I am using Gaussian09 to run a structure minimisation on a molecular system that is radical cationic using the composite method ROCBS-4M. The input structure has already been optimised at UB3LYP/6-31G* level. After two normal terminations, the ROCBS-4M has a convergence failure at step 3 doing the following Job:
Link1: Proceeding to internal job step number 3.
-----------------------------------------------------
#N Geom=AllCheck Guess=TCheck SCRF=Check MP4SDQ/6-31G
-----------------------------------------------------
I have tried using the end structure of this ROCBS-4M for a rerunning of structure minimisation using same composite method but getting the same convergence error. how can i change the input route to ensure that I only change the convergence criteria at step 3, since these are predefined calculations?
Any suggestions are welcome and thank you for your attention!
Look in the Gaussian manual under the keyword scf. There you will find several options that may lead to convergence for your system. I usually turn to scf=(qc,maxcyc=999) first. Some trial-and-error can be expected.
P.S. You do know about ROCBS-4M=restart, I hope. That will keep you from having to redo the first two steps.
Hi Bojidarka, Thanks for your reply. I couldn't find a proper documentation on this command and some says that this should never be used and that it is only used for debugging purposes. could you explain what it does to the calculations please? Thanks.
Hi David, thanks for your suggestion. I took a look at the gaussian manual on qc which states that:
"The SCF=QC option is often helpful with difficult conversion cases. For difficult-to-converge ROHF wavefunctions, where QC cannot be used, add Use=L506 to the route section." Nonetheless thank you for pointing me in the right direction and letting me know about ROCBS-4M=restart option.
# ROCBS-4M=restart Use=L506
returns error for my system:
"ILSW ROHF field is now 1.
GVB not converged.
Error termination via Lnk1e"
I can't see to find a comprehensive explanation on how to overcome this error..
On a side note, will I be able to use other ab initio methods that could give better energy for the system, e.g. CCSD and CCSD(T), when ROCBS-4M wave function does not converge? I have tried the default Unrestricted Open-shell CBS-4M but the wavefunctions are highly spin-contaminated. are there any other ways to avoid spin-contamination? I read that spin-projection methods (https://www.researchgate.net/post/How_can_I_eliminate_the_spin_contamination_in_unrestricted_states
http://www.ccl.net/cca/documents/dyoung/topics-orig/spin_cont.html) can be used, but could anyone advise how this is done in Gaussian? Many thanks!
[just to confirm it, i also used
# ROCBS-4M=restart scf=(qc,maxcyc=999)
but it returns the error: No SCF=QC or DM with ROHF. (as per manual) ]
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry