labels for high-L AOs defy convention

MolSSI-BSE / basis_set_exchange

A repository for quantum chemistry basis sets

https://molssi-bse.github.io/basis_set_exchange/

BSD 3-Clause "New" or "Revised" License

151 stars 44 forks source link

labels for high-L AOs defy convention #292

Closed evaleev closed 7 months ago

evaleev commented 7 months ago

the conventional labels for AOs of L=7,8,9,10,11,... are KLMNO..., whereas BSE uses JKL.... J is skipped to presumably avoid confusion with the total angular momentum operator.quantum number symbols. See any atomic term symbol table and pretty much any research article, e.g. DOI 10.1063/1.438187. Convention is more formally encoded here:

IUPAP Red Book: https://archive2.iupap.org/wp-content/uploads/2014/05/A4.pdf , page 13
IUPAC Green Book: https://iupac.org/wp-content/uploads/2019/05/IUPAC-GB3-2012-2ndPrinting-PDFsearchable.pdf , page 32

susilehtola commented 7 months ago

Not an issue, the Basis Set Exchange handles this correctly; see https://github.com/MolSSI-BSE/basis_set_exchange/blob/f8c059b9ecf34571b358b04a5303b4b4bd8c9268/basis_set_exchange/lut.py#L109C29-L109C29

The issue is that some basis set formats (Gaussian'94) are defined such that L=7 is J.

susilehtola commented 7 months ago

The solution is to use formats that explicitly encode the numerical value of the angular momentum...

addendum: this is actually supported in the (up-to-date?) Gaussian'94 format: instead of using characters for angular momentum, you can explicitly write e.g. L=7 in the basis set.

evaleev commented 7 months ago

I see, this is indeed an issue in Gaussian (and other programs), not BSE issue.