Water does partially dissociate on the perfect TiO${}_2$(110) surface: A quantitative structure determination

There has been a long-standing controversy as to whether water can dissociate on perfect areas of a TiO${}_2$(110) surface; most early theoretical work indicated this dissociation was facile, while experiments indicated little or no dissociation. More recently the consensus of most theoretical calculations is that no dissociation occurs. New results presented here, based on analysis of scanned-energy mode photoelectron diffraction data from the OH component of O 1$s$ photoemission, show the coexistence of molecular water and OH species in both atop (OH${}_{\mathrm{t}}$) and bridging (OH${}_{\mathrm{br}}$) sites. OH${}_{\mathrm{br}}$ can arise from reaction with oxygen vacancy defect sites (O${}_{\mathrm{vac}}$), but OH${}_{\mathrm{t}}$ have only been predicted to arise from dissociation on the perfect areas of the surface. The relative concentrations of OH${}_{\mathrm{t}}$ and OH${}_{\mathrm{br}}$ sites arising from these two dissociation mechanisms are found to be fully consistent with the initial concentration O${}_{\mathrm{vac}}$ sites, while the associated Ti-O bond lengths of the OH${}_{\mathrm{t}}$ and OH${}_{\mathrm{br}}$ species are found to be 1.85 ± 0.08 and 1.94 ± 0.07 Å, respectively.

Figure 1

Schematic diagram of the interaction of molecular H${}_2$O with TiO${}_2$(110) on (left) a surface containing bridging oxygen vacancies, O${}_{\mathrm{vac}}$, to produce two bridging hydroxyl species, OH${}_{\mathrm{br}}$ and (right) on a perfect surface to produce atop and bridging hydroxyl species, OH${}_{\mathrm{t}}$ and OH${}_{\mathrm{br}}$. The diagrams show both of the component parts of a water molecule approaching the dissociation sites and the resulting pair of surface OH species produced. At low temperatures coadsorbed intact molecular water also occupies Ti-atop sites.

Figure 2

Comparison of the experimental O 1$s$ (OH) PhD spectra with theoretical simulations for the best-fit structure discussed in the text.

Figure 3

Dependence of the level of agreement between experiment and theory for the normal emission PhD spectrum as a function of the fractional occupation of OH${}_{\mathrm{t}}$ species. On the left are shown theory/experiment comparisons for 0$\%$ and 30$\%$ OH${}_{\mathrm{t}}$ occupation; on the right is shown the dependence of the $R$ factor on this parameter.