Optimization of photosensing properties in organic photonic synaptic transistors via grain morphology engineering

Organic photonic synaptic transistors (OPSTs) represent a promising class of neuromorphic devices for future intelligent optoelectronic applications, owing to their ability to simultaneously sense, process, and store optical information, as well as the intrinsic advantages of their constituent materials. In this study, we systematically investigated the influence of grain morphology and structural characteristics within an organic semiconductor (OSC) thin film—modulated via deposition rate—on the electrical and photosensitive performance of OPSTs. Pentacene, a widely adopted photoactive OSC material in OPSTs, was thermally deposited at rates ranging from 0.1 Å/s to 4 Å/s, resulting in diverse molecular interactions that produced distinct grain structures, including dendritic, pyramidal, and lamellar morphologies. Among these, OPSTs fabricated at 1 Å/s exhibited optimized performance, achieving the highest photosensitivity and photoresponsivity due to a favorable balance between efficient charge transport and photogenerated charge trapping. Various forms of synaptic plasticity were successfully demonstrated, including the formation of 25 multilevel conductance states via optical potentiation and electrical depression. When integrated into an artificial neural network for MNIST digit recognition, the optimized OPSTs achieved a recognition accuracy of 88.8%, closely approaching the ideal software-based accuracy. These results underscore the critical importance of grain engineering in optimizing OPST performance and provide a practical strategy for the development of high-efficiency, light-responsive neuromorphic devices for applications in artificial vision systems and edge computing.