A Novel Method to Simulate Molecular Collisions

Although usually ignored for a more intuitive macroscopic understanding, molecular collisions are the foundation of almost every chemical event - from chemical reactions to transport phenomena. On the molecular level, a chemical reaction is an exchange (or sometimes knocking off) of atoms between molecules as a result of energetically significant collisions. The current simulation paradigm is based on simulating bulks of molecules, requiring high computational resources. In such simulations sometimes the molecules are not allowed to react, as in the case of classical MD (Molecular Dynamics), or require significant computational resources and time to conclude, as in the case of ab initio MD. Without any fine-tuning to allow a more outcome-oriented approach, the results of such simulations require even more computational resources and sometimes human effort to analyze for meaningful results. To answer these limitations, a novel methodology is developed based on a more fine-tuned approach for simulations. Its Python implementation based on current state-of-the-art libraries, and supporting applications based on globally accessible web technologies, are presented as a software framework named collider.py. It focuses on fine-tuning the collision parameters before running the simulation. Individual linear velocities, positions, and orientations of the molecules can be defined as well as the impact points on each molecule. The result of a proper collision usually involves re-arranging the atoms into new species, which reflects the output of such a chemical reaction. Our framework is based on Atomic Simulation Environment - a computational chemistry library for Python and consists of a web application to design the collisions in detail.