Analytical and Bioanalytical Chemistry

, Volume 402, Issue 5, pp 1827–1846

MIP sensors – the electrochemical approach

Authors

    • Laboratorio di Chimica Analitica, Dipartimento di Scienza dei MaterialiUniversità del Salento
  • Elisabetta Mazzotta
    • Laboratorio di Chimica Analitica, Dipartimento di Scienza dei MaterialiUniversità del Salento
  • Rosaria A. Picca
    • Laboratorio di Chimica Analitica, Dipartimento di Scienza dei MaterialiUniversità del Salento
  • Alessandro Poma
    • Cranfield HealthCranfield University
  • Iva Chianella
    • Cranfield HealthCranfield University
  • Sergey A. Piletsky
    • Cranfield HealthCranfield University
Review

DOI: 10.1007/s00216-011-5405-5

Cite this article as:
Malitesta, C., Mazzotta, E., Picca, R.A. et al. Anal Bioanal Chem (2012) 402: 1827. doi:10.1007/s00216-011-5405-5

Abstract

This review highlights the importance of coupling molecular imprinting technology with methodology based on electrochemical techniques for the development of advanced sensing devices. In recent years, growing interest in molecularly imprinted polymers (MIPs) in the preparation of recognition elements has led researchers to design novel formats for improvement of MIP sensors. Among possible approaches proposed in the literature on this topic, we will focus on the electrosynthesis of MIPs and on less common hybrid technology (e.g. based on electrochemistry and classical MIPs, or nanotechnology). Starting from the early work reported in this field, an overview of the most innovative and successful examples will be reviewed.

Keywords

Molecularly imprinted polymersElectrochemistrySensorsNanotechnology

Introduction

Molecular imprinting [1] (MI) is an exciting field of science recruiting more and more scientists. Entering “molecularly imprinted polymers” (MIPs) as search term in the SciFinder database gives more than 4000 items in the time span 1984–present [2], with exponential growth of original papers published on this topic (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig1_HTML.gif
Fig. 1

Literature on MIPs from a search carried out in the SciFinder database [2] in the time span 1984–present

This trend may be explained by the high demand of research and the market for materials alternative to natural receptors (enzymes, antibodies, etc.) in applications in which selectivity is needed (e.g. separation science, assays and sensors, catalysis) [3]. In particular, MIPs respond to the needs for robust, versatile, cost effective systems [4, 5].

In brief, an MIP may be generally prepared from functional and cross-linking monomers that co-polymerise in the presence of a target molecule (the imprint), which acts as a molecular template. According to the type of binding interactions (covalent or non-covalent) that occur during pre-assembly, two main approaches can be distinguished [6, 7]. The subsequent step of template removal leaves complementary cavities which should, in theory, be capable of rebinding (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig2_HTML.gif
Fig. 2

General scheme of MIP preparation

It should be pointed out, however, that MIPs prepared by bulk polymerisation have proved to be poorly compatible with transducers even if ground before use. In fact, this approach may be limiting in terms of mass transfer and rebinding kinetics in sensor applications [8]. Moreover, a general problem which should be addressed in this case is the integration between recognition element and transducer [9]. This is also an explanation for the less extensive use of MIPs in sensors in comparison with other fields, for example separation science. As a result, novel strategies are being continuously developed to overcome these drawbacks promoting formation of selective binding sites closer to the sensor surface and/or enhancing the diffusion kinetics. To this end, several authors have proposed the in-situ polymerisation of imprinted thin films by different methods (spin-coating [10], layer-by-layer deposition [11, 12], electropolymerisation [13], grafting [14]), the preparation of imprinted micro/nanobeads [15], and more recently also mixed approaches based on nanotechnology [16] and other techniques [17, 18]. However, MIPs synthesis and their assembly in sensors are still key issues of MI technology [19]. These aspects are closely related to the selected transduction system.

Throughout the last decade, many reviews have been devoted to MIP-based sensors in general [4, 2023] or to a specific class of MIP sensors (e.g. electrochemical sensors) [9, 2426]. Although reviews have also addressed the application of peculiar syntheses of molecularly imprinted materials in sensor development [17, 2729], the electrochemical methodology has not been considered. This review is focused on MIP-based materials prepared by the electrochemical approach. Under this topic we would include electrosynthesis of MIPs and its combination with up-to-date technologies (e.g. nanotechnology) and strategies based on the use of electrochemistry to integrate chemically prepared MIPs and the transducer surface.

Reports on (selective) responses of electrosynthesised polymers to species used for polymer preparation (i.e. monomers, electrolytes) or to structurally related species have been reported since 1988. Dong et al. [30] evaluated the selectivity of potentiometric response to chloride of Cl-doped polypyrrole (PPy) grown in the presence of LiCl electrolyte. Similarly, Boyle et al. [31] in the final part of a short paper described the potentiometric response to adenosine triphosphate (ATP) of electropolymerised ATP-doped PPy films on Pt. In both cases, however, no imprinting process was invoked or demonstrated, and the selectivity of the response was not evaluated. More attention was paid by Vinokurov et al. [32, 33] to response selectivity of synthesised polyanilines (homopolymers and copolymers with phenols) toward the corresponding aromatic amines and phenols, establishing theoretical and experimental relationships between selectivity coefficients and voltammetric peak potentials. Potentiometric response of PPy to pyrrole was also reported by Vinokurov [33].

The first example of apparent coupling MI technology and electrochemically synthesised polymers was described by Hutchins and Bachas [34]. They developed a nitrate potentiometric sensor by “electrochemically mediated molecular imprinting”, i.e. by electrosynthesising PPy in NaNO3 aqueous solution. However, no deliberate removal of nitrate from prepared doped PPy was applied before the sensing tests, as is usually done in every MIP procedure. They showed how electrochemical factors may affect selective nitrate recognition, and reported strong interference from thiocyanate and hydroxide anions.

Spurlock et al. in 1996 [35] exploited an electrochemical MIP procedure consisting in the incorporation of neutral (adenosine and inosine) or charged (ATP) templates in PPy films during synthesis on glassy carbon electrodes, followed by overoxidation. The authors expected that this last step could remove template (either neutral or charged), without modifying binding sites. Selected templates were structurally related to adenosine which was the analyte of prime interest. Unfortunately, the effect of templating on film selectivity and sensitivity was small in comparison with the not-imprinted counterpart.

A more significant attempt to prepare an MIP by electrosynthesis was reported by Panasyuk et al. [36] who electrochemically deposited poly-[Ni(protoporphyrin IX) dimethyl ester] in the presence of nitrobenzene. However, a slightly appreciable specific response was achieved only for nitrobenzene at concentrations one hundred times higher than that employed in synthesis.

It was in 1999 that the electrochemical approach was successfully used to develop MIP-sensors by two different groups independently [13, 37]. Malitesta et al. [13] proposed the electropolymerisation of o-phenylenediamine (OPD) in the presence of glucose (neutral molecule) in an aqueous environment on a gold-coated quartz crystal to assemble a piezoelectric sensor. An insulating polyphenol layer imprinted with phenylalanine (electrosynthesised on a thiol-modified gold electrode) was developed as a capacitive sensor by Panasyuk et al. [37]. In the same year, Deore et al. [38], in a rapid communication, reported the electrochemical preparation of PPy imprinted with l-glutamate anion. Also in this case removal of the template was accomplished by overoxidation. The polymer chiral recognition capability was demonstrated by voltammetric and electrochemical quartz crystal microbalance (EQCM) experiments.

Starting from these challenging results, the electrochemical approach coupled to MIP has constantly been exploited and we want to highlight here the general aspects of this fascinating topic and its evolution over recent years.

Electrosynthesised MIP-based sensors

This section presents an overview of typical monomers used in the electrosynthesis of MIPs for sensor applications.

