From the readme file.
Lachlan's Lame GSAS Batch Processing Scripts for Mass Unit Cell Refinement.
- 29th November 2003.
Lachlan M. D. Cranswick
Neutron Program for Materials Research (NPMR),
National Research Council (NRC),
Postal Address:
NPMR, NRC,
Building 459, Station 18,
Chalk River Laboratories,
Chalk River, Ontario,
Canada, K0J 1J0 Fax: (613) 584-4040
Tel (work): (613) 584-8811 Office: ext 3719 ; C2 diff: ext 3039
Email: [email protected] WWW: http://neutron.nrc.ca/
Tel (home): (613) 584-4226 WWW: http://lachlan.bluehaze.com.au/
Home Email: [email protected] Mobile/Cell phone: 613 401 3433
These DOS/Windows batch files are optimised for peforming mass structure
based unit cell refinement, by using previous GSAS EXP file and applying
it to the next file. The scripts go inside the EXP file and substitute
in the new file name. It should be relatively easy to change them to do
other things and for other types of analysis.
Be very careful on how you use batch scripting as it is very easy to
do something dodgy that is not caught by blindly running the batch
scripts - affect of spurious peaks or impurity peaks are an example
of this. Routine graphical spot checking at least some of the datasets
at each temperature is important. As well as perusing the GSAS EXP
files to make sure nothing too beserk has happened. Normally, routine
graphing up of the results will show most problems, but it will not
show all problems.
callkall.bat (has list of gsas data files to use when running
the kall.bat file. These can be generated in Excel or
another spreadsheet. Just use a text editor such as PFE
to replace TAB and "0." with spaces prior to running).
kall.bat (runs GSAS using the parameters provided by callkall)
(to work properly with "mtr" and "grep", the first thing it
does is rename the GSAS data file to be within a 8 x 3
filename) mtr is used to substitute the data file name
within the GSAS EXP file.
filter.bat (obtain the unit cell, volume, and offset values with esds
from the GSAS *.lst output file using Grep. kall.bat deletes
the lst file prior to doing a single cycle of refinement, thus
you only have a single cycle of results in the lst file.)
(use an editor such as PFE to get rid of surrounding text
prior to inserting into a spreadsheet)
clean.bat (remove some of the GSAS output files)
cleaner.bat (remove all of the GSAS output files)
Data Collection and Refinement strategy:
One thing to consider is that 2-theta offset should be a constant for a PSD
detector as is being used here (neutron diffraction in capillary geometry).
(and if you are using a good, reproducable scanning point detector)
Thus do the refinement in two stages:
- Collect lots of short datasets at each temperature rather than
a single long one. (data can be merged on the scans where the
sample has reached thermal equilibrium if counting stats need to
be improved). This will allow you to determine when the sample
has reached thermal equilibrium by plotting the unit-cell results
as a function of scan number. You should do this on some initial
expendable samples so you can find out this information in advance.
Consider collecting both going up in temperature and down in
temperature to make sure the data-point at each temperature at
the same (if they are not, you have not reached thermal equilibrium -
or some other effect is going on). If expecting to go through a
non-reversable phase transition, you can using the following type
of data collection strategy:
start at 25 C
up to 40 C
up to 60 C
down to 40 C
up to 80 C
down to 60 C
up to 100 C
down to 80 C
up to 120 C
down to 100 C,
up to 140 C
down to 120 C, etc, etc. (you get each temperature on the up and down)
If time is valuable and you don't want to collect at the same temperature
twice, you can use a "hop-scotch" method. E.g.,
start at 25 C
up to 60 C
down to 40 C
up to 100 C
down to 80 C
up to 140 C
down to 120 C,
up to 180 C
down to 160 C, etc, etc.
- Refine the 2-theta offset with all the other parameters.
Use the varience in the 2-theta offset as a function of
temperature (or whatever you are measuring) to indicate if the
sample is moving in the beam. There will be some scatter due
to correlation between 2-theta offset and unit-cell (especially
with poor counting stats). Be wary that big jumps in 2-theta
could be indicating that your furnace mounts and/or sample are
annealing and moving in the beam.
Graphically spot-check refinements using Brian Toby's EXPGUI
and LIVEPLOT.
- Using the average "2-theta offset" value for the ambient data,
fix the 2-theta offset at this value and re-refine the data.
Graphically spot-check refinements using Brian Toby's EXPGUI
and LIVEPLOT.