skoswas
I am fitting in sequential mode (Rietveld and PDF simultaneously) a series of thirty-one diffraction
patterns collected at Diamond. Everything seems to go fine,
but in three data sets the R_Bragg indicator jumps to over 99 for no apparent
reason. The fitted parameters vary in a smooth fashion for the whole series,
even for the patterns that display this strange behaviour. Could anyone find
an explanation for this, or spot what I am doing wrong?
I am attaching a list of the most relevant parameters.
Thanks in advance.
Xabier Turrillas (ICMAB-CSIC)
Temp(K) Rwp_riet Rwp_PDF R_Brag a err c err xy1 err zy2 err xf1 err zf1 err xf2 err zf2 err xf3 err Y_biso err K_biso err F_biso err
104.6 6.68065 29.65662 2.37529 8.15456 0.00166 11.49200 0.00454 0.26397 0.00029 0.25364 0.00067 0.22258 0.00097 0.37732 0.00115 0.16681 0.00075 0.15832 0.00150 0.33362 0.00149 0.27743 0.02789 0.19779 0.12262 0.05013 0.07451
158.3 6.53591 28.12024 2.26936 8.15614 0.00164 11.49656 0.00447 0.26384 0.00029 0.25330 0.00065 0.22318 0.00097 0.37713 0.00116 0.16680 0.00074 0.15811 0.00148 0.33361 0.00147 0.24534 0.02691 0.54997 0.13071 0.04463 0.07321
210.7 6.56947 27.68444 2.29121 8.15677 0.00195 11.50711 0.00531 0.26390 0.00029 0.25320 0.00066 0.22340 0.00101 0.37732 0.00120 0.16673 0.00077 0.15821 0.00156 0.33345 0.00155 0.27142 0.02720 0.73892 0.13741 0.10457 0.07422
262.2 6.65546 27.65978 2.36595 8.16273 0.00181 11.50712 0.00491 0.26389 0.00029 0.25323 0.00068 0.22329 0.00105 0.37780 0.00124 0.16679 0.00076 0.15814 0.00154 0.33358 0.00153 0.31547 0.02810 0.85696 0.14193 0.13084 0.07466
312.9 6.53701 27.03154 2.31474 8.17172 0.00165 11.51066 0.00452 0.26369 0.00029 0.25306 0.00068 0.22325 0.00107 0.37793 0.00126 0.16678 0.00074 0.15787 0.00150 0.33356 0.00149 0.37914 0.02901 1.14370 0.14931 0.19785 0.07591
362.5 6.75107 26.97486 2.49434 8.17714 0.00163 11.51745 0.00445 0.26385 0.00030 0.25312 0.00070 0.22319 0.00113 0.37845 0.00133 0.16676 0.00076 0.15777 0.00154 0.33352 0.00152 0.42976 0.03022 1.19773 0.15364 0.24784 0.07732
411.3 6.69419 26.70483 2.48172 8.18230 0.00161 11.52071 0.00441 0.26371 0.00030 0.25304 0.00071 0.22322 0.00117 0.37875 0.00137 0.16670 0.00076 0.15761 0.00153 0.33340 0.00152 0.47667 0.03091 1.39545 0.15870 0.30950 0.07808
460.0 5.89273 26.79748 2.24005 8.18331 0.00107 11.48815 0.00298 0.26261 0.00030 0.25239 0.00068 0.22673 0.00133 0.37936 0.00151 0.16622 0.00066 0.15556 0.00131 0.33244 0.00131 0.58719 0.03070 2.30040 0.16942 0.69948 0.07562
469.4 6.42751 26.92994 2.57618 8.18177 0.00114 11.48873 0.00318 0.26274 0.00033 0.25322 0.00075 0.22635 0.00143 0.37963 0.00163 0.16595 0.00073 0.15653 0.00147 0.33190 0.00146 0.67678 0.03306 1.93220 0.16892 0.72008 0.07781
