Thermal, mechanical and optical modelling of VCSEL packaging

Hybrid OECB (Opto-Electrical Circuit Boards) are expected to make a significant impact in the telecomm switches arena within the next five years, creating optical backplanes with high speed point-to-point optical interconnects. OECB's incorporate short range optical interconnects, and are based on VCSEL (Vertical Cavity Surface Emitting Diode) and PD (Photo Diode) pairs, connected to each other via embedded waveguides in the OECB. The VCSEL device is flip-chip assembled onto an organic substrate with embedded optical waveguides. The performance of the VCSEL device is governed by the thermal, mechanical and optical characteristics of this assembly. During operation, the VCSEL device will heat up and the thermal change together with the CTE mismatch in the materials, will result in potential misalignment between the VCSEL apertures and the waveguide openings in the substrate. Any degree of misalignment will affect the optical performance of the package. This paper will present results from a highly coupled modelling analysis involving thermal, mechanical and optical models. The paper will also present results from an optimisation analysis based on Design of Experiments (DOE).

Parallel bandwidth characteristics calculations for thin avalanche photodiodes on a SGI Origin 2000 supercomputer

An important factor for high-speed optical communication is the availability of ultrafast and low-noise photodetectors. Among the semiconductor photodetectors that are commonly used in today’s long-haul and metro-area fiber-optic systems, avalanche photodiodes (APDs) are often preferred over p-i-n photodiodes due to their internal gain, which significantly improves the receiver sensitivity and alleviates the need for optical pre-amplification. Unfortunately, the random nature of the very process of carrier impact ionization, which generates the gain, is inherently noisy and results in fluctuations not only in the gain but also in the time response. Recently, a theory characterizing the autocorrelation function of APDs has been developed by us which incorporates the dead-space effect, an effect that is very significant in thin, high-performance APDs. The research extends the time-domain analysis of the dead-space multiplication model to compute the autocorrelation function of the APD impulse response. However, the computation requires a large amount of memory space and is very time consuming. In this research, we describe our experiences in parallelizing the code in MPI and OpenMP using CAPTools. Several array partitioning schemes and scheduling policies are implemented and tested. Our results show that the code is scalable up to 64 processors on a SGI Origin 2000 machine and has small average errors.