.⁸ As a result, MIPs have been widely used in many areas, such as adsorbents,⁹ membranes,¹⁰ solid-phase extraction,¹¹ biosensors,¹² isolation,¹³ chromatography stationary phase for efficient separation and specific recognition.¹⁴ The outstanding performance of MIPs depends on the shape memory towards the template via hydrogen bonding, and hydrophobic interactions. MIPs has been manufactured in various morphologies, including membrane, microspheres. Recently, core–shell magnetic MIPs have attracted much attention. It would be a promising solution to fully utilize the advantages of the magnetic NPs for efficient separation and enrichment, and recognition of the template molecule via a simple imposed magnetic field from a complex system. MIP coatings on magnetic Fe₃O₄ NPs with a core–shell structure (Fe₃O₄@BHb–MIPs) have been synthesized via combination of surface imprinting and sol–gel techniques for recognition and enrichment of proteins with accessible recognition sites, high capacity and favorable selectivity.¹⁵ Magnetic indole-3-acetic acid (IAA) imprinted polymer beads have been synthesized by suspension polymerization of 4-vinylpyridine (4-VPy) and a trimethylolpropane trimethacrylate (TRIM) crosslinker having an improved recognition ability to IAA with the aid of a silanized β-CD arm.¹⁶

In recent years, β-cyclodextrin (β-CD) has been widely utilized to recognize target molecules of compatible size based on its hydrophobic cavity together with a hydrophilic external surface.¹⁷ Therefore, it would be a promising solution through the preparation of MIPs with β-CD to contain hydrophobic and hydrophilic functionalities in a single material. The imprinting efficiency of traditional MIPs with an individual functional monomer, such as MAA, 4-VP, will be improved.

In the process of synthesis of MIPs, many methods have been used, such as bulk polymerization,¹⁸ suspension polymerization,¹⁹ emulsion polymerization,²⁰ and precipitation polymerization.²¹ In recent years, precipitation polymerization has been developed as a heterogeneous polymerization technique for the synthesis of polymer micro-/nanospheres with uniform size, shape and various functionalities in the absence of any added surfactant or stabilizer.²² Unfortunately, the abhorrently long polymerization time is inevitable. In order to resolve the problem, ultrasound-assisted preparation polymerization²³ was utilized. As a result, the polymerization time was greatly shortened. Herein, the magnetic MIP beads with a core–shell structure were designed and synthesized by the microwave-assisted precipitation polymerization of methacrylic acid functional monomer (MAA) and ethyleneglycol dimethacrylate (EGDMA) crosslinker in the presence of the silanized β-CD arm coated magnetite nanoparticle (Fe₃O₄) as a template. The mechanism for the construction of Fe₃O₄–β-CD@BAI–MIPs core–shell microspheres and the primary recognition of the MIPs for baicalein with a facile recovery via a magnetic separation process have been elucidated. The imprinting efficiency between individual functional monomer and assisted β-CD for conventional MIPs has been discussed.

Experimental

Materials

Baicalein (BAI) and quercetin (Qu) were purchased from Ciyuan Biotechnology Co Ltd (Shaanxi, Chin). Chloramphenoicol (CAP) was bought from Wuhan Pharmaceutical Chemical Co Ltd (Wuhan, China). Methacrylic acid (MAA), dimethylsulfoxide (DMSO), acetonitrile (AN), methanol, acetic acid, ethanol and dimethyl formamide (DMF) were purchased from Tianjin Kermal Chemical Reagent Co Ltd (Tianjin, China). N,N′-Azo-bis-isobutyronitrile (AIBN) and β-cyclodextrin (β-CD) were provided by Damao Reagent Plant (Tianjin, China). Ethyleneglycol dimethacrylate (EGDMA) and glycidoxypropyltrimethoxysilane (GPTMS) were afforded from Corel Chemical Plant (Shanghai, China). Ferric chloride hexahydrate (FeCl₃·6H₂O), ferrous chloride tetrahydrate (FeCl₂·4H₂O), sodium hydroxide and sodium hydride were bought from Tianjin Guangfu Fine Chemical Research Institute (Tianjin, China). Polyethylene glycol (PEG M_w = 6000) was obtained from Tianjin Tiantai Chemical Reagents (Tianjin, China). MAA was purified by vacuum distillation and AIBN was recrystallized in methanol before utilization.

Preparation of PEG-modified Fe₃O₄ nanoparticles

The Fe₃O₄ magnetic nanoparticles were first prepared by the co-precipitation technique according to the literature.^24,25 Briefly, 2.7 g FeCl₃·6H₂O (1.0 mmol Fe(III)) and 1.0 g FeCl₂·4H₂O (0.5 mmol Fe(II)) were dissolved in 150 mL of de-ionized water under nitrogen with mild stirring for 10 min. Then the mixture was heated to 50 °C and 20 mL of sodium hydroxide aqueous solution was added dropwise into the reaction system with vigorous mechanical stirring. After 30 min, the reaction was heated and maintained at 90 °C for 60 min. After the reaction, the black precipitate was collected with the aid of the outer magnetic field and washed several times by ethanol and distilled water successively. Finally, the resultant Fe₃O₄ nanoparticles (500 mg) were modified with PEG aqueous solution (1.0 g, 25 wt% in distilled water) and then mixed with ultra-sonication as magnetic ferro-fluid (Fe₃O₄–PEG) for further utilization.