478.1 6.25583 26.69004 2.46015 8.18149 0.00111 11.48672 0.00310 0.26265 0.00031 0.25286 0.00072 0.22680 0.00140 0.37950 0.00160 0.16600 0.00070 0.15593 0.00141 0.33199 0.00140 0.65468 0.03218 2.14714 0.17049 0.76769 0.07750
487.3 5.97566 26.58332 2.28426 8.18183 0.00107 11.48469 0.00302 0.26219 0.00032 0.25279 0.00073 0.22778 0.00144 0.37930 0.00163 0.16565 0.00071 0.15597 0.00142 0.33131 0.00141 0.64399 0.03116 2.49326 0.17277 0.84357 0.07682
496.8 5.80083 26.60054 2.19282 8.18222 0.00105 11.48292 0.00295 0.26190 0.00032 0.25262 0.00073 0.22815 0.00145 0.37929 0.00164 0.16564 0.00069 0.15578 0.00140 0.33129 0.00139 0.65122 0.03088 2.66497 0.17395 0.87035 0.07685
506.0 5.76904 26.64563 2.19040 8.18229 0.00105 11.48292 0.00295 0.26178 0.00032 0.25256 0.00074 0.22826 0.00147 0.37934 0.00166 0.16548 0.00070 0.15576 0.00141 0.33096 0.00140 0.66180 0.03110 2.70730 0.17518 0.88333 0.07717
515.5 5.46299 26.90586 1.93526 8.18267 0.00104 11.48349 0.00294 0.26121 0.00031 0.25192 0.00074 0.22971 0.00148 0.37848 0.00166 0.16528 0.00071 0.15545 0.00143 0.33055 0.00141 0.63080 0.02981 3.34567 0.18284 1.01360 0.07903
524.5 5.34556 27.42264 1.86198 8.18279 0.00109 11.48562 0.00307 0.26071 0.00032 0.25136 0.00076 0.23062 0.00151 0.37752 0.00169 0.16514 0.00074 0.15525 0.00149 0.33029 0.00147 0.63597 0.03009 3.82219 0.19453 1.11695 0.08449
533.7 5.34831 27.89207 1.84998 8.18305 0.00083 11.48965 0.00227 0.26019 0.00034 0.25100 0.00081 0.23216 0.00162 0.37682 0.00181 0.16473 0.00082 0.15552 0.00167 0.32945 0.00164 0.62160 0.02988 4.43761 0.21128 1.25838 0.08996
543.1 6.19516 26.48954 2.54625 8.18272 0.00117 11.49276 0.00327 0.26228 0.00032 0.25220 0.00074 0.22667 0.00146 0.38002 0.00163 0.16591 0.00069 0.15467 0.00137 0.33183 0.00137 0.75263 0.03390 2.24422 0.17576 0.82287 0.08074
552.1 6.38441 26.27536 2.65446 8.18458 0.00121 11.49270 0.00339 0.26214 0.00035 0.25317 0.00081 0.22636 0.00163 0.38094 0.00180 0.16549 0.00076 0.15600 0.00153 0.33099 0.00152 0.78580 0.03463 2.25398 0.17667 0.84011 0.07985
561.4 6.55330 26.25395 2.77144 8.18713 0.00115 11.49128 0.00322 0.26230 0.00035 0.25320 0.00081 0.22571 0.00160 0.38116 0.00178 0.16567 0.00074 0.15592 0.00149 0.33133 0.00148 0.80421 0.03532 2.15512 0.17699 0.80988 0.08051
570.8 7.07119 26.47107 3.19337 8.18870 0.00117 11.49230 0.00327 0.26270 0.00037 0.25359 0.00084 0.22456 0.00165 0.38204 0.00181 0.16597 0.00075 0.15597 0.00151 0.33194 0.00150 0.84759 0.03756 1.81216 0.17722 0.73596 0.08223
579.2 6.82279 26.16836 2.96915 8.19083 0.00112 11.49205 0.00314 0.26251 0.00035 0.25331 0.00081 0.22499 0.00160 0.38177 0.00176 0.16590 0.00072 0.15575 0.00146 0.33179 0.00145 0.82821 0.03633 2.02281 0.17742 0.78258 0.08111
