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?