Synthesis of silylated β-CD

1.0 g of β-CD was dissolved in 30.0 mL of anhydrous DMF solution, 0.1 g of NaH was added into the mixture with vigorous stirring for 90 min. After the reaction, the excess NaH was rapidly filtered. Then 0.5 mL of GPTMS was added into the reaction solution dropwise. The silanation of β-CD was performed at 80 °C for 15 h. The resultant precipitate was washed with anhydrous DMF, methanol and acetone sequentially and dried at 100 °C in a vacuum oven to a constant weight.

Synthesis of Fe₃O₄–β-CD@BAI–MIPs

DMF (50 mL) and NaOH solutions (25 wt%, 2 mL) were successively added into the above resultant materials. Then, the mixture solution was left to react for 12 h with vigorous stirring. The target products (Fe₃O₄@PEG/β-CD) were separated via an external magnetic field, and rinsed sequentially with DMF, water, and methanol. After drying under vacuum at 50 °C, 0.27 g (1.0 mmol) of BAI, and 0.17 g (2.0 mmol) of MAA were dissolved in 15 mL of anhydrous DMF and agitated by rolling the bottles in a horizontal position at approximately 140 rpm on an SHA-B shaker for 12 h. Then 0.80 g (4.02 mmol) of EGDMA crosslinker and 20 mg (2 wt%, relative to the monomer and crosslinker) of AIBN initiator were dispersed in 100 mL of DMF/deionized water (8/2, v/v) and added into the above mixture solution. Afterwards, the precipitation polymerization was performed at 60 °C with an ultrasonic irradiation for 6 h. After the polymerization, the resultant Fe₃O₄–β-CD@BAI–MIPs were collected via utilization of an applied magnetic field. The BAI templates were selectively removed from the Fe₃O₄–β-CD@BAI–MIPs by a Soxhlet extraction technique with a mixture solvent of methanol and acetic acid (8/2, v/v) as eluant until the absence of BAI molecules in the eluent, which was determined by UV spectrometry with the absorption peak at 275 nm. The resultant nanoparticles (Fe₃O₄–β-CD@MIPs) were finally rinsed in ethanol to remove the residual acetic acid and dried at 60 °C to a constant weight in a vacuum oven before use.

For comparisons, Fe₃O₄@MIPs in absence of the β-CD arm and Fe₃O₄–β-CD@non-imprinted polymers without BAI template (Fe₃O₄–β-CD@NIPs) were also prepared for evaluation of the imprinting efficiency.

Adsorption experiments

To investigate the adsorption capacity of the resultant functional polymer beads, batch rebinding experiments were implemented as follows. 50 mg of the resultant polymer beads, including Fe₃O₄@MIPs, Fe₃O₄–β-CD@NIPs and Fe₃O₄–β-CD@MIPs, were suspended in 5 mL of methanol solution with various BAI concentrations (0–5.0 mmol L⁻¹). The suspension was gently agitated by shaking the bottles in a horizontal position at around 150 rpm on a SHA-B shaker at room temperature for 20 h. These polymer beads adsorbed with BAI molecules were separated from the suspension with the aid of an external magnetic field. The concentration of the BAI in the supernatant was detected by UV spectroscopy with the absorption peak at 275 nm. The average adsorption capacity Q (μmol g⁻¹) of BAI was repeatedly calculated according to the formula:

Q = (C₀ − C_t)V/M (1)

where, C₀ (mmol L⁻¹) is the initial concentration of BAI, C_t (mmol L⁻¹) is the final concentration of BAI after adsorption, V (mL) is the volume of the solution, M (g) is the mass of the polymers.

Based on the above experiment, the dynamic adsorption experiment was implemented. 50 mg of the resultant beads were suspended in 5 mL 2 mmol L⁻¹ BAI methanol solution. After different adsorption time (5–210 min), the concentration of BAI in the supernatant was detected as same as for isothermal adsorption.

Selectivity adsorption experiments

The selectivity test of the above mentioned beads was performed between BAI and the structurally similar compound quercetin (Qu) as well as the dissimilar compound chloramphenoicol (CAP). The resultant Fe₃O₄–β-CD@MIPs beads (50 mg) were placed in 6.0 mL methanol solutions containing 2 mL 2 mmol L⁻¹ of BAI, Qu and CAP, respectively. The adsorption of the resultant beads for BAI was performed with mild rolling for 20 h on a SHA-B shaker. After the polymer beads adsorbed with the different molecules were separated from the system under an extra magnetic field, the concentrations of the residual BAI, Qu, and CAP in the solution were measured by UV-vis spectrometry with the determination of peaks at 275, 256, 272 nm, respectively.

