Home

USER'S GUIDE to PROGRAM DBWS-9807a RIETVELD

1. JOBTYP 1 INSTRM 1 IDATA 0 The first line contains the variables START STEP STOP and DATAID in 3F8 A56 format START beginning angle in 2 STEP step size in 2 STOP last angle in 2 DATAID An alphanumeric string identifying the data The rest of the file consists of the data and the no of counters in 10 F2 F6 format NCOUNT no of counters Y I Average intensity DBWS programs use these data to reconstruct the originally observed total count at each step and from that to assign the statistically correct weight to each observed step intensity Yio II C Background data file 1 Tape Unit 3 If NBCKGD 1 ICF line 2 this file contains background data to be subtracted from the diffraction pattern data Tape 4 The start angle step size stop angle and data format must be the same as they are in Tape 4 In the present version of the program this requirement probably means that this option for Tape 3 can only be used when the experimental pattern data are in the standard DBWS format 8 F7 0 1X If NBCKGD lt 1 Tape 3 is used only as a scratch file 40 DBWS 9807a User s Guide 20 8 00 If NBCKGD 22 a Tape 3 is used only as a scratch file and b the background will be interpolated from the NBCKGD number of intensity angle points listed in the ICF by the operator according to the instructions in ICF line 6 2 Tape Unit 11 If NBCKGD 1 this file contains background d
2. C Selectable models C 1 Profile shape models functions F The profile function to be used in a given run is selected with the control variable NPROF in line 2 of the ICF Input Control File see ILA The currently available functions are listed below The origins and performances of most of these functions are discussed in Young and Wiles 1982 The TCHZ function is discussed in Young and Desai 1989 13 DBWS 9807a User s Guide 20 8 00 NPROF FUNCTION NAME 0 NC exp Co 27 2 x Hx Gaussian G HxVp DNC LU 1 C 27 27 7 Hx Lorentzian L p Hk 2 2NCy 1 1 C2 2 2 K HET Mod 1 Lorentzian p Hk 3 NCz_ 1 1 C3 27 2 H47 2 Mod 2 Lorentzian 2 Hk 4 Split Pearson VII SPVII Low angle side Split Pearson VII SPVII L 1 A Cal1 2 1 20 20 f 77 High angle side Split Pearson VII SPVII H Call 1 A 27 1 29 28x 7 A in this function is a refinable asymmetry parameter The shape parameters m and my can be refined individually as a function of 2 in the same way as can m in the single PVII profile function 6 These formulations of the normalized SPVII functions are taken from Toraya 1986 14 DBWS 9807a User s Guide 20 8 00 5 L G pseudo Voigt pV The mixing parameter can be refined as a linear function of 2 wherein the refinable variables are NA and NB NA NB 2 Hastings et al 1984 have shown how the FWHM s of the individual L and G compon
3. ZER DISP TRANS pqrt CODEWORDS BACKGROUND CODEWORDS PHASE NUMBER 1 ATMS FU AFOPA PREFDIR ISWT SPACE GROUP LBL M NTYP x y z B So CODEWORDS BETAS CODEWORDS LBL M NTYP x y z B So CODEWORDS BETAS CODEWORDS SCALE Bo OVERALL UVWCTZXY CELL PARAMETERS PREF1 PREF2 R RCF_ASYM NA NB NC MIX_PARAMS NA NB NC HIGH SIDE PEARSON ASYM FACTOR PHASE NUMBER 2 ATMS FU AFOPA PREFDIR ISWT SPACE GROUP LBL M NTYP x y z B So CODEWORDS BETAS CODEWORDS LBL M NTYP x y z B So CODEWORDS BETAS CODEWORDS SCALE Bo OVERALL UVWCTZXY CELL PARAMETERS PREF1 PREF2 R RCF_ASYM DBWS 9807a User s Guide 20 8 00 5850 0000 0000 NA NB NC MIX_PARAMS 00 00 00 0000 0000 0000 NA NB NC HIGH SIDE 00 00 00 0000 PEARSON ASYM FACTOR 00 IL B The observed data file Tape Unit 4 Tape 4 contains the observed experimental data from the diffractometer The format can be in any of several formats depending on the type of instrument and the software used on it The acceptable formats are here divided into categories which differ according to whether the incident beam intensity is fixed or varying and whether a single or multiple detectors are used All require steps consisting of equal increments in 2 The data format type is specified in the ICF by a combination of flags in line 2 For case a of Category 1 and for categories 2 and 3 set the flag IDATA 0 That is the flag in position 11 in line
4. 2 or 3 For further information about what these flags control see III D 2 Flag No Flag setting and output called for 1 IOT 1 observed amp calculated intensities at each step 2 IPL 1 line printer plot file 3 IPC 1 Istructure factors amp R Bragg 2 as for 1 plus IFl IFl amp R F with the phases with the phases reported directly as phase angles 3 as for 1 plus IFl F amp R F with the Fo and Fezie written in A iB form to include the phase information NOTE that this flag must be set on i e non zero i to cause the inclusion of the possible Bragg reflection positions in the PLOTINFO file and ii for the DMPLOT display of the calc d patterns of the separate crystalline phases to work 4 MAT 1 correlation matrix 5 NXT 1 input file updated from last cycle results 6 LSTI 1 initial reflection list with indices multiplicities breadths positions 27 DBWS 9807a User s Guide 20 8 00 LP factors and mixing parameters 7 LST2 1 data list as corrected e g for background 8 LST3 1 merged reflection list 9 IPL1 1 symmetry operators 10 IPL2 1 Not used Reserved for possible alternate plot file creation 11 IPLST 1 Stacked summary of the cycle by cycle value of each parameter applied shift and R value 2 output file as for IPLST 1 plus a separate summary of the last cycle parameters and their e s d S The following options 12 18 to output the specif
5. 36 347 369 or after evaluation of equation 3 in the T aylor amp Matulis paper 1991 Appl Cryst 24 14 17 Also see IID2 PREF preferred orientation direction in reciprocal space expressed as M iller indices ISWT wt of the total sample contributed by the internal standard N ote In previous versions of DBW S there was a problem with the preferred orientation correction that could be significant in cases of very high preferred orientation and high symmetry T hat problem has now been corrected 11 3 20A1 SYMB space group symbol eg P6 m P 63 M P2 2 2 P 21 2121 Pb PB and sometimes P 1 1B etc P3 P 3 Fm3m F M 3M N ote the upper case in this symbol Itis needed for some other space groups also T he program treats the lower case m in this example as specification of an hexagonal space group and tries to be accommodating For rhombohedral space groups add H to the symbol to be sure that hexagonal axes will be used e g R 3 C H and add R to demand rhombohedral axes 11 4 Atom specific parameters 4 lines for each of the N atoms 11 41 A4 1x 14 1x A4 2x 5F8 LABEL identification characters for atom M multiplicity of the particular site as given in International T ables for X ray Crystallography Vol 1lorVolA Itis not refinable NTYP link to scattering data for atom use all caps either a0 the atom s name from line 8 1 which will access its manually added scattering factor set or b0 it
6. Dr Marciniak s address is included on one of the screens you will see when you use his program The usual format of the RR plots is that the observed pattern is plotted in the upper field as a series of unconnected symbols dots squares or e s d bars the calculated pattern is plotted in the same field as a solid line formed by straight line connections of adjacent points the difference pattern observed minus calculated is plotted in the lower field as a continuous curve and the positions of the geometrically possible Bragg reflections are indicated as a series of short vertical lines in the intermediate field To let them be distinguished the Bragg markers for each phase are plotted at different levels in the intermediate field to let them be distinguished However the user may choose other positioning of the plots and other 48 DBWS 9807a User s Guide 20 8 00 symbols or connections The RR plots in the FIGURES section are screen dumps of plots produced with DMPLOT The jagged appearance of the enlarged plot sections is due to the pixel size not to any problem with DMPLOT Had we used the output to plotter feature of DMPLOT this jagged appearance would not be present 2 Adroit use an example In the following figures one may see an example of the use of RR plots to guide fine tuning of a refinement from a maybe we can quit here stage to a very much better stage The data used are in the QTEST subdirectory in the dist
7. for all phases p plus the background b ya SRIZy Sp Ab Sx Fx F 27 27 As Lx Pxl p Yoi 12 DBWS 9807a User s Guide 20 8 00 where Sp is the scale factor for phase p note that quantitative phase analysis depends on this Sr is a function to model the effects of surface roughness A choice of 4 models is provided IC4 Ab is an absorption factor here left at 1 00 That is it is not implemented That is acceptable for 1 the x ray case of an infinitely thick flat specimen in a diffractometer using fixed slits and Bragg Brentano geometry the most common case and 2 the neutron case of any shape of specimen with negligible absorption and completely bathed in the neutron beam Fx is the structure factor F is a reflection profile function which approximates the effects of both instrumental and possibly specimen features A choice of 8 analytical functions is provided See I C 1 As is a profile asymmetry function two choices are available Lx contains the Lorentz polarization and multiplicity factors Px is a preferred orientation function The user has a choice of two models See I C 2 b is the background contribution The user has a choice of four ways to represent it The fourth method has a number of subsets 281 C 4 The ratio of the intensities for the two a wavelengths if used is absorbed in the calculation of Fx so that only a single scale factor is required for each phase
8. provided by Paiva Santos uses the formulation of Hill and Howard 1987 to produce the phase fraction by weight that each modelled phase constitutes of the total weight of the modelled phases The calculation is based on the refined values of the scale factors lattice parameters and atom site occupancies as updated in each least squares cycle plus the incorporated table of atomic weights The atomic weights table covers all atoms for which scattering factors are included in their incorporated table If other atoms are used their atomic weights must be supplied by the user in line 8 1 of the ICF Input Control File see IIA If the mol fraction is also wanted the number of formula units per cell must be provided by the user It is important to note that in the version 9411 form the phase fraction calculation is made WITHOUT any correction for microabsorption In x ray work especially that can be a serious problem requiring separate analysis But see 9 below and 11 D2 B What is new in the 9807 and 9807a versions The changes from version 9411 made to produce the current version 9807 include the following Listed in arbitrary order 1 The format of the ICF Input Control File has been changed particularly with respect to atom site multiplicity and site occupation number In the ICF for the 9411 version the parameter N for the site occupancy multiplier for an atom was in the last position in the first of the four lines for the a
9. 0000 0000 0000 0000 38 62 116 96 134 91 00 00 31 0000 41 0000 181 0000 0000 0000 Alpha quartz 2 3 1 0000 00 001 00 00 P 32 2 1 SI1 3 SI 4 47002 47002 00000 54982 91 00 91 00 00 190 50 00490 00270 00490 00140 00010 00002 00 00 00 00 00 00 ol 6 O 1 14855 41914 11983 1 14715 101 00 111 00 121 00 200 50 01430 00000 01900 00850 00320 00420 00 00 00 00 00 00 658E 02 0000 11 00 00 01959 00327 00594 00000 00000 00000 160 25 00 80 25 00 00 00 4 9139 4 9139 5 4052 90 0000 90 0000120 0000 61 00 61 00 71 00 00 00 61 00 1 00000 00000 16680 00 00 130 50 5100 0000 0000 140 50 00 00 0000 0000 0000 00 00 00 0000 00 PHASE 2 ALPHA ALUMINA 2 6 1 00001 001 00 00 00 R 3CH AL1 12 AL 3 00000 00000 34998 80000 00 00 220 50 00 00000 00000 00000 00000 00000 00000 00 00 00 00 00 00 ol 18 0 1 31309 00000 25000 1 00000 230 10 00 00 00 00000 00000 00000 00000 00000 00000 00 00 00 00 00 00 469E 04 3 2945 51 00 240 50 03121 00327 02430 00000 00000 00000 170 10 00 150 30 00 00 00 4 7609 4 7609 12 9962 90 0000 90 0000120 0000 00 00 00 00 00 00 1 00000 00000 00000 00 00 00 36 1 0000 0000 0000 00 0000 1 00000 00 1 03409 210 40 00000 00 1 00000 00 1 00000 00 00000 00 2 24 LINE 2 1 LINE 3 35 0000 0000 CYCLS EPS RELAX P_CALC PARAMS REFINED
10. ALPHA BETA GAMMA cell anglesa B in degrees 11 72 6F8 CA CB CC CALPHA CBETA CGAMMA codewords for the cell constants Note that within the operations of the program the angles are expressed in terms of cross products of the reciprocal space cell edge vectors in the program Thus for example in a trigonal or an hexagonal system the codeword CGAMMA must be the same as CA CB 11 8 11 81 3F8 Enter G G P where G and G are the preferred orientation parameters see formulae below and P is the asymmetry parameter see RLIM in line 4 for the 20 range over which P is effective 11 82 3F8 CG CG codewords for preferred orientation parameters CP codeword for asymmetry parameter If IASYM 0 in line 2 the asymmetry function is that given by Rietveld 1969 33 DBWS 9807a User s Guide 20 8 00 If IASYM 1 the asymmetry function is that given by Riello et al 1995 For both asymmetry models P and its codeword should be set O when the split Pearson VII profile function number 4 is used GI G2 are the preferred orientation parameters used in the formulae see Part 2 81C2 a if IPREF 0 see line 2 1 lar l G2 H1 G2 exp G ax Rietveld T oraya case G 0 for no preferred orientation NOTE Expect strong correlation between the Rietveld G and the phase scale factor if strong preferred orientation is present b if l P R E F 1 l cor l o G cos H1 G sin x gt March Dollase c
