About ts(transition state) and NEB

[fegg7502]

hello all, I have two questions for help.
1. Is the TS search method using NEB is actually a kind of Linear Synchronous Transit (LST), or some Quadratic Synchronous Transit (QST) is implictly utilized. In other words, is there a partial optmization during the insertion of points?
2.Concering different TS search methods, the NEB, LCLIMB and LTANGENT, which is better?
thanks in advance

[graeme]

The Linear and Quadratic Synchronous Transit methods are popular for quantum chemistry calculations of small molecule reactions. In these methods, a linear or quadratic path between reactants and products is used to estimate the location of a saddle point. Then (typically) a Newton's method optimization is used to locate the stationary point(s) along the path.

The NEB can also be started from a linear path between reactants and products, but then it is a totally different method from LST/QST. In the NEB, the entire pathway is optimized to a minimum energy path. There are some advantages of the NEB for large and complex systems: you don't need to calculate the Hessian matrix, the saddle point does not need to be close to the initial pathway chosen, and you get the full pathway which can show intermediate minima and the connectivity between the stationary points.

The climbing image NEB (LCLIMB) is not implemented in the standard version of vasp, but there is code available for doing this at http://theory.cm.utexas.edu/vtsttools/ . The advantage of this version of the NEB is that the highest energy image will converge to the saddle point. If you use this method, it will automatically use the appropriate tangent definition. There are also some more efficient optimizers for converging NEB calculations in this VTSTTools package, as well as scripts to help extract results from vasp TS calculations.

[fegg7502]

thanks a lot~~!!