Calculation of the selectivity adsorption

The static distribution co-efficient index (K_d), selectivity factor (a) and imprinting efficiency (β) were calculated by the following formulae:

K_d = C_p/C_s (2)

a = K_d(BAI)/K_d(reference) (reference = Qu, CAP) (3)

β = Q_m/Q_n (4)

where, C_p (mmol L⁻¹) is the initial concentration of adsorbed solution; C_s (mmol L⁻¹) is the residual BAI, Qu and CAP concentrations in the solution; K_d(BAI) and K_d(reference) are the static distribution co-efficient indexes of the BAI template and references of Qu and CAP molecules; Q_m and Q_n represent the adsorption capacities of Fe₃O₄–β-CD@MIPs and Fe₃O₄–β-CD@NIPs, respectively.

Characterization

The morphologies of the nanoparticles were determined by field emission scanning electron microscopy (FESEM) using an S-450 FESEM microscope and Transmission electron microscopy (TEM) on an H-7650 TEM microscope (Hitachi, Japan).

The Fourier infrared spectra (FT-IR) were recorded on a TENSOR-37 spectrometer (Brucker, Germany) ranging from 4000 to 400 cm⁻¹ with a potassium bromide pellet.

Thermogravimetric analysis (TGA) was performed on a SAT 409 PC (Netzsch, Germany) via heating from room temperature till 800 °C with a rate of 15 °C min⁻¹ under a nitrogen flow.

The magnetic properties of the nanoparticles were measured with a vibrating sample magnetometer (VSM) (LDJ 9600-1, America) at room temperature.

The concentrations of BAI, Qu and CAP were determined by UV-vis spectroscopy on a UV spectrometer (Tu-1901, China).

The silica content of silylated β-CD was measured by Inductively Coupled Plasma (ICP, JY/ACTIVA-M, France).

Results and discussion

The magnetic BAI molecularly imprinted polymer beads with a silylated β-CD arm coated PEGylated magnetite (Fe₃O₄–β-CD@BAI–MIPs) were designed and prepared as shown in Scheme 1. The whole procedure for this synthesis mainly involved three steps: (i) preparation of PEGylated magnetite nanoparticles; (ii) preparation of silylated β-CD; (iii) preparation of Fe₃O₄–β-CD@BAI–MIPs via the ultra-sonication assisted precipitation polymerization of the functional super-monomer MAA-self-assembled with the BAI endorsed with the silylated β-CD and the EGDMA crosslinker in presence of the PEGylated magnetite nanoparticle as the core. The recognition cavities of the resultant Fe₃O₄–β-CD@MIPs for BAI were afforded by a Soxhlet extraction to selectively remove the imprinted BAI molecules.

Scheme 1 Synthesis of Fe₃O₄–β-CD@BAI–MIPs.

Mechanism for the synthesis of Fe₃O₄–β-CD@BAI–MIPs

The magnetic indole-3-acetic acid imprinted polymer beads with β-CD as a functional arm have been successfully synthesized by a microwave heating initiated polymerization and their application has been investigated for trace analysis of auxins in plant tissues.¹⁶ However, the mechanism for the synthesis has not been clearly stated in this work. In our work, the mechanism for the synthesis of Fe₃O₄–β-CD@BAI–MIPs has been described systematically as follows.

The magnetite (Fe₃O₄) nanoparticles were prepared by the chemical co-precipitation of Fe(III)/Fe(II) (2/1 in molar ratio) cations in a basic solution (1.0 M NaOH) at 60 °C via a sol–gel process. These magnetite nanoparticles were prone to aggregate in the solvent and during the drying process due to the high specific surface area and strong interparticle interaction. This was confirmed by the FESEM micrograph in Fig. 1A with presence of the agglomerate consisting of several magnetite nanoparticles with the average diameter of 10 nm. The formation of the PEGylated Fe₃O₄ nanoparticles not only improved the stability of the composite nanoparticles as shown by the FESEM and TEM micrograph in Fig. 1B, but also endowed the hydrophilicity and the hydroxyl groups on the surface for the further functionalization and modification of the resultant magnetic nanoparticles.²⁵ The mass ratio of the PEG components in the PEGylated magnetite nanoparticles was as high as 22% determined by TGA as shown in Fig. 5b.

Fig. 1 FESEM and TEM micrographs of (A) Fe₃O₄ nanoparticles; (B) PEGylated Fe₃O₄ nanoparticles; (C) Fe₃O₄–β-CD@BAI–MIPs.

Molecularly imprinted polymers (MIPs) have special recognition performance based on the specific interaction/reaction between the functional monomer and the template molecule. Therefore, the selection of functional monomer was the most important aspect to achieve a good imprinting efficiency for the MIPs. However, the imprinting efficiency of traditional MIPs with a simple monomer system was not satisfactory enough for recognition after the selective removal of the template molecule. Herein, the β-CD, as a functional bridge between the BAI molecule and the magnetite core, was introduced to improve the imprinting efficiency during the recognition. The natural β-CD has certain limitations in practical applications,²⁶ such as poor water-solubility and fewer effective functional sites. In order to improve the compatibility of β-CD, GPTMS was used to modify the β-CD molecule.²⁷ The silylated β-CD was prepared via the inter-molecular condensation of the hydroxyl groups on the hydrophilic external surface of β-CD and the methoxy groups of GPTMS with presence of sodium hydride as shown in Scheme 1. The silica content in the silylated β-CD was 6.9% determined by ICP, which was consistent with the theoretical value of 7.7% according to the chemical structure in Scheme 1. The incorporation of the reactive methylsilane group via this modification provided the possibility for silylated β-CD as a functional bridge between the PEGylated magnetite core and the BAI imprinted polymer shell as shown in Scheme 1.