11. a monochromator in the diffracted beam put M O N 1 1 and put M ON 2 0 if amonochromator works on the incident beam ee aes MON2 monochromator band pass parameter for a focusing graphite monochromator MON 1 could be equal to 4 and M O N2 could be equal to 2 NOTE MONO1 and MONO2 are needed mostly when Compton scattering or other scattering which is not confined to a narrow band of wavelengths is important to the experiment For a description of these parameters and of their values and use see Riello Canton amp Fagherazzi 1997 10 21 3F8 0 FLAM codeword for amorphous i e Tape 11 data scale factor FLMON 1 codeword for the first monochromator constant MON1 FLMON2 codeword for MON2 10 3 6F 9 4 BACK background coefficients in standard NBCKGD 0 polynomial representation 10 31 6F9 4 FBACK codewords for background coefficients 11 First group of phasespecific parameters NPHASE as specified in Line 2 sets of these lines numbered 11 are needed i e one set for each phase 11 1 A70 PHSNM phase name and number 31 DBWS 9807a User s Guide 20 8 00 11 2 214 F8 0 3F 4 0 F 7 2 N number of atoms in the asymmetric unit FU number of formula units in the unit cell used only for reporting molar phase fraction AFQPA particle absorption factor t for each phase used in the microabsorption correction T he value can be taken from table V page 368 in the Brindley paper 1945 Philos M ag
12. angle for the calculations and ii assigning the desired ending angle to ST OP 37 DBWS 9807a User s Guide 20 8 00 Case 1 Free format IDATA 1 The first line contains the variables START STEP STOP and DATAID in 3F8 2 A48 format just as in case 0 The rest of the file consists of the intensity data only in free format Any number of columns of any width are allowed The delimiters are commas or spaces Angle data are not allowed Case 2 GSAS standard format IDATA 2 The DBWS program searches through the header material that starts with BANK to find the start angle step size and stop angle The rest of the file consists of the intensity data in format 10 I2 F6 0 No commas or angle data are allowed Note that in the standard GSAS format for the data all rows must be filled out to 80 characters Use spaces if necessary to fill out the row s Case 3 Philips UDF format IDATA 3 The program searches through the voluminous header material for the lines DataAngleRange and ScanStepSize from which it gets the start angle stop angle and step size information It then looks for the start of the data file immediately following the RawScan line The intensity data are expected to be in format 8F8 1X where X in this case is actually a comma Case 4 Scintag text format IDATA 4 The program searches through the multi line header for the start angle step size and stop angle It then ignores the axis
13. as otherwise specified e g in Profile 7 the angular dependence of H is modeled in accord with the ICDD recommendation Lowe Ma et al 1997 as 16 DBWS 9807a User s Guide 20 8 00 Hx U tan V tan W CT cot where U V W and CT are refinable parameters C 2 Preferred orientation Px models In position 8 of line 2 in the ICF a 0 selects the Rietveld Toraya model and a 1 selects the March Dollase model The Rietveld Toraya model Toraya s modification of the function Rietveld put in his original code is Pk G2 1 G exp Giax and the March Dollase model is Px G cos a 1 G sin a where G and G are refinable parameters and ax is the angle between dk and the presumed cylindrical symmetry axis of the texture e g fiber axis direction We note that the value of G corresponding to no preferred orientation is 0 in the Rietveld Toraya function and 1 00 in the March Dollase function It may be noted here that in the Rietveld Toraya model the G parameter tends to be correlated with the scale factor Sp highly so if the preferred orientation is strong C 3 Profile asymmetry models Two choices are available Setting flag 9 in ICF line 2 to O selects the usual Rietveld asymmetry model As 1 P sign 2 2 k 2 2 tan k Setting that flag to 1 selects the Riello Canton and Fagherazzi 1995 model Cu 26 20 10 265 As 20 29x 1 E wx tan x 2 wg 17 DBWS 9
14. as their positions and importantly because their shapes may be inherently different Consider for example how d would be small for a symmetric calculated Lorentzian profile used to fit a symmetric Gaussian experimental profile even if the two profiles had the same peak positions peak heights and areas 2 A Comment about interpretation of R s R p and R wp are not directly comparable to the R values obtained in single crystal refinements R Bragg and R F come closer to that mark but suffer from the fact that I is not directly observable where reflection profiles overlap the usual case The actual observed intensity in a clutch of reflections is parceled out among the component reflections in proportion to the calculated reflection intensities to produce the L s As is well known to experienced users R p and R wp are also very sensitive to the background level if the background is included in the yo s above It is so included by DBWS 9807 when the refinable background model is used It is obvious that R wp for example will be decreased as the background level is increased signal to noise level decreased because it is easier to get a good fit to a function that varies slowly and near to monotonically with angle as the background does than to one that has many sharp maxima along the way as the pattern of the Bragg reflections does Experienced users therefore do not expect to judge the relative quality of Rietveld refinements made with
15. blocks and goes to the four column main output angle count time and e s d The DBWS program then reads the intensity data from column 2 and ignores the other three columns Case 5 Siemens UXD format IDATA 5 As in cases 2 4 above the program searches through the header material to find the start angle step size and stop angle or step ignores the rest of the heading material and proceeds to read just the intensity data in counts from the rest of the file 38 DBWS 9807a User s Guide 20 8 00 Case 6 Rigaku ASCII format IDATA 6 The program proceeds as in 5 Category 2 Varying incident x ray flux e g synchrotron x radiation and a single detector collecting data in equal increments of 2 JOBTYP 0 INSTRM 1 IDATA 0 DBWS programs use the normalized intensity data both directly as the observed data and with the other data provided in the file to reconstruct the originally observed total count at each step and from that to assign the statistically correct weight to each observed step intensity Yio The first line contains the variables DATE for date or place and DATAID in A5 A56 format Line Format Content 1 A5 A56 DATE Identifies the format of the data Currently supported values are a OCT85 FEE respectively DATAID an alphanumeric string identifying the data 2 18 F8 G10 NRANGE No of blocks in which the data are written CHMBRI Value of incident beam counts per step
16. factor for an amorphous or other component represented by whole pattern intensity data on Tape 11 C 5 Models for Surface Roughness effects Sn As was dramatically demonstrated by Sparks et al 1992 surface roughness has a far more important effect on the relative intensities than had been generally appreciated Here we offer the user a choice among four models in the spirit of experimentation no guarantees that any of them are particularly 18 DBWS 9807a User s Guide 20 8 00 appropriate Their refinable parameters are likely to be highly correlated with each other and of course with the overall scale factor However they do represent a gesture in the right direction We would be grateful to learn of your experiences in using one or more of these options The user s choice among the 4 different models is indicated by the digit entered for IABSR in line 2 of the ICF The choices of model are IABSR Model 1 Combination of the Sparks and Suortti models 2 Sparks et al 1992 model straight line 3 Suortti 1972 model exponential 4 Pitschke et al 1993 model The models as recast and normalized to 1 00 at 90 by us are 1 Combination model Sr rf 1 0 p expl q p expl q sin 1 5 1 p 2 2 Sparks et al model Sr 1 0 t p 2 in radians 3 Suortti et al model Sr 1 0 p exp q p expl q sin 4 Pitschke et al model Sr 1 0 pq 1 0 q p
17. further information about this program NOTES 1 Some of the test case data files are rather long e g 20 140 2 in 0 02 steps You may wish to use only a portion of each e g 20 60 to save computer time when you are making your first trial runs See the discussion Part 2 II B of the data file to find how to do this 2 Atthe beginning of refinement with a new data set it is advantageous to refine just a few parameters at first then those plus a few more etc It is therefore advantageous to assign the lowest numbered codewords to those first to be refined parameters The choice of codeword sequence in the Input Control Files Part 2 IIA for these test cases has been influenced but not fully controlled by that 10 DBWS 9807a User s Guide 20 8 00 thought PART 2 The Rietveld refinement program I General Information A General plan and data types accommodated This program is designed to perform Rietveld 1967 1969 analysis on x ray or neutron nuclear scattering only powder diffraction data collected i with a fixed slits 2 diffractometer Bragg Brentano geometry operated in a step scan mode equal steps in 2 with a single detector and with either one or two e g a doublet wavelengths in the constant intensity incident x ray beam Data in this category are now accepted in six specific formats and a free format ii as for i but with a varying incident x ray beam intensity as from a sync
18. given in Part 2 IIAL 50 DBWS 9807a User s Guide 20 8 00 Execution Modify the INPUT CONTROL FILE ICF and DATA files as needed using the instructions in Part 2 of this guide for the Input Control File Tape 5 and the observed data file Tape 4 To run the program in the MSDOS environment on the command line type DBWS DATA INPUT OUTPUT Tape 11 Tape 3 using single spaces between the entries the actual file names you have for the executable program the observed data file the Input Control File and your choice of output file name respectively Tape 11 and Tape 3 relate to a data representing an amorphous component and b a predetermined or assumed whole pattern background respectively For most work they are not used 2 Using DBWS in a Windows environment Although they are DOS programs the executables provided for DBWS can run in the Windows 95 also 98 and NT environment in a DOS window automatically called up when the program execution starts Advantage can thereby be taken of some of the drag and drop and file handling conveniences of Windows 95 98 In a Windows environment you may either 1 work in a DOS window into which you can drag and drop the command line files in sequence or 2 you can drag and drop them onto the command line that is brought up when you select run from the Windows Start Up menu The second option here takes more advantage of Windows features and eliminates retyping of the s
19. in the Int l Tables or a set of lines of the form Posi Scat where Posi sin_ A__and Scat f T he set is terminated by a line with 100 in the first position If the first form is desired A2 can not 0 b Line 8 for the neutron case 8 1 A4 2F8 0 NAM symbol identifying this set DFP b scattering length in units of 10 cm AW atomic weight must be included as part of the set 9 18 MAXS number of parameters to be refined in this pass 10 Refinable global parameters 10 1 7F8 0 ZER offset of the 2 zero point in 2 DISP sample displacement TRANS parameter for effect of sample transparency on apparent 2_ P surface roughness parameter p 30 DBWS 9807a User s Guide 20 8 00 surface roughness parameter q surface roughness parameter r surface roughness parameter t ADD 10 11 7F8 0 FLZER codeword for 20 zero point offset codeword makeup is described near end of I of Part 2 FLDISP codeword for sample displacement FLTRNS codeword for transparency coeff FLGP codeword for surface roughness parameter p FLGQ codeword for surface roughness parameter q FLGR codeword for surface roughness parameter r FLGT codeword for surface roughness parameter t See 2 1 C 5 this part of this Guide for explanation of usage and how to nullify the surface roughness correction 10 2 _ 3F 8 0 AM scale factor for amorphous or other pattern on T ape 11 MON1 monochromator band pass parameter if there is
20. is the number of FWHM s given in the heading The note resembling a fraction that appears there i e 5 21 is the number of cycles in this pass number of parameters varied From the following table one can see which parameters have codeword numbers small enough so that they were included in the number of parameters varied 21 in this case Figure 6 is the RR plot corresponding to data in the subdirectory FTEST on the distribution diskette Fig 6a shows the plot for the range of the fluorapatite data used 15 90 2 Figure 6b shows a closeup of the most troubled looking portion of Figure 6a On such inspection and noting that the R exp is unusually small one sees that the refinement appears to be better than would be expected from the GofF S the Goodness of Fit alone Figure 7 is an enlarged portion of a RR plot from a recent PZT study of Dr Paiva s It is an example of how DMPLOT can display the separate calculated patterns in a multiphase refinement In this case there are four phases C Execution procedure with a PC 1 In DOS A possible Start up routine Create a subdirectory on the hard disk Copy all related files in to that subdirectory including the executable versions of the programs yc Make sure that you are using a version of your ICF that is compatible with this version of the DBWS program You can use DBWS 9807 to convert an ICF used for DBWS 9411 to a format compatible with DBWS 9807 See the instructions
