Low Power Consumption CsPb<sub>0.5</sub>Sn<sub>0.5</sub>Br<sub>3</sub> Quantum Dot-Based Photoelectric Synaptic Transistors for Neuromorphic Computing
Jiangdong Zhang, Jia Liu, Haichuan Geng, Wenwen Wang, Yiying Wei, Menghan Chen, Jiahao Kang, Jinjin Zhao
Abstract
With the rapid development of artificial intelligence and neural networks, synaptic transistors have great potential and advantages in the field of neuromorphic computing. However, power consumption is the biggest challenge for traditional electrically driven synaptic transistors. To overcome this limitation, perovskite quantum dots have emerged as an ideal material for fabricating low-power photoelectric synaptic transistors owing to their high photoresponsivity and tunable band gap characteristics. Here, we propose a photoelectric synaptic transistor based on CsPb 0.5 Sn 0.5 Br 3 quantum dots (CPSB QDs)/poly(3-hexylthiophene) (P3HT) composite functional layer. The incorporation of Sn ions successfully reduced lead toxicity in the perovskite structure while enhancing the current response of the device. Benefiting from the excellent photoelectric properties and charge transport capabilities of the composite film, our device achieves a series of fundamental synaptic behaviors and synaptic plasticity, and exhibits an ultralow power consumption of 0.1 fJ per synaptic event under an extremely low read voltage ( V D = 1 μV). Furthermore, we apply this photoelectric synaptic transistor to biomedical image recognition tasks, achieving 90.12% accuracy on the BloodMNIST dataset using a four-layer convolutional neural network. This work provides a viable pathway for developing environmentally friendly and energy-efficient neuromorphic computing systems.