Synaptic Characteristics and Neuromorphic Computing Enabled by Oxygen Vacancy Migration Based on Porous In₂O₃ Electrolyte-Gated Transistors

The migration of oxygen vacancies is decisively affecting the modulation of channel conductance. However, for electrolyte-gated transistors (EGTs), the modulation of channel conductance driven by the migration of oxygen vacancies can be easily concealed by the existence of an electrical double layer (EDL). Here, we first observed the modulation of channel conductance caused by the migration of oxygen vacancies, marked by a clockwise transfer hysteresis, by keeping the top gate electrode at a far distance from the source/drain electrodes. The contribution degree of the migration of oxygen vacancies and the EDL on the modulation of channel conductance can be adjusted by changing the distance between the top gate and source/drain electrodes. The results were supported by X-ray photoelectron spectroscopy measurements and first-principles calculations. Due to the migration of oxygen vacancies, the proposed porous In 2 O 3 EGT could emulate biological synaptic functions, such as long-term potentiation/depression, paired-pulse facilitation/depression, image learning, and memorizing. A convolutional neural network based on EGTs is constructed to identify the traffic sign data sets with a recognition accuracy of 91.3%. The porous nature of the developed EGT improved both the long-term memory and the neuromorphic computing properties. From the provided results, a deep understanding of the underlying mechanism of the modulation of channel conductance composed of EGTs is provided. Moreover, our work paves the way for the fabrication of next-generation high-performance artificial synapses.