In Figure 5, different stages of the growth have been imaged by in situ STM, up to a final Ge coverage of 12 monolayers (MLs). It can clearly be seen that three-dimensional structures selectively form inside the trenches; the three-dimensional mounds grow and coalesce until the whole trench is selleck screening library completely filled up, leading to the formation of a long in-plane wire. High-resolution
images, displayed in Figure 6, reveal that the wires are bounded by lateral 113 facets. Moreover, following the underlying mesh of the trenches, the wires show micrometer-length straight sections (Figure 6d) which alternate with junction nodes selleck products (Figure 6e). Cross-sectional TEM measurements clearly confirm the presence of the shallow trenches
under the wires (Figure 3b) and also show the absence of any subsurface dislocation defect close to the substrate/wire interface. This indicates that only the presence of the trench is enough to bias the growth of Ge to heterogeneous nucleation. Figure 5 Wire formation. (a , b , c , d , e , f) STM images showing different stages of the formation of the wires. The total Ge coverage is 12 MLs. Figure 6 Wire faceting. (a , b , c , d , e) STM images showing the morphology of the wires. The bottom insets of (c) show, respectively, (left panel) Selleckchem Ilomastat the line profile and (right panel) the FP of the wire in (c). Being the result of homoepitaxial growth, the wires are totally strain-free. We now show that epitaxial strain introduced by Si deposition dramatically alters
the growth morphology, determining a shape transition from wires to dots. As soon as Si is deposited, we notice the formation of faceted squared and rectangular dots along the wires (Figure 7). These dots progressively grow at the expense of the wires, until the latter completely disappear. By carefully analyzing the STM images of the dot assembly, it is still possible, however, to notice the residual imprint of the wires, appearing as a shallow mound along which the dots are aligned (Figure 7e). Table 1 summarizes the morphological parameters of wires and dots obtained 17-DMAG (Alvespimycin) HCl from a statistical analysis of STM and AFM images. It can be noticed that, during the shape transition, the total volume of nanostructures is preserved: The micrometer-long wires are replaced by a large number of dots, which show a bimodal size distribution. By inspecting in details the morphology of the dots (Figure 8), it can be seen that the islands are either squared or elongated pyramids (huts), again bounded by 113 facets, as indicated by the FP analysis (Figure 8c).This suggests that the observed shape change is not driven by the appearance of new stable facets with strain, but rather by a more efficient strain relaxation or a better surface/elastic energy gain which favors the islands over the wires.