By adsorption, if you mean to find the binding energy, then by using Gaussian 16, you can optimize the molecular system under study, i.e., cerium + other molecules. Then, you can optimize the individual molecules.
You will get the binding energy by taking the total electronic energy difference between the entire system and the sum of the energies of individual molecules. Depending on the calculation level, this energy should give you a qualitative or quantitative idea of adsorption.
If you want to model adsorption on a solid-state surface, Gaussian 16 might not be ideal.
I hope this helps!
Note:- The above answer suggests calculating binding energy, which was the earlier question (before editing).
You can go to the Basis Set Exchange website (https://www.basissetexchange.org), select the atoms you have in your system and look for basis sets that contains your atoms and that is of sufficient quality for your desires.
Usually it is recommended to employ relativistic corrections for heavy atoms (like Ce), so I'd recommend using ZORA/DKH2 and also basis sets that are parametrized for this application (SARC2-QZV-DKH2 for example). Depending on your system you could mix different basis sets, but try to go beyond TZ for lanthanides and always consider relativistic effects.
Hope I was of some help, even though there's no simple answer like "use this or that" when the subject is basis sets.