Decoding Halide Perovskite for Neuromorphic and Memristive Devices

Abstract

The von Neumann architecture serves as the foundation for computers by storing data and instructions in the same memory space; however, it limits the data transfer between the CPU and memory. The human brain is an avant-garde organic machine that connects electrical systems with its network of neurons and synapses. We have 80-100 billion neurons, each connected to >1000 other neurons called synapses, thus totalling 100 trillion that excel in our decision-making and learning. Neuromorphic engineering aims to create brain-like devices that operate effectively with low power consumption to supersede von Neumann for faster computation. Despite their efficacy, neuromorphic chips built with CMOS circuits are complex in replicating biological processes. Neuromorphic computing led to the development of memristors to improve performance, flexibility, and scalability. Wonder materials like halide perovskite with both ionic and semiconductive properties mimic synaptic behaviour. Halide perovskites have exceptional ion transport properties, enabling rapid resistive switching for neuromorphic advancement, and further, respond to various stimuli like light and temperature, offering the potential for emulating complex synaptic behaviours. Halide perovskite can modulate functionalities through structural variations, and dimension reduction endorses versatility in neuromorphic computing and future semiconducting technology. Further, we uncover the mechanisms through which halide perovskite emulates synaptic functions in neuromorphic systems.