發光二極體（Light Emitting Diode；LED），為21世紀的新型光源，具高效率、壽命長以及不易破損等優點。在要求高品質與高效率的時代中，晶片封裝製程技術為影響LED 的品質與效能之關鍵。
覆晶接合（Flip-Chip Bonding）具備輕、薄、短、小封裝優勢，為封裝製程技術發展之焦點。熱超音波覆晶接合（Thermo-Ultrasonic Flip Chip Bonding）製程，直接將晶片上的金墊（Gold Pad）與基板上的金凸塊（Gold Bump）接合在一起，金/金介面具有較佳的接合強度、電性及可靠度，且製程簡單、無鉛污染問題，並可在較低溫度及壓力下進行接合，具備高精度及細間距（Fine Pitch）的發展潛力。
本論文在研究LED 熱超音波覆晶接合製程，研究其多重品質特性之製程參數最佳化。因各品質特性間的相關性大及品質特性數目多，要決定最佳因子水準組合有其困難性，因為在取捨各因子水準時，可能會造成各品質特性之間相互衝突的問題。
本研究使用田口直交表（Taguchi Orthogonal Array Table）規劃實驗，實驗完成後，以主成份分析法（Principal Component Analysis）去除品質特性間的關聯性，再以灰關聯分析（Grey Relational Theory）找出最佳因子水準組合。最後結合倒傳遞類神經網路（Back-Propagation Neural Network）與Levenberg-Marquardt 演算法建構LED 覆晶接合製程之預測系統，將控制因子設為網路之輸入參數，而品質特性設為輸出參數，經過網路學習訓練，本系統預測誤差率在5% 以內，證明本預測系統有極佳的預測能力。

The light emitting diode (LED), a new light source in the twenty-first century, contains many advantages, such as high efficiency, long lifespan, and bold structure. The technology for chip packing is the focal point to increase LED quality and performance.
Flip-chip bonding, a kind of chip packing technology, allows produced chip to be lighter and smaller. Thermo-ultrasonic flip chip bonding is a way that directly joins gold pad and gold bump. The gold-gold connecting interface has better strength and electric reliability along with simple production process with no lead pollution. This kind of packaging also can be done in low pressure and low temperature with high accuracy and fine pitch.
This study focuses on LED thermo-ultrasonic flip chip bonding process to find the multi-quality properties with the optimization of LED processing parameters. The process will be more complex as qualities have strong correlation to each other. With many different quality characteristics, we will have to determine the combination of processing factors to yield best performance.
We first used the Taguchi orthogonal array table and then the principal component analysis to remove the relativity of each quality characteristics. We then used grey relational theory to obtain the best LED processing parameter and then used the back-propagation neural network and Levenberg-Marquardt algorithm to establish the LED Flip-Chip Bonding processing-quality predicting system. The error of the result is within 5%, indicating a high forecast capability.

【1】鐘文仁,陳佑任 編著,“IC封裝製程與CAE應用”, 全華科技圖書股份有限公司, (2005).
【2】McLaren T., Sa Yoon Kang, Wenge Zhang, Hellman D., The-Hua Ju and Lee Y. C.,”Thermosonic Flip-Chip Bonding for 8x8 VCSEL Array,” Electronic Components and Technology Conference, Vol. 45, pp.393-400, Las Vegas, NV, USA, (1995).
【3】Tan Q., Zhang W., Schaible B., Bond L. J., th Ju, Lee Y. C.,”Thermosonic Flip Chip Bonding Using Longitudinal Ultrasonic Vibration,” IEEE Transactions on Components, Packaging, and Manufacturing Technology, Part B, Advanced Packaging, Vol. 21, No. 1, pp.53-58, (1998).
【4】Mickelson A. R., Basavahally N. R., Lee Y. C.,”Optoelectronic Packaging,”John Wiley & Sons Inc., chap.12, New York, (1997).
【5】Lee Charles, Yeo Alfred, Tan Ai Min,“Flip Chip Interconnection Systems and its Reliability Performance,”Electronics Systemintegration Technology Conference, Vol. 2, pp.1032-1039, Germany, (2006).
【6】Tan Ai Min, Sharon Pei-Siang Lim, Yeo A., Lee C.,”Influence of Bump Geometry, Adhesives and Pad Finishings on the Joint Resistance of Au Bump and A/NCA Flip Chip Interconnection,”Proc. 5th Electronics Packaging Technology Conf, Vol. 2., pp.797-801, (2005).
【7】廖慶安,“銅晶片以熱音波覆晶銲著於硬性與軟性基板之研究”, 國立中正大學機械研究所碩士論文, (2007).
【8】Luk C. F., Chan Y. C. and Hung K. C.,”Development of Gold to Gold Interconnection Flip Chip Bonding for Chip on Suspension Assemblies,”Microelectronics Reliability, Vol. 42, No. 3, pp.381-389, (2002).
【9】徐嘉彬,劉俊賢,“熱超音波覆晶鍵合製程設備技術”, 工業技術研究院之機械工業雜誌258期, (2004).
【10】黃國興,廖仕傑,張鈞豪,“次世代光源製造及設備技術介紹”, 工業技術研究院之機械工業雜誌270期, (2005).
