Searching for New Magnetocaloric Materials—A High-Throughput Approach

Today, approximately 10% of total energy consumption is related to cooling, i.e. space cooling and food preservation with significant variations between countries and climate zones. Due to climate change and an expected huge increase in air cooling devices especially in emerging countries, we face a rapid rise in energy demand the coming years. Within the next 35 years the energy share needed for cooling is expected to surpass the one needed for heating. Meeting this demand with conventional cooling technology is questionable from the sustainable point of view even with more and more efficient cooling devices.
Magnetic refrigeration is viewed as an eco-friendly and efficient alternative for vapor cooling technologies. With suitable caloric materials, it can be used not only for room temperature applications but also for applications at extremely low temperature such as hydrogen liquefaction. However, magnetic refrigeration has been sidelined in the past, which is partially due to the lack of sufficient magnetocaloric materials. The challenge is to identify non-hazardous, low-cost, abundant materials being suitable for magnetic refrigeration devices. Aiming to identify materials with potential for magnetocaloric application we perform a high-throughput search combined with ab initio calculations and spin-dynamics simulations which is a faster and more resource-efficient approach compared to experiment.
Systems that undergo a magneto-structural phase transition (MST), such as Heusler alloys [1] or MnNiSi are interesting for magnetocaloric applications, provided that the structural and magnetic transition fall in the same temperature range. A recent example is the MnNiSi system where a MST can be achieved by Al and Fe doping such that for Mn_1-xFe_xNiSi_0.95Al_0.5 a giant magnetocaloric effect is observed at room temperature [2,3]. Encouraged by these observations, we designed a high-throughput search method to detect more candidate phases with MST.
In a first step, we performed several benchmark studies for known magnetocaloric systems to identify the most suitable and efficient description of the entropy contributions arising from the electronic, lattice, and magnetic degrees of freedom [4,5]. While the electronic entropy can be straight forward derived using ab initio methods, various approaches were tested to capture the magnetic and the lattice contributions to entropy changes. We observed that, other than commonly assumed, the lattice entropy for hcp Gd is non negligible and finite temperature effects needed to be included for accurate quantitative descriptions [5].
Building on these findings, we combined big data searches with ab initio methods and spin dynamics simulations aiming to identify materials with MST suitable for magnetocaloric applications. Scanning the materials project database [6] for binary and ternary compounds without expensive and hazardous elements yielded 3410 systems. Applying a set of magnetic and structural filter criteria, about 20 systems with several polymorphs survived, including some known systems such as Heusler compounds, which underpins the validity of our search algorithm. Several candidates show potential for an MST near room temperature. Their magnetic properties have been studied in our theoretical approach and will be discussed as well as the expected magnetocaloric performance.

References
[1] V. D. Buchelnikov et al., Physical Review B (2010) 81, 094411.
[2] A. Biswas et al., Acta Materialia (2019) 180, 341
[3] S. Ghorai, under review (https://doi.org/10.48550/arXiv.2307.00128).
[4] R. Martinho Vieira et al., Journal of Alloys and Compounds (2021) 857, 157811.
[5] R. Martinho Vieira et al., Materials Research Letters (2022) 3, 156.
[6] A. Jain et al, APL Materials (2013), 1, 11002.