| In recent years,with the rapid development of the Internet,in various fields gradually enter the information age,scientific research workers focus on the stability,speed and miniaturization of basic components.Memristor was first proposed by Leon O.Chua in 1971.The amount of charge flowing through the memristor determines the resistance value of the memristor.Many researchers have carried out extensive research on it.In the field of memory,resistance random access memory（RRAM）is an important application research.Compared with resistive memory,it is easy to fabricate and has a simple device structure,especially in nonvolatile neural morphology calculation and its adjustable cut-off frequency for filter research.However,in the development process of resistive memory,because the switching voltage is randomly distributed,in order to solve the problem of large dispersion of switching voltage,many researchers have carried out a lot of work in recent years.At the same time,for circuit applications,memristor is still in the early simulation stage.We can also use memristor to mimic the behavior of human brain to test the biological characteristics of nervous system,especially in the aspects of paired pulse facilitation（PPF）and spiking timing dependent plasticity（STDP）.This paper focuses on the performance improvement of memristor,focusing on the electrical characteristics and application of memristor.Firstly,in order to improve the dispersion of memristor switching parameters,graphene oxide quantum dots with different sizes are studied in this chapter through the comparative analysis of different sizes of GOQDs,combined with the traditional oxide material Hf O2 in the functional layer,different sizes of GOQDs were spin coating on Pt substrate to prepare Ag/Hf O2/GOQDs（different mesh）/Pt devices.According to the test and comparative analysis,it was found that the device with 100 mesh size of GOQDs is the most stable.The device can be Set and Reset in both positive and negative directions,that is,it has reversible resistance switch（RS）behavior.After analysis,the Vset/Vreset of the device is low,which is0.13/-0.14 V.Further statistics and Gaussian fitting of the Set and Reset voltages show that the threshold voltage range of the device is very concentrated,which indicates that the device has high homogeneity.In addition,the switching speed of the device is greatly improved,up to 7 ns/25 ns.Finally,the physical mechanism of the memristor is analyzed.The electrical characteristics of the memristor are improved by embedding 100 mesh GOQDs to enhance the local electric field,which provides an important device foundation for the application of memristor in neural bionics.Secondly,the new memristor with continuously adjustable conductance can mimic the continuous change of synaptic weight,which is the key factor of memristor used in synaptic bionics.Taking advantage of the advantages of Ag/Hf O2/100 mesh GOQDs/Pt device,such as good stability,fast switching speed and excellent retention ability,the structure of the device is adjusted and optimized.After analysis,it is found that the electrical characteristics of the device will be significantly improved if 100 mesh GOQDs are insert in the device.Based on this,a new type of memristor with Ag/Hf0.5Zr0.5O2（HZO）/Ag/HZO/Ag/HZO/GOQDs/Pt structure is prepared.The device exhibits a slow switching behavior.Due to the influence of low energy pulse,the device has unique advantages in adjusting the linear conductivity.Based on bidirectional progressive conductivity control,a novel memristor is used to mimic biological synaptic functions,such as STDP and PPF.The physical mechanism of the device is analyzed.The results show that the tunneling mechanism and electrochemical metallization effect of the device are important factors to control the conductivity of the device continuously and gradually,and create favorable conditions for further research on the application of memristor circuit.In the third experiment,based on the outstanding electrical characteristics of the memristor,the application research of the memristor is explored creatively.And in view of the above discussion,the application of memristor in information processing is studied.It is found that it has great potential in digital recognition and filter circuit application.The accuracy of 90.91%is achieved after 500 training with memristor as digital recognition technology.At the same time,the filter circuit is designed into a real object,and the applications of low pass,high pass and band-pass filter are realized.Taking high pass filter as an example,the adjustable range of cut off frequency is achieved from 0.16-1.429 GHz.This work provides a new foundation for intelligent control and the establishment of memristor in artificial intelligence,and provides the possibility for further application of memristor in information processing. |
PDF Full Text Request