[ADF-LIST] TD-DFT计算1°)SAOP与B3LYP; 2°)近乎退化

Erik Van Lenthe([email protected] 支持SCM.com
2月25日12:47:32 CET 福彩3d字迷

Dear Jean-Marie,

1. B3LYP is much more expensive than SAOP.
The time is proportional to the number of requested excitations.
The timings in each step can vary if the number of trial vectors is 
close to the maximum number of allowed trial vectors.
After this maximum has been reached the space will be collapsed in the 
Davidson routine.

General remark:
In ADF for degenerate representations such as the 2-dimensional 
E-representations or the 3-dimensional T-representations, the occupation 
should be either fully occupied or zero,
in case one want to calculate excitation energies.
Degeneracy can be lifted by using a lower symmetry, which does not have 
multi-dimensional irreps.
Use integer occupations in this lower symmetry.

Possibly one could use a high symmetry calculation as a restart in the 
lower symmetry calculation.

Best regards,

You wrote:
> Dear all,
> I have encountered lately two difficulties with TD-DFT calculations, 
> and I would
> appreciate any insight or experience about the two following points :
> 1°) I wanted to compute TD-DFT transitions for an organic (CNH 
> containing) dimer
> of about 60 atoms. I had no problem with the *saop *XC potential (a 
> few hours on a
> 8-cores WorkStation). When I tried running the same with *B3LYP *(as 
> prescribed by
> some in the literature, especially for mono-cations), it took much, 
> much more time
> in the Davidson procedure, with different times for each cycle ... and 
> I gave up.
> Is it also your experience (*saop versus B3LYP*) ? Did I miss akeyword 
> or trick here ?
> 2°) For totally different systems (but still organic CNH-based), I 
> encountered a
> *degeneracy *problem for a (D(3) symmetrical) cation species (LUMO, 
> HOMO and
> HOMO-1 are nearly degenerate). I read on the manual that all 
> representations had
> to have filled or empty orbitals. But mine had one electron in the 
> HOMO, E1 (beta)
> representation, and I got the message : "Total number of occupied 
> orbitals not equal
> to the number of valence electrons" ... as expected. How to circumvent 
> this technical
> problem ? I tried the following trick : run the saop TD-DFT 
> calculation with NOSYM
> and a small lshift value. For a small molecule (benzene cation), it 
> worked and, for
> various lshift values, the transitions are shifted as expected (i.e. 
> all calculations yield
> the same transitions after correction for the lshift values). It is 
> trickier for bigger molecules:
> is there an alternative way to achieve convergence AND following 
> *TD-DFT for systems**
> **exhibiting near degeneracies* ?
> Thanks for any input !
> Jean-Marie (France)
> -- 
> Dr. Jean-Marie Mouesca
> Laboratoire de Résonances Magnétiques
> Service de Chimie Inorganique et Biologique
> C5-312 / Bâtiment 51C
> C5 (office) +33 4 38 78 30 13
> 51C (office) +33 4 38 78 65 71
> _______________________________________________
> ADFlist mailing list
> ADFlist at scm.com
> http://lists.tofoba.com/mailman/listinfo/adflist

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <//lists.tofoba.com/pipermail/adflist/attachments/福彩3d字迷0225/b509cf19/attachment-0001.html>