基于深紫外光激发的低功耗Ga2O3纳米突触器件用于神经形态计算

Synaptic nano-devices have powerful capabilities in logic,memory and learning,making them essential compo-nents for constructing brain-like neuromorphic computing systems.Here,we have successfully developed and demonstrated a synaptic nano-device based on Ga2O3 nanowires with low energy consumption.Under 255 nm light stimulation,the biomimetic synaptic nano-device can stimulate various functionalities of biological synapses,including pulse facilitation,peak time-dependent plasticity and memory learning ability.It is found that the artificial synaptic device based on Ga2O3 nanowires can achieve an excellent"learning-forgetting-relearning"functionality.The transition from short-term memory to long-term memory and retention of the memory level after the stepwise learning can attribute to the great relearning functionality of Ga2O3 nanowires.Furthermore,the energy consumption of the synaptic nano-device can be lower than 2.39×10-11 J for a synaptic event.Moreover,our device demonstrates exceptional stability in long-term stimulation and storage.In the applica-tion of neural morphological computation,the accuracy of digit recognition exceeds 90％after 12 training sessions,indicating the strong learning capability of the cognitive system composed of this synaptic nano-device.Therefore,our work paves an effective way for advancing hardware-based neural morphological computation and artificial intelligence systems requiring low power consumption.