Author Topic: escf for open-shell  (Read 592 times)

cdicker2

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
escf for open-shell
« on: August 25, 2020, 06:44:11 PM »
Hi,
Does turbomole v6.6 support escf for open-shell systems?  If not, does any version?

And if not, can I use the two-component SO included in v6.6 (soghf) when performing spin flip single point calculations?
« Last Edit: August 25, 2020, 08:51:30 PM by cdicker2 »

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 486
  • Karma: +0/-0
Re: escf for open-shell
« Reply #1 on: August 26, 2020, 11:47:28 AM »
Hello,

the 6.6 version from June 2014 can be used for open shell systems using escf (use $scfinstab urpa for unrestricted TDDFT calculations, check the documentation), two-component TDDFT calculations are also possible as this was one of the new features of the 6.6 release.


cdicker2

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Re: escf for open-shell
« Reply #2 on: August 27, 2020, 01:18:33 AM »
Thank you!  I am currently using stuttgart ecp with def2-TZVPP basis set for an f transition.  Is this sufficient?  I read in the manual it recommends 2-c ECPs for scf?

I am noticing the scf energy when there is 1 alpha electron vs . scf energy when there is 1 beta electron are very close in energy.  For Yb+3 I would have expected larger difference, around 1.3 eV.  Does the spin-orbit operator not capture this in Turbomole or am I choosing the incorrect basis set?
« Last Edit: August 28, 2020, 09:19:52 PM by cdicker2 »