Thermal behavior of indium nanoclusters in ion-implanted silica

Fused silica substrates were implanted with $2\times {10}^{17}{\mathrm{In}}^{2+}∕{\mathrm{cm}}^2$ ions at $320\mathrm{keV}$. Indium crystalline nanoclusters with an average size of about $15–20\mathrm{nm}$ were found in the as-implanted samples. The thermal behavior of the nanoclusters was studied by performing heating-cooling cycles in vacuum and by using in-situ techniques based on glancing-incidence x-ray diffraction and transmission electron microscopy. The precipitates showed both superheating and supercooling. Moreover, no evidence of clusters growth or reorientation during the thermal cycle was found. A detailed study of the heating sequence showed that the melting temperature of the Indium precipitates depended on their size, i.e., the smallest particles melt first and at a temperature which is about $7\mathrm{K}$ below the bulk melting point, while the largest ones were superheated until about $13\mathrm{K}$ above it. Moreover, a remarkable stability of the In cluster well above their melting temperature (up to about $980\mathrm{K}$) was evidenced by in-situ transmission electron microscopy analysis. From a thermodynamic point of view, the experimental results were explained by considering two effects acting on the clusters: the thermodynamic size effect and the pressure of the silica matrix.

Figure 1

X-ray diffraction scans collected around the In(101) peak and performed at the following temperatures (from right to the left): $T=\mathrm{RT}\left(h\right)$, 373, 413, 423, 425, 430, 433, 436, 438, 443, 448, $453\mathrm{K}$ and again at $\mathrm{RT}\left(c\right)$, after the cooling of the sample.

Figure 2

(a) Normalized integrated intensity of the nanoclusters In (101) GIXRD peak versus the temperature for the thermal cycle between RT and $453\mathrm{K}$. Solid and open circles represent data obtained during the heating and cooling sequences, respectively. The solid line connecting the experimental points is only a guide for the eye. The vertical solid line at $T_0$ indicates the In bulk melting point. (b) Normalized integrated intensity of a In film evaporated on a glass substrate, as comparison. The vertical solid line at $T_0$ indicates the In bulk melting point.

Figure 3

SAED patterns showing the evolution of the electron diffraction from In clusters during the thermal cycle. The patterns were collected at the temperatures of $298\mathrm{K}$ (RT), $423\mathrm{K}$, and $443\mathrm{K}$ during the heating sequence (label “$h$”) and at $447\mathrm{K}$, $318\mathrm{K}$, and $298\mathrm{K}$ (RT) during the cooling sequence (label “$c$”). The white arrow indicates a crystalline spot in the diffraction pattern in the cooling cycle at $318\mathrm{K}$ (onset of recrystallization).

Figure 4

Histograms of the size distributions of In clusters at $298\mathrm{K}$ (RT) and $443\mathrm{K}$ obtained from the Bright-Field TEM cross-section analysis.

Figure 5

Bright-Field TEM cross-sectional micrographs of the implanted sample collected at the following temperatures: $298\mathrm{K}$ (RT), $443\mathrm{K}$, $888\mathrm{K}$, and $983\mathrm{K}$.

Figure 6

Pseudo-Voigt integral breadth (solid triangles) of the In(101) GIXRD peak and the corresponding clusters volume averaged size $⟨D_V⟩$ (solid circles) versus the temperature. The open symbols are referred to the measurements done at RT after the thermal annealing. The error bars are the estimated standard deviations (e.s.d.’s) obtained from the fit.

Figure 7

Calculated diameters distributions of the In clusters versus the temperature. The clusters distributions were calculated by considering a log normal size distribution.

Figure 8

Experimental melting temperature (solid circles) of the In clusters versus the diameter of the clusters. The solid line connecting the experimental points is only a guide for the eye. In the inset it is shown the linear fit of the higher temperature data corresponding to the melting interval.