posted on 2017-01-15, 23:13authored bySingh, Aashish
In recent times the use of silicon for fabricating active photonic devices has received widespread attention. Many photonic devices such as lasers, amplifiers, modulators and wavelength converters which were once conceived to be impossible to fabricate in silicon have been demonstrated. These developments are promising from the point of view of photonic related applications as well as for Opto-Electronic Integrated Circuits (OEIC). The convergence of electronic and photonic components can result in a cost effective and compact silicon chip. The application areas of such an OEIC can range from optical communications, through optical interconnects in Integrated Circuits (IC's), digital signal processing and sensing to biomedicine.
The experimental demonstrations of silicon based photonic devices are promising. However, the performance of these devices needs further improvement to challenge their commercially-available counterparts made from III-V group elements. To enable performance optimization, development of suitable device physics and mathematical models are indispensable. The physics of the silicon photonic devices is electro-optic in nature. However, the mathematical models reported in the literature optimize the device performance from the optical point of view and account for the associated electronics in a very limited sense. The inclusion of the associated electronics is usually done through an extremely simplified term G=N/'tau', i.e., generation rate (G) of the free-carriers is equal to the recombination rate (N/'tau'), with 'tau' being the effective carrier lifetime. The transport of the free carriers is mostly neglected from the modeling equations. Such simplification may be valid for short optical pulses of picoseconds or less. However, in devices where the optical beam is a continuous wave, such an approximation is invalid as shown in this thesis. Further, the fabrication of the P-i-N diode structure in the silicon waveguide is a common method to control the free-carrier density in the waveguiding medium. The applied bias controls the free carrier density via the transport of the free carriers into and out of the waveguide. Hence it is intuitively clear that the modeling of these devices should exhaustively account for both the electronic and the optical physics.
To accomplish such a coupled electro-optic model, three tasks were systematically carried out. In Chapter 3, the physical phenomena coupling the electronic and optical device physics was thoroughly studied. The coupling phenomenon is the plasma dispersion effect (PLDE)
omenclature{PLDE}{Plasma dispersion effect}. The PLDE relates the linear dielectric constant of silicon to the density of the free charge carriers. In Chapter 4, an extensive study of P-i-N photodiode device physics was carried out. A P-i-N diode rib waveguide used in silicon photonic applications closely resembles a P-i-N photodiode. Thus, a thorough understanding of the carrier transport modeling and the P-i-N photodiode device physics helped in developing physical insight into the electro-optic process. This assisted in developing a novel sliced waveguide model in Chapter 5. The model enabled solution of the initial-boundary value differential equations, describing the electro-optic device physics of silicon photonic devices. The model was validated by reproducing the results of published work in the literature. The model was used to study the influence of the inclusion/neglect of the carrier transport on the CW Raman amplification. The discrepancy observed between the modeling results and the published experimental data were examined. The potential causes of the discrepancy were studied in Chapter 6. In particular space charge electric field strength, diffusion in the rib region and the distribution of the electric field in the depletion region were examined. The analysis resulted in some practical guidelines which may be helpful in the performance enhancement of the silicon photonic devices.