[molpro-user] two questions about virtual orbitals

[Dan Kats]

Dear Karl

I don't have a right answer either, but if you calculate CCSD through
FCI with these "frozen" orbitals, you will get the same result, which
is sort of reassuring. Besides, I have tried to calculate MP2 with
this kind of frozen orbitals and compared to MP2 results with orbital
energies for the virtuals in question set to 10^6 (with and without
zeroing out the orbitals itself) - and it gives the same result, which
is again reassuring. If you need more evidence, you can in principle
compare fci with zeroed orbitals with mcscf with an appropriate active
space...

Best wishes
Daniel

2013/11/6 Irikura, Karl K. Dr. <karl.irikura at nist.gov>:
> Dear Daniel,
>
> Thank you!  It appears to be working (with MATROP inside two nested loops), although I don't have any "right" answers for comparison.
>
> Best regards,
> Karl
>
>
> -----Original Message-----
> From: Dan Kats [mailto:D.Kats at bristol.ac.uk]
> Sent: Wednesday, November 06, 2013 11:39 AM
> To: Peterson, Kirk
> Cc: Irikura, Karl K. Dr.; molpro-user molpro-user
> Subject: Re: [molpro-user] two questions about virtual orbitals
>
> Dear Karl
>
> You can "freeze" the orbitals by setting them to zero in matrop, e.g.
> for orbital 10.1:
>
> hf
> {matrop
> LOAD,orbs,ORB,2100.2
> ADDVEC,orbs,10.1,orbs,10.1,-1
> SAVE,orbs,2100.2,ORBITALS
> }
>
> and then run ccsd calculation without orbital orthogonalization:
> {ccsd
> thresh,orth=false
> }
>
> http://www.molpro.net/pipermail/molpro-user/2012-December/005147.html
>
> Best wishes
> Daniel
>
> 2013/11/6 Peterson, Kirk <kipeters at wsu.edu>:
>> Hi Karl,
>>
>> sorry, no way to freeze virtuals in Molpro. I think also no on your
>> 2nd question.
>>
>> best,
>>
>> -Kirk
>>
>> On Nov 6, 2013, at 6:23 AM, Irikura, Karl K. Dr.
>> <karl.irikura at nist.gov>
>> wrote:
>>
>> Hi,
>>
>>
>>
>> Is there a way to freeze virtuals in CCSD calculations?  I'd like to
>> compare valence-space FCI and CCSD.
>>
>>
>>
>> Also, is there a way to improve the virtuals (like IVO), to make them
>> better correlating orbitals?  Then I might not need orbital
>> optimization before attempting the above.
>>
>>
>>
>> Thanks for any tips!
>>
>>
>>
>> Karl
>>
>>
>>
>>
>>
>> -------------------------------------------------------------------
>>
>> Dr. Karl K. Irikura
>>
>> National Institute of Standards and Technology
>>
>> 100 Bureau Drive, mail stop 8320
>>
>> Gaithersburg, MD 20899-8320
>>
>> voice:  301-975-2510   fax:  301-975-3670
>>
>> email:  karl.irikura at nist.gov
>>
>> http://www.nist.gov/mml/csd/informatics_research/index.cfm
>>
>> -------------------------------------------------------------------
>>
>>
>>
>> _______________________________________________
>> Molpro-user mailing list
>> Molpro-user at molpro.net
>> http://www.molpro.net/mailman/listinfo/molpro-user
>>
>>
>>
>> _______________________________________________
>> Molpro-user mailing list
>> Molpro-user at molpro.net
>> http://www.molpro.net/mailman/listinfo/molpro-user