Interfacial Switching-Based, Bioinspired, Highly Stable, and Reliable Synapse for Neuromorphic Applications

In this study, an interfacial switching (IFS)-based bioinspiredTiN/AlON/TaON/Pt resistive random-access memory (RRAM) device wasfabricated to investigate its conduction mechanism and synaptic behaviorfor neuromorphic computing. This device exhibited excellent dc enduranceover at least 10000 cycles, an ac pulse endurance of 1 M, and long-termretention (10(6) s) at 150 & DEG;C with no degradation. Thedevice also showed multilevel characteristics. The RESET stop voltageof the device varied from -0.9 to -1.3 V. The devicewas highly stable over 250 potentiation and depression cycles, whichare crucial in Hopfield neural network (HNN)-based neuromorphic systems.High nonlinearity (1.13 for potentiation and -1.75 for depression)was achieved using the device potentiation and depression functions.Experimental potentiation and depression data were used to train anHNN based on the fabricated device to recognize an input image of28 x 28 pixels that contained 784 synapses. The HNN had a trainingaccuracy of higher than 93% in 22 iterations. The experimental resultsindicate that the fabricated IFS-based TiN/AlON/TaON/Pt RRAM deviceis highly suitable for neuromorphic devices that mimic synaptic characteristicsfor neuromorphic systems.