High-throughput computational design of Heuslerenes

Heuslerenes, the two-dimensional (2D) analogs of Heusler alloys, exhibit rich electronic, magnetic, and topological properties, making them promising candidates for advanced material applications. In this work, we computationally cleaved monolayers from bulk cubic Heusler structures to create a database of approximately 2000 Heuslerenes. We developed a data-driven machine learning model to predict their static and dynamic stability, as well as the topology of the electronic bands near the Fermi energy. Using the low-energy effective Hamiltonian derived from downfolding with maximally localized Wannier functions, we conducted systematic Berry curvature analysis and proposed an automated approach for the computational design and discovery of 2D Heuslerene-based topological insulators.