Reconfigurable artificial synapse and neuron based on 2D organic-inorganic hybrid perovskite memristor

Memristors have attracted significant attention for their unique conductive characteristics and potential in neuromorphic computing. Reconfigurable memristors can further remarkably enable simplification of circuits and the realization of complex functions. However, fabricating high-performance reconfigurable memristors remains a great challenge. Two-dimensional (2D) organic–inorganic halide perovskites are regarded as promising candidates for the resistance switching (RS) layer due to the diverse atomic structures. In this work, 2D perovskite film of (PEA)2MA3Pb4I13 is developed for applications in reconfigurable artificial synapses and neurons. By adjusting the compliance current, both volatile and non-volatile RS states have been achieved. Furthermore, the non-volatile memristor enables the simulations of synaptic behaviors including long-term potentiation/depression (LTP/LTD), paired-pulse facilitation (PPF) and spike-timing-dependent plasticity (STDP). A 92% accuracy in handwritten digit recognition is achieved using an artificial neural network. Moreover, the device demonstrates the implementation of adjustable excitation threshold Leaky Integrate-and-Fire (LIF) functionality due to its volatile properties. These capabilities underscore the potential of inorganic–organic hybrid memristor for advanced neuromorphic computing systems.