"Calculating XPS Peak Shifts:
Is it Easy and is it Reliable? "
Dr. Chris Somerton, Sowerby Research Centre, BAe
| |
 |
|
 |
|
Figure 1:
Examples of C 1s spectra |
For simple systems such as reference materials, its relatively easy to
breakdown the XPS spectra into there individual peak component , but the analysis of real
systems is difficult, particularly if they have been exposed to harsh environments.
So how can the analyst enhance the level of information he gives his
customer? The answer is to predict from first principles, the binding energy of the
individual peak components. Fine, but how?
Below is the screen from the Molecular simulation software Cerius2,
which can perform semi-empirical and ab-initio calcalutions, having first defined the
molecules of interest. The answer is:
 |
Determine the energy difference between the
initial and final states
|
|
Principle of equivalent cores.
|
 |
Figure 2:
Screenshot from Cerius2 molecular simulation software |
|
|
How does this work?
The calculations take between 30 seconds and 3 hours using a Silicon
Graphics workstation (R10000, O2 with 500mb of RAM).
How accurate is the data you obtain from this method of analysis? The
best way to test the accuracy is to compare the results against a well established data
base such as that provided by Dave Briggs and Graham Beamson.
Dr Somerton used the systems shown below, using the energies given in
the database. Comparing the two sets of results he obtained the following data
correlation.
The results (below) are very encouraging, although it should be
stressed that there is a zero error. Nevertheless, the results are encouraging and
indicate that the two techniques can be used together in order to give a more detailed
description of the data
|
| |
|
Figure 3:
Comparison of theoretical and experimental results |
| |
|
