Investigation On Spectral Properties Of Rough Surface And The Inverse Problem

Spectral properties of rough surface are important for both academic research and practical applications. Its inverse problem has also attracted much attention due to its wide applications. This dissertation focuses on forward calculation of spectral properties of rough surface and its inverse problem. The detailed contents are listed below:(1) Theories for analysis of spectral properties of micro/nano-scaled rough surfaceSpectral properties of two-dimensional micro/nano-scaled rough surface are investigated based on the Maxwell equations. To quantitatively study the spatial distribution of energy flux reflected from rough surface, bidirectional reflectance distribution function is introduced. The finite difference domain time algorithm is proposed to solve the Maxwell equations. Thermal radiation energy flux is calculated based on the Poyting theory.(2) Numerical analysis on spectral properties of two-dimensional rough surfaceThe impacts of different factors on spectral properties of two-dimensional rough surface are analyzed. The factors involve surface structure, incident angle and polarization state. It is indicated that surfaces with lower reflectance possess two major features:larger roughness and scattering into a wide angle range. However the influence of polarization state on spectral properties of rough surface is limited.(3) Experimental measurements on spectral properties of two-dimensional rough surfaceBRDFs of surfaces with different structures were measured under different incident angles. The results, combined with the numerical analysis ones, suggest that the peak of BRDFs are remarkable for smoother surfaces, and larger roughness engenders side peaks of BRDFs.(4) Calculation model of the inverse problemBased on optimization, the inverse problem to reconstruct surface based on the spectral properties is analyzed. The particle swamp optimization algorithm is proposed to simulate the optimization process and is applied to the inverse calculation model.(5) Numerical analysis on the inverse problemThe Kirchhoff approximation and FDTD algorithm are respectively applied to describe the spectral properties when the size of object is larger than or small than the wavelength. The model maintains good performance in the presence of noise and a large number of parameters. Then the efficient method to choose appropriate input information of the inverse calculation model is discussed. The wavelet toolbox of MATLAB is adopted in denoising process. The inverse model is viable in reconstruction of one-dimension rough surface profile, one-dimension grating, two-dimension rough surface structure and material parameter. The numerical solution process to the inverse problem is applied to inverse design of antireflection film.This work is beneficial to optimizing the structure of optoelectronic devices.