An efficient hybrid design to prepare highly dense imprinted layer-coated silica particles for selective uptake of trace metsulfuron-methyl from complicated matrices

Abstract

This work presents an approach called “hybrid design” which combines “grafting from” and “grafting through” methods, in order to prepare highly dense imprinted layer-coated silica particles for improved binding capacity to metsulfuron-methyl (MSM). The imprinted conditions including the kind of solvent and molar ratio of template to functional monomer were carefully optimized by quantum mechanics method. The obtained imprinted materials were evaluated by transmission electron microscopy and rebinding experiments, exhibiting well-controlled shell thickness (up to ∼30 nm) and high binding affinity to MSM. Moreover, the rebinding amount of MSM-imprinted particles to MSM was nearly 1.6- and 1.4-folds that of normal surface-imprinted particles, respectively. When the MSM-imprinted particles were used as dispersive solid-phase extraction materials, the recoveries (RSDs) of MSM determined by high-performance liquid chromatography (HPLC) were 93.4% (1.5%), 86.8% (3.2%) and 88.4% (2.7%) in the spiked water, soil and wheat samples, respectively. Therefore, the method will provide new opportunities for enrichment and detection of trace MSM from complicated matrices.

---

1 Introduction

Sulfonylureas, such as metsulfuron-methyl (MSM), are environmentally deleterious substances that have been of societal security concern, because they are widely used for a variety of crops to control weeds.¹ Compared with other herbicides, sulfonylureas have been used with a lower dose of 10–40 g h m⁻²,² and are more rapidly degraded in soil.³ So the concentration of these herbicides is usually found very low (ppt range) in soil, water or corn samples. These greatly increase the difficulty of their detection in chromatographic analysis. Presently, the determination of sulfonylurea herbicides (SHs) mainly rely on high performance liquid chromatography (HPLC),⁴liquid chromatography-mass spectrometry (LC-MS),⁵capillary electrophoresis (CE),^6,7gas chromatography (GC)⁸ and immunoassay⁹ method, which are usually of low efficiency owing to the lack of specific adsorptive materials. Moreover, the complicated matrices can seriously interfere with the enrichment and detection of SH residues to reduce the reliability and reproducibility of SH analysis. Thus, the exploration in enrichment and detection for trace SHs from environment or food has attracted considerable research interest in recent years. And, the stable antibody-like materials with specific binding properties have been developed. For example, molecular imprinting polymers (MIPs) with specific recognition sites may make effective enrichment of trace substances possible.

Although molecular imprinting for the recognition of SHs has already been reported,^10–12 the most widely used method is bulk polymerization. The obtained MIPs exhibit some disadvantages including incomplete template removal, slow mass transfer and irregular material shapes.^13–15 To overcome the above drawbacks, the surface-imprinting method has recently been put forward.^16–24 In the imprinting process, the surface of inorganic materials are generally functionalized with polymer chains either by the “grafting from” and “grafting to” or “grafting through” methods. As the “grafting from” method, the initiator groups were immobilized at the surface of supports. The imprinting polymerization could be carried out at the surface of initiator-modified support particles suspended in a mixture of the monomers and solvent. A high grafting density of polymer chains can be achieved because the steric barrier to incoming polymers imposed by the in situ grafted chains does not limit the access of smaller monomer molecules to the active initiation sites.^25–30 However, as the initiator is bound to the surface of the substrates only by one end, and the second part of the initiator can diffuse into solution and start a polymerization reaction there, not all polymers which had formed are attached to the substrate, leading to a low graft ratio and low yield of imprinted sites.³¹ In the manner of “grafting to”, often end-functionalized polymers are first synthesized and then react with appropriate surface sites. Another straightforward technique, which was similar to the “grafting to”, is the “grafting through” approach. But the link of polymer chains or networks to the surface is a random process in the “grafting through” manner.^32–34 Once the surface becomes significantly covered, additional polymer molecules, which are trying to reach the surface, have to diffuse against the concentration gradient built up by the already deposited polymer chains. This leads to a strong kinetic and steric hindrance for the attachment of additional polymer chains to the surface and impedes further imprinted layer growth. Therefore, the formation of surface-bound polymer monolayers by such a “grafting through” method is intrinsically limited to low graft density and low imprinted layer thickness.³⁵

Compared with traditional MIPs, surface-imprinting polymers have several advantages, such as the complete removal of templates, good accessibility to the target species, and low mass-transfer resistance. However, the surface-imprinting materials have limited (or few) imprinted sites, resulting in a low rebinding capacity.

In this paper, an efficient hybrid design which combines “grafting from” and “grafting through” methods was firstly developed to prepare highly dense imprinted layer-coated silica particles for improved binding capacity to MSM. Simultaneously, a quantum mechanics method was used to predict the suitable polymerization solvent and the molar ratio of template to functional monomer, and to identify functional monomers capable of interacting with MSM. The volume ratio of two silylated reagents was carefully optimized. The obtained MSM-imprinted particles showed higher binding capacity, compared to normal surface-imprinted particles.

2 Experimental

2.1 Chemicals

3-Aminopropyltriethoxysilane (APTS), 3-(methacryloxy)propyltrimethoxysilane (MPTS), trimethylolpropane trimethacrylate (TRIM) and 4,4′-azobis(4-cyanopentanoic acid) (ACPA) were purchased from Sigma-Aldrich (Steinheim, Germany). Tetraethylosilicate (TEOS), ammonia, azo(bis)-isobutyronitrile (AIBN), methacrylic acid (MAA), thionyl chloride and acetic acid were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Analytical grade toluene, acetonitrile, dichloromethane, chloroform, methanol and ethanol were purchased from Nanjing Chemical reagent Co., Ltd. (Nanjing, China). AIBN was purified by recrystallization from ethanol before use. All the other chemicals were used as received. Metsulfuron-methyl (MSM), chlorsulfuron (CS), bensulfuron-methyl (BSM), tribenuron-methyl (TBM) and thifensulfuron-methyl (TSM) were purchased from Sigma-Aldrich (Steinheim, Germany). The methanol of HPLC grade was purchased from Merck (Darmstadt, Germany). Commercial C18 was from YMC Co. Ltd. (Shimogyo-ku, Japan). HPLC water was doubly distilled. Tap water samples were collected from a suburb of Nanjing in China. Soil samples were from a suburb of Nanjing in China. Wheat samples were from a farm of Xuzhou in China.

