Self-rectifying resistive memory in passive crossbar arrays

Abstract

Conventional computing architectures are poor suited to the unique workload demands of deep learning, which has led to a surge in interest in memory-centric computing. Herein, a trilayer (Hf0.8Si0.2O2/Al2O3/Hf0.5Si0.5O2)-based self-rectifying resistive memory cell (SRMC) that exhibits (i) large selectivity (ca. 10(4)), (ii) two-bit operation, (iii) low read power (4 and 0.8 nW for low and high resistance states, respectively), (iv) read latency (<10 <mu>s), (v) excellent non-volatility (data retention >10(4)s at 85 degrees C), and (vi) complementary metal-oxide-semiconductor compatibility (maximum supply voltage <= 5V) is introduced, which outperforms previously reported SRMCs. These characteristics render the SRMC highly suitable for the main memory for memory-centric computing which can improve deep learning acceleration. Furthermore, the low programming power (ca. 18 nW), latency (100 mu s), and endurance (>10(6)) highlight the energy-efficiency and highly reliable random-access memory of our SRMC. The feasible operation of individual SRMCs in passive crossbar arrays of different sizes (30x30, 160x160, and 320x320) is attributed to the large asymmetry and nonlinearity in the current-voltage behavior of the proposed SRMC, verifying its potential for application in large-scale and high-density non-volatile memory for memory-centric computing. Memory-centric computing refers to computing designs where the memory, rather than the processor is central in the architecture. Here, the authors demonstrate a self-rectifying resistive memory cell that exhibits impressive endurance, and low power consumption, making it suitable for memory-centric applications.