About the relation between Paritition and LnPartition for NH3

[w_wonder]

From https://github.com/iRASPA/RASPA2/blob/master/examples/Tutorial/ReactionEnsembleAmmonia/simulation.input

"

Component 2 MoleculeName NH3 

 

MoleculeDefinition Local 

 

LnPartitionFunction 253.69"


From

https://github.com/iRASPA/RASPA2/issues/4 and
https://github.com/iRASPA/RASPA2/files/3927153/RASPA.Ammonia.Synthesis.Partition.Function.pdf

Table 2,
"NH3 1.8118E+02 1.9951E+02 1.1422E+00 4020.5 1.66002E+08 1.498133E+08"

The partition function seems to be ~ 1*10^8. I calculated the partition function of NH3 using quantum chemistry software and got ~ 10^8 value.


I am unable to relate 253.69 to ~1*10^8, for example, from https://www.rapidtables.com/calc/math/Ln_Calc.html

ln(1.498133e8) gives me ~ 18.82.

May I know how the partition function and LnPartitionFunction are related? Thanks

[David Dubbeldam]

These value come from: https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.7b00092/suppl_file/ct7b00092_si_001.pdf
and to reduce under/overflow you input them as the log of these numbers in Table S3.

[w_wonder]

Thanks a lot. Sorry I am still confused  :'(

Would you provide any page number for McQuarrie8 used in Table S3 and/or calculation steps? Molecular thermal wavelength from https://en.wikipedia.org/wiki/Thermal_de_Broglie_wavelength is about 10^-11 m ~ 0.1 angstrom at 298K. q/\Lambda^3 ~ q/0.1^3 is unlikely to provide 10^100 for NH3, even varying mass and temperature.

[David Dubbeldam]

For a detailed explanation see:
https://homepage.tudelft.nl/v9k6y/thesis-Rahbari.pdf
Page 262 and following, and see Table  A5.

[w_wonder]

Thanks.

Looks like exp(D0/kT) and total atomic energy are the keys. I managed to verified ~ 20%.

[w_wonder]

Hi, one more question, in https://github.com/iRASPA/RASPA2/blob/master/examples/Tutorial/ReactionEnsembleAmmonia/simulation.input
ExternalPressure 4e7

but, for the data in Table A5 https://pure.tudelft.nl/ws/portalfiles/portal/71296377/Thesis_Ahmadreza_Rahbari.pdf
I can reproduce (~20% accuracy) with 1e5 pressure (standard) based on Gaussian output (multiplying /V and exp(beta T) terms; V = RT/p), if I use 4e7, that will lead to orders of magnitude difference.

Question 1: for the data in Table A5, what pressure was used?
Question 2: if the answer to question 1 is the standard pressure, does the partition function needs to be consistent with ExternalPressure? If the answer to question 1 is the ExternalPressure, how it was evaluated?
Thanks

[w_wonder]

These value come from: https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.7b00092/suppl_file/ct7b00092_si_001.pdf
and to reduce under/overflow you input them as the log of these numbers in Table S3.

One more question. Does RASPA have two conventions in the partition function, that previously, the exp(D0/kT) term was not included
as in https://pubs.acs.org/doi/suppl/10.1021/acs.jctc.5b00160/suppl_file/ct5b00160_si_001.pdf
later the exp(D0/kT) was included, as in
https://pure.tudelft.nl/ws/portalfiles/portal/71296377/Thesis_Ahmadreza_Rahbari.pdf

Thanks.

[David Dubbeldam]

You could email Reza Rahbari and ask him how he did it. But note there is no single convention and the results will differ depending on what assumptions and choices you make.