Structure of Si(111)-In Nanowires Determined from the Midinfrared Optical Response

The anisotropic optical response of $\mathrm{Si}(111)-(4\times 1)/(8\times 2)$-In in the midinfrared, where ab initio studies predict significant changes in the band structure between competing models of this important quasi-1D system, has been measured using infrared spectroscopic ellipsometry (IRSE) and reflection anisotropy spectroscopy (RAS). Both IRSE and RAS of the ($8\times 2$) phase show that the anisotropic Drude tail of the ($4\times 1$) phase is replaced by two peaks at 0.50 and 0.72 eV, which appear in ab initio optical response calculations for the hexagon model of the ($8\times 2$) structure, but not the trimer model.

Figure 1

Schematic top views of the hexagon and trimer model of the LT Si(111)-In nanowire array.

Figure 2

IRSE spectra of the RT and LT In-Si phases, for planes of incidence parallel and perpendicular to the In chains: ($4\times 1$) parallel ($\circ$), ($4\times 1$) perpendicular ($+$), ($8\times 2$) parallel (•), ($8\times 2$) perpendicular ($\times$). The clean Si spectra at 300 and 70 K are shown at the bottom.

Figure 3

Enlarged region for the ($8\times 2$) structure from $2500{\mathrm{cm}}^{-1}$ to $3100{\mathrm{cm}}^{-1}$, from which a linearly increasing background has been subtracted. Upper figure: the effect of cooling. Lower figure: the anisotropic response.

Figure 4

RAS spectra of $\mathrm{Si}(111)\mathrm{\text{−}}(4\times 1)\mathrm{\text{−}}\mathrm{In}$ at RT (300 K) and $\mathrm{Si}(111)\mathrm{\text{−}}(8\times 2)\mathrm{\text{−}}\mathrm{In}$ at LT (70 K): upper, experiment; lower, theory. Note the different scales.

Figure 5

Band structure of the hexagon model for $\mathrm{Si}(111)\mathrm{\text{−}}(8\times 2)\mathrm{\text{−}}\mathrm{In}$ calculated within DFT-LDA. Pronounced optical transitions showing up in the RAS spectra are marked. Gray regions correspond to the projected Si bulk bands. The bulk valence band maximum is chosen as energy zero. The Fermi level is indicated.