Simulation of Resistive Switching in Memristor Structures Based on Transition Metal Oxides
- PDF / 909,944 Bytes
- 11 Pages / 612 x 792 pts (letter) Page_size
- 12 Downloads / 232 Views
lation of Resistive Switching in Memristor Structures Based on Transition Metal Oxides O. O. Permyakovaa, b, * and A. E. Rogozhina, ** aValiev bMoscow
Institute of Physics and Technology, Russian Academy of Sciences, Moscow, 117218 Russia Institute of Physics and Technology (State University), Moscow oblast, Dolgoprudny, 141701 Russia *e-mail: [email protected] **e-mail: [email protected] Received January 30, 2020; revised January 31, 2020; accepted February 19, 2020
Abstract—The main approaches to the simulation of resistive switching in systems with transition metal oxide layers are discussed. The algorithms and results of ab initio simulation of such systems, in particular, the energy barrier heights for generation, recombination, and migration of defects in the HfOx and Ta2O5 films, are briefly described. The approaches to the finite element and Monte Carlo simulation and its results are described in detail; special attention is paid to the change in the switching character in a memristor structure upon variation in the kinetic barrier for the diffusion of oxygen ions, which is determined by the interface between the layers involved in resistive switching. Keywords: memristor, resistive switching, Monte Carlo simulation DOI: 10.1134/S106373972004006X
INTRODUCTION Recently, there has been growing interest in nonvolatile resistive random-access memory (ReRAM). The high switching rate allows this memory to be considered as an alternative to dynamic random-access memory (DRAM). Of even greater interest is the possibility of creating neuromorphic systems based on memristors [1–4]. The main advantage of ReRAM is its ability to scaling: the minimum cell size of such a memory is 4f 2 (f is the cell’s element size) [5, 6]. The resistive memory should meet the following technological requirements: an operating voltage below 1 V, energy consumption of ~10 pJ/switching, switching time of shorter than 10 ns, greater than 109 switching cycles, a state preservation time of more than 10 years, and a cell area of
Data Loading...