To elucidate the mechanism of the precipitation polymerization and the role of β-CD for the resultant Fe₃O₄–β-CD@BAI–MIPs, the UV-spectra were carefully and systematically investigated for BAI, MAA–BAI, BAI/β-CD, and MAA–BAI/β-CD in DMF solutions as shown in Fig. 2. The UV-spectrum of BAI in DMF has two peaks at 275 and 323 nm, which may be attributed to the hydrophilic cinnamoyl group and the hydrophobic benzoyl group, respectively. The UV-vis spectra of MAA–BAI with different ratios in DMF were illustrated in Fig. 2A. It indicates that the peak at 323 nm was gradually weakened with the molar ratio of MAA–BAI increasing from 0/1 to 4/1, while the peak intensity at 275 nm was weakened from 1.33 to 1.23 in the presence of MAA. According to the structure of BAI and the changes in the UV spectra, we deduced that there was a strong hydrogen-bonding interaction between the carboxyl groups of MAA and the hydroxyl groups in the BAI molecule as illustrated in Scheme 1(3). After the formation of hydrogen-bonding, the density of electron clouds decreased, the π-electron mobility increased, and a lower energy was required for the transition of electronic energy levels. Consequently, there was a significant weakening of the spectral band. In Fig. 2A, there was an isoabsorptive point at 350 nm. It revealed that two coordination interactions between BAI and MAA were formed in the process of self-assembly. Fig. 2B shows the UV-spectra of BAI/silylated β-CD in DMF with different ratios ranging from 1/0 to 1/4 in molar ratio. It demonstrates that the peak at 323 nm was considerably red-shifted from 323 to 329 nm with decreasing the ratio of BAI/silylated β-CD from 1/0 to 1/4, while the peak at 275 nm slightly changed. These red-shifts may originate from the efficient interaction between the hydrophobic benzoyl group and hydrophobic cavity of β-CD as described in Scheme 1(3), which reduced the energy gap for the π–π* transition to arise a better stabilization of excitons. With the BAI/silylated β-CD (1/1) as an example, the influence of the MAA on the UV-vis spectra were determined via addition of different amounts of MAA functional monomers as shown in Fig. 2C. For the MAA–BAI/silylated β-CD tri-component systems, both the peaks at 283 and 322 nm of the UV-vis spectra were red-shifted. Addition of more and more MAA in this tri-component system had a more significant influence on the absorption peak at 322 nm in UV-vis spectra as observed in Fig. 2C, which was attributed to the hydrogen-bonding interaction between MAA and BAI/β-CD complex. The UV-vis peak at 322 nm almost disappeared, which may be due to the formation of a core–corona micelle for MAA–BAI/silylated (7 : 1 : 1) tri-layer system. In such a case, the hydrophobic part may be located inside the micelle as core, which was shielded by the hydrophilic core during UV-vis irradiation. All these results demonstrated that the self-assembled MAA–BAI/silylated β-CD super-monomers were formed as shown in Scheme 1(3) via the strong hydrogen-bonding interaction between the MAA group and the phenol group of BAI together with the efficient hydrophobic interaction between the phenyl group of BAI and the hydrophobic cavity of β-CD, which would facilitate the precipitation polymerization and the improvement in recognition property of the Fe₃O₄–β-CD@BAI–MIPs.

Fig. 2 The UV-spectra of (A) BAI/MAA, (B) BAI/β-CD and (C) BAI/MAA/β-CD with different molar ratios in DMF solution.

The MAA monomer, silylated β-CD, and BAI were mixed in DMF with a BAI–β-CD–MAA molar ratio of 1 : 1 : 2 in a round bottom flask with mild agitation on a SHA-B shaker overnight to form a super-MAA monomer. The Fe₃O₄–β-CD@BAI–MIPs were synthesized by precipitation polymerization of the super-MAA as functional monomer and EGDMA as crosslinker, in presence of PEGylated nanoparticles as the core with the aid of ultra-sonic irradiation in DMF/deionized water solvent. The FESEM micrograph of the resultant Fe₃O₄–β-CD@BAI–MIPs in Fig. 1C shows that these microparticles have a spherical shape and rough surface with the average diameter of 300 nm. The inserted TEM micrograph in Fig. 1C indicates that Fe₃O₄–β-CD@BAI–MIPs have a typical core–shell structure with a deep contrast inner core containing Fe₃O₄ nanoparticles and a slight contrast of MIP polymer shell. For comparisons, the molecularly imprinted polymer beads in absence of β-CD arm (Fe₃O₄@BAI–MIPs) were similarly prepared by precipitation polymerization of MAA functional monomer and EGDMA crosslinker in the presence of BAI as the template molecule and PEGylated Fe₃O₄ nanoparticles as the core. The molecularly non-imprinted polymer beads in absence of the BAI template (Fe₃O₄–β-CD@NIPs) were prepared by precipitation polymerization of MAA functional monomer and EGDMA crosslinker with presence of PEGylated Fe₃O₄ nanoparticles as the core.

