Hey!
I am looking for any program package which can read output of Gaussian program package and can perform Bader Charge analysis. If anyone has any idea about this, please share. Your suggestions would be highly appreciated.
Thank you!
~ Megha
The "Bader" code mentioned by Adam W. Pelzer is a known code, however, if you are working for molecule systems, in my point of view, it is deprecated. "bader" code may be a good choice for first-principle calculation based on plane-wave (pseudopotentials is always employed), but not ideal for quantum chemistry study of molecular systems. I would like to talk more about this point.
It is highly important to note that in most quantum chemistry calculations for molecular systems, pseudopotential is not employed for the elements in the first three or four rows, in this case the magnitude of electron density around nuclear position is very high and varies sharply, especially for relatively heavy atoms. Simply based on the evenly distributed grid of electron density that loaded from cube file, as the "bader" does, it is in principle impossible to get quite acceptable AIM charges, because the integration accuracy around nuclear position is really too poor (much denser grid must be employed than valence region).
In order to circumvent this severe problem in the uniform grid based method employed by "bader", about five years ago I proposed a new algorithm for AIM basin integration using mixed grid and implemented it in Multiwfn. This algorithm employs atom-center grid for integrating electron density around nuclear position while still uses uniform grid for valence region integration, at the meantime a smooth switch function is introduced to combine the two parts. Numerous tests showed that the AIM integration accuracy of this new algorithm is significantly superior to the algorithm employed by "bader" code. (Multiwfn also supports the original algorithm used by the "bader" code, but it is not recommended due to above mentioned reason)
In addition, although both Multiwfn and "bader" employ uniform grids for valence region, Multiwfn supports refinement of basin boundary, which can noticeably improve the accuracy of the resulting AIM charges, while "bader" code doesn't support this important feature.
BTW: In Multiwfn, the calculation of grid data is completely automatic, therefore much less steps are needed compared to using the "bader" code (for which you need to manually use main function 5 of Multiwfn or the cubegen utility in Gaussian package to yield cube file of electron density first).
Using Multiwfn to compute AIM charge is fairly easy and flexible. In general, after loading a wavefunction file, you simply need to input below commands (more details can be found in manual):
17 // Basin analysis
1 // Use electron density to partition basins (i.e. generating AIM basins)
2 // Medium quality grid
7 // Integrate real space functions in AIM basins with mixed type of grids
2 // There are three algorithms, 2 corresponds to most time-consuming but most accurate algorithm
1 // The function to be integrated is electron density
After a while, you will obtain AIM charges:
The atomic charges after normalization and atomic volumes:
1 (C ) Charge: 0.726302 Volume: 67.306 Bohr^3
2 (C ) Charge: 0.726306 Volume: 67.306 Bohr^3
3 (N ) Charge: -1.549343 Volume: 121.392 Bohr^3
4 (C ) Charge: 0.726325 Volume: 67.306 Bohr^3
...
Multiwfn can also run in command-line mode as "bader" code, thus you can easily write a shell-script to automatically calculate AIM charges for a batch of molecules, see Sections 5.2 and 5.3 of Multiwfn manual on how to do this.
It is worth to mention that Multiwfn can also calculate AIM charges purely based on a cube file of electron density, however in this case the quality of AIM charges will be identical to "bader", which is not satisfactory.
In quantum chemistry field, Bader charge is often known as AIM charge. Gaussian itself has a build-in code to perform AIM analysis, however the code is unstable and no longer under maintained since very long time ago, thus it is highly deprecated. There are several known codes that can calculate AIM charges based on wavefunction produced by Gaussian, such as Multiwfn, AIMALL, AIM2000. The Multiwfn can freely obtained from http://sobereva.com/multiwfn, for Gaussian users, produced .wfn or .fch file can be directly used as input file, the standard way of calculating AIM charge in this code is clearly illustrated in Section 4.17.1 of the manual.
Tian Lu Thank you, Sir!
I will try to do AIM Analysis using Multiwfn code.
http://theory.cm.utexas.edu/henkelman/code/bader/
This reads in the cube file and does Bader analysis. I haven't used it in a while, but Henkelman is a well respected researcher, so give it a try if you don't like the Multiwfn version.
The "Bader" code mentioned by Adam W. Pelzer is a known code, however, if you are working for molecule systems, in my point of view, it is deprecated. "bader" code may be a good choice for first-principle calculation based on plane-wave (pseudopotentials is always employed), but not ideal for quantum chemistry study of molecular systems. I would like to talk more about this point.
It is highly important to note that in most quantum chemistry calculations for molecular systems, pseudopotential is not employed for the elements in the first three or four rows, in this case the magnitude of electron density around nuclear position is very high and varies sharply, especially for relatively heavy atoms. Simply based on the evenly distributed grid of electron density that loaded from cube file, as the "bader" does, it is in principle impossible to get quite acceptable AIM charges, because the integration accuracy around nuclear position is really too poor (much denser grid must be employed than valence region).
In order to circumvent this severe problem in the uniform grid based method employed by "bader", about five years ago I proposed a new algorithm for AIM basin integration using mixed grid and implemented it in Multiwfn. This algorithm employs atom-center grid for integrating electron density around nuclear position while still uses uniform grid for valence region integration, at the meantime a smooth switch function is introduced to combine the two parts. Numerous tests showed that the AIM integration accuracy of this new algorithm is significantly superior to the algorithm employed by "bader" code. (Multiwfn also supports the original algorithm used by the "bader" code, but it is not recommended due to above mentioned reason)
In addition, although both Multiwfn and "bader" employ uniform grids for valence region, Multiwfn supports refinement of basin boundary, which can noticeably improve the accuracy of the resulting AIM charges, while "bader" code doesn't support this important feature.
BTW: In Multiwfn, the calculation of grid data is completely automatic, therefore much less steps are needed compared to using the "bader" code (for which you need to manually use main function 5 of Multiwfn or the cubegen utility in Gaussian package to yield cube file of electron density first).
Using Multiwfn to compute AIM charge is fairly easy and flexible. In general, after loading a wavefunction file, you simply need to input below commands (more details can be found in manual):
17 // Basin analysis
1 // Use electron density to partition basins (i.e. generating AIM basins)
2 // Medium quality grid
7 // Integrate real space functions in AIM basins with mixed type of grids
2 // There are three algorithms, 2 corresponds to most time-consuming but most accurate algorithm
1 // The function to be integrated is electron density
After a while, you will obtain AIM charges:
The atomic charges after normalization and atomic volumes:
1 (C ) Charge: 0.726302 Volume: 67.306 Bohr^3
2 (C ) Charge: 0.726306 Volume: 67.306 Bohr^3
3 (N ) Charge: -1.549343 Volume: 121.392 Bohr^3
4 (C ) Charge: 0.726325 Volume: 67.306 Bohr^3
...
Multiwfn can also run in command-line mode as "bader" code, thus you can easily write a shell-script to automatically calculate AIM charges for a batch of molecules, see Sections 5.2 and 5.3 of Multiwfn manual on how to do this.
It is worth to mention that Multiwfn can also calculate AIM charges purely based on a cube file of electron density, however in this case the quality of AIM charges will be identical to "bader", which is not satisfactory.
Tian Lu
Dear Professor Lu,
Too many thanks for your much valuable, informative, and comprehensive rationalizations.
Regards,
Saeed
Dear Tian Lu Sir,
Thank you very much for such a valuable description.
Regards,
Megha
- Badges
- Science topic