Dual-neuromorphic device based on charge trapping effect and electric double-layer effect and preparation method of dual-neuromorphic device

Abstract

The neuromorphic circuit comprises numerous neuromorphic devices, and at the moment, the more accurate the device weight update (conductivity adjustment and control) is, the better the conciseness is, that is, the difficulty of "learning" is reduced. At present, adjustment and control of weight updating face the scientific problems that conductance is in non-linear / asymmetric change, bipolar electric pulse (positive and negative voltage) adjustment and control must be adopted, and the difficulty of weight updating of a neuromorphic device is increased; according to the invention, a dual-neuromorphic device with a charge capture effect and an electric double-layer effect is prepared, the adjustment and control of unipolar electric pulses on weight updating are realized by utilizing the difference of the adjustment and control rates of the two charge effects on a semiconductor layer, and the two charge effects form a mutual inhibition effect to enhance the linearity and symmetry of conductivity adjustment and control; the problem that weight updating in a neuromorphic circuit is difficult is solved.

Description

Technical field [0001] The present invention relates to the field of electronics and material learning, and more particularly to a double neuromete device based on charge capture effect and a two-layer effect, a preparation method and application thereof. Background technique [0002] The conventional electronic chip is more than a silicon wafer as a substrate, which is difficult to apply in a wearable, implanted electronic field. Flexible electronic devices have the advantages of bent, superior, light, easy to carry, and easy integration. [0003] The traditional "von. Nogan" calculation structure is limited by the storage unit and the calculation processing unit, which has a problem of high power consumption, slow deposition speed, and is inspired by the human brain, and the human brain neurological device is due to its It has the advantages of calculation, efficient, energy-saving, and quickly simulating human brain function, completing the calculation and storage task, andcom...

AI Technical Summary

Problems solved by technology

In hardware neuromorphic circuits, the higher the level of integration, the greater the number of individual devices. At this time, it is required to update the weight of a single device as concisely as possible, that is, the easier the "learning" process of the device, the better, otherwise it will become very cumbersome. , which is not conducive to the regulation of large-scale circuits

Updating the weight of neuromorphic devices is generally to apply electrical pulses to the device to change the conductance of the device. However, at present, when using electrical signals to regulate the weight update, the conductance is nonlinear and asymmetrical, and bipolar (positive and negative) must be used. Voltage) pulse and other scientific issues are not ideal; this has both internal and external reasons. The external reasons include the accuracy of the test instrument, the magnitude of the applied pulse amplitude, the pulse frequency and the pulse width, etc., and the common organic thin film field effect transistor Taking memory as an example, the conductance change of the device is related to the degree to which the device traps charge

Images