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 HfO_x 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.

Memristive structures are among the most promising options to be components of neuromorphic devices. However, the formation of HfO₂-based devices in crossbar arrays requires considerable time since electroforming is a single stochastic operation. In this study, we investigate how Ar⁺ plasma immersion ion implantation (PI) affects the Pt/HfO₂ (4 nm)/${\mathrm{HfO}}_X{\mathrm{N}}_Y$ (3 nm)/TaN electroforming voltage. The advantage of PI is the simultaneous and uniform processing of the entire wafer. It is thought that Ar⁺ implantation causes defects to the oxide matrix, with the majority of the oxygen anions being shifted in the direction of the TaN electrode. We demonstrate that it is feasible to reduce the electroforming voltages from 7.1 V to values less than 3 V by carefully selecting the implantation energy. A considerable decrease in the electroforming voltage was achievable at an implantation energy that provided the dispersion of recoils over the whole thickness of the oxide without significantly affecting the ${\mathrm{HfO}}_X{\mathrm{N}}_Y$/TaN interface. At the same time, Ar⁺ PI at higher and lower energies did not produce the same significant decrease in the electroforming voltage. It is also possible to obtain self-compliance of current in the structure during electroforming after PI with energy less than 2 keV.

The results of the research on the resistive switching in Pt/HfO2/HfOxNy/TiN, Pt/HfO2/TaOxNy/TiN and Pt/Al2O3/HfO2/TaOxNy/TiN structures are presented. Active layers of the structures were deposited by atomic layer deposition. The formation of conductive filaments in all three structures was demonstrated using conductive atomic force microscopy. A full cycle of resistive switching with a microscope probe in these structures was also demonstrated. The properties of filaments formed at different voltages were studied, and the distributions of the filament density by conductivity and size were presented. The characteristics of the studied structures were compared. It appears that the Pt/Al2O3/HfO2/TaOxNy/TiN structure has the greatest potential for resistive random-access memory application

The mechanism of resistive switching in Pt/HfO₂(8 nm)/HfO_XN_Y(4 nm)/TiN structures, in which there are two resistive switching modes: bipolar resistive switching and complementary resistive switching. We demonstrate that resistive switching without external current compliance is possible. It is shown experimentally that the conductivity in the low-resistance state corresponds to the space-charge-limited current. A qualitative model is proposed that describes the transition from bipolar resistive switching to complementary resistive switching using Schottky barrier modulation at the metal-insulator interface. Based on this model, an explanation is given for the degradation of the low-resistance state during endurance measurements.

The linearity of the memristor conductivity (or weight) variation affects the training accuracy of analog neuromorphic systems. Usually, the weight update linearity can be achieved by changing the waveform, but this requires the inclusion of additional peripherals in the analog circuit. The paper considers the influence of the signal pulse amplitude on the long-term potentiation of a HfO2-based memristor. It is observed that the pulse amplitude decrease leads to the higher device's linearity. The linearity optimization procedure is shown. Additionally, we consider the effects of unusual conductance changing during consecutive cycles of long-term potentiation and long-term depression with a constant voltage amplitude. We demonstrate the possibility of a linear weight update of the device once identical pulses are applied due to unusual behavior.

Rapid progress of artificial neural network applications in recent years has led to the issue of unprecedented energy consumption. It can be solved by the implementation of energy-efficient hardware based on non-von Neumann architectures, which requires the development of electronic components emulating the behavior of synapses and neurons. While research of synaptic elements is vast, the technology for the fabrication of scalable and highly reproducible neuronal elements is far less developed. In this work, we demonstrate an artificial neuron with multiple functionalities based on filamentary switching Ag/Hf0.5Zr0.5O2 (HZO) memristors. To improve the parameters of the memristors, we propose a two-step annealing method, which allows for better control of the crystallization of the functional dielectric layer (HZO) as well as of the diffusion of the active electrode (Ag) atoms. Furthermore, we demonstrate leaky integrate-and-fire neuronal behavior in multiple spiking modes: time-to-first-spike, number of spikes, and firing rate coding. Moreover, the neuron operation does not require the additional electronic overhead and is supported solely by an Ag/HZO memristor with a current limiting resistor connected in series. The presented results pave the way for the development of next-generation energy-efficient neuromorphic hardware operating on the principles of spiking neural networks.