· The Becker-Coppens formulation of extinction (single crystal case) has beenimplemented and now is under testing.
· FullProf is now able to treat "flipping ratios" (polarized neutrons under magnetic field without polarization analysis, ) as input data. The program may now be used for magnetic form factor refinement using a variety of models: spherical (linear combinations of , , , ), multipolar and orbital occupation. All this part is now under testing and the manual being written.
· A bug concerning the calculation of angles was introduced as a consequence of internal re-arrangements: some angles were not calculated in spite of the conditions selected with the parameters "dmax" and "dangl". This has been corrected in the new release. New information about the distances involved in a particular angle is also written.
· A bug concerning the microstructural analysis was introduced as a consequence of internal re-arrangements in May: the instrumental contribution components of the FWHM was not ordered so the reflections had wrong instrumental part. This bug affected only the cases where the instrumental resolution was read in a separate file. This has been corrected in the new release.
· A bug concerning the re-writing of the PCR file in case of using cubic splines for treating the background is corrected. Now the list of background points is re-written.
· The current version of the program FullProf is 1.9c May2001
· The height of the background points given for linear interpolation can now be refined. The code of each background point appears now besides the actual value in the PCR file. Putting "1.00" as code in automatic mode (AUT=1) is sufficient to make the program select the appropriate code. If the background point is within an excluded region the refinement code will be fixed to zero by the program.
· Due to problems with memory in computers having less than 64Mb RAM, a version of the program compiled with static allocation of arrays is also provided. The executable code of this version is nearly half that compiled with dynamic allocation.
· A 12 terms cosine Fourier series to model the background has been included, according to the expression:
Putting Nba= -4 and 12 background coefficients in the usual place (like for the case Nba=-3) this option is accessed. This option is more robust than the usual polynomial modelling.
· The value of ATZ is now calculated automatically when it is given equal to zero. This value is output to the new PCR file. Multiplicity of each site, content of the unit cell and other useful information for quantitative analysis has been included in the output file CODFIL.out.
· A bug concerning reading background from external file (CODFIL.bac or FILEDAT.bac)has been corrected. Also the Fourier filtering was not working in the last version due to a side effect of dynamic allocation, this has also been corrected.
· Some output formats have been changed to take into account the high resolution of synchrotron powder diffraction machines.
· A bug in the initialisation has been corrected in the version using dynamic allocation. In some cases the use of the same PCR and DAT file in a run (answering OK to the question "Continue with the same data?") gives rise to a huge error because the reflection list was corrupted. Sorry!
· The current version of the program FullProf is 1.9b Mar2001
· The current version can handle up to 16 phases and 7 diffraction patterns.
· Soft constraints on angles are now available (it was disabled in previous versions). The automatic generation of strings containing information about angle constraints in done by using jdist=3 and angl /= 0. See the appendix of the manual.
· In multipattern case there is a new flag array telling to the program if a particular pattern is used or not for the refinement session. This allows to work with the same PCR file without removing information about the patterns.
· It is now possible to use other extension than ".dat" for data files if they are provided inside the PCR file (new multipattern format). The list of files may content the same name. That means all the diffraction patterns to be treated are contained in a single file. This works only for files given for formats INSTR=0,6,12.
· The program uses dynamic allocation for several arrays. In the case of not existing the file called "fullprof.dim" in the current directory, the program search for the environment variable FULLPROF containing the path where it is the executable program and, eventually,the file "fullprof.dim". In case of unsuccessful search for this last file the program adopts default dimensions. An example of "fullprof.dim" file is:
MaxPOINT = 70000 ! Maximum number of data points per pattern
MaxREFLX = 20000 ! Maximum number of reflections
MaxPARAM = 350 ! Maximum number of free parameters
MaxOVERL = 2096 ! Maximum number of reflections contributing to a point
What is important for this file to work properly is the order of items one per line and in the above order) and the integer numbers that should be given between the characters " = " and " ! " (without quotes). A file like the following is also valid
= 30000 !
= 10000 !
= 100 !
= 800 !
The default values (used when unsuccessfull search of "fullprof.dim"):
MaxPOINT = 30000 ! Maximum number of data points per pattern
MaxREFLX = 9000 ! Maximum number of reflections
MaxPARAM = 250 ! Maximum number of free parameters
MaxOVERL = 1024 ! Maximum number of reflections contributing to a point
· The current version of the program FullProf is 1.8a Dec2000
· It can be obtained by direct ftp access from a Web browser by typing:
· The program has been changed internally during the last 10 months without any public information because I had no time for doing that. I'll try to update the information in order to keep informed the users of FullProf of the new capabilities, changes, correction of bugs, projects, etc.
· A new manual is being written. A preliminary version of the appendix of the manual in PDF format is already available. A complete new manual is expected to be totally prepared by the end of january 2001. The user may consult the document "App_fp2k_man.pdf" for details concerning the changes described below.