Pyrrole

One of the most widely used polymers in electrochemical imprinting is polypyrrole (PPy) (Table 1). As discussed in the Introduction, early attempts to imprint PPy can be seen in the development of electrochemical sensors for anionic species based on the memory of PPy for the anion of the electrolyte used in its preparation [30, 31]. Potentiometric sensors for chloride [30] and for ATP [31] were prepared in this way, although no selectivity test was performed and even the sensing mechanism was still considered unclear by the authors. Using the same approach, some years later, Hutchins and Bachas [34] developed an electrochemically mediated imprinting of polypyrrole for the potentiometric detection of nitrate but also other anionic species, because the sensor response was not specific. It should be noted that in these cases the anion “template” was left in the polymeric matrix for additional interaction with the anions [35]. Polymer recognition properties were created by entrapment of the anion within the polymer during electropolymerisation—leading to doped PPy—and to the creation of ion sieving channels when the polymer passes from the conducting (oxidised doped film) to the non-conducting (reduced de-doped film) form [39].
Table 1

MIP sensors based on electrosynthesised PPy

Sensing layer

Template/Analyte

Transduction

Ref.

oPPy

Adenosine, inosine, ATP

Voltammetric

[35]

oPPy

l-Glutamic acid

Voltammetric, EQCM

[38, 40]

oPPy

l-Aspartic acid

EQCM

[41]

oPPy

l-Tryptophan

EQCM

[42]

oPPy

1-Naphthalenesulfonic acid

EQCM

[47]

oPPy

Taurocholate

EQCM

[48]

oPPy

ATP

Amperometric

[49]

oPPy

2,4-Dichlorophenoxyacetic acid

Voltammetric

[50]

oPPY

Paracetamol

Voltammetric

[51]

oPPY

Sulfamethoxazole

Voltammetric

[52]

PPy

Caffeine

QCM

[53, 54]

PPy

Sodium dodecyl sulfate

QCM

[55]

PPy

Trichloroacetic acid

Voltammetric, conductometric, QCM

[56]

PPy

Caffeine, theobromine

Amperometric

[57]

PPy

Ascorbic acid

Voltammetric

[58]

PPy

Zearalenone

SPR

[59]

PPy

Caffeine

Voltammetric

[60]

PPy

Bovine leukaemia virus glycoprotein

Amperometric

[61]

PPy/Poly(3-methylthiophene)

Bacillus endospores

Voltammetric, impedance spectroscopy

[62]

Poly(N,N-ethylenebis[N-[(3-(pyrrole-1-yl)propyl) carbamoyl)methyl]glycine])

Cd(II)

Voltammetric

[63]

Poly(ethylenediaminetetra-N-(3-pyrrole-1-yl)propylacetamide)

Cu(II), Hg(II)

Voltammetric

[64]

This approach has been further explored leading to the development of electrosynthesised imprinted films based on overoxidised PPy (oPPy) which were successfully applied for the detection of large anionic templates. In this imprinting scheme, the anionic template is entrapped within the polymer during electrosynthesis and subsequent overoxidation replaces the conventional washing step as it determines the template expulsion and the creation of complementary nanocavities within the polymer [26].

Overoxidation actually removes the net positive charge on the polymer by a complex mechanism, enabling release of the anionic template. At the same time the introduction of oxygen-containing carbonyl and carboxyl groups conferred peculiar permselectivity properties on the film, enabling selective recognition of the template [26].

A further development in this direction was made by Spurlock et al. [35] who prepared oPPy templated with charged and neutral molecules, for example adenosine, inosine and ATP. These authors, however, did not observe a remarkable imprinting effect, because little success was achieved in the recognition of these template molecules.

The first successful example of imprinting of oPPy was reported by Deore et al. [38, 40] who demonstrated the enantioselectivity of imprinted oPPy for l-glutamic acid. The model proposed to explain the selective uptake of the template is schematically represented in Fig. 3. Template entrapment/removal and subsequent rebinding were based on acid–base equilibria of glutamic acid. The key role of pH in promoting the selective interaction between oPPy and the cationic template was in fact verified.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig3_HTML.gif
Fig. 3

Mechanism of l-glutamic acid de-doped PPy and overoxidation (A) and subsequent removal/rebinding to oPPy (B). Reproduced, with permission, from Ref. [40]. Copyright 2000 American Chemical Society

The same scheme based on the effect of pH in modulating the interaction between oPPy and the template in the ionic form has been applied to the detection of analytes such as l-aspartic acid [41] and l-tryptophan [42] and has been extended also to chemically synthesised colloids of oPPy imprinted with lactate for the enantioselective uptake of l-alanine [43, 44] and with 1-naphthalenesulfonate [45] and also to chemically prepared oPPy nanowires imprinted with glutamic acid [46]. In other cases [4750], the template was present in the anionic form both in PPy electrosynthesis and in the rebinding steps, and the specific interaction with the MIP was due mainly to the precisely controlled shape-complementary cavity created by the overoxidation process. A slightly different binding mechanism was explored for the electrosynthesis of oPPy for paracetamol [51] and sulfamethoxazole [52]: the template was neutral and interacted with PPy through hydrogen bonding between the N–H groups of the pyrrole units and some functionalities of the template molecule (S = O group and amino group of sulfamethoxazole; C = O group, hydroxyl group, and N–H group of paracetamol). Nevertheless, the role of overoxidation in determining the removal of the neutral template was not exhaustively explained by the authors.

It is important to mention here that the use of overoxidised PPy requires an additional step in the imprinting scheme, not only enabling template removal but also modifying to some extent the chemical properties of the polymer leading to partial degradation of the PPy polymeric backbone and to introduction of oxygen-containing carbonyl and carboxyl groups into the MIP moiety that are responsible for its ability to recognise the imprinted molecule [26].

A more conventional imprinting strategy based on conducting PPy has been explored for the development of selective materials for caffeine [53, 54], surfactant [55], trichloroacetic acid [56], theobromine [57], ascorbic acid [58], and the mycoestrogen zearalenone [59]. The imprinting effect in these cases was because of the formation of monomer–template complexes through hydrogen and/or ionic interactions. Conventional washing procedures were used for subsequent removal of the template. Recently a novel application of imprinted PPy has been explored [60] proving the possibility of modulating MIP thickness and thus the number of imprinted sites for the template under electrical stimuli. The proposed mechanism was based on electrochemical control of the degree of swelling of the polymer associated with ion transport in and out of the polymer. This aspect could have great potential for the development of smart sensors where the sensitivity needs to be dynamically optimised for a specific test solution.

Electrochemical molecular imprinting technology based on PPy has been extended also to the bio-imprinting field for label-free detection of biological macromolecules [61, 62]. By this strategy, a receptor for the amperometric detection of bovine leukaemia virus glycoprotein has been developed [61]. The suitability of electrosynthesised PPy for this purpose was verified, although much work was still required for elimination of non-specific interactions and for enhancement of sensor performance. Another attempt to combine electrochemistry with bio-imprinting led to the fabrication of an imprinted composite film based on PPy and poly(3-methylthiophene) for detection of Bacillus endospores [62]. The work provided proof-of-concept in the fabrication of microbial imprinted films using conducting polymer composite membranes, although in this case also the selectivity issue was not addressed. The literature on this topic is relatively scarce so far, although it is clear that these preliminary efforts could open the way to the design of innovative electrosynthesised MIPs as receptors to selectively recognise biological macromolecules and microorganisms.

Some efforts have also been made in the electropolymerisation of pyrrole derivative monomers in the ion imprinting strategy [63, 64]. Modified electrodes based on a complexing imprinted poly(pyrrole-EDTA like) film were prepared and used for the electrochemical detection of metal cations. It is worth mentioning that an interesting approach based on the electropolymerisation of metal complexes to give polypyrroles-containing cavities had already been proposed by Bidan et al. [65, 66]. They prepared metal complexes from ligands bearing pyrrole rings and, after demetallation, the reloading of different metals was followed by cyclic voltammetry. A specific templating effect of the metal centre was observed as the demetallated polymer had higher loading than the film prepared in the absence of the metal. The ligand cavity was thus preserved, although selectivity among different metals was very limited.

