summary: MO & state symmetry in G94/98



 Dear CCL,
 this is my summary regarding my question posted last week. I tried
 to make GAUSSIAN determine the symmetry of the MOs in a DFT
 calculation. The result, in short: it is not possible, at least not for my
 molecule ([UO2Cl4]2-, D4h, with an ECP basis set on U.)
    I got a number of interesting suggestions, and I tried them all but nothing
 helped. This seems to be an artefact of GAUSSIAN, and other programs
 don't have the problem (e.g. ADF or Jaguar, cf. also answer #5 below.)
 Here is what I did try (without success, as I said already)
 - "Symm=Loose"
 - "Conver=11", together with IntRep
 - various finer integration grids
 - fine integration grids combined with "SCF=(Conver=11, Intrep)
 - "QC" combined with all of the above
 Thanx to all who replied, and best regards,
 Georg
 -----------------------------------
 The detailed answers:
 #1: Original question:
 Dear all,
 while running a DFT single point on a symmetric (D4h) molecule
 using Gaussian98, I get the message that the program is unable to
 determine the symmetry of some orbitals, and consequently of the
 electronic state. This is problematic since I try to calculate excited
 states, and if the MO symmetries are unknown, then so are the
 excited state symmetries.
 Is there a reason for this behavior? Is there a way to make the program
 find the appropriate symmetry labels for the orbitals? -- I tried already
 SCF=(IntRep,Tight) but it had no effect.
 Thank you very much! -- I'll summarize if there is any interest.
 Best regards, Georg
 -----------------------------------
 #2
 From: theis { *at * } kalium.kiku.dk
 Date: Mon, 22 Feb 1999 09:09:42 +0100
 Dear Dr. Schreckenbach,
 I would be very intersted in a summary. I once asked Doug Fox almost the
 same question, but I never got a reply that enabled me to work my way
 around the problem. I have a feeling that only thing one can to do is to
 determine the symmetry of the state manually.
  All the best,
    Theis Soelling,
 **  **  **      **  Mr. Theis Ivan Solling
 **  **   ** ** **       Department of Chemistry
 ******    ******        University of Copenhagen
 **  **      **          DK-2100 Copenhagen
 **  **      **          Ph. 45-3532-0187/Fax. 45-3532-0212
 -----------------------------------
 #3
 Date: Fri, 19 Feb 1999 17:07:57 -0600
 From: John McKelvey <mckelvey { *at * } ncsa.uiuc.edu>
 Is there a symmetry threshold value available? i.e. Symmetry is determined
 by comparing ao coefficients. The "same" or "not the same"
 is measured by
 how large the difference between the coefficients is... Just looked in the
 manual... try SYMMETRY=LOOSE ....  You might say this is a workaround, but
 the differences judged ma be of the order of 10**-5 or smaller.. is this
 important?
 John McKelvey
 NCSA
 -----------------------------------
 #4
 From: hofmann { *at * } thkin.pci.uni-leipzig.de
 Date: Mon, 22 Feb 1999 16:20:35 +0000 (CUT)
 Hello Georg,
 we've had a similar problem. It was solved by doing a very tight
 SCF-Calculation. We used scf=(conver=11,intrep) and the unresolved
 orbitals become nice.
 Good luck
 Alex
 -------------------------------------------------------------------------
 |                                    |                                  |
 | Alexander Hofmann                  | Phone (office) +49-341-97-36337  |
 | Wilhelm-Ostwald-Institut fuer      | Fax   (office) +49-341-97-30099  |
 | Physikalische und Theoretische     |                                  |
 | Chemie                             |                                  |
 | Augustusplatz 10-11                | hofmann { *at * }
 thkin.pci.uni-leipzig.de |
 |                                    |                                  |
 | D-04109 Leipzig                    | http://www.uni-leipzig.de/~quant |
 |                                    |      /hofmann/alex.html          |
 -------------------------------------------------------------------------
 -----------------------------------
 #5
 From: George Vacek <vacek { *at * } schrodinger.com>
 Date: Wed, 24 Feb 1999 13:39:05 -0800 (PST)
 Georg,
 %% using Gaussian98, I get the message that the program is unable to
 %% determine the symmetry of some orbitals, and consequently of the
 %% electronic state.
 I'm not sure why Gaussian has this problem.  It has been around since
 the early days of Gaussian (certainly 92, most probably earlier).
 Maybe it is an artifact of the fact that the early versions of
 Gaussian didn't use symmetry at all.  In principal orbital symmetry
 identification is something that is easy to do right.  We've gotten
 compliments on our Jaguar program because it does properly identify
 orbital symmetry and everyone is used to the crappy behaviour of
 Gaussian.
 Regards,
 George
 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 George Vacek                    + Schrodinger, Inc.
 (503) 299-1150                  + 121 SW Morrison, Suite 1212
 vacek { *at * } schrodinger.com           + Portland, OR 97204
 http://www.schrodinger.com/~vacek/
 The phrase "rule of thumb" is derived from an old English law
 which stated that you couldn't beat your wife with anything
 wider than your thumb.
 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=