Full Year 2006
----------------------
>> 27 November 2006: Corrections for multiple patterns
----------------------
- The program was not working with the maximum number of patterns (50) available
in the previous distributed version. Errors reading the contribution of the
phases to the different patterns occurred because the length of the string for
reading was too small. Moreover the calculation of reduced chi2 of each pattern
was also wrong because the number of degrees of freedom was sometimes negative
(the division of the number of free variables by the number of patterns was not
done in some places). These bugs have been corrected in the current version.
- Help for debugging: when a problem reading the PCR file is found and no message
is given for some reason, the user can force the program to write the read lines
of the PCR file as soon as they are read by including the keyword "print_pcr"
(all in lower case) within the global title (first line of the PCR file). When
the program encounter this keyword it outputs the read lines in a file called
"debug_PCR.lst". The content of this file may help to determine the place where
the problem is found.
- The current version of FullProf has been updated to:
**********************************************************'
** PROGRAM FullProf.2k (Version 3.76 - Nov2006-ILL JRC) **'
**********************************************************'
----------------------
>> 6 October 2006: Effective moments for incommensurate structures
----------------------
- The nature of the variables used as components of Fourier Coefficients
of magnetic moments can now be tuned by the user. In the normal mode of
operation the variables (R, I) used for describing incommensurate magnetic
structures correspond to "amplitudes" (Remember the factor 1/2 in the
expression of Fourier coefficients: Sk = 1/2 (Rk + iIk) exp{-2pii Phk}) .
The fact that a propagation vector of type k=1/2H is equivalent to -k
means that there is a single term in the Fourier series for calculating
magnetic moments (Sk = Rk) and magnetic reflections are generated using
a single satellite per nuclear reflection. If k is slightly different from
half a reciprocal lattice vector then k is not equivalent to -k, so that
two satellites are generated per nuclear reflection. This produces a
discontinuity (by a factor of sqrt(2)) in the values of the amplitudes
when a commensurate-incommensurate phase transition occurs. In order
to avoid this, the user can force the program to use "effective moments"
that correspond to the values of the magnetic moments in the case of
k=1/2H. In such a case, if the difference k-1/2H is greater than 0.00001,
one can put Nvk=-1, for instance when ky=0.49993, and the values of the
components of Fourier coefficients are the same as those obtained
when we put Nvk=1 and ky=0.5. For using "effective moments" the user
should write the keyword EFF_MOMENT within the commands lines associated
to the corresponding magnetic phase in which Nvk=-1.
Example:
. . . . . . . . Relevant portion of the PCR file
Magnetic phase name
!
COMMANDS
eff_moment
END COMMANDS
!Nat Dis Mom Pr1 Pr2 Pr3 Jbt Irf Isy Str Furth ATZ Nvk Npr More
3 0 0 0.0 0.0 1.0 1 -1 -1 0 0 0.000 -1 7 0
. . . . . . . .
- The current version of FullProf has been updated to:
**********************************************************'
** PROGRAM FullProf.2k (Version 3.75 - Oct2006-ILL JRC) **'
**********************************************************'
----------------------
>> 3 October 2006: Magnetic domains in single X-tal refinements
----------------------
- The presence of magnetic domains of any type can now be handled easily
with the current version of FullProf.
Within the commands associated with a particular phase the new keyword
MAGDOM has been created. To introduce a set of domains one has to write
a line per domain with the keyword MAGDOM (including the domain 1 with
identity matrix) followed by rotational operators, a real number (as
for a magnetic symmetry operators used in MSYM) a colon symbol and the
values and refinement codes for the populations.
Two artificial examples are given below
. . . . . . . . Relevant portion of the PCR file
My Magnetic phase name
!
COMMANDS
magdom u, v, w, 0.0 : 0.650 121.00
magdom -v, u, w, 0.0 : 0.250 131.00
magdom u,-v, w, 0.0 : 0.100 0.00
END COMMANDS
!Nat Dis Mom Pr1 Pr2 Pr3 Jbt Irf Isy Str Furth ATZ Nvk Npr More
2 0 0 0.0 0.0 1.0 1 4 -1 0 0 0.000 -1 0 0
. . . . . . . .
In the example above there are: three domains in total.
Notice that the first domain corresponds to that of the magnetic model
described in the PCR file. The sum of the populations are always 1.0
The refinement code of the last domain should always be zero.
It is calculated as a function of the previous values in order to ensure
that the sum is always 1.
. . . . . . . . Relevant portion of the PCR file
My Magnetic phase name
!
