Artificial nerve synapse device and preparation method thereof

Abstract

The invention belongs to the related technical field of intelligent devices, and discloses an artificial neural synapse device and a preparation method thereof. The device comprises a substrate; a bottom electrode layer arranged on the substrate; a dielectric layer arranged on the bottom electrode layer, wherein the dielectric layer comprises a first resistive oxidation layer and a second resistive oxidation layer, the first resistive oxidation layer is an undoped oxide, and the second resistive oxidation layer is an oxide doped with oxygen vacancies; and a top electrode layer arranged on thedielectric layer. The charges of the device can pass through a dielectric layer material under the stimulation of external excitation, the oxygen vacancies in the dielectric layer migrate and recover,opening and closing are achieved, the structure is simple, in addition, the device is prepared through the 4D printing technology, and a microstructure unit can be prepared, so that the artificial nerve synapse device has more excellent resistance change sensitivity and interlayer bonding performance.

Description

technical field [0001] The invention belongs to the technical field related to smart devices, and more specifically relates to a neurosynaptic device and a preparation method thereof. Background technique [0002] With the development of society and people's increasing demand for intelligent robots, modern science and technology pose new challenges to robotics. In the future, robots will abandon the stereotyped image of "reinforced and iron-framed" industrialization, and move toward the true meaning of "flesh and blood" that is light and agile. The development of humanoid "engineering life forms" on the Internet. The sensor system is a key part of it. It is the source of information for the robot, and the nervous system, as a system that gathers all sensor information, performs information perception, transmission and processing. Neuromorphic devices can convert sensor digital signals into analog Neural simulation signals can provide core components for building future huma...

AI Technical Summary

Problems solved by technology

Chinese patent CN108110137A discloses a SiOxNy-based light-reading neural protrusion device and its preparation method. The patent application document uniquely uses light intensity and phase to represent synaptic weights; Chinese patent CN109037442A discloses a SPR neurosynaptic device with resistance effect and its preparation method. This patent application invents a neurosynaptic device with electrical modulation and optical reading, which has the advantages of low energy consumption; although it can simulate synapses to a certain extent, its resistance changes The value cannot be adjusted, the plasticity is poor, and it cannot be used in artificial neural networks; in addition, most of the preparation methods of the above devices use methods such as sputtering and deposition to form each thin film layer, which cannot meet the needs of customized preparation of complex structures.

Images