Assessing Correlations Between Phonon Features and Cation Migration Barriers in Multivalent Solid Electrolytes

Batteries based on the redox of multivalent cations (Mg2+, Ca2+, Zn2+, Al3+, etc.) offer potential advantages over today’s lithium-ion batteries, but their development is hindered by the sluggish migration of such ions in solid electrodes and electrolytes. Computational screening can accelerate the discovery of more conductive materials, provided that ionic conductivity can be estimated with sufficient accuracy and efficiency. The present study examines whether vibrational properties can be used to predict energetic barriers for cation migration in 24 prototypical multivalent solid electrolytes. Phonon band centers (i.e. mean frequencies), which have been previously used to predict Li-ion conductivity, are calculated using density functional theory. Band centers alone are found not to correlate with migration barriers (R^2 = 0.02), perhaps due to poor alignment of low-frequency phonon eigenmodes with ion migration pathways in some materials. A new metric that incorporates both frequencies and alignments—the mean alignment-weighted frequency—is more strongly correlated with migration barriers (R^2 = 0.25). Materials in this study with the lowest migration barriers consistently exhibit the lowest mean alignment-weighted frequencies, suggesting the utility of this metric for filtering out materials with high barriers in screening efforts. Comparisons to previous studies suggest that phonon band centers may be correlated with migration barriers only in compositionally similar materials, and that adding alignment information may enable more reliable predictions among more diverse sets of materials. These results quantify the promise of using phonon frequencies and alignments, perhaps in combination with other properties, to efficiently screen for materials with high multivalent ionic conductivity.