Photoactive Metal-Organic Framework Heterojunctions for Optoelectronic Synapse and Neuromorphic Computing.

Metal-organic frameworks combine the structural tunability of organic materials with the periodic order of inorganic lattices, making them attractive candidates for neuromorphic devices. However, most MOF-based synaptic devices reported to date have been limited to electrically driven memristors because conventional MOFs generally suffer from poor conductivity and inadequate film quality. Here, we report the in situ preparation of photoactive ZnO/Zn3(HHTP)2 heterojunction that enables optoelectronic synaptic plasticity and neuromorphic computing. Two-dimensional Zn3(HHTP)2 framework is synthesized via the coordination reaction between 2,3,6,7,10,11-hexahydroxytriphenylene and Zn2+ derived from sacrificial ZnO nanorods, yielding a porous conductive MOF integrated directly with ZnO. The resulting ZnO/ZnHHTP heterojunction forms type-II band alignment, facilitating efficient separation and transport of photo-generated carriers. Combined with defect-assisted trapping and relaxation processes, the device emulates key synaptic functions, including paired-pulse facilitation, spike dependent plasticity, and transition from short-term plasticity to long-term plasticity. Benefiting from dual-wavelength optical responses under 445 and 520 nm illumination, the device also realizes all-optical "AND" and "OR" logic operations. Furthermore, when integrated into a reservoir computing framework, the nonlinear transient responses of the heterojunction enable handwritten digit recognition with an accuracy of ∼95% on the MNIST dataset. This work establishes conductive MOF-based heterojunctions as promising platforms for neuromorphic information processing.