動態蒙地卡羅模擬應用於二維材料電阻記憶體之研究

Abstract

本研究透過實驗和模擬研究了二硫化鎢（WS2）、二硫化鉬（MoS2）和六方氮化硼（h-BN）等二維材料電阻式記憶體（2D-material-based RRAM）。深入討論了主動層厚度的影響，從而找到具有最佳關鍵績效指標（KPIs）的二維材料電阻式記憶體厚度。這項工作揭示出 2D-material-based RRAM 和傳統氧化物 RRAM 之間的根本區別。此外，此研究採用動力學蒙特卡羅（KMC）模型來提取 2D-material-based RRAM 元件的物理參數。與傳統氧化物RRAM 相比，二維材料沿垂直方向具有較低的擴散活化能，從而降低RRAM 元件的能耗並縮短切換時間。值得注意的是，在透過化學氣相沉積（CVD）生長的2D-material-based RRAM 樣品中觀察到其擴散活化能低於機械剝離樣品的擴散活化能，表明基於 CVD 的2D-material-based RRAM 具有更卓越的元件性能。我們的結果表明，MoS2 在三種二維材料中具有最快的切換速度。

In this study, we conducted investigations on two-dimensional resistive random access memory (2D-material-based RRAM) materials, specifically tungsten disulfide (WS2), molybdenum disulfide (MoS2), and hexagonal boron nitride (h-BN), employing a combination of experimental and simulation methodologies. The influence of the active layer thickness is thoroughly discussed, leading to the determination of the optimal thickness for 2D-material-based RRAM materials exhibiting superior key performance indicators (KPIs). Our research uncovers a fundamental distinction between 2D-material-based RRAM and conventional oxide RRAM. Moreover, the kinetic Monte Carlo (KMC) model is employed in this study to extract the physical parameters of the 2D-material-based RRAM devices. In comparison with traditional oxide RRAM, the 2D materials demonstrate a lower diffusion activation energy along the vertical direction, thereby enabling reduced energy consumption and shorter switching times for RRAM devices. Significantly, the diffusion activation energy observed in the 2D-material-based RRAM sample grown via chemical vapor deposition (CVD) is lower than that of the mechanically exfoliated sample, indicating superior device performance for CVD 2D-material-based RRAM. Our results indicate that among the three considered 2D materials, MoS2 exhibits the fastest switching speed.