太陽能多晶矽晶錠良率之產能改善

Abstract

目前多晶矽太陽能電池仍是市場上佔有率最高的太陽能電池。而為使太陽能電池發電為達到市電平價（grid parity）的目標，目前各太陽能大廠皆朝向提升晶片品質、降低生產成本及增加產出這三方面而努力。 此研究的目的在於防止多晶矽晶錠在固化生長時與石英坩堝產生化學反應而出現黏堝，而此一現象易在拆除石英坩堝時使晶錠缺角甚至破裂，就算晶錠完整無缺，但因石英坩堝在固化時會附著於晶錠上，造成該區晶格結構不完整，此現象非常容易在後續晶錠的開方破錠製程中造成Brick出現裂痕，而影響產出，這也將會是影響產能最重大的因素。 本研究計畫將以增加產能為出發點，進行三項實驗: 1. 在石英坩堝內壁噴塗厚度不同的氮化矽塗層來找出最適合長晶條件的塗層厚度。實驗結果顯示，氮化矽塗層4層，厚度約在0.08mm左右，晶錠外觀最為完整，產品良率約在96%。 2. 依據實驗1的結果，在相同的條件下，先在石英坩堝內部刷上一層矽溶膠，並再將矽溶膠加入氮化矽中噴塗於石英坩堝內壁上，厚度同樣約為0.08mm左右，來驗證矽溶膠對晶錠良率是否有助益。實驗結果顯示，實驗2的晶錠外觀更勝實驗1的晶錠平滑，且晶錠側邊皆無沾黏現象，產品良率可達98%，證明添加矽溶膠的塗層是有益於晶錠良率提升的。 3. 實驗不同的融熔速率對晶錠良率的影響。實驗結果顯示，在Melting功率 2100kw~2500kw的區間中，並無法明確的表現出其差異性，所以溫度控制對於良率的改善並無明顯的影響。

The silicon (Si) solar cell still has highest market share nowadays. To achieve the goal of grid parity, at present the solar wafer manufacturers are facing challenges to enhance quality, to reduce production costs and to increase output and work. The purpose of this study is to prevent reactive between the polysilicon ingot and the silica crucible. This phenomenon will cause chipping or breakage of ingot during the removal of the quartz crucible. Even if the ingot is intact, attached the quartz will result in a damaged lattice structure area, which is likely to cause cracks in brick in the subsequent ingot manufacturing process. This will affect the amount of output and loss on capacity. The work to increase output was carried out in three experiments: 1.On the quartz crucible wall, spray coating with different thickness of silicon nitride, and identify to the most appropriate coating thickness. Experimental results showed that when silicon nitride coating thickness was about 0.08mm, the ingot appearance was most complete and the product yield is highest. 2.Based on the results of Experiment 1, using the same conditions, first brush applied layer of silicon colloidal on the quartz crucible, then added silicon colloidal in the silicon nitride and coating on the inner wall of the quartz crucible with the same thickness of about 0.08mm to verify the ingot yield. The results showed that the appearance of Experiment 2 ingot was far better then Experiment 1. The ingot was the guest coherence problem was not detached , product yield went up to 98%. So we showed that by adding silicon colloidal, ingot yield was improved. 3.We also study the different melting rate with the ingot yield. Experimental results show that with power setting during Melting between 2100kw ~ 2500kw, no different was observed, so the temperature control to improve yields had no significant impact.