The individual stock (10 mmol L⁻¹) solutions were prepared in chloroform, and the standard solutions of lower concentration were prepared by the serial dilution of stock solutions. All stock and standard solutions were stored at −18 °C in a refrigerator and re-prepared every month.

2.2 Apparatus and analytical methods

The instruments used in this study were as follows: UV-2450 UV-vis spectrophotometer (Shimadzu, Kyoto, Japan), Bruker Vector 27 FT-IR spectrometer (Bruker, Ettlingen, Germany), JEM-1010 transmission electron microscope (JEOL, Tokyo, Japan), thermogravimetric analyzer Linseis L81 (Linseis, Selb, Germany) and a Shimadzu HPLC system (Shimadzu, Kyoto, Japan) equipped with LC-20AT pump, SPD-M20A detector, CTO-20A column oven and Shimadzu shim-pack C18 column (5 μm, 250 mm × 4.6 mm I.D.). The column temperature was 30 °C. The mobile phase was methanol/water (65 : 35, v/v), the pH of which was adjusted to 3.0 with acetic acid, at a flow rate of 1.0 mL min⁻¹. MSM was monitored at 230 nm. The injection volume was 20 μL. Sample solutions were filtered through a nylon 0.22 μm filter before use.

2.3 Chlorination of 4,4′-azobis(4-cyanopentanoic acid) with thionyl chloride

ACPA (1.40 g) was dispersed in dichloromethane (150 mL), followed by the addition of thionyl chloride (2.0 mL) within 20 min. The reaction mixture was stirred for 12 h while in an ice bath under high-purity nitrogen. After reaction, the solvent was evaporated and 4,4′-azobis(4-cyanopentanoyl chloride) (ACPC) was finally obtained. The product was washed with dichlormethane to remove unreacted thionyl chloride, and then dried at 4 °C under vacuum overnight.

2.4 Synthesis and chemical modification of silica particles

Bifunctional modification of silica particles. Uniform spherical silica particles with a size of ∼400 nm were prepared by the hydrolysis of TEOS with aqueous ammonia according to the reported method.³⁶ Then, the monodispersive silica particles were chemically modified by two silylated reagents. Typically, dried silica particles (0.3 g), APTS (2.0 mL) and MPTS (4.0 mL) were added into toluene (50 mL), refluxed for 12 h under high-purity nitrogen. The resultant SiO₂@(MPTS+APTS) were separated by centrifuge, washed with toluene, and redispersed in toluene.

Subsequently, the amino end groups of APTS monolayer were further acryloylated by ACPC. Typically, ACPC (2.66 g) was dispersed in SiO₂@(MPTS+APTS) toluene solution (50 mL), and anhydrous potassium carbonate was added into this reaction system as a catalyst. The mixture was vigorously stirred for 12 h while remaining in ice bath under high-purity nitrogen. The product was separated by centrifuge and washed with toluene, water and ethanol, in that order, and dried under vacuum at room temperature. Finally, SiO₂@[MPTS+(APTS-ACPC)] was obtained.

Monofunctional modification of silica particles. The SiO₂@(APTS-ACPC) was prepared by immobilized APTS-ACPC monolayer at the surface of silica particles. The preparation process was identical with that of SiO₂@[MPTS+(APTS-ACPC)], except that the total volume (6.0 mL) of MPTS and APTS was replaced by 6.0 mL of APTS.

The SiO₂@MPTS was prepared by immobilized MPTS monolayer at the surface of silica particles. Typically, dried silica particles (0.3 g) and MPTS (6.0 mL) were added into toluene to make the mixture solution (50 mL). The mixture was refluxed for 12 h under high-purity nitrogen. The resultant SiO₂@MPTS was separated by centrifuge and washed with toluene.

2.5 Molecular modeling studies

Quantum mechanics (QM) was employed to select the optimal solvent and molar ratio of template to functional monomer. The solvent effects on the electronic structures of the studied systems were evaluated through polarized continuum model (PCM) (dielectric constant ε = 35.69, 2.37, 4.71 and 7.43 for acetonitrile, toluene, chloroform and tetrahydrofuran, respectively). This model had been successfully employed to predict the surface morphology and molecular recognition properties of MIPs.^37–39 The full geometry optimization of template, monomer and their complexes with different ratios were carried out with density functional theory (DFT) at B3LYP/3-21G level. Then, the corresponding energetic calculations were performed at the B3LYP/6-31G(d,p) level. The binding energy (ΔE_bind) of MSM with the monomer in solvent was calculated according to the following formula:

ΔE_bind = E_complex – E_template – ΣE_monomer (1)

where E_template and E_monomer were the energies of the template and functional monomers, respectively, and E_complex was the total energy of the template–monomer complex. All the calculations were performed with the Gaussian 09 program.⁴⁰

2.6 Preparation of MSM-imprinted particles

Prior to polymerization, the prearranged solution was prepared by MSM (0.625 mmol) and MAA (2.5 mmol) dissolved in 25 mL of mixture solvent with toluene/acetonitrile (4 : 1, v/v) and placed in the dark for 12 h. Subsequently, 0.1 g of SiO₂@(APTS-ACPC), SiO₂@MPTS and SiO₂@[MPTS+(APTS-ACPC)] particles were dispersed in 25 mL of mixture solvent with toluene/acetonitrile (4 : 1, v/v) by ultrasonic vibration, respectively. The obtained three mixture solutions were mixed with the prearranged solution, TRIM (2.5 mmol) and different amounts of AIBN (i.e., 0 mmol, 0.10 mmol and 0.06 mmol), respectively. And the mixing solutions were purged with high-purity nitrogen while cooled in ice bath. The polymerization was first done at 50 °C for 6 h, then at 60 °C for 24 h. The products were further aged at 85 °C for 6 h obtaining a high cross-linking density,²⁰ and separated by centrifuge and rinsed with methanol, respectively. To remove template molecules, the imprinted particles were ultrasonically cleaned by methanol/acetic acid (9 : 1, v/v) and dried under vacuum at room temperature. The obtained MSM-imprinted particles were termed SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3.

