The implementation of new materials on Si substrates has been a subject of intensive research since decades. Recently, group IV germanium tin (GeSn) and silicon germanium tin (SiGeSn) have triggered huge scientific interest due to the unique features of these materials [1,2]. However, their epitaxial growth is a complex task due to the low solid solubility of Sn in Ge and the limited thermal stability of such alloys. The aim of this work was first to set up a relatively fast and non-destructive experimental protocol to characterize SiGeSn ternary layers by combining X-ray diffraction (XRD) with wavelength dispersive X-ray fluorescence (WDXRF). The surface morphology of the grown layers was otherwise assessed through atomic force microscopy (AFM). Afterwards, different growth conditions were tested in order to optimize growth conditions.

Binary and ternary alloys were grown on 2.5 µm thick strain relaxed Ge buffers on 200 mm Si(001) substrates in a reduced pressure chemical vapor deposition reactor from Applied Materials. Close to 30 nm thick GeSn and GeSi pseudomorphic layers were characterized first through XRD. Figure 1 shows the omega-2theta scans around the symmetric (004) reflections. Through these scans, the layer thickness and composition could be calculated. Afterwards, WDXRF measurements were carried out on the same samples and the intensity of the specific element drawn as function of the layer thickness multiplied by this element composition in the layer, as shown in Figure 2. The fitting lines served afterwards to find the composition of each element in the SiGeSn alloys.

Ternary alloys were then grown at 349°C and 100 Torr using Si₂H₆, SnCl₄ and Ge₂H₆ precursors and H₂ as a carrier gas. In that first set of samples, we investigated the influence of the Si₂H₆ flow. Thus, we varied the F(Si₂H₆ )/F(H₂) mass-flow ratio from 0 up to 5.2×10^-3. Figure 3 shows the omega-2theta scans around the (004) XRD order for the various SiGeSn/Ge/Si(001) heterostructures grown. Interestingly, the SiGeSn XRD peak did not significantly shift even though the Si₂H₆ flow varied considerably. Using Figure 2 XRF calibration curves and the layer thickness extracted from XRD scans, the concentration of Si and Sn in the layers could be calculated. Resulting values are shown as functions of the F(Si₂H₆)/F(H₂) mass-flow ratio in Figure 4. As expected, the increase of the Si₂H₆ flow leads to an increase of the Si content in the layers (from 0% up to nearly 21%). However, this increase is accompanied by an increase of the Sn content in the layer, from ~ 5% up to ~ 11% (Figure 4). This likely explains the stability of the SiGeSn peak’s angular position in XRD profiles. The initial compressive strain in the layer (for Si₂H₆ = 0 sccm i.e. for the GeSn layer) is kept despite Si₂H₆ flux variations. Figure 5 shows the growth rate as a function of the F(Si₂H₆)/F(H₂) mass-flow ratio. There is a clear drop of the growth rate when the Si₂H₆ flux increases (from ~ 32 down to ~ 20 nm min.^-1). The surface morphology was otherwise assessed using AFM. Results are provided in Figure 6 and Figure 7. The AFM images highlight that the surface roughness increases when the Si₂H₆ flux increases. Above a certain F(Si₂H₆)/F(H₂) value, a catastrophic degradation of the surface morphology occurred which could be related to Sn segregation (cf. Figure 6f).

In order to better understand the growth of ternary SiGeSn pseudomorphic layers, we have explored a wide range of growth conditions. First, the growth temperature was varied from 313°C to 409°C in 12°C steps, all the other parameters being fixed (pressure (100 Torr), fluxes…). This study allowed us to determine the optimal growth temperatures. Afterwards, three growth temperatures were selected (349°C, 337°C and 325°C) and the Si₂H₆ flow varied, while keeping constant the SnCl₄ and Ge₂H₆ flows. The layers grown under these conditions are currently under investigations and detailed results will be presented during the conference.

[1] S. Wirths et al, Nat. Photonics 9 (2015) 88.

[2] S. Wirths et al, Thin Solid Films 557 (2014) 183.