If you have literature references for similar particles with oxidation state assignments, in the best case with a detailed description of their data evaluation process, use the same substrate in order to ensure comparability.
Otherwise, a decently conducting substrate which forms as little chemical bondings as possible to your particles would be an obvious choice. Among metals, Au(111) would be an option since gold doesn't voluntarily bind oxygen. If you don't want to measure on metal, you might consider graphite(001), but if you deposit the particles from solution/dispersion, there is of course always a risk of intercalation and exfoliation.
The answer is here https://www.researchgate.net/post/What_kind_of_substrate_would_you_suggest_for_an_XPS_measurement_of_TiOx_thin_films#:~:text=As%20XPS%20generally%20probes%20~10nm,substrate%20shouldn't%20matter%20much.&text=Depending%20on%20how%20the%20layers,coefficient%2C%20silicon%20or%20magnesium%20oxide.&text=I%20have%20used%20Si%20wafer,measurements%20and%20it%20worked%20fine.
In addition to what Jürgen Weippert said, I would also check for potential overlaps of the photoelectric peaks of the substrate and the transition metal. Gold, platinum and the other noble metals have many xps peaks in almost any energy range, and the transition metal oxides show extended x-ray excited auger transitions which may overlap. Thus I feel that carbon might be the best opportunity. Best regards, Dirk