Vacancy complexes in nonequilibrium germanium-tin semiconductors

Abstract

Understanding the nature and behavior of vacancy-like defects in epitaxial germanium-tin (GeSn) metastable alloys is crucial to elucidate the structural and optoelectronic properties of these emerging semiconductors. The formation of vacancies and their complexes is expected to be promoted by the relatively low substrate temperature required for the epitaxial growth of GeSn layers with Sn contents significantly above the equilibrium solubility of 1 at.%. These defects can impact both the microstructure and charge carrier lifetime. Herein, to identify the vacancy-related complexes and probe their evolution as a function of Sn content, depth-profiled pulsed low-energy positron annihilation lifetime spectroscopy and Doppler broadening spectroscopy were combined to investigate GeSn epitaxial layers with Sn content in the 6.5-13.0 at.% range. The investigated samples were grown by chemical vapor deposition method at temperatures between 300 and 330 °C. Regardless of the Sn content, all GeSn samples showed the same depth-dependent increase in the positron annihilation line broadening parameters, relative to that of epitaxial Ge reference layers. These observations confirmed the presence of open volume defects in as-grown layers. The measured average positron lifetimes were found to be the highest (380-395 ps) in the region near the surface and monotonically decrease across the analyzed thickness, but remain above 350 ps.All GeSn layers exhibit lifetimes that are 85 to 110 ps higher than those recorded for Ge reference layers. Surprisingly, these lifetimes were found to decrease as Sn content increases in GeSn layers.

All‐Group IV Transferable Membrane Mid‐Infrared Photodetectors

“…A larger residual compressive strain ε || ≈ −1.3% is, however, measured in this set of samples (Figure 1e) without compromising the crystalline quality and composition uniformity. [ 35,45 ]…”

“…The dark current as a function of the applied bias for all samples is shown in Figure 4c, indicating a steep increase as Sn content increases, eventually reaching a factor ≈8 increase in Ge 0.83 Sn 0.17 . The higher leakage current in GeSn is believed to originate from the presence of point defects, such as vacancies and vacancy complexes, [ 52 ] which have been suggested to be responsible for the unintentional p‐type doping in GeSn that is an order of magnitude higher than in epitaxial Ge. [ 55–57 ] It is worth mentioning that other factors could play a role in the measured dark current, such as the absence of a surface passivation layer [ 43 ] and the parallel conductance associated with the Ge‐VS/Si stacking.…”

“…For example, photocurrent signal at wavelengths beyond 3.0 µm would only be associated to the optical absorption in Ge 0.83 Sn 0.17 , while at shorter wavelengths the underlying lower Sn content layers also contribute to the absorption thus enhancing the PD responsivity. Trapping centers in GeSn are mainly located in the defective low Sn content layers, [ 45 ] thus they are expected to play a minor role in the PD performance. As an indication of this, similar responsivities were obtained at a fixed detection wavelength in GeSn PDs with different layer thickness and composition (see 2.0–2.5 µm range in Figure 4d).…”

“…Si (100) wafer in a low‐pressure CVD reactor using ultrapure H 2 carrier gas, 10% monogermane (GeH 4 ) and tin‐tetrachloride (SnCl 4 ) precursors. [ 35,45,64 ] First, the Ge‐VS was grown using a two‐temperature step process at 450 and 600 °C with a thickness of 1.6 µm for the Ge 0.895 Sn 0.105 and Ge 0.87 Sn 0.13 samples, while in the Ge 0.83 Sn 0.17 sample the Ge‐VS was grown at 450 °C with a thickness of 0.6 µm. Next, a postgrowth thermal cyclic annealing (>800 °C) process was performed to improve the crystalline quality of the Ge‐VS.…”

“…Indeed, GeSn semiconductors are inherently metastable and typically exhibit a compositional gradient and a large compressive strain resulting from the canonical, lattice‐mismatched growth on Ge. [ 44,45 ] This compressive strain increases the Sn critical content to reach the indirect‐to‐direct bandgap transition thus exacerbating the growth hurdles. Moreover, at a fixed Sn content the corresponding optical emission and detection range of compressively strained layers shifts to shorter wavelengths in comparison to fully relaxed layers.…”