TDDFT non-adiabatic calculation with Gaussian 09 does not produce non-adiabatic coupling terms

986 views
Skip to first unread message

agdij...@gmail.com

unread,
Nov 22, 2013, 10:40:16 AM11/22/13
to new...@googlegroups.com
Dear Newton-X users,

We are trying to run the TDDFT non-adiabatic test calculation (MD-SICH4-G09-TDDFT-NAD) with Gaussian 09 (Rev. A.02) and Newton-X 1.4.0-2.  Newton-X and Gaussian seem to be running fine. However, the calculated non-adiabatic coupling terms remain all zero for all time steps.

from nx.log:

Nonadiabatic coupling terms v.h (a.u.):
     0.0000000
     0.0000000
     0.0000000


Does anybody know what can cause this problem?

Thanks.

Rachel Crespo

unread,
Nov 28, 2013, 11:55:27 AM11/28/13
to new...@googlegroups.com
Do you get 0 couplings for all steps or for the first one? Send me the files and I can take a look. 
Best, 
Rachel 

agdij...@gmail.com

unread,
Nov 29, 2013, 2:53:57 AM11/29/13
to new...@googlegroups.com
Dear Rachel,

Thank you for your reply! We get zeros for all time steps. I sent you some of the files by e-mail.

Thanks,

Arend.

Felix Plasser

unread,
Nov 29, 2013, 3:25:46 AM11/29/13
to new...@googlegroups.com
With Turbomole you'd get this type of output when the double-molecule calculation (in TEMP/tmp.old/overlap) failed. My guess is that it may be the same thing here?

Error handling for this double-molecule calculation is kind of difficult, since this is an artificial step, which can't actually produce a reasonable result in the sense that the program performing this calculation would believe. With Turbomole error handling happens in run_cioverlap_turbo.pl
line 574:
 if ( $ao2 != 2*$ao ){
      die "$mdle Internal inconsistence in ao count: ao2=$ao2; ao=$ao.\n";
    }

Maybe something like this would also be needed for Gaussian.

Best regards,
Felix
--
You received this message because you are subscribed to the Google Groups "Newton-X" group.
To unsubscribe from this group and stop receiving emails from it, send an email to newtonx+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Rachel Crespo

unread,
Nov 29, 2013, 4:09:38 AM11/29/13
to new...@googlegroups.com
Dear Ared, 

Please check if you are using the same basis set for the dynamics and the couplings, check that the basis set is the same
in the "gaussian.com" file and in the "basis" file inside JOB_AD. You are using 3-21G* and in the original file the basis
set is 3-21G, if you didn't change "basis" you can have that problem, let me know if this is the case. 

Best wishes, 
Rachel  

Message has been deleted

agdij...@gmail.com

unread,
Dec 6, 2013, 10:34:03 AM12/6/13
to new...@googlegroups.com
Dear all,

Rachel helped a lot in identifying and solving the problem. The revision of Gaussian we are using (A.02) does not support a keyword used by Newton-X that is available in later versions. This caused the zeros in the non-adiabatic coupling terms. There is an IOp that can be used to solve the issue, and the program now appears to be working. Thanks a lot for your follow-up.

best regards,

Arend.

Shiladitya Banerjee

unread,
Jun 10, 2015, 6:55:07 AM6/10/15
to new...@googlegroups.com
Dear Arend

Hi, I am having the same problem while calculating NAC vectors with TDDFT-g09 and Newton-X. It will be very kind if you could please help me. Which IOP should be used for solving the problem?

Thanks a lot :-)
Best regards
Shiladitya

Rachel Crespo-Otero

unread,
Jun 10, 2015, 7:50:13 AM6/10/15
to new...@googlegroups.com
Dear Shiladitya, 

Substitute the $NX/bin/run_cioverlap_g09.pl for this one. It should work.  

Best Regards, 

Rachel 
run_cioverlap_g09.pl

Shiladitya Banerjee

unread,
Jun 12, 2015, 3:50:00 AM6/12/15
to new...@googlegroups.com
Dear Rachel

Many thanks for your kind reply :-) . I had a problem with the wrong 'vdoth' option, but now it is okay. Thanks a lot.

Best wishes and regards
Shiladitya

--
You received this message because you are subscribed to a topic in the Google Groups "Newton-X" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/newtonx/KTqmQUC8eg8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to newtonx+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Pedro Castro

unread,
Oct 18, 2015, 9:34:54 PM10/18/15
to Newton-X
Dear all,
I have a similar problem with  the NAC, but these are calculated only relating excited states (S1 and S2), but the NAC related with S0 always are zero.
I download newton-x recently (OCT-2015), since I can calculate the NAC for the excited states I guess I am using one version updated in order to include the IOp in gaussian. But, what about the NAC relating S0 with the excited states? Since I can not calculate them, I can not get population transfer from S1 to S0.
May be I am doing something wrong, I would appreciate any suggestion and comments.
Best regards,
Pedro
Reply all
Reply to author
Forward
0 new messages