The chemical structures of Fe₃O₄–β-CD@BAI–MIPs were confirmed by the FT-IR spectra in Fig. 3. The presence of a peak at 2929 cm⁻¹ in the FT-IR spectrum of Fe₃O₄–β-CD@BAI–MIPs (Fig. 3a) proved that the β-CD was incorporated in the imprinted polymer beads, which was consistent with the imprinted polymer using acryloyl-β-CD and acrylamide as functional monomers in the literature.²⁸ The presence of the peaks at 1728 cm⁻¹ attributed to the stretching vibration of the carbonyl components from PMAA as well as PEGDMA network were clearly observed in Fig. 3a, c and d for all these three polymer beads. The formation of P(EGDMA-co-MAA) shell-layer over Fe₃O₄nanoparticles for Fe₃O₄–β-CD@BAI–MIPs and Fe₃O₄–β-CD@NIPs was attributed to the reaction between the tri-methoxy group from GPTMS and the hydroxyl groups of PEGylated magnetic nanoparticles, which was confirmed by Ji et al.²⁹ The P(EGDMA-co-MAA) shell-network in Fe₃O₄@BAI–MIPs was formed by ultrasonic assisted precipitation polymerization with the aid of efficient hydrogen-bonding interactions between the carboxyl group of PMAA and the hydroxyl group of the PEGylated Fe₃O₄ nanoparticles, which was similar to the construction of P(DVB-co-AA)@PAA@P(DVB-co-AA) (DVB: divinylbenzene) tri-layer microspheres in our previous work.³⁰ These peaks ranging from 1630 to 1450 cm⁻¹ in Fig. 3a proved the incorporation of the BAI template molecules in Fe₃O₄–β-CD@BAI–MIPs, which was contributed from the stretching vibration of C–C bonds in BAI species. The strength for this wide peak was considerably decreased in Fig. 3b after the extraction of the BAI with a mixed solvent of methanol and acetic acid (8/2, v/v) using the Soxhlet extraction technique, which confirmed the selective removal of the BAI templates from Fe₃O₄–β-CD@BAI–MIPs through the damage of hydrogen bonding interaction between the monomer and BAI.

Fig. 3 FT-IR spectra: (a) Fe₃O₄–β-CD@BAI–MIPs; (b) Fe₃O₄–β-CD@MIPs; (c) Fe₃O₄@MIPs; (d) Fe₃O₄–β-CD@NIPs.

The magnetic properties of the Fe₃O₄, Fe₃O₄@BAI–MIPs, Fe₃O₄–β-CD@BAI–MIPs, and Fe₃O₄–β-CD@BAI–NIPs were determined by VSM at room temperature, as shown in Fig. 4. The saturation magnetization value of the particles was 15.99, 3.67, 2.43, and 2.0 emu g⁻¹, respectively. The saturation magnetization values for the functionalized nanoparticles were significantly smaller than that for magnetite core nanoparticles, which was due to the considerable mass decrease of the neat magnetite species after the functionalization. No magnetic hysteresis loops were observed from the field dependent magnetization plots in Fig. 4 for all these samples. In other words, the inexistence of remanence when magnetic field was removed, implies that all the magnetic particles retained paramagnetic properties originating from the magnetite core nanoparticles at room temperature. With the existence of an external magnetic field, the black suspension with these magnetic nanoparticles became transparent and clear within a short time. All these beads were attached to the walls of the vial as shown by the inserted digital photo in Fig. 4. As a result, it would be possible to facilely separate these magnetic particles during the recovery of BAI recognition for these molecularly imprinted beads.

Fig. 4 Magnetic properties by VSM technique: (a) Fe₃O₄ nanoparticles; (b) Fe₃O₄@MIPs; (c) Fe₃O₄–β-CD@BAI–MIPs; (d) Fe₃O₄–β-CD@NIPs.

The amount of the inorganic components in Fe₃O₄, PEGylated Fe₃O₄, Fe₃O₄@BAI–MIPs, Fe₃O₄–β-CD@BAI–MIPs and Fe₃O₄–β-CD@NIPs were determined by TGA as shown in Fig. 5. These results demonstrated that the magnetite contents were 29.5 mass% for Fe₃O₄@MIPs, 19.3 mass% in Fe₃O₄–β-CD@BAI–MIPs and 16.1 mass% for Fe₃O₄–β-CD@NIPs, as the residual silica from β-CD was very low after pyrolysis of these hybrid beads. The significant decrease of magnetite content was consistent with the decrease of the saturation magnetization value from 15.99 emu g⁻¹ to 2.43 emu g⁻¹ for these particles.

