Multiwfn forum

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,624

Website

Questions on CT analysis

Today a Multiwfn user asked me a series of question about charge-transfer analysis of electronic excitation process:

I am trying to calculate CT during electron excitation based on electron density difference according to the manual (4.18.3). I am wondering if I can use fchk files for EX & GS to do this than wfn file. I cannot find how to generate wfn file.
...
The q_CT, D_CT, and t-index in the output look like similar to the values from the D-CT calculation using Gaussian. Are they the same?
...
I am wondering if I can calculate ratio of CT and LE given certain excited state using Multiwfn. For example, the molecule has 80% of CT & 20% of LE for S1 state and 40% CT and 60% of LE for T1 state.

Below is my reply, I think the content should be also useful for other users:

Please carefully check beginning of Chapter 4, where how to generate .wfn and .fch file in proper way is detailedly described.

Using .wfn file to realize the calculation illustrated in Section 4.18.3 is staightforward. Firstly using e.g. "# B3LYP/6-31G* out=wfn" to get .wfn file for ground state, then using "# B3LYP/6-31G* TD(nstates=5,root=2) out=wfn" to get .wfn file corresponding to the 2nd excited state, then follow the procedure described in Section 4.18.3.

Also you can use .fch to carry out this analysis. In order to get .fch containing excited wavefunction, you need special step to store natural orbitals corresponding to the excited state of interest to the .fch file, as explicitly mentioned at the beginning of Chapter 4:
"To analyze wavefunction for post-HF wavefunction or TDDFT excited state wavefunction, analysis should be done using natural orbitals (NOs) at corresponding level, there are two ways to yield them:..."

Finally, it is worth to note that Multiwfn is also able to directly generate .molden file containing excited state wavefunction, please check Section 4.18.13 (note that I just updated the manual yesterday, please download the latest version of manual).

It seems that Gaussian16 is also able to generate q_CT etc., the result should be essentially the same as Multiwfn, but I never tested this feature, because I strongly believe using Multiwfn is more convenient, and Multiwfn is able to print much more useful information than Gaussian during the analysis.

Calculating ratio of CT and LE using Multwfn is fairly easy, I wrote a blog article: http://sobereva.com/398. But the article was written in Chinese. You can try to use Google translator to read it. If you fail to understand the content, I will describe basic idea of this article to you in English.

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,624

Website

Re: Questions on CT analysis

His next question:

Calculating ratio of CT and LE using Multwfn is based on the charge difference between EX & GS, right?
You suggest to use ADCH charge or Mulliken charge. Is it okay to use q_CT from the output to calculate CT/LE ratio?

My reply

The q_CT mentioned in the example of Section 4.18.3 should not be used in evaluating CT/LE ratio, because this term only reflects the amount of electrons *perturbed* during the excitation, it doesn't directly correspond to the amount of electrons *transferred* from one fragment to another fragment. Commonly, you should use variation of fragment charge (can be evaluated using various atomic charge calculation methods) to evaluate CT/LE ratio. Using the amount of CT evaluated by the IFCT analysis is also completely OK, the IFCT analysis is illustrated in Section 4.18.8.

To avoid confusion about the physical meaning of the qCT, in fact there is a paragraph in Section 3.21.3:
The transferred charge qCT is the magnitude of the integral of rho+ and rho− over the whole space.
It is important to correctly recognize the physical meaning of this quantity. qCT only corresponds to the total amount of charge whose distribution is perturbed during electron excitation, it does not correspond to net charge transfer from one fragment to another fragment (e.g. from donor group to acceptor group)