[molpro-user] explicitly corr. methods and core electron correlation?

[Kirk Peterson]

Dear Raphael,

yes, Grant Hill and I have recently developed core-valence correlation  
consistent basis sets for use in F12
calculations for core correlation (1st and 2nd row atoms). Overall it  
works as well as one might expect for F12
(our new CVDZ-F12 basis sets  yield core-valence correlation effects  
on things like Re and De that are as accurate
as conventional CVTZ or CVQZ) but one has to pay extra attention to  
the value of geminal exponent.  We should be submitting
a paper to JCP on this over the next couple of weeks and can send you  
a preprint if you're interested. In collaboration with
Hans-Joachim Werner's group in Stuttgart, we are also actively working  
on all the post-d elements that include outer-core
correlation effects. So there will be lots of information in the  
literature in the near future.

Of course without core correlation, there have already been a few  
papers published that have used F12 for weak interactions
(check out Werner's recent pubs as well as Klopper).

-Kirk

On Sep 29, 2009, at 5:29 AM, Berger Raphael wrote:

> Dear Readers,
>
> has any one got experience with core electron correlation in  
> combination
> with F12. Especially it would be interesting to know how the
> intermolecular BSSE would be affected by approaching the basis set  
> limit
> using F12. For instance in case of a method like
> CCSD(T)-F12b/aug-cc-pVQZ + core electron correlation
>
> Thank You for reading
> 	R. Berger
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user