ABSTRACT
Furthermore, using B3LYP/6-311+G**//B3LYP/6-31G* and including solvent effects in the energy calculations of complexes in the pre-polymerization mixture, Diñeiro et al. [34, 35] were able to rationally choose both the functional monomer and porogen to use for the successful preparation of a homovanillic acid imprinted polymer. An extended strategy for optimization of both functional monomer and solvent was presented by Dong et al. [36], who complemented B3LYP/6-31G(d) calculations with a preceding step in which candidate systems were chosen using molecular dynamics simulations. Another use of electronic structure methods has been for cases where a template that differs from the target structure of the MIP is used. Here, these methods can also assist in choice of the template itself. One example of such a study was published in 2005 by Rathbone et al. [37] and concerns the design of a MIP-based binding mimic of cytochrome CYP2D6. Here, the templates to be imprinted were chosen on the basis of superposition studies of a series of PM3-optimized candidate molecule geometries on known CYP2D6 substrates with the selected templates. This strategy yielded MIPs with recognition of such substrates. A further example of the use of quantum-chemical calculations in molecular imprinting is the design of an ester hydrolysis catalyzing polymer based on PM3 calculations. These calculations provided support for the hypothesis that the template used in the MIP synthesis is a mimic of the transition state of the reaction to be catalyzed [38].Finally, using a quite different approach, Voshell and Gagné [39] used HF/6-31G* and AM1 computational studies to determine the conformational rigidity of a dendritic system used in the imprinting of BINOL. Based on the results, a more rigid dendrimer structure was chosen to lower the binding-site heterogeneity and the enantioselectivity of the resulting MIPs was improved. The primary limitation in the application of electronic structure methods to MIP systems is the difficulty in handling the large numbers of atoms necessary to provide a comprehensive picture of the pre-polymerization or polymer system. One particular complication appears to be the problems that can arise through the lack of explicit solvent [40, 41]. For electronic structure methods, the inclusion of a reasonable number of solvent molecules makes the calculation rather time consuming; as a result, solvent effects are often omitted completely. However, methods such as PCM provide the possibility
of including solvent effects without the inclusion of explicit solvent molecules [42]. Nonetheless, the validity of this assumption appears limited in light of a recent report [40]. The method was recently applied by Wu et al. [43], in an MP2/6-311+G*//B3LYP/6-311G* study of the pre-polymerization mixture of a nicotinamide-imprinted polymer. It was shown that the model works well for predicting the influence of different solvents on the retention and selectivity characteristics of the polymer, as long as the solvent itself is aprotic. A significant limitation of the PCM method is that not all solvents can be modeled adequately due to the inability of the PCM method to include the effect of hydrogen bonds to solvent molecules, which compete with the hydrogen bonds formed between the template and functional monomers. This was also exemplified by Liu et al. [44] employing a B3LYP/6-31+G(d,p) level of theory in which the solvation energies obtained using PCM were compared to the energies of interaction between the template and individual solvent molecules. A different approach of using solvation energies obtained from B3LYP/6-31+G(d,p) including a PCM model was demonstrated in a study by Dong et al. [45]. Here the solvation energies of template and functional monomer molecules in different solvents were directly used as a measure of potential competition for interactions from the solvent. A further application of electronic structure methods is the evaluation and characterization of a given MIP. Based on the optimized geometries and Mulliken charges calculated for a series of different substrates using B3LYP/6-311G**, Wang et al. [46] were able to suggest a recognition mechanism explaining the selectivity of an N-(4-isopropylphenyl)-N’-butyleneurea imprinted polymer toward such compounds. Jacob et al. [47] were able to suggest a model for the interactions between polymer and template using both ab initio and density functional theory (DFT) methods in combination with different basis sets. Similarly, by using PM3 calculations of template-monomer interactions, Wu and Li [48] could explain the failure of imprinting an MAA-based polymer using picolinamide. They subsequently developed a Cu(II) complex based system that did show selectivity for the imprinted compound. The system has also been extended to the recognition of small organic acids [49]. The mode of interaction of semiquinone radicals with an imprinted polymer system was elucidated by Christoforidis et al.
