Molecularly imprinting: a tool of modern chemistry for analysis and monitoring of phenolic environmental estrogens

Selection of template

The MIPs are usually synthesized for a specific analytical use that implies the selection of a given template. As expected, the imprinting of the template itself can automatically leave a binding site that is tailored for the template. Suitable candidate templates should be chosen carefully in order to produce the highest number of well-defined binding sites. The criteria to consider when selecting a candidate template are its availability, its cost, and its chemical functionalities defining its ability to strongly interact with functional monomers (Pichon and Chapuis-Hugon 2008). In the non-covalent approach, the interactions are weak; thus, the candidate template should possess multiple functional sites for the purpose of increasing the strength of the template-functional monomer interactions. Ideally, those interactions should be strong so that the recognition mechanism after the synthesis of the MIPs could be enhanced.

Based on the non-covalent interaction in the synthesis of the PEEs-MIPs, the main considerations in the selection of template are the stability, the ability to establish hydrogen bonding, and the absence of polymerizable groups to avoid reaction with the newly free radicals (Figueiredo et al. 2016). In most cases the template is the target molecule that is to be selectively retained by the MIPs, and most of the PEEs-MIPs were synthesized using the same molecule as target and template. For example, Yang et al. (2015) prepared MIPs for selective SPE of eight bisphenols from human urine samples using BPA as template. For difficult-to-achieve template or considering the risks of template leakage during applications, a structural analog can be used as dummy template for the synthesis. This dummy template must resemble the target analyte in terms of size, shape, and functionalities. The resulting MIPs should have the ability to bind the target analyte. However, the dummy approach was not so explored in the considered studies involving MIPs and PEEs. An example is described by Li et al. that the dummy MIPs were synthesized by using tetrabromobisphenol A (TBBPA), whose structure was similar to that of BPA, as the dummy template. The resulting MIPs were evaluated and applied as a sorbent for SPE combined with high-performance liquid chromatography (HPLC) to detect BPA (Li et al. 2013).

Selection of other reagents

To make good MIPs, the selection of suitable functional monomers, cross-linkers, and porogen solvents requires careful consideration (Yu and Lai 2010, Wei et al. 2015). In most of the studies related to the synthesis of PEEs-MIPs for analytical applications, the non-covalent interaction is described (Table 1). In this case, the functional monomer is one of the most important factors, so as to allow the maximization of the template-functional monomer association determining effective molecular recognition. The studies on MIPs provide a detailed list of the functional monomers suitable for the approach (Mahony et al. 2005, Vasapollo et al. 2011), but 4-vinylpyridine (4-VP), methacrylic acid (MAA), and acrylicacid (AA) are the most common choices for synthesizing PEEs-MIPs. In addition, some novel functional monomers were successfully prepared and employed in the synthesis of MIPs selective to PEEs. Pan et al. successfully synthesized 3,4-ethylenedioxythiophene-gold nanoparticles (EDOT-Au) as functional monomers to prepared 4-Octylphenol (OP)-MIPs sensors. The resultant sensors presented high sensitivity, stability, and selectivity in the voltammetric determination of OP (Pan et al. 2015).

The choice of the cross-linker is also important to obtain selective MIPs, because the cross-linker is responsible for the stabilization of the morphology, the binding cavities, and the mechanical stability of the polymers, and the binding capacity of the polymers increases with the degree of cross-linker (Spivak 2005, Fang et al. 2011). Many reviews have highlighted the importance of providing lists of cross-linkers that are suitable for molecular imprinting technology (Gong et al. 2008). Ethylene glycol dimethacrylate (EGDMA), divinylbenzene, and trimethylolpropane trimethacrylate (TRIM) are the most common cross-linkers for MIP preparation. When it comes to the synthesis of MIPs for the analysis of PEEs, EGDMA has been most used as a cross-linker. Tetraethoxysilane (TEOS) has also been significantly used, despite being exclusively suitable for the sol-gel process. Han et al. demonstrated that a multifunctional OP-MIP based on CdTe/CdS quantum dots (QDs), magnetic Fe₃O₄, and graphene oxide (GO) was prepared by sol-gel using TEOS as a cross-linker. The resultant MIPs presented high adsorption capacity and fast sensing rate for OP (Han et al. 2015).

