Career: Photonic Crystal Waveguides and Device Integration

这是一个设计、建造和评估光子晶体纳米光波导元件和光学系统的建议。该计划的主要目标是开发实用的纳米光子技术，并将光子技术的进步知识融入课堂。在过去的几年里，人们对光子带隙结构进行了大量的研究。然而，所有这些工作都是作为对不同介电晶格或单个光子元件的光学性质的基础研究而完成的。我在南加州大学的研究包含几个项目，其中的目标是开发单独的纳米光子设备或设计其中一个设备的特定功能。该提案旨在建立在这些计划中正在进行的工作，并将结果整合到集成光通信技术中。当然，波导管本身就很重要，但它们也是联合收割机组成功能光路所必需的关键元件之一。在这样做的时候，必须注意尽量减少耦合和插入损耗。在这个项目中有几个关键问题需要研究。光子晶体波导损耗的来源是什么？我们如何在薄膜图案中设计有效的引导？一维和二维周期性引导结构之间的权衡是什么？我们如何从单个器件发展到使用光子晶体的集成光学技术？什么样的建模算法可以有效地处理这些结构，有没有系统级的建模算法？

This is a proposal to design, build, and evaluate photonic crystal nanophotonicwaveguide components and optical systems. The broad goals of this program are todevelop a practical nanophotonic technology, and to incorporate knowledge of theadvancements in photonics technology into the classroom. There has been a great deal ofresearch into photonic bandgap structures in the last few years. All of this work,however, has been done either as basic research on the optical properties of differentdielectric lattices or on single photonic components. My research at USC containsseveral programs in which the goals are to develop individual nanophotonic devices or toengineer a particular feature of one of these devices. This proposal is intended to buildon work ongoing in those programs and to integrate the results into an integrated opticalcommunication technology. Waveguides, of course, are important in their own right, butthey are also one of the key elements necessary to combine devices to form functionaloptical circuits. In doing this, care must be taken to minimize coupling and insertionlosses as well. There are several key questions to be investigated in this program. Whatis the origin of photonic crystal waveguide loss? How do we design efficient guiding inthin patterned films? What are the trade-offs between one-dimensional and two-dimensionalperiodic guiding structures? How do we go from single devices to an integrated optics technology using photonic crystals? What modeling algorithms can handle these structures efficiently, and are there system level modeling algorithms possible?