From owner-chemistry@ccl.net Thu May 16 09:51:00 2013 From: "Johannes Hachmann jh=chemistry.harvard.edu" To: CCL Subject: CCL:G: Orbital energies for neutral and cation species in vacuum and solution Message-Id: <-48702-130516090641-23733-pGDiVp1ighV498Yf1rQN7w]=[server.ccl.net> X-Original-From: "Johannes Hachmann" Content-Language: en-us Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="US-ASCII" Date: Thu, 16 May 2013 09:06:27 -0400 MIME-Version: 1.0 Sent to CCL by: "Johannes Hachmann" [jh^-^chemistry.harvard.edu] Hi Scott, The decision between U and RO is tricky and in my opinion there is no golden bullet - both approaches have their problems :). The U results may not give you a good quantum state, in particular if you have excessive spin contamination, while RO is not very flexible and does not allow for DODS, i.e., the R part cannot adequately account for the interactions with the unpaired electron. In my experience, U results tend to be better, but one has to be more careful how to interpret them. If you have very large spin polarization this indicates that your electronic structure is too complicated for a single-reference approach and you may want to try something more sophisticated than DFT. All bets are off then for the DFT results: you know that RO will be bad, while U results *may* still be ok (but somewhat artificially so). If you use U for the cation, you should try using U for the neutral molecule as well (don't forget to break spin symmetry in your guess!). If the molecule is complicated and the neutral one already shows spin polarization, you *may* get some error cancellation, since both species are treated on the same footing. I would then also use U for the respective geometry optimizations. Hope this helps Best Johannes ----------------------------------------------- Dr. Johannes Hachmann Research Associate Harvard University Department of Chemistry and Chemical Biology 12 Oxford St, Rm M104A Cambridge, MA 02138 ----------------------------------------------- > -----Original Message----- > From: owner-chemistry+jh==chemistry.harvard.edu++ccl.net [mailto:owner- > chemistry+jh==chemistry.harvard.edu++ccl.net] On Behalf Of Scott McKechnie > jsm78]![cam.ac.uk > Sent: Thursday, 16 May, 2013 07:28 > To: Hachmann, Johannes > Subject: CCL:G: Orbital energies for neutral and cation species in vacuum and > solution > > Dear Dr. Hachmann and Dr. Kumar, > > > Many thanks for your very helpful replies and for the excellent references. > > > Did you set up your delta-SCF as restricted open shell or unrestricted? > > I used the default settings in Gaussian09 which is unrestricted for the cation, > but I would appreciate any advise here. I notice that the delta SCF values only > differ by 0.07 eV but the orbital energies are significantly different (cation gas > phase alpha HOMO and LUMO values of -6.48 eV and -6.36 eV as compared > with the unrestricted beta gas phase HOMO and LUMO values of -10.65 eV and > -8.76 eV). > > > I had two follow up questions here. > > > 1. Using ROB3LYP results in degenerate alpha and beta orbitals? > > > 2. Is it advised to do cation optimization with UB3LYP instead of B3LYP? > > > Thanks again and very best wishes, > > Scott > > > > > On 15 May 2013 04:30, Ramesh Kumar rameshchitumalla^gmail.com chemistry,+,ccl.net> wrote: > > > > Sent to CCL by: Ramesh Kumar [rameshchitumalla(a)gmail.com] > Dear Scott, > > The reply from Dr. Johannes Hachmann is really very useful. > The orbital energy of a cation is much lower than that of its > corresponding neutral molecule and again it comes very near to the > case similar to neutral molecule in the presence of a polar solvent. > For some numbers you can see CPL 396 (2004) 43-48. > > > > -- > With Best Regards: > > CH. Ramesh Kumar > Senior Research Fellow, > Computational Chemistry Lab, > Indian Institute of Chemical Technology(IICT), > Tarnaka, > Hyderabad.> > > > E-mail to subscribers: CHEMISTRY,+,ccl.net or use:> > E-mail to administrators: CHEMISTRY-REQUEST,+,ccl.net or use> Conferences: > http://server.ccl.net/chemistry/announcements/conferences/> > > > >