Jayakanth Ravichandran1
University of Southern California1
Jayakanth Ravichandran1
University of Southern California1
Neuromorphic computing has gained a lot of interest for the potential to significantly improve the energy efficiency of computing associative tasks. The two key components of this computing methodology focus on mimicking the operation of neurons and synapses. Phase change materials that undergo abrupt volatile and non-volatile changes in conductivity as a function of temperature, and applied electric field are prime candidates to mimic neurons and synapses in an energy and footprint efficient manner. In this talk, I will introduce a novel phase change material, BaTiS<sub>3</sub>, which belongs to a broader class of phase change materials known as quasi-1D hexagonal chalcogenides. This material demonstrates two properties can be leveraged to mimic simplified forms of neuronal and synaptic processes. At room temperature, the material shows resistive switching behavior, presumably the switching of the ferrielectric dipole moments. Below ~180 K, the material shows voltage tunable and stable changes in conductivity, suggesting the possibility of multistep memory characteristics. The former can be leveraged to mimic low power neurons, and the later as tunable synaptic weights in a neuromorphic computing architecture. This work setups up the potential for novel materials research in creating functionalities necessary to achieve neuromorphic hardware.<br/><br/>Reference:<br/>1. H. Chen, B. Ilyas, B. Zhao, E. Ergecen, J. Mutch, G. Y. Jung, Q. Song, C. A. Occhialini, G. Ren, S. Shabani, E. Seewald, S. Niu, J. Wu, N. Wang, M. Surendran, S. Singh, J. Luo, S. Ohtomo, G. Goh, B. C. Chakoumakos, S. J. Teat, B. Melot, H. Wang, D. Xiao, A. N. Pasupathy, R. Comin, R. Mishra, J. -H. Chu, N. Gedik, J. Ravichandran, “Unconventional Charge-density-wave Order in a Dilute <i>d</i>-band Semiconductor”, https://arxiv.org/abs/2207.11622 (2022).