Intensitiywise, Most variables can be classified as withdrawing or adding. In the former group, we can find the thermal parameters, the profile coefficients and the scale factor;
Would it be a valid strategy to turn the scale factor flag off at some point to make sure positive thermal paramenters are obtained?
This is not how I would think about the fitting process. A large Debye-Waller parameter removes intensity from the high Q part of a pattern. Coupled with the idea that we do not know the scale factor and have to fit it also, Uiso values effectively shift intensity between the high and low Q sections of the pattern.
This does not create a problem, when reflections are well separated across the pattern, so that the background is well defined. However, in many cases, one wants to use data where reflections clump together (typically at high Q) so that the background location is not uniquely determined across the entire data set. In this case, one can find a range of fits that produce more or less equivalent results with a range of background, profile, Uiso and scale factor values. Smaller than expected Uiso values for all atoms commonly result from this. (This will not produce a negative Uiso for a single atom where all other values are in a reasonable range.) Under such circumstances, where the data do not allow all parameters to be fit independently, I will typically fix the background values and then note this in the paper, but I will never fix the scale factor, since that is completely arbitrary in almost all experiments.
Brian ******************************************************************** Brian H. Toby, Ph.D. office: 630-252-5488 Materials Characterization Group Leader, Advanced Photon Source 9700 S. Cass Ave, Bldg. 433/D003 work cell: 630-327-8426 Argonne National Laboratory secretary (Kristy): 630-252-5453 Argonne, IL 60439-4856 e-mail: brian dot toby at anl dot gov ********************************************************************
