基于AlNx/Nb2O5/NbOy结构的自选择阻变器件及其物理机制研究

Three-dimensional vertical resistive random access memory (3D VReRAM) is a very promising memory scheme due to the advantages of high density, low cost, etc compared with the other architectures. Data misreading induced by sneak current in memory arrays seriously limits the capacity of 3D VReRAM. Conventional one-selector-and-one-resistor (1S1R) scheme could not address the sneak current issue in 3D VReRAM. Self-selecting resistive switching (RS) devices without a middle conducting layer are an extremely attractive answer for the issue. Nevertheless, how to achieve high-performance such a device suitable for large-capacity 3D VReRAM is still a big question. In this project, we propose the idea of using AlNx as an ion-migration barrier layer in RS devices and plan to study self-selecting RS devices based on the AlNx/Nb2O5/NbOy structure and their mechanism. The focus of the project is to study the impact of the direct electron tunneling, intrinsic defect-mediated electron conduction, and dopant-mediated electron conduction together with the corresponding chemical-bond components of the AlNx barrier layer on the properties of the RS devices, especially the characteristics of the switching processes, the electrical transport mechanism of different resistance states, and self-selecting properties. Based on the above investigation, the physical mechanism on the relationship between the properties of the AlNx barrier layer and the characteristics of the RS devices will be established. The AlNx modulation methods will also be obtained to increase the selectivity and on/off ratio and reduce the leakage current of the self-selecting RS devices. The implementation of the project would promote the studies on self-selecting RS devices and establish the foundations for the achievement of large-capacity 3D VReRAM.

三维垂直型阻变存储器（3D VReRAM）具有存储密度高和成本低等优势，是极具前景的信息存储技术。旁路电流引起的数据误读严重限制了3D VReRAM的存储容量。常规的1S1R方案无法解决3D VReRAM的旁路电流问题。本项目针对可克服大容量3D VReRAM旁路电流问题的自选择阻变器件，创新提出将无定型AlNx作为离子迁移阻挡层的方法，研究基于AlNx/Nb2O5/NbOy结构的自选择阻变器件及其物理机制。重点探究AlNx阻挡层的直接隧穿、本征缺陷辅助导电和杂质辅助导电三种导电类型及其相应的化学键组成对器件的阻态转变特性、阻态导电机制和自选择性的影响规律，建立AlNx阻挡层与器件电学特性关系的物理模型，获得提高器件选择比、开关比和减小器件漏电的AlNx阻挡层调控方法，为3D VReRAM的实现奠定基础。

基于过渡金属氧化物的阻变（RS）器件结构简单，尺寸小，开关速度快，成本低廉且特性可调，引起了人们的广泛关注，已成为神经形态芯片中的电子突触和下一代非易失存储器的有力候选者之一。在众多过渡金属氧化物阻变器件中,铌氧化物器件表现出阻变或阈值开关行为，甚至即有阻变也有阈值开关行为。阻变和阈值开关在单个铌氧化物器件中的共存在无需选择器的情况下就可解决无源交叉阵列的潜行路径问题。此外，基于二氧化铌相变的阈值开关可用于产生人工神经元的脉冲信号。这些现象有非常大的应用价值，但是氧化铌阻变器件的物理机制仍然不够清楚，其特性调控的方法还不明确。.本项目围绕Nb2O5/NbOx阻变器件的物理机制和特性调控开展了一系列系统的研究工作。首先研究了Nb2O5/NbOx阻变器件中电极诱导的阻变极性反转。电形成之后，具有惰性Pt顶电极的器件显示顺时针RS行为。然而，具有活性W顶电极的器件表现出逆时针RS行为。研究表明，活性电极通过界面氧化还原反应在导电丝的活性区域附近提供了一个氧离子储蓄区。该储蓄区在导电丝的断裂和恢复过程中起主导作用，导致了阻变极性的反转。可见氧离子在导电丝尖端附近的分布和迁移对氧化物阻变器件的阻变过程具有至关重要的影响。然后研究了氧空位储蓄层对Nb2O5/NbOx阻变器件电学特性的影响。研究表明，随着氧空位储蓄层中氧含量的增大，初始状态的漏电流减小，电形成电压增大，设置电压增大，重置电压增大，惰性电极的开关比增大。这些规律为我们之后调控器件的阻变特性，实现性能优异的存储器件和神经形态器件奠定了坚实的基础。.深入研究了氮化硅阻变器件的物理机制，建立了氮化物阻变器件的基于氮离子迁移与转移的物理模型，成功地解释了电极储氮能力、电极厚度调控器件特性的机理。通过金属掺杂降低了氮化硅阻变器件的功耗并提高了器件的稳定性。这为今后进一步提升氮化物阻变器件的性能奠定了良好的基础。通过PbS量子点调控了金属电极与AlN阻变层的接触面积和位置。随着量子点密度的增加，接触面积不断缩小。接触点定位和金属间断电场增强效应使得器件的稳定性显著改善。该研究为氮化物阻变器件稳定性的提升提供了一条有效的途径。