2026-10-19 · arXiv Quantum Physics
arXiv:2603.17194v2 Announce Type: replace-cross
Abstract: We introduce a first-principles-guided framework for predicting statistical TLS distributions from mobile impurities in disordered materials. Machine-learned interatomic potentials, with DFT and minimum-action-path validation, connect impurity landscapes to tunnel-splitting and asymmetry distributions, TLS densities, and dielectric loss. Applied to amorphous Nb$_2$O$_{5-\delta}$ and Ta$_2$O$_{5-\delta}$, the framework implicates interstitial hydrogen as a source of microwave loss. Loss tangents match Nb-pentoxide measurements at inferred H concentrations, while weaker H uptake in Ta reproduces the Nb/Ta TLS contrast.
原文链接:https://arxiv.org/abs/2603.17194
查看更多资讯请见 AI 行业资讯。