Initial stages of the growth of CaF2 on Cu(111) visualized by STM

The epitaxial growth of insulators on semiconductor surfaces is a topic that has attracted widespread interest in recent years. In particular, a strong effort has been devoted to study the growth of CaF 2 on Si(111) and (100) surfaces [1] because of the potential interest for novel Si-based quantum effect devices, such as Resonant Tunnelling Diodes operating at 300 K [2]. Much less is known concerning the epitaxial growth of insulators on metal surfaces. We briefly report here on the early stages of the growth of CaF 2 on Cu(111), a highly mismatched system (the lattice parameters are 0.543 nm versus 0.361 nm) as visualized by STM. Deposition of 0.4 TL at 300 K produces initially the decoration of the steps of the substrate from the lower side. Once that the steps are saturated, islands irregular in shape and 1.2-1.9 nm in height (4-6 TL) nucleate at the terraces as shown in Figure 1. The density of islands on the terraces is 3 × 10 10 cm-2. The Cu(111) lattice can be atomically resolved in the patches of the surface between the islands, showing that, unlike Si(111) [1, 2], there is no CaF wetting layer with the Ca atoms bonded to Si, onto which the following molecules can grow epitaxially. Most of the CaF 2 molecules stay intact as they contact the Cu(111) surface and diffuse on it to nucleate and form the observed islands. There are, however, indications (not shown) of a very limited reaction between the Cu(111) substrate and the CaF 2 molecules in the form of some dark, triangular shaped defects of unknown nature, but not detected in the clean Cu(111) surface [3]. It is important to note that in order to decorate the steps of the Cu substrate, the diffusion length of CaF 2 molecules at 300 K has to be larger than the average distance between steps, i.e. 100 nm. Fig. 2 shows that increasing the amount of CaF 2 evaporated at 300 K to 1.7 TL simply increases the density of islands nucleated on the terraces to 1.2×10-11 cm-2 , but not their height, which is still 4-6 TL. The islands show rami¯ed or dendritic shapes, indicating that the diffusion along the steps at 300 K is not fast enough to produce islands with compact shapes. After deposition at 550 K the steps are still decorated but the islands at the terraces have now …