o-Phenylenediamine: imprinted homopolymers and copolymers

Although the volume of published work on OPD cannot be compared with that on pyrrole, this molecule (also known as 1,2-benzenediamine) is a well established monomer in electrosynthesised imprinted polymers. As already discussed, the homopolymer (PPD) based on OPD was one of the first successful examples of electrosynthesised MIPs [13]. Afterwards, other works were proposed on PPD [54, 6776] and on the copolymerisation of OPD with other electroactive monomers [7784]. Examples of MIP sensors based on the electropolymerisation of this monomer are reported in Table 2. PPD films grow compact and rigid and are regarded as mechanically stable. The typical preparation occurs by cyclic voltammetry in aqueous buffered solutions at different pH, although acetate buffer at pH 5.2 is the most used [13, 67, 75, 76]. Following detection of glucose by use of a piezoelectric sensor [13], some authors applied this strategy to other compounds (phenylalanine [67], sorbitol [69], caffeine [54]). In particular, the possibility of imprinting an amino acid instead of its derivative [67] (as is often performed in classical MIP preparation [7]) was noticeable. On the other hand, a very interesting feature of PPD is its nonconducting nature at pH 5.2, which makes it a good candidate in MIP sensors based on capacitance changes [68, 70, 71]. In fact, in 2001 Cheng et al. [68] reported the first capacitive sensor with a PPD layer imprinted against glucose. A slight but significant variation of the procedure in Ref. [13] was proposed to improve the insulating property by immersing the modified electrode in a solution of long-chain 1-dodecanethiol overnight [68]. Results were comparable with those in previous work [13] in terms of linear range and sensitivity. An “equilibration time” of 20 min was required to obtain a stable response. The need for good insulation in capacitive sensors led to a further modification of the approach in Ref. [68]. As a result, in Ref. [70] a preliminary step was introduced for modification of the gold surface by a self-assembled monolayer of 2-mercaptoethane sulfonate (MES) before carrying out OPD polymerisation in the presence of glutathione (the analyte) and coating with 1-dodecanethiol. The response in this case was stable after 20 min.
Table 2

MIP sensors based on electrosynthesised homopolymers and copolymers of OPD

Sensing layera

Template/analyte

Transduction

Ref.

PPD

Glucose

QCM

[13]

PPD

Caffeine

QCM

[54]

PPD

Phenylalanine

QCM

[67]

PPD/1-dodecanethiol

Glucose

Capacitive

[68]

PPD

Sorbitol

QCM

[69]

SAM of MES/PPD/1-dodecanethiol

Glutathione

Capacitive

[70]

PPD

O,O-Dimethyl-α-hydroxylphenyl phosphonate

Capacitive

[71]

PPD

Sarin acid/DMMP

SAW

[72]

PPD

Salicylic acid

Voltammetric

[73]

PPD

Triclosan

Voltammetric

[74]

PPD/OTC-enzyme

OTC

Voltammetric

[75, 76]

Poly(OPD/An)

Atropine sulfate

QCM

[77]

Poly(OPD/resorcinol)

2,4-D

Voltammetric

[78]

Poly(OPD/An)

Paracetamol, metamitron, DNOC

Voltammetric

[79, 81, 83]

Poly(OPD/DA)

l-Glutamic acid

Capacitive

[82]

Poly(OPD/resorcinol)

Dopamine

Voltammetric

[84]

aMIP is highlighted in bold if the sensing layer consists of different materials

Few publications have reported electrochemical sensors based on PPD. A recent example on the imprinting of PPD with triclosan to develop a voltammetric sensor was described by Liu et al. [74]. The authors showed how this analyte may be detected by involving a “so-called” indirect method based on the use of ferrocyanide as a probe for testing of the binding between the MIP on the electrode and the target molecule in solution. In fact, when triclosan binds to the complimentary cavities present in PPD a decrease in ferrocyanide diffusion to the electrode surface occurs. In other words, the increment of triclosan concentration corresponds to the overall current decrement of the Fe(CN)64−/Fe(CN)63− redox couple. It should be pointed out that incubation of the MIP sensor in the analyte solutions for 15 min was necessary before detection. When reviewing the performance of imprinted PPD-based sensors it is generally found that stable responses to target analytes may be achieved after rather long times, irrespective of the detection method and the type of molecule tested. This may be explained by possible slow diffusion and binding kinetics, as also stated in other reports on MIP sensors [10, 85]. In particular, the latter factor may be related to site heterogeneity and quite weak interactions involved between binding sites and analyte, as reported in some papers [13, 67]. Despite these problems, PPD has been proposed also for use in a surface acoustic wave (SAW) gas sensor for dimethyl methyl phosphonate (DMMP) [72] and more recently in an elegant strategy for oxytetracycline (OTC) detection [75, 76] in combination with an enzyme amplifier. Li et al. prepared an OTC-imprinted PPD film on gold electrodes and then determined its concentration in samples via an “isolation–incubation–competition” three-step procedure. In detail, after template removal all binding sites were occupied again by immersing the MIP sensor in OTC solution during the “isolation” step, an incubation step was then performed to replace OTC with enzyme-labelled OTC (using either horseradish peroxidase [75] or glucose oxidase [76]) and finally a competition step was carried out by incubating the sensor in OTC samples for a specific time to replace enzyme–OTC with OTC alone. The last event may be followed by differential pulse voltammetry (DPV) as a current decrease in the hydroquinone–hydrogen peroxide–buffer solution (reaction catalysed by both enzymes). The complexity of the scheme, however, may limit this approach to analytes that require extremely high sensitivity.

The OPD-based copolymers were also considered in MIP preparation. The first imprinted copolymer of OPD was prepared with aniline (An), in the presence of atropine sulfate, by Peng et al. [77], for detection of this alkaloid with a bulk acoustic wave sensor. The optimum An/OPD molar ratio was 40:1 indicating that a polyaniline-like structure with OPD insertion was created. Moreover, template removal (and sensor response recovery in general) was achieved only by several washing steps with water and sulfuric acid. Extensive work on the development of voltammetric sensors based on electrosynthesised copolymers of OPD and An was carried out by Barrio’s group [79, 81, 83]. They prepared poly(OPD/An) films on carbon fibre microelectrodes (CFME) to detect paracetamol [79], metamitron [81], and 4,6-dinitro-o-cresol (DNOC) [83] by square wave voltammetry (SWV). The copolymer was generated in sulfuric acid aqueous solution in the first two cases, and 50% (v/v) methanol was also used as solvent [83] because of the poor solubility of DNOC in water. In general, they observed that OPD was responsible for the insulating property of the films whereas An enabled incorporation of the template into the polymeric chain. As in Ref. [77], An/OPD ratios of approximately 30–40:1 were considered optimum in imprinting electropolymerisation. On the other hand, template removal was carried out by contact in stirred media (chosen according to imprint) for less than 1 h followed by rapid “electrochemical washing” in the detection medium (i.e. by cycling the MIP sensor in a suitable potential range for few scans to remove the remaining template) [79, 81]. Besides showing good selectivity of the developed sensors [83] the authors also characterised the binding properties of MIP for DNOC by analysing different adsorption isotherms. They compared typical models proposed in the MIP literature (Langmuir, bi-Langmuir, and Freundlich–Langmuir) by fitting the experimental data. It was found that the Langmuir isotherm was the best model which is associated with a quite homogeneous population of binding sites, while generally non-covalent bulk MIPs have some heterogeneity. This appealing feature may be explained by process control achieved with electrochemical techniques. In particular, the copolymerisation of An and OPD is responsible for the formation of a more rigid and compact film (compared with homopolymers) thus promoting site homogeneity.

Resorcinol was also used as co-monomer. In fact, Weetall and Rogers [78] copolymerised resorcinol and OPD electrochemically (in an equimolar aqueous solution) in the presence of three different molecules (the dyes fluorescein and rhodamine, and the pesticide 2,4-dichlorophenoxyacetic acid, 2,4-D) on graphite electrodes. However, extensive washing in methanol was necessary to remove the templates. Moreover, dye-imprinted films deposited on electrodes were used in “classical rebinding batch experiments” performing dye elution in methanol and recording its fluorescence. 2,4-D MIP-modified electrodes were instead used as voltammetric sensors to evaluate the response to sequential additions of the pesticide (every 15 min) in solution by SWV, though no selectivity study was performed. A subsequent paper by the same group [80] reported a slightly modified preparation of this copolymer in the presence of 2,4-D with enhanced sensor response to the pesticide. However, no template removal (an essential step in molecular imprinting) was performed indicating that a 2,4-D doped copolymer (not MIP) is used in this case. In contrast, a similar copolymer was imprinted with dopamine (DA) and a voltammetric sensor for this neurotransmitter was successfully developed by Song et al. [84]. The need for ultrathin insulating films in capacitive sensors for the enantioselective recognition of l-glutamic acid (Glu) led to the design of a copolymer based on OPD and DA [82]. In particular, OPD provided rigidity to the zigzag chain, rich in functional groups, formed by DA and also ensured smoothness of the MIP surface. Hence, cavities with higher affinity than those in the corresponding homopolymers (PPD and polydopamine) were obtained, giving good responses to Glu.

