MLatom Software Ecosystem for Surface Hopping Dynamics in Python with Quantum Mechanical and Machine Learning Methods

Abstract

We present an open-source MLatom@XACS software ecosystem for on-the-fly surface hopping nonadiabatic dynamics based on the Landau−Zener−Belyaev−Lebedev algorithm. The dynamics can be performed via Python API with a wide range of quantum mechanical (QM) and machine learning (ML) methods, including ab initio QM (CASSCF and ADC(2)), semiempirical QM methods (e.g., AM1, PM3, OMx, and ODMx), and many types of ML potentials (e.g., KREG, ANI, and MACE). Combinations of QM and ML methods can also be used. While the user can build their own combinations, we provide AIQM1, which is based on Δ-learning and can be used out-of-thebox. We showcase how AIQM1 reproduces the isomerization quantum yield of trans-azobenzene at a low cost. We provide example scripts that, in dozens of lines, enable the user to obtain the final population plots by simply providing the initial geometry of a molecule. Thus, those scripts perform geometry optimization, normal mode calculations, initial condition sampling, parallel trajectories propagation, population analysis, and final result plotting. Given the capabilities of MLatom to be used for training different ML models, this ecosystem can be seamlessly integrated into the protocols building ML models for nonadiabatic dynamics. In the future, a deeper and more efficient integration of MLatom with Newton-X will enable a vast range of functionalities for surface hopping dynamics, such as fewest-switches surface hopping, to facilitate similar workflows via the Python API.

Universal and Updatable Artificial Intelligence-Enhanced Quantum Chemical Foundational Models

“…This is true for both 0 th iteration of UAIQM which knew nothing about GMTKN55 and for later iterations of UAIQM which we allowed to improve on a limited number (six) of GMTKN55 subsets it was performing badly (see SI note 5 UAIQM provides a totally practical solution and can easily perform various types of simulations with high accuracy. While above we showed its prowess for energy calculations, dynamics, and IR spectra simulations, the methods should be applicable in other simulations, e.g., for emission spectra 45 and excited-state dynamics 46 as was shown for the platform's predecessor AIQM1 which is part of the library of UAIQM models. The final remark is that the online availability of the UAIQM platform is enabling calculations around the world for research groups with varying access to computational resources.…”

Surprising new dynamics phenomena in Diels–Alder reaction of C60 uncovered with AI

“…The initial conditions were sampled from the harmonic quantum Boltzmann distribution 27,45 at 298 K using the normal modes analysis of TS. In addition, the TS state of the Diels-Alder reaction of C60 with 2,3-dimethyl-1,3-butadiene has several low-frequency modes which can result in very distorted structures in normal mode sampling (as observed before for similar sampling procedures 44 ), hence, we set these frequencies as 100 cm −1 if they are smaller than this value; this procedure was introduced here but later found useful also in the context of nonadiabatic dynamics submitted earlier than this work 46 .…”