The choice of porogen solvent is also one of the key factors determining effective molecular recognition (Zhang et al. 2009, Mu et al. 2011). An inappropriate selection of porogen solvent may lead to poor selectivity of the resultant MIPs. The ideal porogen solvents should have good dissolution for the template, strengthen the interaction between the template and the functional monomer, and produce large pores to obtain good flow-through property (Tan et al. 2012). MIPs are generally prepared using a moderately polar and aprotic solvent, such as toluene, chloroform, dichloromethane, and acetonitrile as porogen solvent. Therefore, strong polar interactions such as electrostatic interactions and hydrogen bonds for template can take place in these organic media (Pichon and Chapuis-Hugon 2008). Toluene and acetonitrile have been most used as porogen solvent for PEEs-MIPs prepared. A mixture of good and poor solvents with proper ratios is always adopted to tune solubility, polarity, and the ability to generate pores to a satisfactory degree. Tarbin and Sharman (2001) prepared Hexestrol (HES)-MIPs for application in the trace residue analysis area using the mixture solvents of cyclohexanol and dodecanol as porogen solvent.

Polymerization protocol

Different methods of polymerization have been used to obtain MIPs. Conventionally, MIPs are prepared by the traditional free radical polymerization. In this approach, all species, template, functional monomer, cross-linker, and radical initiator are dissolved in the porogen solvent. Prior to the polymerization, the mixture is filled with nitrogen to remove oxygen whose presence can retard polymerization (Hu et al. 2013). Among the free radical polymerization, the bulk polymerization is the most widely used procedure owing to its simplicity. Using this procedure, the polymer monolith must be crushed, ground, and sieved to obtain useful particles mainly in the 10–25 μm range. However, despite the reliability and simplicity of this method, particles with irregular shape and size are significant drawbacks associated with the bulk technique (Pichon and Haupt 2006). The resultant particles therefore invariably have a heterogeneous size distribution with poor binding-site accessibility for the target analyte.

As alternative to bulk polymerization, several other radical polymerizations have been developed to better control particles size and porosity (e.g. suspension, multi-step swilling, and precipitation polymerization). The surface imprinted technique, grafting a thin imprinting layer on the surface of a supporting substrate (e.g. silica particle, polymer supports, and magnetic nanoparticles), is another useful way to improve the MIP preparation. The resultant polymers reveal high binding capacities, fast mass transfer, and rapid binding kinetics due to the easy accessibility to recognition sites and the homogeneous distribution of binding sites (Jiang et al. 2015). Recently, grafting an imprinting layer onto the surface of nanoparticles to increase the interfacial surface area has attracted a great deal of attention. Xie et al. prepared the Diethylstilbestrol (DES)-MIPs by a surface imprinting strategy using magnetic nanoparticles as supporter. The obtained MIPs exhibited not only outstanding magnetic property but also high adsorption capacity and selectivity for four EDCs (Xie et al. 2016).

In the recent years, controlled/living radical polymerization (CLRP) has shown great potential to synthesize nanostructured films with well-controlled molecular weight and composition (Fischer 2001, Chen et al. 2011). Benefiting from the negligible chain termination and slower rate, the CLRP can overcome the intrinsic limitations of the traditional free radical polymerization (Bompart and Haupt 2009, Hu et al. 2012). Due to its versatility and simplicity, reversible addition-fragmentation chain transfer (RAFT) polymerization is an ideal candidate for the CLRP method, where a wide range of functional monomers can be employed and the use of transition metal catalysts can be avoided (Moraes et al. 2013). With its controlled/living property of RAFT polymerization, the well-defined MIP layer can be easily introduced on the surface of support and thus significantly facilitates the applications in complicated samples (Li et al. 2010b, He et al. 2014). Recently, MIPs with different formats have been successfully synthesized via RAFT polymerization. Chen et al. demonstrated that the novel DES-MIPs with a homogeneous molecularly imprinted shell were successfully synthesized by surface-initiated RAFT polymerization. This material reveals high potential for the efficient enrichment of EDCs from complicated matrices (Chen et al. 2015).