588.1 6.97023 26.12191 3.10097 8.19239 0.00115 11.49429 0.00322 0.26261 0.00036 0.25341 0.00082 0.22454 0.00162 0.38193 0.00179 0.16583 0.00074 0.15587 0.00149 0.33167 0.00148 0.85144 0.03729 1.90169 0.17789 0.77666 0.08211
597.2 7.44620 26.31384 99.84540 8.19322 0.00120 11.49875 0.00338 0.26296 0.00038 0.25378 0.00086 0.22336 0.00167 0.38256 0.00184 0.16610 0.00077 0.15615 0.00155 0.33220 0.00154 0.90965 0.03960 1.60768 0.17830 0.71367 0.08467
606.1 7.30788 26.08231 99.25965 8.19542 0.00116 11.49919 0.00328 0.26288 0.00037 0.25370 0.00085 0.22338 0.00164 0.38249 0.00180 0.16610 0.00075 0.15606 0.00151 0.33220 0.00150 0.90283 0.03885 1.71308 0.17792 0.73959 0.08392
614.7 7.14435 25.94753 3.39971 8.19672 0.00114 11.50242 0.00319 0.26277 0.00036 0.25345 0.00083 0.22357 0.00161 0.38226 0.00177 0.16603 0.00074 0.15597 0.00150 0.33207 0.00149 0.89707 0.03822 1.85946 0.17869 0.78348 0.08370
623.7 6.59525 26.04390 2.92050 8.19611 0.00115 11.51099 0.00323 0.26246 0.00034 0.25264 0.00079 0.22373 0.00146 0.38103 0.00162 0.16598 0.00073 0.15554 0.00146 0.33197 0.00145 0.96010 0.03764 1.88980 0.17531 0.83512 0.08490
632.6 6.73343 25.54420 2.97863 8.19744 0.00117 11.51414 0.00327 0.26261 0.00034 0.25301 0.00080 0.22417 0.00155 0.38155 0.00172 0.16585 0.00075 0.15583 0.00152 0.33169 0.00151 0.88990 0.03638 2.15915 0.17900 0.90559 0.08370
641.2 6.43371 25.26453 2.77363 8.20023 0.00110 11.51362 0.00306 0.26241 0.00033 0.25264 0.00077 0.22441 0.00146 0.38095 0.00163 0.16580 0.00071 0.15557 0.00144 0.33159 0.00143 0.89732 0.03545 2.31532 0.17749 0.95903 0.08296
684.7 6.44499 24.49716 2.82143 8.21056 0.00098 11.52013 0.00277 0.26247 0.00032 0.25258 0.00074 0.22361 0.00140 0.38105 0.00156 0.16588 0.00068 0.15571 0.00138 0.33177 0.00137 0.93081 0.03543 2.30843 0.17598 1.02438 0.08283
727.0 6.74725 23.30142 3.20232 8.21893 0.00096 11.53620 0.00268 0.26266 0.00032 0.25290 0.00076 0.22215 0.00141 0.38186 0.00156 0.16602 0.00070 0.15616 0.00142 0.33205 0.00141 1.00519 0.03648 2.04645 0.17253 1.07963 0.08411
768.3 6.89651 21.88400 99.89733 8.22730 0.00088 11.55461 0.00244 0.26277 0.00032 0.25325 0.00076 0.22084 0.00141 0.38272 0.00156 0.16610 0.00072 0.15673 0.00146 0.33220 0.00144 1.06207 0.03628 1.85865 0.16653 1.16653 0.08399
johnsoevans
Nothing obvious springs to mind.
Do you get the same behaviour if you run one of the misbehaving refinements again?
What happens if you change the finish_X slightly?
alancoelho
R-Bragg in terms of the observed intensity Io and calculated intensity Ic is defined as:
R-Bragg = Sum[ Abs(Io(hkl) - Ic(hkl)), hkl ] / Sum[Io(hkl), hkl ]
where
Io(hkl) = Sum[ Peak(i) Yobs(i)/Yc(i), i]
Peak(i) = a scaled hkl peak
and the summations are over all hkls
To get high R-Bragg values then Sum[Io(hkl), hkl ] is probably close to zero. Which means that all of Peak(i) intensities are close to zero.
