極紫外光微影（extreme ultraviolet lithography， EUVL）因線寬持續縮小，對反射式光罩中吸收層之反射率、相位控制與結構穩定性提出更高要求。吸收層若無法有效降低反射光，或與下方多層膜反射鏡產生非理想干涉，將可能影響成像品質與臨界尺寸控制。因此，吸收層材料選擇、膜厚設計與界面品質，皆為 EUV 反射式光罩結構設計中的重要議題。 本研究以離子束濺鍍法製備 B₄C/Mo/B₄C/Si 週期性多層膜反射鏡，並作為後續吸收層設計之基底。透過 EUV 反射率量測、TEM、EDS mapping、XRR 與 IMD 擬合分析，評估實際鍍製結構與理論模型之差異。結果顯示，多層膜反射鏡之實測反射率低於理論設計值，顯示膜厚偏差、界面粗糙度、Mo/Si 界面混合與 B₄C 超薄層連續性，皆可能影響反射鏡之光學表現。因此，後續吸收層設計需同時考量理論光學響應與實際沉積後之結構特性。 針對原始 Al/Ni/Mo/Ni 吸收層中可能出現之 Ni/Al 界面混合與 Al 膜層粗糙化問題，本研究導入 B₄C 作為金屬層間阻隔層，並比較不同 B₄C 厚度對界面分層之影響。TEM 與 EDS 結果顯示，較薄 B₄C 阻隔層之界面改善效果有限，而 1.5 nm B₄C 可形成較清楚之分層結構，具有較佳阻隔效果。進一步以 Ta 取代 Al 後，實際鍍製之 B₄C/Ta/B₄C/Ni/B₄C/Mo/B₄C/Ni 多層吸收層，可視為以 B₄C 作為層間阻隔之 Ta/Ni/Mo/Ni 吸收層結構；其粗糙化現象獲得改善，並可透過光學模擬維持低反射率、相位調控與等效折射率設計需求。綜合 XRR、TEM、EDS mapping 與 EUV 反射率量測結果，Ta/Ni/Mo/Ni 吸收層結構可作為後續吸收層厚度修正與模型優化之可行設計方向。最終實際鍍製之 B₄C/Ta/B₄C/Ni/B₄C/Mo/B₄C/Ni 吸收層經 EUV 反射率量測後，低角度區域最高反射率約為 9%，相較於多層膜反射鏡基底於 6° 入射角之實測反射率 45.71%，已呈現明顯反射率抑制效果。;As the linewidth in extreme ultraviolet lithography (EUVL) continues to shrink， the reflectance， phase control， and structural stability of the absorber layer in reflective masks have become increasingly important. If reflected light is not sufficiently suppressed， non-ideal interference between the absorber stack and the multilayer mirror may affect aerial image quality and critical dimension control. Therefore， absorber material selection， thickness design， and interface quality are key issues in EUV reflective mask structures. In this study， a periodic B₄C/Mo/B₄C/Si multilayer mirror was fabricated by ion beam sputtering and used as the substrate for absorber-stack design. EUV reflectance measurements， TEM， EDS mapping， XRR， and IMD fitting were used to compare the deposited structure with the theoretical model. The measured reflectance was lower than the design value， indicating that thickness deviation， interfacial roughness， Mo/Si intermixing， and the continuity of ultrathin B₄C layers may affect the mirror performance. These results show that absorber design should consider both ideal optical response and actual deposited structure. To reduce Ni/Al interfacial mixing and Al-induced roughening in the original Al/Ni/Mo/Ni absorber stack， B₄C was introduced as an interlayer barrier. TEM and EDS results showed that thinner B₄C barriers provided limited improvement， whereas a 1.5 nm B₄C layer produced clearer layer separation. After replacing Al with Ta， the deposited B₄C/Ta/B₄C/Ni/B₄C/Mo/B₄C/Ni multilayer absorber can be regarded as a Ta/Ni/Mo/Ni absorber structure with B₄C interlayer barriers. This structure reduced roughening while maintaining the intended low reflectance， phase control， and effective refractive index design targets， providing a feasible basis for further model optimization. The fabricated B₄C/Ta/B₄C/Ni/B₄C/Mo/B₄C/Ni absorber stack exhibited a maximum EUV reflectance of approximately 9% in the low-angle region， which was significantly lower than the measured reflectance of the multilayer mirror substrate at 6° incidence， 45.71%.
