From owner-chemistry@ccl.net Mon Apr 8 04:19:01 2019 From: "Susi Lehtola susi.lehtola * alumni.helsinki.fi" To: CCL Subject: CCL:G: G09-Excessive mixing of frozen core and valence orbitals Message-Id: <-53673-190408041808-26172-lkxDwg8gY1FUKqfcM0H56A(-)server.ccl.net> X-Original-From: Susi Lehtola Content-Language: en-US Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=utf-8; format=flowed Date: Mon, 8 Apr 2019 11:17:56 +0300 MIME-Version: 1.0 Sent to CCL by: Susi Lehtola [susi.lehtola%a%alumni.helsinki.fi] On 4/6/19 1:19 PM, Jacob Berenbeim jacob.berenbeim*|*york.ac.uk wrote: > > Sent to CCL by: "Jacob Berenbeim" [jacob.berenbeim###york.ac.uk] > I'm getting inconsistent results with TD-DFT calculations performed > on organic alkali M+(Na+, K+, Rb+) clusters. These systems have > previously been optimized and confirmed to be minima at the > wb97xd/defsvp level using Gaussian 09. A small number of my low > energy cluster structures are terminating early when running TD-DFT > from these optimized chk files at the same computational level. I've > included the termination dialog below. I'm not using a pseudo > potential for most of the terminating cases. I'd appreciate any > advice or tips. Dear Jacob, this is a well-known issue in calculations with alkali metals: the definition of core orbitals by the orbital eigenvalue can lead to the wrong orbitals being included in the active space. This problem has been described e.g. by Petrie in J. Phys. Chem. A 102, 6138 (1998) and by Rassolov, Pople, Redfern and Curtiss in Chemical Physics Letters 350, 573 (2001). The problem is that the semi-core orbitals of the alkali metals can have orbital energies that are higher than the valence orbitals of other species, which means that the valence orbitals end up frozen while the semi-core orbitals are included in the active space of the TD-DFT calculation. This can be solved either by redefining the core orbitals by studying the orbital character visually or via Mulliken analysis like Rassolov et al, or by increasing the size of the active space such that all valence orbitals are included therein. Pierre Archirel already suggested using td=full, which means dropping the active space altogether. A less costly option would be to use just a slightly larger active space e.g. the inner noble gas core. -- ------------------------------------------------------------------ Mr. Susi Lehtola, PhD Junior Fellow, Adjunct Professor susi.lehtola**alumni.helsinki.fi University of Helsinki http://susilehtola.github.io/ Finland ------------------------------------------------------------------ Susi Lehtola, dosentti, FT tutkijatohtori susi.lehtola**alumni.helsinki.fi Helsingin yliopisto http://susilehtola.github.io/ ------------------------------------------------------------------