Modeling and Design Optimization of High Frequency Micro and Nano Structures Exploiting Adjoint Sensitivity Analysis

在过去几年中出现了一些高频技术，提供了有趣的功能。 例如，超材料引入了具有自然界中不存在的特性的材料。 它们用于构建在不同频率范围内工作的小型化天线，传感器和滤波器。 等离子体激元器件利用金属-电介质界面来引导低于衍射极限的光。 这允许实现具有小得多的尺寸的光子器件，从而有望实现光子处理单元的更多集成。 纳米天线（nantennas）被提出用于高频能量收集和视线通信。太赫兹（THz）的应用也出现了由于最近的太赫兹源和探测器的可用性。 所有这些技术都利用其性质可能取决于源值（非线性）、取决于频率（色散）或取决于激发方向（各向异性）的材料。 这些特性的不同组合可以存在于这样的高频结构中。 这些复杂的特性使得这些设备的精确建模耗费大量时间和内存。 这些结构的优化通常需要估计目标响应相对于所有参数的灵敏度。 估计这些敏感性的经典方法需要重复模拟这些时间密集型结构，这可能是禁止的。 在本提案中，我们将利用我们在伴随灵敏度分析方面的专业知识，开发通用的伴随灵敏度方法。这些方法将适用于一般的高频结构，其材料可能表现出任何组合的非线性，色散和各向异性。使用我们的方法，所需的响应或目标函数的灵敏度将估计使用最多一个额外的模拟，无论参数的数量。 伴随灵敏度将应用于超材料结构、等离子体结构、纳米天线、太赫兹滤波器和频率选择表面的设计优化。 我们将把我们开发的伴随技术应用到宽带隐身领域，在该领域，隐身衣可以拥有大量的可优化参数。 我们还将利用伴随灵敏度分析来提高不同样品的太赫兹成像和光谱学的准确性。 拟议的研究涉及加拿大和世界的重要研究领域，包括能量收集、太赫兹医学成像和电大物体的宽带隐身。它将产生新技术，加拿大公司将利用这些技术创造新产品或开发现有产品。 它还将培养大量高素质的人才（HQP），用于先进高频微纳米结构的建模，设计优化，制造和表征。

A number of high frequency technologies emerged over the past few years that offer interesting functionalities. Metamaterials, for example, introduced materials with properties that do not exist in nature. They are used in building miniaturized antennas, sensors, and filters that operate in different frequency ranges. Plasmonic devices utilize metal-dielectric interfaces to guide light below the diffraction limit. This allows the implementation of photonic devices with much smaller dimensions thus promising more integration of photonic processing units. Nano antennas (nantennas) were presented for high frequency energy harvesting and line-of-sight communication. Terahertz (THz) applications also emerged thanks to the recent availability of THz sources and detectors. All these technologies utilize materials whose properties may be dependent on source values (nonlinear), dependent on frequency (dispersive), or dependent on the direction of excitation (anisotropic). Different combinations of these properties may exist in such high frequency structures. These complex properties make the accurate modeling of these devices time- and memory-intensive. Optimization of these structures often requires estimating sensitivities of the target response with respect to all parameters. The classical approach for estimating these sensitivities requires repeated simulations of these time-intensive structures which may be prohibitive. In this proposal, we will build on our expertise in adjoint sensitivity analysis to develop universal adjoint sensitivity approaches. These approaches will apply to general high frequency structures whose materials may exhibit any combination of nonlinearity, dispersion, and anisotropy. Using our approaches, the sensitivities of the desired response or objective function will be estimated using at most one extra simulation regardless of the number of parameters. Adjoint sensitivities will be applied to the design optimization of metamaterial structures, plasmonic structures, nantennas, and THz filters and frequency selective surfaces. We will apply our developed adjoint techniques to the area of wideband cloaking where a cloak can possess a large number of optimizable parameters. We will also utilize adjoint sensitivity analysis to improve the accuracy of THz imaging and spectroscopy of different samples. The proposed research addresses important areas of research for Canada and the world including energy harvesting, THz medical imaging, and wideband cloaking of electrically large objects. It will result in novel technologies that will be utilized by Canadian companies to create new products or develop existing ones. It will also train a good number of high quality personnel (HQP) on modeling, design optimization, fabrication, and characterization of advanced high frequency micro and nano structures.