COMMANDS
magdom u, v, w, 0.1 : 0.250 0.250 121.0 131.0
magdom u,-v, w, 0.1 : 0.200 0.300 141.0 0.0
END COMMANDS
!Nat Dis Mom Pr1 Pr2 Pr3 Jbt Irf Isy Str Furth ATZ Nvk Npr More
3 0 0 0.0 0.0 1.0 1 4 -1 0 0 0.000 -1 0 0
. . . . . . . .
In the example above there are: two domains + two chirality domains, four
domains in total. Notice that the first domain corresponds to that of the
magnetic model described in the PCR file. A value of the real number following
the operator greater than 0.001 means that we have chirality domains (second
population value and second refinement code). The sum of all populations should
be equal to 1.0 and the last one should not have a refinement code.
The rotational operators should be those of the paramagnetic group that are
lost in the transition to the ordered state. Be careful with trying to refine
all domain populations. There are situations where two domains give exactly the
same module of the magnetic interaction vector so that the relative population
cannot be determined using non-polarised neutron diffraction.
- A bug (sign) in the derivative with respect to one imaginary component of magnetic
Fourier coefficients in the case of using Jbt=1 (but not for Jbt=-1) has been
corrected. The bug was producing poor convergence in case of refinement of "Iy".
----------------------
>> 18 September 2006: Lambda/2 in single X-tal refinements
----------------------
- Changes in the code concerned with the contribution of Lambda/2 harmonics
in single crystal refinements have been introduced. Now the contribution
of Lambda/2 is explicitly written in the output file and the forbidden
reflections are not eliminated from the refinement when (2h, 2k, 2l) is
allowed. The flag "Excl" is still written but it refers to (hkl) and not
to (2h,2k,2l).
- A bug introduced in the latest versions of Bond_Str, when using disordered
compounds has been corrected. Now the bond valence of a chemical species,
occupying partially a site is correct.
- The current version of FullProf has been updated to:
**********************************************************'
** PROGRAM FullProf.2k (Version 3.71 - Sep2006-ILL JRC) **'
**********************************************************'
---------------------
>> 20 August 2006: Simulated annealing using profile intensities
---------------------
- A new cost function, using the profile intensities, for simulated annealing has been
included in FullProf. The program generates a new output file after a Le Bail fit
when global Ipr=-1 and Cry=0 with extension ".spr" together with the integrated
intensity file (*.int) when More=1 & Jvi=11. In such a file the profile intensities
of selected points together with the profile (and other constants) contributions
of each reflection of the proper phase are written and will be used to calculate
the new cost function for subsequent simulated annealing jobs
To run a simulated annealing (Cry=3) job, the user should make the usual PCR file
and he(she) has just to put Ipr=-1 if he(she) wants to use the new cost function.
function. The new cost function corresponds to the profile Rp (unweighted) of selected
points. When there is only one phase, then the quantity "yraw(i)-bac(i)" is used as
observed profile intensity (yobs(i) in the SPR file). If the Le Bail fit is used for
a second (or higher) phase (magnetic contribution, for instance) then "yobs(i)"
is taken as the "calculated profile" of the corresponding phase.
It is expected that this function improves the success ratio for structure solution
because it contains more information than the clusters of integrated intensities.
---------------------
>> 27 July 2006: Simulated annealing using rigid body and angles in degrees
---------------------
- A bug in the output of the solution found by simulated annealing when using
a rigid body description and angles in degrees has been corrected. The values
of the angles were multiplied by the factor converting radians to degrees twice
before the output to the new PCR and "*.so1" files. This has been corrected.
There was no problem when using radians.
Up to now the limits of the angles had to be given in radians even if the angles
were provided in degrees, this has been changed to be consistent with the global
description using degrees. The program uses internally radians as the unit for
angles but this is transparent for the user. Comments on that are now output in
the OUT file. For working properly the name of the Euler angles should contain
the string "hRGB" like "ThRGB_NC1","PhRGB_NC1" and "ChRGB_NC1". This is attributed
automatically by the program, but for starting the Simulated Annealing work one
has to take care of putting that at the end in the lines of the box limits.
---------------------
>> 24 July 2006: Additional background profile code-names provided in the PCR file
---------------------
- The codes of the files containing background profiles can now be arbitrary.
Remember that to access this option one has to put MORE=2,3 in the same line
as the ZERO point appears.
According to the note of 28 July 2003, the only accepted names were of the
form: "filedat_n.bac". Where "filedat" was the code of the data file corresponding
to the diffraction pattern to be refined and the index "n" was the number of the
contributing profile. This is still true by default. Otherwise the user can provide
the individual codes of the contributing profiles as given in the following example:
........
!
