Multivariate analysis and experimental design in the screening of combinatorial libraries of molecular imprinted polymers

The combination of an experimental design/multivariate analysis method and a high-throughput technique was developed in the screening and evaluation of molecular imprinted polymers (MIPs). Combinatorial libraries of MIPs were prepared automatically at a reduced scale and screened for high affinity with regard to six experimental factors with a large impact in the MIP synthesis: the amounts of the functional monomer, cross-linker and template; the volume of the porogen, the amount of the initiator and the types of initiation. With the objective of optimizing a polymer imprinted with the non-steroidal anti-inflammatory drug piroxicam, different polymer libraries were prepared by varying the six factors according to a two-level fractional factorial design of resolution VI. This optimization resulted in polymers showing high affinity as a result, mainly, of reduced binding to the non-imprinted polymers. The type of cross-linker showed a strong influence on the specific binding of the template when assessed by equilibrium rebinding experiments. The cross-linking monomer divinylbenzene gave rise to an enhancement in the binding capacity of the MIP that may be attributed to its similar structure with the functional monomer considered (4-vinylpyridine). The binding capacity of the MIP and blank polymers was related to the six experimental factors using a partial least squares regression (PLS) method. The regression models of the polymer libraries showed the same trend in the MIP and blank polymers, suggesting a higher specific binding capacity when the polymerization was photoinitiated at low temperature; the amounts of the functional monomer and the cross-linker were similar. The proposed procedure is time and cost effective and can be used as a general tool for preparing MIPs for different analytes. The proposed method provides new insight into the influence and interaction of the main factors that affect the MIP performance that could facilitate improvements in the MIP process design in the future.