Hello everyone, I am performing a BSSE (Basis Set Superposition Error) calculation using the Counterpoise=2 option to calculate the hydrogen bonding energy between two small molecules with gaussian 16.
For this purpose, I first performed Opt and Freq calculations using the following input:
# OPT(MaxCycle=200) Freq MP2/aug-cc-pVDZ NoSymm pop=npa geom=AllCheckpoint
Then, I conducted the BSSE calculation with the following input:
# SP MP2/aug-cc-pVDZ Counterpoise=2 NoSymm pop=npa
I have two questions regarding this process:
1. Is it generally correct to include the Counterpoise=2 option only in the SP calculation, as shown above? Should the BSSE option also be included in the Opt and Freq calculations?
2. I would like to obtain the partial charge information of the atoms from the BSSE calculation results. However, while the BSSE calculation converges, an error occurs during the NPA (Natural Population Analysis) process. Is it not possible to use the pop=npa option in BSSE calculations? If so, are there any alternative methods to obtain the partial charges aside from NPA?
If anyone has answers to these two questions, I would greatly appreciate your input.
Thank you very much!
Hallo, Kim!
1. You can use the keyword Counterpoise=2 with Opt and Freq. In such case the BSSE corrected energy is minimized.
2. The BSSE correction is related to the energy only and not to the el. structure. The population analysis can be done without BSSE correction only for the BSSE optimized structure.
Best wishes
mb
If you don’t include the counterpoise=2 in your opt and freq calculations, your end structure and frequencies will be affected by BSSE. At this point, using the counterpoise keyword will only give you the correction for a bsse-contaminated structure.
Note also that if you do want to determine as accurate a binding energy as possible, you will also need to do two further calculations: opt and freq on each of the individual fragments in the monomer basis sets to determine any distortion energies and zero-point correction. (In my experience, however, this is generally pretty minimal for hydrogen bonding, so if you're just after a ballpark figure then you might omit this step.)
I’m afraid i can‘t comment on your second question, but you should have a basic Mulliken analysis in any Gaussian calculation and, as noted by Martin Breza, I don’t see that adding the counterpoise=2 keyword for an NPA analysis would make any difference.