Live Demonstration: Electroforming of TiO<sub>2–x</sub> Memristor Devices using High Speed Pulses

Foster, Patrick; Huang, Jinqi; Serb, Alexantrou; Prodromakis, Themis; Papavassiliou, Christos

doi:10.1109/iscas45731.2020.9180955

Cited by 1 publication

(1 citation statement)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This process is commonly referred to as ''electroforming'' or ''forming'' and can affect fundamental parameters such as the SET-RESET cycling repeatability and multi-level switching response of the devices [41]. Therefore, different methods to carry out this task have been explored [14], [42], [43]. In voltage-based experimental setups, forming is typically carried out using a voltage ramp that might reach much higher voltages than the nominal operating voltages for SET/RESET.…”

Section: Device Preparation: Current-based Forming and Set-reset Cyclingmentioning

confidence: 99%

Effective Current-Driven Memory Operations for Low-Power ReRAM Applications

et al. 2023

View full text Add to dashboard Cite

Resistive switching (RS) devices are electronic components which exhibit a resistive state that can be adjusted to different nonvolatile levels via electrical stressing, fueling the development of future resistive memories (ReRAM) and enabling innovative solutions for several applications. Most works so far have used voltage-based driving schemes for both WRITE and READ operations. However, results from current-driven WRITE operations have shown high uniformity in switching performance, and thus constitute a valid alternative to consider, but current-driven READ operations have rarely been explored. In this context, here we tested a current-based READ/WRITE memory driving scheme on commercial self-directed channel (SDC) devices, while operating constantly at low current levels between tenths of nA and 1.5 uA. We propose a novel method to carry out efficient READ operations exploiting the transient response of the voltage on the current-driven ReRAM memory cells. For READ operations performed at 100 nA, we calculated the cumulative probability distribution of the standard deviation of the measured voltage (σ V ) on the devices and we observed a ratio σ V−HRS /σ V−LRS ∼ 10×. Moreover, the HRS and LRS states were distinguishable in all the tested devices with less than 0.5% error. Finally, the calculated energy consumption (E SET ≈ 10 nJ, E RESET ≈ 30 nJ, and E READ between 80-400 pJ) was competitive even when the duration of the READ/WRITE current pulses was suboptimal in the millisecond range. Therefore, the presented results validate the promising characteristics and the power-efficiency of the proposed READ method for current-driven ReRAM circuits and applications.

show abstract