· The most important changes made on the code of FullProf.2k are:
1. The old code generating symmetry operators from the Hermann-Mauguin symbols and the code generating reflections has been totally removed. This code was written many years ago by Alen C. Larson. Everything has been substituted by Fortran 90 modules, developed by myself and Javier Gonzalez-Platas, that are gathered in a crystallographic library. The source code of this library will be in the public domain under a GNU licence after its publication. The consequences of this change is that repetition of some equivalent reflections in rhombohedral and trigonal groups are no more happening. The generation of symmetry operators can be made using Hall symbols, Hermann-Mauguin symbols or a list of generators. At present only the notation of H-M and Hall (preceded with the keyword HALL) can be used. In case of troubles (messages like: Sum(Iobs)=0, means that the symbol is wrong) try removing the comment after the space group symbol ("<- Space...").
2. The calculation of distances and angles, as well as bond valence sums can be done automatically, without using external programs (e.g. BONDSTR). This is obtained using JDIST=3,4. Output files with extension "dis" contain all the relevant information for the different phases. Output files of names <dconstr"n".hlp> (n stands for the number of the phase) contain lines that can be directly pasted to PCR files for soft constraints.
3. An option for helping to solve crystal and/or magnetic structures has been included. This is a simulated annealing module able to handle two types of algorithms. This work, at present, only with integrated intensities. The user must create first the corresponding integrated intensity files using FullProf in profile matching mode using the option (Jvi) JVIEW=11 and adjusting RMUB and RMUC. The Simulated Annealing option is accessible using ICRYG=3.
4. The handling of refinement codes has been changed. Now the user may select between the traditional mode changing by hand the numbering of codes or select the AUTOMATIC mode. This last mode works with the flag AUTO=1. In this case the user may put by hand the codes for making constraints as usual, and just put "1.00" to inform the program that the corresponding parameter will be refined. The program will attribute automatically the codeword. In the automatic mode there is no "holes in the matrix" and the number of refined parameters may be different from the specified by the user in the corresponding line. The automatic mode is useful when one has to fix a parameter in the middle of many others: just put the codeword (including the multiplier) to zero. Sometimes the message "hole in the matrix" still appears. In such cases you have just to increase artificially the number of parameters to be refined or just put it equal to zero, or, in the worst case, suppress a large number of codewords (just leaving the multiplier 1.0).
5. The profile function NPROF=7 has been changed in order to include the Larry Finger treatment of the axial divergence. The two additional parameters (S_D,S_L) are included together with the asymmetry parameters that may still be used in combination with S_D and S_L. If strain models are used an additional line must be included for the lorentzian component of the strain broadening.
6. An optimisation mode has been introduced in order to speed up the calculations. Depending of the kind of work the user is performing, the program will be able to accelerate the calculations or not. This mode suppress the calculation of the Bragg-R factors for each cycle. If something is getting wrong suppress the optimisation mode... this means the modifications are not ready for the particular problem the user is handling.
7. New models for strain and size broadening have been included. If ISTR=1 the notation of Peter Stephens is used as default. The parameter IstrainModel is just equal to the number of the Laue class. A reorganisation of the old models have been made. New models for size broadening including spherical harmonics have been introduced.
8. A new mode for refining magnetic structures has been included. It may be accessed using ISY=-2. The free parameters are then the coefficients of the basis functions of the irreducible representations of the propagation vector group. This mode can also be used with the simulated annealing optimisation.
9. A new file containing information about the microstructure is output. This happends only if the user provides a file containing the instrumental resolution function. The use of the profile function NPRO=7 is imperative. This works only for constant wavelengths modes at present. The information is, at present, quite limited. This is still under construction.
10.A new way of calculating the T.O.F. as a function of the d-spacing proposed by Jason Hodges (Argonne) has been implemented. Improvements in the algorithms for calculation of the peak shape function have also been introduced.
11.For single crystal work a new parameter for the contamination with can be refined.
13.A general formulation for peak shifts due to defects or to residual stresses have has been implemented.
14.In the present version of FullProf the comments starting with "!" appear automatically after re-writing the PCR file, even if the keyword COMM is absent. In fact, suppressing the keyword COMM produces a PCR output with a format similar to that used for the treatment of multiple diffraction patterns. PCR files starting with COMM and used for a single pattern retain its traditional format after re-writing.
15.ILOR=-1 means that the Lorentz-polarisation correction is not applied. This supposes that the profile has been corrected for LP.
16.The extension of data files are no more restricted to be ".dat". If the user provides another extension the program is able to work
17.The option concerning asymmetry correction, applied when NPHASE is negative, following the approximation of C.J.Howard, J.Appl.Cryst.,15. 615-620 (1982) with the Simpson formula for five points, HAS BEEN SUPPRESSED. The existing asymmetry corrections are largely more adequate.
18.The Symmetry Adapted Spherical Harmonics for form-factors are now fully implemented. The case of C60 is provided in the examples file using cubic harmonics.
19.The user must provide the indices of the harmonics to be used according to the symmetry of both the site and the molecule.