Atomic Layer Deposition of CeOx Nanoclusters on TiO2

Titanium dioxide has a band-gap in the ultra violet region and there have been many efforts to shift light absorption to the visible region. In this regard, surface modification with metal oxide clusters has been used to promote band-gap reduction. CeO_x-modified TiO₂ materials have exhibited enhanced catalytic activity in water gas shift, but the deposition process used is not well-understood or suitable for powder materials. Atomic layer deposition (ALD) has been used for deposition of cerium oxide on TiO₂. The experimentally reported growth rates using typical Ce metal precursors such as β-diketonates and cyclopentadienyls are low, with reported growth rates of ca. 0.2-0.4 Å/cycle. In this paper, we have performed density functional theory calculations to reveal the reaction mechanism of the metal precursor pulse together with experimental studies of ALD of CeO_x using two Ce precursors, Ce(TMHD)₄ and Ce(MeCp)₃. The nature and stability of hydroxyl groups on anatase and rutile TiO₂ surfaces are determined and used as starting substrates. Adsorption of the cerium precursors on the hydroxylated TiO₂ surfaces reduces the coverage of surface hydroxyls. Computed activation barriers for ligand elimination in Ce(MeCp)₃ indicate that ligand elimination is not possible on anatase (101) and rutile (100) surface, but it is possible on anatase (001) and rutile (110). The ligand elimination in Ce(TMHD)₄ is via breaking the Ce-O bond and hydrogen transfer from hydroxyl groups. For this precursor, the ligand elimination on the majority surface facets of anatase and rutile TiO₂ are endothermic and not favourable. It is difficult to deposit Ce atom onto hydroxylated TiO₂ surface using Ce(TMHD)₄ as precursor. Attempts for deposit cerium oxide on TiO₂ nanoparticles that expose the anatase (101) surface show at best a low deposition rate and this can be explained by the non-favorable ligand elimination reactions at this surface.