Silicon is the basis for a multibillion dollar industry and has been the preferred platform for integrated electronics for decades. Unfortunately, the band structure of Si has limited its use in infrared applications. The 1.12 eV band gap restricts the range of Si photodetectors to wavelengths of 1100 nm or lower, and the indirect nature of the band gap has prevented the realization of a Si-based laser. Although there are devices available that cover these wavelength ranges, they are typically made from exotic materials such as InGaAs or HgCdTe. These materials are incompatible with Si processing, so the device cannot be integrated with other Si-based electronics on the same chip. Furthermore, many of the devices require cryogenic temperature operation. These unfortunate conditions mean that the devices and systems that use them are prohibitively expensive, which prevents their widespread adoption for both military and commercial applications. Si-compatible materials that have desirable optical properties beyond the band gap of silicon could produce inexpensive devices and systems with applications in telecommunications, low-light imaging, targeting and acquisition, photonic integrated circuit. A number of approaches have been attempted for producing infrared devices on silicon, but they have achieved limited success. The integration of III-V materials on Si has been an active research area for a number of years, and it has yielded limited success. The flip chip wafer bonding and packaging process still requires a separate facility for production of the III-V material, so the costs cannot be brought down to the level that would be achieved with a Si-compatible material. Another approach has been to use other Group IV materials that are grown epitaxially on Si. SiGe alloys were explored for this purpose, but for the concentrations of Ge that can readily be achieved, the band gap still does not extend far enough into the infrared. Furthermore, SiGe is still an indirect band gap material, so it is not suitable for producing a laser. Pure Ge grown epitaxially on Si has also been a recent area of research, and this has resulted in some success. Detectors operating at 1550 nm have been achieved by using tensile strain to lower the direct band gap to enhance absorption at that wavelength, but this approach cannot be extended to further wavelengths due to the limits of achievable strain. Tensile-strained Ge has also been used to demonstrate optically- and electrically-pumped lasers. A new class of Group IV materials that are grown on Si or Ge-buffered Si is the GeSn alloy system. Pure Ge is an indirect band gap semiconductor, but the difference between the conduction band minima at the L- and Γ-points is only 140 meV. The diamond cubic structure of Sn, α-Sn, is a semimetal. When Sn is substituted for Ge in the Ge lattice, a hybrid band structure is formed, yielding a material with a tunable band gap. For Sn concentrations around 10% or lower, the band structure is still Ge-like with the lowest conduction band minimum at the L-point. However, the energy difference between the L- and Γ-minima reduces with increasing Sn concentration. For higher Sn concentrations, the minimum at the Γ-point becomes the lowest minimum, thereby creating a direct band gap semiconductor. In this talk, I will discuss our efforts to enhance optical emission in GeSn through the use of waveguide architectures and the introduction of n-type dopants. Our results obtained from optically-pumped waveguides indicate that we have achieved amplified spontaneous emission, where we see an exponential dependence of output power versus input power, but we do not see a narrowing of the optical emission spectrum. Instead, the entire spontaneous emission spectrum is amplified. This does produces some evolution in the spectrum, and we attempt to understand the emission spectrum through mathematical and physical modeling.
Read full abstract