Abstract
本論文旨在探討以氫稀釋法(diluted-hydrogen method) 及氫原子處理 法(hydrogen-atom-treatment method)成長之氫化矽(hydrogena- ted silicon)薄膜微細結構的演化情形。本論文以核磁共振(NMR) 為矽氫鍵結 組態特性量測之主要工具, 並配合紅外線吸收光譜、拉曼散射 (Raman Scattering)光譜、X-射線繞射光譜、光學能隙(optical band gap)及導 電度進行微細結構之分析。本實驗發現, 以氫稀釋法在氫氣流量比例 Xc(H2/(H2+SiH4) 大於 90% 時成長之氫化矽試樣的氫核磁共振光譜出現 一很尖銳譜線, 此結果顯示了結構之改變。當 Xc為95%以上時, 核磁共振 光譜的尖銳譜線呈現半高寬變窄且其它譜線消失之結果, 由拉曼散射光譜 及X-射線繞射光譜顯示試樣含有微晶相矽 (microcrystalline phase silicon )。此時紅外線吸收光譜亦顯示試樣中 SiH2 之鍵結量增加及氫 含量減少, 同時亦伴隨著暗導電度升高、光/暗導電度比降低及光學能隙 之降低。以氫原子處理法成長時依處理條件不同亦可得到結構之改變, 當 氫原子處理時間不足時並無微晶相矽生成。 而氫原子處理時間在40 秒時 可生成與氫稀釋法中相似之微晶相矽。以較高rf功率進行氫原子處理時間 只能提供少許之改善, 而較長之氫原子處理時間如 50 秒反而破壞微晶相 矽的生成, 此效應可歸因於氫電漿過度之"侵蝕"(etching)所致。本文中 討論了 NMR 之尖銳譜線的來源與微晶結搆之形成的關係。吾人認為 NMR 之尖銳譜線應來自氫化矽薄膜內的氫分子 (molecular hydrogen)。 吾人亦提出一解釋微晶相矽在氫稀釋及氫電漿中生成之成長模型, 此模型 基本上支持"侵蝕"之機構。同時此結果亦建議氫化矽結晶程度(degree of crystallinity) 是可以系統化調變的。 The microstructure evolution of hydrogenated silicon film deposited by hydrogen-dilution and hydrogen-atom-treatment methods were studied. Proton nuclear magnetic resonance(NMR) were employed as the major tool to characterize the silicon hydrogen bonding configuration. FT-IR, Raman scattering, X-ray diffraction, optical bandgap, and conductivity measurements were also used for the microstructure characterization. A sharp Lorentzian line-shape in the NMR spectra appeared in samples deposited by the diluted-hydrogen method with the fraction of hydrogen flow rate Xc(H2/(H2+SiH4)) increased beyond 90%. The sharp line-shape became dominant and its full width at half maximum became narrower as Xc was increased to higher than 95%. Under this condition, the Raman spectra and X-ray diffraction identified the formation of microcrystalline silicon (uc-Si) phase. When hydrogen-atom-treatment(HAT) was employed, microstructure of Si:H films could also be changed depending on the treatment conditions. With insufficient HAT time, the Si:H film showed no uc-Si phase. A HAT time of 40s at the fraction of hydrogen flow rate of 90%, was found to be sufficient in creating the uc-Si:H phase in our case. The uc-Si phase was found to be destroyed by longer HAT time of 50s in this experiment. The relation between the origin of the sharp line-shape in NMR spectra and the formation of microcrystalline phase was also discussed. It was believed the the sharp line-shape in the NMR spectrum should come from molecular hydrogen incorporated inside the Si:H films. A growth model supporting the "etching" mechanism was presented to illustrate the formation procedures of uc-Si under the influence of plasma hydrogen and hydrogen dilution. From these results, it was suggested that the degree of crystallinity of hydrogenated silicon films can be systematically adjusted.