extract localized orbitals


Clicked A Few Times
Dear NwChem Users,

I would like to use the localization scheme from NwChem and extract localized orbitals [LO] (occupied + virtuals) for further processing in a different code. Nwchem by default doesn't print the localized orbitals in the output file. Here are my questions:

1. Is there a switch enabling the printing of the localized orbitals to the output file?

2. In one of the previous posts I've read that the localized orbitals are printed into the .movecs file. Is is possible to extract the coefficients from there? If so how could I do this? Are the LO's given expanded over AO's?

3. If its troublesome to extract the LO's from the .movecs file could you point me to the place in the code where I could print the LO's to the output or separate file in ASCII format? I've checked the /src/ddscf/localize.F file but couldn't find the place where the orbitals are printed to the .movecs file.

Cheers,
Lukasz

Forum Vet
1. There is no switch to print the orbitals to the output file, and even if it does it would not have the needed accuracy.

2. The localized molecular orbitals are written to the .movecs file, and the coefficients can be extracted from this file. These are regular MOs so yes they are expanded into AOs. Key is to make sure you have the same basis function (AO) ordering (for example make sure dxx,dxy,dxz..etc. ordering is the same) and normalization in both codes. We have a small code in contrib/mov2asc which will convert the binary .movecs to an ascii file you can use.

3. MOs are written in scf_movecs_write

Clicked A Few Times
Thanks
Thanks Bert. The mov2asc program was precisely what I needed.

Lukasz

Clicked A Few Times
I am trying to check the ordering of the AO in NWCHEM.
I am using spherical angular functions, which I imagine are real AOs.
This is confirmed for the p orbitals which are given with the standard notation px, py, pz.
But, this is not the case for higher angular orbitals (they seem to be given simplify by m = -L .. L).
Here is an example:

   Basis function labels
---------------------
No. Atm El Type
--- --- -- ----
1 1 H s
2 1 H s
3 1 H s
4 1 H px
5 1 H py
6 1 H pz
7 1 H px
8 1 H py
9 1 H pz
10 1 H d -2
11 1 H d -1
12 1 H d 0
13 1 H d 1
14 1 H d 2

This is probably a dumb question, but what does these correspond to in terms of the standard notation for real orbitals (d_{xz}, d_{yz}...)? Should I assume that d-2 = d_{xy}, d 2 = d_{x^2-y^2}, etc?

Forum Vet
Sphericals
That's an interesting question. We do not print this info. From what I can see, it's

d-2 = d(xy)
d-1 = d(yz)
d0 = d(2z^2-x^2-y^2)
d1 = xz
d2 = d(x^2-y^2)

There is a piece of code in src/NWints/auxil that can be compiled and when run will print out the linear combinations for each lm. Try "make spherical". It should build a code called spherical that give you want you want.

Bert

Clicked A Few Times
spherical


Forum >> NWChem's corner >> NWChem functionality