The non-imprinted particles (SiO₂@NIP1, SiO₂@NIP2 and SiO₂@NIP3) were also prepared using an identical procedure but without the addition of a template.

2.7 Binding properties of MSM-imprinted particles

Twenty milligrams of MSM-imprinted particles (SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3) and non-imprinted particles (SiO₂@NIP1, SiO₂@NIP2 and SiO₂@NIP3) were suspended in chloroform (5.0 mL) with various MSM concentrations (0.3–10 mmol L⁻¹), and the samples were shaken on a reciprocating shaking-table at room temperature, respectively. The particles in the solution were removed through a 0.22 μm microporous membrane after reaching adsorption equilibrium, and the equilibrium concentrations of MSM-imprinted and non-imprinted particles were determined with UV-vis spectrophotometer. The equilibrium adsorption amounts of MSM-imprinted and non-imprinted particles to MSM were calculated according to eqn (2), the binding isotherm was figured. The specific recognition property of MSM-imprinted particles was evaluated by imprinting factor (α), which was calculated according to eqn (3), and the binding constants of binding sites of MSM-imprinted particles were calculated according to the Langmuir eqn (4).where Q_e (μmol g⁻¹) was the equilibrium adsorption amount of MSM. C₀ (mmol L⁻¹) and C_e (mmol L⁻¹) were the initial and equilibrium (final) concentrations of MSM, respectively. V (L) was the volume of the MSM solution. m (g) was the mass of the polymer.where Q_MIP and Q_NIP were the adsorption capacity of MSM on MSM-imprinted and non-imprinted particles, respectively.where Q_e (μmol g⁻¹) was the adsorption capacity of MSM at equilibrium concentration, C_e (mmol L⁻¹) was the equilibrium concentration of MSM, K_e (L mmol⁻¹) was the equilibrium binding constant and K_s (μmol g⁻¹) was the saturation constant.

Meanwhile, the binding kinetics was tested by monitoring the temporal evolution of MSM concentration in the solutions. 20 mg of MSM-imprinted particles (SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3) and non-imprinted particles (SiO₂@NIP1, SiO₂@NIP2 and SiO₂@NIP3) were suspended in chloroform (5 mL) with MSM (2 mmol L⁻¹) solution, shaken on a reciprocating shaking-table at room temperature, respectively. The mixtures (including MSM-imprinted/non-imprinted particles and chloroform with MSM solutions) were taken suction after 5, 10, 15, 20, 30, 45, 60 and 120 min, then filtered through 0.22 μm microporous membranes. And the concentrations of MSM in the filtrate were determined with a UV-vis spectrophotometer. The dynamic binding curve was plotted.

The selectivity of SiO₂@MSM-MIP3 was measured using the structure analogues CS, BSM, TSM and TBM. 20 mg of SiO₂@MSM-MIP3 particles and SiO₂@NIP3 particles were suspended in chloroform (5.0 mL) with MSM, CS, BSM, TSM and TBM individual solutions at the 2 mmol L⁻¹ level, respectively. The equilibrium concentrations of SiO₂@MSM-MIP3 and SiO₂@NIP3 were determined with UV-vis spectrophotometer, and the equilibrium adsorption amounts of SiO₂@MSM-MIP3 and SiO₂@NIP3 were also calculated according to eqn (2). The selectivity factor (β) was defined as expressed in eqn (5).

where α_tem and α_ana are the imprinting factors toward MSM and the analogues, respectively.

2.8 Detection of MSM from the spiked samples

The wheat and soil were taken and crushed into homogenates. The homogenates (20 g) of wheat and soil were mixed 20 mL of MSM standard solution (0.015 μmol L⁻¹), incubated for 10 h and filtered, respectively. The filtrate was evaporated to dryness. Then, the residue was re-dissolved in chloroform (5.0 mL) to form the spiked samples. Water samples (20 mL) were spiked at a concentration of 0.015 μmol L⁻¹ with MSM. Subsequently, the mixture was filtered, evaporated to dryness under vacuum and then the residue was re-dissolved in chloroform (5.0 mL). So, the concentration of MSM in the three spiked samples was about 0.060 μmol L⁻¹.

SiO₂@MSM-MIP3, SiO₂@NIP3 and commercial C18 (20 mg) were dispersed in the spiked solutions (5.0 mL) and incubated for 60 min at room temperature. The particles were collected through a 0.22 μm microporous membrane, washed with chloroform (3.0 mL) to reduce nonspecific adsorption, and then eluted with 5.0 mL of methanol/acetic acid (9 : 1, v/v). The particles in the desorption solution were separated through 0.22 μm microporous membrane. The filtrate was dried by nitrogen and re-dissolved in methanol (0.5 mL), and then analyzed with HPLC by the procedures and conditions as mentioned in Section 2.2.

3 Results and discussion

3.1 Preparation and characterization of modified silica particles

To obtain MSM-imprinted particles with high binding capacity, two silylated reagents (APTS and MPTS) were simultaneously used in this work. The amount of APTS and MPTS modified at the surface of silica particles is a key factor for imprinting performance. At low graft density, few recognition sites caused by the low degree of cross-linking decrease rebinding capacity. With the increasing of graft density, the high degree of cross-linking may enhance the difficulties of the accessibility for template molecules in polymerization. Therefore, a suitable amount of grafting yield is critical for maximum adsorption capacity. With the fixed amount of silica particles (0.3 g) and the total volume of APTS and MPTS (6.0 mL), the volume ratio of MPTS to APTS from 1.0 : 2.0 to 3.0 : 1.0 was optimized using rebinding capacities. As shown in Table 1, at the region of low values of volume ratio, the rebinding capacities increased as the value of the volume ratio was increased. When the ratio was 2.0 : 1.0, the rebinding capacity of SiO₂@MSM-MIP to MSM reached to the maximum at 117.0 μmol g⁻¹. Further increasing the volume ratio led to the decrease of rebinding capacity. This is likely because the dense vinyl groups decreased the steric space and the polymer chains could not easily access the reactive sites to copolymerize the with cross-linker. Thus, a ratio of 2.0 : 1.0 of MPTS to APTS was chosen to obtain SiO₂@MSM-MIP with the maximum adsorption capacity.

