The same method can be used to optimize transition states, if the
starting structure is not too far from the saddle point (MOPAC
In that case the negative value for the reaction coordinate should be found, plus some other negative values for other normal coordinates.
When sufficiently close to the TS, these values will be closer to zero than the one belonging to the reaction coordinate. Using this distinction, the lowest, most negative eigenvector should be followed, to maximize the energy, and the other ones in the opposite direction, minimizing the energy. This algorithm can be made to converge at a stationary point, where one negative value remains.
See the MOPAC manual for a more detailed description.
Back to Chapter 1B.