Magnetoelastic coupling descriptor for high-throughput ab initio search of magnetocaloric materials

Magnetocaloric effect results in a heat transfer between a ferromagnetic material and its surroundings in response to a change in applied magnetic field. This has potential applications in the development of energy efficient magnetic cooling technologies provided suitable high-performance magnetocaloric materials can be found. High-throughput ab initio searches for new magnetocalorics are challenging since the magnetocaloric figure of merit, isothermal entropy change, is not directly available from standard electronic structure calculations. Here, we use the Landau theory of phase transitions to design a computable magnetoelastic coupling descriptor that can identify magnetic materials on the brink of first-order transition, which can lead to a strong magnetocaloric effect. The descriptor requires structural optimization in the paramagnetic state, which can be modeled using special quasi-random spin configurations. The descriptor correctly identifies many known magnetocaloric materials and we use it for high-throughput ab initio screening of new magnetocalorics among the L₁₂ family of magnetic materials.