真空紫外光（Vacuum Ultraviolet， VUV，約 120 nm）反射鏡為半導體檢測與紫外太空望遠鏡之關鍵元件。此波段幾乎所有材料皆強烈吸收，可用之反射層材料僅餘鋁，但鋁一經接觸氧氣即氧化而使反射率大幅下降，故必須於高真空中鍍製並覆以氟化鎂（MgF₂）保護層。膜面之粗糙度則會以散射造成額外損耗，且波長越短影響越劇烈。因此本論文主要探討 Al/MgF₂ 反射鏡自基板前處理至保護層鍍製之各道製程條件應如何選擇。 本論文使用電子槍熱蒸鍍系統，於石英玻璃基板（另以 B270 光學玻璃為比對基準）鍍製 Al 反射層與 MgF₂ 保護層。鍍膜前先以高溫長時間預烘烤腔體去除殘留水氣，待降溫至室溫、背景壓力達高真空後，以 hBN-TiB₂ 坩鍋高鍍率沉積鋁膜。分析則以原子力顯微鏡（AFM）於兩種掃描範圍量測均方根粗糙度 Rq，並以真空紫外光譜儀量測反射率，作為判斷製程優劣之依據。 接著逐道討論製程參數：基板前處理比較兩種基板於不同離子源轟擊電壓下之形貌，選定最佳電壓；鋁反射層比較不同基板溫度所得之粗糙度，並以 Thornton 結構區域模型討論成長行為後選定沉積溫度；MgF₂ 保護層則針對鋁膜不耐熱與 MgF₂ 需高溫成膜之矛盾，採「室溫抗氧化層＋高溫保護層」二階段溫控製程，比較不同保護層溫度下之形貌後選定鍍製溫度。 最後以選定之條件鍍製完整 Al/MgF₂ 膜堆，量測其粗糙度並委外量測真空紫外反射率，比較不同基板材質與鋁膜厚度之結果，並與模擬光譜及文獻數據比對，藉粗糙度與反射率之對應關係，探討以表面粗糙度作為真空紫外光反射鏡製程判據之可行性。;Vacuum-ultraviolet (VUV， ~120 nm) mirrors are key components of semiconductor inspection systems and of ultraviolet space telescopes. Almost every material absorbs strongly in this band， leaving aluminum as the only usable reflective material， yet aluminum oxidizes on contact with oxygen and its reflectance drops sharply， so the film must be deposited under high vacuum and capped with a magnesium fluoride (MgF₂) protective layer. Surface roughness causes a further loss through scattering， and the shorter the wavelength the stronger the effect. This thesis therefore investigates how the process conditions of an Al/MgF₂ mirror， from substrate pretreatment to the deposition of the protective layer， should be chosen. Al reflective layers and MgF₂ protective layers were deposited on fused-quartz substrates， with B270 optical glass as a reference， in an electron-beam thermal evaporation system. The chamber was pre-baked at high temperature for an extended period to remove residual water vapor and then allowed to cool to room temperature; once the background pressure had reached high vacuum， the aluminum layer was deposited at a high rate from an hBN-TiB₂ crucible. For analysis， the root-mean-square roughness Rq was measured by atomic force microscopy (AFM) over two scan ranges and the reflectance with a VUV vacuum spectrometer， both serving as criteria for judging each process step. The process parameters were then examined in turn. For substrate pretreatment， the morphology of the two substrates under different ion-source bombardment voltages was compared and the optimal voltage selected. For the aluminum reflective layer， the roughness obtained at different substrate temperatures was compared， the growth behavior was discussed in terms of the Thornton structure-zone model， and the deposition temperature selected. For the MgF₂ protective layer， a two-step process consisting of a room-temperature anti-oxidation layer plus a high-temperature protective layer was adopted to reconcile the poor thermal tolerance of aluminum with the high temperature required for MgF₂ growth; the morphology obtained at different protective-layer temperatures was compared and the optimal deposition temperature selected. Finally， complete Al/MgF₂ stacks were deposited under the selected conditions; their surface roughness was measured and their VUV reflectance measured by an external laboratory. Results for different substrate materials and aluminum thicknesses were compared with each other and with a simulated spectrum and published data. Through the correspondence between roughness and reflectance， the feasibility of using surface roughness as the process criterion for VUV mirrors was assessed.
