[CP2K:8650] Re: questions about TDDFPT2

[Hanning Chen]

Matt,

  Thanks for your clarification. I agree with you that the coupling term
between occupied-virtual orbital pairs may deviate the excitation energy
from HOMO-LUMO gap. I was just surprised by the rather large deviation on
an isolated molecule which may stem from its small basis set as well as its
small box size, as your suggested. I will make the system bigger with a
larger basis set.

Best,

Hanning


On Mon, Feb 6, 2017 at 10:22 AM, Matt W <MattWa... at gmail.com> wrote:

> Hi Hanning,
>
> I've not had a chance to look at your particular calculation. Such a small
> basis set, and small box could lead to strange results in any case. More
> generally, there is no reason, even if the transition is a pure single
> determinant HOMO to LUMO, for the energy to be that of the HOMO_LUMO gap.
> There are still additional terms in the TDDFT kernel that will shift the
> transition energy.
> For GGAs the shifts can be in either direction, there is no general rule.
> We have checked small molecule transition energies vs gamess-us and were
> pretty happy with the comparison.
>
> HTH,
>
> Matt
>
>
> On Sunday, February 5, 2017 at 8:50:49 PM UTC, NUCP2K wrote:
>>
>> Dear CP2K Developers and Users,
>>
>>   I just tried out the latest implementation of time-dependent density
>> functional perturbation theory (TDDFPT) in CP2K, and found some serious
>> inconsistencies on the calculated excitation energy.
>>
>>   I tested the TDDFPT2 on an isolated water (H2O) molecule using the
>> attached input file. Interestingly, the calculated excitation energy of
>> 10.075 eV  is much much greater than the HOMO-LUMO gap of 9.137 eV, even
>> though the transition is dominated (>99.9%) by the electron transfer from
>> HOMO to LUMO.
>>
>>   It seems to me that the diagonal terms of the TDDFPT response matrix
>> are not appropriately assigned, leading to the apparent inconsistency.
>> Perhaps, the subroutine "tddfpt_init_by_energy_diff" in
>> qs_tddfpt2_methods.F should be fixed to yield correct orbital energy gap.
>>
>>   Please correct me if my understanding is wrong.
>>
>>   Thanks.
>>
>> Hanning
>>
>>
>>          &FORCE_EVAL
>>   METHOD QS
>>   &PROPERTIES
>>     &TDDFPT
>>      NSTATES      1
>>      MAX_ITER    100
>>      MAX_KV 10
>>      CONVERGENCE [eV] 1.0E-3
>>      &DIPOLE_MOMENTS
>>       DIPOLE_FORM LENGTH
>>      &END DIPOLE_MOMENTS
>>      &MGRID
>>       COMMENSURATE
>>       CUTOFF 50
>>      &END MGRID
>>     &END TDDFPT
>>   &END PROPERTIES
>>   &DFT
>>     BASIS_SET_FILE_NAME BASIS_SET
>>     POTENTIAL_FILE_NAME POTENTIAL
>>     &MGRID
>>       COMMENSURATE
>>       CUTOFF 50
>>     &END MGRID
>>     &POISSON
>>      PERIODIC NONE
>>      POISSON WAVELET
>>     &END POISSON
>>     &SCF
>>       SCF_GUESS ATOMIC
>>     &END SCF
>>     &XC
>>       &XC_FUNCTIONAL PBE
>>       &END XC_FUNCTIONAL
>>       &XC_GRID
>>         XC_DERIV SPLINE2_SMOOTH
>>       &END XC_GRID
>>     &END XC
>>   &END DFT
>>   &SUBSYS
>>     &CELL
>>       ABC 6.0 6.0 6.0
>>       PERIODIC NONE
>>     &END CELL
>>     &COORD
>>    O     3.000000     3.000000     3.000000    H2O1
>>    H     3.000000     3.000000     4.000000    H2O1
>>    H     3.942809     3.000000     2.666667    H2O1
>>     &END COORD
>>     &KIND H
>>       BASIS_SET SZV-GTH
>>       POTENTIAL GTH-PADE-q1
>>     &END KIND
>>     &KIND O
>>       BASIS_SET SZV-GTH
>>       POTENTIAL GTH-PADE-q6
>>     &END KIND
>>     &TOPOLOGY
>>     &END TOPOLOGY
>>   &END SUBSYS
>> &END FORCE_EVAL
>> &GLOBAL
>>   PROJECT WATER
>>   PRINT_LEVEL MEDIUM
>>   RUN_TYPE ENERGY
>> &END GLOBAL
>>
> --
> You received this message because you are subscribed to the Google Groups
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to cp2k+uns... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at https://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170207/3e150d1c/attachment.htm>