Table 1 The rebinding capacities of SiO₂@MSM-MIP particles to MSM

Volume ratio of MPTS to APTS	Rebinding capacity (μmol g^−1)	RSD (%) (n = 5)
1.0 : 2.0	29.14	3.3
1.0 : 1.0	56.69	2.3
1.7 : 1.0	101.7	1.7
2.0 : 1.0	117.0	1.5
2.3 : 1.0	100.6	2.1
3.0 : 1.0	45.02	2.7

---

Fig. 1 showed the infrared spectra of SiO₂@(APTS-ACPC) (a), SiO₂@MPTS (b), SiO₂@[MPTS+(APTS-ACPC)] (c) and pure silica (d). Compared with the infrared data of pure silica, the SiO₂@(APTS-ACPC) displayed a characteristic peak at 1715 cm⁻¹ (due to the stretching vibration of C=O in amide group). In Fig. 1(b), the C=O band centered at 1735 cm⁻¹ (typical stretching vibration of C=O in ester group). The characteristic peaks of SiO₂@[MPTS+(APTS-ACPC)] appeared at 1710 cm⁻¹ and 1737 cm⁻¹ (attributed to the stretching vibration of C=O in amide group and ester group, respectively). These characteristic peaks found in SiO₂@(APTS-ACPC), SiO₂@MPTS and SiO₂@[MPTS+(APTS-ACPC)] verified the successful grafting of the silane coupling agents or azo-initiators to the surface of particles.

Fig. 1 FTIR spectra of SiO₂@(APTS-ACPC) (a), SiO₂@MPTS (b), SiO₂@[MPTS+(APTS-ACPC)] (c) and pure silica (d).

3.2 Molecular modeling of template–monomer interaction

Currently, a typical imprinting protocol, such as the selection of functional monomer and polymerization solvent, is tedious and time-consuming. Now, we have developed a method, i.e., quantum mechanics (QM) for selecting the most suitable solvent and optimizing the optimal molar ratio of template to functional monomer.

In order to understand the electric structure of MSM molecule, the Mulliken charges are investigated at B3LYP/6-31G(d,p) level and the results are displayed in Fig. 2a. It can be seen that there were several potential binding sites on the MSM molecule, which can interact with the MAA molecule through hydrogen bonding or electrostatic interactions. For example, the three nitrogens of the triazine ring and the four oxygens with negative partial charges in the MSM molecule were likely to become the hydrogen bond acceptors for the hydrogen atoms on the hydroxyl group of the MAA molecule (Fig. 2b). Because of the higher negative charges on the nitrogen atoms of the triazine ring than the oxygen atoms, the nitrogen atoms were firstly selected as the most preferable binding site for MAA.

Fig. 2 The charges obtained by molecular modeling of atoms on the MSM molecule (a), MAA molecule (b) and the complex formed between the MSM molecule and four molecules of MAA (c). Nitrogen is represented in blue, carbon in gray, oxygen in red, hydrogen in white and sulphur in yellow.

To select the optimal solvent, the binding energies (ΔE_bind) of MSM with the monomer in various solvents, including acetonitrile, toluene, chloroform and tetrahydrofuran, were compared. As shown in Table 2, the interaction energies between MSM and MAA in different solvents were ordered as follows: ΔE_bind (acetonitrile) > ΔE_bind (toluene) > ΔE_bind (chloroform) > ΔE_bind (tetrahydrofuran). It indicated that the MSM–MAA complexes in acetonitrile are more stable than those in toluene, chloroform and tetrahydrofuran. Furthermore, because the boiling point (b.p.) of acetonitrile, chloroform and tetrahydrofuran are 81 °C, 61 °C and 65 °C, respectively, which were lower than that of the aged temperature (85 °C) of the polymers, only toluene (b.p. 110 °C) can be used as the solvent. However, the MSM molecule was difficult to dissolve in toluene. So, a mixture of toluene/acetonitrile (4 : 1, v/v) was chosen as the polymerization solvent.

Table 2 Calculated binding energy (ΔE_bind) of template with functional monomer in different solvents

Solvent	E template (a.u.)	E monomer (a.u.)	E complex (a.u.)^a	ΔEbind (kcal mol^−1)
a The Ecomplex was the total energy of template-monomer complex with the molar ratio of 1 : 3.
Acetonitrile	−1656.74226	−304.78948	−2571.23402	−77.39
Toluene	−1656.72606	−304.79397	−2571.21701	−68.42
Chloroform	−1656.73351	−304.79550	−2571.22466	−65.66
THF	−1656.73688	−304.79616	−2571.22821	−64.53

---

In addition, the influence of molar ratio between MSM and MAA are also studied by QM method. The MSM-MAA complexes with different molar radios were investigated in toluene. The detailed geometric parameters and the interaction energies were given in Fig. 2c and Table 3, respectively. It can be seen that MSM–MAA complexes with molar radios of 1 : 1, 1 : 2, 1 : 3 and 1 : 4 are formed through intermolecular hydrogen bonding interactions. The average hydrogen bonding lengths (angle) are 1.57 Å (174.9°), 1.57 Å (172.8°), 1.60 Å (167.4°) and 1.64 Å (167.2°), respectively. With the addition of MAA molecules, there are more hydrogen bonds between MSM and MAA, resulting in the increased ΔE_bind. Further additions of MAA will not lead to a greater level of complexation. So, the optimal molar ratio of MSM to MAA was 1 : 4.

