Magnetic refrigeration is an emergent technology that promises to be more energy efficient and eco-friendly than conventional cooling devices. At the basis of this technology is a thermodynamic cycle that exploits the magnetocaloric effect (MCE). In general, all magnetic materials experience an adiabatic temperate change upon removal of an applied magnetic field, this is the MCE. For real world devices however, an adequate material must meet several conditions, such as a strong MCE due to a magnetic transition near the room temperature, low hysteresis and absence of critical (rare-earths) or dangerous elements (arsenic). To date the search for search materials has mostly been an intensive experimental effort. An ongoing topic in our group has been to search for optimized or novel compounds that would serve as enablers of widespread magnetic refrigeration. We focus on two approaches: high-throughput search for new compounds and detailed calculations of phase transitions from first principles.

A common tactic in optimizing magnetocaloric (hexagonal and full) Heusler compounds, is to induce magnetic transitions coupled with discontinuous structural transitions. Often this is done by substitutional disorder. In this sense we can picture this substitution as an interpolation between different isostructural compounds, to tune the magnetic and structural properties. As part of our work in the hexagonal Heusler system we are also conduction high-throughput search for novel parent phases, so that we can later add substitutional disorder by using Special Quasi-random Structures supercells and the Coherent Potential Approximation within DFT. To this end we use our in-house high-throughput environment.

A central aspect of high-performance magnetocaloric materials is the presence of a first order transition, for our purposes of a magneto-structural type. To determine the stable phase at a given temperature one must minimize the free energy as a function of volume for each competing phase. The total free energy can be expressed by considering the internal energy (E0), the vibrational (Fvib) and magnetic (Fmag) contributions, the first two can be determined by Density Functional Theory. For the Fmag term on must go further and consider either the mean-field approximation or use Monte Carlo methods, using exchange interactions in turn obtained by DFT. An instance of this approach is shown below for the transition of the magnetocaloric Mn-Fe-Ni-Si-Mn hexagonal Heusler system.