In other words the phase is non-existent. You may instead want to plot the scale parameter for that phase and see how that changes.
cheers
alan
skoswas
I have run the sequential procedure one hundred times, and the behaviour is absolutely the same.
Now I am going to modify the final 2theta, and also to check, as Alan suggests, the scale factor variations.
Thanks for your help
Cheers,
Xavier
skoswas
By restricting the angular domain the thing has improved. From the initial range 1.2 to 17, now the computations were restricted to 1.2 to 12 2theta. This way only the last pattern exhibits a Rwp over 99.
I have checked the scale factors and remain within the same range, 2.8 10^-8. The phase did not vanish.
Still the fact is puzzling...
Thank you ver much for your help.
skoswas
Sorry for the mistake in previous post, I meant R Bragg instead of Rwp.
skoswas
One last observation. When restricting the angular domain to 1.3 to 17.0 instead of 1.2 to 17.0, all R Bragg factors are coherent except for the last pattern.
I am including a list of temperatures, R Bragg and scale factors. Th two previously problematic Rb are marked with ???
Temp(k) R Bragg Scale factor
104.5 2.33 2.889 x10-8
158.2 2.24 2.911 x10-8
210.6 2.27 2.852 x10-8
262.2 2.34 2.755 x10-8
312.9 2.31 2.794 x10-8
362.5 2.46 2.946 x10-8
411.3 2.46 3.008 x10-8
459.9 2.26 2.804 x10-8
469.4 2.56 2.749 x10-8
478.1 2.45 2.772 x10-8
487.3 2.31 2.871 x10-8
496.8 2.21 2.925 x10-8
505.9 2.22 2.915 x10-8
515.5 2.02 2.926 x10-8
524.4 1.94 2.905 x10-8
533.6 1.94 2.898 x10-8
543.1 2.55 2.819 x10-8
552.0 2.64 2.789 x10-8
561.4 2.75 2.762 x10-8
570.7 3.19 2.645 x10-8
579.2 2.95 2.673 x10-8
588.1 3.09 2.734 x10-8
597.2 4.46??? 2.644 x10-8
606.0 3.71??? 2.665 x10-8
614.6 3.42 2.699 x10-8
623.6 2.89 2.858 x10-8
632.5 2.95 2.818 x10-8
641.2 2.75 2.913 x10-8
684.6 2.80 2.928 x10-8
726.9 3.21 2.923 x10-8
768.2 99.94 2.911 x10-8
rowlesmr
Can you provide the data and input files for
579.2 2.95 2.673 x10-8
588.1 3.09 2.734 x10-8
597.2 4.46??? 2.644 x10-8
606.0 3.71??? 2.665 x10-8
614.6 3.42 2.699 x10-8
623.6 2.89 2.858 x10-8
?
johnsoevans
If changing 2theta or d range had an effect it might be worth plotting esd vs X. Are there any rogue data points or esds in the xye file?
skoswas
Dear Mattew,
I am attaching the inp file and the data in two compressed files.
Please let me know any improvement that could be introduced in the script. Also point out any possible mistake.
I would really appreciate it.
alancoelho
Hi Xavier
I got the XY files but I don't see the INP file attached. Can you e-mail me the INP file please
cheers
alan
rowlesmr
The data are all XY.
When I ran the data/file as-received, it all worked. (in TA6; it wouldn't run in TA7 - Not Responding)
If I changed the lower limit from 1.3 to 1.2, I duplicated the 99 Rbragg for the final dataset. There are negative intensities in that data from 1.02 -- 1.24. Maybe that triggered it? But there are also negative intensities from 1.90 -- 2.00...?
The peaks are most definitely there, and there is only one str.
.
Ha!
(I think) It's a bug triggered by having the initial intensities being negative. I can make Rbragg be 99 by making the initial intensities be negative.
alancoelho
Thanks for looking into that Matthew.
