Applying Absorption Corrections to BT-1 Data

At the NCNR, we usually apply absorption corrections directly to diffraction data. Assuming one knows the neutron wavelength, approximate sample volume, mass and composition of the sample, (you really should know all of these,) the effects of absorption can be computed "good enough for government work".

1) How much absorption do I have?

You can plug the composition, wavelength and density into the BT-1 Compute Neutron Attenuation and Activation calculator (http://www.ncnr.nist.gov/instruments/bt1/neutron.html). Note that incoherent scattering, like absorption, makes neutrons unavailable for diffraction, so this calculator sums both effects and provides a "Linear Attenuation Factor" in units of cm-1.

Multiply this "Linear Attenuation Factor" by the sample radius in cm (common BT-1 can sizes: A: 0.3 cm, B: 0.46 cm, C: 0.54 cm, D: 0.62 cm, E: 0.78 cm). This provides a unit-less constant, "mu.R". When mu.R << 1, absorption is negligible. When mu.R <= 1, absorption will shift atomic displacement parameters ("temperature parameters") somewhat, but corrections are probably not needed. If mu.R > 1, an absorption correction is certainly needed, but a better choice is to do the experiment better: Use a smaller diameter can, or even better use an annular can, to reduce the absorption effects and increase the scattering intensities.

2) Obtain Absorption Correction Factors

Using the "Mu.R" value, one can look up absorption correction factors for the cylindrical geometry (A*). Table 5.3.5B in Volume II of the International Tables for X-Ray Crystallography (p. 295ff) tabulates A* for 0 <= mu.R <= 20, while Table 6.3.3.2 in Volume C of the International Tables for Crystallography (p. 523 in the 1995 ed.) tabulates A* for 0 <= mu.R <= 2.5

It is possible to compute absorption correction factors using the DAVE program, for both cylindrical and annual geometries. See the DAVE documentation for information on how to do this.

3) Fit the A* vs theta values to a Chebyshev polynomial

This is done with a short tcl script, absfit.tcl, which can be downloaded from the NCNR: ftp://ftp.ncnr.nist.gov/pub/cryst/bt1data/process/absfit.zip

Note that the A* values can be loaded from a file that looks like this: 

2.4
35.3, 34.2, 31.4, 27.7, 24.1, 20.9, 18.2, 16.0, 14.2, 12.7, 
11.5, 10.5, 9.76, 9.09, 8.57, 8.17, 7.87, 7.69, 7.64
where the first line contains the Mu.r values and the remain numbers in the file (separated by commas, spaces or new lines) are the A* values at theta = 0, 5, 10, 15,... 90. Press the "Load" button to read a file.

Once Mu.r and A* values are read in or entered, one can fit the values to a polynomial by pressing the "Fit" button. This will create a file, named abs.corr.XXX, with contents like this: 

Mu.R = 2.4, fit by absfit.tcl
6,17.797340,-14.382910,4.586276,-0.044471,-0.709454,0.562135,
note that the quality of the fit can be improved by adding more fit terms.

To run the absfit.tcl script in Unix or Mac OS X (assuming Tcl/Tk is loaded) simply type in a X-term window: 

wish absfit.tcl
In windows, this is a bit more complex. You will either need to establish a shortcut or type as a run command: 
C:\gsas\tcl823\bin\wish82.exe c:\absfit\absfit.tcl

4) Run gformat using the Chebyshev polynomial values

The proprep & gformat programs can be found in ftp://ftp.ncnr.nist.gov/pub/cryst/bt1data/process/. For example, windows executables (windows install notes).

for Windows, place the data files and programs in the same directory and open a "DOS" window (cmd.exe or command.exe) in this directory. For other OS's, place the programs in a diretory in your path and change your working directory to the data file location.

To apply an absorption correction,

  1. first run proprep on the BT-1 file(s): 
    proprep file.bt1
    or 
    proprep file1.bt1 file2.bt1 ...
    This creates .raw versions of the .bt1 files.
  2. Then copy the abs.corr.XXX file to file abs.corr in the same directory as the data files.
  3. Run gformat on the .raw files, using the -a option to trigger the program to apply the absorption correction: 
    gformat -a file.raw
    or 
    gformat -a -s file1.raw file2.raw ...

Neither the author nor the U.S. Government makes any warranty, expressed or implied, or assumes any liability or responsibility for the use of this information or the software described here. Brand names cited here are used for identification purposes and do not constitute an endorsement by NIST.

Comments, corrections or questions: crystal@NIST.gov

$Revision: 1.1 $ $Date: 2004/06/15 19:08:07 $