Search for: All records

Nickel and aluminum ohmic contacts were formed on p-doped GeC and GeCSn epitaxial films with ∼1%C. When a 40 nm p-GeC contact layer was added to p-Ge, annealed contact resistivity (Rc) dropped by 87% to 9.3 × 10−7 Ω cm2 for Al but increased by 32% to 2.9 × 10−5 Ω cm2 for Ni. On the other hand, thick films of GeCSn, which showed lower active doping, had contact resistivities of 4.4 × 10−6 Ω cm2 for Al and 1.4 × 10−5 Ω cm2 for Ni. In general, Al contacts were better than Ni, regardless of anneal, and were further improved by adding carbon. Annealing reduced Rc for both Ni and Al contacts to GeCSn by 4×, 2× for Al on GeC, and 5 orders of magnitude for Ni on GeC. It is speculated that C forms bonds with Ni that inhibit diffusion of Ni into the Ge, thus preventing the formation of low-resistance nickel germanide. Adding C, either as bulk GeCSn or as GeC contact layers, seems to significantly reduce the contact resistivity for Al contacts when compared to bulk Ge of comparable doping. 

Direct bandgap group IV materials could provide intimate integration of lasers, amplifiers, and compact modulators within complementary metal–oxide–semiconductor for smaller, active silicon photonics. Dilute germanium carbides (GeC) with ∼1 at. % C offer a direct bandgap and strong optical emission, but energetic carbon sources such as plasmas and e-beam evaporation produce defective materials. In this work, we used CBr4 as a low-damage source of carbon in molecular beam epitaxy of tin-free GeC, with smooth surfaces and narrow x-ray diffraction peaks. Raman spectroscopy showed substitutional incorporation of C and no detectable sp2 bonding from amorphous or graphitic carbon, even without surfactants. Photoluminescence shows strong emission compared with Ge. 

GeSnC alloys offer a route to direct bandgap semiconductors for CMOS-compatible lasers, but the use of CBr4 as a carbon source was shown to reduce Sn incorporation by 83%–92%. We report on the role of thermally cracked H in increasing Sn incorporation by 6x–9.5x, restoring up to 71% of the lost Sn, and attribute this increase to removal of Br from the growth surface as HBr prior to formation of volatile groups such as SnBr4. Furthermore, as the H flux is increased, Rutherford backscattering spectroscopy reveals a monotonic increase in both Sn and carbon incorporation. X-ray diffraction reveals tensile-strained films that are pseudomorphic with the substrate. Raman spectroscopy suggests substitutional C incorporation; both x-ray photoelectron spectroscopy and Raman suggest a lack of graphitic carbon or its other phases. For the lowest growth temperatures, scanning transmission electron microscopy reveals nanovoids that may account for the low Sn substitutional fraction in those layers. Conversely, the sample grown at high temperatures displayed abrupt interfaces, notably devoid of any voids, tin, or carbon-rich clusters. Finally, the surface roughness decreases with increasing growth temperature. These results show that atomic hydrogen provides a highly promising route to increase both Sn and C to achieve a strongly direct bandgap for optical gain and active silicon photonics.