SUMMARY: G98: How to calculate excited singlet states with ROHF ?

Dear CCL members,
 a few days ago I posted a question about the possibility to do restricted
 open shell calculations on singlet excited states with Gaussian98.
 Since the resonance was unexpectedly low (only two answers) I also
 asked a somewhat more specialized question at
 help ^at^ gaussian.com.
 Furthermore, I was asked for a summary -- here it is.
 Thanks to Uwe Richter, Doug Fox and David Shobe.
 To summarize it shortly, there is NO possibility to do RO singlet
 excited state calculations within Gaussian 98. (I intended to do RO
 DFT calculations with the B1LYP functional. Of course, there are
 other possibilities to handle excited singlet states).
 ORIGINAL POSTED MESSGAGE:
 =====================================================
 Johannes Weber <Johannes.Weber ^at^ uni-koeln.de> writes:
 > Dear CCL members,
 >
 > I have a Gaussian 98 related question:
 >
 > (How) is it possible to do calculations on excited singlet states
 > different from the ground state symmetry with the restricted open
 > shell method ?
 >
 > There is no problem with this for triplet states since there must be
 > at least two unpaired electrons in different MOs and one can change
 > these MOs by guess=alter.  Alas, for the singlets the program always
 > seems to follow the aufbau principle resulting in a closed shell
 > system. Is anybody aware of a workaround?
 >
 > Thanks in advance, :-)ohannes.
 >
 >
 > PS: On demand I will summarize the answers.
 ==========================================================
 Answers:
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 1. Uwe Richter wrote:
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 Hi Johannes,
 ...
 G98 erlaubt keine ROHF-Rechnungen fuer
 Singuletts. Versuch's mal mit Gamess (US und UK koennen es beide).
 Ich glaube, mit Turbomole geht es auch.
 Ach so, wenn du angeregte closed-shells berechnen willst (macht
 nicht allzuviel Sinn, aber es geht), kannst du mit guess=alter
 die Orbitale so hintauschen, dass du die gewuenschte Konfiguration
 hast und dann scf=symm setzen, damit die Orbitale waehrend der
 SCF-Prozedur nicht wieder zuerueckflippen. Jaja, Molcas macht
 das besser.
 Gruss,
 Uwe
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 2. Douglas F. Fox wrote:
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    Dr. Weber,
    The RO code is not setup for open shell singlets.  There is an
 option in the GVB code for spin restricted open shell singlets but
 this is setup for HF not for DFT.
    The IOP(4/17) option applies to semi-empirical calculations, l402,
 again not DFT and the IOP(5/26) option applies to l504, which works
 for HF but not for DFT.
    The best I can suggest is to use a RB1LYP ground state with the
 TD option and select singlet states.  It will give you vertical
 excitation energies not optimizations on the excited states but they
 will be spin restricted.
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 3. David Shobe wrote:
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 You can use guess=alter, and pray. :-)  Actually, if you can afford
 correlated methods, CIS does a reasonable job at finding excited states.
 >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
 Comment on Davids first sentence: An extraodinary good connection to
 heaven seems to be necessary to perform these calculations - so I
 don´t have to wonder why it didn´t work for me ;-) . More seriously,
 only excited closed shells, i.e. excitation of an electron pair, can
 be calculated in this way (see Uwes mail).
 Regards and once
 again thanks to all
 who read the mail,     :-)ohannes.
 ----------------------------------------------------------------------
 | Johannes Weber                     | Email:                        |
 | Universitaet zu Koeln              | Johannes.Weber ^at^ uni-koeln.de   |
 | Institut fuer Physikalische Chemie |                               |
 | Lehrstuhl Prof. Dr. G. Hohlneicher | tel: 0049-(0)221-470-4816     |
 | Luxemburger Str. 116               | fax: 0049-(0)221-470-5144     |
 | 50939 Koeln                        |                               |
 ----------------------------------------------------------------------