Search for: All records

Ge1−x−ySixSny alloys were grown on Ge buffers via reactions of SnH4 and GeH3Cl. The latter is a new CVD source designed for epitaxial development of group-IV semiconductors under low thermal budgets and CMOS-compatible conditions. The Ge1−x−ySixSny films were produced at very low temperatures between 160 and 200 °C with 3%–5% Si and ∼5%–11% Sn. The films were characterized using an array of structural probes that include Rutherford backscattering, x-ray photoelectron spectroscopy, high-resolution x-ray diffraction, scanning transmission electron microscopy, and atomic force microscopy. These studies indicate that the films are strained to Ge and exhibit defect-free microstructures, flat surfaces, homogeneous compositions, and sharp interfaces. Raman was used to determine the compositional dependence of the vibrational modes indicating atomic distributions indistinguishable from those obtained when using high-order Ge hydrides. For a better understanding of the growth mechanisms, a parallel study was conducted to investigate the GeH3Cl applicability for synthesis of binary Ge1−ySny films. These grew strained to Ge, but with reduced Sn compositions and lower thicknesses relative to Ge1−x−ySixSny. Bypassing the Ge buffers led to Ge1−ySny-on-Si films with compositions and thicknesses comparable to Ge1−ySny-on-Ge; but their strains were mostly relaxed. Efforts to increase the concentration and thickness of Ge1−ySny-on-Si resulted in multiphase materials containing large amounts of interstitial Sn. These outcomes suggest that the incorporation of even small Si amounts in Ge1−x−ySixSny might compensate for the large Ge–Sn mismatch by lowering bond strains. Such an effect reduces strain energy, enhances stability, promotes higher Sn incorporation, and increases critical thickness. 

We report the synthesis of Ge1−ySny films containing 6%–13% Sn directly on Si(100) for monolithic integration applications, circumventing the use of conventional Ge-buffer layers. The films are produced in a gas source molecular epitaxy chamber at ultralow temperatures of 185–210 °C and a pressure of 10−5 Torr by the reactions of pure vapor Ge4H10 and SnD4 or SnH4 without carrier gases. Very small amounts of Si, incorporated via the Si4H10 precursor, can be used to improve the structural properties. All samples were characterized by XRD, RBS, IR-ellipsometry, AFM, and TEM, indicating the formation of monocrystalline single-phase films with relatively low defectivity and flat surfaces. A notable highlight is that the residual strains of the alloy layers are much lower compared to those grown on Ge buffers and can be further reduced by rapid thermal annealing without decomposition, indicating that growth on bare silicon should produce bulklike, high Sn content alloys that cannot be accessed using Ge buffers. N-type analogs of the above samples doped with phosphorus were also produced using P(SiH3)3 as the in situ dopant precursor. The results collectively illustrate the potential of our chemistry-based method to generate good quality Ge1−ySny layers directly on large area Si wafers bypassing Ge buffers that typically lead to complications such as multiple hetero-interfaces and epitaxial breakdown at high Sn concentrations. 

We describe an alternative strategy to the fabrication of Ge–Sn based materials on Si by using chlorogermane (GeH 3 Cl) instead of the specialty Ge hydrides (Ge 2 H 6 , Ge 3 H 8 , Ge 4 H 10 ) currently employed as ultra-low temperature sources of Ge. This simpler and potentially more practical chlorinated derivative is obtained in high yields and in research-grade purity by direct reactions of commercial GeH 4 and SnCl 4 and exhibits favorable physical and chemical properties that make it an effective source of Ge for a wide range of chemical vapor deposition (CVD) processing conditions. As a proof-of concept, we have employed GeH 3 Cl to demonstrate deposition of pure Ge and GeSn hetero-structures on large-area Si wafers, at conditions compatible with current specialty methods for next generation technologies but with higher deposition efficiency, ensuring an optimal use of the Ge feedstock. In the case of pure Ge, GeH 3 Cl has enabled growth of thick and uniform Ge layers with flat surfaces and relaxed microstructures at 330–360 °C, exhibiting lower residual doping than obtained by alternate Ge hydride methods. GeH 3 Cl allows for in situ doping with the same facility as the Ge hydrides, and this has enabled the design and fabrication of homo-structure pin photodetectors exhibiting low dark current densities and closer to ideal optical collection efficiencies when compared to devices produced by other Ge-on-Si approaches. In the case of GeSn, the high reactivity of GeH 3 Cl toward Sn hydrides has enabled the formation of mono-crystalline alloy layers at ultra-low temperatures between 200–300 °C and conditions akin to molecular beam epitaxy (MBE). Combined, these results suggest an intriguing potential for this new CVD process in the device-application space. The deployment of GeH 3 Cl as a highly reactive low-temperature Ge-source could not only improve on the current wasteful methods that use GeH 4 , but also eliminate the need for the higher-cost polygermanes.