在化學工業、特用化學品、食品工業、製藥業等各相關產業中，目標產品之純化與分離是 製造過程中十分關鍵的部分，化學工程師在設計及改良的化工程序必須使用大量的熱力學性 質與相平衡數據，如：利用液-液相平衡數據篩選萃取製程所需之有機溶劑。傳統上都是透過 實驗量測獲取所需的熱力學相平衡數據，配合經驗法則與試誤法（trial-and-error）來篩選與決 定製程所需的有機溶機。然而近年來隨著電腦計算能力與理論計算的快速發展，少數預測型 熱力學模型（如：我們過去開發之COSMO-SAC 活性係數模型與PR+COSMOSAC 狀態方程 式）在相平衡的預測上已能達到定性甚至是定量的精確度，雖然熱力學模型預測結果終究無 法完全取代實驗量測結果，但是將可以協助化學工程師快速的初步篩選製程溶劑，並大幅減 少實驗量測所耗費的時間與金錢。 本三年研究計畫的主要目標為開發一個結合預測型熱力模型與實驗量測快速篩選超臨界 二氧化碳流體製程之有機共溶劑的方法。由於含有超臨界流體系統的實驗數據量測難度高， 因此相關數據相較於常壓下的系統非常稀少，也是導致超臨界二氧化碳製程較不普及的主因 之一。本研究計畫將分為三個方向：(1)持續開發預測型熱力學模型並將其應用於預測超臨界 流體系統之相平衡；(2)架設相關實驗設備與實驗數據量測並用以驗證預測結果；(3)選擇一個 有經濟價值潛力的固體溶質來進行專案研究，利用預測型熱力學模型進行萃取製程有機共溶 劑之初篩，決定二至三個最佳有機共溶劑後，再以實驗量測所需相平衡數據來驗證預測結果 並提供程序設計所需資訊。本研究所開發之快速篩選有機共溶劑的方法，預期可應用於各種 製程之溶劑篩選，除了大幅減少目前試誤法所花費的時間與金錢，亦能夠更有效率地找到符 合製程需求之最佳有機溶劑。 ;The separation of mixtures and the purification of final products are key steps in industries, such as chemical-, food-, specialty chemicals-, and pharmaceutical industry. The information of fluid phase equilibrium and thermodynamic properties are crucial for chemical engineers to design and optimize the separation and purification processes, such as choosing a suitable organic solvent for the extraction process from liquid-liquid equilibrium data. The required phase equilibrium data are typically obtained from experimental measurements, and the suitable extraction solvent is determined by the tedious trial-and-error method. However, taking advantage of the fast development of computing power and theoretical computing methods, a few thermodynamic models, such as the COSMO-SAC model and PR+COSMOSAC EOS developed by our research group, can provide convincing fluid phase equilibrium prediction (from qualitative to quantitative results). Even though experimental data are the most reliable information for process designs, the predictive thermodynamic model with such accuracy still has great potential to be used to fast screen the suitable solvent candidates. And thus it will significantly reduce costs and time for doing all trial-and-error measurements. In this 3-year project, we propose to develop an approach to combine the predictive thermodynamic model and experimental measurements to fast identify the suitable organic co-solvent candidates for supercritical fluid processes containing carbon dioxide. One of the major reasons that the supercritical fluid processes have not been widely used in industry is the lack of experimental data since it is costly and difficult for doing experiments under high pressure conditions. The proposed work will focus on the following three major topics: (1) Improve the accuracy of predictive thermodynamic models and apply them in predicting phase behavior of supercritical fluid systems containing carbon dioxide. (2) Set up the apparatus to investigate phase behavior of supercritical fluid systems containing carbon dioxide and use them to validate prediction results. (3) Conduct a case study of a solid solute having potential commercial value. Two or three co-solvent candidates are first determined by using the predictive thermodynamic model to screen all possible organic co-solvents. Then, the required phase equilibrium data are generated from experiments for process design and validation of prediction results. We believe that the proposed approach will replace the current tedious trial-and-error method with a systematic and highly efficient way to identify the suitable organic solvents for the separation processes.
