品質檢測為台灣光電產業及半導體產業兩大高科技產業的重要探討議題之一。因其投資成本甚鉅，故如何藉由提升製程品質檢測效能進而提升生產力便成為其重要的議題之一。一個良好的品質檢測方法必須具備其良好的診斷及分類預測能力，尤其是在異常品僅佔成品極少數的高科技產業中，一套穩健不受限於任何資料型態的品質檢測方法必然是不可或缺的。馬氏-田口系統為一針對多變量資料所提出之診斷與預測技術，其相異於其它針對分析資料學習的分類法，而是以透過量測尺度的建立來建構分類模型，故其較不受資料分佈型態影響。本研究藉由馬氏-田口系統對不平衡資料的穩健分類能力及特徵挑選能力，針對光電產業薄膜電晶體液晶顯示器濺鍍製程膜厚一致性品質檢測問題及半導體產業IC封裝凸塊製程品質檢測效率改善，進行品質檢測系統分類模式之建構及檢測位置之挑選。
在濺鍍製程膜厚一致性品質檢測問題中，本研究有別於一般僅以膜厚平均值或CV值來作為該量測濺鍍製程品質之品質特性，進而採用膜厚平均值、標準差及全距三項品質特性來量測該製程之品質，並試測多種不同比例的不平衡資料，其結果發現藉由馬氏距離方法可有效的解決該製程之膜厚一致性品質及此案例中所呈現之資料不平衡分類問題。
而在半導體產業IC封裝凸塊製程品質檢測效率改善中，本研究藉由縮減檢測位置來減少該製程之檢測時間同時降低與檢測數量成正比之檢測成本。在此製程中進行了兩階段的品質檢測，首先進行單一品質檢測之品質檢測系統建構及檢測位置之挑選，在該實驗中，我們採用SPC技術中的管制圖來有效且有系統的選取建構馬氏空間之觀測值用以取代傳統的專家意見及重複的選取實驗，並利用以二分法為基之門檻值搜尋法決定其門檻值作為分類診斷的依據。其實驗結果顯示利用該方法所選取的馬氏空間可有效的將檢測位置由原先的10個縮減為6個，其檢測時間減少了40％，同時維持同樣高的準確率及相對敏感度。
在進行完單一品質檢測實驗後，本研究進一步將凸塊製程之單一品質檢測延伸至多維品質檢測。在此實驗中，為了使3-D雷射及2-D攝影機之凸塊檢測位置之縮減位置一致，我們運用馬氏距離之概念將多個檢測品質特性轉換成單一檢測指標後，再利用馬氏-田口系統之MTGS方法進行檢測位置縮減，藉以挑選出適合的檢測位置。該實驗結果顯示在不影響其分類準確率的情況下，其檢測位置由原先的10個縮減為7個，經由檢測位置的縮減，該檢測系統之3-D雷射的檢測時間及2-D攝影機的檢測時間皆減少了30％。
另一方面，觀察藉由馬氏-田口系統所建構的各個縮減模型中可逐步的發現各檢測位置的重要性，檢測人員可以此作為檢測位置挑選之參考，依序設定適當的檢測位置，藉此在能獲得相同凸塊製程之可接受的品質條件下，減少該製程之檢測位置。而在本實驗中除採用的SPC之觀點來建構馬氏空間之外，亦可用其來監控製程，當製程出現異常時，馬氏空間將被重新建構以維持該製程之品質。

