posted on 2024-01-12, 12:33authored byShuyao Han, Lue Zhou, Junli Huang, Yuncheng Mu, Yuhao Xie, Shu Zhou
Halide
perovskites are promising for energy-efficient and reconfigurable
artificial synapses due to the unique charge transport characteristic,
i.e., ionic-electronic coupling. This work attempts to detangle the
size and surface chemistry effects on perovskite-based synaptic memristors
that remain elusive to date. Herein, three different perovskite-based
memristors were constructed for mimicking the synaptic behavior: device
1 based on CsPbBr3-nanocrystal (NC) film terminated with
short alky ligands, device 2 based on naked CsPbBr3-NC
film, and device 3 based on CsPbBr3 bulk film. In addition
to emulating a series of essential synaptic functions, our results
suggest that perovskite-NC films outperform their bulk counterpart
in light of paired pulse facilitation (PPF) index and gain due to
an increased trap density and enhanced ion transport. The incorporation
of short alkyl ligands into the NC surface decreases the film conductance
without suppression of charge transport, thereby reducing the on-state
energy consumption. Using the best-performed NC device comprising
a neuron network, we did a preliminary proof-of-concept demonstration
of the potential for brain-inspired neuromorphic computing which was
proved effective for digit recognition with a high accuracy of 94%.
Our study signifies the importance of grain size and surface chemistry
engineering for the development of high-performance robust synaptic
memristors based on halide perovskites.