First Principles Studies of Small Molecular Adsorbates on The Anatase TiO2 (101) Surface

Abstract

The (101) surface of the anatase polymorph of titanium dioxide (TiO2) has found wide application in (photo-)catalysis due to its unique surface properties. We have studied the nature of active sites and their influence on surface adsorptions, especially in the presence of intrinsic defects. Density functional theory (DFT) is used in the simulations of the electronic and spectroscopic properties, at several different functional levels.

The adsorption configurations and vibrational frequencies of formic acid on the anatase (101) surface have been investigated. Dissociation of formic acid is highly promoted by (sub-)surface oxygen vacancies (VO's). Description of the vibrational properties, notably of the band separation between the two C-O stretching modes (Δasy-sym), relies on an accurate description of the surface-adsorbate hydrogen bonds. Our results show that the hybrid PBE0 functional provides a better description of hydrogen bonds than the GGA-PBE functional, and predicts adsorption geometries that can explain the bands in infrared spectra.

The CO stretching frequency (&nuC-O) is an effective probe of surface active sites. We have studied CO adsorption on the anatase (101) surface with intrinsic subsurface defects, notably an oxygen vacancy (VO) or a Ti interstitial (Tii). Both these defects give rise to excess electrons in the material. Excess electrons induce a redshift in &nuC-O compared to the oxidized surface due to the π back donation mechanism. This effect decays with increased distance from the defect, and becomes more pronounced when the number of excess electrons increases, i.e., the red shift is larger with a Tii (4 e-/defect) than with a VO (2 e-/defect). Our results also show that the Linear-Response method predicts vibrational frequencies in much better agreement with experimental values than the finite difference method.

Moreover, our PBE0 functional calculations predict that VO is more stable in the subsurface region than at the very surface, in agreement with the PBE results. However, our DFT+U calculations suggest that the introduction of a Fe impurity stabilizes the surface VO; in addition, the Fe captures at least one of the two electrons generated by VO. Finally, we studied the adsorption of molecular oxygen (O2), a well-known electron scavenger capable to attract a subsurface VO to the surface to form a bridging oxygen dimer. In the presence of a Fe impurity, the oxygen dimer traps the excess charge introduced by the defect and has stronger electron scavenging ability than Fe on the anatase (101) surface.