5
$\begingroup$

I am currently studying the UV-visible spectrum of ethylene using GAUSSIAN-09. I optimized the ground state(# opt b3lyp/6-31g geom=connectivity) then used the logfile to calculate the energy without optimizing it (# td=(root=10) b3lyp/6-31g geom=connectivity), but when I get the UV-visible spectrum from the log-file using Gaussview I do not know whether it is an absorption or emission spectrum, how can I distinguish between the two from my results?

When I try to run an optimization on the excited-state(#p opt freq=savenm td=(nstates=6,root=1) rb3lyp/6-31g(d) geom=connectivity), I get an error (Error termination request processed by link 9999). What causes this error? How can I solve it? And is it necessary to optimize the structure in the excited state to have the UV-visible spectrum?

How can we generate an absorption spectrum and an emission spectrum separately using GAUSSIAN09? What keywords do we have to add?

Can you recommend articles, books, or tutorials about GAUSSIAN software?

$\endgroup$
2
  • 3
    $\begingroup$ +1 Welcome to Matter Modelling SE! Please note that you can render codes with backticks like this: ` `. Also note the formatting edits. About your question, I think Gaussiview should be able to show both absorption and emission spectra. And please also consider uploading the output file because it is difficult to understand what's going wrong with just an error code. $\endgroup$ – S R Maiti Jun 15 at 14:13
  • $\begingroup$ Thank you for your comment and for welcoming me into Matter Modeling. $\endgroup$ – sarah bnm Jun 15 at 19:29
7
$\begingroup$

Roughly speaking, absorption spectra are obtained by TDDFT calculations on optimized ground-state geometries, while emission spectra are obtained by TDDFT calculations on optimized excited state geometries. These are the results of the Franck-Condon approximation, which says that there is a high probability that a molecule is near its equilibrium geometry, and electronic transitions are much faster than geometry changes of the molecule, so that during an electronic transition, the molecular geometry can be viewed as staying unchanged in the equilibrium geometry of the initial state. Since you did a ground-state geometry optimization, the spectrum is an absorption spectrum. If you want to calculate the vibronic features of the absorption spectrum, then you may also need to optimize the excited state geometry, but this is an advanced issue.

With a simple "Error termination request processed by link 9999" we cannot analyze the cause of the error. You should post at least a few tens of lines before this line, or better, upload the complete output file. If Gaussian aborts due to an error, the cause of the abort usually cannot be deduced solely from the last few lines, or from the last line that contains the word "error".

It's beneficial to take a careful read of the descriptions of keywords and example input files on gaussian.com. I would also recommend you to refer to general quantum chemistry textbooks that are not specific to Gaussian, as many problems that you encounter when using Gaussian are in fact quantum chemistry problems that exist regardless of what software you use, rather than "what keyword should I use"-type problems that are specific to the software.

$\endgroup$
12
  • 1
    $\begingroup$ +1 for a rapid and thorough answer, which is no doubt helpful to the new user that asked this question! I've just made a slight edit to format some text in a code block. $\endgroup$ – Nike Dattani Jun 15 at 14:28
  • 1
    $\begingroup$ @sarahbnm No, that's not wrong - the Franck-Condon approximation only applies to the structure on which you do your TDDFT calculation (in other words, the final, converged structure of your geometry optimization), not the initial guess structure of your geometry optimization. To find the excited state geometry you certainly have to use something that is not the excited state geometry as the initial guess. And the ground state geometry is the best initial guess if you know nothing about how the excited state geometry should look like. $\endgroup$ – wzkchem5 Jun 16 at 6:36
  • 1
    $\begingroup$ @sarahbnm The point is that, when you open the output file of your TDDFT single point job, GaussView does not know whether your geometry is the ground state geometry or the excited state geometry, so it plots both an absorption spectrum and an emission spectrum. Only one of them makes sense, and you should pick out the correct one by yourself. $\endgroup$ – wzkchem5 Jun 19 at 10:33
  • 1
    $\begingroup$ @sarahbnm I mean, if you are doing the TDDFT calculation on the ground state geometry, then Gaussview's absorption spectrum makes sense and can be used, but Gaussview's emission spectrum is useless; to obtain the correct emission spectrum, you have to do another TDDFT calculation on the excited state geometry, for which Gaussview again gives you two spectra, one absorption and one emission, yet this time only the emission spectrum makes sense and the absorption spectrum means nothing at all. $\endgroup$ – wzkchem5 Jun 19 at 11:08
  • 1
    $\begingroup$ @sarahbnm Besides there are two problems with your input file: (1) fluorescence emission usually occurs from the S1 state (even if the S1 state is a dark state), so you should write root=1 instead of root=10, in the excited state geometry optimization calculation. If you have strong reason to believe that the emission occurs from a higher state (for example, S2), then you may choose a "root" value other than 1, but 10 is essentially never the correct answer, as high-lying states usually decay exclusively via nonradiative pathways and you don't see emission from these states at all. $\endgroup$ – wzkchem5 Jun 19 at 11:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.