Popov E, Nevière M
Institute of Solid State Physics, Sofia, Bulgaria.
J Opt Soc Am A Opt Image Sci Vis. 2000 Oct;17(10):1773-84. doi: 10.1364/josaa.17.001773.
Using theorems of Fourier factorization, a recent paper [J. Opt. Soc. Am. A 13, 1870 (1996)] has shown that the truncated Fourier series of products of discontinuous functions that were used in the differential theory of gratings during the past 30 years are not converging everywhere in TM polarization. They turn out to be converging everywhere only at the limit of infinitely low modulated gratings. We derive new truncated equations and implement them numerically. The computed efficiencies turn out to converge about as fast as in the TE-polarization case with respect to the number of Fourier harmonics used to represent the field. The fast convergence is observed on both metallic and dielectric gratings with sinusoidal, triangular, and lamellar profiles as well as with cylindrical and rectangular rods, and examples are shown on gratings with 100% modulation. The new formulation opens a new wide range of applications of the method, concerning not only gratings used in TM polarization but also conical diffraction, crossed gratings, three-dimensional problems, nonperiodic objects, rough surfaces, photonic band gaps, nonlinear optics, etc. The formulation also concerns the TE polarization case for a grating ruled on a magnetic material as well as gratings ruled on anisotropic materials. The method developed is applicable to any theory that requires the Fourier analysis of continuous products of discontinuous periodic functions; we propose to call it the fast Fourier factorization method.
利用傅里叶分解定理,最近一篇论文[《美国光学学会志A》13, 1870 (1996)]表明,在过去30年光栅微分理论中使用的不连续函数乘积的截断傅里叶级数在TM偏振中并非处处收敛。结果表明,只有在无限低调制光栅的极限情况下,它们才处处收敛。我们推导了新的截断方程并进行了数值实现。计算得到的效率相对于用于表示场的傅里叶谐波数量,其收敛速度与TE偏振情况大致相同。在具有正弦、三角和层状轮廓的金属和介质光栅以及圆柱形和矩形棒上都观察到了快速收敛,并且给出了100%调制光栅的示例。新的公式为该方法开辟了广泛的新应用范围,不仅涉及TM偏振中使用的光栅,还涉及圆锥衍射、交叉光栅、三维问题、非周期物体、粗糙表面、光子带隙、非线性光学等。该公式还涉及在磁性材料上刻划的光栅以及在各向异性材料上刻划的光栅的TE偏振情况。所开发的方法适用于任何需要对不连续周期函数的连续乘积进行傅里叶分析的理论;我们建议将其称为快速傅里叶分解方法。