Quality inspection is one of important issues in the optical electrical and semiconductor industries which are two of high-tech industries in Taiwan. These two industries need huge investments. Therefore, how to increase the inspection efficiency of processing quality to raise the production capacity has become an important issue. A good quality inspection method must have a good diagnosis and forecasting capability, especially in the high-tech industries, which reveals very limited numbers of nonconforming items. A robust quality inspection method is necessary and the method must be not influenced by data distribution. MTS is a new diagnosis and forecasting technique for multivariate data. MTS establishes a classification model by constructing a continuous measurement scale rather than learning from training data set. Therefore, MTS is not influenced by data distribution. In many applications, Mahalanobis-Taguchi System (MTS) verified that it has a more robust classification and enhanced features to allow improved selection. In this study, MTS is used to solve the sputtering process thin-film thickness uniformity quality problem in the optical electrical industry and to improve the bumping process inspection efficiency in the semiconductor industry by using its classification and ability to selection important features.
To solve the sputtering process thin-film thickness uniformity quality problem, the mean, the standard deviation and the range of the thickness are adopted as the quality characteristics, as opposed to the general approach that only considers the thickness mean in the measurements of thin-film thickness uniformity. The present study proposes Mahalanobis distance (MD) as a method to detect the nonconforming items in the sputtering process. The empirical results demonstrate that the MD method is more accurate and efficient in solving a multivariate, class imbalance problem.
Moreover, two stages quality inspection is proceeded in order to improve the bumping process inspection efficiency and reduce cost by selecting suitable inspection positions. In the first stage, the quality inspection system of the single quality is constructed and the suitable inspection positions are selected; in the second stage, the quality inspection system of the multidimensional quality is constructed and the suitable inspection positions are selected. In the experiment with first stage, the concept of control chart is used to generate a suitable MS and the bisection algorithm is used to determine the threshold value. The empirical result demonstrates that the numbers of bump height inspection features are significantly reduced from 10 to 6 without losing classification accuracy; and inspection time can be reduced by 40% in the single quality inspection system.
In the experiment with second stage, the concept of MD is used to convert multidimensional inspection quality characteristics into a single-dimensional inspection index and subsequently MTS is used to reduce costly inspection time and maintain the quality of the bumping process. The empirical result demonstrates that the numbers of bump height inspection features are significantly reduced from 10 to 7 without losing classification accuracy; and inspection time can be reduced by 30% in the multidimensional quality inspection system.
Furthermore, by virtue of reducing the bump height inspection position, the time of bump inspection was reduced. The inspection staff can select the suitable inspection position in sequence according to the significance of features selected by the MTS method. Meanwhile, they can reduce the number of inspection positions when checking acceptable quality of bump height. On the other hand, SPC has been used in this study, to select a suitable normal example for the full MS measurement; while it can also monitor the process. If the process detects defects that are distinct from preceding shift, the MS will be re-calculated in order to maintain the quality of the process.

陳順宇，民93，多變量分析，華泰書局。

蘇朝墩，民91，品質工程，中華民國品質學會。

Barandela, R., Sánchez, J. S., García, V., and Rangel, E., 2003, Strategies for learning in class imbalance problems, Pattern Recognition, 36, 849-851.

Brewer, R. F., 1996, Design of Experiments for Process Improvement and Quality Assurance, Engineering & Management Press, Norcross, Ga.

Chao, P.-S., Huang, J.-T., Hsu, H.-J., and Shih, S.-H., 2007, Fabrication on uniform plating-based flip chip solder bumps after reflow process using a polishing mechanism, Microelectronic Engineering, 84, 60-71.

Chawla, N. V., Bowyer, K. W., Hall, L. O., and Kegelmeyer, W. P., 2002, SMOTE: Synthetic minority over-sampling technique, Journal of Artificial Intelligence Research, 16, 321-357.

Chen, K.-M. and Chiang, K.-N., 2003, Impact of probing procedure on flip chip reliability, Microelectronics Reliability, 43, 123–130.

Das, P. and Datta, S., 2007, Exploring the effects of chemical composition in hot rolled steel product using Mahalanobis distance scale under Mahalanobis–Taguchi System, Computational Materials Science, 38, 671-677.

Engin, M., Fedakar, M., Engin, E. Z., and Korürek, M., 2007, Feature measurements of ECG beats based on statistical classifiers, Measurement, 40, 904-912.

Fan, Z., Chu, P. K., Cheung, N. W., and Chan, C., 1999, Thickness uniformity of silicon-on-insulator fabricated by plasma immersion ion implantation and ion cut, IEEE Transactions on Plasma Science, 27, 633-636.

