Ultralow-power consumption bimodal synaptic transistors for high-efficiency neuromorphic vision system.
Journal:
Journal of colloid and interface science
Published Date:
Sep 12, 2025
Abstract
Artificial vision synaptic devices, characterized by low power consumption and high parallelism efficiency, hold significant promise for artificial visual systems. However, nearly all reported low-power consumption synaptic devices lack the capability for bimodal optoelectronic coordinated regulation. Furthermore, a prevalent but often overlooked issue in most electrically stimulated low-power consumption synaptic transistors is current backflow. This phenomenon leads to an underestimation of actual power consumption and obscures the true synaptic signal. Here, an ultralow-power consumption bimodal organic optoelectronic synaptic transistor is fabricated using a double insulating layer consisting of a proton-conductive material and a slow polarization material, as well as a p-type/n-type polymer blend. This study pioneers the systematic investigation of current backflow within low-power consumption synaptic transistors. Moreover, bimodal synaptic response is demonstrated with ultralow energy consumption, exhibiting a minimum energy expenditure of 8.3 and 2.2 fJ per synaptic event under electrical and optical stimulation, respectively. In addition, core functionalities of the visual system, including optoelectronic synaptic plasticity, image erasure/enhancement, real-time signal preprocessing, and dynamic information storage are successfully emulated. This work provides a viable approach for developing ultralow-power consumption bimodal optoelectronic synaptic transistors, supporting the advancement of energy-efficient neuromorphic devices and artificial intelligence systems.
Authors
Keywords
No keywords available for this article.