Machine-learning-assisted discovery of lattice dynamics signatures of sodium superionic conductors

Sodium superionic conductors are key to the development of all-solid-state sodium batteries. Discovery of new superionic conductors has traditionally relied on insights from material defect chemistry and the transition/hopping theory, while the role of lattice vibrations, <i>i.e.</i>, phonons, remains underexplored. We identify key lattice dynamics signatures that govern ionic conductivity by analyzing the phonon mean squared displacement (MSD) of Na⁺ ions. By high-throughput screening of a dataset of 3903 Na-containing structures, we establish a strong positive correlation between phonon MSD and diffusion coefficients, providing a quantitative correlation between lattice dynamics and ion transport. To accelerate this discovery, we incorporate machine learning (ML) into our screening workflow, using phonon-derived descriptors to rapidly predict ionic transport properties across a broad structural space. Our findings reveal that low acoustic cutoff phonon frequencies, low center vibrational density of states of Na⁺ ions, slightly higher than the acoustic cutoff frequencies, and enhanced low-frequency vibrational coupling between Na⁺ ions and the host sublattice promote superionic conductivity. Phonon mode analysis further demonstrates that only a small subset of low-frequency acoustic and optic modes contribute dominantly to large phonon MSDs and Na⁺ ion migration, while higher-energy modes contribute negligibly. These insights enable the integration of lattice dynamics descriptors, phonon MSD, Na⁺ VDOS center, acoustic cutoff frequency, and low-frequency phonon coupling into machine learning frameworks, accelerating the discovery and rational design of high-performance sodium superionic conductors.