CAREER: Decipher the Mechanism of High-performance Novel Memristors by Phase-field Simulation

University Of Texas At Arlington, Arlington TX

Investigators

Abstract

NONTECHNICAL SUMMARY This award supports integrated research, educational, and outreach efforts on understanding and designing high performance memristors as potential device candidates for next generation data-centric computing applications. Memristor is a type of material device that can change its resistance states under external fields and maintain such changes via the formation and evolution of electrically conductive channels in the host materials. However, high-performance memristors that exhibit low-energy and gradual resistive switching, good retention, and uniformity in switching in a single device have not yet materialized. In this project, the PI and his team aim to develop a computational framework to elucidate the key factors of materials used in memristors as well as the roles of ion migration, electrochemical reaction, surface energy, and mechanical strain that can potentially be tuned to enable low-current and gradual switching. Based on this, the PI will identify new mobile species/host materials and predict microstructures that can achieve gradual, uniform, and stable switching performances. These performances are key to memristors used for accurate and efficient neural network training and analog computing. Knowledge obtained from this study will provide guidance to the experimental synthesis, characterization, and testing of novel memristors, which will be used to further benchmark the computational approach. To integrate the educational and outreach activities with the research works, the PI will compile the developed computational tools into a user-interface software and disseminate it to the broader research community. The project will encourage underrepresented minorities, especially Hispanic and first-generation college students in the Dallas−Fort Worth area, to pursue science and engineering related projects through the proposed “Mentor−Mentee” program. The K-12 outreach activities will involve continued collaborations with local universities through leveraging the Texas Pre-Freshman Engineering Program and be carried out at the annual “Engineers Week” at local museums and libraries, promoted by the National Society of Professional Engineers, to attract talented students into STEM field. TECHNICAL SUMMARY This award supports integrated research, educational, and outreach efforts on developing and applying computational models to understand a novel resistive switching mechanism in Ruthenium-based memristor, and to identify new mobile species and host materials to achieve needed performance. Memristors, which can change their electrical resistance and maintain such changes, enable in-memory computing, making them promising candidates for next-generation data-centric computing applications, such as nonvolatile memory, neural network training, and analog computing etc. However, high-performance memristors that exhibit low-energy and gradual resistive switching, good retention, and uniformity in switching have not yet been achieved, and the mechanisms that underpin these switching behaviors are not fully understood. The main objective of this research is to establish accurate theories and advance the knowledge of the mechanism of these highly desirable resistive switching, and to identify potential mobile species and host materials that enable these needed performances which are unachievable with existing memristors. It is hypothesized that the interactions among reduction-oxidation reaction, charge transport, and interfacial and strain effects result in a dynamic free energy landscape of the memristor system, which can be engineered by selecting proper mobile species and designing the microstructure of the switching layer to achieve needed switching performance. This project will integrate the experiment-validated phase-field model to understand both gradual and sudden switching dynamics driven by the chemical, electrical, interfacial, and mechanical energy competitions, the high-throughput calculations and machine learning to identify new mobile species/host materials, and the combined atomistic and mesoscale modeling of the microstructure of the host material and its effect on the stability and uniformity of resistive switching. Knowledge obtained from this theoretical study will guide experimental synthesis, characterization, and measurement of the novel memristors, which will be used to further benchmark and complement the simulation work. To integrate the educational and outreach activities with the research works, the PI will compile the developed computational tools into a user-interface software and disseminate it to the broader research community. The project will encourage underrepresented minorities, especially Hispanic and first-generation college students in the Dallas−Fort Worth area, to pursue science and engineering related projects through the proposed “Mentor−Mentee” program. The K-12 outreach activities will involve continued collaborations with local universities through leveraging the Texas Pre-Freshman Engineering Program and be carried out at the annual “Engineers Week” at local museums and libraries, promoted by the National Society of Professional Engineers, to attract talented students into STEM field. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.