Fast high-order perturbation of surfaces methods for simulation of multilayer plasmonic devices and metamaterials.
Fast high-order perturbation of surfaces methods for simulation of multilayer plasmonic devices and metamaterials.
复制标题
用于模拟多层等离子体器件和超材料的表面快速高阶扰动方法。
DOI：
10.1364/josaa.31.001820
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
Sang‐Hyun Oh
中科院分区：
文献类型：
--
作者：
D. Nicholls;F. Reitich;T. Johnson;Sang‐Hyun Oh
周期介质对时谐线性波的散射在材料科学、无损检测、遥感和海洋学等领域有着广泛的应用。在这项工作中，我们考虑到光学方面的应用，更具体地说，等离子体，以及表面等离子体激元，它们是非凡光学传输、表面增强拉曼散射和表面等离子体共振生物传感等非凡现象的核心。在本文中，我们开发了鲁棒，高精度和极快的数值求解器，用于在频率范围内近似解决光栅散射问题，其中这些问题通常使用。对于这些应用中常见的分段常数介电常数，表面公式显然具有优势，因为它们假设了仅在材料界面处支持的未知数。我们在这里开发的算法是高阶曲面摄动方法，并推广了以前的方法，以利用这些算法在部分或全部界面平凡（平坦）时可以显着加速的事实。更具体地说，对于具有一个非平凡接口（和一个平凡接口）的配置，我们描述了一个与两层求解器具有相同计算复杂度的算法。通过数值模拟和与实验数据的比较，我们证明了我们的新算法的速度、准确性和适用性。
The scattering of time-harmonic linear waves by periodic media arises in a wide array of applications from materials science and nondestructive testing to remote sensing and oceanography. In this work we have in mind applications in optics, more specifically plasmonics, and the surface plasmon polaritons that are at the heart of remarkable phenomena such as extraordinary optical transmission, surface-enhanced Raman scattering, and surface plasmon resonance biosensing. In this paper we develop robust, highly accurate, and extremely rapid numerical solvers for approximating solutions to grating scattering problems in the frequency regime where these are commonly used. For piecewise-constant dielectric constants, which are commonplace in these applications, surface formulations are clearly advantaged as they posit unknowns supported solely at the material interfaces. The algorithms we develop here are high-order perturbation of surfaces methods and generalize previous approaches to take advantage of the fact that these algorithms can be significantly accelerated when some or all of the interfaces are trivial (flat). More specifically, for configurations with one nontrivial interface (and one trivial interface) we describe an algorithm that has the same computational complexity as a two-layer solver. With numerical simulations and comparisons with experimental data, we demonstrate the speed, accuracy, and applicability of our new algorithms.