Abstract:

The anatase phase of TiO₂ is often considered to have the highest reactivity among TiO₂ polymorphs. Since the anatase TiO₂(001) surface has a relatively high surface energy, it is expected to be active; however, because of its high surface energy, the surface generally forms a (1 × 4) reconstructed structure. A model named, "ad-molecule" (ADM) model, has been suggested for this (1 × 4) reconstruction, theoretically predicting that the surface retains a high reactivity. However, several recent experimental results have shown that the (1 × 4) reconstructed surface is not as active as expected, leading to a controversy about the actual atomic geometry of the reconstruction. Recent theoretical work suggests that the introduction of strain in the anatase TiO₂(001) surface may enhance its reactivity by distorting the surface lattice. Thus, understanding the surface structure under strain may be the key to resolving these existing challenges with this material. Herein, we present a systematic study of the epitaxial growth of anatase TiO₂(001) films on BaTiO₃(001)/SrTiO₃(001) substrates using pulsed laser deposition, characterized using X-ray diffraction (XRD), X-ray photoemission spectroscopy (XPS), scanning transmission electron microscopy (STEM), and scanning tunneling microscopy (STM). A thin layer of BaTiO₃(001) was epitaxially grown on several SrTiO₃(001) substrates to introduce strain in anatase TiO₂(001) films by leveraging the relatively large lattice mismatch between the anatase TiO₂(001) and BaTiO₃(001). The XRD and STEM results showed that strain was partially introduced in the films when the thickness of the BaTiO₃ layer was ~4–6 nm. The XPS results showed that a suitable thickness of the anatase TiO₂(001) films was at least 15 nm, inducing a negligible concentration of outwardly diffused Sr and Ba from the substrate to the surface, and minimizing their possible effects on the surface structure. Dominant Ti⁴⁺ oxidation state was observed, indicating that the anatase TiO₂(001) surface was fully oxidized. The surface structure as characterized by STM showed that the (1 × 4) reconstruction remained as films grew on the SrTiO₃(001) substrate. However, ridges in the (1 × 4) reconstructed surface showed additional super-periods typically shown as dim features in the ridges separated by 2–5 lattice distances. Considering the high-resolution STM images and fully oxidized surface, we propose that these dim features may have been caused by "TiO₂" vacancies in the ridges. This is consistent with the ad-oxygen model (AOM) for the fully-oxidized (1 × 4) reconstructed surface of anatase TiO₂(001). In the AOM model, Ti atoms in the ridges were coordinated fivefold, in contrast with the fourfold coordination in the ADM model. We find direct evidence that the strains introduced in anatase TiO₂(001) films can significantly modify ridge structure in the (1 × 4) reconstructed surface, providing key insights into the complicated surface structure, and suggesting important implications for furthering our understanding of the reactivity of this commonly used surface.

Key words: Anatase TiO₂(001) surface, Thin film, Scanning tunneling microscopy, Surface reconstruction, Surface strain