Understanding Synaptic Mechanisms in SrTiO3 RRAM Devices

In this paper, we investigated the performance of SrTiO3 (STO) resistive random-access memory (RRAM) as a candidate to realize short-term synaptic elements in neuromorphic circuits. Ionic defect movement within oxides is postulated to play a significant role in resistive state reconfiguration of these devices. This relatively small movement in oxides was sensed by developing a novel measurement technique using differential voltammetry. Using this technique, the widely hypothesized belief of drift and diffusion of ionic defects in oxides was systematically studied and linked to the state reconfiguration in STO RRAM. Next, the devices’ state retention was examined and a consistent decay of the resistive state was observed. The decay rate was found to be a function of the conditioning voltage, allowing us to control the state decay via operational voltages. We then developed a test battery that quantified how these devices operate within a spike-rate encoded learning paradigm. By utilizing a carefully tuned test battery, we were able to quantify the magnitude of potentiation and depression due to varying spike amplitudes and spike rates. These tests provide a systematic methodology to capture the constantly changing resistive state of short-term synaptic elements.