21. molar plus various numerical 43 DBWS 9807a User s Guide 20 8 00 criteria of fit e g R s and e s d s and related values are printed out See IILE for definitions of the various criteria and comments about their interpretation 2 Optional outputs The following optional outputs mostly in the main output file are selected 1 2 or 3 or suppressed 0 by the settings of the flags in line 3 of the ICF input control file They are numbered below in the order of their appearance in ICF line 3 Flag Items output if flag setting is 1 or for flags 3 and 11 1 or 2 1 Observed and calculated intensities After the last cycle a table of the observed and calculated counts at each step can be printed 2 Line printer plot For use when there is nothing else available 3 It is this same flag which must be set on to cause the inclusion of the possible Bragg reflection positions in the PLOTINFO file Different choices of the flag IPC value produce these results in the main output file IPC 1 Istructure factorsl amp R Bragg 3 as for 1 plus Fles Fl amp R F with the Fop and Fic Written in A iB form to include the phase information 2 as for 3 except that the phases are reported in terms of the phase angles 4 Correlation matrix 5 This flag controls creation of the updated input control file discussed in III B above which is out 6 Reflection list A list of the reflections generated from each
22. of the package elements transferred d Date of transfer 3 Please inform the recipient that we can not accept any responsibility for missing parts of the primary distribution package such as documentation and test case files nor for any modifications that might have been made in the program nor for the currency of the version so distributed Unfortunately such problems do seem to be common in secondary distributions 4 Please pass on to the recipient a copy of this notice who should then pass on a copy with each copy of the program package he might distribute etc 5 From all users we request that proper credit citation be given in all written works for which this program played a role The best citation to DBWS 9807a is as an upraded version of DBWS 9411 described by Young Sakthivel Moss and Paiva Santos 1955 The full reference appears in the list of references in this Guide and in the print out of the main output file 59
23. than 8 phases with good data Even the 15 now accommodated may not always be enough The program uses the Newton Raphson algorithm to minimize the quantity 11 DBWS 9807a User s Guide 20 8 00 Sy S wifi ya where Wi 1 yi yi observed gross intensity at the i step Ya Calculated intensity at the i step and the sum is over all data points The normal matrix elements formally given by dYa dYa dYa Mix Si 2wil yi Yai Cae 00 9 9 where the are the adjustable parameters are approximated by deletion of the Dya Yi Ya term 07 07 Parameters are stored internally in arrays XL I J GLB I and PAR LJ XL contains the data for the atoms The first index runs over the atoms the second over the parameters for the atom GLB contains global parameters hose which apply to all phases such as surface roughness 2 zero specimen displacement background and others PAR contains crystalline phase dependent parameters such as lattice constants scale factor profile shape parameters and m and preferred orientation parameters The first index runs over the phases There are corresponding arrays LP LGLB and LPAR which map the parameters to the normal matrix elements This mapping is determined by the user through the use of codewords See ID below B Calculated intensities The calculated counts ya are determined by summing the contributions from neighboring Bragg reflections K
24. 2 For other cases in Category 1 the IDATA flag will be set to other values as is shown below For case a of category 1 and for categories 2 and 3 IDATA 0 the data format is further identified in ICF line 2 by the combination of the settings of flags 1 JOBTYP and 7 INSTRUM Category 1 Constant incident x ray or neutron flux and a single detector collecting x ray or neutron data at equal increments in 20 JOBTYP 00r 1 INSTRUM 0 Case 0 The standard D BW S observed data file T ape Unit4 IDATA 0 The first line contains the variables START STEP STOP and DATAID in 3F8 3 A48 format START beginning angle in degrees 20 ST EP step size in degrees 20 ST OP last angle in degrees 20 DATAID An alphanumeric string identifying the data T he rest of the file consists of the intensity data themselves in 8 F 7 0 1X format T here may be but need not be a ninth column listing the 2 angle for the last datum in each row NOTE The START angle assigned here must correspond to that of the first datum not in an excluded region The STOP angle can be any angle less than that of the last datum in the file One can select any desired angular range for the refinement without removing any data from the file by i declaring an excluded region NEXCRG in line 2 of the ICF and specifying its bounds line 7 of the ICF as beginning at the START angle declared in the first line of the data file and ending at the desired beginning
25. 2 CNA CNB CNC Codewords for NA NB NC 11 93 3F 8 4 NA NB NC for the high angle side of SP7 Split Pearson VII 11 94 3F 8 2 Codewords for NA NB NC on the high angle side of SP7 11 95 F 8 4 A the SP7 asymmetry factor 11 96 F8 2 CA codeword for A NOTE When the Split PVII function is not being used the parameters and codewords for both m and A on the high angle side should be set to zero W hen the Split PVII function is being used one should start with small parameter values and proceed cautiously as this function tends to run away easily One strategy that has been used successfully is that of using the results from refinement with PVII for the initial values in SPVII and then alternating refinement of the parameters for the H igh side and the Low side NEXT PAGE Example of an Input Control File for a simple two phase case as it appears on the screen T hisis a copy of the QIN file in the QTEST subdirectory on the distribution diskette Other examples are provided in the test cases T he ICF for the test case with an internal standard and an underlying amorphous component may be particularly interesting for advanced users 35 DBWS 9807a User s Guide 20 8 00 QDATA 2 PHASE TEST CASE QUARTZ alumina 8FWHM 13 5 99 0 5 2 0 0 0 00211 10000 10000 000 0 1 0 1 0 0 1 54050 1 54430 50000 90 0000 8 0000 8009 2 10 95 95 95 95 24 0000 0183 0000 0000 0000 1 0000 0000 21 0000
26. 807a User s Guide 20 8 00 in the regions 2 min lt 2 lt 90 RLIM and 90RLIM 2 2 max so RLIM must be 0 lt RLIM 90 SW M easured sample T hickness cm 5 14 5F4 3F8 0 Some operational choices M CYCLE number of cycles can not be 0 or no calc n will be done EPS run terminates when all calculated shifts not the actually applied shifts shown in the output are lt EPS e s d Relaxation factors for shifts by parameter groups RELAX 1 co ordinates amp isotropic atom displacement temperature factors and site occupancies RELAX 2 anisotropic atom displacement parameters RELAX 3 profile width asymmetry overall atom displacement temperature preferred orientation parameters lattice parameters and overall scale factor RELAX 4 26 zero specimen displacement specimen transparency surface roughness amorphous scale factor and monochromator band pass parameters N eeded for pattern calculation only THMIN starting angle _ 26 for the pattern to be calculated STEP stepsize 2 THM AX ending angle _ 28 6 2F8 2 if NBCKGD gt 2in ICF line 2 enter NBCKGD lines here specifying the background intensities measured at those NBCKGD points POS position in degrees 2 BCK background counts at this position 7 2F8 2 if NEXCRG in line 2 is gt 0 enter NEX CRG lines with ALOW low angle bound 2 AHIGH high angle bound 8 H ere isthe place to enter scattering factors and scattering lengths which are
27. 807a User s Guide 20 8 00 where Cy is the refinable parameter 2 x is the position of the peak of the K reflection profile 2 2 x is the displacement of the observed peak from the calculated one and w is the refined value of the FWHM of the K reflection C 4 Background representation The background intensity y at the i step may be obtained by any of several methods The user s choice is indicated by the value of NBCKGD entered in the fourth position in line 2 of the ICF The choices are 1 an operator supplied table Tape 3 of background intensities NBCKGD 1 or 2 linear interpolation between operator selected points in the pattern NBCKGD n where n is the number of points or 3 a specified background function NBCKGD 0 or 4 the alternative background representation of Reillo et al NBCKGD 1 If the background is to be refined in the NBCKGD 0 option yi must be obtained from a refinable background function The one available in the current version of DBWS is 5 Ybi S Bn 2 BKPOS 1 m 0 where BKPOS is user specified in line 4 of the Input Control File ICF See IIA If the alternative background NBCKGD 1 is to be used there are several choices to be made 1 Absorption correction for Bragg Brentano geometry 2 Compton and disorder diffuse scattering contribution for each phase using either individual isotropic or overall atom displacement factors 3 Refinement of the scale
28. DBWS 9807a User s Guide 17 5 99 USER S GUIDE to PROGRAM DBWS 9807a for RIETVELD ANALYSIS OF X RAY AND NEUTRON POWDER DIFFRACTION PATTERNS with a PC and various other computers R A Young Allen C Larson and C O Paiva Santos School of Physics Georgia Institute of Technology Atlanta GA 30332 17 5 99 20 8 00 Copyright notice 1998 1999 2000 R A Young AII rights reserved Current Addresses 1 Dr Allen C Larson 14 Cerrado Loop Santa Fe NM 87505 8248 2 Prof C O Paiva Santos Instituto de Quimica UNESP C P 355 14800 900 Araraquara SP BRAZIL DBWS 9807a User s Guide 20 8 00 CONTENTS OF THIS GUIDE Item Part 1 Preliminaries I Foreword A History of DBWS B What is new in the 9807 version IL Platforms and Environments A Portability B Windows environment III Contents of the Distribution Package Part 2 The refinement programs I General information A General plan and data types accommodated B Calculated intensities C Selectable models 1 Profile shape models functions 2 Preferred orientation models 3 Profile asymmetry models 4 Background representation four types 5 Models for surface roughness effects D Quantitative phase analysis 1 Procedure a With no internal standard b With an internal standard 2 Microabsorption E Codewords and constraints F Array size selection and executable program size II Description of input files A The Input Contro
29. I assembly program 57 B Copyright and Fair Use 59 FIGURES 60 74 PART 1 Preliminaries I FOREWORD History of DBWS and what is new in this version A History The program is designed to carry out Rietveld refinements with X ray or neutron powder diffraction data in digitized form collected under any of several of the most commonly used instrumental conditions Fixed wavelength s and equal increments in the scattering angle 2 are required conditions The current program is the latest in a long chain of versions each updated and upgraded from its predecessor and each distributed in a distribution package containing the source code a User s Guide plus data and Input Control Files for test cases For the 1990 the date of the first PC version and later distributions executable versions of PC compatible MSDOS plot programs were included These plot programs were offered by third parties as Shareware items The chain started with DBW2 9 Young and Wiles 1981 which was written in FORTRAN IV and incorporated some parts of Rietveld s 1969 original code particularly the codeword system A W Hewat s code for anisotropic thermal parameters and A C Larson s code for dealing with space group symmetries and reflection multiplicities However it was otherwise written from scratch and incorporated many new features some quite major It was written to be used with x ray as well as neutron data to be modular single pass and portable
30. LOTINFO with any of several other generally available software packages 2 PLOT FILES that can be generated when NBCKGD 1 Riello et al case and one or more of flags 12 18 in ICF line 3 is are set to 1 Filename Flag no Description PLOTOSS 12 T otal observed plot file corrected for absorption scattering expressed in counts PLOTCAL 13 T otal calculated plot file obtained by summing the background PLOTPOL 14 Background representation obtained from refining it with a polynomial PLOTCOM 15 Plot file for total Compton scattering contribution to the background corrected for the monochromator band pass function if a monochromator is present on the diffracted beam PLOTDIS 16 Total uncorrelated calculated disorder contribution to the 42 DBWS 9807a User s Guide 20 8 00 background PLOTAM 17 Plot of Tape 11 data e g amorphous corrected for transmissivity PLOTBIG 18 M ultiple columns file to do plots using commercial programs like Origin _ Excel etc It contains in each column respectively PLOTOSS PLOTCAL PLOTAM PLOTAM PLOTPOL PLOTAM PLOTPOL PLOTCOM PLOTAM PLOTPOL PLOTCOM PLOTDIS and weighted residuals defined as X Y cal Y obs ZZ Y obs NOTE ifthe PLOT BIG flagin ICF line 3 issetto 1 the other GPLOT files will not be created T o have those plotfiles created set PLOT BIG 0 B Anupdated Input Control File Tape 5 which is updated with the final cycle parameters is output if
