Neuromorphic phototransistors based on ZnO/MAPbI3−xClx perovskite heterostructures with tunable synaptic plasticity
摘要
We report neuromorphic phototransistors based on MAPbI3−xClx perovskite channels interfaced with ZnO nanocrystal underlayers. Comparing hybrid and control devices reveals that the oxide/perovskite interface is critical for synaptic operation. The ZnO layer enhances electron extraction and induces a negative threshold voltage shift. Under illumination, the device exhibits persistent photoconductivity with slow dynamics driven by charge trapping at ZnO oxygen vacancies. Notably, increasing ZnO resistance extends the dielectric relaxation time, significantly amplifying hysteresis and residual current. This behavior effectively mimics the biological transition from short-term to long-term memory: rapid perovskite photo-generation provides the initial synaptic influx, while slow ZnO detrapping encodes temporal information as a ‘memory trace’. These dual-timescale dynamics and gate-tunable conductance position the ZnO/perovskite heterostructure as a robust platform for in-sensor learning and neuromorphic vision. Our findings highlight the multifunctional role of oxide interlayers as active modulators of interfacial trap dynamics and synaptic memory behavior.