- 學位論文
鋁誘發結晶製備結晶／磊晶薄膜應用於太陽能電池之研究
Crystalline/Epitaxy Thin Film Approach for Solar Cell Application by Aluminum-Induced Crystallization
摘要
近十年來，太陽光電能源的需求因全球能源匱乏危機而大幅提升，在太陽能產業發展中，降低發電成本並提升光電轉換效率始終是各家投注心力之研究主流。結晶矽基類的太陽能電池具有低材料成本、相對高的轉換效率、穩定且可長效使用等優點，使其產能及各界投入之研究能量始終難以被超越。鋁誘發結晶(aluminum-induced crystallization, AIC)為一種藉由將非晶矽與鋁金屬接觸來轉換為結晶矽之製程，此製程能製備出晶粒尺寸超過10微米之多晶矽(poly-Si)薄膜，亦可在晶種層(seeding layer)上藉由此製程在低於673K之溫度來達成低溫磊晶。AIC製程因具備了低溫、低材料及設備成本並且對環境無害等優點，使得此製程極有機會克服目前在太陽能元件以及其他光電固態元件領域中所面臨之諸多關鍵議題。 在本研究當中，為了能夠達到太陽能電池結構設計或工業化之需求，我們投入研究並突破了許多AIC製程之本質限制，甚至更進一步將再結晶的材料系統由一元材料的純矽改變為二元合金材料，使得此製程除了矽晶太陽能元件外，更能廣泛應用，推展至各類前瞻固態元件當中。除了應用面之研究結果外，我們也針對改良之製程深入分析反應機制，並建立理論模型。 針對改良太陽能或光電固態元件之關鍵議題，本論文之研究可分為三部分進行探討。首先在第一部分，我們將特定濃度的矽摻雜在初始的鋁膜當中，當非晶矽沉積在矽摻雜鋁膜上並經過退火後，將能使再結晶反應加速將近五十倍，藉此將能大幅提升鋁誘發結晶製程之產率，我們將此稱為Si-AIC製程。與一般AIC製程相較，Si-AIC製程因預摻雜矽原子於鋁膜中，而改變了晶粒成核及成長之行為，所製備出之poly-Si膜之晶粒尺寸及晶向皆與一般AIC有明顯差異。此部分我們針對矽晶粒之成核機制以及成長速率藉由穿透式電子顯微鏡(transmission electron microscope, TEM)以及能量散佈光譜儀(energy dispersion spectroscope, EDS)進行時間解析之分析與比較，建立反應機制之模型，並發現Si-AIC反應之活化能較一般AIC降低了約0.7eV。 在第二部分，我們針對在低溫製備重摻雜之背面場效層(back surface field, BSF)進行研究。一般AIC製程可將p型(p-type)摻雜之poly-Si在低於773K之低溫製備於玻璃基版之上，然而，因為鋁在矽中之溶解度不高，而使得在低溫下，poly-Si膜之電洞濃度僅能達到約31018 cm-3。在此階段的研究當中，我們建立了稱為B-AIC之技術來提升電洞濃度。B-AIC藉由預先摻雜特定濃度之硼原子於初始鋁膜當中，非晶矽沉積於硼摻雜鋁膜上，在經過673K之低溫退火製程後，將可使製備出之poly-Si薄膜具有超過1019 cm-3之高摻雜濃度，達成低溫高濃度摻雜，製備出p型重摻雜(p++)之poly-Si。除此之外，我們將基版由玻璃改為單晶矽基版，將可使B-AIC製備出之p++-Si層經由固態磊晶(solid phase epitaxy, SPE)機制，磊晶成長於矽晶圓上，此部分之研究將建立AIC-SPE之熱力學理論模型。 最後一部分，我們藉由AIC-SPE之磊晶技術，將非晶矽鍺(Si1-xGex)薄膜再結晶並磊晶於單晶矽基版上，發展出低溫異質磊晶技術。此階段之研究中，我們針對化學成分、原子結構以及AIC-SPE之分解-擴散-結晶機制進行詳盡的分析並建立理論模型。此SiGe-AIC-SPE技術具有能良好控制縱斷摻雜分布(doping profile)之優勢，使其能易於製備出階梯漸變結構來形成低缺陷密度之虛擬基版，除此之外，此技術具備製程簡單、低材料成本以及低溫製程等優點，使此技術在虛擬基版領域極具發展潛力。而矽鍺虛擬基版則可推展應用至高效率多接面太陽能電池以及固態照明元件等前瞻元件領域中，對產業之發展極具影響。 此論文的最後，我們也將幾項針對AIC製程之基礎研究工作收錄於附錄中，以幫助我們更詳細了解AIC反應。
並列摘要
The requirement of photovoltaic (PV) energy has vastly grown in last decade. To achieve high conversion efficiency with low production cost is the mainstream in the solar cell technology. Crystalline Si (c-Si) based solar cell is very popular since it combines the low-cost material system with the stable and high conversion efficiency for long term usage. Aluminum-induced crystallization (AIC) process is an approach to crystallize amorphous Si (a-Si) by means of contacting Al layer with a-Si. This approach is capable of achieving large grain sizes of over 10 μm or realizing the epitaxial growth on a seeding layer at a very low temperature of 673K. For improving several critical issues in the present solar PV technology and solid state devices, AIC process is very promising due to the features of low reaction temperature, low cost in precursors and facilities, and environmental friendly. In this study, the several limitations of standard AIC process were broke to adapt to the solar cell or industrial requirement. Furthermore, the material system in AIC process was modified to extend the application to other advanced solid