Phenols

MIP sensors based on phenols are listed in Table 3.
Table 3

MIP sensors based on electropolymerised phenols

Sensing layera

Template/analyte

Transduction

Ref.

SAM of 4-mercaptophenol/Polyphenol/octanethiol-1

Phenylalanine

Capacitive

[37]

Polyphenol

Rifamycin SV

Voltammetric

[86]

Polyphenol

Theophylline

Capacitive

[87]

Polyphenol

Methyl viologen

Voltammetric

[88]

Poly(o-aminophenol)

2,4-Dichlorophenoxy acetic acid

QCM (impedance analysis)

[89]

Poly(m-aminophenol)

Tegafur

Capacitive

[90]

Poly(o-aminophenol)

Dopamine

Voltammetric, amperometric

[91]

Poly(o-aminothiophenol)

Nitrobenzene

Voltammetric

[92]

Poly(o-aminothiophenol)

Metolcarb

Amperometric

[93]

SAM of 1,3-propanedithiol/Polyresorcinol

N-(1-Pyrenyl)maleimide/pyrene

Voltammetric

[94]

aMIP is highlighted in bold if the sensing layer consists of different materials

As mentioned above, one of the first successful examples of electrosynthesised MIPs was based on the imprinting of polyphenol for phenylalanine [37], leading to the development of the first capacitive sensor possessing a receptor layer consisting of an MIP. Since this first application the electrochemical imprinting of polyphenol has attracted increasing research interest, resulting in the assembly of sensing devices for the antibiotic rifamycin SV (RSV) [86], theophylline [87], and methyl viologen [88]. Blanco-López et al. [86] observed that such a non-conductive imprinted film for RSV had selectivity properties based on size-exclusion effects combined with charge differentiation. An interesting approach was used by Willner’s group [88] for imprinting of methyl viologen on a polyphenol film as a new method to generate an MIP. The π donor–acceptor interactions between the electrogenerated film and the template were used as the driving force for formation of the imprinted sites.

Also, phenols bearing an amino group have been shown to be suitable materials for the design of electrosynthesised MIPs [8991], because of the possibility of creating functionalised imprinted cavities thus providing more specific interaction with the template molecule. A first attempt at electropolymerising MIPs from a self-assembled layer of o-aminothiophenol on a gold electrode was reported for the electrochemical sensing of nitrobenzene [92]. Very recently this approach has been further explored leading to the development of an amperometric sensor for the herbicide metolcarb [93]. In this case, it was shown that the application of a cathodic potential to the electrode before electropolymerisation is an effective way to promote monomer–template interactions and ensure monomer proximity at the electrode surface. Another application of functionalised phenols as functional monomers in the design of electrosynthesised MIP was proposed by Luo et al. [94] who investigated the electrochemical detection of pyrene by an imprinted electrochemically deposited polyresorcinol film on a gold electrode. Enhancement of imprinting effect was achieved as a consequence of anchoring pyrene to the electrode surface through a thiol-linked derivative before MIP electropolymerisation.

MIPs based on thiophene derivatives

More recently, thiophene (Th) derivatives have been proposed as functional monomers in MIP electrosynthesis [95105] as summarised in Table 4. In 2005 Yeh and Ho developed an amperometric sensor for morphine (MO) based on a poly(3,4-ethylenedioxythiophene) (PEDOT) film potentiostatically grown in the presence of the alkaloid on an indium tin oxide (ITO) electrode [95]. Imprinted PEDOT was not only a good and stable receptor for MO but also had electrocatalytic behaviour for MO oxidation which occurred at lower potential than on bare ITO. The device was later miniaturised by the same group [96]. In that case, MO-imprinted PEDOT was deposited on to a Pt microelectrode and assembled in a microfluidic system. Moreover, a Nylon 6 membrane imprinted for ascorbic acid (typical interfering molecule) was placed up-stream to the sensor. In this way, sensitivity, detection limit, and response time were significantly improved by miniaturisation and interference limitation. EDOT has also been chosen as comonomer in electrochemical imprinting, because of its hydrophilic properties and ability to polymerise into stable and homogeneous films. In a recent paper [99] EDOT was used with thiophene acetic acid (AAT) to grow a copolymer, poly(EDOT-co-AAT), in the presence of the herbicide atrazine (At). The authors suggested that the functional monomer AAT could interact via the carboxylic moiety with aminic and triazinic nitrogen atoms forming H-bonds, whereas EDOT was considered a conjugated linker to create a stable conducting film with good compatibility with aqueous media. Voltammetric responses to At were recorded using MIP and non imprinted polymer (NIP) poly(EDOT-co-AAT) and the imprinting effect was verified. Similarly, a homopolymer based on AAT was also proposed for imprinting At [100]. Nuclear magnetic resonance (NMR) spectroscopy confirmed the ability of this monomer to bind At effectively. X-Ray photoelectron spectroscopic (XPS) analysis on the electrosynthesised MIP furnished evidence of entrapment of the template in poly(AAT) and its successful removal upon washing. Moreover, it was observed how a cross-linker (3,3-bithiophene) could be helpful in improving the performances of electrosynthesised MIPs.
Table 4

MIP sensors based on electropolymerised thiophene derivatives

Sensing layera

Template/analyte

Transduction

Ref.

PEDOT

Morphine

Amperometric

[95, 96]

poly(EDOT-co-AAT)

Atrazine

Voltammetric

[99]

Poly(Bis(2,2′-bithienyl)-5,5-dimethyl-2-phenyl-[1,3,2]dioxaborinanemethaneco-Bis(2,2′-bithienyl)-benzo-[18-crown-6]methane)

Histamine, adenine

QCM

[97, 103]

Poly(Bis(2,2′-bithienyl)-benzo-[18-crown-6]methane-co-3,3′-Bis[2,2′-bis(2,2′-bithiophene-5-yl)]thianaphthene)

Melamine

QCM

[98]

Poly(meso-(3,4-dihydroxyphenyl)-bis(2,2′-bithienyl)methane-co-Bis(2,2′-bithienyl)-benzo-[18-crown-6]methane)/poly(bithiophene)

Dopamine

QCM

[102]

Polyterthiophenes based on mono-, bi-functional monomers bearing orthogonal moieties (i.e. –COOH, –OH, –NH2)

Theophylline, paracetamol, naproxen

SPR

[101, 104]

Poly(3,5-bis[[2-(2,5-dithiophen-2-ylthiophen-3-yl)ethoxy]phenyl]methanol)

Folic acid

QCM

[105]

aMIP is highlighted in bold if the sensing layer consists of different materials

A very promising approach to the imprinting of polypyrrole/poly(3-methylthiophene) (PPy/PMT) conducting film with Bacillus subtilis endospores was reported by Namvar and Warriner [62]. In this work the endospores were first adsorbed on a glassy carbon electrode upon which a two-step electropolymerisation of PPy and PMT was carried out. Template removal was achieved by soaking in dimethyl sulfoxide. The compactness ensured by PMT was essential to create more efficient binding sites for the spores compared with PPy films. In particular, it was shown that direct monitoring of the species was possible on the basis of the change in film susceptance upon contact with the spore suspension. Alternatively, endospore binding could be monitored after their germination via the release of dipicolinic acid which affects the changes in film charge on cycling between −0.3 and 0.4 V vs. Ag/AgCl. Unfortunately, the main drawback of this system was the inability to restore sensor response after first use. Moreover, selectivity and calibration were not addressed.

