Perovskite based Low Power Synaptic Memristor Device for Neuromorphic application

Abstract

In artificial intelligence, high speed neuromorphic computing architectures are needed to perform various operations such as learning, transferring information, and processing of data. Due to high power dissipation, high operating energy, and lower density of integration CMOS device has limited application in neuromorphic computing in nanoscale domain. On the other hand memristor devices are promising candidates for implementing synaptic devices in a neuromorphic computing architecture due to their swift information storage, high-speed processing of data and high density with lower power consumption. To the best of our knowledge this paper proposes the first studies made on a perovskite $\left( C H _ { 3 } N H _ { 3 } P b I _ { 3 } \right)$ based photovoltaic memristive device with $I T O / S n O _ { 2 } / C H _ { 3 } N H _ { 3 } P b I _ { 3 } / A u$ structure in the dark condition. This perovskite based memristor is able to mimic the neuromorphic learning and remembering process same as the biological synapses. The proposed synaptic memristor device has potential to operate at low energy, low cost, solution processability, low activation energy, high efficiency and used as a power-on-chip synaptic device in artificial neural network.