定向固化多晶矽生長熱流場模擬與驗證

Abstract

目前多晶矽太陽能電池仍然是目前太陽能產業的主流，本研究針對目前工業上生產大尺寸矽晶所使用的定向固化法(Directional Solidification Method)做2D的熱流模擬，使用的軟體為Comsol Multiphysics，這套軟體使用有限元素法，主要用來模擬多物理場的耦合，本研究引入國內廠商的鍋爐模型以及材料參數做為模擬的條件，驗證模擬和實驗的溫度點以及長晶速率，並研究不同的矸鍋熱傳導值，發現矸鍋熱傳導值越大，會讓整體的長晶速率變快，但同時側向長晶也會變嚴重，另外，為了達到節省長晶的能源成本以及好的長晶品質，設計了4種不同的鍋爐結構來達到目的，第一種將矸鍋正下方的絕熱層打開，第二種為矸鍋底部角落加裝絕熱塊，第三種為支撐絕熱塊做延伸，第四種則結合第二種和第三種，研究結果發現，將底部打開，在長晶前期雖然會增加溫度梯度，但是增加的幅度沒有比長晶速率大，而加裝側絕熱塊可以減慢側邊熱量的散失，並且減少側向長晶，對於中央的長晶速率不會有太大影響，而下絕熱塊延伸可以改變側邊的熱場，避免側邊的熱量散失，大幅增加軸向溫度梯度而中央的長晶速率會隨之增加，增加的幅度比溫度梯度小，而最後結合側絕熱塊以及下絕熱塊的延伸並沒有更好的結果，原因在於下絕熱塊改變了原本矸鍋側邊的熱傳方向，熱量由原本應該由矽湯傳出來的方向因為下絕熱塊延伸的效果，改為傳入矸鍋內矽湯，而側絕熱塊的功用也因為熱傳方向的改變，功用由原本的保護矸鍋內部的熱量不從側面散失，轉變為阻止熱量進入矸鍋內部，因此，軸向溫度梯度增加的幅度比較小。

Currently, polysilicon is the mainstream technology applicable for solar cell. In this thesis, efforts are made toward the 2D thermal dynamic simulation to produce large size polysilicon by directional solidification method. The simulation software is Comsol Multiphysics. This software adopts finite element method to simulate the heat, mass, and solidification phenomena. The simulation takes into account the transient phenomenon of solidification of silicon and the variation of material properties. The simulations are compared with the real crystal growth process of the test data from some private company with satisfactory agreements. Also in this study, effect of the natural convective heat transfer coefficient of the crucible ambient on the crystal growth is also examined. The results showed that when the crucible’s heat transfer coefficient is large, velocity of crystal growth becomes fast. In addition, lateral crystal growth also increases because lateral heat loss is comparatively large. On the other hand, this study also investigates four kinds of method to improve the configuration during solidification of melt/crystal interface. The first one is to open the insulation layer underneath crucible. It can shorten the period of crystal growth but the growth velocity may be too fast to achieve a better crystal quality. The second one is to set up heat insulation block outside the crucible corner. Through these designs, the flow velocity in the crucible on the corner is slowing down. It also decreases lateral crystal growth without affecting the overall rate of crystal growth. The third one is to extend heat insulation support on the bottom of crucible. The design impairs heat transfer decreases temperature difference on the crucible. Keeping the temperature of the side wall of crucible will increases axial temperature gradient. Therefore, this design can achieve the lowest V/G value and have negligible lateral crystal growth. The last design is to combine the second and the third design. However, it is not so effective to improve the lateral crystal growth because the third design can keep the temperature and transfer heat into the crucible. But the second design will block heat into crucible.