Effect of indium and niobium segregation on the surface vs. bulk chemistry of titanium dioxide (rutile)

Since the landmark paper in 1972 by Fujishima and Honda [1], TiO 2 has become one of the most promising candidates of a new generation of solar energy materials capable of generating clean hydrogen fuel using only sunlight (photo-electrochemically) to dissociate water. TiO2 has both bulk properties and surface properties which contribute to its functional performance. Considering that all of the electrochemical reactions induced by light o ccur at the surface of TiO2, it becomes clear that understanding the surface properties of TiO 2 is of crucial importance for its performance; specifically the conversion of solar energy into chemical energy. The surface phase of TiO 2 can be substantially different from that of the bulk phase as a result of a phenomenon known as segregation. Segregation involves the transport of certain lattice species from the bulk phase to the surface, driven by excess surface energy. To date, developments in the understanding of TiO 2 solid solutions and related properties have mainly been centred on bulk properties. In comparison, relatively little work has been reported on segregation in T iO2 solid solutions and its influence on functional properties, such as reactivity and photoreactivity. The present work has studied the effect of indium (acceptor-type ion) and niobium (donortype ion) segregation on the surface chemistry of well-defined In-doped and Nb-doped TiO 2 solid solutions. Specifically, examining the relationship between imposed sample processing conditions, such as the gas phase oxygen activity, on segregation-induced surface enrichment. This was achieved using a range of complimentary analysis techniques including X-ray photoelectron spectroscopy (XPS), secondary ion mass spectrometry (SIMS), Rutherford backscattering (RBS) and proton-induced X-ray emission (PIXE). In-doped TiO2 The incorporation of dopants in TiO 2 results in a change of Fermi level and related semiconducting properties. However, to achieve well-defined solid solutions, in which the dopant is homogeneously distributed throughout the bulk phase, requires knowledge of the dopant ion diffusion coefficients [2-7]. The present work included the determination of the selfdiffusion coefficient of indium in TiO2 single crystal (rutile) in the temperature range 1073 1573 K and p(O2) = 21 kPa.   1 2 4 316 exp 10 4 . 7 2           s m RT mol kJ D TiO In A.J. Atanacio Abstract | Page 12 The determination of this new data enables us to predict the processing conditions required to incorporate In into the TiO 2 lattice. The effect of oxygen activity on the segregation-induced surface concentration of indium in In-doped TiO2 was studied for various annealing times, ranging between 5 h and 120 h, at 1273 K. It was shown that the equilibrium segregation of indium in oxidising conditions, p(O2) = 21 kPa, can be established within ~20 h. However, in highly reducing conditions, p(O2) = 10 Pa, equilibrium could not be established due to indium evaporation, which becomes substantial at p(O2) < 1.8x10 kPa. The present work examined the effect of indium (acceptor) segregation on the surface vs. bulk composition of In-doped TiO 2. The data determined that annealing of 0.3 at% In-doped TiO2 at 1273 K in oxygen activity, p(O 2) = 75 kPa and p(O 2) = 10 Pa, resulted in an enrichment of the surface to the level of 2.95 at% In and 2.61 at% In, respectively . It was postulated that substantial In surface enrichment leads to the formation of a low-dimensional surface structure and a sub-surface layer resulting from the interactions of titanium vacancies and interstitial indium ions. Nb-doped TiO2 The present work also studied the effect of niobium (donor) segregation on the surface and near-surface composition of Nb-doped TiO 2 (0.18 at% Nb and 0.018 at% Nb). The data showed that 0.18 at% Nb-doped TiO 2 specimens annealed at 1273 K in oxidising conditions, p(O 2) = 75 kPa and p(O2) = 10 Pa, resulted in segregation-enriched surface concentrations of 2.83 at% and 2.35 at%, respectively. However, annealing the same specimen in strongly reducing conditions, p(O2) = 10 Pa, resulted in a depletion of the surface to the level of 0.05 at% Nb (desegregation). The theoretical model postulated in this work considered the predominant driving force for niobium segregation in oxidising conditions is the negative surface charge associated with titanium vacancies formed in the surface layer. However, the effect of desegregation is induced by a positive surface charge related to a Magneli-type surface structure formed in strongly reducing conditions. The present work provides new well-defined empirical data on segregation. The derived theoretical models describing the effect of processing conditions on indium and niobium segregation/desegregation can be used as a technology for the imposition of controlled surface composition. The related data can be used for imposition of a chemically-induced electric field required for charge separation and controlled surface composition that is required to achieve desired reactivity of TiO2 as photocatalyst and photoelectrode for solar energy conversion. A.J. Atanacio Abstract | Page 13