[CP2K-user] [CP2K:21022] Re: CAM_B3LYP usage

[Frederick Stein]

Dear Miguel,
I added a hint to the manual to set the HF section explicitly in case of 
hybrid calculations. It is available in the current master trunk.
Best,
Frederick

Miguel RP schrieb am Dienstag, 7. Januar 2025 um 00:08:04 UTC+1:

> Hi Beliz,
>
> Thank you so much for your response. It has been incredibly helpful, as I 
> am now obtaining reasonable values for my titania nanoclusters.
>
> I believe the addition of the "INTERNAL POTENTIAL" subsection is crucial 
> and should be specified in the CP2K manual, particularly in the XC part (
> https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/XC/XC_FUNCTIONAL.html
> ).
>
> I'm still a bit puzzled because, in calculations involving formaldehyde, I 
> did not explicitly define this subsection, and yet everything seemed to 
> work well.
>
> Anyway, once again, thank you so much for your assistance, it has been 
> invaluable to me!
>
> Cheers,
>
> Miguel
>
>
> El domingo, 5 de enero de 2025 a las 19:05:09 UTC+1, Beliz Gökmen escribió:
>
>> Hi Miguel,
>>
>> Your setup of the XC_FUNCTIONAL is correct, you are only missing this 
>> section inside HF:
>>          &INTERACTION_POTENTIAL
>>             POTENTIAL_TYPE MIX_CL_TRUNC
>>             CUTOFF_RADIUS ${cutoff}
>>             OMEGA 0.33
>>             SCALE_COULOMB 0.19
>>             SCALE_LONGRANGE 0.46
>>             T_C_G_DATA t_c_g.dat
>>          &END
>> This is needed for the mixed short-range long-range part of the 
>> functional. Furthermore, the HF fraction should be set to 1.0, otherwise 
>> all these values in the interaction potential will be multiplied by the 
>> number you specify. The cutoff value should be around half the cell length.
>>
>> It is not straightforward to compare the cp2k values to gaussian, as 
>> gaussian uses VWN3 local correlation functional whereas cp2k uses VWN5 for 
>> B3LYP. 
>>
>> Another comment is that your cutoff value for the vdw pair potential is 
>> larger than your cell size, you might want to check this. Also, since you 
>> have a runtype geo_opt, I assume you are interested in optimising the 
>> geometry of the first excited state. Then you should include
>>   &EXCITED_STATES
>>     STATE 1
>>   &END EXCITED_STATES
>> in DFT, where the state should be the state of interest.
>>
>> Best,
>> Beliz
>>
>> On Sunday, 5 January 2025 at 03:52:15 UTC-6 Miguel RP wrote:
>>
>>> Hi All,
>>>
>>> I am attempting to perform a TD-DFT calculation using the CAM-B3LYP 
>>> functional on a simple 3-atom titania system. However, the energies I am 
>>> obtaining differ significantly from those produced by a different code 
>>> (Gaussian), even if I use the same functional and all-electron basis-set. 
>>> Interestingly, when using the B3LYP functional, the results align well with 
>>> Gaussian outputs. This suggests that the discrepancy is likely due to 
>>> incorrect usage or configuration of the CAM-B3LYP functional with this 
>>> titania system (in the past I got reasonable values with same functional 
>>> applied to a formaldehyde molecule)
>>>
>>> Below is the relevant section of my input file, specifically the XC 
>>> definition:
>>>
>>>   &XC
>>>       &XC_FUNCTIONAL
>>>         &HYB_GGA_XC_CAM_B3LYP
>>>           _AC 0.81
>>>           _ALPHA 0.65
>>>           _BETA -0.46
>>>           _OMEGA 0.33
>>>         &END HYB_GGA_XC_CAM_B3LYP
>>>       &END XC_FUNCTIONAL
>>>       &HF
>>>         &SCREENING
>>>           EPS_SCHWARZ 1.0E-12
>>>         &END
>>>         &MEMORY
>>>           MAX_MEMORY  10000
>>>           EPS_STORAGE_SCALING 0.1
>>>         &END MEMORY
>>>         FRACTION 0.19
>>>       &END
>>>       &VDW_POTENTIAL
>>>          POTENTIAL_TYPE PAIR_POTENTIAL
>>>          &PAIR_POTENTIAL
>>>             PARAMETER_FILE_NAME dftd3.dat
>>>             TYPE DFTD3
>>>             REFERENCE_FUNCTIONAL B3LYP
>>>             R_CUTOFF [angstrom] 16
>>>          &END
>>>       &END VDW_POTENTIAL
>>>     &END XC
>>>
>>> I am uncertain about the correct value for the "FRACTION" parameter in 
>>> the HF section. Should this parameter align somehow with the _ALPHA, _BETA, 
>>> and _OMEGA values specified for the CAM-B3LYP functional?
>>>
>>> I would greatly appreciate any guidance or suggestions on correctly 
>>> configuring the CAM-B3LYP functional for this calculation. I attach my 
>>> input and output files.
>>>
>>> Thank you for your help!
>>>
>>> Best regards,
>>> Miguel
>>>
>>

-- 
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+unsubscribe at googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/f4c977ce-ae90-402f-a848-1fd1354698f7n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250109/034041dd/attachment.htm>