The tremendous potential of electrochemical imprinting have been demonstrated by developing novel bis(biTh)-based functional monomers with increased complexation properties toward target analytes [97, 98, 102, 103]. In particular, researchers electrosynthesised advanced MIPs on Pt-coated quartz crystals assembling very sensitive microgravimetric sensors for analytes bearing an amino group (i.e. histamine [97], melamine [98], adenine [103], dopamine [102]). It is worth mentioning that the acid–base equilibrium involving target molecules may help in the creation of more selective MIP cavities and determine the choice of the optimum conditions for template extraction and analyte rebinding. In fact, in the papers cited above one monomer consisted of bis(biTh) bearing a 18-crown-6 substituent (crown-bis(biTh)) which enables interaction with the protonated amine. Thus, the protonation process was essential for effective formation of the monomer–template complex which was achieved by adding an appropriate amount of acid to the non-aqueous polymerisation medium (acetonitrile). Moreover, in this work [97, 102, 103] a second point beneficial for binding was also created by selecting another modified bis(biTh). In this study [97, 103] the second monomer was a bis(biTh) with a dioxaborinane moiety which enables coordination between the lone pair of nitrogen of the imidazole ring of histamine [97] or in the third position of the purinic ring of adenine [103] (as reported in Fig. 4) whereas bis(biTh) modified with benzene-1,2-diol enabled H-bond formation with enediol on dopamine [102].
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig4_HTML.gif
Fig. 4

Acid–base equilibrium of adenine (template) (1). Basic structure of bis(biTh)-based monomers (2) with dioxaborinane moiety (3) or 18-crown-6 substituent (4). Typical interfering species for adenine: 2-aminopurine (5), guanine (6), ascorbic acid (7). Reproduced from Ref. [103], Copyright 2011, with ermission from Elsevier

In these cases, however, a preliminary step of electrosynthesis of poly(bis(biTh)) was essential to prevent template electro-oxidation on bare Pt. Instead, in another paper [98], this film was not grown because the pH of the pre-polymerisation mixture was suitably adjusted to achieve complete protonation of amino groups of melamine thus avoiding its electro-oxidation. Moreover, in this work crown-bis(biTh) was electrochemically copolymerised with 3,3′-bis[2,2′-bis(2,2-bithiophene-5-yl)]thianaphthene as cross-linker to furnish a relatively open but fixed 3D structure to enhance site accessibility in the detection step. A further improvement was achieved by using an ionic liquid (trihexyl(tetradecyl)phosphoniumtris(pentafluoroethyl)-trifluorophosphate) as supporting electrolyte and porogen (a typical feature of classical MIP synthesis) thus improving film porosity and homogeneity. Finally, removal of the templates was completed in all the examples by washing for a long time with a strong base (0.01 mol L−1 NaOH for 12 h) to shift the acid–base equilibrium in order to release the imprints. An important feature of the described sensors was also the lower response time in comparison with previously reported similar MIP sensors [54, 106]. This was explained, in authors’ opinion, by good diffusion of the analytes inside the MIP layer and by the high stability constants (Ks) of MIP–analyte complexes determined for these polymers [97, 98, 102]. The last aspect also favoured sensor selectivity because interfering species had lower Ks.

Even more recently, functionalised terthiophenes were also numbered among monomers for MIP electrosynthesis to sense theophylline, paracetamol, and naproxen by surface plasmon resonance (SPR) [101, 104] and folic acid by QCM [105]. In fact, Advincula’s group used mono and bi-functional terthiophenes bearing orthogonal moieties (e.g. –COOH, –OH, –NH2) and capable of complexing template molecules by H-bonding, generally in the form of a 2:1 monomer:template complex (as also calculated by theoretical modelling). Moreover, terthiophene may polymerise at lower oxidation potentials than mono and bithiophenes. In particular, the authors claim that the process involved resembles cross-linking polymerisation, suggesting that the regular α,α-coupling (2,5 position) of the monomer units results in an ordered conjugated polymer network (CPN) film [101].

It was also observed how monomers bearing –COOH and –OH, compared with –NH2, conferred a higher MIP rebinding ability. MIPs based on bifunctional units also gave a greater response compared with monofunctional ones (although the response time was also significantly higher [104]). Interestingly, the authors used in all their work an extremely long “pre-polymerisation step” which consisted in 24 h (or even 48 h) contact between monomer and template in the synthesis medium before electrochemical polymerisation. The authors considered this strategy essential for successful formation of selective sites. Furthermore, potential-induced template removal at 0.4 V in acetonitrile was successfully applied, reducing the time of the washing step to 10 min. It was hypothesised that an increase in swelling ability of polymer films could lead to an easier release of the imprint. Finally, good selectivity properties were demonstrated in all the cases, indicating the feasibility of this approach. On the other hand, all the systems had an average 15 min response time, which is comparable with that of previous simpler MIPs (see also examples presented in other sections of this review). This behaviour still needs to be fully understood, because several factors, for example film thickness, porosity and MIP–analyte interactions, may play a significant role.

Electrosynthesised MIPs based on other monomers

The electrosynthesis of MIPs has been shown to be a versatile approach in the choice of functional monomers, as shown by the wide range of molecules successfully used to this end. In addition to the commonly used monomers mentioned above, electrochemical imprinting has recently been extended to MIPs electrogenerated from novel monomers bearing both electrosynthesisable moieties and functional groups interacting with the template molecule. Among these, poly(3-aminophenylboronic acid) has been used because of the variety of reversible complexation possibilities afforded by the species, principally through electrostatic and covalent interactions of the B(OH)2 moiety [107]. Sensing devices based on this imprinted material have been developed for d-fructose [108], lysozyme, and cytochrome c [107], and, more recently, also for large molecules such as Staphylococcal enterotoxin B [109] and T-2 toxin [110].

The use of electropolymerised poly(mercaptobenzimidazole) (PMBI) as an imprinted sensing material has been the focus of several papers [94, 111115]. Its application has been proposed for the electrochemical detection of electroinactive molecules, for example cholesterol [111, 112] and pyrene [94]. In contrast with the insulating NIP films, the current response of the imprinted polymers was related to the diffusion of an electrochemical probe (i.e. ferricyanide) into the imprinted cavities. The sensing mechanism was thus based on the decrease of this current in the presence of the analytes. The electropolymerisation of an imprinted PMBI on a self-assembled monolayer of MBI has been explored for fabrication of a capacitive sensor for fenvalerate [113]. Yang et al. also used imprinted PMBI on gold electrodes to detect the herbicide mefenacet [114]. Recently PMBI has been shown to be suitable also for ion imprinting applications able to promote the selective detection of Hg(II) by stripping voltammetry [115].

A novel imprinting strategy combining electropolymerisation with metal-ion coordination has been explored by designing an MIP based on a Co(II)-porphyrin for electrocatalytic detection of a chlorinated pesticide [116]. This approach led to the development of a novel imprinted polymer featuring both MIP selectivity properties and electrocatalytic activity of the metal complex, providing at the same time advantages related to MIPs electrochemical synthesis. Also MIPs based on complexes of Ni(II) and of Fe(III) have been assembled for imprinting of nitrobenzene [36], fenitrothion [117], and dopamine [118] although the ion mediated recognition mechanism was not involved in the imprinting scheme.

Electrogenerated imprinted polyacrylamide membranes have been used as sensing material for detection of a histidine isomer [119], nucleotides, and monosaccharides when copolymerised with acrylamidephenylboronic acid [120]. In the last case, the remarkable selectivity observed was attributed to the high degree of cross-linking of the copolymer and to the occurrence of cooperative binding interactions of the templates with boronic acid ligands and of H-bonds with the acrylamide units. The application of polyacrylamide-based electrosynthesised MIPs has been successful also in the imprinting of large molecules such as haemoglobin [121].

An original imprinting scheme has been proposed [122] involving the electropolymerisation of poly(methylene green) (PMG) in the presence of DNA acting as “pseudo-template”: after DNA removal, the adsorption of DNA aptamers in the imprinting cavities was promoted and the prepared material was used for electrochemical sensing of thrombin because of the occurrence of aptamer–thrombin interactions. The electrosynthesis of PMG had previously been investigated for fabrication of an electrochemical sensor for theophylline [123] by adopting a conventional electrochemical imprinting approach.

Investigation of electrosynthesised MIPs has been extended also to the imprinting of polydopamine for nicotine [124] and of polycarbazole for some drug molecules, for example naproxen, paracetamol, and theophylline [104].

The role of electrochemistry in the integration of chemically synthesised MIPs in sensors

An interesting approach in which electrochemistry meets imprinting technology has led to the design of composite sensing materials based on the entrapment of chemically prepared MIP particles into electropolymerised films [125127]. Kriz et al. [125] investigated the preparation of morphine-imprinted acrylate polymer particles hosting chemically synthesised PPy in their pores. Such composite particles were then immobilised on a gold substrate by PPy electropolymerisation with the objective of achieving close contact between the transducer and the recognition sites within the polymer. Nevertheless, the advantages related to this novel imprinting scheme were not exhaustively proved. In fact, although the electrical conductivity of PPy/MIP particles was verified, the sensing ability in the electrochemical detection of morphine was not shown. The important conclusion was made however, on the basis of competitive radioligand binding, that the morphine-specific molecular recognition properties were not significantly altered compared with pure MIP particles.

