I have a problem with QST2 calculations. My job has failed with error: Old curvilinear step not converged, using linear step. I compare initial geometries, atoms of the reagent and product are in the same order.
Can anyone help me with this issue? I have attached my file. Thank you so much!
Hai Olga Kovalenko,
Initially i have faced this problem.... In actual scenario, u should use optimized coordinates of reactants and products...but you will face the error wharever you got... So, what i practicing is....
Take any one geometry either reactant or product...use that geometry as such....for example, take geometry of reactant and use that coordinates. now open the geometry of reactant that u have considered using gaussview, and now try to modify that geometry so as to get product geometry by simple breaking and making bonds....keep in mind, in doing so dont use clean option...The geometry should resemble your product...take the coordinates of the modified geometry.Now use these two coordinates in input file of QST2....it will run...I hope you got it....
all the best....
This problem: " Old curvilinear step not converged, using linear step. " does not mean that you have wrong order of atoms in coordinates specification; if you had it gaussian will die much faster.
In the first step of qst2 gaussian tries to guess the geometry of the transition state by doing very primitive calculations based on superposition of your reactants and products. For a non-trivial transition state this step fails in many cases, since transition state geometry is rarely just a simple midpoint between the reactants and products.
There is a few ways to deal with it. First, change the geometries of reactants and products. In most cases they should not be optimized stationary points, but rather some virtual structures which are before and after reaction, but just barely "before" and "after". Also, adding a reasonable guess of the ts and switching to qst3 helps a lot.
Hai Olga Kovalenko,
I dont think your reaction is a single step concerted process. try it in complex reaction pathway. i think your first step involves breaking C-O bond and making C-N bond and diene will be in cis confirmation (cis confirmation is your intermediate here). and your second step will be transformation from cis to trans....
in relevance to my answer, i have performed calculations with PM6 level to finish fast...i will attch the results here....you can check them and you may follow them with your original level of theory.... This is for your understanding.....
all the best.
Dear Ch Venkata Surya Kumar, if I understood, as a result of this calculations we obtain an optimized cis form, but IRC calculations don't match to this cis form because after IRC calculations we optimized a cis. Am i right?
- Badges
- Science topic