Multinary hydrides offer great promise for achieving high-temperature conventional superconductivity under ambient pressure. However, the rich variety of structural configurations in hydrides poses a formidable challenge for searching their immense phase space. In this work, we used our developed artificial intelligence (AI) search engine (InvDesFlow) to perform extensive investigations regarding ambient stable superconducting hydrides. We assessed the thermodynamic stability of hydrides by our AI model. For ${\mathrm{K}}_2{\mathrm{GaCuH}}_6$, the AI-predicted formation energy difference along the synthesis path is 0.146 eV/atom, closely matching the density functional theory result of 0.137 eV/atom. This agreement confirms the thermodynamic stability of hydrides and validates the reliability of AI method. Several quaternary hydrides with high superconducting temperature ($T_c$) are predicted. In particular, the superconducting $T_c$ of ${\mathrm{K}}_2{\mathrm{GaCuH}}_6$ and ${\mathrm{K}}_2{\mathrm{LiCuH}}_6$ are calculated to be 68 and 53 K under ambient pressure, respectively, which shows a significant enhancement in comparison with that of ${\mathrm{K}}_2{\mathrm{CuH}}_6$ ($T_c\sim 18\text{K}$). We also find that intercalating atoms could cause phonon softening and induce more phonon modes with strong electron-phonon coupling. Hence, we propose that intercalating atoms is a feasible approach in searching for superconducting quaternary hydrides.

Close

PrevNext

View figure in article

• Superconductivity
• Superconductors
• Hydrides
• Artificial intelligence
• Deep learning
• Density functional calculations