31. Some of the other new features were choice of a variety of profile functions a refinable background model a multiple phase capability and a number of other features listed in Wiles and Young 1981 DBW2 9 which incidentally had several bugs was soon followed by DBW 3 2 1982 written in FORTRAN V DBW3 2 was superseded in 1987 by DBW3 2S a rather major revision made by A Sakthivel and more nearly conforming to ANSI 77 standards Next followed versions DBWS 8711 November 1987 and in turn DBWS 8804 April 1988 Version DBWS 8804 had a small bug which allowed only 2 phases rather than the intended 8 to be refined correctly That error was corrected in version DBWS 8804a In the next major modification the program was adapted to run on PC type computers and was named DBWS 9006PC This version with the CalComp instruction routine omitted was fully ANSI77 Subsequently various of the several hundred users of record have successfully compiled and run DBWS 9006 on a variety of computers both small and large The principal differences between successive versions up to this point were listed in the appendices to the relevant versions of the User s Guide and for the DBWS 8804 to DBWS 9006 differences in the Foreword of the User s Guide for DBWS 9006 The principal differences between the next version DBWS 9411 and DBWS 9006 were 1 DBWS 9807a User s Guide 20 8 00 rearrangement of the input control file ICF to make it more user frie
32. Yoi Ye S Wi Yo the weighted pattern R factor R expected 100 N P C S w yoi where N is the total number of data points observations P is the number of parameters adjusted and C is the number of constraints applied S is the goodness of fit the ratio R wp R expected Calling for the reflection list flag 3 in line 3 of the Input Control File causes R Bragg and R F to be printed out at the end of the final cycle where 45 DBWS 9807a User s Guide 20 8 00 R Bragg 100 SI I Ll S Io which is the R value for Bragg intensities and R F 100 SI Props IFeaicll S IF obs l which is the R value based on the deduced observed and calculated structure amplitudes Some workers find this useful in ab initio structure determination c D W Dis the Durbin Watson statistic d Hill and Flack 1987 It is included in the output from NDELTA N d Ay Ay 7 L ye i 2 i each cycle as a possibly useful additional indicator of the progress of the refinement It is given by where y Wi Yio Yic In DBWS 9411 and 9807 the w s are all taken to be 1 00 in order to yield the unweighted form That is preferred by some statisticians over the weighted form discussed by Hill and Flack 1987 and given in Schwarzenbach et al 1989 The weighted form is used in DBWS 9006 This d is sensitive to misfit of the calculated and observed reflection profiles both because their areas may differ as well
33. ame command line With a good text editor that will operate in Windows and not leave tracks e g control characters and anything else not compatible with ASCII text one may enjoy essentially all of the benefits of the Windows operating system while using DBWS The text editor that we are presently using with considerable satisfaction is the Programmer s File Editor PFE32 from Alan Phillips at Lancaster University UK It is available free from his website http www lancs ac uk people cpaap pfe We find it is very helpful to keep a Shortcut to Pfe32 exe icon on our desktop preferably in or near one corner of the screen where it can be accessed with little or no movement of the windows that are open 51 DBWS 9807a User s Guide 20 8 00 The first step in the procedure is to prepare all of the files that will be needed for the next set of Rietveld refinement cycles the next run and to put them in one subdirectory not far from the root directory The results worth keeping can be transferred later from this working directory to another more comprehensive file directory The purpose of the not far specification is to maximize the probability that files generated during the refinement run will find their way back to your working directory rather than going off someplace where you have to hunt for them e g in the Windows directory The files that are always wanted in the working directory are 1 the exec
34. ase G 1 000 for no preferred orientation where ax is the acute angle between the scattering vector e g dx and the orientation direction e g fiber axis direction specified in line 11 2 A special feature With the IPREF 0 model setting G to any number gt 99 0 for a phase causes the program to generate for it only those reflections for which d is parallel to the preferred orientation vector PREF specified in line 11 2 as though that phase had a 100 complete fiber axis texture Note however that since the LP factor has not been corrected for the texture the relative intensities among orders of a reflection so generated are not correct but are the same as those for a randomly oriented powder This is true for all preferred orientation corrections in this program Line Format Description 11 9 MIXING SHAPE PARAMETERS NOTE for the usual pseudo Voigt NPROF 5 the mixing parameter can vary with 2 as __ NA NB 2 where NA and NB arerefinable and NC must be set 0 34 DBWS 9807a User s Guide 20 8 00 For the Pearson V II functions N PRO F and NPROF 6 the shape parameter m is calculated as m NA NB 20 NC 20 m m or m or ms where NA NB and NC are refinable parameters For NPROF 5 NC and its codeword must be set 0 For all other profiles listed 1 3 7 NA NB amp NC and their codewords must all be set to zero 11 91 3F 8 4 NA NB NC for pV and PVII or the low angle side of SP7 11 92 3F 8
35. at they would also run under some older versions of MSDOS They also ran very well in the Windows 95 98 environment See C2 of Part 3 B Plots 1 Production and character Although it has no plotting code embedded in it DBWS 9807 does produce the output file PLOTINFO which can be used by other programs to make Rietveld refinement RR plots The separate plotting program offered as SHAREWARE for which the MSDOS executable code is here included is DMPLOT EXE It will use PLOTINFO or specified other file with the same structure to make a plot with selectable X and Y magnifications either on screen or for better resolution with many HP and similar plotters of the observed and calculated patterns the difference and the Bragg reflection position markers for all phases The on screen plots can then be dumped to a laser or dot matrix printer The program works with various commonly available graphics adapters for both monochrome and color monitors of various resolution capabilities DMPLOT has many very useful features including separate plots for the phases multiple zoom d and I readouts etc See the Read me file in the DMPLOT subdirectory on the distribution diskettes DMPLOT is the property of its author Dr H Marciniak The executable version is offered here as a Shareware item which means that you the user should try out the program and IF you find it really useful you should send the requested fee to the author
36. ata for which the scale factor can be refined in the Rietveld refinement process These data may be partly calculated e g Compton and disorder diffuse scattering or real data from an amorphous or other sample or both See Reillo Fagherazzi et al 1995 For the Tape 11 data i e for the amorphous or other material for which only the scale factor is to be refined The first line contains the variables START 1 STEP1 STOP1 TMV SW1 DATAID 1 6F8 2 A 16 START 1 beginning angle in degrees 2 it must be equal to START in the main data file tape 4 ST EP 1 step size in degrees in 26 it must be equal to ST EP in the main datafile ST OP1 last angle in degrees 26 it must be equal to STOP in the main data file T MV linear absorption coefficient for the amorphous phase cm SW 1 sample thickness cm DATAID1 an alphanumeric string identifying the data The rest of the file consists of the intensity data in free format These background data from Tape 11 are called for by inclusion of the file name in the fifth position on the command line used to start a refinement For example PGM DATA ICF OUT TAPELI 41 DBWS 9807a User s Guide 20 8 00 III Description of Output files A Plot files 1 PLOTINFO A file called PLOTINFO identified internally as Unit 9 is generated by DBWS It is an alpha numeric file containing the title information from line 1 of the Input Control File a separately labeled listing of the geo
37. ation for phase abundance then becomes Taylor and Matulis 1991 S ZMWY 1 PO N y s ZMV q jel E Codewords and constraints This effective codeword system has been carried over from Rietveld s original code A codeword is entered for each parameter Each codeword has two parts 1 the designation of a matrix position for the parameter and 2 specification of the fraction of the calculated shift that is to be applied to this parameter The codewords have the form sdddc cc where the ddd digits specify the matrix position for the particular parameter The c cc digits specify what fraction or multiple of the calculated shift is to be applied to update the parameter and s is the sign or desired for the applied shift A positive shift will be applied if no minus sign is present Separately from the codewords the user may also supply a group relaxation factor to be applied to the 21 DBWS 9807a User s Guide 20 8 00 calculated shifts for all parameters in each of four different groups of parameters See line 5 of the Input Control File As an example assume that one wishes to refine the x y z coordinates of an atom and that y x 2 is required by the crystallographic symmetry If we let the codewords be given as x 131 00 y 130 50 and z 141 00 then x and y will be assigned to the 13 normal matrix parameter and z to the 14 Also 1 00 times the calculated shift in the 13 parameter will be ap
38. atly simplified in modern versions of DBWS One need only specify the numbers a single time in the file PARAM INC and then keep that file in the same subdirectory with the source code file or in the path when the compiling is done If the refinement stops and an error message to the effect that IDSZ is less than MSZ MAXS appears it will be necessary to increase IDSZ or reduce MSZ or MAXS until IDSZ is equal to or greater than MSZ MAXS The reason is that the space in memory set aside for IDSZ is also used at different times by MSZ MAXS MAXS is the number of parameters being varied line 9 in the ICF Obviously as the matrix size and number of parameters being refined increases one could quickly run into the 640K barrier characteristic of MSDOS and PCDOS Compilers are commonly available that can overcome that problem and also make use of the full 32 bit architecture in today s PC s One such is the compiler is a part of Microsoft s FORTRAN PowerStation It makes use of a separate memory manager program DOSXMSF EXE which then must be in the same directory or in the path with the exe program being used The executable versions of DBWS 9807a in the distribution package have been so prepared Even under Windows 95 98 the DOSXMSF EXE file is needed because the actual carrying out of the operations is still subliminally a DOS activity 23 DBWS 9807a User s Guide 20 8 00 II Description of input files A The input contr
39. be the working directory it is necessary either 1 to have the out file in the working directory before it is dragged over and dropped on the DOS window in which the DBWS program is about to run or 2 to type out in the command line being constructed in the window the full path for the out file which of course gives but little advantage over the plain DOS procedure With option 1 above the out file thereby brings its path information with it This feature has worked well for the PLOTINFO and PLOTINFO BIN files in our PC operating system WIN95 98 with a PII chip 53 DBWS 9807a User s Guide 20 8 00 REFERENCES Caglioti G Paoletti A and Ricci FP 1958 Nuclear Instruments and Methods 35 223 228 Cheary R W and Cline J P 1994 Adv in X ray Analysis 38 75 82 Cox D E and Papoular R J 1996 Mat Sci Forum v 228 231 pp 233 238 Hastings JB Thomlinson W and Cox DE 1984 J Appl Cryst 17 85 95 Hill RJ and Flack HD 1987 J Appl Cryst 20 356 361 Hill RJ and Howard CJ 1987 J Appl Cryst 20 467 474 Jansen E Schafer W and Will G 1994 J Appl Cryst 27 492 496 Lowe Ma C K Cline J P Crowder C E Kaduk J A Robie S B Smith D K and Young R Pitschke W Hermann H and Mattern N 1993 Powder Diffraction 8 74 83 Pitschke W Mattern N and Hermann H 1993 Powder Diffraction 8 223 228 Riello P Canton P and Fagherazzi G 1995 Powder Diffracti
40. d this feature to model an amorphous component of the background with a pattern of an amorphous material stored on Tape 11 The possibility is being explored that this scheme may work for a crystalline component for which the diffraction pattern is known but not the structural data needed for regular Rietveld refinement In version 9807 a modest step has been taken to deal with this problem 17 Plots can be made of the separate contributions to the background e g Compton and disorder as well as the entire modeled background New in version 9807a Il Platforms and Environments A Portability The source code provided can be used to carry out Rietveld refinements on various mainframes workstations and PC s as well as MacIntosh personal computers The user must comment out the parts not applicable to the platform at hand For example it might be necessary to verify that the bitwise mapped logic statements AND OR and XOR are properly recast in all occurrences The DBWS source code now provided is fully ANSI 77 as passed by Microsoft FORTRAN PowerStation 1 0a The executable code provided for DBWS is DOS executable but can be run rather advantageously in the Windows 95 98 environment See IIB below Only executable code is provided for the Shareware plotting program included in the distribution package for making the plots of the Rietveld results see Note 3 below The DBWS program does output two files which can be used for plot