【11】田口玄一,“田口實驗計劃法”, 滄海出版社, (1997).
【12】Byrne D. M., and Taguchi, S.,”The Taguchi Approach to Parameter Design,”Quality Progress, pp.19-26, (1987).
【13】Chen R. S., Lee H. H., and Yu C. Y.,”Application of Taguchi Method on the Optimal Process Design of an Injection Model PC/PBT Automobile Bumper,”Composite Structures, Vol. 39, No. 3-4, pp.209-214, (1997).
【14】Anderson S. W., and Sedatole K.,”Designing Quality into Products: The Use of Accounting Data in New Product Development,”Accounting Horizons, Vol. 12, No. 3, pp.213-233, (1998).
【15】Chien C. H., Chen T. P., Lin W. B., Hsieh C. C., Wu Y. D., Yeh C. H.,”Experimental and Statistical Study in Adhesion Features of Bonded Interfaces of IC Packages,”Microelectronics Reliability, (2007).
【16】Pearson K.,”On Lines and Planes of Closest Fit to Systems of Points in Spaces,”Philosophical Magazine, Vol. 2, pp.559-572, (1901).
【17】Hotelling H.,”Analysis of a Complex of Statistical Variables into Principal Components,”Journal of Educational Psychology, Vol. 24, pp.498-520, (1933).
【18】Tong L. I. and Su C. T.,”Multi-Response Robust Design by Principal Component analysis ,”Total Quality Management, Vol. 8, No 6, pp.409-416, (1997).
【19】Kaiser H. F.,”The Application of Electronic Computers to Factor Analysis,”Educational and Psychological Measurement , Vol. 20, pp.141-151, (1960).
【20】鄧聚龍,“灰色系統基本方法”, 華中理工大學出版社, (1987).
【21】Deng J. L.,”Grey Informational Space,” The Journal of Grey System, Vol. 2, No. 1, pp.103-117, (1989).
【22】Deng J. L.,”Control problems of Grey System,”Systems and Control Letters, Vol. 1, pp.288-294, (1982).
【23】Chen Chie-bein and Lin Chin-Tsai and Chang Che-Wei and Ho Cheng-pin,”Grey Relation for Solving Multi-Quality Characteristics Problems of Taguchi Methods,”Journal of Technology, Vol. 15, No. 1, pp.25-33, (2000).
【24】Chester M.,”Neural Network : A Tutorial,”Prentice-Hall, (1993).
【25】葉怡成,“類神經網路模式應用與實作”, 儒林圖書, (2003).
【26】John F. C. Khaw, Lim B. S., Lennie E. N. Lim ,”Optimal Design of Neural Networks Using the Taguchi Method,”Neurocomputing, Vol. 7, No. 3, pp.225-245, (1995).
【27】Jean M. D., Liu C. D. and Wang J. T.,”Design and Development of Artificial Neural Networks for Depositing Powders in Coating Treatment,”Applied Surface Science, Vol. 245, pp.290-303, (2005).
【28】Fun M. H., Hagan M. T.,”Levenberg-Marquardt Training for Modular Networks,”IEEE International Conference on Neural Networks, Vol. 1, pp.468-473, Washington, DC, USA, (1996).
【29】Lera G., Pinzolas M.,”A Quasi-Local Levenberg-Marquardt Algorithm for Neural Network Training,”The 1998 IEEE International Joint Conference on Neural Network, Vol. 3, pp.2242-2246, Anchorage, AK, USA, (1998).
【30】Zhao J. and Wang F.,”Parameter Identification by Neural Network for Intelligent Deep Drawing of Axisymmetric Workpieces,”Journal of Materials Processing Technology, pp.387-391, (2005).
【31】Singh V., Gupta I. and Gupta H. O.,”ANN-Based Estimator for Distillation Using Levenberg-Marquardt Approach,”Engineering Applications of Artificial Intelligence, pp.249-259, (2007).
【32】Kermani B. G., Schiffman S. S., Nagle H. T.,”Performance of the Levenberg-Marquardt Neural Network Training Method in Electronic Nose Applications,”Sensors and Actuators B: Chemical, Vol. 110, pp.13-22, (2005).
【33】蘇朝墩,“品質工程”, 中華民國品質學會, (2005).
【34】林清山編,“多變項分析統計法”, 東華書局, (1980).
【35】李金泉,“SAS/PC 應用手冊-多變量應用統計與研究分析實務”, 松崗電腦圖書資料股份有限公司, (1996).
【36】曾國雄,鄧振源,“多變量分析(一)”, 松崗電腦圖書資料有限公司, (1986).
【37】鄧聚龍,“灰色系統理論與應用”, 高立圖書有限公司, (2003).
【38】王進德,蕭大全,“類神經網路與模糊控制理論入門”, 全華圖書, (2005).
【39】羅華強,“類神經網路-MATLAB 的應用”, 清蔚科技, (2001).
【40】EIA/JEDEC Standard,”Wire Bond Shear Test Method,”Electronic Industries Alliance, EIA/JESD22-B116, American, (1998).
【41】Sedra A. S., Smith K. C., ”Microelectronic Circuits,”Oxford, chap.3, New York, (1998).
【42】燦圓光電, http://www.forepi.com.tw/