聚對苯二甲酸二丁酯(Polybutylene terephathalate, PBT)工程塑料為某公司的主力產品之一，PBT品質的關鍵問題是產品中的A含量及B含量，而A的生成是PBT生產製程中最主要的副反應，尤其是當製程以對苯二甲酸(TPA)為原料時，製程中產生的大量A將使得產品品質下降而失去競爭力，造成對製程重要的品質挑戰。
本文主要分成兩個部分，第一部份為提出不同的方法量測產品中A含量及B含量。A量測方式藉由氣相層析儀(gas chromatography, GC)取得不同A濃度的面積值，利用面積值與濃度建立檢量線。B量測方式則是利用液相色譜法-質譜聯用(liquid chromatography–mass spectrometry, LCMS)，對某公司PBT樣品C、D、E與它牌競爭品進行B含量分析，所得出的數據可用來判斷產品的優劣。上述的方法能避免繁雜的實驗步驟，進而縮減時間，能在短時間取得多組數據。
第二部分為主成分分析(principal components analysis, PCA)，利用第一部份所提取的數據及某公司提供的製程數據做主成分分析。主成分分析是一種分析、簡化數據集的技術，經常用於降低數據集的維數，同時保持對變異數貢獻最大的特徵。因此，低階成分能夠保留住數據最重要的訊息，如果一個多元數據集能夠在高維數據空間座標系中被顯現出來，PCA的作用就是提供較低維度的圖像，這幅圖像即為在訊息最多的點上原對象的一個投影。這樣就可以利用少量的主成分使得數據的維度降低了，藉此判斷各操作條件對品質的影響性，而PCA能夠反映更多的微觀細節資訊，本研究計畫將透過上述的量測方式，經由長春所提供的樣品，提取數據，經由主成分分析，判斷哪些變數影響PBT品質，進而改變製程參數，提升PBT品質。

Polybutylene terephthalate (PBT) is a thermoplastic engineering polymer that is used as an insulator in the electrical and electronics industries. The quality of PBT mainly refers to its content of A. However, synthesis of PBT by terephthalic acid (TPA) process always accompanies the generation of A by-product, especially, the excessive A content will downgrade the commercial value of PBT.
The main purpose of this research project is to improve the manufactoring technologies of PBT process. In first part, we have been established a new measurement method for determining A and B content in PBT product. The A measurements were performed on a gas chromatography (GC). We can use the calibration curve to estimate the amount of that analysis in a sample of unknown A concentration. The B content of the reaction mixtures was determined by liquid chromatography–mass spectrometry (LCMS). We use three PBT products from company to measure quality.
In second part, we aim to use principal component analysis (PCA) for enhancing the quality of PBT. PCA is a technique used to emphasize variation and bring out strong patterns in a dataset. It's often used to make data easy to explore and visualize. PCA is a statistical procedure for identifying a smaller number of uncorrelated variables, called principal components. The goal of principal components analysis is to explain the maximum amount of variance with the fewest number of principal components. PCA is commonly used in the social sciences, market research, and other industries that use large data sets. Based on PCA, we attempt to develop analysis for reducing amount of A generation and amount of B generation during the PBT process.

摘要 I
Abstract II
謝誌 III
目錄 IV
圖目錄 VI
表目錄 VII
第一章 緒論 1
1.1 前言 1
1.2 文獻回顧 3
1.3 某公司現有量測技術回顧 5
1.4 研究動機與目的 6
1.5 文章架構 7
第二章 A檢測技術 8
2.1 A一次萃取檢測技術 8
2.1.1 實驗原理 8
2.1.2 實驗材料 8
2.1.3 實驗設備 9
2.1.4 實驗步驟 10
2.1.5 實驗結果 11
2.2 PBT泡水加熱50°C實驗 15
2.2.1 實驗原理 15
2.2.2 實驗設備 15
2.2.3 實驗步驟 15
2.2.4 實驗結果 16
2.3 A高溫減壓檢測技術 17
2.3.1 實驗原理 17
2.3.2 實驗材料 17
2.3.3 實驗步驟 17
2.3.4 實驗結果 17
第三章 B反溶劑檢測技術 19
3.1.1 實驗原理 19
3.1.2 實驗材料 19
3.1.3 實驗設備 19
3.1.4 實驗步驟 21
3.1.5 實驗結果 21
第四章 反應器製程數據分析 23
4.1.1 研究目的 23
4.1.2 研究方法 23
4.1.3 模型建立方法及在線監控方法 25
4.1.4 分析流程 26
4.1.5 實驗結果 26
第五章 結論 31
第六章 未來展望 32
第七章 參考文獻 33

