Any chance using user-defined or program-fitted auxiliary basis functions?

[葉哲和]

Dear Molpro developers,

Recently I have encountered a scenario where my molecule is very large and therefore density fitting methods seemed like the only way out.

It seemed to me that Molpro uses a pre-defined library of auxiliary functions like AVDZ/JKFIT or AVDZ/MP2FIT. But say for example if I am still not happy with one layer of diffuse functions but instead attempted to use d-aug-cc-pVDZ in density fitting methods, the program would abort to an error saying

 ? Error

 ? Cannot find D-AUG-CC-PVDZ-JKFIT-FIT in basis catalogue

 ? The problem occurs in binput

 I wanted to use more delicate basis sets like the d-aug-cc-pVDZ mentioned above, or even better, user-defined basis sets with density fitting. Is it possible in Molpro? (Maybe a relatable example is ORCA's "autoaux" keyword)

Best regards,

Che-Ho Yeh

[Tatiana Korona]

Dear Che-Ho Yeh,

It is possible, e.g.:

basis={
default,vtz

set,mymp2fit
default,avqz/mp2fit

set,myjkfit
default,vqz/jkfit
}

df-hf,basis=myjkfit
df-mp2,basis=mymp2fit

Best wishes,

Tatiana

--
You received this message because you are subscribed to the Google Groups "molpro-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/molpro-user/338dc4c9-344a-4f8a-aaaa-76243011fb56n%40googlegroups.com.

[葉哲和]

Dear Tatiana,

Thank you for the quick response.

I get it that setting different auxiliary functions is possible by what you have shown. But my main problem is actually that I would like to add more layers of diffuse functions to the basis, like d-aug-cc--pVDZ did by having even-tempered reduction of the orbital exponents, OR, even better, to set customary (user-defined) basis sets, for which it seems like Molpro does not have the functionality to. It seems to me that the best I can do so far is to use

basis={

default,d-aug-cc-pVDZ

set,myjkfit

default,aug-cc-pVDZ/jkfit

}

which is to use auxiliary functions for aVDZ to help faciliate the calculations of d-aug-cc-pVDZ. But I doubt the fact that having no dedicated auxiliary functions for the larger d-aug-cc-pVDZ basis set will somehow result in accuracy problems.

Do you have any idea that if Molpro can actually calculate the auxiliary functions on-the-fly? Or that I can have my own fitted auxiliary functions loaded in to use them? I do need very diffuse basis functions to study some very far-stretching Rydberg states.

Best regards,

Che-Ho Yeh

27....@gmail.com 在 2026年2月25日 星期三凌晨3:49:34 [UTC+8] 的信中寫道：

[Grant Hill]

Hi,

You can specify whichever (auxiliary) basis functions you like within the set,myjkfit construction. This is done in a similar way to the explicit basis set input example in the manual:

https://www.molpro.net/manual/doku.php?id=basis_input#examples

You can also point towards a text file that contains the basis set definition. For example,

basis={

default,d-aug-cc-pVDZ

set,myjkfit

include,/path/to/basis-set.txt

}

Hope this helps,

Grant

To view this discussion, visit https://groups.google.com/d/msgid/molpro-user/82740af1-e570-49ba-8753-d397c863c4b6n%40googlegroups.com.