cc-pCV6Z for 2nd row (Al-Ar) in error

From NWChem

Viewed 1455 times, With a total of 13 Posts
Jump to: navigation, search

Clicked A Few Times
Threads 4
Posts 10
Hi,

looks like the cc-pCV6Z basis sets for the whole 2nd row are not quite right. For NWChem format they contains an extra s-type contraction which makes them linearly dependent. The molpro format is completely off. Do I need to download another set to replace them?

-Kirk

Forum Vet
Threads 9
Posts 1536
Kirk
Could you please upload a new set?
Thanks, Edo

Clicked A Few Times
Threads 4
Posts 10
Sorry Edo, I didn't realized that you'd replied to this. I just tried to upload a replacement but I get a "Stream not marked" error on step 2 of the upload.

Forum Vet
Threads 9
Posts 1536
Kirk
The whole problem for the Al-Ar series seems to be due to the fact that the BSE entry is built by grafting extra functions on top of the cc-pV6Z basis and the cc-pV6Z basis contains that (third?) contractions that causes the linear dependence you mention.
I will see what I can do to fix the problem. Let me know if my understanding of the problem is correct

Forum Vet
Threads 9
Posts 1536
BSE updated
Kirk
I have just introduced a new entry for cc-pCV6Z and renamed the old one to cc-pCV6Z (old).
Could you please check if the Molpro and NWChem output make sense?
Thanks, Edo

Clicked A Few Times
Threads 4
Posts 10
Edo,

it still looks wrong to me, e.g., the NWChem entry has 2 HF p contractions and the Molpro ones look different but odd. Can you define the cc-pCV6Z set independently of cc-pV6Z?
thanks, -Kirk

Clicked A Few Times
Threads 4
Posts 10
Edo,

you can see what I generate for Molpro (also NWChem) on my website (which is essentially the same Molpro input as what I re-submitted last week): http://tyr0.chem.wsu.edu/~kipeters/basissets/cv6z_basis.html

-Kirk

Clicked A Few Times
Threads 4
Posts 10
Edo,

sorry, I wasn't looking from the contributor side before. Now I see your new ones. I think these are now good to go!

thanks, -Kirk

Forum Vet
Threads 9
Posts 1536
New try. I have uploaded your data.
The new finding is that the bse.pnl.gov seems to neglect coefficient smaller than a certain threshold (10-3 ?)

Clicked A Few Times
Threads 4
Posts 10
The one on the contributor site marked cc-pCV6Z (2nd row) seems to be correct to me.

-Kirk

Forum Vet
Threads 9
Posts 1536
Quote:Kipeters Oct 17th 10:55 am
The one on the contributor site marked cc-pCV6Z (2nd row) seems to be correct to me.

-Kirk

Sure, that's the one you uploaded. Are you happy with the truncation of the small coefficients, too?

Forum Vet
Threads 9
Posts 1536
new update
Kirk
Could you please have a look at what you get right now using the "cc-pCV6Z" basis set?

Clicked A Few Times
Threads 4
Posts 10
Edo,

The Molpro format looks a bit strange to me. For example, the s-type functions have two contractions running from exponents 1 to 21 and then instead of uncontracting functions 12 through 21, the primitives are duplicated (yielding 22 through 31). A similar thing also happens with the p's. Compare this to what I uploaded in cc-pCV6Z (2nd row).

Forum Vet
Threads 9
Posts 1536
Kirk
I have merged your changes (cc-pCV6Z (2nd row)) into cc-pCV6Z


Forum >> BSE: Basis Set Exchange >> General BSE Topics



Who's here now Members 0 Guests 1 Bots/Crawler 0


AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC