Author Topic: Totally symmetric vibrations in NumForce  (Read 3601 times)


  • Jr. Member
  • **
  • Posts: 18
  • Karma: +0/-0
Totally symmetric vibrations in NumForce
« on: February 06, 2017, 10:43:40 AM »
Dear Users and Developers!

I wander if it is possible to make NumForce calculate only the totally symmetric vibrations. I am aware that for excited states NumForce works properly only in C1, but the symmetry information would be used only in producing only such displacements, for which the molecular symmetry is preserved.

Why would it be useful? In cases of closely lying excited states of different symmetries, keeping of the molecular symmetry would preven the states from interacting with one another, even though formally the electronic calculations are done without symmetry. So at least the totally symmetric vibrations could be obtained with reasonable accuracy, and these are often the most important for vibrational activity that determines bandwidths and bandshapes.

If such a possibility is not available in Turbomole, perhaps you could tip me off, whether a code exists that includes this functionality and can do CC2 at the same time. However, for the latter Turbomole seems to be the best choice by far.

Yours sincerely
Marcin Andrzejak


  • Developers
  • Sr. Member
  • *
  • Posts: 239
  • Karma: +0/-0
Re: Totally symmetric vibrations in NumForce
« Reply #1 on: February 09, 2017, 11:59:34 AM »
Dear Marcin,

it is indeed possible to let NumForce produce displacements in symmetry for excited state vibrational modes.
If your lowest excited state is 1b1, you would optimize it with such settings:
  geoopt model=adc(2) state=(b1 1)
  irrep=b1  multiplicity=  1  nexc=  1
After jobex is done, change it to:
  geoopt model=adc(2) state=(a 1)
  irrep=a  multiplicity=  1  nexc=  1
And then run "Numforce -level cc2 -c". That  should do the trick.
The exited state energies and gradients will then be calculated in C1, but the displacements are done in symmetry.