Negative intensities would play havoc in the Io formula:
Io(hkl) = Sum[ Peak(i) Yobs(i)/Yc(i), i]
I think it would be best to set the intensities to zero.
skoswas
Dear Colleagues,
I am going to try to summarise the situation on the R Bragg higher than 99 issue.
Matthew Rowles kindly run the inp script with the whole series of data sets. He has a straight Windows machine to run Academic TOPAS v.6. First he run the script with angular domain set between 1.3 and 17. Everyting was fine; all R Bragg within bounds.
Then he decreased the angle, starting at 1.2 2theta, and when executed, the last pattern (#31 at 768 K) run astray. He examined the points and, indeed, found rogue intensities, as John pointed out: Negative values between 1.02 and 1.24 and also between 1.9 and 2.0 2theta. He simply removed the negative signs between 1.02 and 1.24 and run the script again. This time all diffraction patterns exhibited R Bragg within bounds and the presence of negative values between 1.9 and 2.0 didn´t seem to affect.
Matthew suggested that I tried to reproduce his findings, but I couldn´t. I always obtained Rb larger than 99 for the last pattern no matter I removed the negative signs. Now I have to explain that I use a Mac machine with MAC OS Catalina and Parallels emulation of Windows 7 64 bits to run Academic TOPAS v6.
So, the provisional conclusion is that there is no issue for an execution of the script in a straight Windows machine provided that the rogue points are duly corrected.
Being aaware of this facts and noticing that the last pattern exhibited the lowest background of the series, I decided to attempt something weird. You may hit the rough though. I increased the background of last pattern by adding 100 units to all pints. Same thing to the two other problematic patterns but with 40 units. Notice that the highest intensity is higher than 3000 counts.
Then I run the script --between 1.2 and 17-- and presto! All Rb were within bounds as you can see below.
Temp(K) Rwp Rie Rwp PDF R Bragg Scale factor (x10^8)
104.5 6.68622 29.82854 2.38 2.892
158.2 6.53265 28.72710 2.28 2.910
210.6 6.56902 27.71072 2.29 2.851
262.2 6.65435 27.60508 2.36 2.753
312.9 6.52946 26.97072 2.31 2.797
362.5 6.75275 26.90068 2.48 2.947
411.3 6.69387 26.74042 2.48 3.010
459.9 5.89235 26.76382 2.23 2.804
469.4 6.42861 26.94508 2.57 2.743
478.1 6.25695 26.68817 2.45 2.766
487.3 5.97527 26.58315 2.28 2.867
496.8 5.80009 26.60056 2.19 2.925
505.9 5.76818 26.64935 2.19 2.914
515.5 5.46224 26.90207 1.93 2.924
524.4 5.34491 27.42164 1.86 2.906
533.6 5.35406 27.90878 1.84 2.894
543.1 6.19468 26.49461 2.54 2.816
552.0 6.38588 26.30064 2.65 2.793
561.4 6.55292 26.24389 2.77 2.761
570.7 7.07137 26.50039 3.19 2.645
579.2 6.82192 26.16378 2.96 2.672
588.1 6.96488 26.11820 3.09 2.733
597.2 6.96411 26.29118 3.23 2.646
606.0 6.83986 26.07572 3.13 2.665
614.6 7.14681 26.07962 3.41 2.709
623.6 6.59576 26.04801 2.92 2.857
632.5 6.73250 25.53428 2.97 2.815
641.2 6.43783 25.26765 2.77 2.914
684.6 6.44327 24.50503 2.82 2.927
726.9 6.74651 23.25795 3.20 2.910
768.2 6.02061 22.06586 2.89 2.913
Perhaps this behaviour could provide clues of what is going on. Now the question is what is the "real" Rb value for sets at 597, 606 and 768 K?
In any case I would like to thank everyone for the contributions and hints provided. If Alan wishes to have a closer look into this issue I could furnish the patterns. Also, in due time (after being accepted for publication) you could use these results for teaching purposes if you want.