There is a generalized restraint where you could write your own function that will “push” your Frac values to be positive, but this would not be a good idea. If your Frac value(s) are negative by more than 2 sigma, then something is wrong. A value within 2 sigma of zero is experimentally indistinguishable from zero.
Usually this happens when you have site sharing and the refinement is trying to put more scattering density that 100% of the “heavy” atom or too little and the density is less than that of the “light” atom. In the “heavy” case are you sure about the site assignments? In the “light" case could you have vacancies? It is also possible that your data are insensitive to the scattering from that site and any value for the occupancy gives about the same fit. It is also possible your model is very far from correct and the model is not good enough for occupancies to be refined. It is perfectly reasonable to set the occupancy as 100% of one atom type and not refine that. Look to see if the Uiso value is reasonable. If you are refining the occupancy of for a single atom that does not share a site and the occupancy goes below zero then with x-ray data you either have a serious problem with some other part of your model or there is no scattering density in that site and you should take that atom site out. With neutrons also consider an atom type with the opposite scattering phase (negative b). In any case, a constraint would serve to hide that something is wrong with your result. A restraint would make it harder to do that, but still is not addressing the problem. Brian On Nov 19, 2019, at 10:17 AM, Hao Liu via GSAS-II <[email protected]<mailto:[email protected]>> wrote: Hi, Is there a way to constrain the atomic occupancy to be non-negative? Thanks, Hao -- Hao Liu, Ph.D. Assistant Professor Chemistry Department, Binghamton University 4400 Vestal Parkway East, NY 13902 Tel: +1 (607) 777-4671 Group website<http://chemiris.chem.binghamton.edu/LIU/liu.html> _______________________________________________ GSAS-II mailing list [email protected]<mailto:[email protected]> https://mailman.aps.anl.gov/mailman/listinfo/gsas-ii
_______________________________________________ GSAS-II mailing list [email protected] https://mailman.aps.anl.gov/mailman/listinfo/gsas-ii
