Hi Mathieu,

another way to get IRC results but not use the IRC keyword, is by using the "manual displacement" on the transition state structure, however this works only if you have Gauss View.

1)If you are using GaussView to view your molecules, you will find that upon opening the log file that contains the vibrational frequencies of your transition state, there is an option of "manual displacement".

2) The "manual displacement" for the imaginary frequency is set at 0, however you can slide it to the left or right of 0. Check the "manual displacement" box and then slide either left or right of 0, and then save the structure on either side of 0 separately to generate a guess structure of the local minimum that the TS should theoretically converge to. And you can then optimize these two structures to their respective local minimum.

3) I have found this option to work well whenever my IRC doesn't work out.

All the best.

You can increase the corrector steps by irc( maxcycles=100). Default is 20. Reduce stepsize to 5 it sometime works. 

However, if you do not need very accurate determination of each point along the path,  turn off the "recorrection" with "IRC=ReCorrect=Never"

Hello Debasish. I have tried to use maxcycles=100, stepsize=5 and several combinations of both with scf=qc or scf=conver=4 but nfortunately, nothing has helped. However, thank you for your advice.

Dear Jakub,

It will be helpful if you please mention the magnitude of the imaginary frequency of the converged TS calculation.

Of course, the value of the frequency is -120.35 cm-1 (I use Avogadro and there imaginary frequencies are signed with "-") and its intensity is 356.775 km/mol. I have rather added the output .log file from the optimization of the TS.

I suggest that you should add ReCalc=-N in IRC option. If you believe your transition state is true, this problem may be not  geomtric convergence on the IRC step. The ReCalc=-N means the analytic Hessian matrix every N step in correction step.

Hello, Zhao Chongyang,

thank you for your advice. If I use ReCalc=-20, SCF=QC and I do the calculation for vacuum, I will obtain the IRC path, but only for forward direction. And if I use ReCalc=-20, SCF=(Conver=4) and I do the calculation for vacuum, I will obtain the IRC path, but only for reverse direction. And in advance, both IRC paths I have obtained, for forward and reverse direction respectively, look identically.

Hi Jacob,

may I ask a question : are you interested in knowing the path IRC takes to optimize to a minimum on either side of the transition state (TS) or are you interested in knowing what each of the minima on either side of the TS look like? Because if its the latter then there s another way to get the structures of the potential minima.

Abhigya Mookherjee Could you tell me what would be your strategy?

Hi Mathieu,

another way to get IRC results but not use the IRC keyword, is by using the "manual displacement" on the transition state structure, however this works only if you have Gauss View.

1)If you are using GaussView to view your molecules, you will find that upon opening the log file that contains the vibrational frequencies of your transition state, there is an option of "manual displacement".

2) The "manual displacement" for the imaginary frequency is set at 0, however you can slide it to the left or right of 0. Check the "manual displacement" box and then slide either left or right of 0, and then save the structure on either side of 0 separately to generate a guess structure of the local minimum that the TS should theoretically converge to. And you can then optimize these two structures to their respective local minimum.

3) I have found this option to work well whenever my IRC doesn't work out.

All the best.

Abhigya Mookherjee That is awesome, you are a genius! Thank you very much miss.

Abhigya Mookherjee after optimize these two structure (from "manual displacement" of imaginary frequency of TS) should it required to run IRC again?

@Saikat Mandal - sorry for the delayed reply. An IRC calculation is not reuiqred after you have optimized the structures by manual displacement.