Heterogeneous photocatalysis over TiO2 using the sunlight seems to be a promising technology for waste-water treatment and drinking water production. However, the overall efficiency of TiO2 under natural sunlight is limited to the UV-driven activity (λ < 400 nm), accounting only to ~ 4% of the incoming solar energy on the Earth's surface. The increase of the TiO2 absorption towards wavelengths more abundant on the planet's surface has, thus become recently an interesting challenge. In this framework, the main objective of this work was to study the modification of TiO2 by non metallic doping (N and S) in order to increase its light absorption in the visible region (45% of the light hitting the terrestrial surface). Commercial TiO2 powders were doped by a simple preparation method consisting of manual grinding them with a N or S precursor (thiourea) and then annealing at 400 and 500 °C. At both temperatures N, S co-doped commercial TiO2 powders with visible response were obtained. However, the annealing step produced a reduction on the specific surface area and dehydroxylation causing a detrimental effect on the photocatalytic UV-driven activity principally on photocatalyst annealed at 500 °C. E. coli bacteria inactivation, phenol and di-chloroacetate (DCA) oxidation was achieved when the co-doped photocatalyst annealed at 400 °C was illuminated upon UV light the •OH radical being the main oxidative species, that was detected by Electronic Spin Resonance (ESR) spin-trapping with DMPO. In contrast, upon visible light irradiation, N, S co-doped TiO2 powders showed a diminished oxidation power since phenol and DCA were not oxidized. On the other hand, it was possible to inactivate E. coli cells demonstrating that under these conditions, the photocatalytic mechanism was different. In order to elucidate this mechanism, characterization by Diffuse Reflectance Time Resolved Spectroscopy (DRTRS), Low Temperature-ESR and ESR-spin trapping measurements were done. By DRTRS measurement, it was found that under visible light excitation, electrons would be promoted from N, S localized states within the band gap to the conduction band. These electrons were probably trapped on shallow traps such as oxygen vacancies (Vo) allowing them to reach lifetimes of ms. LT-ESR data revealed that N, S co-doping probably could benefit the formation of Vo. ESR spin trapping with TMP-OH as a singlet oxygen quencher, revealed the formation of singlet oxygen as the main oxidative species instead of the •OH radical. Thus, it was suggested that a photo-promoted electron, instead of the localized hole on N, S states, would be the carrier charge playing the main role in photocatalytic reactions on N, S co-doped commercial TiO2 powders. When the electron is trapped on Vo, it could react with molecular oxygen previously adsorbed producing superoxide radical (•O2-). Finally, the oxidation of •O2- by localized holes seems to be thermodynamically favored leading the singlet oxygen (1O2) production. It is well known that singlet oxygen is a Reactive Oxygen Species with a lower oxidative power than the hydroxyl radical. Singlet oxygen is not able to oxidize organic substances such as phenol and DCA but this species is very toxic to microorganism producing lipid peroxidation reaction on biological membranes. It was also concluded that N, S co-doped TKP 102 annealed at 400 °C did not present an enhancement on their photocatalytic activity towards phenol oxidation and E. coli inactivation when simulated solar light was used. Undoped Degussa P-25 was the commercial powder with highest photocatalytic activity. Evidences are reported about the classical photocatalytic process where •OH radicals are produced mainly by oxidation of water or hydroxyl ions with photo-induced valence band holes prevails upon simulated solar light exposition. Localized states induced by N or S-doping and responsible of visible light absorption did not play an important role on the photocatalytic activity of these novel materials under the experimental conditions used.