gonzalezma
commented
8 months ago Nothing against, but just two comments:
- Using the masses to identify the elements/isotopes is not always unambiguous, as the known problems with the Lammps converter show. It would be good to have a list of the different isotopes that have the same mass (within the typical tolerance that a user could use in the input file, I would say 0.001 or even 0.01 amu) and some way of discriminating between them when the mass alone cannot be used to choose the right element.
- I would not remove the atom transmutation. I think you can still have a single simulation and want to calculate the expected output for two different isotopes without needing to redo the simulation or change manually the input files to modify some masses.
Is your feature request related to a problem? Please describe. In DL POLY atomic masses are specified in the field file. As far as I can tell MDANSE ignores this and uses the masses it has stored internally in the atoms database. I'd imagine something similar occurring in a lot of other convertors as well. This seems like an issue since for example I may run a MD simulation with heavier isotopes by changing its mass, MDANSE could then pick the wrong isotope if it reads the output data from the atomic symbol used.
Describe the solution you'd like Perhaps the atom mapping as suggested in #334 should be done for all convertors to ensure no assumption are made. This also has the benefit that the atom_transmutation could be removed since this could be done using the mapping.