Dear Lukasz, I dont no which version of Topas do you use. Just for your information - the current version Topas 5 enables you to calculate the correlation matrix directly from GUI. Just click on "correlation matrix button" a run calculation.
Frantisek Laufek Czech Geological Survey ----- Původní zpráva ----- Odesilatel: "Peter Y. Zavalij" <pzava...@umd.edu> Příjemce: "Łukasz Kruszewski" <lkruszew...@twarda.pan.pl> Kopie: "Cline, James Dr. (Fed)" <james.cl...@nist.gov>, rietveld_l@ill.fr Datum: 2016-07-27 19:23 Předmět: Re: CrySize error in TOPAS If you switch from GUI mode to launch mode and perform refinement, you will see correlation matrix at the end of .inp file and can tell which parameters correlate a lot and figure out what's causing such large error. I also think that this error is not strange for such large cryst.size as it is basically beyond the limt of the methods. Peter .................................................................... Dr. Peter Y. Zavalij Director - X-ray Crystallographic Center University of Maryland, College Park, MD 20832 On Wed, Jul 27, 2016 at 1:07 PM, "Łukasz Kruszewski" <lkruszew...@twarda.pan.pl> wrote: Hello James. I just read the error that is reported for each refined parameter in TOPAS. I obtain value like, e.g., 9000 nm, and the error is, e.g., 205006. Best regards, Łukasz Kruszewski > By what metric are you determining that the CrySize value is in error? > > > James P. Cline > Materials Measurement Science Division > National Institute of Standards and Technology > 100 Bureau Dr. stop 8520 [ B113 / Bldg 217 ] > Gaithersburg, MD 20899-8523 USA > jcl...@nist.gov > (301) 975 5793 > FAX (301) 975 5334 > > -----Original Message----- > From: rietveld_l-requ...@ill.fr [mailto:rietveld_l-requ...@ill.fr] On > Behalf Of "Lukasz Kruszewski" > Sent: Wednesday, July 27, 2016 10:22 AM > To: rietveld_l@ill.fr > Subject: CrySize error in TOPAS > > Dera friends, > > I'm doing Rietveld refinements in TOPAS; I get rather good fit of the > reflections, rather good wt.%, but I've observed large errors for the > CrySize (Lorentzian) values for some introduced Structures. I've tried to > change it by refining CrySize (Gaussian), but it only changed in few > particular cases. I'm rather sure that the intrumental parameters > (geometry of the diffractometer was analyzed with the use of LaB6, Si, and > other standards) and corrections (sample displ. etc.) are OK and I suppose > these factors shouldn't influence (?) the CrySize values. I always > constrain the minimum (20 starting value) and maximum (10000 nm) value for > CrySize; I usually refine strain, but I've noticed that it doesn't > influence the results (at least the wt.% calculated). Rwp, goodness of fit > and Durbin-Watson statistics R values are OK; also, the calculated > background line is "flat", i.e., there is no mistaking of the background > with reflections; the shapes of the reflections in the calculated > diffractograms seem to be OK, too. I was thus wondering if these high > errors coming from the CrySize are that important for the refinement? > > Best regards, > > > -- > Łukasz Kruszewski, Ph.D., adjunct > Polish Academy of Sciences > Institute of Geological Sciences > Twarda 51/55 str. > 00-818 Warsaw > Poland > -- Łukasz Kruszewski, Ph.D., adjunct Polish Academy of Sciences Institute of Geological Sciences Twarda 51/55 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/ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 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/ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
