Next: 6.2 Further Examples
Up: 6 Examples
Previous: 6 Examples
%------------------% autobk.inp %-------------------% title = Cu 10K, with standard from feff data = cu10k.dat % xmu data output = cu.dat standard = cu-feff.chi % = chi.dat from feff %------------------% end of autobk.inp %-------------------%
The autobk.log you get from this input file should be similar to this:
--------------------- automatic background removal---------------------
autobk: version 2.61 17-jan-1995
-----------------------------------------------------------------------
Cu 10K, with standard from feff
--------------------------------------
input xmu data file name and skey: cu10k.dat ASCII
first document line: Cu foil at 10K
--------------------------------------
input theory chi file name and skey: cu-feff.chi ASCII
first document line: Cu metal chi.dat from feff601.a
--------------------------------------
output chi file : cu.chi
--------------------------------------
----------fitting parameters----------
initial value of e0 = 8977.580078
final value of e0 = 8982.461914
pre-edge range = -50.000000 -200.000000
pre-edge line = -.502631E-03 * Energy + .541864E+01
edge step = 2.302949
energy range = 8977.580078 11362.469727
k range = .000000 24.950001
k weight = 1.000000
fourier transform window:
sills: dk1, dk2 = .000000 .000000
# of knots in spline = 15
background r range = .000000 .981748
the theory was scaled by = .310267
1st shell r range = .981748 2.945243
-----------------------------------------------------------------------
Don't worry too much if there are slight differences in the actual numbers
you get -- precisions vary for different machines. The important numbers
are
(which should be within a few eV of 8982eV), the number of knots
in the spline, (which should be exactly 15), and the upper limit of
R -range for the background, (which should be 0.98Å, the
R -grid point just below 1Å, the default value for
). The
edge step and the slope and intercept of the pre-edge line are given (here
Energy is the total energy, not relative to
, in eV). Output
data files from this running of autobk will be cuk.chi,
containing the
data, and cue.bkg, containing the
data on the same grid as the input data in cu10k.dat.
Figure 1 shows the low energy part of the the original
and the
function. Note that there is a large spike in the
function at very low energy. Such spikes at low-energy are not
unusual from autobk, but do not cause much of a problem. In this case,
the entire spike is below 1
, which does not get very much weight
in a Fourier transform into R -space, and so is not very important
in the fit done by autobk. Such spikes do imply that the autobk
method may not be optimal for getting near-edge structures, but that's
another story. The resulting
data is shown in Fig. 2,
which looks quite good even at low-k , the hard part of background
removal. The Cu data actually extends reliably out to about 22
.