Closed OJ-0908 closed 6 months ago
In order to compute rate coefficients and thermochemistry at a particular level of theory Arkane usually wants frequency scaling factors (for both) and atom energy corrections and bond additivity corrections for thermochemistry. Strictly however of these it only requires the atom energy corrections, without which most resulting calculations would be useless.
Likely RMG does not have corrections for your level of theory in the RMG-database (although it's worth checking if it has it and the way you specified the theory in the input file is causing RMG to misinterpret the level of theory). If the RMG-database does not have the corrections. You have three primary options: 1) There are IPython notebooks in RMG-database/scripts for automating computation of atom energy corrections (AECs) and bond additivity corrections (BACs). Although, they will require installation of ARC. 2) Find AECs, frequency scaling factors and ideally BACs in literature and add them to RMG-database/input/quantum_corrections 3) Manually run a set of calculations and fit at least AECs to them and then add the result to RMG-database/input/quantum_corrections
@mjohnson541 thanks for the very thorough set of next steps! I will add that the log has this warning before the errors
Warning: No frequency scaling factor found for LevelOfTheory(method='m062x',basis='6311++g(2df,2p)',software='gaussian'). Assuming a value of unity. ...
Should have Arkane still have successfully executed here, assuming some default value for the corrections (however erroneous it may be)?
So in general for computing properties:
Atom Energy Corrections | Frequency scaling factors | Bond Additivity Corrections | |
---|---|---|---|
Thermochemistry | Necessary | Improves Accuracy | Improves Accuracy |
Rate Coefficients | No Impact | Improves Accuracy | No Impact |
In abstract, the primary function of the atom energy corrections is to take the quantum chemistry energies and reference them to the standard state. So unlike frequency scaling and bond additivity corrections which only refine the numbers to better match experiments the atom energy corrections can be very large. One would be unlikely to get a sane thermochemistry computation without them. See https://doi.org/10.1002/kin.21637.
However, I think if a user were to compute all species of interest at the exact same level of theory the relative thermochemistry for all of those species would be correct without atom energy corrections. So one could imagine allowing a user to deliberately turn off atom energy corrections, but in such a system you could never mix data because the references would be different.
Yeah, in hindsight atom energy "correction" may not have been the best choice of term from the kinetics community.
@mjohnson541 ,@Jackson Burns, thank you very much for your response. Following mjohnson541's suggestion, we have successfully incorporated the selected Gaussian method's BACs into the program, allowing it to run smoothly and generate reaction rate curves.
General area which your question is related to.
[√ ] Arkane (formerly CanTherm)
Question
Dear RMG Team, cal rate.zip
I am writing to seek your guidance regarding an issue I encountered while attempting to use Arkane to calculate reaction rates. Despite following the standard procedures, I encountered an error during the calculation.I have attached the Gaussian frequency and single point energy calculations for the reactants, products, and transition state. However, I am unable to identify the cause of the error.I would greatly appreciate any assistance or suggestions you can provide to resolve this issue.Thank you very much for your time and support.