在高科技產業中，晶圓表面粗糙度的檢測是非常重要的品質參數，而散射光量測技術是一個強而有力的技術可以用來量測晶圓表面上的奈米級特徵，其可以滿足高科技產業對於高生產率、高靈敏性之檢測儀器需求。晶圓表面的散射光會由許多散射源產生，例如微粒污染物、表面形貌粗糙度、加工後表面殘餘物和表面缺陷等，當光線打在這些表面特徵上都會產生散射光，因此想要有效的分析散射光特性藉以來精確的量測晶圓上奈米級表面粗糙度值是非常複雜的工作。因為雙向反射分佈函數可以詳細的描述一個雷射光源入射到物體表面上時，物體表面特徵散射光的發散情況，所以它是一個有效的工具可以用來精準的分析物體表面特徵之散射光。 本計畫將自行推導散射光數值模型和實驗量測來探討由晶圓上奈米級表面粗糙度所發出的全場散射光之三維偏振特性和其角度相依性。在實驗方面，本計畫將自行規劃、設計和架設一套高速大面積多角度調整散射光偵測儀來量測由不同奈米級表面粗糙度所發出的散射光，分析其散射光之偏振特性，將量測實驗結果與理論計算結果比較。在理論方面，本計畫將採用米氏表面交互影響模型，來推導和計算奈米級表面粗糙度全場散射光的雙向橢圓參數之角度相依性，也就是雙向橢圓參數隨著散射光極角和方位角變化的情形，這些偏振特性可以用來更有效的分析奈米級表面粗糙度的大小，此研究結果將可有效解決高科技產業中奈米級表面粗糙度的量測問題。 A major concern to the high technology industry for surface qualities is the detection and characterization of surface roughness on silicon wafers. Light-scattering has been shown to be a powerful diagnostic technique for characterizing optical surface qualities. Instruments based on light-scattering satisfy many of the requirements of the high technology industries, such as high throughput rate and high sensitivity. However, scattered light may arise from a number of sources, such as particulate contamination, surface topography, surface residue, and subsurface defects. Analyzing the scattered intensity to accurately determine the value of the nanoscale surface roughness is complicated, but possible. The bidirectional reflectance distribution function of a surface, describing the angular distribution of radiance scattered by a surface normalized by the irradiance incident on the surface, is a great tool to accurately analyze the scattering from surface features. In this proposal, we will explore the angular dependence and polarization of three-dimensional scattering from nanoscale surface roughness on wafers. The high speed goniometric optical scatter instrument will be developed to measure out-of-plane scattering from surface roughness. Experimental results from measurements of various surface roughnesses will be presented and compared to theoretical light-scattering calculations. Theoretical calculation based on the Mie-surface double-interaction model describes the angular dependence of bidirectional ellipsometric (BE) data qualitatively and predicts the general trend of the BE data as functions of the polar and azimuthal angles. The angular dependence and the polarization of light scattered by surface roughness can be used to determine the value of nanoscale surface roughness of wafer. Exploitation of this knowledge should allow a substantial increase in the detection sensitivity for surface topography on nano-rough surfaces in the high technology industry.
