Er sensitization by a thin Si layer: Interaction-distance dependence

From photoluminescence measurements on sensitized erbium in $a$-Si/${\mathit{SiO}}_2$:Er/${\mathit{SiO}}_2$ multilayers, we determine the characteristic interaction length of the sensitization process from the silicon-layer sensitizer to the erbium-ion receiver to be $0.22\pm 0.02$ nm. By using sufficiently low temperatures in the fabrication steps, we ensure that diffusion of erbium ions does not affect our results. In addition, we demonstrate that saturation of the erbium ${}^4{I}_{13/2}\to {}^4{I}_{15/2}$ transition may lead to an exaggerated estimate of the interaction distance.

Figure 1

(a) Schematic view of the sample preparation steps. (b) Transmission electron micrograph of the sample with a nominal layer thickness $d$ $=$ 3.0 nm after annealing at 700${}^{\circ }\mathrm{C}$. The dark horizontal lines correspond to the Si layers.

Figure 2

The optical characterization apparatus. The sample is mounted in a sample holder and subjected to pump light ($p$ polarized), a focus of which can be adjusted down to approximately 100 microns by translation of a cylindrical lens. Different spectral parts of the emitted light is collected simultaneously by two monochromators. One is equipped with a charge-coupled device (CCD), and the other with a photomultiplier tube (PMT).

Figure 3

The emitted spectrum from the sample with nominal Er-layer thickness of 3.0 nm under continuous pumping at 405 nm with intensity $I={10}^4\mathrm{W}/{\mathrm{m}}^2$. The part below 1400 nm has been multiplied by a factor of ten for clarity.

Figure 4

The photoluminescence yield at 1535 nm (circles) and in the range 700–950 nm (squares) vs pumping intensity for the sample with 3.0 nm Er-layer thickness. The vertical position of each data set is arbitrary. The straight lines denote proportionality fits of the data, the dashed line is fitted to the five lowest-intensity points only. The solid line through the circles represents a fit to the expression in Eq. (1) with $A$ fixed at the slope of the dashed line and the saturation intensity fitted to $I_{\mathit{sat}}=33\mathrm{W}/{\mathrm{m}}^2$. The data points corresponding to the low and high intensities, $I_{\mathit{low}}=3.0\mathrm{W}/{\mathrm{m}}^2$ and $I_{\mathit{high}}=0.94\times {10}^4\mathrm{W}/{\mathrm{m}}^2$, are marked on the graph.

Figure 5

The dependence of the 1535-nm normalized PL yield on the Er-layer thickness. Panels (a) and (b) show the luminescence yield obtained at low and high pump intensities, respectively. The data in panel (c) is the PL yield ratio $Y_{I=I_{\mathit{high}}}/Y_{I=I_{\mathit{low}}}$, i.e., the data of panel (b) normalized to the data of panel (a). The solid line in panel (c) is a fit to Eq. (2), and the dashed lines in panels (a) and (b) show the in-growth function $y=A[1-\exp (-x/x_0)]$ with $x_0$ fixed at the fitted values $x_{\mathit{low}}$ and $x_{\mathit{high}}$, respectively.

Figure 6

The dependence of the integrated photoluminescence yield in the range 700–950 nm on the Er-layer thickness. Panels (a) and (b) show the yield obtained at low and high pump intensities, respectively. The solid lines are fits to a double-exponential model.

Figure 7

(a) Schematics of the diffusion model. An initial distribution of Er is located at $x_0$ within the thickness $d{x}_0$. A subsequent heat treatment broadens the Er concentration (solid line, exaggerated in magnitude) according to Eq. (A1). The part, which falls below $x=0$, is assumed to become optically active. (b) The solid line denotes the analytical sheet carrier density of Eq. (A3). The circles show a simple exponential in-growth with characteristic length $L$.