KY EPSCoR: FY23 Department of Energy EPSCoR State Match: University of Louisville Subawards

Controlling reversible phase transitions in rare-earth nickelates for novel memory devices Badri Narayanan, Mechanical Engineering, University of Louisville (Principal Investigator) Dillon Fong, Materials Science Division, Argonne National Laboratory (Collaborator) Subramanian KRS Sankaranarayanan, Center for Nanoscale Materials, Argonne National Laboratory (Collaborator) Resistive switching in strongly correlated perovskite rare-earth nickelates is lucrative for emerging applications in neuromorphic computing, and densely-scaled non-volatile memory. In particular, colossal changes in electrical conductance (up to 7 orders of magnitude) can be achieved in these complex oxides by introducing oxygen vacancies via a combination of crystal field splitting and filling-controlled Mott- Hubbard electron-electron correlations in the Ni 3d orbitals. More importantly, resistance states can be manipulated by controlling the spatial distribution of oxygen vacancies, which can migrate under applied bias. Nevertheless, the promise of using defect-driven Mott transitions in these emergent materials for a new paradigm in energy-efficient computing is far from being realized. Most of the challenges thwarting progress stem from a lack of fundamental understanding of the dynamical processes that determine the spatio-temporal evolution of oxygen vacancies (and other extended defects) over nano-to-mesoscopic length/timescales under applied electric field. Here, we propose to address this knowledge gap by developing accurate classical interatomic models for rare earth nickelates by using data-centric machine learning (ML) methods on a large dataset derived from first principles. Such models can accurately describe the correlations between subtle structural distortions and the oxidation state of Ni, treat localized charge carriers, and give good descriptions of defect/ion transport, chemical reactions, and microstructural evolution. In collaboration with precision synthesis, multi-modal X-ray imaging experiments and machine learning-based image analysis at Argonne National Laboratory, atomistic simulations based on these newly developed models will greatly advance the current understanding of microstructural evolution in rare-earth nickelates under applied bias. Such an understanding will enable precise control over hierarchical defect structures and unravel new routes to manipulate resistance states in rare-earth nickelates. This, in turn, will lead to the design of novel neuromorphic devices with a prescribed set of neural functionalities, high-speed densely-scaled resistive random access memory (RRAM) technology, and other emerging microelectronic devices. Successful execution of this project will also stimulate local economic development in Louisville through collaboration with local companies engaged in the development of energy-efficient microelectronics. Finally, the diverse and collaborative environment associated with this project will enable successful training of graduate students and postdoctoral researchers in various facets of strongly correlated materials and significantly contribute to their professional development.