Anion Topology and Cation–Anion Coupling Govern Phase Transitions and Ionic Transport in Closo-Hydroborate Solid Electrolytes

High-temperature closo-hydroborate phases have emerged as promising solid electrolytes owing to their excellent thermal stability and high ionic conductivity. However, stabilizing these high-temperature phases at ambient conditions remains a significant challenge. A fundamental understanding of the phase-transition mechanisms and their correlation with ionic transport behavior is therefore crucial for guiding the design of closo-hydroborate electrolytes with enhanced room-temperature performance. Here, we developed a machine learning force field (ML-FF) for both Na2B10H10 and NaCB9H10 via active learning, enabling large-scale molecular dynamics (MD) simulations with near-density functional theory (DFT) accuracy. The ML-FF predicts phase transitions at ∼410 K for Na2B10H10 and ∼318 K for NaCB9H10. Na2B10H10 transforms from a low temperature monoclinic phase to a high ionic conducting fcc phase by anion orientational disorder, whereas NaCB9H10 undergoes a diffusionless bcc→hcp lattice rearrangement. In both materials, Na+ occupies tetrahedral interstitial sites, but in NaCB9H10 roughly half of those sites are vacant, forming a three-dimensional Na+ migration network. Consequently, NaCB9H10 exhibits ∼4× higher Na+ conductivity and ∼3× faster anion rotation than Na2B10H10. These findings reveal how anion topology and cation–anion coupling govern ionic transport, and demonstrate the effectiveness of ML-FF in uncovering mechanistic insight into complex solid electrolytes.