Fátima Vales Silva, M. and Nicholls, J. R., 2001, A model for calculating the thickness profile of TiB2 and Al multilayer coatings produced by planar magnetron sputtering, Surface and Coatings Technology, 142-144, 934-938.

Gerald, C. F. and Wheatley, P. O., 2004, Applied Numerical Analysis, 7th edition, Addison-Wesley, Boston, MA, USA.

Grouven, U., Bergel, F., Schiltz, B., and Schultz, A., 1995, A PC program for unbiased and predictive linear and quadratic discriminant analysis, Computers in Biology and Medicine, 25, 425-430.

Guerra, E. and Villalobos, J. R., 2001, A three-dimensional automated visual inspection system for SMT assembly, Computers & Industrial Engineering, 40, 175-190.

Han, J. and Kamber, M., 2006, Data Mining : Concepts and Techniques, Morgan Kaufmann, San Francisco, CA, USA.

Hayashi, S., Tanaka, Y., and Kodama, E., 2002, A new manufacturing control system using Mahalanobis distance for maximizing productivity, IEEE Transactions on Semiconductor Manufacturing, 15, 442-446.

Ho, W. K., Lee, L. L., Tay, A., and Schaper, C., 2002, Resist film uniformity in the microlithography process, IEEE Transactions on Semiconductor Manufacturing, 15, 323-330.

Hsieh, K. L. and Lu, Y. S., 2008, Model construction and parameter effect for TFT-LCD process based on yield analysis by using ANNs and stepwise regression, Expert Systems with Applications, 34, 717–724.

Itagaki, M., Takamiya, E., Watanabe, K., Nukaga, T., and Umemura, F., 2007, Diagnosis of quality of fresh water for carbon steel corrosion by Mahalanobis distance, Corrosion Science, 49, 3408-3420.

Jaing, C. C., Cheng, M. H., Chen, J. S., Tsai, C. H., Yeh, P. S., Kao, J. S., and Hsiao, H. Y., 2001, Studying layer uniformity of sputter coatings by intensity distribution of plasma spectrum, Applied Surface Science, 169-170, 649-653.

Japkowicz, N. and Stephen, S., 2002, The class imbalance problem: a systematic study, Intelligent Data Analysis, 6, 429-449.

Jeang, A. and Li, H.-C., 2009, An empirical study on the robust design of a semiconductor-bumping process, Journal of Materials Processing Technology, 209, 2986-2993.

Kubat, M. and Matwin, S., 1997, Addressing the curse of imbalanced training sets: one-sided selection, Proceedings of the International Conference on Machine Learning, Nashville, TN, USA, 24, 179-186.

Kuzanyan, A. S., 2003, PLD of large area films onto substrate undergoing translational motion by mask method, IEEE Transactions on Applied Superconductivity, 13, 2868-2870.

Lau, J. H., 1996, Flip Chip Technologies, McGraw-Hill, New York.

Liau, L. C.-K. and Chen, B. S.-C., 2005, Process optimization of gold stud bump manufacturing using artificial neural networks, Expert Systems with Applications, 29, 264-271.

Lee, J. H., Shim, C. M., Park, S. K., Jung, D. G., and Park, S., 1997, A possible method for large area deposition of a-Si:H thin films using electron cyclotron resonance plasma-enhanced chemical vapor deposition, Thin Solid Films, 295, 67-72.

Lee, L. L., Schaper, C. D., and Ho, W. K., 2002, Real-time predictive control of photoresist film thickness uniformity, IEEE Transactions on Semiconductor Manufacturing, 15, 51-59.

Lin, J. T., Wang, F. K., and Peng, C. C., 2008, Lot release times and dispatching rule for a TFT-LCD cell process, Robotics and Computer-Integrated Manufacturing, 24, 228-238.

Lozano, M., Cabruja, E., Collado, A., Santander, J., and Ullán, M., 2001, Bump bonding of pixel systems, Nuclear Instruments and Methods in Physics Research A, 473, 95–101.

Montgomery, D. C., 2001, Introduction to Statistical Quality Control, 4th edition, John Wiley, NY, USA.

