*From*: "Dr. Lars Goerigk" <lars.goerigk_-_chem.usyd.edu.au>*Subject*: CCL:G: Unexpected error in G09 Rev B.01*Date*: Sun, 13 Jan 2013 09:42:42 +1100

Hi Sam and Massimo, just a comment to Sam's comment. Using CIS for DFT does make sense and several programs also allow to do that. Then one usually speaks of TDA-DFT, where TDA stands for the Tamm-Dancoff approximation. From another point of view, if you take the RPA equation (TD-HF) and evoke TDA, then the resulting equation is the CIS equation. As linear response TD-DFT is very similar to the TD-HF equations, you can do the same for TD-DFT and get TDA-DFT and I assume that is what Massimo wanted to do when he tried to evoke CIS/B3LYP. However, I am not sure whether Gaussian allows it. I know that NWCHEM and ORCA allow to do TDA-DFT. Excitation energies are very similar to those of full linear response TD-DFT, at a lower computational cost. Transition dipole moments were reported to be more effected by this approximation, however. I hope this could help. Cheers, Lars --
Dr. Lars Goerigk School of Chemistry (Building F11) The University of Sydney, NSW 2006 Australia Profile: http://www.researchgate.net/profile/Lars_Goerigk List of publications: www.researcherid.com/rid/D-3717-2009 |