# what ways of TS searches are available in Gaussian09

I have a system for which I could not find the transition state using Gaussian09 corresponding to the experimental product formation. I have tried

opt=modredundant


route command fixing the two atoms in the (C-C) bond formation. It worked well showing the vibration in the bond formation in this direction (one -ve frequency of about -200), however, when i release the constraint and do a TS search, it gave a TS for a remote bond rotation (no longer along C-C bond formation).

I have tried to do a PES scan along the C-C bond forming direction varying the bond length from 3 Angstrom to 1.2 Angstrom, there is no maxima (but a minimum) within this range; the energy of the system only increases when the C-C bond approaches too close to each other.

What other ways can I try?

• I'm interested in this, can you explain how you want to find a TS using frozen coordinates? Every time you set a coordinate to a non-minimum configuration you will get at least one negative frequence, this doesn't say anything about if that's a TS. And did you do berny afterwards? – DSVA Nov 17 '16 at 9:55
• yes, you are right, i think. when you freeze the atoms for which you expect the bond formation and do an opt=modredundant, you are checking if you could get a TS after you release the constraint. indeed a berny optimisation for TS search is done after you have confirmed the existence of a negative vibrational frequency in the direction of the desired bond formation. – X Zhang Nov 17 '16 at 23:46

You could also try the Synchronous Transit-Guided Quasi-Newton (STQN) method requested by Opt=QST2 or Opt=QST3 in the route section. It is described in many different places, including the Gaussian official web-site.

Note that

QST2 requires two molecule specifications, for the reactant and product, as its input, while QST3 requires three molecule specifications: the reactant, the product, and an initial structure for the transition state, in that order.

Another important thing to whatch for is that

The order of the atoms must be identical within all molecule specifications.

• thanks Wildcat! yes indeed! this is one method that I forgot! will try and update if it works for my system! – X Zhang Nov 10 '16 at 11:44
• it did not work for my system; it gave a C-C rotation outside the C-C bond-forming carbon atoms. This is the same as the remote bond rotation that I obtained when I release my constraint in TS search starting from the modredundant results.. – X Zhang Nov 11 '16 at 17:31
• @XZhang, without details it would be difficult to suggest something. – Wildcat Nov 11 '16 at 18:16