Pan, J.-N., Pan, J., and Lee, C.-Y., 2009, Finding and optimising the key factors for the multiple-response manufacturing process, International Journal of Production Research, 47(9), 2327-2344.

Pendharkar, P. C., Rodger, J. A., Yaverbaum, G. J., Herman, N., and Benner, M., 1999, Association, statistical, mathematical and neural approaches for mining breast cancer patterns, Expert System with Applications, 17, 223-232.

Riho, T., Suzuki, A., Oro, J., Ohmi, K., and Tanaka, H., 2005, The yield enhancement methodology for invisible defects using the MTS+ method, IEEE Transactions on Semiconductor Manufacturing, 18, 561-568.

Serel, A. D. and Moskowitz, H., 2008, Stochastics and statistics joint economic design of EWMA control charts for mean and variance, European Journal of Operational Research, 184, 157-168.

Singer, P., 1994, Flat panel displays: an interesting test case for the US, Semiconductor International, 17, 78-88.

Srinivasaraghavan, J. and Allada, V., 2006, Application of Mahalanobis distance as a lean assessment metric, International Journal of Advanced Manufacturing Technology, 29, 1159-1168.

Su, C. T. and Hsiao, Y. H., 2007, An evaluation of the robustness of MTS for imbalanced data, IEEE Transactions on Knowledge and Data Engineering, 19, 1321-1332.

Su, C. T., Wang, P. S., and Hsu, C. C., 2005, Effective approaches for low temperature polysilicon TFT-LCD post-mapping yield control problem, IEEE Transactions on Automation Science and Engineering, 2, 198-206.

Su, Y. C., Hung, M. H., Cheng, F. T., and Chen, Y. T., 2006, A processing quality prognostics scheme for plasma sputtering in TFT-LCD manufacturing, IEEE Transactions on Semiconductor Manufacturing, 19(2), 183-194.

Taguchi, G., Chowdhury, S., and Wu, Y., 2001, The Mahalanobis-Taguchi System, McGraw-Hill, New York City, NY, USA.

Taguchi, G. and Jugulum, R., 2002, The Mahalanobis-Taguchi Strategy: A Pattern Technology System, John Wiley, New York City, NY, USA.

Tsukahara, H., Nishiyama, Y., Takahashi, F., Fuse, T., Ando, M., and Nishino, T., High-speed 3D inspection system for solder bumps, Proceedings of SPIE, 2597, 168-177, Oct. 23-25, Bellingham, WA, USA, 1995.

Tsukahara, H., Nishiyama, Y., Takahashi, F., Fuse, T., Nishino, T., and Ando, M., 2000, High-speed solder bump inspection system using a laser scanner and CCD camera, Systems and Computers in Japan, 31, 94-102.

Tzeng, Y. C., 2008, Distance weighted multiple classifiers systems applied to remote sensing images classification/data fusion, Journal of the Chinese Institute of Engineers, 31, 639-647.

Wu, X., Chung, W. K., Cheng, J., Tong, H., and Xu, Y., 2009, A parallel-structure solder paste inspection system, IEEE/ASME Transactions on Mechatronics, 14, 590-597.

Yang, T. and Cheng, Y.-T., 2010, The use of Mahalanobis-Taguchi System to improve flip-chip bumping height inspection efficiency, Microelectronics Reliability, 50, 407-414.

Yang, T. Cheng, Y.-T., and Shen, Y.-A., The use of Mahalanobis distance in solving the sputtering process thin-film thickness uniformity quality problem, Journal of the Chinese Institute of Engineers, to appear.

Yang, T., Tsai, T.-N., and Yeh, J., 2005, A neural network-based prediction model for fine pitch stencil-printing quality in surface mount assembly, Engineering Applications of Artificial Intelligence, 18, 335-341.

Zhong, Z. W. and Lye, T. M., 2004, Effects of profiles and conditioning methods of heater surfaces on deposited TiN film thickness and uniformity, Microelectronic Engineering, 75, 405-412.