Electrochemical ohmic memristors for continual learning

Research Square

Abstract

Abstract Developing versatile and reliable memristive devices is crucial for advancing future memory and computing architectures. The years of intensive research have still not reached and demonstrated their full horizon of capabilities, and new concepts are essential for successfully using the complete spectra of memristive functionalities in industrial applications. The physicochemical complexity of these nanoscale systems makes control over performance and functionalities difficult where fundamental interactions and mechanisms are not fully understood. Here, we report on the discovery of a new switching mechanism that in contrast to other memristive devices uses low-work-function electrodes to create metal/oxide interfaces with minimal Schottky barrier heights. The novel two-terminal Ohmic memristor operation is based entirely on localized electrochemical redox reactions. The device is characterised by essential advantages such as ultra-stable binary and analogue switching with high OFF/ON ratio, broad voltage stability window, low forming voltages and high temperature stability. We demonstrate the multifunctional properties enabled by the new mechanism can be effectively used to overcome the catastrophic forgetting problem as a significant and fundamental issue in conventional deep neural networks as connectionist models. Our findings represent a new milestone in the resistive switching fundamentals and provide a new approach for the design of a memristive system, expanding the horizon of functionalities, enabling more effective emulation of the metaplasticity concept in neuroscience.

Authors 6

1. Ilia Valov Aachen Institute of Materials in Electrical Engineering 2 (IWE2)

   RWTH Aachen University · Forschungszentrum Jülich

   Affiliation as printed

   Forschungszentrum Juelich

   Institute of Materials in Electrical Engineering 2 (IWE2), RWTH Aachen University, Aachen, Germany

2. Shaochuan Chen Aachen Institute of Materials in Electrical Engineering 2 (IWE2)

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University

   Institute of Materials in Electrical Engineering 2 (IWE2), RWTH Aachen University, Aachen, Germany

3. Zhen Yang

   Peking University

   Affiliation as printed

   Peking University

   Beijing Advanced Innovation Center for Integrated Circuits, School of Integrated Circuits, Peking University, Beijing, China

4. Heinrich Hartmann

   Forschungszentrum Jülich

   Affiliation as printed

   Research Centre Juelich

5. Astrid Besmehn

   Forschungszentrum Jülich

   Affiliation as printed

   Research Centre Juelich

6. Yuchao Yang

   Peking University · Chinese Institute for Brain Research

   Affiliation as printed

   Peking University

   Beijing Advanced Innovation Center for Integrated Circuits, School of Integrated Circuits, Peking University, Beijing, China

   Center for Brain Inspired Chips, Institute for Artificial Intelligence, Peking University, Beijing, China

   Center for Brain Inspired Intelligence, Chinese Institute for Brain Research (CIBR), Beijing, China

   School of Electronic and Computer Engineering, Peking University, Shenzhen 518055, China