Table 3 The H-bond geometries and binding energies between MSM and MAA with different molar ratios

MSM:MAA	Intermolecular distance (Å)						Intermolecular angle (°)						Binding energy (kcal mol^−1)
	r N1⋯H2	r O5⋯H1	r O2⋯H3	r O1⋯H4	r O3⋯H5	Average	∠N1⋯H2-O4	∠O5⋯H1-N2	∠O2⋯H3-O6	∠O1⋯H4-O7	∠O3⋯H5-O8	Average	ΔEbind
1 : 1	1.54	1.59				1.57	176.4	173.3				174.9	−77.39
1 : 2	1.55	1.59	1.56			1.57	176.2	172.1	168.8			172.8	−68.42
1 : 3	1.56	1.57	1.57	1.68		1.60	176.0	172.6	169.1	151.0		167.4	−65.66
1 : 4	1.58	1.57	1.60	1.80	1.64	1.64	176.1	171.4	168.4	157.8	166.6	167.2	−64.53

---

3.3 Preparation of MSM-imprinted particles

Recently, several methods have been adopted to increase the thickness of the MIPs layer coated on silica particles to improve the rebinding capacity of imprinted particles to template molecules, for instance, by tuning the amount of surface end vinyl groups³⁶ and by controlling the ratio of acrylamide-APTS-silica to the polymerization precursors.²⁰ But the critical scale of the shell with the most imprinted sites was only ∼25 nm. Thus, in this work, we attempted a novel and straightforward strategy for the preparation of MSM-imprinted particles through combining “grafting from” and “grafting through” methods.

Here, the organic modifying layer on the silica support is very important. For SiO₂@(APTS-ACPC), APTS-ACPC with an azo group (initiator) was introduced on the surface of silica particles. When polymerization occurred at the silica particle’s surface, only one end of the initiator was bound to the surface, the second part dispersed into solution and started a polymerization reaction and formed some nonbonded polymers (as indicated with swallow tail arrow in the inset TEM image of Fig. 3A). Thus, the formation of surface-bound polymers by the “grafting from” technique was intrinsically limited to low graft ratio and thin shell thickness (Fig. 3A-a).

Fig. 3 Outline of fixation of different functional monomer at silica particles and the coating MIP layer at silica particles by the “grafting from” method (A), the “grafting through” method (B) and the combination of “grafting from” and “grafting through” methods (C). SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c).

For SiO₂@MPTS, MPTS with an active end group was modified onto the surface of silica particles, ensuring that the imprinting polymerization selectively occurred at the particle's surface. With the growth of surface polymer chains, additional polymers had difficultly reaching the silica particle’s surface due to steric crowding (as indicated with curved arrow in Fig. 3B). This impeded further the MIPs layer growth.³⁶ So, the surface-imprinted layer formed by the “grafting through” technique contained a limited number of recognition sites and its thickness was thin (Fig. 3B-b).

For SiO₂@[MPTS+(APTS-ACPC)], APTS-ACPC and MPTS were simultaneously modified on the surface of silica particles. And the copolymers at the silica particle’s surface included two parts. One part (polymer chains or netwoks) which had mainly formed by the “grafting through” method, entered in the space of the substrate's surface, as indicated with hollowed arrow in Fig. 3C. The other formed in the manner of “grafting from” distributed in a far space from the surface, as indicated with solid black arrow in Fig. 3C. Compared with SiO₂@MPTS, the vinyl groups on the SiO₂@[MPTS+(APTS-ACPC)] surface were not so crowded that most of the polymer chains were attached to the surface. The nonbonded polymers in solution could enter reactive sites at the silica particle’s surface (as indicated with curved arrow in Fig. 3C), crosslinked with excessive vinyl groups. Therefore, both the graft density and graft ratio of the polymers were improved. As a result, the imprinted layer of SiO₂@MSM-MIP3 contained so many recognition sites and was thick (Fig. 3C-c). The mechanism mentioned above is only qualitative. Detailed quantitative analysis including the determination or characterization of shell thickness, weight loss, binding capacity and binding affinity of the three imprinted materials can be seen in Section 3.4 and 3.5.

3.4 Characterization of MSM-imprinted particles

The obtained SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3 were characterized by transmission electron microscopy (TEM) and thermogravimetric analysis (TGA), respectively.

By TEM observation (Fig. 4), the thicknesses of imprinted layer at the surface of SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c) were about 18 nm, 20 nm and 30 nm, respectively.

Fig. 4 TEM images of SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c).

Additionally, Fig. 5 displayed the TGA curves of SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c). TGA curves of SiO₂@MSM-MIP1 (a) and SiO₂@MSM-MIP3 (c) approximately plateaued at the range 100–150 °C, showing weight losses of 5.7 and 5.5%, respectively, due to the thermal decomposition of the remaining azo groups and the evaporation of the adsorbed water.⁴¹ In a successive heating from 150 to 800 °C, the steep decrease was attributed to the further thermal decomposition of the imprinted layer. And the weights (wt%) of imprinted layers of SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c) were about 7.4%, 9.2% and 12.8%, respectively. The above data showed that the thicker the imprinted layer is, the higher the weight of the imprinted layer is, possibly resulting in the higher binding capacity.

Fig. 5 Thermogravimetic weight loss curves of SiO₂@MSM-MIP1 (a), SiO₂@MSM-MIP2 (b) and SiO₂@MSM-MIP3 (c).

