Chain-of-State Ends Pre-Optimization Using Spherical Optimization

Chain-of-states(COS) methods like nudge elastic band(NEB) method are robust and efficient methods with double eneded, whereas two reliable ends are needed which in some cases cannot be acquired like molecule dissociation, or the two ends are too far so that more replicas are needed for convergence of minimum energy path(MEP) and transition state(TS). A method based on spherical optimization(SOPT) is presented here for pre-optimizing ends used in COS methods, in which regular geometry optimization is executed with root mean square distances(RMSD) of two ends keeping fixed. We show how this method works in two examples: (1) CHOH dissociation channel and (2) the first step of Aldol reactions.In all cases, SOPT makes MEP smoother and the convergence faster.