Origin of universal H/T scaling in disordered frustrated magnets

In $\mathrm{H_3LiIr_2O_6}$ and several other disordered frustrated magnets the data obey $C/T \sim T^{-\gamma} F_C(\mu_B H/k_B T)$ and $M \sim H^{(1-\gamma)} F_M(\mu_B H/k_B T)$ with a material-dependent $\gamma$ between 0 and 1 (Kimchi, Sheckelton, McQueen, Lee, Nature Communications 2018), attributed to a power-law distribution of singlet energies $P(J) \sim J^{-\gamma}$. Derive $\gamma$ and the scaling functions $F_C, F_M$ from a microscopic disordered model of at least one compound, and determine whether $\gamma$ is fixed by a 2D random-singlet fixed point or is nonuniversal. An answer is a model whose computed $\gamma$ and $F_C$ agree with the data without fitting $\gamma$.

Strong-disorder renormalization or quantum Monte Carlo of a calibrated random-exchange model that reproduces the measured $\gamma$ and scaling functions of one compound.