41. different data sets and even different materials by the relative R wp or R p values Those R values are however good 46 DBWS 9807a User s Guide 20 8 00 indicators of the progress of a particular refinement Obviously it would be helpful in some cases if the DBWS programs output the various R values as calculated both with and without the background included Coding for that would probably not be very difficult but it has not yet been done Users particularly those to whom this effect of background is a new thought may find interesting the Jansen Sch fer and Will 1994 paper in which are given specific examples of the effects of excluding vs including the background in R value calculations Almost all users may find interesting the Cox and Papoular 1996 paper on R factors Errors and Significance tests Although Synchrotron Data is a part of the title the content of the paper has much wider potential applicability 47 DBWS 9807a User s Guide 20 8 00 PART 3 Further comments on using DBWS 9411 AND 9807 with a PC A Compiling The executables provided in this distribution package and used for the test case examples given here were produced with the Microsoft FORTRAN PowerStation compiler It has the capability of producing executable versions which can make use of nearly all of the RAM available i e not limited to 640K We have run some of the executable versions under MSDOS 6 0 We suppose but have not demonstrated th
42. e R values that appeared on the screen during the run At some point you will want to look at the out file perhaps print parts of it all easily done with Pfe32 exe and perhaps check the magnitude and pattern of the parameter shifts the calculated e s d s for them and the correlation matrix in order to guide your choice of changes to make in the ICF for the next run When you have finished setting up the ICF for the next run you can press the Start button and then run again The command line you used before now appears If the file names have not changed you need do nothing but press the OK button and the computer takes off on the next set of refinement cycles while you look on with delight at seeing all this work being done for you while you just sit there If you do not keep all of the needed files including program files in the working subdirectory the problem may then occur that a file created during the operation of the program will go to another directory That will usually be the Windows directory where at t he expense of some effort and exasperation it can be captured and brought back to the working directory However DBWS 9807 includes some special programming that is supposed to avoid that problem by sending a newly created file e g PLOTINFO to the same place that the main output file e g out or whatever other name was specified in position 4 of the command line goes To assure that the place to which it goes will
43. e user supplies the number of formula units per unit cell in ICF line 11 2 b With internal standard DBWS can also make use of an internal standard in Quantitative Phase Analysis The fact that an internal standard is to be used is signified with a number in the last position the twelfth in line 2 1 of the ICF The crystallographic data for the internal standard phase are entered as a set of 11 lines just as is done for the other phases present The sequence number of this internal standard phase among all of those entered in this section of the ICF is the ISPHASE number to be entered in line 2 1 The wt which this phase contributes to the total sample weight is entered in line 11 2 at the last position in the line in F72 format For example if the standard constitutes 28 42 of the weight of the entire sample the number 28 42 will be entered there CAUTION As they stand above these QPA calculations do not take into account any microabsorption effects see for example Taylor and Matulis 1991 They will be important unless the linear absorption coefficients are the same or nearly so for all phases being analyzed Therefore the neglect of microabsorption effects will generally be a much more serious error in the xray powder diffraction case than in the neutron powder diffraction case 2 Microabsorption A beginning effort at providing a microabsorption correction has been made in DBWS 9807 It involves the use of a particle absorptio
44. ents can be recovered from the values in this profile function 6 Cs Hx 1 4 2 1 2 22 Hk Pearson VII PVIT and m can be refined as a function of 2 as m NA NB 2 NC 2 Y where the refinable variables are NA NB and NC We note that PVII Lifm landGifm 7 Modified Thompson Cox Hastings pseudo Voigt Mod TCH pV TCHZ Thompson et al 1987 Note The G used in the description of this function is simply a symbol related to the profile breadth and is not to be confused with the gamma function represented by G m in the definition of the normalizing constants C4 and Cs in the PVII functions TCHZ L 1 G where 1 36603q 0 47719q 0 1116q q G G G G AG G BGG CGG DGgG G Hx 15 DBWS 9807a User s Guide 20 8 00 A 2 69269 B 2 42843 C 4 47163 D 0 07842 Go U tan V tan W Z cos 7 G X tan Y cos The modification consists of addition the parameter Z to provide a component of the Gaussian FWHM whichis constant in d as is the Y component of the Lorentzian FWHM Young and Desai 1989 In the above profile functions the refinable parameters are those in m Hk and in the SPVII functions A The normalizing constants are C 4 In2 Ci 4 Cz 42 1 Cys Ahha 2 I A C4 i Jr Hk Al m 27m _ J _ 20 my 2 1 gt a Tm 0 5 Hk is the full width at half maximum FWHM of the K Bragg reflection Except
45. flag 5 in line 3 of the ICF is set at 1 Routinely one does want this file to be generated so that the next step of the refinement process can proceed from the last ending point without extra keyboarding NOTE The updated ICF is written over the old one which is then lost C Output to terminal during run This file is assigned to the terminal when the program is run interactively e g when the program is run ona PC or workstation It displays the program name DBWS and version the sizes declared for the five most frequently redimensioned arrays the title information for the run the number of cycles called for the number of parameters being refined the START and STOP angles plus the step size and a dynamic series of dots marking the progress of the program through the current cycle At the end of each cycle it reports R p R wp R exp and S GofF We thank Dr H Marciniak for his very useful additions of much of the information content and the dynamic display to this output to terminal A compilation of R p and R wp cycle by cycle values are saved to the working directory as Unit 7 D The main output file Tape 6 1 Standard parts of the main output file All control variables and structure parameters are printed out so that a given run can be reconstructed unambiguously At the end of each cycle the parameters with their shifts applied not the full calculated ones the quantitative phase analysis result wt and
46. ha Faculty of Chemistry Jagellonian University ul Ingardena 3 30 060 Krak w ROLAND have implemented XRS 82 to run on 386 486 PC s A Rietveld program is incorporated See their note in J Appl Cryst 27 201 202 1994 Among the PC versions if one is willing to pay for a program that by Schneider WYRIET might be especially interesting because of its improved speed its use of extended memory and all of the ancillary programs that come with it e g Fourier electron density mapping bond distance and angle calculations menu driven preparation of the input control file and various kinds of plots It is available from Margarita Schneider E D V Vetrieb Starnbergerweg 18 D 8134 P cking GERMANY 2 Programs ancillary to DBWS a Input file Control File preparation programs As distributed the Input Control File is to be prepared with a full screen editor For the experienced user that method is more convenient and faster than a menu driven system such as might be preferred by the beginning user Such a menu system is provided as WYIN EXE and WYIN HLP in J Schneider s WYRIET package Free distribution of WYIN EXE has been permitted 2 DB2dI assembly program The ICDD International Centre for Diffraction Data is always interested in adding more high quality diffraction patterns of well characterized materials to its Powder Diffraction File PDF 2 in the established searchable format The PDF 2 now has gt 100 000 pat
47. he multiplicity of the general site for the space group involved is printed in the first part of the main output file on the line between the first mention of the space group and the listing of the initial parameters 3 The variety of data formats accepted by DBWS in Category 1 See 2 II B has been increased to include in addition to the standard DBWS format a free format the GSAS standard format Philips UDF format a Scintag text format a Siemens UXD format and a Rigaku ASCII format All of the additions are based on equal steps in 2 a single detector and constant incident beam intensity 4 The output of structure factor magnitudes called for by flag 3 in line 3 of the ICF can now be called to report the structures factor phases as well Setting the flag at 2 calls for the phase angles to be reported as such for use in Fx IFxl exp G f K where f x is the phase angle and i is the square root of 1 Setting the flag at 3 calls for this output to include phase angle information by reporting F in the A iB form In the process of making these changes to the structure factor outputs an earlier error was corrected so that the IFl s are now reported to 3 decimal places and do not differ for the al and a2 wavelengths 5 The number of phases that can be handled at once has been increased from 8 to 15 The main rational for this increase is to accommodate quantitative phase analyses which can benefit from inclusion
48. hrotron source or iii with fixed incident neutron beam intensity or monitor count and wavelength and with multiple counters which do not all contribute at every step such as the HRNPD instrument at Brookhaven National Laboratory USA and the DIA and D2B instruments at ILL France It can also be used with data collected in HDS geometry from non absorbing specimens It can be used with other geometries if an external data preparation step is taken e g to compensate for the differences e g the effects of a varying divergence slit or the angle dependent absorption in absorbing specimens used in HDS or Guinier geometry The program can also be used in a pattern calculation only mode The program is currently set up to handle up to fifteen phases simultaneously The previously used number eight was chosen on the presumption that most powder diffraction patterns are not likely to have enough information in them to support refinement of structural details in more than eight phases at once However there is now widespread and growing recognition that multiphase Rietveld refinement is the very best way to do quantitative phase analysis QPA if the structures are known or are known well enough to be refined Therefore cases will occur commonly in quantitative phase analysis in which only a few parameters often only the scale factor need be refined for most of the phases In such cases sensible QA results can be expected for many more
49. ied plot files are intended primarily for use in the NBCKGD 1 case 12 13 14 15 16 17 18 PLOSS 1 Observed data plot file corrected for absorption scattering PLCAL 1 Calculated data plot file PLPOL 1 polynomial background plot file PLCOM 1 Compton scattering plot file PLDIS 1 uncorrelated disorder plot file PLAM 1 Plot file of data in amorphous diffraction data file Could be for a polycrystalline rather than an amorphous material PLOTBIG 1 Observed corrected for absorption amorphous amorphous plus Compton C NOTE if PLOTBIG is set to 1 PLOSS PLCAL PLPOL PLCOM PLDIS and PLAM will not be created To have these plotfiles created set PLOTBIG 0 9F8 0 Fixed parameter values to be supplied by the operator WAVELENGTH 1 of the incident radiation WAVELENGTH 2 of the incident radiation RATIO intensity ratio e g a a BK PO S origin of polynomial for background in 20 W DT width range on either side of peak max of calc profile in units of H CTHM monochromator coeff in polarization term of the LP factor 14 CT H M cos 2 sin cos TMR linear absorption coefficient cm needed for the transparency correction RLIM A 2 value that determines the 2 range s over which the asymmetry correction is applied IfIASYM 0 Rietveld model only the profiles below RLIM are corrected If IASYM Riello et al model the asymmetry correction is applied 28 DBWS 9
50. indices and separate phase patterns QDATA the observed data file QIN the Input Control File ICF as rewritten at the end of the last cycle 4 In the subdirectory QPATEST you find the files for a somewhat more advanced case involving quantitative phase analysis 3 phases one of which is the internal standard plus an amorphous component To be determined are the wt present of two phases plus the amorphous component QPADATA the observed data 3 crystalline phases plus an amorphous component QPA IN the input control file ICF QPA OUT TAPE11 whole pattern observed data for an amorphous silica sample PLQPA thePLOTINFO file for this case note how well the amorphous component is fit PLQPA BIN 5 In subdirectory DMPLOT files for using this proprietary program offered as a SHAREWARE item DMPLOT CFG Configuration for DMPLOT DMPLOT EXE Version 3 48 DMPLOT TXT Notes on DMPLOT 6 In the subdirectory XCUTABLE there are three DOS executable versions of DBWS 9807 They differ only in the choices of dimensions of the five arrays listed in PARAM INC See the accompanying Read me file for details 7 The current distribution program includes another diskette with a PC executable program named DB2dI provided by the International Center for Diffraction Data ACDD The ICDD hope you will use it to provide good patterns properly formatted for acceptance in their PDF Powder Diffraction File See 2b of Appendix A p 54 for