! Number and coefficients of profiles contributing to background of Pattern# 1
3 1.2310 0.3310 2.3133
51.0000 61.0000 71.0000
! Codes of files containing additional background profiles
myfile1 qwerty21 file33389
............................
The program read the line following the codewords of the refined weights and if
it finds the word "Codes" then the codes of the files are read in the next line.
In the above example three files containing background profiles are read. The
complete names of the three files are: myfile1.bac, qwerty21.bac and file33389.bac
The maximum length of the codes for this kind of files is of 9 characters.
The restrictions described in the note of 28 July 2003 (format of files and number
of points) are still valid.
---------------------
>> 22 July 2006: Treatment of Single Crystal Electron Diffraction Data
---------------------
- The current version of FullProf has been updated to:
**********************************************************
** PROGRAM FullProf.2k (Version 3.70 - Jul2006-LLB JRC) **
**********************************************************
- The new version of FullProf can handle automatically data taken form electron
diffraction patterns provided they are close to kinematical scattering. The
access to this option is exactly the same as for treating single crystal X-rays
or neutron data except that the parameter Job should be equal to 4 (Job=4 or Jtyp=4).
The program applies kinematical theory using the Mott-Bethe formula for calculating
the scattering factor of the atoms or ions. If the parameter Cry is put equal to
4 (Cry=4) the program simulates a single crystal electron diffraction pattern using
the provided wavelength and the values of 2theta-init (Thmin) and 2theta-end (Thmax)
given in the PCR file (see note of 10 May 2003 for details).
For refining of solving crystal structures by simulated annealing, experimental
data should be given as structure factors squared. It is supposed that the program
handling the diffraction images makes the proper corrections in order to obtain
h,k,l,F2 and sigma(F2).
---------------------
>> 20 July 2006: Refinement of magnetic structures with SXtal data and magnetic groups
---------------------
- The refinement of magnetic structures with k=0 using time reversal operators for
single crystal together with the crystal structure is now possible without skipping
the systematic extinction due to the crystallographic space group.
Up to now putting JBT=10 and using time reversal operators was impossible due to the
fact that the program suppressed automatically the reflections of the list that were
forbidden for the space group (but not for the magnetic group). Now the reflections
are not suppressed and only the magnetic contribution is calculated. These reflections
appear in the output list with Icod=-1.
---------------------
>> 30 May 2006: The official Web site of the Fullprof Suite
---------------------
- The new official Web site for the FullProf Suite has been installed in the
web page of the Diffraction Group at the Institute Laue-Langevin.
The page is:
http://www.ill.fr/dif/Soft/fp/
All previous web sites (not mirroring this site) contain non updated versions
of the programs. I cannot access anymore the CEA site and I know that some
files have been corrupted in particular the GBasIreps program for windows
is not working.
Please inform me (fullprof@ill.fr) of possible errors in using the programs
distributed from this page.
---------------------
>> 25 May 2006: Another way of using quadratic form of anisotropic size
---------------------
- It is now possible to use a quadratic form for lorentzian anisotropic size
broadening (HL = FZ(h,k,l)/cosTheta), as defined by the expressions below
cte_siz=0.001*360*lamda/pi^2
FZ(h,k,l,...)= cte_siz*1000/D, D in same units as Lambda
FZ(h,k,l,...)= cte_siz*d(hkl)^2*(c1 h^2+ c2 k^2+ c3 l^2+ c4 2hk+ c5 2hl+ c6 2kl),
for whatever case of strain model.
Up to now the use of this expression in microstructure refinement was limited
by putting Str=2 or 3, now the only condition is to put Isizemodel=30. In
particular it is possible to use spherical harmonics (Str= 4 + StrainModel) or
the quartic form for strains (Str= 1 + StrainModel) together with the quadratic
form for size broadening (Isizemodel=30). Up to six coefficients can be refined.
An example of PCR file in which appear the size model 30 is given below:
................................................................................
! Scale Shape1 Bov Str1 Str2 Str3 Strain-Model
10.000 0.00000 0.00000 0.00000 0.00000 0.00000 2
11.00000 0.000 0.000 0.000 0.000 0.000
! U V W X Y GauSiz LorSiz Size-Model
0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 30
0.000 0.000 0.000 0.000 0.000 0.000 0.000
! a b c alpha beta gamma #Cell Info
4.871308 2.979645 9.287657 90.000000 95.000000 90.000000
0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
! Pref1 Pref2 Asy1 Asy2 Asy3 Asy4 S_L D_L
1.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
! Generalised size parameters (quadratic form):
! h2 k2 l2 2hk 2hl 2kl
0.52036 2.4900 0.069229 0.0000 0.0000 0.0000
231.00 241.00 121.00 0.00 0.00 0.00
! Y00 Y20 Y22+ Y22- Y40 (Strain)
88.000000 338.000000 117.000000 0.000000 -10.000000
61.00 51.00 71.00 0.00 41.00
! Y42+ Y42- Y44+ Y44- (Strain)
11.000000 117.000000 0.000000 0.000000
31.00 21.00 0.00 0.00
! Lorentzian strain coeff.+ code
0.20000 0.00000
.................................................................................
