[molpro-user] *** [SPAM] *** Re: Frequency analysis of electronic excited states

Tatiana Korona tania at tiger.chem.uw.edu.pl
Wed Aug 16 23:41:27 CEST 2017


Dear Pablo,

These problems are not related to state degeneracies, nor to some vibronic 
interactions. The point group or water (C2v) is Abelian, so the states cannot be 
degenerate (irreps are one-dimensional).

In the case of finite field calculations (like numerical calculations of 
Hessian) the symmetry of the system can be lowered, so you should specify the 
lowest symmetry common for all displacements by dx, dy, or dz. The safest way is 
to give just C1 (nosym), and apparently it is default. This means however that 
there is only one irrep (i.e. irrep no 2 does not exist, i.e. it is illegal). 
You should specify the number of your state in C1 instead.

For the combination 3.1+forced C2v for freq you got probably a saddle point.

Best wishes,

Tatiana


  On Wed, 16 Aug 2017, Pablo Avaria wrote:

> Dear Tatiana,
> thank you for your reply. I forgot to put  "variable, energy(2);" in the frequency card.
> As I understand, it is expected that the optimization results 
> are different if different point groups were chosen? 
> For 2.2 state I experienced the following problems. 
> Molpro chooses the C2v symmetry for H2O for optimization by default.However, for freq calculations it chooses C1 group and then failswith the error: 
>  ?ILLEGAL SYMMETRY ON EOM CARD: 2 
> If I force Molpro to use the C2v symmetry for freq calculations,then it fails with different error:  Attempt to include non-existent symmetry X 
>  ... ? The problem occurs in zmatrix.f:zmat_evaluate
> For 3.1 state C2v group works for freq analysis, but gives imaginary frequencies.It may be related to the fact that optimization for C2v group convergesto a saddle point.  
> The only successful way was when I chose the nosym option both for opt and freq analysisfor 3.1 state, in this case I got real freqs and no warning about the gradient.
> Are these problems related to state degeneracies or some vibronic interactions?
>
> Saludos cordiales,Pablo Avaria
>
>    On Monday, August 14, 2017 6:29 AM, Tatiana Korona <tania at tiger.chem.uw.edu.pl> wrote:
>
>
> Dear Pablo,
>
> If you want frequencies for the 2.1 state, you should specify this explicitly,
> otherwise they are calculated for the default (ground) state. You did it for the
> optimization, but forgot to do the same for the frequency job.
>
> Best wishes,
>
> Tatiana
>
>   On Fri, 11 Aug 2017, Pablo Avaria wrote:
>
>> Hello,
>>
>> I was trying to use EOM-CCSD for finding harmonic frequencies of H2O molecule for excited states. First, I optimized the geometry for 2.1 state, see first attachment.Then, I did freq analysis, see second attachment.
>> However, it looks like that the freqs are calculated for a different electronic statethan was used for optimization.
>> Optimization: Results for state  2.1: Excitation energy         0.36677468 au  Freq: Results for state  2.1: Excitation energy         0.34742486 au  Moreover, the freq output contains the line  ?WARNING: GRADIENT NORM NOT ZERO! which confirms that the geometry used is not the equillibrium one. Should I set some parameters manually in the freq input?I suspect it is related to the symmetry somehow, but I don't know how to manage it. 
>> Regards,
>> Pablo
>
> Dr. Tatiana Korona http://tiger.chem.uw.edu.pl/staff/tania/index.html
> Quantum Chemistry Laboratory, University of Warsaw, Pasteura 1, PL-02-093 Warsaw, POLAND
>
>

Dr. Tatiana Korona http://tiger.chem.uw.edu.pl/staff/tania/index.html
Quantum Chemistry Laboratory, University of Warsaw, Pasteura 1, PL-02-093 Warsaw, POLAND


More information about the Molpro-user mailing list