51. ine instructions for setting up the Input Control File ICF The line by line instructions follow A star before a line number indicates that the line s existence depends on the value of a control variable Line Format Description 1 A70 TITLE any 70 characters to be used to label the main output file and the plots 2 Selection of modules in the Refinement model 2 1 1214 JOBTYP 0 X ray case 1 Neutron case nuclear scattering only 2 pattern calculation only X ray 3 pattern calculation only Neutron NPROF Profile selection 0 Gaussian 1 Lorentzian Cauchy 2 Mod 1 Lorentzian 3 Mod 2 Lorentzian 4 Split Pearson VII asymmetric 5 pseudo Voigt pV 6 Pearson VII symmetric 7 Modified Thompson Cox Hastings TCHZ pV lt 0 If you set NPROF lt 0 an ICF from DBWS 9411 will be converted to DBWS 9807 format and given the name you specified for the output file See IIA1 p 24 for further how to information NPHASE number of phases up to 15 possible NBCKGD background model control 0 background to be refined 5th order polynomial in 20 1 background to be read from file tape 3 2 3 N background to be determined by linear interpolation between the N given points see instructions for ICF line 6 1 Alternative physically based background Riello Fagherazzi Clemente amp Canton 25 DBWS 9807a User s Guide 20 8 00 1995 NEXCRG number of excluded regions NSCAT number of atomic scatter
52. ing factor sets to be added manually INSTRM Laboratory 26 X ray synchrotron X ray or neutron diffractometers using a fixed wavelength radiation 0 Laboratory _ 26 X ray data or single detector neutron data 1 varying incident intensity data e g synchrotron X ray and a single detector or multiple counters and fixed e g constant monitor counts per step incident intensity e g neutron CW data IPREF preferred orientation function 0 Rietveld T oraya function 1 March Dollase function IASYM 0 for usual Rietveld asymmetry model 1 for Riello Canton and Fagherazzi 1995 model IABSR Choice of surface roughness model 1 Combination model 2 Sparks et al model 3 Suortti model 4 Pitschke et al M odel IDATA Additional identification of the format of the particular i nput diffraction data file for the 7 cases in Category 1 single detector constant incident beam intensity and wavelength See section II B Category 1 for further description of these formats and how D BW S utilizes them the standard DBW S format free format see IIB category 1b GSAS format Philips UDF format Scintag text format Siemens U X D format as converted from aRAW filebyX CH program ver 1 4 DOS Rigaku ASCII format HAUBWNEFE Note It will sometimes be possible to utilize data in some other formats by reducing the format to one of the above particularly the free format case 1 ISPH ASE Sequence numbe
53. l File ICF 1 What it is and how to create and update it 2 Line by line instructions 3 Example Page 11 12 13 17 17 18 18 19 20 20 21 22 24 25 35 36 DBWS 9807a User s Guide 20 8 00 B The observed data files Category 1 Constant incident flux and a single detector 7 data formats accepted Category 2 Varying incident x ray flux e g synchrotron and a single detector Category 3 Fixed wavelength and multiple detectors e g CW neutron sources C Background data files 1 Tape 3 2 Tape 11 III Description of output files A Plot files 1 PLOTINFO file for general use 2 PLOT FILES when NBCKGD 1 B Updated Input Control File C Output to terminal during run D The main output file 1 Standard parts of the main output file 2 Optional outputs to this and other files E Definitions and interpretations of the numerical criteria of fit 1 e s d s R s GofF DW d 2 Comment about interpretation of R s Part 3 Further comments on using DBWS 9411 and 9807 with a PC A Compiling B Plots 1 Production and chacteristics 2 Use An adroit example C Execution procedure 1 In DOS 2 Using DBWS in Windows REFERENCES APPENDICES A Other programs 1 Other Rietveld refinement programs 2 Programs ancillary to DBWS a ICF preparation programs 37 39 40 40 40 42 42 43 43 44 44 45 46 48 48 49 50 51 54 56 57 DBWS 9807a User s Guide 20 8 00 b DB2d
54. metrically possible Bragg reflection positions and their Miller indices h k l for each phase involved in the refinement flag 3 in line 3 of the input control file must be on for these positions to be included and a listing of the calculated and observed intensities at each step i e Yic and Yio With PLOTINFO the DMPLOT program offered as a Shareware item with the PC version of the DBWS program can produce a plot of the calculated and observed patterns the difference plot and position markers for the possible Bragg reflections from each of the phases involved either a on screen which can then be dumped to a printer to provide a hard copy or b directly via a plotter With PLOTINFO and the additionally generated binary file PLOTINFO BIN DMPLOT can display on screen 1 the Miller indices for the various reflections and 2 the component patterns one for each phase which make up the total calculated pattern in a multiphase refinement DMPLOT also offers Mouse control dynamic zoom on screen display of 2 and d values at the cursor position and several other advanced features Apart from Rietveld refinement plots DMPLOT can also display up to 4 raw data patterns simultaneously on screen and move them around relative to each other To use DMPLOT in this mode the raw data file names must have the extensions rit for x ray data neu for neutron data It should be a relatively simple matter to make a useful Rietveld refinement plot from P
55. n factor for each phase Necessarily such a correction also involves the sizes and shapes and distributions thereof of the particles An approximation to the needed value for 20 DBWS 9807a User s Guide 20 8 00 this particle absorption factor AFQPA in the ICF can be obtained from table V page 368 of the Brindley paper 1945 Philos Mag 36 347 369 or after evaluation of equation 3 in the Taylor amp Matulis paper 1991 J Appl Cryst 24 14 17 which is ta 1 V lexp u u bar x dV where t is a particle absorption factor for phase a the integral is taken from 0 to Va Va is the particle volume of phase a U is its linear absorption coefficient x is the path of the radiation in particle a when diffracted by the volume element dV and u bar is the mean linear absorption coefficient of the solid matrix of the powder The user needs to be cautious with the ubar According to Brindley 1945 p 349 this value must be the mean value for the solid material excluding the spaces between particles Table V in the Brindley paper was computed for spherical particles for use as a better than none approximation He well recognized of course that a naturally occurring material consisting entirely of spherical particles of identical size would not be often found The value of t herein coded as AFQPA for each phase is then used to correct the scale factor for each phase in the specimen for microabsorption The equ
56. n under MSDOS Distribution on other media can be arranged on special request The distribution package consists of this User s Guide plus the following files on magnetic media in six subdirectories all written under MSDOS 1 In subdirectory FOR DBWS 9807a FOR the DBWS source code PARAM INC DOSXMSF EXE The file PARAM INC is needed at compilation time It conveys the user s choice of the sizes for the five most likely to be redimensioned arrays DOSXMSF EXE must be either in the path or in the same directory with the executable version of DBWS that is running 2 In subdirectory FTEST files for a test case with fluorapatite This is a single phase case with a moderate quantity of lines Bragg reflection profiles FDATA the observed data file FIN the ICF set up for refining 31 parameters as rewritten after the last cycle run PLFAP the PLOTINFO file needed by SPLOT and DMPLOT renamed for saving PLFAP BIN the binary file needed by DMPLOT to display the Miller indices and the calculated patterns for the separate phases FOUT the main output file from a run based on FIN README Comments about this example 3 In subdirectory QTEST files for a two phase case of quartz with alumina as a minor impurity DBWS 9807a User s Guide 20 8 00 QOUT the main output file from a run based on QIN PLQ the PLOTINFO file renamed for saving needed by DMPLOT PLQ BIN the binary file needed by DMPLOT to display the
57. ndly 2 correction of the long standing error noted in the earlier User s Guides of a factor of two in the multiplicities calculated for some Laue groups 3 addition of four surface roughness models with refinable parameters 4 addition of a split Pearson VII profile function to provide another means of modeling still imperfectly profile asymmetry 5 addition of a quantitative phase analysis routine see comments below 6 an improved method for inputting the selectable sizes of certain arrays at compilation time via a param inc file plus readout of those sizes to screen during and to output file after each refinement 7 addition of a dynamic screen display of the progress of the current refinement cycle 8 additions to the output to the plotting file PLOTINFO and addition of PLOTINFO BIN so that the plotting program DMPLOT can display the Miller indices at each reflection location and can display simultaneously the separate calculated patterns for each of the various phases being refined in a multiphase specimen 9 change of the Durbin Watson statistic d to the mathematically preferred unweighted version and 10 some additional diagnostics We are indebted to Dr H Marciniak author of DMPLOT for providing features 6 7 and 8 above The split Pearson VII code is based on the mathematics actually used by Toraya for this function in his program PRO FIT introduced in Toraya 1986 The quantitative phase analysis calculation
58. not in the incorporate If NSCAT gt 0 enter here NSCAT sets of lines in the formats shown below for your case T he x ray scattering factors and neutron coherent scattering lengths that are incorporated here are those listed in the 1974 version of Vol 4 or in some cases the 1995 Vol C of the International Tables for X ray Crystallography 29 DBWS 9807a User s Guide 20 8 00 NOTE that the anomalous dispersion corrections f and f to the atomic scattering factors are now given in this program for 10 Ka wavelengths 2 748510 Ti 2 289620 Cr 1 935970 Fe 1 788965 Co 1 540520 Cu 0 709260 Mo 0 559360 Ag 0 215947 Ta 0 209010 W and 0 180195 Au in A instead of only those of Cr Fe Cu Mo and Ag given in previous versions of DBWS The coefficients provided before were taken from the International Tables for Xray Crystallography Vol 1 1974 and the now added ones come from Vol C 1995 One reason for providing these additional data is so that scattering factors for working wavelengths chosen from a continuum e g synchrotron X radiation might be better approximated a Line 8 for the X ray case 8 1 A4 3F8 0 NAM symbol identifying this set DFP f real part of the anomalous dispersion DFPP f imaginary part of the anomalous dispersion AW atomic weight must be included as part of the set 8 2 9F8 0 Either one line of the form Al B1 A2 B2 A3 B3 A4 B4C the coefficients for the approximation to f as used
59. of many phases for which the structure is known and only the scale factor and one or two other parameters need to be refined Dr Marciniak s DMPLOT also handles 15 phases now 6 The quantitative phase analysis routines have now been extended to include an option for operation 6 DBWS 9807a User s Guide 20 8 00 directly with an internal standard 7 A calculation of the wt of the amorphous component is now offered when an internal standard is used 8 X ray anomalous dispersion coefficients for the atoms have been added for 5 more wavelengths bringing the total to 10 sampling the wavelength range from 2 75 to 0 18 A We are grateful to Ian Madsen and Dr R J Hill of CSIRO Minerals Australia for providing the values and associated coding for these coefficients 9 The preferred orientation correction now works correctly for high symmetry non unique axis cases as well as the others The existence of the now fixed problem and conditions under which it might have had significant effects were pointed out in the User s Guide to the 9411 version 10 FWHM As recommended by the ICDD Lowe Ma et al 1997 and by Cheary and Cline 1994 a term in cot has been added to the Caglioti Caglioti et al 1958 expression for the FWHM Full Width at Half Maximum in profile functions 0 1 2 3 5 and 6 i e for the Gaussian Lorentzian simple pseudo Voigt and symmetric Pearson VII functions 11 A tentative model for microabsorp
60. ol file ICF Tape Unit 5 1 What it is and how to create and update it For compatibility with other programs e g DB2dI which produces d and I files in the format needed by the ICDD for the PDF the name you give this file should include the extension in or inp This also guards against the propensity Windows 95 98 has for opening a file with a program selected on the basis of that file s extension This file the ICF contains the control variables along with the structural and other refinable parameters If output of a new ICFis selected flag 5 for NXT in line 3 the current ICF will be updated at the end of the last cycle The following line by line description tells the user how to set up the ICF control file for any given case Some users have developed menu driven preparation programs with which to set up the input control files A good program of that sort can certainly be a help to the beginner However once one becomes familiar with the format detail then to set up a new ICF using an old one as a template or to make adjustments to the ICF between runs it is probably quicker and easier to do so by editing the ICF directly on screen using a full screen editor Particularly for making adjustments to the ICF between runs of successive batches of cycles using the the full screen approach is quicker and easier than plodding through a menu driven approach which forces one through preset sequences Users who are not fully familia
