Topology and symmetry guided design of framework structures

Materials design and discovery have been gaining ground rapidly due to the combination of strong empirical and theoretical knowledge bases as well as statistical tools combined with computing resources. The above process is strongly dependent upon the crystallographic nature of the problem, and we focus on this aspect in our study. We investigate the relationship between local topology-symmetry and crystal structure in framework structures. We choose the aluminosilicates and aluminophosphates (both mesoporous materials) as being representative of framework materials. Previous work on network materials has focused mainly on classification based on topology or symmetry. We combine both symmetry and topology in our approach by using Wigner-Seitz cells. We also predict structural details on the mesoscale. This is achieved by developing "secondary" descriptors based on Wigner-Seitz cells and by using statistical tools such as Principal Component Analysis (PCA) and Partial Least Squares (PLS). By this approach, we investigate a mesostructural feature (pore size) as well M-O bond length using PLS and PCA on local topology-symmetry descriptors. The success of this approach in mesoporous materials points the way toward extending the investigation to framework materials in general. Multiscale structure relations in these materials can be studied and used for design by appropriate descriptor selection and statistical methods.