Experimental Determination and Computational Prediction of Androstenedione Solubility in Alcohol + Water Mixtures

详细信息    查看全文

• 作者：Weiwei Tang ; Zhao Wang ; Ying Feng ; Chuang Xie ; Jingkang Wang ; Chengsheng Yang ; Junbo Gong
• 刊名：Industrial & Engineering Chemistry Research
• 出版年：2014
• 出版时间：July 16, 2014
• 年：2014
• 卷：53
• 期：28
• 页码：11538-11549
• 全文大小：742K
• ISSN：1520-5045

文摘

This article evaluates the accuracy and applicability of three of the most common solubility models (i.e., Jouyban鈥揂cree, NRTL-SAC, and COSMO-RS) in prediction of androstenedione (AD) solubility in binary mixtures of methanol + water and ethanol + water. The solubilities were measured from (275 to 325) K using medium-throughput experiments and then well represented mathematically by modified Apelblat and CNIBS/Redlich鈥揔ister equations. The computational results show that AD solubility decreases monotonically with increasing water concentration in methanol + water mixtures, but it has a maximum at 0.15鈥?.30 mole fraction of water in the ethanol aqueous solution. Moreover, the performance of three solubility prediction models in this particular case was compared to identify the advantages and disadvantages of each model. The overall average relative deviation (ARD) for solubility prediction is 4.4% using Jouyban鈥揂cree model, while it is 18.3% with NRTL-SAC model. Surprisingly, COSMO-RS model in combination with reference solubility achieves a good performance for solubility prediction in mixed solvents, including the prediction of synergistic effect of solvents, with overall ARD of only 4.9%.