Elucidating Ionic Programming Thermodynamics of Oxide-Ion-Based Electrochemical Random-Access Memory Devices.

Electrochemical random-access memory (ECRAM) enables analog conductance modulation by tuning the carrier concentration in the channel, making it a promising hardware platform for neuromorphic computing. Consequently, elucidating the quantitative relationship between the concentration of ionic defects and the conductivity of the channel material is essential for understanding the synaptic behavior of ECRAMs, which is still largely unexplored. To bridge this gap, we employ La0.6Sr0.4FeO3-δ (LSF) as a model system to quantitatively correlate the applied electrochemical driving force, ionic defect concentration, and modulated electrical conductance. Leveraging this correlation, we fabricate LSF-based ECRAMs with high linearity, symmetry, and stable retention. To better understand the underlying principles in the electronic structures of LSF tuned by ionic defects, we further use operando ambient-pressure x-ray photoemission spectroscopy (AP-XPS) to show that the evolution of conductivity and the shift of the Fermi level occur in high synchronization during conductance modulation. We show that defect chemistry offers a powerful theoretical basis for understanding and predicting the dynamic behavior and informs the rational design of device-level engineering strategies of ECRAMs.