Dear all,
I aim to extract the value of valence band maximum of XPS spectra. Some papers suggest to extract the value by taking the intersection value of conventional linear extrapolation and the baseline. Besides, there is also another possible solution, i.e. by deconvoluting the tail feature with several small width gaussian peaks. I am done with the linear extrapolation and I would like to compare them with deconvolution method. I have tried that with origin by using multiple peaks analyzer and picked several points to be deconvoluted, however, origin did not show a good deconvolution (no gaussian peaks were observed).
Regarding to that issue, does any one have any recommendation out of origin? or perhaps still with origin but different menu option?
Thank you very much in advance.
Best,
Efi
Concerning Origin: if it is still like in the version that I usually used, the multiple peaks fit does not hold your peak width constant which doesn't make too much sense if you acquire XPS width constant analyzer energy (CAE mode), standard. So you would have to define your fit manually in order to secure that all your components have the same width. [If that is different in the most recent version, someone correct me, please.]
Alternative tools would be Igor Pro (I would not recommend that) or Casa XPS (never tried it, but some people seem to be more satisfied with that).
Hello Mr. Weippert..
Thank you for your answer.
I currently use Origin 2017 (the academic version)
Actually what I did was a manual one : Analysis --> Peaks and Baselines --> Multiple Analyzer --> Pick the points to fit manually --> click fit.
However, as predicted, the fitting didn't work well. I was thinking this happened because the feature is a shoulder with tail, so if I aim to perform a good fitting, I would need two gaussian functions (well, the number is a prediction) with different width, instead of several small width of gaussian functions, which I found out in a paper.
Now, I would like to test whether my thought is correct. Yet, I am still curious whether my method is correct or not.
Do you have any suggestions? Thank you very much in advance.
Best,
Efi
The standard models used to fit XPS signal generally involves a convolution of Gaussians and Lorentzians. I might be wrong but setting your fitting algorithm in origin will be not only difficult but also prone to mistakes. And it might result in unnecessary heavy calculations. I personally use CasaXPS which already contains all these built-in functions, and you can easily restrict your fitting parameters (keep all FWHM constant, set your doublet energy step...).
Hope it helps, good luck
- Badges
- Science topic