Multiwfn forum

mbergami

Member

Registered: 2020-01-12

Posts: 11

Electron excitation analysis of a very large wavefunction

Hi all,

I am trying to perform excitation analysis only for the first excited state of a system with 561 atoms. I employed TD-DFT/6-311G(d,p) with LC-wHPBE using Gaussian 16. Now I have the .log and .fchk of this wavefunction. Moreover, I have 300 wavefunctions of different geometries of this system to analyse in the same way, and I need to accelerate the excitation analysis.

I want to perform the analysis with the function “18 Electron excitation analysis” and “1 Analyze and visualize hole&electron distribution, transition density, and transition electric/magnetic dipole moment density” to get fragment contributions using “2 Hirshfeld partition”.

The system and wavefunction are very large, so it’s too slow. Is there any way to accelerate? Is there an alternative using Multiwfn's other functionalities to perform this analysis faster?

In principle, I need the fragments' contributions for hole and electron only of S1, hole and electron overlap, and, if possible, also perform an IFCT analysis to determine the CT or LE character of this excited state.

Any help will be very welcome!

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,144

Website

Re: Electron excitation analysis of a very large wavefunction

1 If you haven't set "nthreads" in settings.ini of Multiwfn, don't forget to set it to actual number of CPU physical cores.

2 Using better CPU will make the calculation faster.

3 Slightly increasing grid spacing (equivalent to reducing number of grids) of hole-electron analysis will reduce computational cost.

4 In the case of IFCT analysis with Hirshfeld partition, you can reduce number of atom-centered integration grids to reduce cost, the method is lowering "radpot" and "sphpot" in settings.ini before booting up Multiwfn.

5 Using IOp(9/40=3) instead of IOp(9/40=4) in Gaussian TDDFT calculation will reduce cost of hole-electron and IFCT analyses.

6 Using a smaller basis set such as 6-31G* for unimportant atoms.

Note that 3,4,5,6 will lower the accuracy.