Example of optimizing to a saddlepoint with the Newton-Raphson method (continued)

The Newton-Raphson step is H^-1g, so

(6)

· This is close to the saddlepoint (which is at x = 0, y = 0). Another step will take us closer still.

· The optimizer gradually converges to the saddlepoint.

· Various methods similar to the Newton-Raphson approach have been developed to optimize saddlepoints in this way, when the starting structure is close to the transition state structure.

Next: so how do we get close to a saddlepoint structure?