Modeling and Optimization of Ultrafast and Low-Noise Thin Avalanche Photodiodes for Optical Communications

University Of Dayton, Dayton OH

Investigators

Abstract

This research program is a theoretical and experimental interdisciplinary effort that will lead to the design of a new generation of ultrafast and high-accuracy photodetectors. The program will focus on the widely-used class of photodetectors known as thin avalanche photodiodes (APDs). This program is motivated by the need to meet the increasing demand for bandwidth in next-generation optical core networks, where there is a need to develop a new generation of low-noise and high responsivity photodetectors with gain-bandwidth products that are far beyond the current state-of-the-art. Moreover, since the capability of a lightwave network is ultimately limited by its architecture and the performance of its components, a thorough understanding of the fundamental performance limits of applicable photodetectors significantly impacts the design of future communication networks. In addition, with the emerging int~rest in using wide-bandgap-material technology in various high- accuracy and ultrafast sensing applications, there is a need for the development of high-performance photodetectors using wide-bandgap materials such as CaN. The first goal of this project is to develop and validate a rigorous renewal-theory-based model for the joint statistics of the gain and the response time of APDs. The model will be applicable to APDS with various structures and materials with special emphasis on thin APDs, which exhibit low multiplication noise and high bandwidth. The theory will specifically capture the important effect of dead space, which plays a principal role in the performance of thin APDs and significantly affects the performance of both ultrafast (intersymbol-interference limited) and lot-power (gain-fluctuation limited) applications. Significant improvements in the noise and bandwidth characteristics is to be achieved by reducing the thickness of the APD's multiplication layer, which is responsible for the device gain. To date, the fundamental limits of the statistics of the gain-bandwidth product for thin APDs, and more notably, the effect of reducing the thickness of the multiplication layer on the fluctuations in the response time remain unknown. These questions will be thoroughly addressed in this research and the fundamental limits of APD performance will be established. The second goal of this project is to utilize the developed model to design and develop next- generation high-performance APD's. Device fabrication and characterization will be carried out at the existing facilities at the Microelectronics Research Center at the University of Texas in Austin. The role played by the thickness of the avalanche multiplication layer of the device will be thoroughly investigated in an effort to design devices with application-specific optimal charac-teristics. Low-noise devices with gain-bandwidth products well beyond 500 CHz (bandwidths of 50-100 GHz) are to be developed in this program. Optimization criteria will include a) maximizing the data transmission rate subject to a fixed bit-error rate, which is applicable to intersymbol-interference-limited communication systems, and B) maximizing the receiver signal-to-noise ratio in power-limited sensing applications. As a tool in accomplishing the above objectives, a CAD tool will be designed consisting of custom-made parallel-computing algorithms intended for the high-performance implementation of the model and the optimization process. The synergy between the investigators in this project, who have a demonstrated record in opto-electronic device modeling and fabrication, can lead to the development of devices with superb per-formance characteristics. The devices developed in this program will be useful for next-generation lightwave systems operating at 40 0Hz (per channel) and beyond.