Advancements in Memristor Device Development: From Material Stack Optimization to Physical Modelling

Abstract

Analog and parallel processing architectures are promising computing paradigms to implement energy-efficient artificial intelligence (AI) hardware systems that resemble biological brains. As result, the development of novel materials/devices and integrating them to neural and synaptic functionalities have been successfully demonstrated by showing analog signal processing of neuron and synapse devices (e.g., oscillatory neurons and analog in-memory computing of synapse arrays) [1]. This presentation delves into the pivotal role of material science in shaping various devices based on Phase Change Materials and Oxide Resistive Random-Access Memories (RRAMs), fundamental components for emulating artificial neural and synaptic functions in neuromorphic computing [2,3]. These devices, grounded in diverse physical mechanisms and materials, necessitate comprehensive understanding through experimental and theoretical approaches for optimization and integration into higher layers of computer architecture. Their rich spectrum of switching and oscillating behaviors, governed by a myriad of factors such as material compositions, fabrication techniques, and device morphologies, offers substantial design flexibility to tailor devices to specific application requirements. We will see examples of how material and device engineering can lead to breakthroughs in device performance. Nonetheless, device variability among other non-idealities still hinders the hardware scaling to large networks required to solve more complex AI tasks. Challenges at the device and hardware level may also be overcome through a complementary research effort to develop more robust, hardware-friendly algorithms and computational models that could compensate for the variability issues of devices. [1] A. Sebastian et al., "Memory devices and applications for in-memory computing", Nature Nanotechnology, 15, 529-544, 2020. [2] From: https://spectrum.ieee.org/analog-ai. [3] G. Csaba and W. Porod, "Coupled oscillators for computing: A review and perspective," Applied Physics Review, 3 January 2020.