| Summary: | Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2009. === Cataloged from PDF version of thesis. === Includes bibliographical references (p. 202-211). === The rapid growing needs for high data transmission bandwidth challenge the metal interconnection technology in every area from chip-level interconnects to long distance communication. Silicon photonics is an ideal platform for the implementation of optical interconnection capable of high bandwidth and low power consumption by integrating electronic and photonic devices on silicon. Many optical components in silicon photonics have been extensively studied, among which a silicon-based laser is arguably the most challenging element. This thesis mainly focuses on using engineered germanium as the optically active material for silicon-based light emitters with many potential benefits: Si-CMOS compatibility (both material and processing), electrical injection capability, and direct gap emission at technologically important 1.55 pm telecommunication band. Tensile-strained n+ germanium is capable of behaving like a direct band gap material owing to the direct band gap shrinkage upon tensile strain and the state-filling in the indirect L valleys with extrinsic electrons from n-type dopants. Our theoretical calculation using a direct band-to-band transition model has shown great benefit of tensile strain and n-type doping on the direct gap optical gain characteristics. By considering free carrier absorption which dominates the optical loss we have proven net gain can be achieved in 0.25% tensile-strained Ge with n-type doping concentration in a range of 1019 to mid-1020 cm-3. The injection threshold of the net gain is about 1018 cm-3 which can readily be achieved with either optical pumping or electrical pumping. === (cont.) The net gain is in favor of the raise of temperature in a large injection range (threshold to mid-1019 cm-3) because of the increased number of high energy electrons in the direct F valley contributing to the direct band-to-band radiative recombination. We have successfully grown single crystalline germanium epitaxially on silicon with a two-step approach. Tensile strain between 0.2% and 0.25% is formed in germanium upon cooling from high growth temperature (or post-growth annealing temperature) to room temperature because of the larger thermal expansion coefficient of germanium compared to that of silicon. Phosphorus are in situ doped in germanium as n-type dopants during the epitaxial growth. By carefully adjusting the growth condition, we have obtained active doping concentration as high as 2 x 1019 cm-3. An in situ doping model built by considering the transportation processes and the reactions of phosphorus-containing species well explains the temperature dependence of the doping concentration. The deviation from the model while analyzing the influence of other growth parameters indicates possible compensation of the dopants. We used photoluminescence (PL) measurement to study to the optical properties of tensile-strained n+ germanium. Room temperature PL was observed from the epitaxial Geon-Si films near the direct gap wavelength of 1600nm. The direct gap PL spectrum exhibits Ge direct band-to-band optical transition properties. The direct gap PL intensity increases with n-type doping concentration as a result of the indirect valley state filling effect which increases the Fermi level leading to higher excited electron density in the direct F valley. === (cont.) The direct gap PL intensity also increases with temperature because of the increased number of high energy direct F valley electrons thermally activated from the indirect L valleys. This effect make germanium light emission robust to inevitable heating effects during operation in practice. The "unusual" n-type doping and temperature dependences of PL are unique properties of the direct gap emission from indirect bandgap Ge. There effects are predicted by our theory, and the observation of these effect in experiments is a strong evidence of validity of the theory. In order to study the electrical injection in Ge, we fabricated Si/Ge/Si heterojunction light emitting diodes (LEDs). Room temperature direct gap electroluminescence (EL) are observed from these diodes. It is the first observation of EL from Ge. The direct gap EL spectrum matches the PL spectrum underlying the same injection mechanism in both electrical pumping and optical pumping. The direct gap EL intensity increases superlinearly with injection current because of the raised quasi Fermi level leading to the increased fraction of the injected electrons in the direct F valley. The internal quantum efficiency of the LEDs is on the order of 10-3 consistent with the finite-element simulation results. This EL efficiency can be improved to 10-1 if doping germanium active region with n-type. The design of n+Ge based heterojunction diodes has been simulated, and an optimal design has been proposed based on the simulation. We used pump-probe spectroscopy to measure material gain of the tensile-strained n+ germanium. === (cont.) We have observed an optical bleaching effect, the reduction of absorption under pumping and the prelude of optical gain, above the direct band gap energy from the engineered Ge. The population inversion factor increases with the n-type doping concentration in Ge, as predicted by the theory. By increasing the injection level using a Ge micro-mesa structure carrier confinement, we have successfully demonstrated the net gain, i.e. population inversion. A peak gain of 50 ± 25 cm-1 at 1605 nm has been obtained from the experiment. It is the first report of observing net gain from germanium. === by Xiaochen Sun. === Ph.D.
|
|---|