Fig. 5 TGA curves: (a) Fe₃O₄ nanoparticles; (b) PEGylated Fe₃O₄ nanoparticles; (c) Fe₃O₄@BAI–MIPs; (d) Fe₃O₄–β-CD@BAI–MIPs; (e) Fe₃O₄–β-CD@NIPs.

Adsorption properties of the resultant polymer beads

A series of BAI standard solutions with concentrations of 0–5.0 mmol L⁻¹ were utilized to determine the adsorption isotherms of these polymer beads via rolling the bottle for 24 h on a SHA-B shaker. The adsorption isotherms are illustrated in Fig. 6. The amount of the adsorbed BAI was increased with increasing of BAI concentration in the solution. For Fe₃O₄–β-CD@NIPs, the saturated adsorption capacity of 50 μmol g⁻¹ was afforded with 2.0 mmol L⁻¹ of BAI (Fig. 6c). The adsorption of BAI for Fe₃O₄@MIPs increased with increasing of the BAI concentration and reached a saturated adsorption capacity of 90 μmol g⁻¹ at the BAI concentration of 3.0 mmol L⁻¹ (Fig. 6b). The adsorption of BAI for Fe₃O₄–β-CD@MIPs increased with enhancing the BAI concentration and reached a saturated adsorption capacity of 140 μmol g⁻¹ at the BAI concentration of 4.0 mmol L⁻¹ (Fig. 6a). The significant effect of imprinted BAI molecules in these systems is attributed to the considerably higher adsorption capacities for Fe₃O₄–β-CD@MIPs (140 μmol g⁻¹, Fig. 6a) and Fe₃O₄@MIPs (90 μmol g⁻¹) than that for Fe₃O₄–β-CD@NIPs (50 μmol g⁻¹). The synergic effect of the β-CD arm enhanced the imprinting efficiency for Fe₃O₄–β-CD@MIPs (140 μmol g⁻¹) with a much higher capacity than that for Fe₃O₄@MIPs (90 μmol g⁻¹) in the absence of β-CD. Fe₃O₄–β-CD@NIPs had an adsorption capacity of 50 μmol g⁻¹, which may be due to the physical adsorption of BAI via the hydrogen bonding interaction between the carboxyl group of the polymer beads and the phenol group of BAI molecules.

Fig. 6 Adsorption isotherms of BAI on functional beads: (a) Fe₃O₄–β-CD@MIPs; (b) Fe₃O₄@MIPs; (c) Fe₃O₄–β-CD@NIPs.

A BAI standard solution with the concentration of 2.0 mmol L⁻¹ was used to evaluate the dynamic adsorption for rebinding capacity of the resultant polymer beads. The adsorption dynamic curves of Fe₃O₄–β-CD@MIPs, Fe₃O₄@MIPs and Fe₃O₄–β-CD@NIPs are shown in Fig. 7. It can be observed that the adsorption capacities of all these three beads increased very quickly during the first 20 min and then leveled off between 30 and 120 min. The adsorption equilibrium state was reached after 120 min. The results indicated that the adsorption capacities of Fe₃O₄–β-CD@MIPs (130 μmol g⁻¹) and Fe₃O₄@MIPs (100 μmol g⁻¹) were significantly larger than that of Fe₃O₄–β-CD@NIPs (50 μmol g⁻¹). These were mainly attributed to the presence of the specific binding cavities, which were afforded by the selective removal of the templated BAI molecules from Fe₃O₄–β-CD@BAI–MIPs and Fe₃O₄@BAI–MIPs via Soxhlet extraction. Compared to Fe₃O₄@MIPs, Fe₃O₄–β-CD@MIPs had a higher adsorption capacity via the incorporation of the β-CD arm in this system, which implied that more imprinted sites were well maintained during the extraction of BAI molecules due to the synergic effect of the silylated β-CD arm.

Fig. 7 Dynamic adsorption of BAI on functional polymer beads: (a) Fe₃O₄–β-CD@MIPs; (b) Fe₃O₄@MIPs; (c) Fe₃O₄–β-CD@NIPs.

The adsorption kinetics of MIPs can be described via utilization of the internal diffusion model and the pseudo-first-order and pseudo second-order kinetic models.^31,32 Through fitting the experimental data from Fig. 6, the pseudo second-order model is more consistent with the kinetics of the adsorption characteristic of MIPs than those of the pseudo first-order according to the internal diffusion model. The kinetic data were calculated by the following formula according to the second-order model:

t/Q_t = 1/(K_sQ_e²) + (1/Q_e)t (5)

where, the Q_e (μmol g⁻¹) and Q_t (μmol g⁻¹) represent the theory adsorption amounts and at time t (min), respectively. K_s (g min⁻¹ mmol⁻¹) is the rate constant of the pseudo second-order.

The dynamic binding data were summarized in Table 1. These results indicated that the binding process could be the rate-limiting step for the whole adsorption process for BAI.

