Manipulating thin film properties by search for substrates over databases

Substrates can have major effects on the growth direction, orientation, nucleation, morphology, field emission in carbon nanotubes, superconducticity, and etc. The influence of substrate on the thin film can be mechanical, electronic, or both. Specific applications require different substrates. In order to tune the characteristics of the thin film to ones needs, one has to explore a large set of possible substrates. As experimentally this work is almost impossible, one has to use computational methods to evaluate possible substrates. In this work we report a flowchart on how to search for optimal substrates and complement the selection by specific electronic structure characterization. We apply our methodology to FeSe, where the superconductor properties are very sensible to the substrate. We employ the available algebraic algorithm provided by MPInterfaces package to search for geometrical matches in the OQMD database and calculate the work-function of the matched substrates as the criteria to select a good substrate. The results from the geometrical matches are input to a machine learning algorithm to increase the search speed in the databases.