[molpro-user] MP2 pair energies

[Matthias Heger]

Hello everyone,

thanks for your replies! I'm still toying around with double-hybrid functionals and the question came up when I looked at my outputs. I didn't have a simple way to see if the energy components I obtained were really the right thing, so I wanted to double check and got confused.

Thanks,
Matthias

--
Dr. Matthias Heger
Department of Chemistry, University of Alberta
11227 Saskatchewan Drive NW
Edmonton, AB T6G 2G2, Canada

Office: CCIS 4-151
Lab: CCIS 4-108
Email: heger at ualberta.ca

Am 06-Dec-2017 um 5:49 AM schrieb Daniel Kats:
> Dear Matthias
> 
> SCSFACS and SCSFACT correspond to coefficients for the opposite-spin and
> same-spin energy contributions, respectively. You can get those from the
> singlet/triplet contributions in the closed shell as
> E_OS=E_singlet+(1/3)*E_triplet
> E_SS=(2/3)*E_triplet
> 
> Best wishes
> Daniel
> 
> On Wed, Dec 6, 2017 at 7:33 AM Matthias Heger <heger at ualberta.ca> wrote:
> 
>> I am a bit confused about the MP2 pair energies reported in my
>> calculations. Specifically, my attached example yields:
>>
>>> SCS-MP2 correlation energy:           -0.198842375046   (PS=  1.200000
>> PT=  0.333333)
>>> (...)
>>>   MP2 singlet pair energy              -0.126127623239
>>>   MP2 triplet pair energy              -0.076321972220
>>
>> Scaling the pair energies manually and adding them yields a different
>> energy (-0.176792... Hartree) than given above.
>>
>> In addition, I tricked the SCS calculation to only pick up the singlet or
>> triplet pairs with SCSFACS and SCSFACT. That yields SCS correlation
>> energies different from the corresponding pair energies in the output.
>>
>> Am I interpreting this wrong?
>>
>> Best,
>> Matthias
>>
>> --
>> Dr. Matthias Heger
>> Department of Chemistry, University of Alberta
>> 11227 Saskatchewan Drive NW
>> Edmonton, AB T6G 2G2, Canada
>>
>> Office: CCIS 4-151
>> Lab: CCIS 4-108
>> Email: heger at ualberta.ca
>> _______________________________________________
>> Molpro-user mailing list
>> Molpro-user at molpro.net
>> http://www.molpro.net/mailman/listinfo/molpro-user
>