I think the references given by Jeffrey Weimer in this old discussion contain what you're looking for:

https://www.researchgate.net/post/In-high-resolution-xps-spectra-of-Si-2p-can-Si2p3-2-have-less-concentration-than-Si2p1-2

Thank you Jürgen Weippert for pointing me to the discussion by @Jeffrey Weimer. It partly answers my question.

If there are two materials existing together (Si and SiO2), it makes sense to deconvolute Si2p core spectrum into two different peaks. What if Si exists in the SiC3N environment, and does it make sense to deconvolute it to Si-C and Si-N peaks? Or just consider it as one peak and assign it to a mixed chemical environment?

You can check what should be applied by looking at the width of the signal: if the width is the same as for pure SiC or Si3N4 measured with the same instrument at the same parameters, then you have one single component and you can say that it's a homogenous mixture. If it's wider than that, you have multiple components and then deconvolution can start making sense.

Hi Jürgen Weippert ,

I appreciate your idea and contribution to this discussion.

I did acquire C 1s, Si 2p, and N 1s high-resolution spectra for silicon carbide (SiC) and silicon nitride (Si3N4) materials as relevant. The FWHM for the C-Si (283.2 eV) component in C1s and the Si-C (100.7 eV) component in Si2p spectra of SiC were found to be 1.42 and 1.73, respectively. For Si3N4, the FWHM for the N-Si (398.1 eV) component in N1s and Si-N (101.3 eV) in Si2p spectra were determined to be 1.49 and 1.95, respectively. The spectra were calibrated to the Ar2p 3/2 component at 241.9 eV.

After comparing the spectra, I fit only one component in Si2p spectra for amorphous silicon carbonitride due to similar FWHM, as obtained for SiC and Si3N4, and ascribe it to mixed environments.

Thanks,

Sai

As a late note to this discussion ...

As you have done, obtain reference spectra for the appropriate carbides and nitrides for silicon. You might also explore making taking reference spectra from physical mixtures of the two compounds at selected molar ratios, for example 1:9, 1:1, and 9:1. Refer also to extensive databases and journal articles that have published XPS spectra on these compounds. Finally, recall that peak height, position, fwhm, and peak shape all contribute in various ways to the confidence we can have when we try to resolve individual component peaks from a spectral peak. In this regard, one can never be faulted for calling one broad peak as an unknown mixture of two components that cannot be resolved, whereas one should always be prepared to justify from objective physical evidence the confidence one has to use two or more component peaks to fit one broad peak.

Hope this helps you complete your study.