Bayesian Optimization of Cold Sintering Process for High-Conductivity LATP Solid-State Electrolytes

Lithium aluminum titanium phosphate (LATP) is a NASICON-type solid-state electrolyte with favorable chemical stability and relevance for safe battery technologies. However, conventional pellet fabrication involves sintering at 800–1000 °C often results in lithium loss and secondary-phase formation, while also increasing energy consumption. The cold sintering process (CSP) provides a low-temperature alternative (<400 °C), yet the coupled effects of temperature, pressure, dwell time, and water content make identification of suitable processing conditions challenging. In this work, LATP was intentionally processed using only CSP, without pre- or post-heat treatment and without additives beyond water, to evaluate the feasibility of a simplified and low-energy fabrication route. The CSP parameter space was systematically explored using Bayesian optimization as an experimental design strategy. Under optimized conditions, CSP alone yielded LATP with ionic conductivity up to 1.94 × 10 −4 S/cm. Incorporation of experimental data from two CSP-based preparation routes further expanded accessible high-conductivity regimes and improved the efficiency of parameter discovery. Microstructural analysis revealed that while densification is necessary, porosity and intergranular connectivity critically govern Li-ion transport, underscoring the importance of balanced process control. These findings demonstrate that appropriately optimized CSP is sufficient to produce high-conductivity LATP at substantially reduced temperatures, establishing a viable low-temperature and additive-free processing pathway for NASICON-type solid-state electrolytes.