3.5 Binding properties of MSM-imprinted particles

Binding isotherms. In accordance with the polarity of polymerization solvent, chloroform was selected as the suitable adsorption solvent of the imprinted particles. The binding capacities of the MSM-imprinted particles (SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3) and non-imprinted particles (SiO₂@NIP1, SiO₂@NIP2 and SiO₂@NIP3) with varying concentrations of MSM (0.3–10 mmol L⁻¹) were evaluated from batch binding studies. Fig. 6 shows that the rebinding capacity continuously increased along with the increase of concentration in range 0.3–5 mmol L⁻¹ for MSM-imprinted polymers and non-imprinted polymers, and reached equilibrium as the concentration was up to 5 mmol L⁻¹. The maximum rebinding amount (Q_max) of SiO₂@MSM-MIP3 to MSM was about 1.6- and 1.4-fold that of SiO₂@MSM-MIP1 and SiO₂@MSM-MIP2, respectively. In addition, from the data of the imprinting factors listed in Table 4, imprinted polymers had a greater uptake for MSM than non-imprinted polymers. These results suggest clearly that the inherent arrangement of functional monomer cause specificity of MSM-imprinted polymers toward MSM molecules.

Fig. 6 Rebinding capacity curves of SiO₂@MSM-MIP1 (▲), SiO₂@NIP1 (▼), SiO₂@MSM-MIP2 (■), SiO₂@NIP2 (●), SiO₂@MSM-MIP3 (◀) and SiO₂@NIP3 (▶) to MSM. The amounts of rebinding MSM were measured by suspending 20 mg of materials in 5 mL of solution with different MSM concentrations. The adsorption time was 60 min.

Table 4 Maximum equilibrium adsorption amounts (Q_max), imprinting factors (α), equilibrium binding constant (K_e) and saturation constant (K_s) for MSM on the imprinted polymers and non-imprinted polymers (n = 5)

Particles	Q max (μmol g^−1) (RSD%)	α ^a	K e (L mmol^−1)^b	K s (μmol g^−1)^b	r^2
a The imprinted factor (α) was calculated from maximum adsorption amounts of SiO2@MSM-MIP (QMIP) and that of SiO2@NIP (QNIP): α = QMIP/QNIP. b The equilibrium binding constant (Ke), saturation constant (Ks) of higher affinity binding sites were calculated by Langmuir equation: 1/Qe = 1/KeKsCe + 1/Ks, having a slope of (KeKs)^−1 and intercept of Ks^−1.
SiO2@MSM-MIP1	93.22 (1.5)	3.31	4.057	13.23	0.99
SiO2@NIP1	28.20 (3.1)		1.567	3.569	0.97
SiO2@MSM-MIP2	105.1 (1.2)	3.26	4.435	13.50	0.99
SiO2@NIP2	32.23 (2.3)		1.752	3.855	0.99
SiO2@MSM-MIP3	146.7 (1.0)	3.41	6.338	19.72	0.99
SiO2@NIP3	43.05 (2.9)		2.021	4.647	0.98

---

To obtain further knowledge of MSM binding on MSM-imprinted polymers, the Langmuir equation was employed to evaluate the binding affinity of the imprinted polymers.⁴² A plot of Q_e⁻¹versus C_e⁻¹ gave a straight line having a slope of (K_eK_s)⁻¹ and intercept of K_s⁻¹. From the slope and intercept of the plot, K_e and K_s of higher affinity binding sites were calculated. Meanwhile, K_e and K_s values are summarized in Table 4. From the data of Table 4, it is evident that the imprinted polymers have a higher binding affinity to MSM than non-imprinted polymers, and MSM interacts more strongly with SiO₂@MSM-MIP3 compared to SiO₂@MSM-MIP1 and SiO₂@MSM-MIP2.

The above results indicate that there were more recognition sites in SiO₂@MSM-MIP3 than that in SiO₂@MSM-MIP1 and SiO₂@ MSM-MIP2. This was because the imprinted layer of SiO₂@MSM-MIP3 was the thickest among three kinds of MSM-imprinted particles (as shown in Fig. 4). In other words, the thicker the imprinted layer is, the more imprinted sites there are, resulting in the higher binding affinity and binding capacity.

Binding kinetics. Adsorption dynamic studies were carried out to investigate the adsorption process and evaluated with 2 mmol L⁻¹ of MSM solution. Fig. 7 shows the time-dependent evolution of MSM amounts bound by MSM-imprinted particles (SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3) and non-imprinted particles (SiO₂@NIP1, SiO₂@NIP2 and SiO₂@NIP3). At the early time, a large number of imprinted cavities existed on the surface or in the proximity of the surface of the particles, so the template could easily reach the specific recognition sites. When the recognition sites were nearly filled up, the rate of adsorption dropped significantly and the adsorption process achieved equilibrium. During the initial adsorption process (the first 20 min), the SiO₂@MSM-MIP1, SiO₂@MSM-MIP2 and SiO₂@MSM-MIP3 took up MSM molecules from the solution phase at a fast rate. For SiO₂@MSM-MIP1 and SiO₂@MSM-MIP2, adsorption reached equilibrium at 30 min. However, for SiO₂@MSM-MIP3, adsorption reached equilibrium at 60 min. One main reason was that the imprinted layer of SiO₂@MSM-MIP3 was thicker than that of SiO₂@MSM-MIP1 or SiO₂@MSM-MIP2. The other reason could be that the residual amino groups of APTS in the imprinted cavities also contributed to the interaction of SiO₂@MSM-MIP3 with MSM molecules. In general, the thicker the imprinted layer is or the more the binding sites are, the slower the binding kinetics is. Thus, the rebinding process of SiO₂@MSM-MIP3 reached equilibrium in a longer time.

Fig. 7 Adsorption time curves of SiO₂@MSM-MIP1 (▲), SiO₂@NIP1 (▼), SiO₂@MSM-MIP2 (■), SiO₂@NIP2 (●), SiO₂@MSM-MIP3 (◀) and SiO₂@NIP3 (▶) to MSM. The amounts of rebinding MSM were measured by suspending 20 mg of materials in 5 mL of 2 mmol L⁻¹MSM solution at different adsorption time.

