I am not sure that I fully understand the question you are asking but it is very important to consider charge balance when doping wide band gap oxides. Perhaps, in this case, it is because S(2-) is incorporated substitutionally for O (2-). Same charge state and ions of similar (ish) sizes. If you were to replace O(2-) with F(-) for example you would be n-doping the TiO2 and that is expensive as TiO2 is a wide band gap semiconductor.
S, N, and C are complex cases where both cation and anion doping (different chemical species as well as charge compensating defects) are possible and typically observed in the same material. F is a more simple case. You can see several new reviews for a summary.
C and N can enter into TiO2 either substitutionally or intersitially. But F can enter into TiO2 as substitutionally. F cannot enter into an interstitial position of TiO2. I don't know the scientific reason behind this. Could you help me to understand this?