台灣LED廠商憑藉擁有完整的上、下游供應鏈和廣大的競爭力，大舉添購大量的金屬有機物化學氣相沉積(Metal-Organic Chemical Vapor Deposition，MOCVD）機台設備，大量生產LED產品，進而提高機台產能，增加整體的獲利。目前金屬有機物化學氣相沉積（MOCVD）薄膜的製程參數調整，是依據工程師經驗採逐一因素法加以調整。但由於影響金屬有機物化學氣相沉積薄膜之物理特性有諸多互相關聯性的製程參數，使得實驗過程變的相當複雜。本研究擬採取實驗設計和反應曲面法來決定LED磊晶之MOCVD製程最佳參數。使用DOE(Design of experiment)可同時改變多個自變數特點來規劃實驗並以變異數分析找出影響亮度顯著因子。應用RSM(Response surface method)來評估其主因子、水準範圍內所呈現的品質特性，求出製程可能的最佳參數。
由反應曲面法實驗分析得知，在P-type超晶格層，鎵流量為4.46 sccm、鎂流量為40 sccm、通入時間為19.7sec時，可得到預測最高亮度為143.7 mcd，最後並執行驗證實驗以確認最佳參數水準之組合。透過此實驗方法，希望藉降低實驗複雜度，讓磊晶工程師能藉此方法得到執行參數調整的參考方案，更可以改善產品之品質。

With the solid LED infrastructures and the worldwide competitive advantage, LED industry in Taiwan has invested a great amount of capital to obtain the facilities of Metal-Organic Chemical Vapor Deposition(MOCVD). Since the LED manufacturing industry has entered the stage of mass-production, and thus machinery capacity and overall profits are increasing accordingly. Currently, the parameter tuning of the MOCVD process is usually adjusted with one-factor at a time experiments by experienced engineers. However, due to the interaction of the physical characteristics of the LED epitaxial and the tunable process parameters, the experiment process is often very complicated. This thesis determines the optimal parameter setting of the MOCVD process and the physical characteristics of epitaxy by using design of both experiments (DOE) and the response surface method (RSM) where DOE is used to plan a group of experiments in which specified changes are made simultaneously to several process inputs and the RSM is used to investigate the quality characteristics within the range of the key factors found in the precious DOE process. Accordingly the optimal parameters which maximize the response can be determined.
The predicted optimal parameter settings for p-SLS layer derived from the RSM regression are 4.46 sccm, 40 sccm and 19.7sec for TMGa flow, Cp2Mg flow and growth time respectively. With this composition, the predicted maximum output luminescence are 143.7 mcd, and the results are justified by field implementation. According to the experimental results, engineers can obtain the information about the parameter setting and use these values as a reference point for tuning the process parameter.

郭峰鳴， “MOCVD 反應器之氮化鎵薄膜成長參數探討”，國立中山大學機械與機電工程研究所碩士論文，(2004)
黃韋誠， “以反應曲面法探討鱈魚魚漿製品最適化之研究”，屏東科技大學食品科學系碩士論文，(2003)
呂紹旭，“全球LED照明產業動態”，光連雙月刊NO.73,pp.38-39，(2008)
Alam, A., Schon, O., Schineller,B., Heuken, M. and Jurgensen, H., “MOVPE Growth optimization using computer supported design of experiments” phys. stat. sol.(a)180,pp.109-114,(2000).
Amano, H., Kito,M., Hiramatsu, K. and Akasaki, I., Jpn. J. Appl. Phys., 28, L2112, (1989).
Amano, H., Sawaki, N., Akasaki, I. and Toyoda, I., “Metalorganic vapor phaseepitaxial growth of a high quality GaN film using an AlN buffer layer”, Appl.Phys. Lett., vol.48, pp.353-355,(1986).
Bar-Ilan, A.H., Zamir, S., Katz, O., Meyler, B. and Salzman, J., “GaN layer growth optimization for high power devices ”, Materials Science and Engineering, A 302, pp. 14-17,(2001).
Bicknell-Tassius, R.N., Lee, K. and Brown, A.S.,“The growth of A1GaAs-InGaAs quantum-well structures by molecular beam epitaxy: observation of critical interdependent effects utilizing the design of experiments approach”, Journal of Crystal Growth 175/176 ,PP.1131-1137,(1997).
Box, G.E.P. and Wilson, K.B., “On the experimental attaimment of optimum condition” Journal of the Royal Statistic Society, B13, 1-38, PP.38-45,(1951).
Chen, J.P., Kim, S.L.and Ting, Y.P.,“Optimization of membrane physical and chemical cleaning by a statistically designed approach” , Journal of Membrane Science. 219 (1–2), PP.27–45,(2003).