Binding selectivity. The selectivity of SiO₂@MSM-MIP3 was measured using the structure analogues CS, BSM, TSM and TBM. It can be seen from Fig. 8 and Table 5 that the SiO₂@MSM-MIP3 exhibited good adsorption selectivity for MSM. The adsorption capacity of SiO₂@MSM-MIP3 to MSM was higher than those of SiO₂@MSM-MIP3 to the analogues. The selectivity factors of SiO₂@MSM-MIP3 to CS, BSM, TSM and TBM were 1.46, 1.73, 1.59 and 2.25, respectively. In the rebinding process, many specific recognition sites for MSM were situated on the surface of imprinted particles, so the MSM was strongly bound to the polymer. As the competitive compound, CS molecule possessed almost identical molecular dimension and functional groups with MSM except the substitute on C–2. Thus, SiO₂@MSM-MIP3 also exhibited a large rebinding affinity to CS. For BSM molecule, a larger substitute on C–1, the volume of BSM was larger than that of MSM. So, the imprinting cavities were not complementary to BSM, leading to a lower rebinding capacity. But for the structure of TSM, the difference was that the phenyl group of MSM was replaced by the thiazole group. The imprinting cavities were also not complementary to TSM, leading to a lower rebinding capacity. For TBM, the hydrogen atom of one amino group was substituted with methyl, leading to fewer recognition sites, so the rebinding capacity was only 46.17 μmol g⁻¹.

Fig. 8 Binding selectivity of SiO₂@MSM-MIP3 and SiO₂@NIP3. The measurements were carried out by suspending samples (20 mg) in chloroform (5 mL) with MSM, CS, BSM, TSM and TBM individual solutions at the 2 mmol L⁻¹ level. Before the measurement, the samples were incubated for 60 min at room temperature on a reciprocating shaking-table.

Table 5 Binding capacities of SiO₂@MSM–MIP3 and SiO₂@NIP3 particles to MSM, CS, BSM, TSM, and TBM (n = 5)^a

Compounds	Q MIP (μmol g^−1) (RSD%)	Q NIP (μmol g^−1) (RSD%)
a The individual solution concentration of MSM, CS, BSM, TSM, and TBM used for binding capacities was 2 mmol L^−1.
MSM	117.0 (1.5)	31.24 (3.2)
CS	93.10 (2.0)	36.41 (2.9)
BSM	61.24 (2.4)	28.40 (3.3)
TSM	77.11 (1.9)	32.85 (3.1)
TBM	46.17 (2.8)	27.73 (3.4)

---

In contrast, the rebinding capacities of SiO₂@NIP3 to target analytes were much lower than those of SiO₂@MSM-MIP3, and SiO₂@NIP3 did not exhibit the obvious difference in the rebinding capacities to MSM, CS, BSM, TSM and TBM. Thus, the small binding capacities of SiO₂@NIP3 to target compounds mainly depended on the physical adsorption. These comparisons further demonstrate that the MSM-imprinted layer at the surface of silica particles have high molecular selectivity for the target species.

3.6 Selective enrichment of MSM from the spiked samples

To assess the application of SiO₂@MSM-MIP3 to selective enrichment of MSM, the wheat, soil and water samples spiked with MSM (0.060 μmol L⁻¹) were analyzed. Simultaneously, SiO₂@NIP3 and commercial C18 were also investigated for comparison. As can be clearly seen from Fig. 9d, no MSM was detected in the three spiked samples due to low concentration. After the dispersive solid-phase extraction by SiO₂@MSM-MIP3, the peak of MSM sharply increased (Fig. 9a) and the enriched concentration of MSM was high enough to be quantitatively analyzed. Under the same conditions, SiO₂@NIP3 and commercial C18 did not show a selectively enriched ability for MSM due to no specific binding sites existing on them, so only a small quantity of MSM was detected after extraction by SiO₂@NIP3 (Fig. 9b) and commercial C18 (Fig. 9c). Most of the MSM was washed during the chloroform cleaning up process, thus, higher sensitivity can be achieved by using SiO₂@MSM-MIP3 as the extraction sorbent. Furthermore, recoveries of MSM in the spiked water, soil and wheat samples were 93.4%, 86.8%, 88.4%; the RSDs were 1.5%, 3.2%, 2.7%, respectively. The results clearly demonstrate that the SiO₂@MSM-MIP3 can be effectively applied in the enrichment and detection of trace MSM in environment and foodstuff.

Fig. 9 HPLC chromatograms of spiked water, soil and wheat samples. (a) extracted with SiO₂@MSM-MIP3; (b) extracted with SiO₂@NIP3; (c) extracted with commercial C18; (d) spiked solution containing 0.060 μmol L⁻¹MSM. Experimental conditions: 5.0 mL of spiked solutions; 20 mg of SiO₂@MSM-MIP3; wash solution, 3.0 mL of chloroform; eluent solution, 5.0 mL of methanol/acetic acid (9 : 1, v/v); re-dissolve solution, 5.0 mL of methanol. HPLC conditions: mobile phase, methanol/water (65 : 35, v/v), the pH of which was adjusted to 3.0 with acetic acid; flow rate, 1.0 mL min⁻¹; column temperature, 30 °C; UV detection, at 230 nm.

Conclusions

In summary, an effective method was developed for coating silica particles with a molecularly imprinted polymervia a hybrid strategy. As a model, the organic MPTS+(APTS-ACPC) monolayer at the surface of silica particles can not only direct the selective occurrence of imprinting polymerization at the surface of supports, but also drive template molecules into the formed polymer shells during the imprinting process. The resultant SiO₂@MSM-MIP3 particles show some attractive characteristics, such as uniform morphology, controlled shell thickness (up to ∼30 nm) and higher binding capacity or binding affinity. It has clearly been shown that the combination of “grafting from” and “grafting through” methods can more significantly improve the rebinding capacity of the imprinted materials to MSM (up to 146.7 μmol g⁻¹), indicating the success of the hybrid imprinting technique used in this study. In addition, the analytical method based on SiO₂@MSM-MIP3 was successfully applied to MSM analysis in spiked water, soil and wheat samples with higher selectivity and sensitivity. The approach reported herein could be potentially exploited in detecting the other environmentally deleterious chemicals.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No.21075066 and 20875048), and the National 863 High Technology Project of China (No. 2008AA10Z421).

