Predicting solid state material platforms for quantum technologies

Semiconductor materials provide a compelling platform for quantum technologies (QT), and the properties of vast amount can be found in databases containing information from both experimental theoretical explorations. However, searching these to find promising candidate technology applications is major challenge. Therefore, we have developed framework automated discovery semiconductor host platforms QT using material informatics machine learning methods, resulting dataset consisting over $25.000$ nearly $5000$ physics-informed features. Three approaches were devised, named Ferrenti, extended Ferrenti empirical approach, label data supervised (ML) methods logistic regression, decision trees, random forests gradient boosting. We that three, approach relying exclusively on findings literature predicted substantially fewer candidates than other two with clear distinction between suitable unsuitable when comparing largest eigenvalues covariance matrix. In contrast expectations focusing band gap ionic character, ML highlighted features related symmetry crystal structure, including bond length, orientation radial distribution, as influential predicting QT. All three all four agreed subset $47$ eligible %(to probability $>50 \ \%$) $8$ elemental, $29$ binary, $10$ tertiary compounds, basis further explorations towards technology.