Adatom surface diffusion of catalytic metals on the anatase TiO2(101) surface†
Titanium oxide is often decorated with metal nano-particles and either serves as a catalyst support or enables photocatalytic activity. The activity of these systems degrades over time due to catalytic particle agglomeration and growth by Ostwald ripening where adatoms dissociate from metal particles, diffuse across the surface and add to other metal particles. In this work, we use density functional theory calculations to study the diffusion mechanisms of select group VIII and 1B late-transition metal adatoms commonly used in catalysis and photocatalysis (Au, Ag, Cu, Pt, Rh, Ni, Co and Fe) on the anatase TiO2(101) surface. All metal adatoms preferentially occupy the bridge site between two 2-fold-coordinated oxygen anions (O2c). Surface migration was investigated by calculating the minimum energy pathway from one bridge site to another along three pathways: two in the  direction along a row of surface O2c anions and one in the  direction between two rows of surface O2c anions. For all adatoms, migration along the  direction is favored over migration along the  direction due to closer packing of the atoms in the  direction and therefore stronger adatom–surface interactions. As the adatom hops along the  direction, it preferentially moves through a metastable OTiO structure in which the adatom partially embeds itself within the surface, with the exception of Au, which remains above the surface. The adatoms migrate with relative activation energies of: Au (0.24 eV) < Ag (0.48 eV) < Rh (0.60 eV) < Co (0.78 eV) < Pt (0.84 eV) < Ni (0.86 eV) < Cu (1.23 eV) < Fe (1.79 eV) along the favored pathway. This preference arises from the strength of adatom–surface bonding and the electronegativity difference between the metal adatom and the TiO2 surface. We found a linear correlation between the binding energy/electronegativity and the activation energy for hopping where stronger binding energies and more oxidized adatoms have higher activation energies for adatom migration. The linear correlation developed in this work enables rapid estimations of the hopping rates of other transition metal adatoms across the TiO2 surface.