The most straightforward way of plotting electron density difference may be using Multiwfn (http://multiwfn.codeplex.com). The Section 4.18.3 of its manual illustrated how to plot electron density difference between excited state and ground state, the .wfn files corresponding to the two states are needed. You can also consult Sections 4.13.6, 4.5.4, 4.5.5 and 4.4.7, in which the procedure of plotting different kinds of electron density difference maps are detailedly described.

Use Gabedit, it is free.

The most straightforward way of plotting electron density difference may be using Multiwfn (http://multiwfn.codeplex.com). The Section 4.18.3 of its manual illustrated how to plot electron density difference between excited state and ground state, the .wfn files corresponding to the two states are needed. You can also consult Sections 4.13.6, 4.5.4, 4.5.5 and 4.4.7, in which the procedure of plotting different kinds of electron density difference maps are detailedly described.

PS: In the case that the transition can be perfectly discribed by a single pair of NTOs, the electron density difference can be written as the difference of square of the two NTOs. Notice that there are also many cases in which a single pair of NTOs is not enough to represent the transition, and therefore the electron density difference cannot be evaluated in this manner. However, the electron density difference can always be expressed as substracting "electron" distribution by "hole" distribution. Multiwfn is able to generate "electron" and "hole" distributions based on MOs and configuration coefficients of CIS/TDDFT calculations, see Section 3.21.1 for theoretical aspects and Section 4.18.1 for practical examples.

Thank you, Tatiana for your recommendation.

To Dr Tian Lu,

Thank you. I am reading the manual of your software. Your second comment is what i was wondering, NTO and electron density difference.

Could you please advice me briefly (to save your time) how can i obtain the EDD map between the first triplet excited state and ground state (at excited state optimized geometry) lets say.

- How to set up Gaussian job.

-Then how to run the job on your software.

(Just brieftly so i can navigate my reading and you dont have to spend much time)

Thank you very much!

Please follow the first half of the example in Section 4.18.3. You need two Gaussian input files, the first one is used to produce ground state wavefunction file (.wfn file), and the second one is used to produce .wfn file for excited state. Such two files used in Section 4.18.3 are extP2.gjf and basP2.gjf in "examples" folder, please consult them on how to write the keywords (the SCRF keyword is not necessary). As the state you are interested in is triplet, you should write "triplet" instead of "singlet" in the TD keyword. Once the two .wfn files are generated, the electron density difference can be very rapidly and conveniently plotted in Multiwfn.

Note that the two states must share the same geometry, otherwise the electron density difference will be meaningless.

If you are a Windows user, to run Multiwfn, just download the program, and double-clicking the Multiwfn.exe icon, then input the commands according to screen prompts.

I guess you could simply get the density from S0 and the density from T1 in independent gaussian runs with the same grid points, and then subtract S0 from T1 in the visualisation program. I don´t know if gaussview does it, but Vesta at least can do it.

Thank you, Hegoi. I will check the option you suggested.

To Dr Tian Lu. Thank you. I am trying to generate the two wfn file on the server at our facility. The job failed at the writing .wfn step even if the directory is specified tmp\account\...

You can send me your Gaussian input and output file to my mail box ([email protected]), I will check the reason.

Thank you, Dr Tian Lu. Bellow is my input file:

%nprocshared=8

%mem=8GB

# rb3lyp/genecp scrf=(iefpcm,solvent=dichloromethane) nosymm out=wfn

0 1

Ir -0.52942104 -0.03211912 0.04413096

C -1.45664184 1.18658154 -1.36754072

C 5.51827999 1.27720725 -1.50713238

H 5.74651011 2.04896298 -2.23849466

C 0.59652874 -0.99654186 1.52706170

N 1.51425018 0.63602031 -0.37064923

Ir 0

SDD

****

C H N 0

6-31G*

****

Ir 0

SDD

hpc\mer\EDD\ppy2quigrd.wfn

Blank line

The directory hpc\mer\EDD contains the input, output and chk files of Gaussian calculation for the job. (I try other option also, just specified ppy2quigrd.wfn in the last line but it didnt work.) The job terminated with the last line in output file was: "Writing wfn file"

Thank you for your time!

-Mer-

It seems that the path of .wfn file was not properly recognized by Gaussian or the path is invalid. Please make sure that the last several lines are:

...

 C 0.59652874 -0.99654186 1.52706170

 N 1.51425018 0.63602031 -0.37064923

[Blank line]

Ir 0

SDD

****

C H N 0

6-31G*

****

[Blank line]

Ir 0

SDD

[Blank line]

hpc\mer\EDD\ppy2quigrd.wfn

[Blank line]

[Blank line]

After altering the format, if the .wfn file is still cannot be generated normally, try to specify an absolute path (e.g. /hpc/mer/EDD/ppy2quigrd.wfn) instead of the relative path.

