Following is what happens typically with
=> PROGRAM FULLPROF (Version 2.6-Oct94-LLB JRC)
Authors : J.Rodriguez-Carvajal (I.L.L.)
based on the original code provided by
D.B. Wiles & R.A. Young & A. Sakthivel
when you perform a
refinement with and without specimen displacement and transparency.
Choosen case : Al2O3, SRM 1976, a sample particularly designed to
test these effects because of light elements ; "|Fobs|" extracting mode.
A) WITHOUT DISP and TRANSP
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
=> No. of fitted parameters: 10
=> Phase No. 1 AL2O3 R -3 C
=> No. of reflections: 124
==> RELIABILITY FACTORS:
=> N-P+C: 3964
=> Rp: 6.05 Rwp: 8.62 Rexp: 5.37 Chi2: 2.58 L.S. refinement
=> Conventional Rietveld R-factors ==>
=> Rp: 8.45 Rwp: 10.9 Rexp: 6.80 Chi2: 2.58
=> Deviance: 0.109E+05 Dev* : 2.750
=> DW-Stat.: 1.0475 DW-exp: 1.9065
=> Phase: 1
=> Bragg R-factor: 2.84 Vol: 254.820 Fract(%): 0.00
=> Rf-factor= 1.79 ATZ: 0.000 Brindley: 1.0000
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Now, you try to refine one supplementary parameter : TRANSP first,
**NOT DISP FIRST : MORE UNSTABLE**. All damping factors set to 0.2 ...
Half the results are given:
=> Conventional Rietveld R-factors ==> Cycle
=> Rp: 8.45 Rwp: 10.9 Rexp: 6.80 Chi2: 2.58 1
=> Rp: 8.55 Rwp: 11.0 Rexp: 6.79 Chi2: 2.63 3
=> Rp: 8.88 Rwp: 11.2 Rexp: 6.79 Chi2: 2.73 5
=> Rp: 9.03 Rwp: 11.3 Rexp: 6.79 Chi2: 2.77 7
=> Rp: 9.04 Rwp: 11.3 Rexp: 6.80 Chi2: 2.77 9
=> Rp: 8.97 Rwp: 11.3 Rexp: 6.80 Chi2: 2.75 11
=> Rp: 8.87 Rwp: 11.2 Rexp: 6.80 Chi2: 2.71 13
=> Rp: 8.76 Rwp: 11.1 Rexp: 6.80 Chi2: 2.68 15
=> Rp: 8.65 Rwp: 11.0 Rexp: 6.80 Chi2: 2.64 17
=> Rp: 8.53 Rwp: 11.0 Rexp: 6.80 Chi2: 2.61 19
=> Rp: 8.45 Rwp: 10.9 Rexp: 6.79 Chi2: 2.59 21
=> Rp: 8.38 Rwp: 10.9 Rexp: 6.79 Chi2: 2.57 23
=> Rp: 8.32 Rwp: 10.9 Rexp: 6.79 Chi2: 2.56 25
=> Rp: 8.27 Rwp: 10.8 Rexp: 6.79 Chi2: 2.55 27
=> Rp: 8.23 Rwp: 10.8 Rexp: 6.79 Chi2: 2.54 29
=> Rp: 8.20 Rwp: 10.8 Rexp: 6.79 Chi2: 2.53 31
=> Rp: 8.16 Rwp: 10.8 Rexp: 6.79 Chi2: 2.52 33
=> Rp: 8.15 Rwp: 10.8 Rexp: 6.79 Chi2: 2.52 35
=> Rp: 8.14 Rwp: 10.8 Rexp: 6.78 Chi2: 2.52 37
=> Rp: 8.13 Rwp: 10.8 Rexp: 6.78 Chi2: 2.51 39
=> Rp: 8.12 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51 41
=> Rp: 8.12 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51 43
=> Rp: 8.11 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51 45
=> Rp: 8.11 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51 47
You see that Rp increases first, then decreases. Globally Rp
goes from 8.45 to 8.11%.
B) Final result with TRANSP added
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
=> No. of fitted parameters: 11
=> Phase No. 1 AL2O3 R -3 C
=> No. of reflections: 124
==> RELIABILITY FACTORS:
=> Cycle: 16 => MaxCycle: 16
=> N-P+C: 3946
=> Rp: 5.81 Rwp: 8.50 Rexp: 5.36 Chi2: 2.51 L.S. refinement
=> Conventional Rietveld R-factors ==>
=> Rp: 8.11 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51
=> Deviance: 0.106E+05 Dev* : 2.691
=> DW-Stat.: 1.0375 DW-exp: 1.9068
=> Phase: 1
=> Bragg R-factor: 2.83 Vol: 254.824 Fract(%): 0.00
=> Rf-factor= 1.79 ATZ: 0.000 Brindley: 1.0000
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Now, you try to refine one supplementary parameter : DISP,
together with TRANSP. All damping factors set now to 0.1 ....