---------------------
>> 24 May 2006: A new option of anisotropic strains using spherical harmonics
---------------------
- It is possible now to use spherical harmonics (SH) to handle anisotropic strains.
Up to now the use of SH in microstructure refinement was limited to the use as
modelling for anisotropic lorentzian size effects. We have included an option
putting Str=4 and StrainModel=n, where n=1,2,-2,3,4,5,6,7,8,9,10,11,12,13,14.
The value of n corresponds to the Laue classes:
1 2 -2 3 4 5 6,8 7,9,10 11 12 13 14
-1 12/m1 112/m mmm 4/m 4/mmm -3 -3m 6/m 6/mmm m3 m3m
The spherical harmonics are referred to a cartesian frame where the z-axis
is always along the principal symmetry axis. In the case of trigonal system
the hexagonal setting is used and "c" is along the z-axis. The order of
the harmonics has been arbitrarily limited as it is the case for size broadening.
This may be changed in the future.
The input is similar to that used for S_HKL coefficients except that for some
cases the number of parameters is different. There are n1 and n2 values to
be read in different lines. The codewords are always below the corresponding
parameter. After the lines containing the coefficients of the spherical
harmonics an additional line containing the degree of Lorentzian character
of the strain, together with its codeword, is given.
(n1,n2)
( 5, 1) (-1) : coeff. Y00,Y20,Y21+,Y21-,Y22+,Y22- (6 parameters)
( 5, 4) (2/m) : coeff. Y00,Y20,Y22+,Y22-,Y40,Y42+,Y42-,Y44+,Y44- (9 parameters)
( 6, 0) (mmm) : coeff. Y00,Y20,Y22+,Y40,Y42+,Y44+ (6 parameters)
( 4, 4) (4/m) : coeff. Y00,Y20,Y40,Y44+,Y44-,Y60,Y64+,Y64- (8 parameters)
( 4, 2) (4/mmm) : coeff. Y00,Y20,Y40,Y44+,Y60, Y64+ (6 parameters)
( 4, 1) (-3) : coeff. Y00,Y20,Y40,Y43+,Y43- (5 parameters)
( 4, 3) (-3m) : coeff. Y00,Y20,Y40,Y43-,Y60,Y63-,Y66+ (7 parameters)
( 4, 2) (6/m) : coeff. Y00,Y20,Y40,Y60,Y66+,Y66- (6 parameters)
( 4, 1) (6/mmm) : coeff. Y00,Y20,Y40,Y60,Y66+ (5 parameters)
( 5, 0) (m3) : coeff. K00,K41,K61,K62,K81 (5 parameters)
( 4, 0) (m3m) : coeff. K00,K41,K61,K81 (4 parameters)
For instance a case corresponding to the Laue class 2/m is introduced in the PCR
file (remember that Str=4) as follows:
................................................................................
! Scale Shape1 Bov Str1 Str2 Str3 Strain-Model
10.000 0.00000 0.00000 0.00000 0.00000 0.00000 2
11.00000 0.000 0.000 0.000 0.000 0.000
! U V W X Y GauSiz LorSiz Size-Model
0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0
0.000 0.000 0.000 0.000 0.000 0.000 0.000
! a b c alpha beta gamma #Cell Info
4.871308 2.979645 9.287657 90.000000 95.000000 90.000000
0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
! Pref1 Pref2 Asy1 Asy2 Asy3 Asy4 S_L D_L
1.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
! Y00 Y20 Y22+ Y22- Y40 (Strain)
88.000000 338.000000 117.000000 0.000000 -10.000000
61.00 51.00 71.00 0.00 41.00
! Y42+ Y42- Y44+ Y44- (Strain)
11.000000 117.000000 0.000000 0.000000
31.00 21.00 0.00 0.00
! Lorentzian strain coeff.+ code
0.20000 0.00000
.................................................................................
---------------------
>> 21 April 2006: Some new options for single crystal
---------------------
- A version of FullProf without the correction of the bug described in the
note of 21 October 2005 has been distributed inadvertently.
