Python(编程语言)
可视化
计算机科学
工作流程
计算科学
插值(计算机图形学)
标杆管理
算法
趋同(经济学)
力场(虚构)
势能面
势能
过渡状态
国家(计算机科学)
统计物理学
领域(数学)
能源景观
数据可视化
理论计算机科学
作者
Bastiaan van Hoorn,Patrick Norman,Mårten S. G. Ahlquist
摘要
A force field interpolation method for generating initial guesses for transition state optimizations is presented. The user supplies reactants and products as SMILES strings (or XYZ coordinates), after which automatically constructed force fields are interpolated to approximate the potential energy surface (PES) along the reaction coordinate. The transition state guess is obtained by sampling along this approximate PES. Reliance on force fields ensures the method is transparent in its workings and computationally inexpensive. The workflow is implemented in VeloxChem, enabling Jupyter notebook execution with a few lines of Python code. An interactive widget makes visualization and inspection of results easy and intuitive, and a flexible Python API facilitates integration into complex automated workflows. The method is demonstrated on ten chemically diverse systems, including a transition-metal catalytic cycle. Benchmarking across 121 reactions resulted in convergence for 115 transition states, with an average cost of 107 gradient evaluations per reaction-comparable to more demanding double-ended methods.
科研通智能强力驱动
Strongly Powered by AbleSci AI