Multiwfn forum

Hello Prof Tian,
I have run a regular PBE and PBE SCNL calculation with ORCA and obtained the .wfx files from orca_2aim. Now, with the wfx files, I have run AIMAll and found that the AIMAll obtained energies are similar whereas energies obtained from ORCA are very different.
Here is the ORCA input with SCNL

! PBE SCNL VERYTIGHTSCF NORI AIM KEEPDENS
%maxcore 8000
%pal
        nprocs 8
end
%method
        IntAcc 7
end
%basis
        Basis "aug-cc-pVTZ"
end

* xyzfile 0 1 mp2_opt_AVTZ.xyz

With energy = -381.7650100511 Hartree

And ran PBE calculation with the same input as above without "SCNL".
With energy = -381.952393605 Hartree

Now, with the obtained .wfx Files, I had to add
<Model>
PBE
</Model>

Which I did to both SCNL and non-SCNL .wfx files and ran AIMAll.
AIMAll gives total energy as,
PBE-SCNL = -381.952350
PBE = -381.952365
And both are relatively same.

So, does the SCNL keyword change the density? or it is the problem with AIMAll? Such that providing any .wfx files (even produced by M06) with <Model> as PBE will also give the same energy?

Thank you very much
Prasanta

Dear Tian,
I was looking at this paper, by Medvedev et al (Science 355, 49-52 (2017)) where they have compared the differences of rho, gradient and laplacian of rho for many atoms between CCSD(full), MP2, MP3 etc with many xc functionals. I was wondering if I can test the accuracy of some of the functionals suggested by them (B3LYP, PBE0 etc) by means of rho, grad and lap of rho for some molecules. I think I can calculate those values with Multiwfn, right?

Do I need any particular software, as ORCA, Gaussian for them.
How large integration grids in multiwfn is necessary?

Thank you very much
Prasanta

Dear Prof Tian,
I have performed vertical excitation using ORCA (default setting) with RSH/def2-TZVP.
The output is this, (only printing interesting texts)
STATE  1:  E=   0.085477 au      2.326 eV    18760.0 cm**-1 <S**2> =   0.000000
   139a -> 141a  :     0.380347 (c= -0.61672301)
   139a -> 142a  :     0.021636 (c= -0.14709245)
   139a -> 144a  :     0.029537 (c=  0.17186373)
   140a -> 141a  :     0.493712 (c= -0.70264618)
   140a -> 142a  :     0.021983 (c= -0.14826775)
   140a -> 144a  :     0.030323 (c=  0.17413609)

STATE  2:  E=   0.085954 au      2.339 eV    18864.8 cm**-1 <S**2> =   0.000000
   139a -> 141a  :     0.493490 (c=  0.70248856)
   139a -> 142a  :     0.022278 (c= -0.14925782)
   139a -> 144a  :     0.027986 (c=  0.16728926)
   140a -> 141a  :     0.380403 (c= -0.61676836)
   140a -> 142a  :     0.021947 (c=  0.14814422)
   140a -> 144a  :     0.027699 (c= -0.16642982)

STATE  3:  E=   0.132745 au      3.612 eV    29134.1 cm**-1 <S**2> =   0.000000
   139a -> 142a  :     0.123433 (c=  0.35133024)
   139a -> 143a  :     0.221706 (c= -0.47085677)
   139a -> 144a  :     0.117215 (c=  0.34236621)
   139a -> 146a  :     0.039887 (c= -0.19971670)
   140a -> 142a  :     0.072587 (c= -0.26941893)
   140a -> 143a  :     0.269701 (c=  0.51932752)
   140a -> 144a  :     0.061824 (c= -0.24864433)
   140a -> 146a  :     0.039400 (c=  0.19849355)

STATE  4:  E=   0.133335 au      3.628 eV    29263.6 cm**-1 <S**2> =   0.000000
   137a -> 145a  :     0.011128 (c=  0.10548751)
   139a -> 142a  :     0.086160 (c=  0.29353050)
   139a -> 143a  :     0.212033 (c=  0.46047060)
   139a -> 144a  :     0.140644 (c=  0.37502547)
   140a -> 142a  :     0.142376 (c=  0.37732800)
   140a -> 143a  :     0.156057 (c=  0.39504032)
   140a -> 144a  :     0.211858 (c=  0.46028051)

When I try to calculate the transition probabilities, it doesn't give back 100%. The system is a closed shell. Any suggestion? or can it be done in Multiwfn? I tried to plot the Uv-vis spectra and see the transitions by this.
Multiwfn sp.out > 11 > 3 > 15 > 0 > 533.04910
What I get,
Sum of absolute values of all transitions:            41.19493
The individual terms are ranked by magnitude of contribution:
   #Transition     Contribution      %
         2            36.69845     89.085
         1             4.34905     10.557
         3             0.12731      0.309
         7             0.01532      0.037
         4             0.00242      0.006
         8             0.00215      0.005
         6             0.00015      0.000
         5             0.00007      0.000
         9             0.00000      0.000
        10             0.00000      0.000

How does it calculate the transition percentage?

Hi there,
I was wondering if there is a way to export MO cube files using Multiwfn.
We can visualize the HOMO, LUMO etc, but can we export the cube files?

Hi Prof Tian,
For testing purposes, I have calculated a single-point (Hartree Fock) calculation on H2 molecule with ORCA and PSI4, followed by reading the .gbw and .fchk file by Multiwfn and exporting electron density to a .txt file with the high grid.
The steps I have taken are as follows.
Multwfn H2.gbw(or H2.fchk) > 5 > 1 > 3 >3

To integrate the electron density, I wrote a small python code which shows the total value of density as 0.28, which should be 2 (for H2). There can be some problems with my program, but when I cross-checked the data, I observed that the 4th column (density value) has only one maxima. Maxima value I obtained is 0.429159635 (from my program) and by using grep "4.291596" output.txt I got only one value "4.29159635E-001".

For H2 molecule, there should be two equal maximum values of density, right?

What could get wrong, any idea?

Hi Prof Tian,
I want to partition the total molecular density into small fragment densities based on Hirshfeld method. After that, I need those densities to be outputted as four column .txt file.
Can it be done in Multiwfn?

Thanks and Regards,
Prasanta

Hi there, I tried to plot the ESP contour diagram of a molecule. It is not fitted inside the default print area. Let me show by this image.

Are there an method by which I can fit the image inside the rectangle? Or can I increase the size of the rectangle?