ROKS with HFX using ADMM basis sets

[Matt W]

Hi,

On Monday, September 25, 2017 at 11:16:50 AM UTC+8, Xiaoming Wang wrote:
>
> For singlet excited state calculation, one usually sets
>
>     MULTIPLICITY 3
>
>     ROKS
>
>     &LOW_SPIN_ROKS
>
>      ENERGY_SCALING 2.0 -2.0
>
>      SPIN_CONFIGURATION 1 2
>
>      SPIN_CONFIGURATION 1 1
>
>     &END LOW_SPIN_ROKS
>
>
> Can I get the singlet state with the &BS section? 
>
>
>     MULTIPLICITY 1
>
>     LSD
>
>     &BS
>
>      ......
>
>     &END 
>
>
>
the &BS just changes the initial guess. So you can generate a symmetry 
broken initial guess, this may converge to an excited state if you have a 
sensible initial geometry.

There might be some stuff in the constrained DFT  section that would allow 
to generate an approximate excited state geometry.

Alternatively you could try the Maximum Overlap Method, but we find it to 
be hard to converge unless you have a very well separated S1 state.

LR-TDDFT is nearly ready for ADMM calculations, but we don't have forces 
yet.

Matt


 

> Best,
>
> Xiaoming
>
>
> On Saturday, September 23, 2017 at 4:23:31 PM UTC-4, Xiaoming Wang wrote:
>>
>> Hi,
>>
>> I want to calculate the excited singlet state with HFX for which I am 
>> using ADMM basis sets, but was told that the combination of ROKS and ADMM 
>> is not implemented. Any work around over this?
>>
>> Best,
>> Xiaoming 
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170924/6c7a4452/attachment.htm>