Defects at the rutile TiO2 surface
Title: Defects at the rutile TiO2 surface
SNIC Project: SNIC 2014/1-3
Project Type: SNAC Medium
Principal Investigator: Susanne Mirbt <>
Affiliation: Uppsala universitet
Duration: 2014-02-01 – 2015-02-01
Classification: 10304


TiO$_{2}$ is a functional material with widespread applications. It‚s wide bandgap (3 eV) is utilized in sunscreening and chemical solar cells. It has a high refractive index. It is discussed in connection with photocatalytic splitting of water and degradation of organic molecules in polluted air or water. In addition this oxide show a broad spectrum of solid state phenomena. We study phenomena due to oxygen vacancies in the rutile (110) surface. Our calculations reveal that the numerical modeling of this surface is demanding: At least a (4x2) unit cell with at least 4 slab layers is mandatory. In addition, in order to describe the electron localization caused by the vacancy formation, the use of a hybrid functional is mandatory. This implies one has to do a Hartree Fock calculation and mix it with the density functional calculation. And a Gamma-point only calculation we find to give wrong results. First we studied the surface reconstruction at the (110) surface.Ideal rutile TiO2 has orthorhombic D$_{2h}$ symmetry, because not all 6 oxygen-titantium bonds are equal. The (110) surface is the cleavage surface of rutile TiO2. The surface is charge neutral and reconstructs in a (1x1) pattern. It contains both Ti and O ions with two different coordinations. Second we studied the vacancy formation. The removal of one bridging oxygen ion gives rise to an oxygen vacancy. An oxygen vacancy in TiO2 gives rise to an excess of two electrons because every neutral oxygen atom contains 6 electrons but the oxygen ion in TiO2 is populated by 8 electrons. In the ideal rutile lattice these two excess electrons can only populate the conduction band, because the valence band is completely occupied. We calculated that the vacancy formation not only gives rise to a lattice reconstruction around the vacancy but always is accompanied by the formation of a bipolaron. Third we want to study the interaction between two vacancies in order improve the understanding of existing experimental results. Fourth we want to study Ti interstitial defects because they have been proposed to exist more frequent than oxygen vacancies. In order to continue our study in line with the last two points, we have large need of the here applied computer resources.