I am investigating a reaction (A+B -> C) through DFT implemented in Gaussian 09.
I find that E(product)-E(starting materials) computed based on 'Sum of electronic and thermal energies' or 'Sum of electronic and zero point energies' gives different outcomes.
For example, 'Sum of electronic and thermal energies' tells me starting materials (A and B combined) is more stable by ~10kCal/mol and 'Sum of Electronic and zero point energies' tells me product is more stable by 1.5 kCal/mol.
Which one should I consider?
I might be wrong, but I think this might have to do with the fact that my reactant side has 2 molecules, and hence more degrees of freedom?
All geometries were fully optimized and checked for the absence of imaginary frequencies.
However, A and B were calculated independently, i.e. not in the same frame.
Although i am pretty sure that you know how to run calculations for such system I am writing my answer in a detailed manner.
To solve this problem you have to calculate the following:
1-Enthalpic energies (H) for A, B and C.
2-Entropic energies (S) for A, B and C; you should run the calculations with frequency in order to get S.
3-Molecules A and B should be calculated as one system with only one H and one S values.
And the rest you know:
delta [G (c) - G (A, B system)] = delta [ H (c)- H (A, B system) ]-T [delta S (c)- S (A, B system)].
When you write A and B were calculated independently, did you correct for the basis set superposition error?
What I meant is that:
You optimize A and B SEPARATELY, then you place both optimized structures in a distance of at least 3 A for the closest atoms and finally you optimize the whole system (containing both A and B).
Please note that DFT is unlike simiempirical methods in which you can calculate each reactant separately.
@Andreas
I did a test run with basis set superposition error correction, and correction factors came around to ~1.6 kCal/mole in my system.
@Rafik
This is a very interesting suggestion: I have just started to look into it, and I believe this will solve my issue.
I will report updates here shortly.
Thank you both!
Dear Shrinwantu,
I suggest you repeat your calculations with a larger basis set and if this doesn't help, try either a hybrid or post-HF method.
@Ratik
I have never seen the reaction energy calculated this way. Why do you suggest A and B energies to be calculated in one single system?
Bests
Fatemeh
A good introduction to the energy terms in the Gaussian output file from frequency calculations can be found in a white paper on the Gaussian web page.
http://www.gaussian.com/g_whitepap/thermo.htm
What energy you should consider depends entirely on the question you want to answer. There's no right or wrong, but only does the choice match the problem at hand. Please note that the white paper also explains you how to go straight for the Gibbs free energy.
Please keep in mind that the models used by Gaussian strictly apply only to the gas phase. Since most experiments are done in condensed media, you might observe a mismatch between calculation and experiment. This mismatch cannot be resolved by simply choosing more sophisticated methods on the same model for reaction, but require a more refined model setup.
- Badges
- Science topic