Predicting polarization enhancement in multicomponent ferroelectric superlattices

We use ab initio calculations as an input to develop a one-dimensional chain model of the local polarization in epitaxial short-period CaTiO$_3$/SrTiO$_3$/BaTiO$_3$ superlattices grown on a SrTiO$_3$ substrate. The model is then combined with a genetic algorithm optimization technique to predict the arrangement of individual CaTiO$_3$, SrTiO$_3$ and BaTiO$_3$ layers in a superlattice that would simultaneously have the highest possible polarization and a low in-plane lattice constant mismatch with the substrate. This modelling procedure can be applied to a wide range of layered perovskite-oxide nanostructures providing guidance for experimental development of nanoelectromechanical devices with substantially improved polar properties.