The electropolymerisation of a conductive polymer as an easy way to interface MIP particles with the transducer surface was later investigated in more detail [126, 127]. Ho et al. [126] prepared an amperometric sensor for morphine by immobilising MIP particles into an electrosynthesised film of PEDOT. A composite material prepared by entrapment of MIP particles for the alkaloid (−)-ephedrine into a PPy matrix was successfully used as the sensing layer of a voltammetric sensor for (−)-ephedrine [127]. One of the most appealing aspects of this approach is the possibility of decoupling MIP synthesis and immobilisation, thus enabling better optimisation of each step separately. At the same time advantages related to the electropolymerisation (i.e. good adherence of the film to the transducer surface and easy control of film thickness) were manifested [127].

An alternative successful strategy for creation of a path for the conduction of electrons from MIP active sites to the electrode surface has recently been proposed [128]. A novel hybrid material combining a methacrylamide-based MIP grafted to an electrically conductive polymer aniline was prepared. A layer of MIP was photochemically grafted over the electropolymerised polyaniline (PANI) via iniferter (N,N′-diethyldithiocarbamic acid benzyl ester) activation of the methacrylamide groups. In this way a network of “molecular wires” was created enabling the electronic conduction from the MIP to the electrode surface (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig5_HTML.gif
Fig. 5

Preparation of MIP for catechol and grafting via electrosynthesis of a PANI-like film underneath. Reproduced, with permission, from Ref. [128]. Copyright 2009 American Chemical Society

Another original application of electrochemistry in the preparation of MIP-based electrochemical sensors consisted in electrografting of azo-initiators on to the electrode surface [129]. Quercetin-imprinted grafts were prepared by atom transfer radical polymerisation (ATRP) using aryl-modified gold electrodes as macroinitiators. Instead of the traditional method using self-assembled monolayers (SAMs) of thiols with ATRP initiator ends, attachment of ATRP initiators to the gold surface was performed by the electrochemical reduction of a reactive diazonium salt. It was shown that the tandem diazonium salt chemistry and ATRP enabled the development of a robust sensor device based on ultrathin MIP grafts.

Excellent adhesion of chemically prepared MIPs to the electrode surface has been demonstrated in the electrodeposition of imprinted sol–gel films [130133]. This procedure is simple and controllable, enabling film thickness to be adjusted by selecting appropriate deposition conditions, for example deposition time and applied potential [130, 134]. The choice of the potential is related to the possibility of promoting the hydrolysis and condensation of silanes [130] and thus the formation of a polymeric network around the template. In most cases, electrochemical deposition is accompanied by entrapment of nanostructured materials in the imprinted films (multi-walled carbon nanotubes [131, 132] and gold nanoparticles [133]) with the objective of overcoming slow diffusion kinetics and long response time that typically affect sensing devices based on imprinted sol–gel films [132]. A similar scheme has recently been applied to the electrochemical deposition of imprinted polymeric micelles for fabrication of a voltammetric glucose sensor [134]. In this case, positively charged MIP micelles imprinted with glucose were prepared by self-assembly of a photocross-linkable amphiphilic copolymer in a selective solvent. They were then electrodeposited on the electrode and the subsequent photocross-linking “froze” the recognition cavities in the polymer matrix. Electrochemical treatment has also been applied to chemically prepared MIP films for tyrosine, after their casting on the electrode surface, to improve selectivity [135].

A hybrid approach combining both electrochemistry and chemical synthesis in the preparation of imprinted materials has been widely explored by Nagaoka’s research group [4345] who proposed the overoxidation of chemically prepared PPy as a feasible way for template removal, thus extending an imprinting strategy to electrosynthesised PPy [4345]. This methodology involved PPy synthesis in the form of colloids in the presence of a steric stabiliser and produced imprinted materials with satisfactory enantioselective properties for the uptake of amino acids [43, 44] and good selectivity for discrimination of structural isomers [45]. The application of this imprinting scheme succeeded also in the preparation of surface-imprinted oPPy nanowires [46] for glutamic acid. In this way, imprinted binding sites were located on the surface of oPPy nanowires thus enabling easier access of the analyte to the imprinted binding sites, faster mass transfer, and faster binding kinetics.

Recent advances in MIP sensors—nanomaterials in MIP electrosynthesis

In this section an overview of the latest achievements in the application of nanomaterials in several formats (e.g. nanoparticles or electrosynthesised MIP nanostructures) to the development of improved MIP sensors is presented (Table 5).
Table 5

MIP sensors based on different types of nanomaterials

Sensing layera

Template/analyte

Transduction

Ref.

AuNPs/PPD

Theophylline

Voltammetric

[138]

AgNPs/PPD

Dimethoate

Voltammetric

[137]

AuNPs/PPD

Tripeptide glutathione (GSH), GSH-capped Au

QCM

[139]

AuNPs/poly-o-ATP

Tolazoline

Voltammetric

[140]

AuNPs/poly-p-ATP

Chlorpyrifos

Voltammetric

[141]

AuNPs/poly-p-ATP

TNT

Voltammetric

[136]

AuNPs/poly-p-ATP

Picric acid/TNT

SPR

[142]

AuNPs/poly-p-ATP/Cysteine/mercaptoethane sulfonic acid

Several amino acids: l-histidine, l-aspartic acid, l-phenylalanine and L-glutamic acid

SPR

[143]

PPy Nanowires

d or l-Camphorsulfonic acid/d or l-phenylalanine

Faradic impedance spectroscopy

[144]

PPy Nanofilms/Carbon Nanotubes

Caffeine

Voltammetric

[145]

Polyphenol Nanofilms/Carbon Nanotubes

Several proteins: human ferritin and human papillomavirus derived E7 protein and calmodulin

Impedance, voltammetric

[146]

PEDOT(PSS doped) microrods

Avidin, avidin-FITC

Fluorescence binding assay

[148]

Tyrosinase mimicking polymer/derivatised-PANI nanostructures

Catechol

Voltammetric

[149]

aMIP is highlighted in bold if the sensing layer consists of different materials

Composite systems MIP/nanomaterials

For the successful application of MIPs in sensors it is necessary to improve their binding kinetics, shorten analysis times, and achieve complete template removal [136]. The integration of nanomaterials into sensing layers can solve these issues by enhancing the sensor surface area, which in turn should increase the sensitivity of imprinted polymers. Nanoparticles, in particular, have been exploited in electrochemical sensing, enabling the development of ultrasensitive assays. Nanoparticles can dramatically increase the intensity of the electrochemical signal derived from the binding event of the analyte, acting in two ways. They can be used either as electroactive and conductive labels to make signal generation easier, or alternatively they can themselves act as “carriers” to accumulate larger amounts of electroactive analyte and deliver them more easily to the electrode surfaces [137].

An example of the use of gold nanoparticles (AuNP) for increasing the electrical conductivity of MIP layers imprinted with theophylline has been provided by Kan and colleagues [138]. In this work the authors modified a glassy carbon electrode (GCE) by electropolymerising OPD in presence of template. Then AuNP were formed on the imprinted surface by potentiostatic deposition, leading to a remarkable increase of the conductivity of the films (approximately 30 times). Rebinding of the template itself by the MIP layer resulted in a substantial decrease in the conductivity of the imprinted films. Integration of the Au nanoparticles gave the imprinted sensors a wider linear range and a lower detection limit than sensors prepared in the absence of AuNP. The modified sensors had a detection limit of 1 × 10−7 mol L−1 and higher selectivity toward the template than to analogues such as caffeine and guanine. In addition, analysis of real serum samples spiked with the template produced reproducible and consistent results, which are promising for real application of the imprinted sensors in a clinical environment.