1/6 of the results are given:
=> Conventional Rietveld R-factors ==> Cycle
=> Rp: 8.11 Rwp: 10.7 Rexp: 6.78 Chi2: 2.51 1
=> Rp: 23.7 Rwp: 28.8 Rexp: 6.79 Chi2: 18.0 7
=> Rp: 25.3 Rwp: 31.4 Rexp: 6.84 Chi2: 21.1 13
=> Rp: 22.0 Rwp: 27.1 Rexp: 6.86 Chi2: 15.6 19
=> Rp: 18.4 Rwp: 22.3 Rexp: 6.86 Chi2: 10.6 25
=> Rp: 15.9 Rwp: 18.9 Rexp: 6.83 Chi2: 7.69 31
=> Rp: 13.9 Rwp: 16.5 Rexp: 6.81 Chi2: 5.86 37
=> Rp: 12.5 Rwp: 14.9 Rexp: 6.80 Chi2: 4.77 43
=> Rp: 11.3 Rwp: 13.6 Rexp: 6.79 Chi2: 4.00 49
=> Rp: 10.5 Rwp: 12.8 Rexp: 6.78 Chi2: 3.54 55
=> Rp: 9.80 Rwp: 12.1 Rexp: 6.78 Chi2: 3.20 61
=> Rp: 9.23 Rwp: 11.6 Rexp: 6.78 Chi2: 2.95 67
=> Rp: 8.83 Rwp: 11.3 Rexp: 6.77 Chi2: 2.80 73
=> Rp: 8.53 Rwp: 11.1 Rexp: 6.77 Chi2: 2.69 79
=> Rp: 8.30 Rwp: 10.9 Rexp: 6.77 Chi2: 2.62 85
=> Rp: 8.16 Rwp: 10.8 Rexp: 6.77 Chi2: 2.57 91
=> Rp: 8.04 Rwp: 10.8 Rexp: 6.77 Chi2: 2.53 97
=> Rp: 8.00 Rwp: 10.7 Rexp: 6.77 Chi2: 2.51 103
=> Rp: 7.96 Rwp: 10.7 Rexp: 6.77 Chi2: 2.50 109
=> Rp: 7.94 Rwp: 10.7 Rexp: 6.77 Chi2: 2.49 115
=> Rp: 7.92 Rwp: 10.7 Rexp: 6.77 Chi2: 2.49 121
=> Rp: 7.91 Rwp: 10.7 Rexp: 6.77 Chi2: 2.48 127
=> Rp: 7.90 Rwp: 10.7 Rexp: 6.77 Chi2: 2.48 133
=> Rp: 7.90 Rwp: 10.7 Rexp: 6.77 Chi2: 2.48 139
You see that again Rp increases first (a lot), then decreases.
Globally Rp goes from 8.11 to 7.90%.
C) Adding DISP : final result
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
=> No. of fitted parameters: 12
=> Phase No. 1 AL2O3 R -3 C
=> No. of reflections: 124
==> RELIABILITY FACTORS:
=> N-P+C: 3930
=> Rp: 5.66 Rwp: 8.43 Rexp: 5.36 Chi2: 2.48 L.S. refinement
=> Conventional Rietveld R-factors ==>
=> Rp: 7.90 Rwp: 10.7 Rexp: 6.77 Chi2: 2.48
=> Deviance: 0.105E+05 Dev* : 2.678
=> DW-Stat.: 1.0337 DW-exp: 1.9072
=> Phase: 1
=> Bragg R-factor: 2.78 Vol: 254.736 Fract(%): 0.00
=> Rf-factor= 1.77 ATZ: 0.000 Brindley: 1.0000
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Summary : no sample TRANSP TRANSP+DISP
effect
Rp % 8.45 8.11 7.90
damping 0.2 0.1
zero -.0029(2) 0.0051(4) -.049(3)
a 4.75880(1) 4.75883(2) 4.75828(6)
c 12.99297(4) 12.99301(4) 12.99147(18)
TRANSP 0. -.0076(7) -.0039(8)
DISP 0. 0. 0.052(6)
Asym1 -.121(7) -.069(9) -.03(1)
Asym2 0.008(1) 0.007(1) 0.009(1)
X 0.0013(2) 0.0009(2) 0.0008(2)
Eta(p-V) 0.65(1) 0.68(1) 0.70(1)
U 0.0045(1) 0.0044(1) 0.0044(1)
V 0.0009(2) 0.0015(2) 0.0017(2)
W 0.00203(8) 0.00157(8) 0.00144(7)
Without damping factors, the refinement may CRASH. This is to
be tested very ultimately in a refinement. Now, is it significative
or a new bin ? I don't know.
Good Luck,
Armel Le Bail - Laboratoire des Fluorures, CNRS-URA-449,
Universite du Maine, 72017 Le Mans Cedex, FRANCE -
armel@ONE.univ-lemans.fr or lebail@naimn2.cnrs-imn.fr