Structure and Optoelectronic Properties of Atomically Random Sn-Rich Gesn Semiconductors

The monolithic integration of group IV optoelectronic devices on silicon platform is a long-sought-for paradigm holding the key to a wealth of opportunities in ultrafast data transfer, low-power electronics, energy conversion, and sensing, to name a few. This all-group IV integrated photonics has recently gained a renewed interest with the re-emergence of direct band gap germanium-tin (GeSn) semiconductors. GeSn layers with Sn contents above 10% are commonly grown on a Ge virtual substrate (VS) on Si wafers. In recent years, digermane (Ge2H6) has been the canonical precursor to accomplish the growth of various Sn-rich GeSn layers and heterostructures.[1,2] However, Monogermane (GeH4) has recently been proposed as an alternative precursor for large-scale, CMOS-compatible growth of GeSn semiconductor lasers.[3-4] Here, we present the the growth and properties of metastable Sn-rich GeSn layers from the wafer-level down to the atomic scale.[5] The chemical vapor deposition (CVD) growth of GeSn multi-layer structures with Sn content in the 7-18% range was performed on a ~600 nm-thick Ge layer on 4-inch silicon wafers using Monogermane (GeH4) and tin-tetrachloride (SnCl4) precursors. The Sn incorporation is controlled by the change in temperature, ranging from 320°C, 300 °C, and 280 °C for the bottom (BL), middle (ML), and top (TL) GeSn layers, respectively. This leads to a BL/ML/TL stacking in Fig. 1a with Sn contents of 8.2%/11-13.5%/18%. Using transmission electron microscopy (TEM) measurements a TL/ML/BL GeSn layers stacking with a total thickness of ~380 nm grown on a ~600 nm-thick Ge-VS is observed. No threading dislocations propagating toward the GeSn TL are observed, while misfit and edge dislocations are confined at the Ge-GeSn interface and within the first 60-70 nm of the GeSn stacking. Herein, the lower Sn content BL and ML seem to minimize or even suppress the propagation of dislocations towards the defect-free 18% Sn TL. Atom probe tomography (APT) measurements are performed to establish an atomistic-level understanding of the evolution of Sn content across the grown layer. The APT profile for the Sn incorporation in Fig. 1b reveals the presence of abrupt interfaces between the different GeSn layers, with the highest incorporation of 18% in the TL. To provide details regarding the short-range atomic distribution on the scale of few lattice constants a series of four different statistical APT analyses will be discussed: frequency distribution analysis; partial radial distribution function; nearest-neighbor analysis; and iso-concentration surfaces.[5-6] The APT analysis demonstrates the growth of a GeSn random alloy, with a perfect random distribution of Sn atoms within the grown layers, and without the presence of short-range ordering effects and Sn precipitates. Interestingly, the defect-free highest incorporation of Sn is achieved at a growth rate (~1.3 nm/min) that is one order of magnitude lower than other studies where either GeH4 or Ge2H6 precursors were used.[1-3] It is clear that the use of GeH4 allows for additional set of growth parameters were a high incorporation of Sn still be achieved with a drastic decrease reduction (one order of magnitude) in the growth rate. The presence of multiple experimental growth conditions with extremely different growth rates suggests that the growth rate is a marginal factor in the Sn incorporation process, whereas the strain relaxation using lower Sn content buffered layers (BL/ML) will be discussed as a key factor in enhancing the incorporation of Sn. We therefore highlight the importance of reducing strain at the growth surface while keeping a low growth temperature, which can enhance the incorporation of Sn. This behavior will facilitate the growth of GeSn multilayer structures enabling more advanced strain and band structure engineering. The sharp interfaces across GeSn layers pave the way for achieving a high degree of control of the growth of GeSn-based multi-quantum wells heterostructures for mid-IR sources and detectors. Furthermore, by combining theoretical calculations and experiments, the presentation will also discuss the influence of structure and composition on the optoelectronic properties of the as-grown GeSn layers. Acknowledgements The authors thank J. Bouchard for the technical support, and NSERC Canada (Discovery, SPG, and CRD Grants), Canada Research Chair, Canada Foundation for Innovation, Mitacs, FQRNT, and MRIF Québec for support. References [1] S. Wirths, et al., Nat. Photonics 9 (2), 88–92 (2015). [2] J. Aubin, et al., Semicond. Sci. Technol. 32 (9), 94006 (2017). [3] J. Margetis, et al., Mater. Sci. Semicond. Process. 70, 38–43 (2017). [4] J. Margetis, et al., ACS Phot. 5 (3), 827-833 (2018). [5] S. Assali, et al., Under review. [6] S. Mukherjee, et al., Phys. Rev.B 95(16), 161402 (2017). Figure 1

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.