The human retina can provide inspiration for the development of optoelectronic devices for artificial vision systems. It has, in particular, a curved geometry for low optical aberration imaging and is capable of signal preprocessing, in part due to its high synaptic facilitation. However, it remains challenging to create artificial synaptic devices with mechanical softness, curvilinear form factors and superior synaptic facilitation. Here we report a cascaded two-stage optoelectronic synapse that uses silicon photovoltaic cells to modulate sodium-alginate-gated synaptic transistors. The photovoltaic cells and an electric double-layer capacitor convert the initial light signal into a gate voltage (stage one), which then controls the postsynaptic current in the channel of the synaptic transistor (stage two). This cascaded synaptic signal transmission mechanism results in high synaptic facilitation of the postsynaptic signal. The system exhibits stable and long-term linearly potentiated characteristics, enhancing the accuracy of pattern recognition. We use an array of the cascaded optoelectronic synapses to create a curvy, kirigami-structured neuromorphic imager, which mimics the curved geometry and neural signal transmission of the human retina and provides visual information sensing and preprocessing functions.

- Author (Pusan National University): Hyunseok Shim (School of Electrical and Electronics Engineering)

- Title of original paper: A neuromorphic imager based on a cascaded optoelectronic synapse

- Journal: Nature Electronics 

- Web link: https://www.nature.com/articles/s41928-025-01540-w

- Contact e-mail: hshim@pusan.ac.kr