Predicting novel superconducting hydrides using machine learning approaches

The search for superconducting hydrides has, so far, largely focused on finding materials exhibiting the highest possible critical temperatures (${T}_{c}$). This has led to a bias toward materials stabilized at very high pressures, which introduces a number of technical difficulties in experiment. Here we apply machine learning methods in an effort to identify superconducting hydrides that can operate closer to ambient conditions. The output of these models informs subsequent crystal structure searches, from which we identify stable metallic candidates prior to performing electron-phonon calculations to obtain ${T}_{c}$. Hydrides of alkali and alkaline earth metals are identified as especially promising; of particular note, a ${T}_{c}$ of up to 115 K is calculated for ${\mathrm{RbH}}_{12}$ at 50 GPa, which extends the operational pressure-temperature range occupied by hydride superconductors toward ambient conditions.