Perovskite-insulator-perovskite architecture for dynamic recognition of dual-dimensional optical information with the narrow range of mixed incident light

Abstract

Dynamic information recognition and encryption communication have been seriously affected under complex incident light environment conditions, especially the narrow range of mixed incident light. Single material device is not suitable for the intricate ambient light source, resulting in the incapable of accurate identification for the dual-dimensional optical signals. Herein, we design a perovskite-insulator-perovskite (PIP) sandwich structure photodetector through two-dimensional perovskite materials and machine learning to dynamically discriminate the narrow range of complex ambient light including the dual-dimensional optical signals, which enhances the device's light-sensing capabilities for photodetection information recognition and encryption communication. Meanwhile, the machine learning reinforces the complex dynamic recognition of PIP device under the dual-dimensional optical information. The PIP device demonstrates the notable performances along with average responsivity up to 0.13 A W^-1 and detectivity over 3.09 ×10¹¹ Jones. The machine learning-PIP device with the mass and dual-dimensional signal recognition also exhibits the encryption and dynamic decryption of light-based data. This machine learning-assisted PIP architecture device will apply into the reliably optical communication and environmentally adaptive devices landscape.