Electronic structure of CeO2 greatly depends on initial guess of wave function in OT method

[stclair]

Hi, Florian,

Sorry for the misunderstanding.

I restart the calculation from the wave functions obtained from another U 
value: for example, I start the geometry optimization with U=7eV from the 
wave function obtained from U=10eV. The calculation in this way can get 
lower work function which is comparable with experimental value. Also the 
total energy is ~0.86 eV lower than the calculation with 'scf_guess' set as 
'atomic' (same U value used, U=7eV). It seems the later didn't converge to 
a global minimum.

Best,

On Monday, April 6, 2015 at 1:14:15 PM UTC-5, Florian Schiffmann wrote:
>
> Hi,
>
> could you clarify what you mean when you say restarting with a different U 
> value?
>
> If you use a different value for U in the initial and the restarted 
> calculation there will be for sure a difference in energy and computed 
> properties. Maybe I am getting you wrong so clarification would be useful.
>
> Flo
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20150406/8c802ccf/attachment.htm>