Deependra Singh1,2,Pukhraj Prajapat1,S.B. Krupanidhi2,Karuna Nanda2,3,Govind Gupta1
CSIR-National Physical Laboratory1,Indian Institute of Science2,Institute of Physics3
Deependra Singh1,2,Pukhraj Prajapat1,S.B. Krupanidhi2,Karuna Nanda2,3,Govind Gupta1
CSIR-National Physical Laboratory1,Indian Institute of Science2,Institute of Physics3
The successful implementation of artificial synaptic devices in neuromorphic computing systems requires the demonstration of various synaptic features with low energy consumption. Herein, a photonic artificial synapse based on molecular beam epitaxy-grown <i>c</i>-GaN thin film along with its photo memristive properties has been presented. The behavior of the metal-semiconductor-metal device is similar to that of a biological synapse. The adaptable synaptic neuromorphic functionalities such as short-term memory (STM) and long-term memory (LTM), potentiation, and paired-pulse facilitation have been successfully mimicked. In this artificial synapse, STM, LTM, and the transition between STM–LTM behavior have been demonstrated by modulating the number of optical spikes (355 nm). The device also exhibits memristive properties under illumination of 355 nm laser light, with a high resistance state/low resistance state ratio of ~10 at -0.6 V, along with a high cyclic stability for 500 cycles. The synaptic properties have been attributed to the inherent persistent photoconductivity performance, whereas the memristive behavior is due to the volatile resistive switching characteristics. The results pave the way to developing bio-realistic neuromorphic devices based on wide-bandgap semiconductors.