Medium-energy ion scattering investigation of the surface and subsurface structure of three-dimensional ${\mathrm{HoSi}}_{2-x}$ grown on $\mathrm{Si}\left(111\right)$

The surface structure of the three-dimensional (3D) holmium silicide grown on $\mathrm{Si}\left(111\right)$ has been determined using medium-energy ion scattering. Using this technique, it has been possible for the first time to elucidate the structure of the surface bilayer, free from complications due to lower layer atoms. It has been established that this surface generally displays sixfold symmetry which has been attributed to a mixture of reversed and nonreversed buckled bilayers ($B$-type and $A$-type respectively). A threefold surface has also been observed which is due to the preferential growth of an $A$-type buckled bilayer. It is proposed that the underlying vacancy network is the cause of the mixed phase on the surface. Both phases show consistent lattice spacings with the surface Si bilayer in the $A∕B$ mixed phase sitting $1.97\pm 0.01\mathrm{\AA }$ above the first Ho layer, with a vertical buckling separation of $0.79\pm 0.04\mathrm{\AA }$. Such terminations are present on top of the established layered structure of the 3D silicide, with a $\mathrm{Ho}\text{−}\mathrm{Ho}$ vertical spacing of $4.01\pm 0.02\mathrm{\AA }$, which is compressed relative to the bulk value due to the relief of strain at the interface.

Figure 1

The structural model for the 3D RE silicides as proposed by Lohmeier et al.[17]

Figure 2

A top down view of the three azimuths used in this study.

Figure 3

(Color online) The effect of varying the number of layers in the $\left[\overline{1}0\overline{2}\right]∕\left[13\overline{1}\right]$ geometry. Simulations of 5–8 layers are shown with the experimental data offset for clarity.

Figure 4

Normal incidence geometries eliminate contributions from lower Ho layers due to the effects of shadowing if a $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ beam is used. Hence features due to the surface bilayer alone make up the blocking curves for such geometries.

Figure 5

(Color online) (a) The symmetrical MEIS blocking curves obtained by summing data sets from the five samples which displayed sixfold symmetry in the LEED pattern. Data were acquired using $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ ions. The blocking curves are offset along the $y$-axis for clarity. (b) A side view schematic showing which atoms cause each of the observed features.

Figure 6

(Color online) $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ MEIS blocking curves from the surface displaying threefold symmetry, which are typical of an $A$-type Si bilayer termination.

Figure 7

(Color online) The fits obtained by $R_{\chi }$ for a $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ beam, using the vibrations listed in Table . The structural parameters are summarized in Table . (a) The $\left[n\right]∕\left[100\right]$ geometry. (b) The $\left[n\right]∕\left[110\right]$ geometry.

Figure 8

(Color online) The results of an experimental damage survey in the $\left[n\right]∕\left[100\right]$ geometry. The same sample point was exposed to a series of $1.5\mathrm{\mu }\mathrm{C}$ ${\mathrm{He}}^{+}$ doses. The $\mathrm{Si}1$ blocking features accumulate damage more rapidly than the $\mathrm{Si}2$ features.

Figure 9

(Color online) The fits obtained by $R_{\chi }$ using $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ and a simple model of damage evolution. The structural parameters are summarized in Table . (a) The $\left[n\right]∕\left[100\right]$ geometry. (b) The $\left[n\right]∕\left[110\right]$ geometry.

Figure 10

(Color online) A comparison of $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ data from the $A∕B$ mixed phase and the predominantly $A$-type terminated surface. The $A$-type data from both geometries have been summed to restore the sixfold symmetry in the MEIS data and allow a comparison with the data from the $A∕B$ mixed phase. The resulting curves look very similar to the $A∕B$ mixed phase.

Figure 11

(Color online) The fits obtained for the $A$-type termination by $R_{\chi }$ using $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ and a simple model of damage evolution. The structural parameters are summarized in Table . (a) The $\left[n\right]∕\left[100\right]$ geometry. (b) The $\left[n\right]∕\left[110\right]$ geometry.

Figure 12

(Color online) The fits obtained using $100\mathrm{keV}$ ${\mathrm{H}}^{+}$ in a variety of geometries. The structural parameters are summarized in Table . (a) The $\left[\overline{1}00\right]∕\left[110\right]$ geometry, (b) the $\left[0\overline{1}\overline{1}\right]∕\left[100\right]$ geometry, and (c) the $\left[\overline{1}0\overline{2}\right]∕\left[13\overline{1}\right]$ geometry.

Figure 13

Side views of the three different structures used to model damage, which are mixed in the ratio $a:b:c$. (a) A perfect, undamaged surface bilayer, (b) a surface layer which has had the $\mathrm{Si}1$ atoms removed, and (c) a fully damaged surface.

Figure 14

(Color online) (a) The $1.5\mathrm{\mu }\mathrm{C}$ experimental data fitted to simulations with a damage ratio, $a:b:c$, of 50:35:15. (b) The $3.0\mathrm{\mu }\mathrm{C}$ experimental data fitted to simulations with a damage ratio, $a:b:c$, of 0:70:30. Both fits are for $50\mathrm{keV}$ ${\mathrm{He}}^{+}$ in the $\left[n\right]∕\left[100\right]$ geometry.