High-Threshold-Stability GeSe-Based Selector-Only Memory Enabled by Gallium Doping

Selector-only memory (SOM), which utilizes threshold voltage (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {V}}_{\mathbf {\textit {th}}}$ </tex-math></inline-formula>) shift in ovonic threshold switching (OTS) devices for data storage, has emerged as a promising storage-class memory (SCM) candidate due to its low latency, high density, and cost-effectiveness. However, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {V}}_{\mathbf {\textit {th}}}$ </tex-math></inline-formula>instability, including variability and drift, remains a major reliability challenge. Here, we propose a gallium (Ga) doping strategy to enhance <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {V}}_{\mathbf {\textit {th}}}$ </tex-math></inline-formula>stability. Guided by double energy well OTS model, we employed ab-initio molecular dynamics (AIMD) calculations to analyze the effects of various dopants on the coordination numbers (CNs) and bonding characteristics that govern <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {V}}_{\mathbf {\textit {th}}}$ </tex-math></inline-formula>stability in SOM materials, identifying Ga as the optimal dopant. Compared to undoped GeSe-based SOM (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\gamma =35.1$ </tex-math></inline-formula> mV/dec, endurance <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$= {10}^{{6}}$ </tex-math></inline-formula> cycles), Ga-doped SOM exhibits significantly enhanced performance: reduced drift (16.5 mV/dec), improved endurance (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\gt {10}^{{9}}$ </tex-math></inline-formula> cycles), suppressed variability (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sigma ~0\lt 0.2$ </tex-math></inline-formula> V, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sigma 1\lt 0.068$ </tex-math></inline-formula> V, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\sigma $ </tex-math></inline-formula> is defined as the standard deviation between <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${\mathrm {V}}_{\mathbf {\textit {th}}}$ </tex-math></inline-formula>), and enhanced thermal stability (>400°C). Notably, excessive doping degrades performance, aligning with theoretical predictions and validating our screening approach. This work provides both mechanistic insights and practical guidance for developing highly reliable SOM devices.