1844
Stability of Pseudomorphic and Compressively Strained Ge1-XSnx Thin Films under Rapid Thermal Annealing

Thursday, 9 October 2014: 13:40
Expo Center, 1st Floor, Universal 8 (Moon Palace Resort)
B. R. Conley, A. Mosleh, S. A. Ghetmiri, W. Du (University of Arkansas), G. Sun, R. Soref (University of Massachusetts Boston), J. Margetis, J. Tolle (ASM America), H. A. Naseem, and S. Q. Yu (University of Arkansas)
Silicon lacks the higher mobility and direct band gap needed for faster electronics or active photonic applications. This shortcoming has pushed research towards the incorporation of other material systems, such as Ge1-xSnx, for increased mobility and possible direct band gap. Increasing the mole fraction of Sn will eventually lead to an indirect-to-direct crossover around 9 % Sn [1]. Additionally, the growth temperatures and processing steps for Ge1-xSnxfilms are backend Si complementary metal oxide semiconductor (CMOS) compatible.

The challenges facing Ge1-xSnx grown by the chemical vapor deposition (CVD) technique are limitations of the Sn gaseous precursors. The deuterated stannane (SnD4) precursor has demonstrated monolithic growth of Ge1-xSnx directly on Si, but this gas is unstable at room temperature and can decompose within days. A more commercially accessible and stable precursor, stannic chloride (SnCl4), is stable under standard room temperature storage. Its role in Ge1-xSnx growth has started to be incorporated using commercially available CVD reactors [2]. However, an initial Ge buffer layer grown on Si prior to the Ge1-xSnx film is needed to mediate the Si-Cl reactivity. The lattice mismatch at the Ge1-xSnx/Ge interface is reduced compared to Si, which allows compressively strained and pseudomorphic Ge1-xSnxthin films to be grown.

The growth of Ge1-xSnx thin films on Ge is done at temperatures below 400 °C; however Sn inter-diffusion between the Ge1-xSnx and Ge buffer layer during high temperature processing steps presents an unknown for future device reliability. This inter-diffusion of Sn between growth layers was observed already for a 500 °C, 30 minute anneal of Ge1-xSnx/Ge/Si [2]. Other post-growth annealing studies of the thermal stability of thin film Ge1-xSnx/Ge have focused only on single Sn mole fractions and thin films strained below the critical thickness [3-5]. Future applications in Si optoelectronics will require a more detailed study of thicker Ge1-xSnx layers and of multiple Sn mole fractions. In this paper, we investigate on the post-growth RTA stability of Ge1-xSnx/Ge/Si thin films grown with six different Sn mole fractions from 0.009 to 0.07. The Sn mole fraction and film thicknesses are listed in Table 1. The Ge1-xSnx thin films were grown partially pseudomorphic on a Ge buffer layer using an ASM Epsilon®reduced pressured CVD reactor. The lattice constants measured for each sample using X-ray diffraction rocking curves in Fig. 1 show a decrease in strain percentage at increased temperatures. Atomic force microscopy measured the film surface roughness change from a cross-hatched on the as-grown, to clumps of Sn-segregates for samples subjected to high temperature RTA. The samples that were observed to be relaxed as measured by XRD, also showed an increased red-shift for the Ge-Ge LO phonon measured by Raman spectroscopy. This red-shift indicates a higher incorporation of Sn atoms in the lattice for those annealed samples. Room temperature photoluminescence (PL) spectra indicate initial film degradation for short time anneals at constant temperature, but increased time shows film quality improvements [Fig. 2]. An increase in the PL emission intensity points to a reduced number of non-radiative traps.

Table I. Ge1-xSnxsample compositions and thickness

Sn mole fraction

Ge1-xSnx  thickness (nm)

0.009

327

0.026

257

0.027

173

0.032

76

0.032

128

0.06

45

0.07

211

1.     G. Grzybowski, et al., Appl. Phys. Lett. 101, 072105 (2012).

2.     B. Vincent, et al.,  Appl. Phys. Lett. 99, 152103 (2011).

3.     M.-Y. Ryu, et al., Appl. Phys. Lett. 102, 171908 (2013).

4.     H. Li, et al., Appl. Phys. Lett. 102, 251907 (2013).

5.     R. Chen, et al., J. of Crystal Growth 365, 29-34 (2013).