隨著晶錠長晶的尺寸愈來愈大，晶錠在經過切割後，晶圓片表面會產生不平整的高低差也愈來愈明顯。晶圓表面的不平整不只會導致後續的半導體製程所需成本增加，也會浪費更多的材料，在長晶速度緩慢與晶片需求日益增加的情況下，要如何減少材料的浪費與後續製程的成本增加是一個重要挑戰。在線切割中影響晶圓片翹曲度的原因我們判斷有兩種：第一，切割時隨著不同深度晶錠受熱及散熱產生差異，導致晶錠熱膨脹程度不一樣，造成切割後晶圓表面產生高低差。第二，切割時鋸線的震動，使得晶圓表面不平整。我們初步判斷切割時隨著不同深度晶錠受熱及散熱的差異，是造成切割後表面形貌變異的主因，鋸線振動會造成局部形貌的變異，此部分的研究將針對第一種原因深入研究。 本研究建立了晶錠長300mm和長180mm的模型來研究切割熱對翹曲度的影響，將數值模擬得到的溫度與實驗數據作比較，並且得到了晶錠與漿料對熱量的吸收比例，最後再探討如何以更改在各個深度的拉線速度來將不同深度所得到的溫度變為一致，使得晶圓片表面的翹曲度可以改善。 ;As the size of the ingot grows larger, the uneven thicknese difference on the surface of the wafer becomes more and more obvious after wire-sawing. The unevenness of the wafer surface will not only increase the cost of the subsequent semiconductor manufacturing process, but also waste more materials. In the case of slow crystal growth and increasing demand for wafers, how to reduce the waste of materials and the cost increase of subsequent processes is an important challenge. We think that there are two reasons for affecting the flatness of wafers in wire sawing: First, the ingot is heated differently at different depths during wire sawing results in different thermal expansion of the ingot, causing the difference height of the wafer surface after cutting. Second, the vibration of the wire during cutting makes the surface of the wafer uneven. We preliminarily think that the difference in heating of ingots with different depths during cutting is the main reason for the uneven surface. The vibration of the wire will cause local surface unevenness. This research will focus on the first reason. In this study, a model with a thicknese of 300 mm and a thicknese of 180 mm are established to study the influence of cutting heat, and the temperature obtained by the numerical simulation is compared with the experimental data, and the heat absorption ratio of the crystal ingot and the slurry will be obtained. Then, discussing how to change the wire drawing speed at each depth to make the temperatures obtained at different depths the same, so that the flatness of the wafer surface can be improved.
