I want to calculate solvatisation effects on some molecules with the Polarizable Continuum Model (PCM) with Gaussian09. There are two possible ways to do that:
Way A) I first run optimization and frequency jobs for the isolated molecule in the gas phase. Then calculation of SCRF=PCM with the optimized geometry.
Way B) I do optimization and frequency calculation with the SCRF=PCM option.
Comparing both methods I see a difference in energy between both methods for the same molecule (about 4 kJ/mol).
What is the correct way to do these calculations?
Hi Uwe,
The options calculated different things, therefore there is no 'right' way, the queston is what it is that you want to know.
The first option you describe will give you the the difference in energy for the gas phase optimized structure, when referenced to energy of the gas phase optimized structure without the PCM. Here, you should be aware of the fact that the gas phase structure you optimized is only a minimum on the potential energy surface of in the gas phase. As soon as you use the PCM the potential energy surface changes so to speak and you can not be sure that you are still at a minimum. You compare energies for the same geometry (while this geometry is probably not a minium when using the PCM). If you check the energies you should see that the second approach should give the lower energy indicating that you are not at a minimum.
Your second approach described above would compare a minimum energy structure in the gas phase with a minimum energy solvated structure. (You should be able to see differences in the geometries)
The conclusion might be:
If you are sure that the gas phase geometry is right (for what ever reason) then you will probably want to use the first approach as it gives you the effect on that specific geometry.
If you do not know what your geometry looks like you would probably want to compare two minima with each other. (If you say you want to study the effect of solvation, the change in the geometry is one effect)
I hope this helped,
Jan
Hi Uwe,
The options calculated different things, therefore there is no 'right' way, the queston is what it is that you want to know.
The first option you describe will give you the the difference in energy for the gas phase optimized structure, when referenced to energy of the gas phase optimized structure without the PCM. Here, you should be aware of the fact that the gas phase structure you optimized is only a minimum on the potential energy surface of in the gas phase. As soon as you use the PCM the potential energy surface changes so to speak and you can not be sure that you are still at a minimum. You compare energies for the same geometry (while this geometry is probably not a minium when using the PCM). If you check the energies you should see that the second approach should give the lower energy indicating that you are not at a minimum.
Your second approach described above would compare a minimum energy structure in the gas phase with a minimum energy solvated structure. (You should be able to see differences in the geometries)
The conclusion might be:
If you are sure that the gas phase geometry is right (for what ever reason) then you will probably want to use the first approach as it gives you the effect on that specific geometry.
If you do not know what your geometry looks like you would probably want to compare two minima with each other. (If you say you want to study the effect of solvation, the change in the geometry is one effect)
I hope this helped,
Jan
Dear Uwe Böhme,
Usually, solvation effects should be carried out with the optimized geometry. So optimize the molecule first and then run the calculation with SCRF=PCM. If you are using more than multiple job options, Gaussian 09 gives priority depend on the memory and iterative steps. So this is the correct option always to the best of my knowledge.
Dear Jan, Jason, and T. Karthick,
thank you for answering my question.
The explanation of Jan sounds reasonable.
Therefore I will procede after method B, since I want to know the energy changes in solution of an organic molecule in dependence on different conformations.
Thank you all, best greetings
Uwe Böhme.
Hi Uwe,
You have already got very good answers. Just to add my own bit: if you are interested in the Delta G of solvation, you can use the keyword SCRF=SMD. The procedure is recommended by Gaussian and corresponds to Truhlar’s and coworkers solvation model (http://www.gaussian.com/g_tech/g_ur/k_scrf.htm). Then, a separate gas phase calculation is required. Alternatively the DoVacuum keyword can be used in the same route card, SCRF=(SMD,DoVacuum), to compute the energy difference between gas phase and solvated models.
With the keyword scrf=pcm. You can add additional keywods, but depending on the problem the better option is to look into
http://www.gaussian.com/g_tech/g_ur/k_scrf.htm
- Badges
- Science topic