Oral: An ab-initio method for magnetic specific heat using constrained supercells

Across sectors, there is an increased reliance on materials and devices with tunable magnetocaloric properties as critical components in a larger systems stack; yet the discovery and design of these materials simultaneously satisfying other technological constraints remains an open problem. For instance, cryogenic cooling systems require materials that (1) are antiferromagnetic, (2) exhibit high magnetic heat capacity at cryogenic temperatures, and (3) contain earth abundant elements. This example requires reliable ab-initio computational tools to predict the temperature-dependence of the magnetic specific heat (MSH) and other magnetothermodynamic (MT) properties for de novo material candidates. To date, we have not found any reliable ab-initio models to predict magnetothermodynamic properties like MSH and its temperature dependence. Here we present a systematic ab-initio methodology that utilizes a statistical mechanical model parameterized by energetic and structural inputs from density functional theory (DFT) calculations to predict the temperature and volumetric-dependent MSH and related MT properties. We demonstrate the tool’s ability to predict temperature dependent quantities like the MSH and associated critical temperatures for exemplary ferromagnetic and antiferromagnetic materials like Fe and HoCu₂. Finally, we compare our method against existing baselines including semiclassical Monte Carlo techniques and experimental measurements.