Search for: All records

A theoretical model that can be used to simultaneously fit the I–V characteristics and spectral optical responsivity of Ge-like pin diodes is described in detail and validated experimentally using specially fabricated Ge- and Ge1−ySny devices. The model combines a numerical solution of the basic semiconductor transport equations with a rigorous calculation of the optical generation rate that accounts for multiple reflections in the device structure multilayers. The results can be used to quantify the reduction of photocurrent associated with recombination centers for full optimization of the device structure. 

In this review, the nonparabolicity of the light-hole and electron bands at the Γ-point in cubic diamond or zinc blende semiconductors is derived from Kane’s 8×8k→⋅p→ model in the large spin–orbit splitting approximation. Examples of several approximations are given with InSb as an example, and their accuracy is discussed. To determine the temperature dependence of the effective masses and the nonparabolicity parameters, the unrenormalized bandgap must be utilized. This includes only the redshift of the bandgap due to thermal expansion, not the renormalization due to deformation-potential electron-phonon coupling. As an application of this method, the chemical potential and the charge carrier concentration of intrinsic InSb are calculated from 50 to 800 K and compared with electrical and optical experiments. These results are also relevant for other semiconductors with small bandgaps as needed for mid-infrared detector applications. 

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. 

GeSn films were simultaneously deposited on Si (100), Si (111), c-plane sapphire (Al2O3), and fused silica substrates to investigate the impact of the substrate on the resulting GeSn film. The electronic, structural, and optical properties of these films were characterized by temperature-dependent Hall-effect measurements, x-ray diffractometry, secondary ion mass spectrometry, and variable angle spectroscopic ellipsometry. All films were polycrystalline with varying degrees of texturing. The film on Si (100) contained only GeSn (100) grains, 40.4 nm in diameter. The film deposited on Si (111) contained primarily GeSn (111) grains, 36.4 nm in diameter. Both films deposited on silicon substrates were fully relaxed. The layer deposited on Al2O3 contained primarily GeSn (111) grains, 41.3 nm in diameter. The film deposited on fused silica was not textured, and the average grain size was 35.0 nm. All films contained ∼5.6 at. % Sn throughout the layer, except for the film deposited on Al2O3, which contained 7.5% Sn. The films deposited on Si (111), Al2O3, and fused silica exhibit p-type conduction over the entire temperature range, 10–325 K, while the layer deposited on the Si (100) substrate shows a mixed conduction transition from p-type at low temperature to n-type above 220 K. From ∼175 to 260 K, both holes and electrons contribute to conduction. Texturing of the GeSn film on Si (100) was the only characteristic that set this film apart from the other three films, suggesting that something related to GeSn (100) crystal orientation causes this transition from p- to n-type conduction. 

A practical quantitative model is presented to account for the I–V characteristics of pin diodes based on epitaxial Ge-like materials. The model can be used to quantify how the different material properties and recombination mechanisms affect the diode performance. The importance of dislocations, non-passivated defects, and residual intrinsic layer doping in determining the qualitative shape of the I–V curves is discussed in detail. Examples are shown covering literature diodes as well as diodes fabricated with the purpose of validating the theoretical effort. 

Most cubic semiconductors have threefold degenerate p-bonding valence bands and nondegenerate s-antibonding conduction bands. This allows strong interband transitions from the valence to the conduction bands. On the other hand, intervalence band transitions within p-bonding orbitals in conventional p-type semiconductors are forbidden at k=0 and, therefore, weak, but observable. In gapless semiconductors, however, the s-antibonding band moves down between the split-off hole band and the valence band maximum due to the Darwin shift. This band arrangement makes them three-dimensional topological insulators. It also allows strong interband transitions from the s-antibonding valence band to the p-bonding bands, which have been observed in α-tin with Fourier-transform infrared spectroscopic ellipsometry [Carrasco et al., Appl. Phys. Lett. 113, 232104 (2018)]. This manuscript presents a theoretical description of such transitions applicable to many gapless semiconductors. This model is based on k→⋅p→ theory, degenerate carrier statistics, the excitonic Sommerfeld enhancement, and screening of the transitions by many-body effects. The impact of nonparabolic bands is approximated within Kane’s 8×8k→⋅p→-model by adjustments of the effective masses. This achieves agreement with experiments.