Some other methods have also been introduced to prepare MIPs. In the sol-gel process, MIPs are prepared by the hydrolysis and polycondensation of silane-based functional monomers such as tetramethoxysilane or TEOS, which form highly cross-linked silica materials. Unlike the method of radical polymerizations, the sol-gel process can be performed in aqueous media under a mild thermal condition (Dai et al. 2014, Rezaei et al. 2014). The resultant MIPs are thermally stable, structurally porous, and extremely rigid, due to the high degree of cross-linking. Rao et al. (2014) prepared dummy OP-MIPs based on multi-walled carbon nanotubes (CNTs) by a sol-gel technique, which were applied to rapidly separate and detect trace 4-Nonylphenol (NP) residues in environmental water samples with high selectivity and sufficient adsorption capacity. Electrochemical imprinting is also an appealing way to prepare MIPs as it is simple, fast, and easy controlled. Pyrrole and phenylenediamine easily form polymer; therefore they are usually used as functional monomers for electrochemical imprinting. For example, Long et al. demonstrated that a novel molecularly imprinted electrochemical sensor was developed for OP determination using cobalt-nickel bimetallic nanoparticles decorated graphene as a sensitized element to modify the carbon electrode. This proposed sensor showed excellent sensitivity and selectivity toward OP in the real sample (Long et al. 2015).

Nanomaterial MIPs as sorbent for SPE

Over the years SPE has been widely employed for the isolation and preconcentration of target analyses, due to its simplicity and versatility. The widespread application of MIPs as SPE sorbents (MISPE) has attracted many researchers’ attention (Hu et al. 2013, Wackerlig and Schirhagl 2016). In MISPE, the sorbent exhibits high selectivity toward the template and its structural analogs, resulting in the desired compounds being extracted from complicated matrixes while other compounds are not retained. Now, there have been numerous original papers on MISPE applications in the analysis of PEEs. Conventional MISPE sorbents present some limitations such as slow mass transfer or incomplete template removal which negatively affects the number of available sites for rebinding. New initiatives to overcome those limitations have been reported such as the combination of the imprinting technique with nanomaterials (NMs) for faster rebinding kinetics and higher binding capacity. NMs are able to play prominent roles to improve the sensitivity, increase the analytical throughput, and simplify the analytical procedures, but there still are some challenges associated with the application of NMs. First, the novel properties of NMs are a two-edged sword. The same properties that generate exciting benefits in use may also pose significant adverse health effects. Second, more efforts should be made to explore further the potential of NMs in the analysis of complex samples. Finally, the technology of synthesis is still an obstacle hindering the wide application of NMs (Liu et al. 2014).

With the development of MISPE, a technique based on magnetic particles has received increasing attention (Figure 3) (Xie et al. 2015a, Gao et al. 2016). The MISPE sorbent can not only easily isolate samples using an external magnet to replace the centrifugation and filtration steps in a convenient and economical way, but also display higher adsorption ability and excellent recognition selectivity. Co-precipitation is the easiest way to synthesize Fe₃O₄ magnetic nanoparticles by adding a base to aqueous Fe²⁺/Fe³⁺ salt solutions under nitrogen atmosphere and vigorous stirring. Then, the magnetic nanoparticles are usually protected by a coating (e.g. silica, surfactant, metal, polymer and carbon) to prevent oxidation by oxygen. Among these coatings, silica coating is especially interesting because it is easy to prepare (e.g. using the sol-gel process) and is favorable for further functionalization. Hiratsuka et al. (2013) demonstrated that magnetic MIPs for BPA and its structural analogs, which have both magnetic and molecular recognition properties, had been developed for magnetic MISPE of BPA in river water samples.

Figure 3:

The synthetic strategy for MIPs based on magnetic particles. This figure was reprinted from Xie et al. (2015) with permission from Elsevier.

Carbon NMs comprise a number of allotropes. Among them, CNTs and graphene have been widely used in preparation of MIPs due to their large surface area, excellent electrical and optical properties, and good chemical/physical stability (Prasad and Singh 2015, Zhai et al. 2015). CNTs can be prepared by various methods, such as chemical vapor deposition (CVD), arc discharge, and high-pressure carbon monoxide disproportionation. Some of them are also used for graphene (e.g. CVD and arc discharge). Moreover, graphene can be exfoliated to GO by harsh oxidation and sonication, followed by reduction of the GO to the parent graphene state. Notably, GO is highly soluble in water and possesses a large number of functional groups (e.g. carboxyl and hydroxyl) (Liu et al. 2014). Zhang et al. prepared novel magnetic CNTs-MIPs for the selective extraction of BPA. Combined with HPLC detection, the MIPs were successfully applied as adsorption materials for magnetic extraction and determination of BPA in environmental water samples (Figure 4) (Zhang et al. 2014). Han et al. (2015) prepared magnetic GO-based MIPs for selectivity recognition and separation of DES, which possessed excellent adsorption capacity and outstanding selectivity for DES.

