Ab initio workflow for high-throughput multiferroic materials discovery

Ferroics are an important and exciting class of compounds with great potential for applications that exhibit ferroelectricity, magnetism, or both. High-throughput ab initio approaches, in conjunction with materials databases, have been shown effective for screening and discovery of a broad variety of new materials. Here, we build on a previously-developed automated high-throughput workflow [1] to identify a large number of new ferroelectric, antiferroelectric, and multiferroic compounds using density functional theory and the Materials Project. Importantly, our newly-developed workflow does not require the existence of nonpolar insulating “reference” structures in the database. Instead, our more general approach identifies polar insulating materials on or near the convex hull, and assesses possible ferroelectric order, (antiferro)magnetic order, or both by (i) constructing a prospective nonpolar parent phase using a pseudosymmetries-based approach; and (ii) determining the stable magnetic ground state ordering for potential magnetic compounds. The results of applying our new workflow to more than 1500 polar and magnetic materials will be discussed. [1] doi:10.1038/s41597-020-0407-9.

We acknowledge DOE for funding, NERSC for computational resources.