I have some x-ray photoelectron spectroscopy data (XPS). Unfortunately, the system under review, has a rather unpleasant background structure. In order to remove the background I have the following approach:
1) I recorded a clean spectrum, just giving me the background 
2) I would fit the clean background spectrum (2 Gaussian + 1 Line)
3) Then I want to remove the clean background from the "real" (i.e. with peaks) data. However, the background changes a little from spectrum to spectrum. Therefore, I cannot just use the fit for the clean background.
Note: The peaks explicitly shall not be fitted in this procedure.
Thanks for your help in advance.