Long-term memory performance with learning behavior of artificial synaptic memristor based on stacked solution-processed switching layers

In this work, oxide resistive random access memory (OxRRAM) devices with stacked solution-processed (SP) metal oxide (MO) layers were fabricated to investigate artificial synaptic behavior such as long-term potentiation (LTP) and long-term depression (LTD). The stacked RRAM devices exhibited stable and repeated bipolar IV curves with operation voltage lower than the ~0.5 V and a switching ratio larger than 2*104. Also, with the stimuli from external consecutive pulses, the stacked devices demonstrated learning-forgetting-relearning behavior similar to neuron-induced behavior in the human brain. Finally, based on stable long-term memory performance, the pattern recognition system with an artificial neuron network (ANN) algorithm was simulated with the recognition accuracy higher than 95%.