Positioning of Strained Islands by Interaction with Surface Nanogrooves

When strained Stranski-Krastanow islands are used as “self-assembled quantum dots,” a key goal is to control the island position. Here we show that nanoscale grooves can control the nucleation of epitaxial Ge islands on Si(001), and can drive lateral motion of existing islands onto the grooves, even when the grooves are very narrow and shallow compared to the islands. A position centered on the groove minimizes energy. We use as prototype grooves the trenches which form naturally around islands. During coarsening, the shrinking islands move laterally to sit directly astride that trench. In subsequent growth, we demonstrate that islands nucleate on the “empty trenches” which remain on the surface after complete dissolution of the original islands.

Figure 1

Surface morphology after deposition of 5 ML of Ge on Si(001) at $740°\mathrm{C}$. (a),(b) $2\times 2\mu {\mathrm{m}}^2$ AFM images of the sample. (c)–(f) STM images of a dome (c), a pyramid (d), a truncated pyramid (e), and an empty trench (f). The gray scale reflects the local surface slope with respect to the (001) plane in (a) to highlight the islands; the local height in (b) and (f), to highlight the trenches; and a combination of the local surface slope and height in (c)–(e) to show both the islands and the trenches. The sides of the images are aligned parallel to the [100] directions and the scale bars are 50 nm.

Figure 2

(a) Schematic illustration of the geometry used to calculate relaxation energy, for lateral island positions $x=0$, 1, and 1.5 in units of the island half-width. The trench is $1/5$ as wide as the island here. Black insets illustrate corresponding total thickness $h(x)$ of the strained layer; see text for details. (b) Energy vs lateral island position. The plot is independent of island size or slope—these factors only enter through the energy scale, which is left unspecified to emphasize this generality.

Figure 3

(a)–(b) AFM images showing the same surface area (a) before and (b) after etching the sample for 30 s in a BPA solution. (c) Difference of (a) and (b), highlighting that the trenches are partially filled with SiGe. (d) Schematic illustrating the smoothly curved surface expected for the material in the trench.

Figure 4

AFM image of a sample obtained by growth of 6 ML Ge at $840°\mathrm{C}$, followed by 300 s annealing, and then depositing 6 ML Ge at $620°\mathrm{C}$. (a) Overview image showing both large islands from first growth, and small islands formed during second growth. (b)–(d) Closeups of (b) one, (c) four, and (d) a continuous string of small islands that formed on empty trenches during the second growth, with each image centered on a circular trench (barely visible due to its shallowness). (e) Line scan taken along the line marked in (d). (f) Histogram of number of islands nucleating on each empty trench in the second growth step.