Critical Role of a Buried Interface in the Stranski-Krastanov Growth of Metallic Nanocrystals: Quantum Size Effects in Ag/Si(111)-($7\times 7$)

We show that the buried interface between a metallic nanocrystal and its supporting substrate is essential for understanding the stability of the ubiquitous class of nanomaterials that grow on a wetting layer in the Stranski-Krastanov growth mode. Importantly, these new results reveal the broad role played by quantum confinement effects in the growth of thin nanoscale metals. In situ x-ray scattering experiments on Ag/Si(111)-$(7\times 7)$, where the apparent minimum stable thickness of the first two atomic layers on top of the wetting layer has posed a long-standing puzzle, show that the commensurate wetting layer is locally removed by the formation of incommensurate nanoislands, which is unanticipated for the conventional Stranski-Krastanov growth mode. The anomalous lattice expansion that had been previously proposed is not observed, and these new results for Ag are explained by electron confinement effects whose manifestation differs from other metals.

Figure 1

(a) X-ray specular reflectivity and (b) crystal truncation rod measurements from 0.9 ML of Ag deposited on Si(111)-$(7\times 7)$ at 360 K. The inset shows a side view of the structural model where the unknown crystal structure of the first atomic layer beneath the islands is indicated in black. With the use of the bulk Ag lattice constant, the solid curves are the best fit of the model to the data with the island height distribution given in Fig. 2. The dashed curve imposes a 12% lattice expansion on the first bilayer of Ag. The dotted curve is obtained for islands having the bulk Ag lattice constant but that are displaced from the substrate by an amount that is equivalent to a 12% expansion of the first bilayer.

Figure 2

Comparison of the island height distribution, $p_j$, obtained from the specular reflectivity and the crystal truncation rod measurements in Fig. 1. The two measurements yield nearly identical height distributions, indicating that the fcc Ag islands extend to the substrate interface rather than reside on top of the wetting layer.

Figure 3

The island height distribution $p_j$ for 1.8 ML of Ag deposited at different temperatures. As the temperature increases, the average island height and the breadth of the distribution increases, which indicates that mobility plays an important role in the observed distributions. The inset shows $p_j$ for 0.45 ML Ag deposited at 300 K: even though the coverage barely exceeds the wetting layer saturation (about 5% of the surface is covered by Ag islands), there is a clear preference for three atomic layer islands.

Figure 4

The large in-plane orientational disorder of the fcc Ag islands is observed by comparing scans performed azimuthally in the surface plane through the ${(1.328,1.328,0.1)}_H$ peak of the fcc Ag islands and the ${(8/7,0,0.1)}_H$ peak of the Ag/Si(111)-$(7\times 7)$ wetting layer. The Ag island peak extends more than 3 deg, which is 2 orders of magnitude broader than the Ag wetting layer that is commensurate with Si(111)-$(7\times 7)$. The upper inset schematically shows the Ag island mosaicity in the surface plane. The lower inset provides an expanded view of the Ag wetting layer peak.