| Abstract Scope |
The exceptional resistive switching characteristics of complex oxides make them lucrative for memristors – a key building block of emerging neuromorphic computers. Nevertheless, precise control over neural functionalities remains challenging due to paucity of fundamental understanding of defect evolution over nano-to-mesoscopic length/timescales that underlie ultrafast electronic transitions in complex oxides, especially under applied electric field. Here, we integrate density functional theory (DFT) calculations, ab initio/classical molecular dynamics (AIMD/CMD) simulations, machine learning, precision synthesis, and multi-modal X-ray imaging to address this critical knowledge gap. Specifically, we will highlight our recent successes of this approach in (a) designing WO3-x electrodes that enable controlled electroforming owing to localized electric field, and minimize the intrinsic variability in memristors, and (b) elucidating key correlations between oxygen stoichiometry, distribution of oxygen vacancies, and electron-lattice coupling in rare-earth perovskite nickelates. We will discuss these findings in the context of designing reliable memristive devices for artificial intelligence technologies. |