I agree with everything Dave said. I was a little terse in my response. Physically, altering a unit cell's contents is going to alter its size and the intensities.
On Thu, 7 Nov 2024, 20:49 Bish, David L, <b...@iu.edu> wrote: > Matthew has given a nice summary for you, but I'd like to add one thing. > Whenever the site occupancies change, the unit-cell parameters will/should > change (unless the substituting ions have identical radii). You can > actually use the refined unit-cell parameters to estimate compositions in > many cases. You should always refine the uc parameters (with the exception > of a certified standard), and if your materials' compositions change, you > can expect the uc parameters to vary. Even if they change by only a small > amount, they can have a significant effect on the fit. I would not fix the > uc parameters in a refinement unless you are working with materials of a > known, fixed composition, e.g., synthetic, pure materials. Having said > this, if the substitutions are very minor, e.g., ppm level, you will likely > not see any effect. But, as you said, reflection positions and thus uc > parameters vary even for the same materials listed in the PDF and the COD. > > Also, keep in mind that the extent of the effect of site occupancy > variations on intensities will depend on the scattering factors of the > substituting ions. Si-Al substitutions, for example, in minerals such as > feldspars will have a negligible effect on intensities but a larger effect > on uc parameters. On the other hand, Mg-Fe substitution in the olivine > minerals will affect both intensities and uc parameters, and you can use > either or both of these to estimate composition. > > Best of luck with your refinements. > > Dave > ------------------------------ > *From:* rietveld_l-requ...@ill.fr <rietveld_l-requ...@ill.fr> on behalf > of Matthew Rowles <rowle...@gmail.com> > *Sent:* Wednesday, November 6, 2024 11:03 PM > *To:* Łukasz Kruszewski <lkruszew...@twarda.pan.pl> > *Cc:* Rietveld <rietveld_l@ill.fr> > *Subject:* [External] Re: some uncertainties regarding the Rietveld > refinement results in TOPAS > > > This message was sent from a non-IU address. Please exercise caution when > clicking links or opening attachments from external sources. > > Great questions! > > > does the Absorption correction also influences microabsorption? > > No. The absorption correction in TOPAS is only a peak-shape correction to > account for the transparency of the specimen. There is no real way to > correct for microabsorption beyond grinding to a fine particle size, > changing your wavelength, or moving to neutrons. > > > > Is it worthy to use the parameter known as "Sample Thickness", > associated with the Absorption Correction in TOPAS? > > Generally, no. You only need to worry about it if the penetration depth of > your incident beam is larger than the thickness of the specimen. If that is > the case, then "Sample Thickness" and "Scale Intensity" is necessary to > correct for the change in peak shape due to the finite depth, and the > change in diffracting volume. > > > >Is there any difference, to be expected, in the calculated unit cell > parameters, between approximating a phase using (1) a chemically pure > STRUCTURE model, and (2) a STRUCTURE with Occupancy information updated > with the known content of impurities? > > Altering site occupancies should only affect peak intensities, and so unit > cell parameters should only be a second-order effect. (disregarding the > physicality of various unit cell contents vs dimension pairings) > > > > Is the negative intensity in the difference curve... > > It is showing that there is some mismatch between your model and the data. > The intensities could be off because the peaks positions are wrong, the > peak-shape is wrong, the atomic displacement parameters are wrong (don't > refine these from normal QPA data), the LP factor correction is wrong, the > site occupancies are wrong, the atomic scattering factors are wrong. Could > also be that you have forgotten to refine a parameter that you should be. > > > > We have been working using few computers with, supposedly the very same > version of the software > > That does seem concerning. There can be differences between different > versions of TOPAS. You would also need to verify that all the scattering > factors, wavelengths, MACs, and space group files are the same for all > installations, and that nothing has been incorrectly edited. Also that > TOPAS.inc hasn't been touched. > > > Hope this helps. > > > Matthew > > > > > > > > > > > On Wed, 6 Nov 2024 at 19:12, Łukasz Kruszewski <lkruszew...@twarda.pan.pl> > wrote: > > Dear Rietvelders, > > after some years of being a happy user of TOPAS & Rietveld method (after > working-out a good approach, thanks to a great help from users of this > mailing list for which I am really thanful, and having it checked by > attendance in Reynolds Cup) I was recently confronted with few questions > I couldn't answer. I thus have a kind request of a help here: > > - does the Absorption correction also influences microabsorption? This > question arose due to somewhat lowered expected wt.% of pyrite (19 > instead of 25), FeS2, even though all the used parameters (Absorption, > CrySize, Strain - for all the sample components) were checked and proven > to give value > error, i.e., being physically meaningful for the > refinement model. I was wondering how this supposed non-precision for > pyrite could be addressed... > > - Is it worthy to use the parameter known as "Sample Thickness", > associated with the Absorption Correction in TOPAS? I was taught in a > 5-day course of TOPAS that it's not a good idea, especially with the > Bragg-Brentano geometry, but I'm not sure... > > - Is there any difference, to be expected, in the calculated unit cell > parameters, between approximating a phase using (1) a chemically pure > STRUCTURE model (i.e., taken directly from a database / CIF file), and > (2) a STRUCTURE with Occupancy information updated with the known > content of impurities? I always thought that it shouldn't matter, as the > software simply "iteratively" compares the sample' reflections with the > structural information contained in STRUCTURE models (?). I have > compared these two situations, and option "2" actually gave me worse > fit. I thus wondered: does it really matter that much which model > (approximation) of the structure - chemically pure or with added > impurity info - is used? For most substances, the reflection positions > vary even for different standards found in databases like PDF or COD... > > - Is the negative intensity in the difference curve only informing us of > underestimation of the content of a particular phase, suggesting to > replace the Structure model used with another one? > > We have been working using few computers with, supposedly the very same > version of the software. However, even though we've been using exactly > the same Structure models, results were varying depending on the > computer stand used. I was wondering if this could have to do with some > errors in the JAVA script, or maybe a non-identical installation > conditions, or something about the Windows system? > > I would be thankful for any tips. > > Best regards, > > -- > Łukasz Kruszewski, Ph.D., Associate Professor > Polish Academy of Sciences > Institute of Geological Sciences > Twarda 51/55 > <https://www.google.com/maps/search/Twarda+51%2F55?entry=gmail&source=g> > str. > 00-818 Warsaw > Poland > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > Please do NOT attach files to the whole list <alan.he...@neutronoptics.com > > > Send commands to <lists...@ill.fr> eg: HELP as the subject with no body > text > The Rietveld_L list archive is on > http://www.mail-archive.com/rietveld_l@ill.fr/ > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > >
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Please do NOT attach files to the whole list <alan.he...@neutronoptics.com> Send commands to <lists...@ill.fr> eg: HELP as the subject with no body text The Rietveld_L list archive is on http://www.mail-archive.com/rietveld_l@ill.fr/ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