A very similar approach has been used by Du and co-workers [137], who used cyclic voltammetry to create a thin MIP layer, specific for dimethoate, made of OPD, on a gold electrode integrated with silver nanoparticles (AgNPs) as signal amplifiers. The silver used in this work is less expensive than gold and more conductive, thus ensuring a more pronounced amplification effect. Du and colleagues used SPR for monitoring the thickness of the polymer layer, which was estimated to be approximately 25 nm. Such a thin layer is crucial in amperometric sensors, because rapid diffusion across the imprinted matrix is required to obtain high and fast responses. In this sense electrochemical polymerisation is advantageous for coating electrodes, because it enables the production of uniform coatings, with reproducible and easily tuneable characteristics (e.g. thickness or porosity). Ag-modified MIP electrodes also had high specificity and selectivity compared to non-imprinted electrodes, even in the presence of metal ions. The authors concluded that the high selectivity is caused by the rigid structure of the imprinted cavity. Considering that both the fabrication process and the response after consecutive measurements were highly reproducible, this approach might be considered for mass production of the inexpensive MIP devices.

Stobiecka et al. [139] recently synthesised imprinted films for the small tripeptide glutathione (GSH) by exploiting a slightly different approach. In particular, they performed the imprinting process by electropolymerising OPD either on a bare gold electrode, as in previous work, or differently on a layer of AuNP assembled on a gold surface, thus resulting in a rough network with higher surface area. In addition the authors imprinted not only GSH molecules, but also GSH-capped AuNP, in order to increase the accessibility of binding sites at the surface of the electrode. MIP electropolymerisation was performed using a potential step technique (potential pulses). In comparison with a plain gold electrode, the presence of Au nanoparticles as backbone layer resulted in faster polymerisation, because of the higher surface area exposed at the interface with the solution. When GSH-capped AuNP were used as template, an enhanced sensor response was observed, probably because of the greater number of template-shaped binding sites accessible on the surface of polymeric film. Indeed the rebinding capacity of the sensor produced in this way was approximately 2.6 times higher than that obtained through a classical imprinting approach with a small template. Unfortunately the authors did not perform any control experiment in the absence of template or in the presence of template analogues; hence neither the imprinting effect nor the selectivity of the systems could be fully evaluated.

The group of Zhang et al. [140] produced a tolazoline sensor based on AuNP using o-aminothiophenol as functional monomer. This monomer was chosen because the Au–S bond with the gold electrode is highly stable, resulting in better immobilisation of the imprinted film on the sensor surface. Moreover, the amino group and the benzene ring of the monomer can interact with tolazoline through hydrogen bonds and π–π stacking interactions. The authors deposited AuNP both to enhance the sensitivity of the polymer-modified electrode and to produce a structured network free from aggregation phenomena, because of the interaction of the nanoparticles with thiol groups of the monomer. As expected, deposition of the MIP film substantially reduced the conductivity of the electrode, which increased again after deposition of the AuNP. For generation of the sensor signal the authors used ferricyanide as electrochemical probe. In fact, in the presence of the AuNP, typical redox peaks for ferricyanide were obtained from both the imprinted and the non-imprinted sensors. However, only for the imprinted sensor was a further increase of ferricyanide peak current detected after removal of the template tolazoline from the imprinted layer; this effect was then reversed by template re-binding. MIP-modified electrodes were highly sensitive (16 ng mL−1) and selective, and their stability during the regeneration processes and storage was acceptable (up to one month). The authors demonstrated excellent performance of MIP sensors in human urine samples spiked with template, thus revealing good potential of the MIP sensor for the application in clinical area.

Similarly Xie and co-workers [141] prepared an electropolymerised poly-p-aminothiophenol layer imprinted for the electrochemical detection of chlorpyrifos on a GCE with deposited AuNP. The authors exploited the strong Au–S bond between p-aminothiophenol (p-ATP) and deposited gold nanoparticles to build an initial polymerisable SAM directly on the gold surface. This SAM acted as a “wetting agent” for the polymer produced in the subsequent electropolymerisation step. Before the polymerisation, the SAM-modified electrode was incubated with template to localise the imprinted sites on the surface of the MIP layer. Eventually the MIP layer was electropolymerised by cyclic voltammetry, using an electrolyte solution containing p-ATP and template, to further increase the number of imprinted binding sites. Scanning electron microscopy (SEM) revealed the formation of a polymer layer with a rough surface produced by the AuNP lying underneath. As expected, deposition of Au nanoparticles increased the conductivity of the electrode threefold. When rebinding of the chlorpyrifos was monitored by cyclic voltammetry, the developed MIP sensor was shown to be sensitive (detection limit 3.3 × 10−7 mol L−1) and selective, because even a tenfold excess of interfering substances did not cause any relevant change in the signal for the template. In addition, the reproducibility of the measurements and the stability during storage (90% of the sensitivity retained after 30 days) were good. The authors also tested real water samples spiked with template, and the results indicated the possibility application of this system to detection of chlorpyrifos in the environment without any pre-treatment.

Riskin and colleagues [136] recently used a novel approach to produce MIP-coated electrodes for trinitrotoluene (TNT). On a bare Au electrode TNT could be detected at a concentration of 10 μmol L−1, with the detection performed amperometrically by recording the current of the TNT reduction peak at −0.5 V (vs a saturated calomel electrode). After modification of the surface of this electrode with a SAM of p-ATP, TNT molecules were concentrated mostly on the electrode surface as a result of π–π interactions with the p-ATP molecules, thus leading to a 135-fold increase in sensitivity. The authors then tried to increase the sensitivity even further by performing the electropolymerisation of p-ATP-capped Au nanoparticles (p-ATP-Au-NPs) on electrodes already bearing the p-ATP SAM. In this way the sensitivity of the system could be further increased by a factor of 37. The system also had a selectivity factor of approximately 20 for dinitrotoluene (DNT) and 4-nitrotoluene (MNT). Riskin and co-workers then applied an imprinting process to this system by pre-complexing p-ATP-Au-NPs with a TNT analogue, picric acid [136]. They then electropolymerised these capped nanoparticles by cyclic voltammetry and, after extraction of the pseudo-template, performed rebinding of TNT, which was monitored again by amperometry. Imprinted sensors had a sensitivity of 200 pmol L−1 and only 10% loss of response after seven days usage, thus showing good stability. Later, the same group imprinted picric acid using p-ATP-Au-NPs as “functional monomers” to create a detection system for TNT based on a SPR platform [142] (Fig. 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig6_HTML.gif
Fig. 6

Multi-step assembly of a TNT sensor based on poly(p-ATP) imprinted with picric acid and grown directly on Au nanoparticles. Adapted from Ref. [142] and reproduced with permission. Copyright 2009 American Chemical Society

This enabled reduction of the detection limit for TNT to 10 fmol L−1. The authors also confirmed that the electron-acceptor characteristics together with shape complementarity were the main factors responsible for the imprinting effect. In subsequent work, the same sensing platform was used by Riskin and colleagues [143] to develop amplified SPR sensors for chiroselective analysis of glutamic acid. To achieve this objective, the authors co-functionalised the p-ATP-Au-NPs with cysteine units, obtaining functional electropolymerisable AuNP capable of binding amino acids through complementary zwitterionic electrical interactions. The resulting imprinted matrices had chiroselective recognition and binding capabilities toward the imprinted amino acid (either l or d-glutamic acid). It would be interesting to evaluate the ability of all the sensors developed by Riskin and colleagues to detect their target templates in real samples to have a full assessment of their real value.

In summary it seems that application of nanomaterials and, especially, conductive nanoparticles can increase sensitivity of MIP-based sensors. Nevertheless, some work must still be done to optimise production processes.

Electrosynthesis of nanostructured MIPs

Conducting imprinted polymers based on, e.g., PPy are very attractive in sensing applications, because they can directly transduce the binding signal. Their performance can, however, be further enhanced if their format is changed from bulk monoliths to nanoparticles. Huang et al. [144] electropolymerised PPy nanowires using d and l-camphorsulfonic acid (d and l-CSA) as pseudo-templates. Imprinted nanowires had clear tubular morphology and smooth surfaces, with a diameter of approximately 100 nm and lengths ranging from one to several microns. Faradaic impedance spectroscopy (FIS) was used to analyse the changes of the interfacial properties of modified electrodes in the presence of charged molecules bound to their surface. When the sensor was tested with d or l phenylalanine, FIS confirmed that imprinted nanowires had high enantioselectivity, because the electron-transfer resistance decreased only after interaction with the specific imprinted enantiomer. The CSA imprinted nanowires were able to detect phenylalanine enantiomers at 5–200 ppm concentration with a response time of few minutes. However such anions as Cl or SO42− interfered with the recognition process, hence the samples needed to be carefully pre-treated before the analysis.

