Abstract: Increased attention is being focused on engineering micro- and nanostructured materials. Of special technological and scientific interest are engineered materials capable of molecular recognition based on pore size, shape, and functionality. The molecular specificity can be achieved using molecular imprinting techniques. This paper aims to initiate theoretical work to establish criteria for successful molecular imprinting. Using the path-integral mean field theory, we modeled cross-linked polymer networks in the presence of excluded volume particles. Current experimental data is insufficient for a quantitative comparison with our predicted scaling laws. Nevertheless, we were able to predict general experimental trends concerning the positive dependence of the formation of stable molecular templates on the degree of cross- linking and the negative dependence on the size and density of the template molecules. (C) 2002 American Institute of Physics