Optoelectronic In‐Ga‐Zn‐O Memtransistors for Artificial Vision System

An artificial vision system that can simulate the visual functions of human eyes is required for biological robots. Here, In‐Ga‐Zn‐O memtransistors using a naturally oxidized Al2O3 and an ion gel as a common gate stacking dielectric is proposed. Positive charge trapping in the Al2O3 layer can be induced by modulating the gate voltage, which causes the back sweep subthreshold swing (SS) of the device to break the physical limit (≥60 mV per decade at room temperature), and the minimum SS is as low as 26.4 mV per decade. In addition, photogenerated charges in the device are captured at the In‐Ga‐Zn‐O channel/ion gel interface due to the superposition of the additional electric field generated by positive charges trapped in the Al2O3 layer and the external gate electric field. Thus, persistent photoconductivity is observed in the In‐Ga‐Zn‐O memtransistors. Finally, by employing the optoelectronic memristive functions of In‐Ga‐Zn‐O memtransistors, an artificial vision system based on artificial retinal array (ARA) and artificial neural network is proposed. An obvious improvement in the recognition rate and efficiency with the use of ARA for the image preprocessing is achieved. This study provides a new strategy for the realization of artificial vision systems. Optoelectronic In‐Ga‐Zn‐O memtransistors with a sub−60 mV per decade switching is fabricated. By employing the optoelectronic memristive functions of In‐Ga‐Zn‐O memtransistors, an artificial vision system based on artificial retinal array and artificial neural network is proposed. An obvious improvement in the recognition rate and efficiency with the use of artificial retinal array for the image preprocessing is achieved.