Figure 4:

The protocol for synthesis of MIPs based on multiwalled CNTs. This figure was reproduced from Zhang et al. (2014) with permission from Elsevier.

MIPs as electrochemical sensors

Electrochemical sensors are widely employed for their high sensitivity and simplicity, which consists of an electrochemical transducer (chemically modified electrode) coated with a chemical or biochemical film as conducting material (Lawal 2016). PEEs are a class of electrochemically active molecules due to phenolic groups, but the direct determination is difficult because of the weak response of PEEs in conventional electrochemical sensors. Hence, further efforts have been tried to increase the surface area of the electrode to enhance oxidation signals. MIPs have been used as recognition elements in the design of sensors due to their higher thermal stability, selectivity, and reusability, which show advantages such as low cost of manufacture, extended lifetime, easy storage, and the capacity of being applied in critical conditions. However, in most cases electrochemical sensors are still not sufficient for practical applications in routine analysis to compete with reported separation methods, and commercial exploitation of molecular imprinting is still in its infancy. Furthermore, MIPs cannot yet provide a total replacement for biological molecules in terms of capacity, selectivity, and homogeneity of binding affinity (Tiwari and Prasad 2015).

Numerous MIP sensors have been developed for the detection of PEEs. The association of the MIPs with the transducer can be achieved in different ways: in situ electropolymerization, surface coating, chemical coupling of the MIPs, physical entrapment of MIP particles into gel or membrane, etc. (Suryanarayanan et al. 2010). As an example, Zhang et al. developed a [Ru(bpy)₃]²⁺-mediated photoelectrochemical sensor for BPA detection using a MIPs-modified SnO₂ electrode. The MIP membrane was assembled on the electrode by a convenient electropolymerization method. This sensor is highly sensitive, selective, and of low cost, and can be theoretically extended to the analysis of organic molecules that can be photoelectrochemically oxidized by Ru³⁺ (Figure 5) (Zhang et al. 2015).

Figure 5:

Schematic diagram of MIPs based on photoelectrochemical sensor. This figure was reprinted from Zhang et al. (2015) with permission from Elsevier.

Integration of NMs into electrochemical MIP sensors can often dramatically improve the analytical performance. The NMs can enlarge the surface area of the electrode, accelerate electron transfer, adsorb the target pollutants, immobilize bio-active molecules, and serve as electrochemical probes (Li et al. 2010a). Generally, CNTs are applied in electrochemical MIP sensing of PEEs. Chen et al. reported a novel electrochemical sensor for BPA through the combination of the MIPs with the multi-walled CNT paste electrode. The imprinted sensor for BPA exhibited good selectivity, stability, and reproducibility and was successfully used for the determination of BPA in real water samples (Chen et al. 2014). Graphene has also been employed in electrochemical MIP sensing of PEEs. Zhang et al. synthesized magnetic-GO MIP composite, which was used as a new electrode matrix for BPA measurement. The electrochemical sensor shows good sensitivity, selectivity, and reproducibility and proved to be an accurate and reliable method due to the determination of BPA in real samples with satisfactory results (Zhang et al. 2013).

Other applications

Besides the applications mentioned above, MIPs have also been applied in some other techniques to facilitate the analysis of PEEs. In QDs, combining the high selectivity of molecular imprinting technology and excellent fluorescent characteristics of QDs could develop a new method for target analyte recognition (Huang et al. 2015). The QDs-encapsulated molecularly imprinted mesoporous silica nanoparticle was widely used for the recognition of PEEs (Kim et al. 2012). MIPs-solid-phase microextraction has also been widely applied to the analysis of PEEs (Tan et al. 2009), which integrates sampling, extraction, concentration, and injection into one single step, and is simple, time efficient, and solvent free. The stir bars coated with MIP films could selectively extract PEEs from complicated matrixes (Zhan et al. 2012). Compared with the ordinary stir bars, MIP-coated stir bars showed better selectivity and higher enrichment capability.