state devices. Besides the application results, the theoretical models of our approaches had also been established to reveal the mechanisms of the modified processes. There’re three parts in this thesis to investigate several critical issues of AIC process for advanced solid state device application. In the first part, AIC process can be accelerated by a factor of about 50 by the doping of Si atoms into the initial Al layer for improving the throughput. This process is known as Si-doped AIC (Si-AIC). The grain size and crystallographic orientation of the grown polycrystalline Si (poly-Si) thin film produced are modified due to the fact that the presence of excess Si in the initial Al layer alters the nucleation and growth behavior of the Si grains as compared with the traditional AIC process. In this part the nucleation mechanism and growth rate of Si grains for Si-AIC are analyzed and quantitatively compared with those for AIC using time-series TEM/EDS images. It is found that the activation energy for grain growth was significantly reduced in the Si-AIC process, by 0.7eV compared with the AIC process. In the second part we investigate the heavily doping of BSF fabrication at very low temperature. P-type poly-Si film on foreign substrate can be fabricated at temperature lower than 773K by traditional AIC process. However, the ultimate carrier concentration of Si film is limited to approximately 3×1018 cm-3 because of the low solid solubility of Al in Si at temperatures below 773K. In this study, a process called B-AIC is developed in which boron is co-doped with Al to increase the carrier concentration in Si film to ~1019cm-3 at temperature as low as 673K. The carrier concentration can be tuned by the initial thickness of a-Si layer in B-AIC process. Beside the fabrication of poly-Si film on glass, the epitaxial growth of this heavily doped p++-Si film can also be realized on a mono-c-Si wafer via solid phase epitaxy (SPE) mechanism. The AIC/SPE thermodynamic model is also developed in this part. In the final part, the AIC process featured with SPE mechanism are extended to Si-Ge-Al system for heterogeneous epitaxy. A 300 nm Si1-xGex with tunable Ge content thin film can be epitaxially grew onto (100) mono-c-Si wafer at very low temperature of 673K. The chemical composition, atomic structure and the SPE reaction featured with the dissociation-diffusion-crystallization model of AIC process are carefully revealed and established. The development of this SiGe-AIC-SPE process is very promising to fabricate a low defect density virtual substrate by terrace grading structure approach due to its advantage of well controllable doping profile. Furthermore, the features of simple fabrication process, low material cost and low reaction temperature makes this approach predominant in the virtual substrate market. Additionally, our several works for revealing the fundamental of AIC process are appended in the appendix.