Table 1 The dynamic binding indexes for the adsorption of BAI on the functional polymer beads

Beads	Equation	Qe	Qt	Ks	R^2
Fe3O4–β-CD@MIPs	t/Qt = 0.0073t + 0.3441	137.3	133.5	0.00015	0.9949
Fe3O4@MIPs	t/Qt = 0.0091t + 0.0698	110.2	108.5	0.00119	0.9967
Fe3O4–β-CD@NIPs	t/Qt = 0.0146t + 0.3186	68.6	64.2	0.00067	0.9990

---

The thermodynamic adsorption of Fe₃O₄–β-CD@MIPs was investigated under the temperature ranging from 293 to 333 K, which was treated by the thermodynamic equation:

ln K_D = −Δ_rG_m/RT = −Δ_rH_m/RT + Δ_rS_m/R (6)

where, the K_D is the balance dissociation constant, Δ_rG_m (kJ mol⁻¹), Δ_rH_m (kJ mol⁻¹), and Δ_rS_m (J K⁻¹) are the Gibbs free energy, enthalpy and entropy changes of the template molecules interacting with the binding sites of Fe₃O₄–β-CD@MIPs, respectively. R is the molar gas constant. The data was fitted according to calculations using the Van’t Hoff equation:

K_D = Q_e/C_e (7)

The Δ_rH_m was −28.1 kJ mol⁻¹, which was obtained from the straight slope via plotting ln K_D and 10³/T (Fig. 8). This suggested that the process of imprinting adsorption was mainly via a physicochemical adsorption mechanism. The hydrogen-bonding interaction between the resultant polymers and BAI played as a driving force for the construction of Fe₃O₄–β-CD@MIPs.

Fig. 8 The relationship between the dissociated equilibrium constant and temperature for the adsorption of BAI on Fe₃O₄–β-CD@MIPs.

The selectivity of Fe₃O₄–β-CD@MIPs, Fe₃O₄@MIPs, and Fe₃O₄–β-CD@NIPs was investigated by UV measurement of the residual amounts of BAI, Qu and non-analogue CAP in the upper liquid after the separation of the adsorbed beads via external magnetic field from the adsorption system. The adsorption was performed by adding independently these polymer beads into the solution of BAI, Qu and CAP with mild agitation on a SHA-B shaker for 20 h, respectively. The relevant data are summarized in Table 2. The selectivity factor of Fe₃O₄–β-CD@MIPs for Qu and CAP was 3.85 and 4.88, respectively. These results demonstrated that Fe₃O₄–β-CD@MIPs exhibited a higher selectivity for the adsorption of BAI templated molecules than those of Fe₃O₄@MIPs (2.56 and 2.94) and Fe₃O₄–β-CD@NIPs (1.16 and 1.20) for the reference molecules of Qu and CAP. The highest selectivity of Fe₃O₄–β-CD@MIPs was mainly contributed from the molecular size recognition via the hydrogen-bonding interaction between the BAI molecules and the carboxyl groups of PMAA with the synergic effect of β-CD arm in Fe₃O₄–β-CD@MIPs compared to that in Fe₃O₄@MIPs. The higher selectivity of BAI to CAP (4.85, 2.94) than that of BAI to Qu (3.85, 2.56) for the BAI-imprinted Fe₃O₄–β-CD@MIPs and Fe₃O₄@MIPs as chemical structure of Qu was more similar to that of the BAI molecule. The selectivity of Fe₃O₄–β-CD@NIPs (1.16 and 1.20 to Qu and CAP) was very low as there were not specific cavities/sites left in the polymer network for the selective adsorption of BAI molecules with a very low saturate adsorption capacity (50 μmol g⁻¹).

Table 2 The selectivity of BAI for the adsorption of BAI on the functional polymer beads

Substance	Fe3O4–β-CD@MIPs				Fe3O4@MIPs				Fe3O4–β-CD@NIPs
	Q	Kd	α	β	Q	Kd	α	β	Q	Kd	α
BAI	133.5	201	1.00	2.08	108	118.5	1.0	1.69	64	47.31	1.00
Qu	68.5	52	3.85	1.18	63	46.3	2.56	1.10	58	40.65	1.16
CAP	58.1	401	4.88	1.03	57	40.3	2.94	1.02	56	39.31	1.20

---

Conclusion

In this study, Fe₃O₄–β-CD@MIPs core–shell microspheres were synthesized by sonic-irradiation assisted precipitation polymerization of MAA functional monomer and EGDMA crosslinker in the presence of BAI imprinted molecules and the silylated β-CD–Fe₃O₄ nanoparticle as the template. The β-CD played an essential role for the formation of a supramolecular monomer for precipitation polymerization via the hydrophobic interaction with BAI and the intra-molecular condensation with GPTMS. Compared with Fe₃O₄@MIPs and Fe₃O₄–β-CD@NIPs, the resultant Fe₃O₄–β-CD@MIPs incorporated with a β-CD arm had a unique selectivity for the adsorption of BAI molecules, the selectivity factor for Qu and CAP was 3.85 and 4.88, respectively. The results of the adsorption experiments indicated that such MIPs could selectively recognize the imprinting molecules. The imprinted process was accorded with a second-order kinetic model, and was an exothermic reaction. In conclusion, the Fe₃O₄–β-CD@MIPs not only have good selectivity and imprinting efficiency for BAI, but can also be separated conveniently using an external magnetic field and exhibit super-paramagnetic behavior.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant nos 21174065 and 21374049), the Tianjin Municipal Natural Science Foundation (no. 13JCQNJ C02600), and PCSIRT (IRT1257).