I tried with small organic molecule. It works and i can see the EDD map.

Thank you very much for your kind help and your time!

Yes, the real molecules were also calculated and the result were as expected. The software is more user friendly than it appears that at the first sight. Many other function are to be explored.

Thank you very much Dr Tian Lu.

-Mer-

Alternatively, you can use the cubegen and cubeman facilities of Gaussian 09 to generate two electron density cube files and then substract them to generate a new cube file containing the EDD. The resulting EDD can be easily plotted using Gaussview or any other software able to deal with this kind of file. Don't forget using the keyword NoSymm to avoid coordinate axis reorientation, since both electron densities must be referred to the same atomic coordinates system...

Yes. Thank you, Dr José. I just found out that Cubegen and cubman are very helpful in this case, especially for generating the plot.

Anyway, I still want to take this oppotunity to learn more about the Multiwfn, as it can do many other tasks, which i may need later on.

Thanks and Regards,

-Mer-

There are some points about Multiwfn I would like to mention:

1 After using main function 5 to generate the EDD grid data or visualize the EDD isosurface, the EDD grid data can also be exported as .cub file, which can be further visualized by many third-part softwares such as gview, VMD, chemcraft, gabedit, molekel...

2 Multiwfn can do everything that cubegen and cubman can do, but more powerful and easier to use. Its Main function 5 (exemplified in Section 4.5) is able to generate more than 100 important real space functions (see Sections 2.6 and 2.7), the grid can be set up quite flexibly, and the calculation speed of electron density grid data is several times faster than cubegen, mainly because Multiwfn is substantially parallelized. Main function 13 is used to manipulate and process grid data, the capacity of cubman in fact is a subset of its subfunction 11 (see Section 4.13.2 for example).

3 Not only the isosurface of EDD, but also the plane map (e.g. contour map, color-filled map, relief map...) and curve map of EDD can be very easily plotted by Multiwfn via its main function 4 and 3, respectively. Their usages are partially illustrated in Section 4.3 and 4.4 of the manual.

4 During using "custom operation" feature of main function 5 to generate EDD (as what you have done), internally, Multiwfn automatically generates electron density grid data based on the wavefunction files you provided and then get their difference. This way is more convenient than manually using cubegen (or main function 5) to generate .cub file for the two states respectively, and then using cubman (or subfunction 11 of main function 13) to get their difference.

5 "nosymm" keyword can be omitted in present case, because the two states share the same geometry, and therefore their standard orientations are the same. So, even if Gaussian may automatically rotate and translate their input coordinates, their final coordinates in the wavefunction files are still identical. Nosymm is indispensable when you would like to get EDD between complex and its fragments.

If you have any questions about Multiwfn please feel free to contact me.

Thank you very much, Dr Tian Lu, for the very useful highlight.

I would like to ask question if encounter later on. Thank you very much! I really appreciate your time and your kind help.

Best regards,

-Mer-

Can any one of you provide any reference to know the significance of  electron density difference plots?

Take a look: http://onlinelibrary.wiley.com/doi/10.1002/anie.197800801/abstract

https://www.physicsforums.com/threads/electron-density-vs-electron-density-difference.219407/

Thank you so much Muhammad Ali Askari,

Actually I went through those papers and they are quite old. And not related to my work directly.

What I am looking for that the electron density difference plots and their significance in single electron excitation and to predict electron transfer phenomenon between ground and excited states.

If you have any recent related to this issue please let me know.

I tried but could not get any information even I used Scifinder also.

Thank you in advance.

Thank you all for useful discussion,

Here I want to ask how to use cubegen facilitie for generating cube file for visualizing MEP, I am not getting from manual. I use Gaussian software in linux terminal not in window. Can we make it GUI by command.

Shilendra

Dear all,

should I optimize excited state (e.g. 1st) geometry first to get EDD between relaxed S1 and S0? The question arised because my S1 (and Sn so) is dissociative state.

Anton Kuzmin In principle, if you want to study the EDD corresponding to emission process, you have to firstly optimize the excited state under study.

Dear Tian Lu,

thank you for your reply!

A good overview on density difference between ground and excited states

https://chemistry.stackexchange.com/questions/75867/is-there-a-relation-between-transition-density-and-density-differences

A practical approach for visualization

1. Generate formatted check point file from your CIS, TDDFT, ...

2. Using GUI (GaussView, Molden, Gabedit, etc.) generate cubes for total SCF and CI densities

3. Calculate density difference = CI -SCF

4. Plot density difference by itself on on top of SCF density as a mapped surface