61. on 10 204 206 Riello P Fagherazzi P Clemente D and Canton P 1995 J Appl Cryst 28 115 120 Riello P Canton P and Fagherazzi G 1997 Powder Diffraction 12 160 166 Rietveld HM 1967 Acta Crystallogr 22 151 2 Rietveld HM 1969 J Appl Cryst 2 65 71 Schwarzenbach D Abrahams SC Flack HD Gonschorek W Hahn Th Huml K Marsh RE Prince E Ri Smith DK and Young RA 1998 Extracting Peak Height d I data from Rietveld analyses for the Powder diffraction File Presented at the 47 Annual Denver X ray Conference 3 7 August 1998 To be published in Advances in X ray Analyis 42 1999 2000 Sparks CJ Kumar R Specht ED Zschack P and Ice GE 1992 Adv in X RayAnalysis 35 57 62 Suortti P 1972 J Appl Cryst 5 325 331 54 DBWS 9807a User s Guide 20 8 00 Taylor JC and Matulis CE 1991 J Appl Cryst 24 14 17 Thompson P Cox DE and Hastings JB 1987 J Appl Cryst 20 79 83 Toraya H 1986 J Appl Cryst 19 440 447 Wiles DB and Young RA 1981 J Appl Cryst 14 149 151 Young RA and Desai P 1989 Archiwum Nauki o Materialach 10 71 90 Young RA Sakthivel A Moss TS and Paiva Santos CO 1995 DBWS 9411 an Upgrade of the DBWS Young RA and Wiles DB 1982 J Appl Cryst 10 262 269 55 DBWS 9807a User s Guide 20 8 00 APPENDIX A OTHER PROGRAMS 1 Other Rietveld refinement programs There are now many other Rietveld analysis programs available Most are dist
62. onto the command line By doing so you will avoid the tedium and typographical errors of having to type in both the file names and the paths to them If all has gone well when the specified number of cycles is completed you will find in the working directory the new out file the ICF rewritten if you called for that with flag 5 in line 3 of the ICF a PLOTINEFO file a POTINFO BIN file and a file called unit 7 along with the files that have not been changed You will probably want first to look at the Rietveld plot resulting from the run just finished You can do that by dragging PLOTINFO onto DMPLOT EXE and following the screen instructions The PLOTINFO BIN file is the one that makes it possible for you 1 to see all of the reflection indices for all phases by simply pressing the F3 key and 2 to see the individual calculated patterns for all phases superimposed on the standard Rietveld plot You can now look at the content of and edit any text file by dragging it to the desktop and dropping it on the Pfe32 exe icon If you have enough information from the plot to make wise choices now for improving the fit you may wish to go directly to editing the ICF for the next run i e batch of cycles 52 DBWS 9807a User s Guide 20 8 00 You do that of course by dragging the ICF to the Pfe32 exe icon to open the file and then using the Pfe features to do the desired editing You may want to look at the unit 7 file to see again th
63. phase with the initial input profile component mixing parameters 7 Corrected data list A list of the point by point intensity data as corrected for background the assigned weight and the identifying number of the highest angle and the lowest angle reflections which contribute to the data point The reflection list is sorted so that contributing reflections may be indicated by giving a pair of integers K1 is the number of the reflection in the merged list at the smallest 2 angle contributing to the point K2 is the highest angle 2 reflection for contribution 8 Merged reflection list includes hkl FWHM 2 I l amp R Bragg 9 Listing of the symmetry operators used 44 para DBWS 9807a User s Guide 20 8 00 10 Plot files which are stored and printed out separately See part A of this section III 11 A stacked summary of the cycle by cycle values of each parameter another of the R valu generation E Definitions and interpretations of the numerical criteria of fit 1 Numerical criteria a The standard deviation a mathematical statistic that should not be confused with es d sj My S Yoi yicP N P C where M is the corresponding diagonal element in the inverted normal matrix N is the number of observations P is the no of parameters that are refined and C is the total number of constraints b The numbers after R factors are R p 100 Sly Ycil Slyoil the pattern R factor R wp 100 S w
64. plied to x and 0 50 times it to y The full calculated shift in the 14 parameter will be applied to z If on the other hand one were to set up the codewords as x 131 00 y 131 00 then the sum of x and y would remain constant This constraint by codewords feature can be useful in connection with site occupancies also One may let two different atoms occupy the same crystallographic site and then determine how much of each is present in that site by refining their site occupancies under the constraint that the total site occupancy atomic per cent is to remain constant i e by assigning the same codeword to both atoms but beginning one with a negative sign so that the applied shift in the occupancy for one atom is the negative of that for the other Zero values for the three c s in a codeword result in the parameter not being refined However this is not always an advisable way to turn off a parameter If some codewords are non zero but their applied shifts are set to zero the result can be the physically unrealistic one of correlation matrix elements gt 100 If N parameters are to be refined then no codeword with addd value in the range from 1 to N can be missing If any are the program will stop a hole in the matrix error message will be sent to the screen in the PC context along with identification of the missing codeword The missing codeword s will also be specified at the end of the aborted main output file The numbe
65. q sin 1 0 q sin The refinable parameters are p q r and t Note that is expressed in radians If no correction is being applied for surface roughness effects then set and fix p q 0 and r 1 0 if IABSR 1 and p q r t 0 for all other choices of IABSR 2 3 4 In the Pitschke et al model q must not be 0 zero if a codeword is declared for it else a divide by zero error message will appear D Quantitative phase analysis 1 Procedure a No internal standard 19 DBWS 9807a User s Guide 20 8 00 If no internal standard is used and amorphous content is taken to be negligible the program calculates the weight fraction for each phase refined on the assumption that the phases being refined account for 100 of the specimen via the following relation Hill and Howard 1987 _ Sp ZMV p N S ZMV i where p is the value of 1 for a particular phase among the N phases present S is the refined scale factor ZM is the weight of the unit cell in atomic weight units number of formula units Z per cell times the atomic weight M of the formula unit and V is the volume of the unit cell For this calculation the program uses internally tabulated atomic weights and the refined lattice parameters scale factors and atom site occupancies Atomic weights not in the incorporated table may be inserted in line 8 1 of the ICF Mole fractions as well as weight fractions will be reported at the end of each cycle if th
66. r of codewords declared i e entered as non zero in the Input Control File whether or not called for to be refined line 9 of the ICF must not exceed the declared size for the array MSZ see below F Array size selection and executable program size Although the issue is rapidly being made moot with the recent great increases in PC memory and speed redimensioning of certain arrays was often wanted in order to make best use of limited memory 22 DBWS 9807a User s Guide 20 8 00 and time This requires making changes in the array sizes at all occurrences in the source code and then recompiling With today s powerful PC s one may be able to do the bulk of one s Rietveld refinement work by just setting the array sizes to arbitrarily large values and leaving them there However if that is not feasible for some reason the following are the array size statements one would most often wish to modify IDSZ number of data points IRS maximum number of allowed reflections Kal Ka2 NATS maximum number of atoms in the asymmetric unit allowed MSZ maximum number of elements in the normal matrix i e codewords NOTE If the number of codewords declared exceeds the declared value of MSZ the excess will be quietly ignored NOV maximum number of reflections allowed to contribute to the intensity y at a single data point The task of changing the numbers in these statements at every occurrence in the DBWS source code has been gre
67. r of the internal standard phase as it appears in the ICF If 0 is placed here item 12 in line 2 of the ICF no internal standard will be used or sought in the refinement process If an internal standard is specified here its weight of the total sample will be entered for ISW T in the last position in line 11 2 of the ICF 26 DBWS 9807a User s Guide 20 8 00 Line Format D escription 2 2 214 This line applies only if NBCKGD 1 i e for the alternate background All modules on this line are from Riello et al 1995 IAS Absorption correction 0 no correction 1 an absorption really transmissivity model for B B reflection geometry A 1 exp 2mT sin 6 where m coded as T MR is the linear absorption coefficient for the sample and T coded as SW is the effective sample thickness FONDO alternate background module choice 0 Standard background with options 1 amp 2 below the standard polynomial background can also be used and the Compton and disorder scattering will be added to it 1 Individual isotropic temperature factors the IU Cr recommends the term atom displacement rather than temperature or thermal will be used in calculation of the Compton and disorder diffuse scattering contribution of each phase to the background 2 Overall temperature factor will be used in the above calculations 3 4 511 1X O utput control flags A flag is set off with O and on with I
68. r with the format detail of the ICF and either do not have or do not wish to use a menu driven file preparation program will find useful the example ICF placed at the end of this line by line description It is based on the quartz plus alumina case An ICF format conversion subroutine is provided for the convenience of users who have on hand the ICF s from problems for which they used DBWS 9411 This routine will convert an ICF in DBWS 9411 format into one in the format needed for DBWS 9807 It is called up by entering a negative integer for NPROF in line 2 1 of the ICF One then goes through the steps of starting a refinement cycle As is stated elsewhere in this Guide the basic command line to run a Rietveld refinement with DBWS is PGM DATA ICF OUT where the actual names are to be entered for the four files indicated PGM is to be the name for the executable version of DBWS being used DATA is to be replaced by the name of either a real data file as it is in an actual Rietveld refinement run or a dummy name ICF is to be replaced by the name of the ICF being converted and OUT is to be replaced by the name wanted for the converted ICF In special cases another file may need to be specified in the command line that relating a separately determined whole pattern background if it is called for in the ICF being converted See C1 of Part 2 of this Guide 24 DBWS 9807a User s Guide 20 8 00 for further comment on these files 2 Line by l
69. ributed by the authors or their institutions for a nominal distribution cost reimbursement fee or no fee A few are commercial products of which most seem to be based on the DBWS programs The DBWS programs have always been and still are distributed without charge for the software A report prepared for the IUCr Commission on Powder Diffraction Smith D and Gorter S 1991 Powder Diffraction Program Information 1990 Program List J Appl Cryst 24 369 402 lists 25 Rietveld analysis programs of which 5 can be run on a PC That report also lists programs for many other kinds of powder diffraction analyses for a total of 280 programs Of the Rietveld analysis programs listed in that report the two most widely distributed at least on the record during the last decade have been DBWS Young et al and GSAS General Structure Analysis System Allen C Larson and Robert B Von Dreele FULLPROF by Rodriguez France and RIETAN by Izumi JaPan are being increasingly widely distributed and cited GSAS runs under UNIX as well as VMS Recently 8 years it has been adapted to run on a PC It is a very comprehensive about 100 000 lines of code now sophisticated and up to date program which can handle most of the calculations one could wish to do in carrying out and assessing the quality of the results of structure refinement with powders or single crystals It can utilize multiple data sets even of different kinds simultaneously The exec