Dey, A.and Suen, C.Y.,“Optimal Fractional Factorial Plans for Main Effects and Specified Two-Factor Interactions: A Projective Geometric Approach” , Annals of Statistics, Vol. 30, No.5, pp.1512-1523,(2002).
Gao, Y., Gulino, D.A. and Higgins,R.,“Effects of susceptor geometry on gan growth on si(Ⅲ) with a new MOCVD reactor” , MRS Internet J. Nitride Semicon. Res.4s1,G3.53,(1999).
Giovanni, M., “Respones surface methodology and product optimization”, Food Technology,12,41-45,(1983).
Grandusky, J. R., Jamil, M. and Shahedipour-Sandvik, F., “Optimization of the active region of InGaN/GaN 405 nm light emittingdiodes using statistical design of experiments for determination of interaction effects”, J. Vac. Sci. Technol. B, Vol 23, Issue 4, pp. 1576-1581,(2005).
Haitzl, R. , “ Another Semiconductor Revolution: This Time It's Lighting!”, Springer Berlin, Volume 43, pp.67-81,(2003).
Kumakura, K. and Kobayashi, N., “Increased Electrical Activity of Mg-Acceptors in AlGa1N/GaN Superlattices”, Jpn. J. Appl. Phys., Vol. 38 (1999) pp. 1012–1014,(1999).
Kozodoy, P., Hansen, M., DenBaars, S.P.and Mishra, U.K., “ Enhanced Mg doping efficiency in AlGaN/GaN superlattices”, Appl. Phys. Letters, Vol. 74, No.24, pp.3681-3683,(1999).
Kozodoy, P., Keller, S., DenBaars, S.P., Mishra, U.K., “MOVPE growth and characterization of Mg-doped GaN”, Journal of Crystal Growth 195 ,pp. 265-269,(1998).
Kim, K.and Jeon, K.N., “Silicon Delta-Doping Effect on Photoluminescence from InGaN/GaN Multi-Quantum-Well Structures”, Jpn. J. Appl. Phys., Vol. 46, No. 1, pp.131-133,(2007).
Limaye, P., Vandevelde, B., Vries, H.D., “Finite Element Analysis of Ultra Thin BGA Package: First and Second Level Reliability“ ,The 5th.Int.Conf. on Thermal and Mechanical Simulation and Experimentsin Micro-electronic and Micro-systems, pp.225-231 ,(2004).
Mahdi,K. and Gheshlaghi,R. , “Characterization and modeling of a crude oil desalting plant by a statistically designed approach” ,Journal of Petroleum Science and Engineering, 61 ,pp.116–123,(2008).
Montgomery, D.C., Design and Analysis of Experiments, 6th edition, John Wiley & Sons, New York, (2005).
Manasevit,H.M. “Single-crystal gallium arsenide on insulating substrates”, Appl. Phys. Lett, vol. 12, p. 156-159, (1968).
Nataf,G., Beaumont,B., Bouille,A., Haffouz,S., Vaille, M., Cryst,J.,Growth 192, (1998).
Nakamura,S., Isawa, N., Swno,M. and Mukai,T.,Jpn. J. Appl. Phys., 31, L1258, (1992).
Nakanishi, T., “Growth of high-purity GaAs epilayers by MOCVD and their applications to microwave MESFET's”, Journal Crystal Growth, vol. 50, p.255-262, (1981).
Pankove, J.I., Miller, E.A., Richman, D. and Berkeyheiser, J.E., RCA Review 32 ,PP.383,(1971).
Park, S.W. , Kim, C.B and Kim, S.Y. , Seo, Y.J., “Design of experimental optimization for ULSI CMP process Applications” , Microelectronic Engineering, Vol.66, pp.488–495,(2003).
Pawlowski, L. and Bigan, M., “Design of experiments in thermal spraying: A review”, Rhodia Recherches & Technologies, 69-192 Saint Fons, France,(2008).
Santen, H.V., Kleijn, C.R., Harry E.A., Akker,V.D., “Symmetry Beraking in a stagnation-flow CVD reactor”,J. C. Growth 212 ,pp.311-323,(2000).
Wu, L.W., Chang, S.J., Wen, T.C., Su, Y.K., “Influence of Si-Doping on the Characteristics of InGaN–GaN Multiple Quantum-Well Blue LightEmitting Diodes”, IEEE Journal of Quantum electronics, Vol. 38, No. 5,(2005).
Zhang, B., Egawa, T., Ishikawa, H., Liu, Y. and Jimbo, T., “Thermal stability of InGaN multiple-quantum-well light-emitting diodes on an AlN/sapphire template”, Jpn. J. Appl. Phys., Vol. 95, No. 6,pp.3170-3174,(2004).