對化學工程、製藥工程、環境工程、運輸工程或其他與化學相關產業而言，純物質與混合物的熱力學性質，如：蒸氣壓在研發、製程設計、最適化甚至於生命財產安全上有其一定的需求及應用。近年來Hsieh與Lin開發一個結合Peng-Robinson狀態方程式與COSMO-SAC活性係數模型[本文以PR+COSMOSAC(2010)表示之]的方法，此方法在不需要任何與物質相關的參數或實驗數據下，透過量子力學與COSMO-SAC計算，獲得PR-EOS中的能量參數a與體積參數b後，即可用來預測純物質熱力學性質，如：蒸氣壓、昇華壓、臨界性質，亦可預測多成份系統之相平衡，甚至可應用至超臨界流體的系統以及藥物溶解度。 本研究發現將分散項自由能對溫度取自然對數作圖的線性關係會較分散項自由能對溫度的一次方作圖之線性關係來的高，故第一部分針對PR+COSMOSAC(2010)裡的分散項自由能中的溫度函數進行修正，並以三組不同區間的蒸氣壓實驗值來探討實驗值對優化分散項參數之效應；此外，因PR+COSMOSAC所計算出的體積參數與Peng-Robinson狀態方程式所計算出的體積參數存在著多項式關係，故第二部分則探討體積參數修正之效應。結合此兩部分之結果提出PR+COSMOSAC(2017)模型，藉此改善PR+COSMOSAC模型於純物質的蒸氣壓預測精確度並以此方法預測純物質性質與雙成份之氣-液相平衡，包含1125個純物質蒸氣壓(ALD-P = 0.199)、1140個純物質昇華壓(ALD-P = 0.679)、435個純物質臨界溫度(AARD = 4.8%)、351個純物質臨界壓力(AARD = 8.1%)、283個純物質臨界體積(AARD = 18.05%)、1405個純物質沸點(AAD = 14.86 K)、1118個雙成份氣-液相平衡系統(AARD-P = 22.4%、AAD-y1 = 9.2%)，並將預測結果與PR+COSMOSAC(2010)及PR+COSMOSAC(2015)之預測結果進行比較，結果顯示除了臨界溫度、臨界體積、與雙成份氣-液相平衡系統組成預測外，PR+COSMOSAC(2017)預測結果皆優於前兩個版本的模型。 ;The knowledge of thermodynamic properties and phase behavior for pure substances and mixtures over a wide range of temperature and pressure is of importance not only for engineers to design, develop and optimize the equipment and processes in chemical and related industries but for the safety of our daily life and environmental protection. Recently, Hsieh and Lin proposed a method PR+COSMOSAC(2010) which utilizes quantum mechanical and COSMO solvation calculations to obtain the parameters in PR EOS. This method has been applied to predict several types of phase equilibrium, such as VLE, SVE or supercritical fluid systems. The first part of this work is to revise the dispersion energy term (〖∆▁G_i〗^(*dsp)) of PR+COSMOSAC(2010) because the linearity between 〖∆▁G_i〗^(*dsp) and natural logarithm of temperature is better than that between 〖∆▁G_i〗^(*dsp) and temperature. Then, three different sets of dispersion parameters are obtained by the regression of three different regions of vapor pressure data to study which one can provide the lowest deviation of vapor pressure prediction. The second part is to investigate the effect of the modified volume parameter on vapor pressure prediction because a polynominal relationship is found between bPR and Vi,COSMO. Finally, the PR+COSMOSAC(2017) is proposed on the basis of the results in the first two parts. Moreover, we compare the prediction deviation of vapor pressures for 1125 pure substances (ALD-P = 0.199), sublimation pressures for 1140 pure substances (ALD-P = 0.679), critical properties (AARD-Pc = 8.1%, AARD-Tc = 4.8% and AARD-Vc = 18.05% for 351, 435 and 283 pure substances, respectively) and VLE for 1118 binary systems (AARD-P = 22.4% and AAD-y1 = 9.2%) from PR+COSMOSAC(2017) with those from PR+COSMOSAC(2010) and PR+COSMOSAC(2015). The results show that the prediction accuracy of PR+COSMOSAC(2017) is superior to the versions of 2010 and 2015 except for the critical temperature, critical volume and the vapor phase composition of binary VLE systems.
