Scanning tunneling microscopy studies of nucleation and growth in a reactive, epitaxial system: Co/Si(111)

Abstract

We present scanning tunneling microscopy observations of the reaction of cobalt with Si(111)-(7×7). For deposition at 320 °C (reactive epitaxy), flat-topped monolayer islands of triangular shape with vertices along 〈1̄1̄2〉Si nucleate on the faulted side of the 7×7 structure then grow in size attaining edge lengths that are quantized to integer multiples of the 7×7 unit cell. A 2×2 reconstruction with large corrugation occurs on some of the islands, and is believed to be an ordered array of silicon adatoms. The shape and orientation of the islands appears to be determined by energetics, not growth kinetics. They coexist with the 7×7 structure, implying that metal atoms readily diffuse through the silicon matrix before attaching to an existing island. From the areal density of islands we estimate an activation energy for ‘‘surface diffusion’’ of 0.8 eV, and argue that the process involves metal atom transport through near-surface silicon interstitials. At lower temperatures, no ordered silicide forms, while at higher temperatures, multilayer islands form with a predominant layer spacing of 3.1 Å suggesting that they are (111)CoSi2. For postdeposition annealing at 600 °C (solid phase epitaxy), large, multilayer islands of CoSi2 form, in coexistence with a √7 submonolayer phase. General aspects of the nucleation, growth, and coarsening in this reactive, epitaxial system are discussed.

Terrace width dependence of cobalt silicide nucleation on Si(111)-(7×7)

“…Similar FH boundaries have been observed in previous studies of Co/ Si͑111͒ and Si/ Si͑111͒, and it was proposed that this is due to the higher energy required for the overgrowth of adatoms on the FH compared with UH. 26,27 Here we observe less preferential adsorption of the clusters along the terrace edges compared with Figs. 1͑a͒ and 1͑e͒.…”

Co-induced nano-structures on Si(111) surface

“…It is a little higher than the former type. The edges of both two types of islands are aligned parallel to those of the 7 Â 7 unit cell, and the lengths are integer times as large as that of 7 Â 7 unit cell as have been discussed elsewhere [8]. Due to the small percentage of island relative to the whole surface, the structures of the island surface did not reflect clear corresponding spots in LEED pattern.…”

“…These islands can be classified into two typical types. The major one shows approximately triangular, trapezoidal or hexagonal shape with atomically flat top that shows well-known CoSi 2 (1 1 1)-(2 Â 2) structure [8,9]. A typical image of this kind of island is shown in Fig.…”

Thermal evolution of the morphology of Ni/Ge(111)-c(2 × 8) surface

“…A close resemblance between the observed rings and the building elements of a 2 × 2 reconstruction, which develops on tops of iron silicide islands has led the authors to conclude that the rings constitute a very initial formation stage of iron silicide. The occurrence of ring-like clusters has been also reported when 0.25 ML Co, deposited onto Si(111) surface at 373 K was then annealed to 943 K, and interpreted in terms of the formation of an intermediate stage between the clean and silicide-Si interface [25]. Finally, we would like to underline that the appearance of ring-like features has not been reported for the Co/Ge(111) surface [16].…”