References

1. J. F. Liu, J. B. Chao, G. B. Jiang, Y. Q. Cai and J. M. Liu, J. Chromatogr., A, 2003, 995, 21–28.
2. K. V. Penmetsa, R. B. Leidy and D. Shea, J. Chromatogr., A, 1997, 790, 225–234.
3. A. D. Corcia, C. Crescenzi, R. Samperi and L. Scappaticcio, Anal. Chem., 1997, 69, 2819–2826.
4. E. Ayano, H. Kanazawa, M. Ando and T. Nishimura, Anal. Chim. Acta, 2004, 507, 211–218.
5. S. Polati, M. Bottaro, P. Frascarolo, F. Gosetti, V. Gianotti and M. C. Gennaro, Anal. Chim. Acta, 2006, 579, 146–151.
6. A. J. Krynitsky, J. AOCA Int, 1997, 80, 392.
7. G. Dinelli, A. Vicari and V. Brandolini, J. Chromatogr., A, 1995, 700, 201–207.
8. P. Klaffenbach and P. T. Holland, J. Agric. Food Chem., 1993, 41, 396.
9. S. A. Eremin, I. A. Ryabova, J. N. Yakovleva, E. V. Yazynina, A. V. Zherdev and B. B. Dzantiev, Anal. Chim. Acta, 2002, 468, 229–236.
10. Q. Z. Zhu, K. Haupt, D. Knopp and R. Niessner, Anal. Chim. Acta, 2002, 468, 217–227.
11. J. Bastide, J. P. Cambon, F. Breton, S. A. Piletsky and R. Rouillon, Anal. Chim. Acta, 2005, 542, 97–103.
12. K. J. Tang, S. W. Chen, X. H. Gu, H. J. Wang, J. Dai and J. Tang, Anal. Chim. Acta, 2008, 614, 112–118.
13. K. Haupt, Anal. Chem., 2003, 75, 376A–383A.
14. L. Ye and K. Mosbach, Chem. Mater., 2008, 20, 859–868.
15. E. Yilmaz, K. Haupt and K. Mosbach, Angew. Chem., Int. Ed., 2000, 39, 2115–2118.
16. C. Sulitzky, B. Rückert, A. J. Hall, F. Lanza, K. Unger and B. Sellergren, Macromolecules, 2002, 35, 79–91.
17. F. G. Tamayo, M. M. Titirici, A. M. Esteban and B. Sellergren, Anal. Chim. Acta, 2005, 542, 38–46.
18. M. Quaglia, E. D. Lorenzi, C. Sulitzky, G. Massoloni and B. Sellergren, Analyst, 2001, 126, 1495–1498.
19. M. Quaglia, E. D. Lorenzi, C. Sulitzky, G. Caccialanza and B. Sellergren, Electrophoresis, 2003, 24, 952–957.
20. D. M. Gao, Z. P. Zhang, M. H. Wu, C. G. Xie, G. J. Guan and D. P. Wang, J. Am. Chem. Soc., 2007, 129, 7859–7866.
21. X. G. Hu, J. L. Pan, Y. L. Hu, Y. Huo and G. K. Li, J. Chromatogr., A, 2008, 1188, 97–107.
22. L. Schweitz, Anal. Chem., 2002, 74, 1192–1196.
23. J. J. Ou, X. Li, S. Feng, J. Dong, X. L. Dong and L. Kong, Anal. Chem., 2007, 79, 639–646.
24. Y. Peng, Y. Xie, J. Luo, L. Nie, Y. Chen, L. N. Chen and S. H. Du, Anal. Chim. Acta, 2010, 674, 190–200.
25. M. Ejaz, S. Yamamoto, K. Ohno, Y. Tsujii and T. Fukuda, Macromolecules, 1998, 31, 5934–5936.
26. X. Huang and M. J. Wirth, Macromolecules, 1999, 32, 1694–1696.
27. K. Matyjaszewski, P. J. Miller, N. Shukla, B. Immaraporn, A. Gelman and B. B. Luokola, Macromolecules, 1999, 32, 8716–8724.
28. J. B. Kim, M. L. Bruening and G. L. Baker, J. Am. Chem. Soc., 2000, 122, 7616–7617.
29. S. Edmondson, V. L. Osborne and W. T. S. Huck, Chem. Soc. Rev., 2004, 33, 14–22.
30. W. Feng, J. Brash and S. Zhu, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 2931–2942.
31. O. Prucker and J. Rühe, Macromolecules, 1998, 31, 592–601.
32. S. Mona and P. Sébastien, Nat. Chem., 2010, 2, 811–820.
33. N. B. Zhitenev, A. Sidorenko, D. M. Tennant and R. A. Cirelli, Nat. Nanotechnol., 2007, 2, 237–242.
34. C. D. Grande, M.C. Tria1, M.J. Felipe, F. Zuluaga and R. Advincula, Eur. Phys. J. E, 2011, 34, 15–24.
35. Y. Lyatskaya and A. C. Balazs, Macromolecules, 1998, 31, 6676–6680.
36. Q. Lu, X. M. Chen, L. Nie, J. Luo, H. J. Jiang, L. N. Chen and S. H. Du, Talanta, 2010, 81, 959–966.
37. B. Sellergren and K. J. Shea, J. Chromatogr., A, 1993, 635, 31–49.
38. L. Wu, K. Zhu, M. Zhao and Y. Li, Anal. Chim. Acta, 2005, 549, 39–44.
39. R. H. Schmidt, A. S. Belmont and K. Haupt, Anal. Chim. Acta, 2005, 542, 118–124.
40. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria and M. A. Robb, J. R. Cheeseman, Gaussian 09, Revision A.1, Gaussian, Inc., Wallingford CT, 2009.
41. M. Czaun, L. Hevesi, M. Takafuji and H. Ihara, Macromolecules, 2009, 42, 4539–4546.
42. J. J. Hsu, I. B. Huang, C. C. Hwang and M. C. Wu, Indian J. Chem. Technol., 2011, 18, 7–12.

---