Overlap matrix format query

[Zhibo Wang]

What are the column labels for the overlap matrix? I assume that the setup is [orbital number, atom number, atom type, orbital type, matrix ...] but for the orbital type the label does not seem to be n and l. What is the number affixed before the l then?

Overlap matrix output head for CsF:

OVERLAP MATRIX

                                1            2            3            4
     1    1  Cs    2s        1.632390     2.621122     1.273099    -0.000000
     2    1  Cs    3s        2.621122     8.007919     4.342380     0.000000
     3    1  Cs    4s        1.273099     4.342380     2.948680    -0.000000
     4    1  Cs    3py      -0.000000     0.000000    -0.000000     0.966199
     5    1  Cs    3pz       0.000000     0.000000    -0.000000    -0.000000
     6    1  Cs    3px      -0.000000    -0.000000    -0.000000     0.000000
     7    1  Cs    4py      -0.000000    -0.000000     0.000000    -0.124733
     8    1  Cs    4pz      -0.000000     0.000000     0.000000    -0.000000
     9    1  Cs    4px      -0.000000     0.000000     0.000000     0.000000
    10    1  Cs    4d-2     -0.000000     0.000000    -0.000000     0.000000
    11    1  Cs    4d-1     -0.000000     0.000000    -0.000000     0.000000
    12    1  Cs    4d0       0.000000    -0.000000     0.000000    -0.000000
    13    1  Cs    4d+1      0.000000     0.000000    -0.000000     0.000000
    14    1  Cs    4d+2      0.000000    -0.000000     0.000000     0.000000

    15    2  F     2s        0.389074     0.899446     0.553087     0.000000
    16    2  F     3s        0.331211     0.719056     0.306653     0.000000
    17    2  F     3py       0.000000     0.000000     0.000000    -0.065002
    18    2  F     3pz      -0.000000     0.000000    -0.000000     0.000000
    19    2  F     3px      -0.000000    -0.000000    -0.000000    -0.000000
    20    2  F     4py       0.000000     0.000000     0.000000    -0.128646
    21    2  F     4pz      -0.000000    -0.000000    -0.000000     0.000000
    22    2  F     4px      -0.000000    -0.000000    -0.000000    -0.000000
    23    2  F     4d-2     -0.000000     0.000000    -0.000000    -0.000000
    24    2  F     4d-1      0.000000     0.000000     0.000000     0.000000
    25    2  F     4d0      -0.000000    -0.000000     0.000000     0.066629
    26    2  F     4d+1      0.000000     0.000000     0.000000    -0.000000
    27    2  F     4d+2     -0.000000    -0.000000    -0.000000     0.115404

I do not believe F has 3p orbitals and no 2p orbitals nor does Cs stop at n=4.

[Krack Matthias]

Hi

 

The labels for the principal quantum number n in the CP2K printout are rather arbitrary counters. The n value of the first shell is retrieved from the basis set definition. It has no importance for the calculation and therefore it could be in principle any value. The DZVP-SR MOLOPT basis set for Cs is

 Cs DZVP-MOLOPT-SR-GTH DZVP-MOLOPT-SR-GTH-q9

 1

 2 0 2 6 3 2 1

      1.000865707027

Try to change the first “2” in the third line to “6” to get “6s” as a first label in the printout for Cs. The n values for the following p, d, etc. orbital shells, however, might not correspond to physically correct values for the actual chemical element. The proper values depend on the employed pseudopotential, i.e. which orbitals are core and valence orbitals. Moreover, CP2K does not impose any specific order for the orbital shell sets. The printout routine does not take any effort to figure out which is the proper value for n, because it is not that easy to retrieve that information correctly for all possible cases without requesting additional input information just for the printout.

 

Best

 

Matthias

 

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/fed11feb-e8ec-4a7e-a06e-aaf18cdb165cn%40googlegroups.com.

[Zhibo Wang]

I assume this means electron counting is needed to try to guess which orbital is which then.

Addendum:

What are the units associated to the overlap matrix elements? Is it possible to extract a normalized overlap? Trying to compare extents of atomic orbital overlaps for s-s and p-p. I expect the overlap of an orbital with itself to equal 1 and I cannot find a way to compare between different orbital overlaps currently.