[1] Devroede, J. (2007). PhD thesis, Eindhoven University of Technology.
[2] Pilati, F. & Manaresi, P. & Fortunato, B. & Munari, A. & Passalacqua, V. (1981). Formation of Poly(butylene terephthalate): Growing Reactions Studied by Model Molecules. Polymer, 22, 799.
[3] Pilati, F. & Manaresi, P. & Fortunato, B. & Munari, A. & Passalacqua, V. (1981). Formation of Poly(butylene terephthalate): Secondary Reactions Studied by Model Molecules. Polymer, 22, 1566.
[4] Pilati, F. & Manaresi, P. & Fortunato, B. & Munari, A. & Munari, P. (1983). Models for Formation of Poly(butylene terephthalate): Kinetics of The Titanium Terabutylate-Catalysed Reactions: 2, Polymer, 24, 1479.
[5] Pilati, F. & Manaresi, P. & Fortunato, B. & Munari, A. & Munari, P. (1985). Models for Formation of Poly(butylene terephthalate): Effect of Water on The Kinetics of The Titanium Terabutylate-Catalysed Reactions: 3, Polymer, 26, 1745.
[6] Wang, C.-S. & Lin, C.-S. (1999). Catalyst Composition For Production of Polybutylene Terephthalate Via Direct Esterification Process, United States Patent, No. 5900474.
[7] Lum, R. M. (1979). Thermal decomposition of poly(butylene terephthalate). Journal of Polymer Science: Polymer Chemistry Edition, 17, 203.
[8] Padias, A. B. & Hall, H. K. (1981). Oligomeric bishydroxybutyl terephthalates from terephthalic acid. Journal of Polymer Science: Polymer Chemistry Edition, 19, 1021.
[9] Nobutoshi C. & Yoshiyuki M. & Hatuo S. (1989). High-performance size-exclusion chromatography of poly- and oligorthylene terephthalate using a mixture of hexafluoroisopropanol and chloroform as the mobile phase. Journal of Chromatography, 479, 86
[10] Devroede, J. & Duchateau, R. & Koning, C. E. & Meuldijk, J. (2009). The synthesis of poly(butylene terephthalate) from terephthalic acid, part I: The influence of terephthalic acid on the tetrahydrofuran formation. Journal of Applied Polymer Science, 114, 2435.
[11] Devroede, J. & Duchateau, R. & Koning, C. E. (2009) The synthesis of poly(butylene terephthalate) from terephthalic acid, part II: Assessment of the first stage of the polymerization process. Journal of Applied Polymer Science, 114, 2427.
[12] Dayal, B. & MacGregor, J. (1997). Recursive exponentially weighted PLS and its applications to adaptive control and prediction. Journal of Process Control, 7, 169.
[13] Gonzalez, G. & Orchard, M. & Cerda, J. & Casali, A. & Vallebuona, G. (2003). Local models for soft-sensors in a rougher flotation bank. Minerals Engineering, 16, 441.
[14] Tham, M. & Montague, G. & Morris, A. & Lant, P. (1991). Soft-sensors for process estimation and inferential control. Journal of Process Control, 1, 3.
[15] Baffi, G. & Martin, E. Morris, A. (1999). Non-linear projection to latent structures revisited (the neural network PLS algorithm). Computers & Chemical Engineering, 23, 1293.
[16] Qin, S. & McAvoy, T. (1992). Nonlinear PLS modeling using neural networks. Computers & Chemical Engineering, 16, 379.
[17] Passalacqua, V. & Pilati, F. & Zamboni, V. & Fortunato, B. & Manaresi, P. (1976). Thermal Degradation of Poly(butylene terephthalate). Polymer, 17, 1044.
[18] Baffi, G. & Martin, E. & Morris (1999). A. Non-linear projection to latent structures revisited (the neural network PLS algorithm). Computers & Chemical Engineering, 23, 1293.
[19] Qin, S. & McAvoy, T. (1992). Nonlinear PLS modeling using neural networks. Computers & Chemical Engineering, 16, 379.
[20] Jolliffe, I. (2002). Principal component analysis. Hoboken, New Jersey: John Wiley & Sons, Ltd.
[21] Yao, Y. & Chen, T. & Gao, F. (2010). Multivariate statistical monitoring of two-dimensional dynamic batch processes utilizing non-Gaussian information. Journal of Process Control, 20, 1188-1197.
[22] Nomikos, P. & MacGregor, J.F. (1995). Multivariate SPC charts for monitoring batch processes. Technometrics, 37, 41-59.
[23] Steven, S. Z. & Donald, J. D. (Eds). (2012). Chemical Principles(7th ed.) Boston, Massachusetts: Cengage Learning Ltd.
[24] 東洋紡積股份有限公司 (民97)。聚酯彈性體組成物及其製法。日本專利：97114741。