Photoelectric Memristor With Self-Powered for Dynamic Information Recognition

The bionic vision system realizes the functions of optical perception of external information and image information processing. The photoelectric synaptic device is the basic unit of this system. In this work, a HfO<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><roman>2</roman></sub>/Bi<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><roman>4</roman></sub>Ti<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><roman>3</roman></sub>O<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><roman>12</roman></sub> (BIT) heterojunction photoelectric memristor with synaptic behavior is introduced. The device generates a persistent photocurrent (PPC) through the capture and release of photogenerated carriers in the heterojunction. Various synaptic behaviors can be simulated under the stimulation of light pulses, including paired-pulse facilitation (PPF), spike-number-dependent plasticity, and spike-width-dependent plasticity. It can also realize the transformation from short-term memory (STM) to long-term memory (LTM). A 3 × 5 array is constructed and the time dynamic characteristics of the device are used to distinguish the dynamic input. The physical reservoir built by photoelectric synaptic devices can not only improve the speed of data processing but also achieve efficient data storage, which brings new possibilities for AI in the fields of image recognition, speech recognition, and natural language processing.