Here we will be presenting a memristor based on the back-end-of-line (BEOL) amorphous SiC material. The memristors demonstrate excellent binary resistive switching with compliance-free and self-rectifying characteristics which are advantageous for the implementation of high density 3D crossbar memory architectures. The conductance of our SiC based memristor can be modulated gradually through the application of both DC and AC signals. This behaviour is demonstrated to emulate several vital synaptic functions including paired-pulse facilitation, short-term potentiation, and spike-rate-dependent plasticity. The synaptic function of learning-forgetting-relearning processes are successfully emulated and demonstrated using a 3 × 3 array. In addition, we will also demonstrate the modulation of the neuromorphic behaviour through controlling the amorphous SiC film properties. The switching mechanism will also be discussed.
More importantly, our memristor features temporally stored conductance states which decay over time. This short-term memory behaviour can be used to directly process temporal data in recurrent neural networks and is particularly attractive for applications such as speech recognition, classification and time series forecasting. We have implemented a physical reservoir computing system using our SiC-based memristor as the reservoir. This physical reservoir computing system has been experimentally demonstrated to perform the task of pattern recognition. After training, our RC system has achieved an 100% accuracy in classifying number patterns from 0 to 9 and demonstrated good robustness to noisy pixels.