Prediction of Room Temperature Ferroelectricity in Subnano Silicon Thin Films with an Antiferroelectric Ground State

Recent advancements highlight the critical need for ferroelectric (FE) materials compatible with silicon, particularly pure silicon phases exhibiting FE behavior above room temperature that can be readily integrated onto silicon substrates. Here, we systematically predict potential FE silicon films using first-principles structure searching. Furthermore, we develop an advanced machine learning model that simultaneously predicts interatomic potentials and Born effective charges to investigate their FE behavior. In particular, we discover a low-energy FE silicon film merely 1 meV/atom above its antiferroelectric (AFE) ground state. Both phases are confirmed to be dynamically and mechanically stable semiconductors with indirect band gaps of approximately 1.3 eV. Although the ground state exhibits AFE ordering, our simulations with the machine learning model remarkably reveal a robust FE hysteresis loop above room temperature, characterized by a low coercive field of about 0.05 V/Å. An effective Hamiltonian model analysis indicates that this anomalous FE behavior arises from weak AFE coupling between the top and bottom surfaces of the silicon film. These findings of a switchable FE state in pure silicon phases suggest their potential use in memory devices, sensors, and energy converters.