Hello everyone. I want to optimize a molecule using DFT in Gaussian.
It has both positive (N+) and negatively (O-) charged atoms. (The charged fragment is attached below for reference).
Could anyone please suggest me a way to optimize such molecule using DFT in Gaussian?
Also, it would be great if someone could disclose a way, if any, to verify the charge on the atoms after optimization.
Thank you.
Dear Premkumar Ganesan,
In a DFT calculation the electron density is optimized to minimize the energy of the system (so DFT provides the ground state) and you do not have control on the charge distribution.
But I think your system is a zwitterion so you will have with no problem charges in your N (+) and O (-). So just, "# opt freq method/basis" will work.
You can check it through a charge analysis. By default, Mulliken charges are shown in Gaussian, although I think they are not very reliable. As an alternative, you can ask for a NBO analysis in Gaussian (once you get the optimized structure, run a calculation with the keyword "pop=nbo", https://gaussian.com/population/). Then, in the output just search for "natural charges" (be careful, the less reliable Mulliken charges are still in the output!)
Another alternative is the Quantum Theory of Atoms In Molecules (QTAIM or AIM). In order to carry out this analysis, you have to ask for the wavefunction to gaussian (keyword output=wfx, and then specify the location of the .wfx file, https://gaussian.com/output/). With this file, you can use different softwares (e.g. AIMAll, http://aim.tkgristmill.com/index.html) to perform the AIM calculation.
Hope it helps,
Fernando
Thank you very much Fernando Aguilar-Galindo for your kind answer. I optimized the structure and did NBO analysis as well. I've few queries in this regard.
1. In the optimized structure, the oxygen which is supposed to be negatively charged is converted to C=O ( initially C-O in the input file). Is that okay?
2. After NBO analysis, the charges of the two oxygen atoms are -0.649 and -0.779. So, can it be taken that the oxygen with charge -0.779 is the one with minus charge on it [ O(-) ]? If so, why the bond is double bond here (C=O)?
3. After NBO analysis, the charge on Nitrogen atom is -0.296. What could be the reason for having a negative charge on Nitrogen though it has four bonds around it?
Dear Premkumar Ganesan,
The answer to your questions above is:
1. and 2. Because the molecule is not fully symmetric, the atomic charges on the two carboxyl oxygens are not exactly the same. However, you CANNOT say that one oxygen is negatively charged and the other is carbonyl oxygen, because the Lewis structure you draw on paper with two unidentical oxygens is only one resonance structure, which contributes only a part to the whole electronic structure. These two oxygens are the same for the real molecule.
3. The positive charge you draw on the paper is only a Formal Charge, which only reflects the historical view that N donates one of its lone pair and thus seems "positive". It is important that formal charge only helps people qualitatively understand Lewis structure and the octet rule, but does not represent any real electron distribution. N is much more electronegative than C, and thus N is always negatively charged in a NR4+, and the positive charge distributes on C or H atoms.
Thank you Yumiao Ma for your kind answers. I agree with your statement.
Since both the oxygen has double bond with carbon, the valency of the carbon becomes 5 as seen in the attached picture.
So, should the bonds be modified to dashed lines instead of one C-O and C=O in the input file?
Dear Premkumar Ganesan,
Don't worry about the bonding. GaussView judges the bonding type by distance, through a sometimes foolish algorithm, and the automatically determined bonding does not reflect the real bonding mode. Besides, the setting on bonding in a gjf file does not influence the result for a quantum chemistry calculation. In other words, it is just a graph, and you can choose single, dashed, double, triplet, ... bonds as you like. If you want to know the real nature of a certain bond, you need to analyze the bond length like what people do with an XRD result, or calculate the bond order with a program such as Multiwfn (it can also do many other analysis).
Try fragment analysis (EDA) analysis and calculate different fragments separately which may work.
- Badges
- Science topic