A Dual-Mode Image Sensor Using an All-Inorganic Perovskite Nanowire Array for Standard and Neuromorphic Imaging

Zhenghao Long,Yucheng Ding,Xiao Qiu,Yu Zhou,Shivam Kumar,Zhiyong Fan
DOI: https://doi.org/10.1088/1674-4926/44/9/092604
2023-01-01
Journal of Semiconductors
Abstract:The high-density, vertically aligned retinal neuron array provides effective vision, a feature we aim to replicate with electronic devices. However, the conventional complementary metal-oxide-semiconductor (CMOS) image sensor, based on sepa-rate designs for sensing, memory, and processing units, limits its integration density. Moreover, redundant signal communica-tion significantly increases energy consumption. Current neuromorphic devices integrating sensing and signal processing show promise in various computer vision applications, but there is still a need for frame-based imaging with good compatibility. In this study, we developed a dual-mode image sensor based on a high-density all-inorganic perovskite nanowire array. The device can switch between frame-based standard imaging mode and neuromorphic imaging mode by applying different biases. This unique bias-dependent photo response is based on a well-designed energy band diagram. The biomimetic align-ment of nanowires ensures the potential for high-resolution imaging. To further demonstrate the imaging ability, we con-ducted pattern reconstruction in both modes with a 10 × 10 crossbar device. This study introduces a novel image sensor with high compatibility and efficiency, suitable for various applications including computer vision, surveillance, and robotics.
What problem does this paper attempt to address?