- A new option treating twinned crystals has been introduced. For using it
the variable "N_Domain" appearing only when More=1 (in the line after the
name of the phase) should be put equal to the number of domains (number of
twins). In such a case instead of using a scale factor for each domain (as
it is the normal option) a single scale factor is used and the fraction of
N_Domains-1 can be refined. The fraction of the last domain is calculated
with the restriction that Sum(fract)=1.0.
This option is to be used compulsory when a heterogeneous (multidetectors)
data collection has been performed in a twinned crystal and one wants to
use all data simultaneously.
Example: Portion of a PCR file using this new option
!-------------------------------------------------------------------------------
! Data for PHASE number: 1 ==> Current R_Bragg for Pattern# 1: 9.02
!-------------------------------------------------------------------------------
BiSrMnO3 a
!
!Nat Dis Ang Pr1 Pr2 Pr3 Jbt Irf Isy Str Furth ATZ Nvk Npr More
14 0 0 0.0 0.0 1.0 0 4 0 0 0 3085.695 0 7 1
!
!Jvi Jdi Hel Sol Mom Ter Brind RMua RMub RMuc Jtyp Nsp_Ref Ph_Shift N_Domains
0 0 0 0 0 0 1.0000 1.0000 0.0000 0.0000 1 0 0 6
!
I b m m <--Space group symbol
!Atom Typ X Y Z Biso Occ In Fin N_t Spc /Codes
Bi BI 0.00031 0.00000 0.25000 2.50231 0.35000 0 0 0 1
71.00 0.00 0.00 81.00 0.00
..........................................................................
O1 O 0.04823 0.50000 0.25000 3.61239 0.50000 0 0 0 5
161.00 0.00 0.00 181.00 0.00
O2 O 0.75000 0.25000 0.04333 3.09902 1.00000 0 0 0 5
0.00 0.00 171.00 191.00 0.00
....................................................................
! Scale Factor + Domain Fractions
! Scale Dom1 Dom2 Dom3 Dom4 Dom5 Dom6
64.26 0.136345 0.196609 0.181173 0.171022 0.178615 0.136237
11.00 21.00 31.00 41.00 51.00 61.00
..........................................................................
---------------------
>> 6 February 2006: A new version of WinPLOTR for Windows, Linux and MacOS X
---------------------
- A complete re-writing of WinPLOTR has been undertaken in order to make it
portable to other platforms. We have included a beta version of the program
WinPLOTR-2006 in the present distribution of the FullProf Suite. The windows
version of WinPLOTR (written using RealWin) will continue to be distributed
waiting for the availability of all the functions in the new WinPLOTR-2006
(written using Winteracter). Our purpose is to maintain a unique version of
the program that can be run in all common platforms.
The new WinPLOTR-2006 has been developed thanks to the stay of Oscar Baltuano
(from the Peruvian Nuclear Energy Institute: IPEN) at the LLB.
The new WinPLOTR-2006 has new capabilities, absent in the windows version of
WinPLOTR, like the possibility of visualising 3D surfaces using OpenGL. The
preferences can be totally accessed interactively and saved. For the moment
no manual is available. Users of WinPLOTR will find quite easy to use this
new version of the program.
Some comments and hints for using the program are given below:
1: The first time the program is run, a message telling the user that no
"preferences" file is available appears. The user can save the preferences
(Menu: Options) concerning all the aspects of WinPLOTR-2006 either globally
(in the FullProf Suite directory) or locally (in the current directory).
2: The automatic peak and background options (Menu: Calculations) must be first
made "enabled" before using them. After finishing the session of peak/background
search the user should made the option "disabled" in order to do other things.
3: Open data files can be done by selecting several files simultaneously provided
that all of them have the same format. They can be opened in an order that is
not under the program control. The user can re-arrange the files in the appropriate
way by selecting "Sorting" in the menu "Profiles". Once the user is satisfied
with the ordering, the list of opened files can be saved as a buffer file that
can be directly read by WinPLOTR-2006 in subsequent sessions.
4: When using the OpenGL visualisation of a surface representing powder diffraction
patterns as a function of temperature or whatever, the user may see some glitches
or holes in the surface if the data are noisy. This can be repaired by changing
the scale factor along X and Z and/or using a low pass filter (both in the menu
"Processing" of the OpenGL window). Changes in colour of the surface or in the
orientation of the illuminating light can also be performed by selecting the
appropriate menus. In some PCs with old graphic cards this option may not work
properly.
5: The most common programs of the FullProf Suite can be launched from the appropriate
button or menu from WinPLOTR-2006.
6: The fitting procedure is not yet implemented
The MacOS X version of the whole FullProf Suite will be available soon.