# Are multiple transition states possible? Criteria to choose the correct one?

I am studying a reaction using electronic structure methods using various software packages (ORCA,GAMESS, G09) and have found two possible transition structures between product and reactant. Both have imaginary frequencies, but each has a different energy value, frequencies and geometries. Some software coincide in the found transition state while other differs, adding to the confusion. So my questions are:

1. Is it possible to have multiple transition states, if so what is a good criteria to choose between candidate states?

2. Should I expect the frequency to be within certain range? (for example if the reaction is the movement of a H atom within the molecule)

3. Is there some quantitative value(s) (energy, vibration amplitude/frequency, vibration 3d direction vectors, 3d geometry, etc.) that can be used as a reasonable measure to choose the "correct" state? (That is to avoid visualizing the imaginary mode of vibration or computing the IRC).

• Is it possible to have multiple transition states connecting the same two points on a potential energy surface (e.g. reaction pathway)? Sure! For a hydrogen transfer, you would expect to have a an imaginary frequency with a large magnitude associated with that motion. Not sure what you mean by question number 3. Also, have you tried performing an IRC using the transition states you characterized? They should connect the TS to your reactants and products. – LordStryker Apr 10 '14 at 12:32
• Edited: question 3 to clarify. – beangoben Apr 10 '14 at 12:44
• WRT to #3: Not that I'm aware of. The only quantitative (and valid) procedure to ensure a TS connects the poitns on a reaction pathway that you are interested in is to do something like an IRC. – LordStryker Apr 10 '14 at 12:48
• You could look at the frequencies and figure out the displacements (in the X, Y, or Z directions) of each particular atom, but that still doesn't really tell you what you're looking for (plus it takes more effort to do this over visualizing the mode in a program). – LordStryker Apr 10 '14 at 13:12
• One comment: Many TS search algorithms exist, and sometimes they may lead to different points along the PES. I generally use the Freezing String Method (FSM), and in most cases it has helped me successfully locate the actual (physically more meaningful) TS structure. You might want to read more here: scitation.aip.org/content/aip/journal/jcp/135/22/10.1063/… – user3786990 Oct 3 '14 at 20:50