References

1. T. Nishioka, J. Kawabata and Y. Aoyama, J. Nat. Prod., 1998, 61, 1413–1415.
2. D. S. Chou, G. Hsiao, Y. A. Lai, Y. J. Tsai and J. R. Sheu, Free Radical Biol. Med., 2009, 46, 1197–1203.
3. J. A. Wu, A. S. Attele, L. Zhang and C. S. Yuan, Am. J. Chin. Med., 2001, 29, 69–81.
4. H. Sun, Q. M. Che, X. Zhao and X. P. Pu, Eur. J. Pharmacol., 2010, 631, 53–60.
5. W. H. Chang, C. H. Chen, R. J. Gau, C. C. Lin, C. L. Tsai, K. Tsai and F. J. Lu, Planta Med., 2002, 68, 302–306.
6. Q. X. Zhang, J. Li, W. H. Zhang, X. P. Liao and B. Shi, Ind. Eng. Chem. Res., 2013, 52, 2425–2433.
7. E. S. Ong and S. M. Len, Anal. Chim. Acta, 2003, 482, 81–89.
8. B. Gao, Y. Li and Z. J. Zhang, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2010, 878, 2077–2086.
9. K. Matsumoto, A. Torimaru, S. Ishitobi, T. Sakai, H. Ishikawa, K. Toko, N. Miura and T. Imato, Talanta, 2005, 68, 305–311.
10. H. J. Wang, W. H. Zhou, X. F. Yin, Z. X. Zhuang, H. H. Yang and X. R. Wang, J. Am. Chem. Soc., 2006, 128, 15954–15955.
11. F. G. Tamayo, E. Turiel and A. Martín-Esteban, J. Chromatogr. A, 2007, 1152, 32–40.
12. K. Haupt and K. Mosbach, Chem. Rev., 2000, 100, 2495–2504.
13. V. Pichon and F. Chapuis-Hugon, Anal. Chim. Acta, 2008, 622, 48–61.
14. J. P. Lai, X. Y. Lu, C. Y. Lu, H. F. Ju and X. W. He, Anal. Chim. Acta, 2001, 442, 105–111.
15. R. X. Gao, X. Kong, X. Wang, X. W. He, L. X. Chen and Y. K. Zhang, J. Mater. Chem., 2011, 21, 17863–17871.
16. Y. Zhang, Y. W. Li, Y. L. Hu, G. K. Li and Y. Q. Chen, J. Chromatogr. A, 2010, 1217, 7337–7344.
17. E. Schneiderman and A. M. Stalcup, J. Chromatogr. B: Biomed. Sci. Appl., 2000, 745, 83–102.
18. A. G. Strikovsky, D. Kasper, M. Grün, B. S. Green, J. Hradil and G. Wulff, J. Am. Chem. Soc., 2000, 122, 6295–6296.
19. A. G. Mayes and K. Mosbach, Anal. Chem., 1996, 68, 3769–3774.
20. N. Pérez, M. J. Whitcombe and E. N. Vulfson, J. Appl. Polym. Sci., 2000, 77, 1851–1859.
21. J. S. Downey, G. McIsaac, R. S. Frank and H. D. H. Stöver, Macromolecules, 2001, 34, 4534–4541.
22. F. Bai, X. L. Yang and W. Q. Huang, Macromolecules, 2004, 37, 9746–9752.
23. N. Phutthawong and M. Pattarawarapan, Polym. Bull., 2013, 70, 691–705.
24. B. Y. Zu, G. Q. Pan, X. Z. Guo, Y. Zhang and H. Q. Zhang, J. Polym. Sci., Part A: Polym. Chem., 2009, 47, 3257–3270.
25. Y. Y. Zheng, X. B. Wang, L. Shang, C. R. Li, C. Cui, W. J. Dong, W. H. Tang and B. Y. Chen, Mater. Charact., 2010, 61, 489–492.
26. C. T. Rao, B. Lindberg, J. Lindberg and J. Pitha, J. Org. Chem., 1991, 56, 1327–1329.
27. L. Qin, X. W. He, W. Y. Li and Y. K. Zhang, J. Chromatogr. A, 2008, 1187, 94–102.
28. W. Zhang, L. Qin, X. W. He, W. Y. Li and Y. K. Zhang, J. Chromatogr. A, 2009, 1216, 4560–4567.
29. Y. S. Ji, X. Y. Liu, M. Guan, C. D. Zhao, H. Y. Huang, H. X. Zhang and C. M. Wang, J. Sep. Sci., 2009, 32, 2139–2145.
30. G. L. Li and X. L. Yang, J. Phys. Chem. B, 2007, 111, 12781–12786.
31. E. Demirbas, M. Kobya, E. Senturk and T. Ozkan, Water SA, 2004, 30, 533–539.
32. Y. S. Ho and G. McKay, Process Biochem., 1999, 34, 451–465.

---