Author Topic: Egrad - Calculate gradient of a2u state but TM suggests a1g  (Read 4987 times)

martijn

  • Full Member
  • ***
  • Posts: 63
  • Karma: +0/-0
Hi,

This is probably something silly but as it puzzles me, I hope someone can answer this question for me. I'm calculating the gradient on a molecule in the excited state state using TD-B3LYP, Turbomole 6.2 and I'm using symmetry (something I normally rarely do, and why I probably never noticed the following before) so my control file contains the following entry:

$scfinstab rpas
$soes
 a2u          2
$exopt 1
$pop
$denconv 1d-7

And I then run egrad to get the excitation energies of the lowest two excitations belonging to the a2u irrep. For the excitation energies this seems to work fine. The Davidson procedure starts and Turbomole clearly prints that it's looking at the a2u irrep:

 Iteration IRREP Converged      Max. Euclidean
                 roots          residual norm

    1       a2u      0        3.916733017530943D-02

etc etc etc.

And then prints the requested a2u excitation energies:

 ==============================================================================

                              I R R E P   a2u

 ==============================================================================

etc etc etc.

However, when it gets to the gradient step, it suddenly mentions a1g instead of a2u:


                       ----------------------
                        CPKS right-hand side
                       ----------------------


          IRREP   tensor space dimension   number of roots

           a1g            1692                   1

                          ----------------
                           CPKS equations
                          ----------------


 logfile dipl_a1g  exists already

 found      1 converged vectors on dipl_a1g
 read       1 vectors from logfile dipl_a1g


 Switching to coarse grid

 preparing numerical integration ....
          Overall gridpoints after grid construction =          2291


                       Block Davidson iteration


 total number of roots to be determined:   1


 maximum core memory set to   200 MB,
 corresponding to      142 vectors in CAO basis


 maximum number of simultaneously treated vectors (including degeneracy):   1


 Iteration IRREP Converged      Max. Euclidean
                 roots          residual norm

    1       a1g      0        2.470146090332557D-02
 
    2       a1g      0        1.009075158737974D-02
 
    3       a1g      0        1.539081680139929D-03
 
    4       a1g      0        3.409914673031397D-04
 
    5       a1g      0        6.717589632064532D-05
 
    6       a1g      1        9.079853125902758D-06
 

 converged!


Is this is what means to happen? Do the gradients of an irrep span another irrep? I'm confused.

Thanks in advance,

Martijn