Multiwfn forum

alex_benik

Member

Registered: 2021-08-18

Posts: 3

Too many basins (6) for diatomic molecule how to assign charge to atom

Dear Prof. Tian Lu,

I am doing a population analisys (Bader charges etc) with the Multwfn code for the ThAr diatomic molecule.
(The wfn-file of ThAr has been generated with GAMESS.)
And I have found that the Multwfn produces 6 attractors (basins) for it whereas usually it gives only two attractors
associated with each atom (Ar and Th).
I can supply you with my wfn input file if you want and below this letter I give some parts of the Multwfn output for this case.
My problem is that I do not understand now which charge (from which basin) I can associate with each atom (i.e. with Ar and Th).
Can you please clarify this issue from your experience?

Sincerely yours,
Alexander Nikolaev
----------------------------
some parts of output generated by Multwfn:

Coordinate of origin in X,Y,Z is -17.370000 -17.370000 -11.979452 Bohr
Coordinate of end point in X,Y,Z is 17.370000 17.370000 14.060548 Bohr
Grid spacing in X,Y,Z is 0.060000 0.060000 0.060000 Bohr
Number of points in X,Y,Z is 580 580 435 Total: 146334000
Note: All exponential functions exp(x) with x< -40.000 will be ignored

Generating basins, please wait...
Attractor X,Y,Z coordinate (Angstrom) Value
1 -3.63544770 -0.01587532 0.01087355 0.00092233
2 -0.01587532 -3.63544770 0.01087355 0.00092233
3 3.63544770 -0.01587532 0.01087355 0.00092233
4 -0.01587532 3.63544770 0.01087355 0.00092233
5 -0.01587532 -0.01587532 2.99543323 821.05606508
6 0.01587532 0.01587532 0.01087355 5938.80385853
Detecting boundary grids...
There are 1952787 grids at basin boundary
Refining basin boundary...
Generating basins took up wall clock time 25 s
The number of unassigned grids: 0
The number of grids travelled to box boundary: 2070
The number of interbasin grids: 1607978

#Basin Integral(a.u.) Volume(a.u.^3)
1 0.2465339651 7346.68358400
2 0.2465328011 7343.62567200
3 0.2465335221 7345.69948800
4 0.2465348766 7349.98514400
5 17.9579689787 1119.57400800
6 87.2914699853 740.19333600
Sum of above values: 106.23557413
Integral of the grids travelled to box boundary: 0.00000000

Integrating in trust sphere...
Warning: Unable to determine the attractor 1 belongs to which atom!
If this is a non-nuclear attractor, simply press ENTER button to continue. If you used pseudopotential and this attractor corresponds to the cluster of all maxima of its valence electron, then input the index of this atom (e.g. 9). Else you should input q to return and regenerate basins with smaller grid spacing

The trust radius of attractor 1 is 1.151 Bohr

Warning: Unable to determine the attractor 2 belongs to which atom!
If this is a non-nuclear attractor, simply press ENTER button to continue. If you used pseudopotential and this attractor corresponds to the cluster of all maxima of its valence electron, then input the index of this atom (e.g. 9). Else you should input q to return and regenerate basins with smaller grid spacing

The trust radius of attractor 2 is 1.151 Bohr

Warning: Unable to determine the attractor 3 belongs to which atom!
If this is a non-nuclear attractor, simply press ENTER button to continue. If you used pseudopotential and this attractor corresponds to the cluster of all maxima of its valence electron, then input the index of this atom (e.g. 9). Else you should input q to return and regenerate basins with smaller grid spacing

The trust radius of attractor 3 is 1.151 Bohr

Warning: Unable to determine the attractor 4 belongs to which atom!
If this is a non-nuclear attractor, simply press ENTER button to continue. If you used pseudopotential and this attractor corresponds to the cluster of all maxima of its valence electron, then input the index of this atom (e.g. 9). Else you should input q to return and regenerate basins with smaller grid spacing

The trust radius of attractor 4 is 1.151 Bohr

Attractor 5 corresponds to atom 1 (Ar)
The trust radius of attractor 5 is 2.696 Bohr

Attractor 6 corresponds to atom 2 (Th)
The trust radius of attractor 6 is 2.894 Bohr

Integration result inside trust spheres
#Sphere Integral(a.u.)
1 0.0055896817
2 0.0055896817
3 0.0055896817
4 0.0055896817
5 17.4347084614
6 87.3696959859
Sum of above values: 104.82676317

Total result:
#Basin Integral(a.u.) Vol(Bohr^3) Vol(rho>0.001)
1 0.2465282901 7346.684 0.000
2 0.2465271262 7343.626 0.000
3 0.2465278471 7345.699 0.000
4 0.2465292016 7349.985 0.000
5 18.0124284987 1119.373 206.462
6 89.0005665064 740.394 362.901
Sum of above integrals: 107.99910747
Sum of basin volumes (rho>0.001): 569.363 Bohr^3
Integral of the grids travelled to box boundary: 0.00000000

Normalization factor of the integral of electron density is 0.999992
The atomic charges after normalization and atomic volumes:
1 (NNA) Charge: -0.246530 Volume: 0.000 Bohr^3
2 (NNA) Charge: -0.246529 Volume: 0.000 Bohr^3
3 (NNA) Charge: -0.246530 Volume: 0.000 Bohr^3
4 (NNA) Charge: -0.246531 Volume: 0.000 Bohr^3
1 (Ar) Charge: -0.012577 Volume: 206.462 Bohr^3
2 (Th) Charge: 0.998698 Volume: 362.901 Bohr^3

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,768

Website

Re: Too many basins (6) for diatomic molecule how to assign charge to atom

Hello,

If you employed a large core PP for Th, please use a small core PP instead and try again; if this is not the case, please send me your GAMESS-US input file and wfn file via E-mail so that I can check carefully.

Best regards,

Tian

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,768

Website

Re: Too many basins (6) for diatomic molecule how to assign charge to atom

Dear Alexander Nikolaev,

I have received your files, please note two points:

(1) Currently you are calculating the system at MP2 level, however, the occupation numbers in the .wfn file are all integer, that means the orbitals recorded in the .wfn file is Hartree-Fock orbitals, therefore the wavefunction to be analyzed is Hartree-Fock level, which is known to be quite poor. If you insist on using MP2, you should request GAMESS-US to yield MP2 natural orbitals and record them into the .wfn file, so that Multiwfn can analyze MP2 wavefunction. (but I don't know how to do)

(2) The wfn file seems to be problematic, because if integrating electron density over whole space using subfunction 4 of main function 100 of Multiwfn, the result deviates from integer significantly. Originally, the .wfn file format doesn't support angular moment higher than f, while g functions occur in your calculation, probably this is the reason.

As shown in Section 2.5 of Multiwfn manual, output file of GAMESS-US can be directly used as input file if you manually change the file extension to .gms. I suggest you use output file of GAMESS-US and retry.