Research On Graphene/h-BN/graphene Selector And Its Optimization Scheme

Semiconductor memory plays a vital role in the rapid development of the information industry.As a traditional mainstream memory,flash has reached its physical limit in technology.However,among many new types of memory,resistive random access memory is considered to be the most promising new generation mainstream memory because of its advantages of fast speed,high integration,large capacity and low power consumption.The passive cross-array structure is the best structural solution to maximize the integration density of the resistive random access memory,but there is an inevitable crosstalk problem caused by the sneak current in the passive array.Among the many solutions,the 1S1 R structure composed of resistive random access memory cells in series with a non-linear selector is considered to be the simplest and most effective method.It is very important to design a selector with excellent performance to improve the performance of the resistive random access memory.The simplest selector structure is MIM(Metal-Insulator-Metal),and traditional selector insulating layers usually use transition metal oxides.This paper proposes a single-layer horizontal selector(Graphene/h-BN/Graphene)based on the two-dimensional material h-BN with graphene as the electrode.In this paper,the first-principles calculation method based on density functional theory is used to study and explore the working mechanism and performance of the Graphene/h-BN/Graphene selector.First,by comparing the DOS diagrams and calculating the surface energy,the number of lateral layers of the graphene electrode and h-BN is determined.Then the interface distance and interface type of Graphene/h-BN interface are determined by calculation of binding energy.Finally,the interface mismatch rate is calculated,and the selector model is finally determined.After the selector model is determined,the I-V curve of the selector is simulated using the DFTB+ module,showing symmetrical threshold switching characteristics.Then by calculating the electron density,local charge density,TDOS,PDOS,and differential charge density,it is determined that the conduction mechanism of the selector is the boron vacancy conductive filament.After determining the conduction mechanism of the selector,on this basis,different concentrations of boron vacancies were produced in different regions of the insulating layer,and the influence of different boron vacancies in different regions on the performance of the selector and the reasons for the performance of the selector were studied.The study found that in the formation of h-BN conductive filaments and internal regions,boron vacancies will reduce the threshold voltage.However,at a specific concentration in the left interface area of h-BN,the off-state current of the selector will be reduced,the non-linearity will increase by an order of magnitude,and the threshold voltage will increase.For the boron vacancies located in the conductive filament formation area and the internal area,the occurrence of defective states is found through the comparison of the band structure and the DOS diagram,which makes the boron vacancy conductive filaments in h-BN easier to form,which makes the selector easier to conduct.Pass,so the threshold voltage is reduced.However,for boron vacancies existing in the interface area on the left side of h-BN,it is found from the electrostatic potential that an interface barrier will be formed at a specific concentration.The existence of the interface barrier is equivalent to an electronic barrier,making electrons from the side of the graphene electrode to h-BN The transfer becomes more difficult,so the off-state current decreases and the non-linearity increases by an order of magnitude.The threshold voltage increases.In summary,this paper proposes a new type of selector based on the two-dimensional material h-BN,and studies its conduction mechanism and performance,and on this basis,explores the performance parameters of the boron vacancy for the selector.The influence and its reasons provide theoretical guidance for the research and performance improvement of the h-BN-based selector.