Molecularly imprinted polymers (MIPs) have been compared to natural antibodies in that they can specifically bind target compounds in a similar way that antibodies specifically bind to an antigen. The attraction of the MIPs technology is the ease of creating binding elements which are relatively cheap compared with the process of isolating natural antibodies. In this research monoterpenes, such as α-terpineol, were chosen to be the model compounds for investigating the molecular imprinting of small, poorly functionalised organic compounds. The conventional non-covalent approach was mainly used to synthesise these MIPs, but the sacrificial-spacer semi-covalent approach was also investigated. A less widely used method, porogen-imprinting - a variant of non-covalent imprinting - was adapted for α-terpineol. The latter novel terpene MIP appeared to specifically bind α-terpineol, by hydrogen bonding, so the polymer was characterised in detail. The main parameters which were altered for preparing non-covalent MIPs included the template (α-terpineol, (-)-menthol or trans-terpin); the functional monomer (methacrylic acid, 2-hydroxyethyl methacrylate, bilirubin and phenol [for the semi-covalent MIP]); the cross-linking monomer (ethylene glycol dimethacrylate, divinylbenzene and trimethylolpropane trimethacrylate); and also the polymerisation method (block or precipitation polymerisation). The binding specificity and cross-reactivity for all the polymers were tested using a liquid batch-binding setup. The batch-binding setup required the detection of analyte that was not bound in order to calculate by difference the fraction of analyte bound to the polymer. Initially the terpenes were to be detected by a colorimetric method; however attempts to make the method sensitive and reliable were not successful. In comparison, gas chromatography was more reliable for the detection of terpenes and was used for the experiments presented in this thesis. 1H-NMR studies of the interaction between α-terpineol and acetic acid (as a non-polymerisable analogue of methacrylic acid) were investigated as a basis for understanding the binding to the carboxyl functional group moiety employed in many of the non-covalent MIPs that were made. The interaction between (-)-menthol and phenol was also investigated because the phenol moiety was employed in the semi-covalent MIP. Only selected MIPs, which appeared to specifically bind the template, were physically characterised. This included optimising the batch-binding parameters, scanning electron microscopy imaging, surface area and pore radius analysis and in some cases Fourier transform-infrared spectroscopy of the polymers.

The research described in this thesis contributes to the development of new strategies facilitating advanced understanding of the fundamental principles governing selective recognition of molecularly imprinted polymers (MIPs) at a molecular level for the rational optimization of biomimetic materials. The nature of non-covalent interactions involved in the templating process of molecularly imprinted polymers based on the self-assembly approach were investigated with a variety of analytical techniques addressing molecular level interactions. For this purpose, the concerted application of IR and 1H-NMR spectroscopy enabled studying the complexation of the template molecules 2,4-dichlorophenoxyacetic acid, quercetin, and o-, m-, and p-nitrophenol with a variety of functional monomers in the pre-polymerization solution by systematically varying the ratio of the involved components. In aqueous and non protic porogenic solvents, information on the interaction types, thermodynamics, and complex stoichiometry was applied toward predicting the optimum imprinting building blocks and ratios. Molecular dynamics simulations of 2,4-dichlorophenoxyacetic acid and its interactions with the functional monomer 4-vinylpyridine in aqueous and aprotic explicit solvent allowed demonstrating the fundamental potential of computer MD simulations for predicting optimized pre-polymerization ratios and the involved interaction types. The obtained results clearly demonstrate that the application of rapid IR/NMR pre-screening methods in combination with molecular modeling strategies is a promising strategy towards optimized imprinting protocols in lieu of the conventionally applied labor intensive and time-consuming trial-and-error approach. Furthermore, HPLC characterization of the produced MIPs compared to control polymers enabled a systematic approach to imprinting based on advanced understanding of the factors governing the formation of high-affinity binding sites during the polymerization. In addition, the importance of the combination of size, shape, and molecular functionalities for the selective recognition properties of MIPs was investigated. MIPs for the mycotoxins deoxynivalenol and zearalenone and for the antioxidant quercetin were applied as separation materials for advanced sample preparation in beverage analysis. The obtained results demonstrated the potential of MIPs for rapid one-step sample clean-up and pre-concentration from beverages such as wine and beer.