Scholarlink

Abstract <p>We present a framework for exploratory reactive molecular dynamics (MD) that preserves the potential energy surface and canonical bath while selectively increasing the frequency of well-oriented encounters. Sparse, event-synchronized momentum impulses are applied under a Nosé–Hoover chain (NHC) thermostat so that the mean temperature is unchanged and the potential is unmodified. We derive a BAOAB–J operator splitting and a closed-form path weight <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(w=\exp (-\beta W_{\textrm{ext}})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>w</mi> <mo>=</mo> <mo>exp</mo> <mo stretchy="false">(</mo> <mo>-</mo> <mi>β</mi> <msub> <mi>W</mi> <mtext>ext</mtext> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> for recovery of canonical statistics via reweighting. The reweighting formula is validated on the H&#xa0;+&#xa0;H<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(_2\)</EquationSource> <EquationSource Format="MATHML"><math> <mmultiscripts> <mrow /> <mn>2</mn> <mrow /> </mmultiscripts> </math></EquationSource> </InlineEquation> exchange reaction (LEPS potential), where a parameter sweep demonstrates convergence to unbiased reaction probabilities. Equilibrium properties are verified on liquid Ar (Lennard–Jones), and scalability is demonstrated on a 13,000-atom ethylene–sulfuric acid system using a GPU-accelerated machine learning potential.</p> Graphic Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Efficient simulation method for analyzing reaction products via molecular dynamics with tunable reaction probability parameters

  • 期刊论文

  • 发表日期：

    2026年7月29日

查看全文
  • Teruo Hirakawa,
  • Suguru Sakaguchi,
  • Yoshishige Okuno

摘要

Abstract

We present a framework for exploratory reactive molecular dynamics (MD) that preserves the potential energy surface and canonical bath while selectively increasing the frequency of well-oriented encounters. Sparse, event-synchronized momentum impulses are applied under a Nosé–Hoover chain (NHC) thermostat so that the mean temperature is unchanged and the potential is unmodified. We derive a BAOAB–J operator splitting and a closed-form path weight \(w=\exp (-\beta W_{\textrm{ext}})\) w = exp ( - β W ext ) for recovery of canonical statistics via reweighting. The reweighting formula is validated on the H + H \(_2\) 2 exchange reaction (LEPS potential), where a parameter sweep demonstrates convergence to unbiased reaction probabilities. Equilibrium properties are verified on liquid Ar (Lennard–Jones), and scalability is demonstrated on a 13,000-atom ethylene–sulfuric acid system using a GPU-accelerated machine learning potential.

Graphic Abstract