Multi-valued memristor, multi-valued memristor array and preparation process of multi-valued memristor array

Abstract

The invention provides a multi-value memristor, a multi-value memristor array and a preparation process of the multi-value memristor array. The multi-value memristor comprises a substrate, a first electrode and a second electrode which are oppositely arranged on the substrate, and a functional layer arranged on the substrate, wherein the functional layer is located between the first electrode andthe second electrode, the first electrode and the second electrode are in contact connection with the functional layer, each of the first electrode and the second electrode comprises a front end and arear end, the front ends are opposite and close to each other, the rear ends are far away from each other, and the widths of the first electrode and / or the second electrode are gradually increased from the front ends to the rear ends. The memristor electrode pair with the narrow front part and the wide rear part is designed, so that the randomness of the formation of the memristor conductive wireis reduced, and the consistency of the memristor is improved; the number of the conductive filaments is related to the external voltage, so that the resistance state number of the memristor is increased; and the memristor based on the conductive wire has good non-volatility.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a multi-valued memristor, a multi-valued memristor array and a preparation process thereof. Background technique [0002] A memristor is a resistor with memory capability, and its resistance value can be adjusted by applying an external voltage pulse. Applying a positive pulse can increase the resistance value of the memristor, and applying a negative pulse can decrease its resistance value. [0003] At present, the rapid development of artificial intelligence technology puts forward higher requirements for energy-efficient data processing. Brain-like chips that simulate the human brain with outstanding energy efficiency have attracted widespread attention. Memristor, as a resistor with memory function, has plasticity and can perfectly simulate biological synapses. It is considered to be the best choice for realizing brain-like chips. The two metal el...

AI Technical Summary

Problems solved by technology

[0004] However, after more than ten years of development, the non-volatile multi-valued memristor with high consistency has not yet broken through

Most of the multi-valued memristors obtained by introducing a large number of defects in the preparation process of functional materials are based on interface effects and are volatile; non-volatile memristors based on conductive filaments generally do not have multi-valued properties, and the filaments are random. poor consistency

[0005] Therefore, in order to realize the non-volatile multi-valued memristor, we need to solve two problems: one is the randomness of the conductive filaments, and the other is the regulation of the number of conductive filaments

Images