It is shown that compressively strained Ge1−xSnx/Ge quantum wells (QWs) grown on a Ge substrate with 0.1 ≤ x ≤ 0.2 and width of 8 nm ≤ d ≤ 14 nm are a very promising gain medium for lasers integrated with an Si platform. Such QWs are type-I QWs with a direct bandgap and positive transverse electric mode of material gain, i.e. the modal gain. The electronic band structure near the center of Brillouin zone has been calculated for various Ge1−xSnx/Ge QWs with use of the 8-band kp Hamiltonian. To calculate the material gain for these QWs, occupation of the L valley in Ge barriers has been taken into account. It is clearly shown that this occupation has a lot of influence on the material gain in the QWs with low Sn concentrations (Sn < 15%) and is less important for QWs with larger Sn concentration (Sn > 15%). However, for QWs with Sn > 20% the critical thickness of a GeSn layer deposited on a Ge substrate starts to play an important role. Reduction in the QW width shifts up the ground electron subband in the QW and increases occupation of the L valley in the barriers instead of the Γ valley in the QW region.

Unstrained Ge1−xSnx layers of various Sn concentration (1.5%, 3%, 6% Sn) and Ge0.97Sn0.03 layers with built-in compressive (ε = −0.5%) and tensile (ε = 0.3%) strain are grown by molecular beam epitaxy and studied by electromodulation spectroscopy (i.e., contactless electroreflectance and photoreflectance (PR)). In order to obtain unstrained GeSn layers and layers with different built-in in-plane strains, virtual InGaAs substrates of different compositions are grown prior to the deposition of GeSn layers. For unstrained Ge1−xSnx layers, the pressure coefficient for the direct band gap transition is determined from PR measurements at various hydrostatic pressures to be 12.2 ± 0.2 meV/kbar, which is very close to the pressure coefficient for the direct band gap transition in Ge (12.9 meV/kbar). This suggests that the hydrostatic deformation potentials typical of Ge can be applied to describe the pressure-induced changes in the electronic band structure of Ge1−xSnx alloys with low Sn concentrations. The same conclusion is derived for the uniaxial deformation potential, which describes the splitting between heavy-hole (HH) and light-hole (LH) bands as well as the strain-related shift of the spin-orbit (SO) split-off band. It is observed that the HH, LH, and SO related transitions shift due to compressive and tensile strain according to the Bir-Pikus theory. The dispersions of HH, LH, and SO bands are calculated for compressive and tensile strained Ge0.97Sn0.03 with the 8-band kp Hamiltonian including strain effects, and the mixing of HH and LH bands is discussed. In addition, the dispersion of the electronic band structure is calculated for unstrained Ge1−xSnx layers (3% and 6% Sn) at high hydrostatic pressure with the 8-band kp Hamiltonian, and the pressure-induced changes in the electronic band structure are discussed.

The pseudomorphic growth of Ge1-xSnxon Ge causes in-plane compressive strain, which degrades the superior properties of the Ge1-xSnx alloys. Therefore, efficient strain engineering is required. In this article, we present strain and band-gap engineering in Ge1-xSnxalloys grown on Ge a virtual substrate using post-growth nanosecond pulsed laser melting (PLM). Micro-Raman and X-ray diffraction show that the initial in-plane compressive strain is removed. Moreover, for PLM energy densities higher than 0.5 J cm-2, the Ge0.89Sn0.11 layer becomes tensile strained. Simultaneously, as revealed by Rutherford Backscattering spectrometry, cross-sectional transmission electron microscopy investigations and X-ray diffraction the crystalline quality and Sn-distribution in PLM-treated Ge0.89Sn0.11 layers are only slightly affected. Additionally, the change of the band structure after PLM is confirmed by low-temperature photoreflectance measurements. The presented results prove that post-growth ns-range PLM is an effective way for band-gap and strain engineering in highly-mismatched alloys.

No abstract