Hi Grayson, I see you are not starting with an easy system to analyze by SAXS :-) The first question in any scattering analysis would be: “What information are you seeking?” Not having that here are some general thoughts that may be useful. Let me ignore for the moment that you expect your fibrils to be formed of spherical particles and focus on the large object for a minute as this is (I think) the easier one to deal with. As long as you are at low enough Q that the spheres are “too small to see” this should be a reasonable zeroth order approximation I think. For this a rectangular prism (or also the model called parallelepiped in SasView) would be appropriate. However, the length of microns is way beyond the length resolvable by SAXS so just fixing that parameter to something you feel is about right will be necessary (do not try to fit it! The data will not have enough information). This leaves the width and height to fit. The second caveat here will of course be the concentration. If they are concentrated enough (and at microns in length that doesn’t have to be much) then you should see an isotropic to nematic transition in your system? At this point things get harder. However one can estimate the cross section from the position of the peak and the known volume fraction (assuming it is in fact known) and try to fit the model to data at Q higher than the peak to get height and width. Of course the problem with that last bit is that you say your prism is formed of spheres and that you have a crystal structure? My first question would be: what do you mean by crystal structure? If the spheres are reasonably monodisperse and are packing into a close packed structure with the center to center distance being the diameter,D, of the sphere, then I would assume that at Q=2*PI/D there will be a sharp peak? If so the analysis above becomes difficult and one with which I have had little experience. I would need to see a data set to have any more ideas. In principle one can use the modulation of the bragg peak heights to infer something about the underlying form factor but in this case all the form factor information of interest is at lower Q and I would guess suppressed for at least some range of Q. That goes a bit beyond my experience but hope it helps. Maybe someone else has more experience with this? Cheers Paul On Thu, May 7, 2020 at 2:53 PM Grayson Johnson via users < users@lists.sasview.org> wrote:
Dear Users Forum,
I’m new to the process of fitting SAXS data, and I am looking for guidance on what assumptions to make for my models. I have a set of small, probably spherical particles that assemble into micron length rectangular prism shaped structures. I have a known crystal structure from TEM images. Is it better to model the SAXS profile as a parallelepiped (and presumably miss out on the structure factor information that is present) or as spheres? If I do as spheres, is there a good way to use the crystal structure information to help the model along? When I perform a fit with spheres, I don’t seem to get the really low q behavior that I get with the parallelepiped. When I fit with the parallelepiped, the fit takes upwards of 30 minutes, and I haven’t quite been able to dial in on the parameters. The assemblies have a lot of variation in length, but not too much in width and height.
Thanks for any help that you can provide,
Grayson Johnson _______________________________________________ users mailing list users@lists.sasview.org http://lists.sasview.org/cgi-bin/mailman/listinfo/users