70. ribution package The codeword numbers assigned to the various paraemters varied are shown following table Codeword numbers for parameters in sample used for Figures 1 5 I quartz phase Codeword O 0 JO U RBL N m R ON M N o e R R R R R O 6 JOU R Parameter s Scale factor displacement of sample all background coeff 1 background coeff 2 Scale factor a b c W in FWHM x y of Si x ofO y ofO zofO asymmetry profile NA mixing parameter W U U background coeff 3 B for Si B for O II alumina phase Phase I DBWS 9807a User s Guide 20 8 00 21 So occupancy for O I 22 z of Al II 23 x of O II 24 B overall II Figures 5 show the progressive improvement in the refinement done with a part of the QDATA 20 12 60 10 2 as the zoom feature first revealed the character of the remaining misfit at each stage and then appropriate changes were made in the calculated profile tail lengths and in the number and selection of parameters varied In the process the Goodness of Fit parameter went from being gt 2 0 usually not considered to be very good to 1 40 usually considered to be quite good and R wp went from 18 1 to 13 77 From the data in the heading for each plot and the preceding table of the codewords numbers of the parameters varied in this refinement sequence one can deduce just what was changed at each step The length set for the calculated profile tails
71. s Boulevard Newtown Square PA 19073 3273 USA 58 DBWS 9807a User s Guide 20 8 00 APPENDIX B COPYRIGHT AND FAIR USE Program DBWS 9807 is copyrighted by us as is this User s Guide Permission is hereby granted to individual scientists to use the program for their own scientific work but not for commercial exploitation We have no objection to its being shared with other users meeting the same criteria It may not be incorporated into other products which are then sold nor otherwise used for commercial gain without specific arrangements being made with the copyright holder From those individual users sharing the program with colleagues in the aforementioned non commercial spirit we ask the following 1 Please pass on a copy of the entire distribution package not just a bare bones electronic copy of the program alone This package includes all test case files and a hard copy of this User s Guide The requests that we have received and still receive for missing parts of the package show clearly that many perhaps most people who get copies from their friends are really not getting fully functional copies of the distribution package and will surely have unnecessary trouble trying to use the program 2 So that the recipients may receive our future notices of errors and upgrades please communicate to Prof Young address on cover page a Full name and address of recipient b Program number and version transferred c A listing
72. s chemical symbol and valence which will access the incorporated list of f coefficients taken from the Int l Tables NTYP must start in column 11 T he nominal valence is to be indicated for the X ray case e g Ca 2 but not the neutron case 32 DBWS 9807a User s Guide 20 8 00 X y z fractional atomic coordinates B isotropic atom displacement temperature parameter for the atom So site occupancy It should be 1 000 if all equivalent positions are exactly fully occupied Itis refinable 11 42 16X 5F8 CX CY CZ codewords for fractional atomic coordinates CB codeword for the atom s isotropic thermal parameter CSo codeword for site occupancy 11 43 6F8 Betall Beta 22 Beta 33 Beta 12 Beta 13 Beta 23 anisotropic thermal parameters 11 44 6F8 CB11 CB22 CB33 CB12 CB 13 CB23 codewords for the atom anisotropic thermal parameters 11 5 11 51 2F8 SF scale factor Bo overall thermal parameter 11 52 2F8 CS codeword for scale factor CBo codeword for overall thermal parameter for the phase 11 6 11 61 6F8 U V W CT Z X Y FWHM H parameters for the expression H U tan 6 V tan 8 W CTcot_ for NPROF 0 3 5 6 Hx Utan tan 6 W for NPROF 4 split Pearson VII Z X amp Y are used only when NPROF 7 W hen not used their values must be set to zero 0 00 11 62 6F8 CU CV CW CCT CZ CX CY codewords for the FW HM parameters 11 7 11 71 6F8 A B C cell dimensions a b cin Angstroms
73. terns in the form of d s and I s plus much related information The ICDD offers some incentives for submission of patterns See the ICDD home page at www icdd com Every successful Rietveld analysis results in a powder diffraction pattern which is potentially a source of a high quality d amp I type pattern of a well characterized material A direct listing of the d s and I s as they are normally output from a Rietveld refinement program is not sufficient See below On behalf of the ICDD the DB2dI program has been developed D K Smith and R A Young DXC 1998 to extract d s and peak height I s plus other wanted information from the output and input files of a Rietveld refinement RR made with the DBWS 9411 or DBWS 9807 program DB2dI then to 57 DBWS 9807a User s Guide 20 8 00 arrange all of the data for a pattern in a form ready to go into the PDF 2 file A menu driven program DB2dl uses the calculated pattern to locate each resolvable peak and to determine its position and height It then determines all the contributing reflections to each peak and retains for the final listing the hkl s of the two strongest contributors The DB2dI routine obtains other necessary information from the RR files and from interrogation of the user Once all the necessary information is gathered DB2dI prepares an output file in AIDS format ready for direct submission to Editor Powder Diffraction File International Centre for Diffraction Data 12 Campu
74. ting PLOTINFO and PLOTINFO BIN PLOTINFO can probably be used by a number of plotting programs including the formerly distributed Shareware program SPLOT by Sakthivel and the presently distributed Shareware program DMPLOT DMPLOT3 48 as of this writing by Diduszko and Marciniak PLOTINFO BIN is a binary file tailored specifically for the program DMPLOT which then can display the Miller indices and the patterns of the separate phases as noted above The plots can be produced on screen and dumped to a printer or produced directly on a plotter B Windows environment Although they are DOS programs the executables provided for DBWS can run in the Windows 95 8 DBWS 9807a User s Guide 20 8 00 also 98 and NT environments in a DOS window automatically called up when the program execution starts Advantage can thereby be taken of some of the drag and drop and file handling conveniences of Windows 95 98 NT In combination with a windows based file editor that does not leave tracks i e control characters etc incompatible with program the convenience and ease of operation under Windows can be equal to that to be expected if DBWS were written as a Windows based program Further detailed information about how to achieve this convenience and ease is given in Part 3 C3 III CONTENTS OF THE DISTRIBUTION PACKAGE The files in the distribution package for DBWS 9807a are generally provided on two 3 1 2 inch 1 44 MB diskettes writte
75. tion effects relevant to quantitative phase analysis has been provided for testing and user suggestions for improvements 12 Much of the source code primarily that relating to space groups and symmetry operations and multiplicities has been cleaned in the sense of improved readability consistency in formatting and in use of variable names e g floating point and integer arrays can not use the same name even though used in different places and removal of unused code Additions 13 through 17 below have been made at the suggestion of Fagherazzi s group in Italy who made these and other additions to their copy of DBWS 9006 Riello Canton and Fagherazzi 1995 Riello Fagherazzi et al 1995 We thank them for kindly supplying their code for us to work with 13 An alternative model for profile asymmetry has been added 14 Code has been added for calculation of the background contributions of Compton scattering and disorder diffuse scattering This is a part of their Riello et al physically based background 15 An intensity correction for transmissivity low absorption has been added which may be useful for a flat plate sample mounted in forward transmission geometry 7 DBWS 9807a User s Guide 20 8 00 16 A model data set for some component of the sample may now be imported Tape 11 and have its scale factor refined along with the parameters related to the main data set and the instrument Reillo et al develope
76. to which all data are to be normalized DBWS assigns the weights on the basis of the actually observed not normalized number of counts at a step TAUK Coefficient for deadtime correction for the diffracted beam detector in seconds 3 For each of the NRANGE no of blocks 3 1 6F8 ANGMIN beginning angle for this block in 2 STEP step size in 2 identical for all blocks ANGMAX last angle for this block in 2 STPTIM counting time at each step in seconds OFSTIO dark current counts per step in the incident beam monitor OFSTII dark current counts per step in the diffracted beam detector 3 2 IPTS lines of data in one of the following formats where IPTS is the no of data points i e 24X F7 1X F7 for DATE OCT85 OR FEB86 39 DBWS 9807a User s Guide 20 8 00 51X F10 1X F10 for DATE AUG86 37X F9 0 1X F9 0 for DATE SRS83 29X F9 0 10X F9 0 for DATE SRS91 where I and I are the data entered in F format and ly actual total counts from the incident beam monitor at this step I actual total counts from diffracted beam detector at a given step Also I comes before I in the OCT85 and FEB86 formats and vice versa in the other three Category 3 Fixed_wavelength constant intensity or monitor count neutron beam for example incident and multiple diffracted beam detectors but not all detectors contribute to the count at every step e g instruments DLA and D2B at ILL and the HRNPD at BNL
77. tom In the 9807 the current version N has been replaced by the actual refinable site occupancy So in the last position and the actual site multiplicity M not refinable is called for in the newly created third position The user must supply M It is always an integer value which depends on the particular space group and Wyckoff site The values of M for all space groups and Wyckoff sites are given in the 3 DBWS 9807a User s Guide 20 8 00 International Tables for Crystallography They can also be found in the Inorganic Crystal Structure Database for many of those materials for which the crystal structures are included They can also be deduced from the atom s coordinates and the full set of symmetry operations for the space group Setting flag 9 in line 3 of the ICF will cause the full set of symmetry operators to be printed out Categories 2 and 3 remain unchanged 2 A subroutine for converting the format of an ICF for the 9411 version to the format for the 9807 version is provided It is called up by setting NPROF the profile function identifier in the second position of line 2 of the ICF to any single integer value lt 0 Runnning the program as though one were doing a cycle of least squares then produces the converted ICF under the name specified as the output file in the command line The user has then to fill in correct values for M So and other parameters if any not present in the 9411 version As a further aid to the user t
78. utable codes and necessary files are available by anonymous FTP from machine MIST LANSCE LANL GOV or 128 165 83 145 For user type anonymous and for the password give your e mail address or any other identifier Then go to subdirectory DOS and transfer the main files as binary files GSAS is upgraded frequently FULLPROF is a PC based distant descendant of an earlier DBW program that has been much upgraded and extended FULLPROF is rapidly becoming more and more widely used It B available from its author Dr Juan Rodriguez Carjaval Laboratoire Leon Brillouin CEA CNRS Centre d Etudes de Saclay 91191 Gif sur Yvette Cedex FRANCE It has many extra features including models for crystallographic anisotropy in the reflection profile breadths Dr Rodriguez makes no charge for the program if it is to be used strictly for scientific non commercial purposes The beta version of FULLPROF 98 is discussed in CPD Newsletter No 20 Summer 1998 pp 35 37 It is freely available at ftp charybde saclay cea fr pub divers fullprof 98 windows winfp98 zip A Bucky ball Cso module is available in an earlier version of DBWS from Dr Theo Seigrist 1D 348 AT amp T Bell Laboratories 600 Mountain Avenue Murray Hill NJ 07974 U S A 56 DBWS 9807a User s Guide 20 8 00 There are no doubt many other PC based Rietveld refinement programs available from the authors some but not all based on DBWS For example D Mucha and W _asoc
79. utable version of DBWS to be used e g DB9807 exe 2 the DOS extender file DOSxmfs exe 3 the data file for the particular case e g casel dat 4 the input control file the ICF e g casel in 5 the output file named as wanted e g casel out It can be an empty file at this point and 6 the two DMPLOT files i e dmplot exe and dmplot cfg which you will use to examine the refinement results as you proceed Other files that are needed in this working directory for some refinement options are the amorphous data file Tape Unit 11 for which the scale factor will be refinable and a file containing predetermined whole pattern background data Tape Unit 3 which are to be subtracted from the experimental diffraction pattern data The next step is to click on START lower left corner of your screen and select run That opens a command line labeled open into which you will drag and drop the needed files from the working directory in the sequence program file data file ICF and out file For some refinement options you will have to put a fifth and sometimes a sixth file name on the command line These are the file names for the data in Unit 11 and in Unit 3 described in the previous paragraph If both are used the file containing the whole pattern amorphous component data is entered in the fifth place Be sure to put all of the files that will be needed into the working directory first so you can drag and drop them

USER'S GUIDE to PROGRAM DBWS-9807a RIETVELD

Contents

Download Pdf Manuals

Related Search

Related Contents