Gate-Reconfigurable ZnO/ β -Ga2O3 Heterojunction Phototransistor Enabling Dual-Wavelength Ultraviolet Photoresponse and Optoelectronic-Synaptic Behaviors

Abstract A central challenge for next-generation intelligent optoelectronic systems is to co-integrate photodetection and neuromorphic in-sensor computing within a single device to eliminate redundant data transfer between physically separated units. Herein, we demonstrate a gate-reconfigurable ZnO/β-Ga2O3 heterojunction phototransistor that reconciles these two antagonistic functions via a bias-switchable carrier-transport pathway. The device design couples (i) a large interfacial conduction-band energetic step and (ii) oxygen-vacancy-related trap states distributed in the Ga2O3 layer, enabling the external gate voltage to toggle the dominant mechanism between rapid electron transfer/collection and Ov-mediated persistent photoconductivity (PPC). In the photodetection mode (VG = +10 V) under 366 nm illumination, the phototransistor exhibits a responsivity of 7.5 × 104 A/W, a shot-noise-limited specific detectivity of 7.8 × 1014 Jones, and an external quantum efficiency of 3.8 × 107%, with a weak optical intensity of 10 μW/cm2. Complementary measurements at 290 nm further confirm the dual-wavelength ultraviolet response of the device. This mode enables optical ASCII communication and intensity-reconfigurable Boolean logic (AND/OR) within a single transistor. In the neuromorphic synapse mode (VG = –10 V), enhanced electron confinement in the oxygen-vacancy-rich Ga2O3 layer activates pronounced PPC, enabling robust emulation of synaptic plasticity including paired-pulse facilitation (PPF), short-term-to-long-term (STP-LTP) plasticity transition, learning-forgetting-relearning, and Pavlovian conditioning under both 290 and 366 nm excitation. Moreover, leveraging PPC as a bio-inspired resource, we extend functionality to artificial nociceptors (hyperalgesia and allodynia) and dual-wavelength spatiotemporal encoding, highlighting a versatile materials/device strategy for integrated neuromorphic vision and programmable optoelectronic computing hardware.