Author Topic: Rydberg states using molecule-centred rydberg basis set  (Read 2460 times)

marand

  • Jr. Member
  • **
  • Posts: 11
  • Karma: +0/-0
Rydberg states using molecule-centred rydberg basis set
« on: April 25, 2014, 11:48:38 am »
I would like to attempt TDDFT calculations of excited states with inclusion of Rydberg basis set. I would like, however, to avoid using the atom centred basis (like d-aug-cc-pVXZ) . I would rather use the atom-centred basis (like aug-cc-pVXZ) for desription of valence states, adding a hand-made even-tempered basis of very diffuse functions centred in the molecular centre of charge.

Such an approach works very well for the CC2 method.

For DFT, however, there is a problem with grid for a dummy atom, which serves as the artificial atomic centre for the Rydberg basis set.

Is there a way in which I could specify this grid in the control file? If the answer is affirmative, any practical advice on what would be a good grid for such very diffuse orbitals would be most welcome.

Best regards!
Marcin Andrzejak


mpjohans

  • Full Member
  • ***
  • Posts: 26
  • Karma: +0/-0
    • .
Re: Rydberg states using molecule-centred rydberg basis set
« Reply #1 on: May 05, 2014, 08:20:32 pm »
Never tried this, so might not work as expected, but perhaps worth a shot.

Define a large ghost atom, say gold, with a charge of zero as the centre of your Rydberg basis functions. Then in the $dft section, add the diffuse parameter to increase the radial grid. You might want to play around with other grid settings as well. See the manual for the different options for the DFT grid: http://www.turbomole-gmbh.com/manuals/version_6_5/Documentation_html/node336.html

But as I said, this might not work at all...