A Novel Memristor Design based on Hydrogen Movement on Silicon Surfaces

As edge computing advances, memristors are needed for enhanced performance and scalability with minimal energy, but precise resistance switching is a challenge. Silicon (Si) surfaces in a proton-conducting polymer can control memristor behaviour by depositing hydrogen (H) atoms, which increase conductivity via mid-gap traps, forming a memristor unit cell. This study investigates the critical function of H atoms by deliberately altering their positioning and concentration in silicon-based memristor devices. Using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS), we investigate the impact of temperature (T) and electric field (EF) on H migration. We define a polygonal volume of Si and deposit H atoms on its top surface. We apply T and EF to observe diffusion and drift of H atoms. The hopping rate depends on applied T and EF. We establish a relationship between the three-dimensional velocity of H and applied T and EF. We simulate several movement pathways of H atoms over time under varying T and EF acting separately or simultaneously. Therefore, we can determine the required magnitude and direction of EF and T to achieve the desired H position and concentration. Our approach aims to improve edge computing and neuromorphic computing by enhancing memristor functionality. Despite potential scalability issues and the need for precise hydrogen control, this research offers valuable insights into modifying memristor electrical characteristics for low-energy silicon-based neuromorphic devices.