ConspectusAdiabatic demagnetization refrigeration (ADR), which exploits the magnetocaloric effect (MCE), remains the only helium-free refrigeration technology capable of reaching temperatures below 1 K. With the rapid growth of quantum computing and astronomical observation, there is a pressing need for large-capacity ADR systems─underscoring the critical demand for magnetic refrigerants capable of generating substantial magnetic entropy changes (−ΔSm) at millikelvin temperatures. However, a long-standing challenge persists: achieving both large −ΔSm values and low magnetic ordering temperatures (T0). This trade-off has limited the performance of existing refrigerants in the sub-Kelvin regime, thereby hindering ADR advancement. To address this, our work focuses on the rational design of next-generation refrigerants. Rather than relying on readily available materials, we emphasize the importance of tuning key magnetic parameters, T0, exchange and dipolar interactions. This approach began with the discovery that incorporating fluoride (F–) bridges into antiferromagnetic frameworks transforms their magnetic behavior, shifting it from antiferromagnetic to weak ferromagnetic and lowering T0 from 1.4 K to 1.0 K while yielding large −ΔSm values. Building on this concept, we extended the strategy to the Gd(OH)3–xFx. In particular, Gd(OH)F2 combines weak magnetic interactions with high magnetic density, achieving record −ΔSm values and demonstrating a robust route to simultaneously enhance −ΔSm and suppress T0. By integrating the mean-field approximation with quantum Monte Carlo (QMC) simulations, we accurately predicted T0 values in systems with weak exchange and low anisotropy, such as GdCO3F, Gd(HCOO)F2, Gd2(SO4)3·8H2O, GdF3 and Gd(HCOO)3, revealing the dominant role of dipolar interactions in determining T0. Expanding this framework, we introduced Gd3+ into Yb-based compounds characterized by intrinsically weak exchange and dipolar interactions. This doping strategy enabled the realization of low T0 and large −ΔSm values at ultralow temperatures, demonstrating that a balance between competing magnetic interactions and chemical disorder can achieve the coexistence of high −ΔSm and low T0. To translate these insights into practical ADR systems, we synthesized LiGd0.1Yb0.9F4 and investigated its performance in an ADR setup. Notably, LiGd0.1Yb0.9F4 cooled a test sample to 160 mK and delivered a specific cooling capacity more than twice that of the commercial refrigerant CrK(SO4)2·12H2O. Additionally, by integrating high magnetic density with weak exchange and dipolar interactions in a frustrated magnet, we developed KYb3F10, which achieved significant −ΔSm with a low T0. Quasi-adiabatic demagnetization experiments with KYb3F10 reached a minimum temperature of 27.2 mK, highlighting its promise as a next-generation ADR refrigerant. In summary, we proposed and validated rational design strategies for high-performance ADR refrigerants, achieving enhanced −ΔSm across the cryogenic temperatures below 4 K. From theoretical modeling to experimental realization, our work lays a solid foundation for advancing ADR technologies in both fundamental research and applied low-temperature systems.
Xu et al. (Fri,) studied this question.