Abstract

In this work, two types of nano-structured TiO2 layers were obtained by two different methods. First, during an exploratory study, a set of nano-columnar TiO2 films and niobium doped TiO2 films was obtained on stainless steel, by a RF reactive sputtering technique. The argon gas is ionized by a high negative voltage applied to the TiO2 target (cathode), and a plasma is created between the cathode and the substrate (anode). Positively charged ions (Ar+) are accelerated toward the TiO2 target (a co-target Nb is placed beside the TiO2 cathode for Nb doping) and their impact sputters atoms off the target. These atoms travel across the chamber and a fraction of them land on stainless steel, resulting in a film TiO2 and Nb-doped film TiO2. With adjusting the RF power applied on the co – target, different concentrations of Nb in the TiO2 film could be controlled. The results of EIS (Electrochemical Impedance Spectroscopy) mainly showed the presence of a barrier layer (junction metal/semiconductor or metal/oxide/semiconductor) between the TiO2 layer and the stainless steel in the absence and presence of niobium. The carrier density is estimated at almost 1018 cm-3. In the second part of this thesis, two types of TiO2 nano-tubular arrays with nanotubes (NT) aligned perpendicular to the titanium substrate were obtained by anodization of a titanium foil, in two different solutions containing fluoride ions. The NT obtained in a tetrabutylammonium / formamide solution (named: TiO2-NT(TB)) are rough, while the second solution, ethylene glycol , allows to synthesize smoother and denser tubes (named: TiO2-NT(EG)). All the nanotubular arrays were characterized by EIS and showed in the high frequency range a contribution related to the presence of surface states, and a contribution at low frequencies related to the capacity of the space charge layer. In the dark, in the Na2SO4 solution with neutral pH, the EIS study of TiO2-NT (TB) anatase showed that these two contributions vary with the applied potential. The capacity of surface states varied exponentially in a wide range of potential, but in addition the presence of a localized energy state in the gap could be evidenced. This localized state is the signature of the adsorption of molecular water. After UV exposure during 3h, and back to the dark, the quasi irreversible disappearance of the localized energy states is related to photo-induced adsorbed water dissociation at some sites on the surface of NT. Furthermore, the increase by a factor 112 of the capacity of the space charge layer was observed after UV exposure. This increase can be explained by the photo activation of the surface of NT, which was inactive before UV exposure. This activation is related to the dissociation of adsorbed molecular water and the insertion of hydrogen into the walls of NT. A geometric model considering the variation of the band bending taking place inside the wall of NT is proposed to replace the classical Mott-Schottky relation, which is only valid for a plate condensator. This model allows understanding the variation of the space charge layer as a function of the applied potential. With this new model, a carrier density of about 1018 cm-3 et 1020 cm-3 respectively before and after UV illumination were determined, confirming that the photo-induced activation of the wall of NT is linked to the phenomenon of doping due to hydrogen insertion. In the case of TiO2 –NT(EG), the contribution of adsorption of molecular water was not observed. The spectrum of EIS before and after UV exposure did not show a significant change and the capacities of space charge layer after illumination only increased by a factor 8. The simulation with the model shows that the carrier density is about 1020cm-3 before and after illumination. In other words, these tubes of TiO2 –NT(EG), are already activated before illumination and the photo-induced effects are less important compared to the TiO2 –NT(TB) array. For a better identification of the chemical nature of the surface state in the case of rough tubes of TiO2 –NT(TB), EIS measurements were performed in the same manner, but in acidic (pH=3.5) and basic (pH=12.5) media, in order to compare the behaviour to that observed in the neutral Na2SO4 solution. In the alkaline electrolyte, the band bending varies only slightly with the applied potential, reflecting a shift of band edge and the filling/emptying of the surface states during polarization. Moreover, the exponential distribution of capacities of surface states is more spread out (850 meV) in alkaline solution than in the neutral solution (257meV), showing clearly the particular role of the OH groups at the surface of the tubes. In the acidic electrolyte, a phenomenon of diffusion –insertion of protons should be taken into account for interpreting the spectra of EIS. H atoms play the role of electron donors able to increase the carrier density in the wall of the tubes. The comparison between the behaviours in the 3 media clearly associates the surface states to hydroxyl groups. In the last part of this thesis, experiments were performed on TiO2 –NT(TB) in a NaOH electrolyte, using Intensity Modulated Photocurrent Spectroscopy (IMPS), and the results are discussed in comparison with a thin compact film deposed on a titanium foil par PVD.