GRRM is a program to perform a local, semi-global, or global reaction path exploration. It is very useful for various purposes, from conventional individual reaction path calculation to exhaustive reaction path exploration and construction of complex reaction path networks. It has been applied to a variety of reaction systems such as organic reaction, organometallic catalysis, particulate catalysis, radical reaction, photoreaction involving electronically excited states, crystal phase transition under periodic boundary conditions, enzyme catalysis by the QM/MM-ONIOM method, and so on. GRRM has built-in interfaces with Gaussian03/09/16, molpro, GAMESS, ORCA, TURBOMOLE, and SIESTA, and can also be combined with any electronic structure calculation code by providing a simple code. GRRM20 comes with many new features, and four major ones are described below:
1. Kinetic simulation and kinetic navigation using RCMC
In GRRM20, a kinetic analysis method, called the rate constant matrix contraction (RCMC) method, is available, which can be applied to complex reaction path networks. RCMC can be used as (1) a kinetic navigation of SC-AFIR search or (2) an analysis method for reaction path networks obtained by SC-AFIR search. The kinetic navigation allows on-the-fly kinetic simulation of a given system under given experimental conditions (reaction temperature and reaction time). The latter allows for coarse-graining of reaction path networks, calculation of an overall reaction rate constant between the reactant and the product, and extraction of the most kinetically favorable path from reaction path networks, and so on.
2. Periodic boundary conditions (PBCs)
GRRM20 can perform geometry optimization, reaction path calculation, and automated reaction path search under PBCs. When translation vectors are active, both translation vectors and atom positions are optimized simultaneously. In addition, an automated reaction path search on a 2D or 1D slab model can be performed by fixing all or some of the translation vectors.
3. An optimizer for systems with >500 active atoms
GRRM20 provides a geometry optimization algorithm for large systems with many structural variables. The algorithm performs a geometry optimization in all dimensions using a low-dimensional PES expanded by the gradient vectors obtained during the optimization. Its performance was confirmed in systems containing up to 1000 atoms, in combination with a semi-empirical quantum chemical calculation method.
4. Super parallelization
GRRM20 has significantly improved the parallelization efficiency of SC-AFIR calculations. The development team has performed various jobs that execute 100 to 500 path calculations simultaneously. These jobs use multiple cores for each path calculation, for a total of 1000~2000 cores.
About distribution, see: https://global.hpc.co.jp/
About new features, see: S. Maeda, Y. Harabuchi, Exploring paths of chemical transformations in molecular and periodic systems: An approach utilizing force., WIREs Comput. Mol. Sci., 2021, 11, e1538 (23 pages). https://doi.org/10.1002/wcms.1538