Picosecond Time‐Scale Resistive Switching Monitored in Real‐Time

Abstract

The resistance state of filamentary memristors can be tuned by relocating only a few atoms at interatomic distances in the active region of a conducting filament. Thereby the technology holds promise not only in its ultimate downscaling potential and energy efficiency but also in unprecedented speed. Yet, the breakthrough in high‐frequency applications still requires the clarification of the dominant mechanisms and inherent limitations of ultra‐fast resistive switching. Here bipolar, multilevel resistive switchings are investigated in tantalum pentoxide based memristors with picosecond time resolution. Cyclic resistive switching operation due to 20 ps long voltage pulses of alternating polarity are experimentally demonstrated. The analysis of the real‐time response of the memristor reveals that the set switching can take place at the picosecond time‐scale where it is only compromised by the bandwidth limitations of the experimental setup. In contrast, the completion of the reset transitions significantly exceeds the duration of the ultra‐short voltage bias, demonstrating the dominant role of thermal diffusion and underlining the importance of dedicated thermal engineering for future high‐frequency memristor circuit applications.

A Statistical Study of Resistive Switching Parameters in Au/Ta/ZrO2(Y)/Ta2O5/TiN/Ti Memristive Devices

“…[ 28 ] Namely, higher voltages are required for switching at higher sweep rate, especially for the larger switching times observed for reset transitions. [ 29 ] We should also take into account the high‐voltage step for the highest cycling frequency corresponding to 0.4 V, which gives rise to the distorted shape of the I – V curves and a low variability, underestimated due to the limited resolution of measurement circuit.…”

“…The latter factor becomes more significant with an increase in the cycling frequency, and its effect is superimposed on the basic regularities associated with the finite switching time, especially pronounced for the reset process. [ 29 ]…”

Controlling Volatility and Nonvolatility of Memristive Devices by Sn Alloying

“…Memristive devices have already been applied in neuromorphic computing, 2,3,5−9 in-memory computing, 10,11 and reservoir computing. 12,13 They offer high scaling potential, 14,15 fast switching times, 16 high endurance, 17 and good retention. 18 However, they generally suffer from high variability between devices due to the stochastic nature of the switching process.…”

“…Memristive devices have already been applied in neuromorphic computing, ,,− in-memory computing, , and reservoir computing. , They offer high scaling potential, , fast switching times, high endurance, and good retention . However, they generally suffer from high variability between devices due to the stochastic nature of the switching process. , To address this and to enable the ability to tailor device functionalities, proposed solutions aim at either controlling the operation parameters or engineering the material composition of the memristive device.…”

Quantum Transport Properties of Nanosized Ta2O5 Resistive Switches: Variable Transmission Atomic Synapses for Neuromorphic Electronics

“…In this paper, we investigate the quantum transport properties of nanoscale RS filaments formed across a Ta 2 O 5 thin film due to the voltage-induced redistribution of oxygen vacancies or tantalum cations. , Ta 2 O 5 is a prominent representative of transition metal oxides, which are widely considered a competitive material platform for neuromorphic hardware architectures. − In order to track the atomistic changes in the filament structure taking place upon RS, we apply superconducting subgap spectroscopy. ,,,− This method enables the analysis of the distribution of the τ i quantum transmission eigenvalues of the conduction channels, contributing to the conductance of the RS filament. The total filamentary conductance is obtained via the Landauer formula as G = G 0 ∑ i =1 M τ i , where G 0 = 2 e 2 / h is the quantum conductance unit, M is the number of open conductance channels, and τ i is the probability of an electron crossing from one side to the other in the i eigenchannel.…”