I have some XPS spectrums that I am trying to fit (my first time doing so), using XPSpeak. I understand that for spin-orbit splitting the FWHM, line shape (i.e.% gaussian/lorentzian) must be equal (more or less), peak area ratios set (i.e. 2:3 for 3d3/2 and 3d5/2), and the peak separation (relatively) constant. My question is, how much can the FWHM vary from different literature sources? The area I am fitting is the Mo 3d area, if I contrain the FWHM to the values I found at www.xpsfitting.com I get a pretty poor fit. However, if I only constrain the peak FWHM's to be equal (for each doublet), and not explicitly specifiy; I get a good fit. For example, one doublet is the MoO3 3d. The FWHM (Mo 3d3/2) specified at the website I previously mentioned is about 0.86. If I only specify that the doublet must have FWHM invariance (and not explicitly 0.86) the FWHM becomes 2.82 after optimization.