"Data Interchange and Developments in Process
Illustrated by Peak Synthesis Analysis:
Do you need anything else?"
Dr. Martin Seah, NPL, Teddington
Dr Seah showed some recent research from NPL, on the accuracy of
commercial software used for peak synthesis in XPS. The work has been made possible due to
the development of the VAMAS data transfer format. The work focussed on the following
experiments on a set of synthetic spectra:
 |
Gaussian-Lorentzian sum or product function
peaks
|
|
Effectiveness of Shirley background subtraction
|
|
Effectiveness of Tougaard background
substraction
|
|
Peak synthesis with partially resolved peaks
with no background, &
|
|
|
| |
 |
| |
Figure 1:
Comparison of peak fitting results generated by different software packages |
| |
|
Synthetic spectra approximating the C 1s peaks for poly(methyl
methacrylate), poly (vinyl acetate), poly (vinyl chloride) and poly (isobutylene) allowed
the analysis of the accuracy of the data when the spectra were processed using the
different manufacturers software. One set of spectra was constructed with Gauss-Lorentzian
product function peaks and the second with Gauss-Lorentzian sum function peaks, since each
software package was designed for analysing only one of these options.
The fitting of individual peaks is only good for with the software
using the same peak shapes as those of the data. The fit quality will deteriorate markedly
for fitting a shape different to that used in the peak-fitting algorithm. For unresolved
peaks additional information is required to provide meaningful intensity information. The
results obtained from the three commercial systems are shown below in a graph of error
and their standard deviation for the peak intensities versus the percentage error in the
peak intensities.
The three software packages studied all have different facilities but
from the analysis of the data, that the general order of preference in the fittings
algorithms was:
B, A & C
Other aspects of the software packages can favour software package A,
which had a higher level of functionality. For more details see the following paper:
Journal of Electron Spectroscopy and Related Phenomena, 95(1998)71
|