A similar strategy has been developed by Choong et al. [145] who performed electropolymerisation of caffeine-imprinted PPy thin nanofilms on arrays composed of vertically aligned carbon nanotubes (CNTs) with different densities. These latter acted as high-surface 3D scaffolds for the thin MIP layer deposition. In this way, good control over the thickness of the PPy imprinted layer could be achieved. The high conductivity of CNTs ensured good transduction of the electrochemical signal. The authors produced imprinted nanotubes of 1.5 μm length and 100 nm diameter coated with an MIP PPy layer of 60 nm. A denser configuration was also tried, resulting in a honeycomb CNT structure, on to which an MIP layer approximately 20 nm thick was immobilised (Fig. 7).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig7_HTML.gif
Fig. 7

(A). Schematic illustration of the thick MIPPy film grafted around each CNT for large templates, and (B). thin MIPPy film deposited for small target molecules. SEM images of a sparse CNT array (C). before and (D). after caffeine imprinting electropolymerisation. SEM images of a dense CNT array (E). before and (F). after caffeine imprinting electropolymerisation. Reproduced from Ref. [145], Copyright 2010, with permission from Elsevier

Both the sparse and dense CNTs MIP sensors were then analysed using pulsed amperometric detection (PAD). They had improved binding capacity and responses respectively 3.6 and 15-fold higher than the responses of a sensor coated with a conventional MIP thin film. However the polymerisation procedure had to be carefully optimised to avoid the formation of dendrites on the tips of the CNTs. Moreover, in the case of the dense configuration, only very low caffeine concentrations could be detected, because at high concentrations the template solidified and blocked the three-dimensional structure of the membrane. In addition, despite the high surface area of the sensor, a lag time of approximately 1 min was observed during the response, probably because of the slow diffusion of the template inside the dense CNT pattern. Nevertheless, this technique seems to have a good potential, especially for imprinting bulky templates such as proteins.

Cai and colleagues [146] have also used a carbon nanotube array and imprinting technology for ultrasensitive detection of proteins. In this case, instead of using a highly conductive polymer as PPy, the authors electropolymerised a non-conductive polyphenol nanocoating on the tips of the carbon nanotubes in the presence of a template protein (human ferritin or human papillomavirus-derived E7 protein). The authors then relied on measurements of impedance to monitor the recognition event. Upon extraction of proteins from the nanocoating the sensor electrical impedance was greatly reduced, whereas subsequent template protein rebinding was detected as an increase in impedance. A sub-pg L−1 detection limit for human papillomavirus-derived E7 protein was obtained by use of the MIP nanosensor. In addition, the nanosensor was not only highly sensitive and selective, but was also capable of detecting protein conformational changes. This was shown by using, as template, calmodulin, a protein undergoing conformational changes upon Ca2+ binding. The Ca2+–calmodulin complex was detected by differential pulse voltammetry at a variety of Ca2+ concentrations, with the peak current decreasing as a function of Ca2+ free in solution. Although this MIP nanosensor seems very promising for ultrasensitive detection of proteins, testing was mainly done in model solutions and analysis of real samples would be necessary to assess the real value of the technology.

The imprinting of proteins is indeed quite difficult for several reasons, not the least because of the intrinsically poor stability of these types of templates under the imprinting conditions, their size, and the multitude of exposed functional groups [147]. For these reasons, surface-imprinting strategies have given good results in enhancing the imprinting effect for these templates. Menaker et al. [148] introduced a new method to electropolymerise MIP micro or nanowires surface-imprinted with proteins. In particular, they synthesised MIP microrods imprinted with avidin by exploiting a polycarbonate membrane filter as sacrificial mould. The template protein was first physically adsorbed on the hydrophobic surface of the sacrificial matrix. After this step, the polycarbonate membrane was positioned on a gold electrode and imprinted microrods were synthesised by electrochemical polymerisation of PEDOT and polystyrene sulfonate (PSS, doping agent). When the microrods had been synthesised inside the membrane pores, the mould was dissolved with chloroform. Use of an electrochemical polymerisation process enabled good control over polymer growth and deposition. In addition, the mild aqueous synthetic conditions used in this work were beneficial for preventing denaturation of protein templates. Synthesised microrods had characteristics similar to the mould pores, i.e. approximately 8 μm in diameter and 7 μm in height. Their imprinting properties were confirmed by fluorescence binding assays, which showed the dissociation constant to be 394 nmol L−1. The assays were performed using a fluorescent derivative of avidin (avidin-FRTC). Imprinted microrods were also approximately twice as selective for the imprinted template than for bovine serum albumin.

Piletsky and co-workers also used a “sacrificial mould” approach for synthesis of nanostructured MIP [149] and compared its performance with the grafted sensor mentioned previously in this review [128]. The same monomer with orthogonal polymerisable functionality comprising an aniline group and methacrylamide was used in both examples. In the first work electropolymerisation of the aniline moiety resulted in the formation of a conductive film on a gold electrode surface. Then, after activation of the methacrylamide groups with a living initiator, an MIP film capable of oxidising and therefore detecting catechol was photochemically grafted on the conducting layer, creating a hybrid-MIP sensor (Fig. 5). Detection of catechol by the resulting sensors was sensitive (228 nmol L−1) and specific, because the only other compound detected to a lesser extent was dopamine. In subsequent work, Piletsky and colleagues investigated the effect of nano-features in the same type of hybrid-MIP. Alumina nanoporous membranes were used as sacrificial nano-mould for preparation of the polyaniline nanostructure. To do this, the same monomer with orthogonal functionality used in the earlier work was electropolymerised inside the nanopores of the alumina membranes after sputtering a gold layer on one side of the membrane to obtain conductivity. At the end the nano-mould was removed by dissolving the alumina membrane with NaOH, leaving behind MIP grafted polyaniline nanowires attached to the gold layer. SEM photographs of the polyaniline before MIP grafting and after the alumina membrane dissolution are shown in Fig. 8.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-011-5405-5/MediaObjects/216_2011_5405_Fig8_HTML.gif
Fig. 8

SEM images of (A) top view of polyaniline nanowires after nano-mould dissolution; (B) vertically aligned polyaniline wires after nano-mould dissolution, cross-sectional view. Reproduced from Ref. [149], Copyright 2010, with permission from Elsevier

After grafting and removal of the template, the resulting nano-hybrid MIP had selectivity similar to that of the hybrid-MIP film sensor, but sensitivity was higher, with a detection limit of 29 nmol L−1. This demonstrates that the presence of nanostructures increases the ability of the sensor to detect lower concentrations, probably because of enhancement of the surface area.

In conclusion despite being still in its early optimisation phases, electrosynthesis of nanostructured MIPs already seems to be a very promising production strategy, particularly suitable for a rapid creation of highly sensitive detection systems. In addition, its compatibility with surface-imprinting strategies may lead to the production of a new generation of sensing systems which might compete with current commercial sensors based on biological molecules.

Conclusions

Molecular imprinting based on the electrochemical approach has been shown to be an appealing strategy for development of advanced sensors for different applications. In comparison with preparation of all-chemical MIP sensors, MIP electrosynthesis, and electrochemical entrapment of chemically synthesised MIPs, enables the control of film thickness via electrochemical conditions and may often be carried out in aqueous environments. On the other hand, some site heterogeneity, often encountered in such MIPs, seems now addressed by the use of novel electropolymerisable monomers, which bear different functional groups. Hybrid approaches are emerging as an interesting option to develop more sophisticated sensitive devices. In this respect, the suitable coupling of electrochemical strategies with nanomaterials and/or chemically synthesised MIPs may be crucial in further progress of MIP sensors. However, the field has still to mature, particularly, in reference to electrosynthesis of MIPs in which the trial-and-error approach in the choice of suitable monomers and experimental conditions for selected templates should be replaced with a more rational approach. We expect that the application of tools, for example computational and/or combinatorial tools, already developed for other synthetic routes to MIPs, with the extension to this field of the conspicuous theoretical and applied knowledge available in electrochemical preparation of thin polymeric films, could result in a decisive improvement of the electrochemical approach to MIP sensors.

Copyright information

© Springer-Verlag 2011