[CPMD-list] problems with TDDFT geometry optimizatioin

[Juerg Hutter]

Hi

newer versions of CPMD have improved geometry optimizers
with constraints (LBFGS).
There were also improvements to the TDDFT code.

However, I don't think your actual problem would be solved
this way.
I see two possible reasons for the breakdown
1) your system has a crossing of ground and 1. excited state
   at 90 degrees (either real physical or only within TDDFT
   using adiabatic approximation and local functionals).
   This happens e.g. for the rotation of double bonds.

2) A breakdown of the TDDFT diagonalizer, resulting in a
   spurious solution with dE=0. Newer versions of CPMD
   are slight more stable in this respect.

regards

Juerg Hutter

----------------------------------------------------------
Juerg Hutter                   Phone : ++41 44 635 4491
Physical Chemistry Institute   FAX   : ++41 44 635 6838
University of Zurich           E-mail: hutter at pci.unizh.ch
Winterthurerstrasse 190
CH-8057 Zurich, Switzerland
----------------------------------------------------------


On Fri, 10 Jun 2005, Eneritz Muguruza gonzalez wrote:

> CPMD users,
>
>
>        I am doing TDDFT torsion constraint geometry optimization
>  calculations for an isolated molecule. I am calculating the relaxed
>  structures within an excited state at several torsional angles. For the
>  angles 180, 135, 45 and 0 everything goes more or less well, but for
> the twisted tau=90 structure the geometry optimization crashes. At the
> beginning of the geometry optimization the forces and energies go down
>  well, but at a certain step the energy drops drastically (to a value
>  similar to the one in the ground state) while the forces increase very
>  much (1 order of magnitude). Looking at the output file, at this point
>  the 2nd order perturbation energy goes to 0.
>
>        Does anybody know why this is happening? Is there some problem
> with the code? How can it be solved?
>
>  I would really appreciate if somebody could give me a hint. Thanks!
>
>
>                       Eneritz
>
>
>  I attach the output file so you can have a look.
>