Abstract The knowledge of solubility of solid solute in supercritical carbon dioxide is crucial for the design and optimization of relevant processes in food and pharmaceutical industries. We systematically examine the accuracy of a first-principles thermodynamic model PR + COSMOSAC EOS in predicting solid solute solubility in supercritical carbon dioxide and compare to that of a group contribution model PSRK EOS. Molecular structure, heat of fusion, and melting temperature of the investigated compound are the only required information for the solubility calculation. All other information can be obtained from quantum mechanical solvation calculations. The average logarithmic deviation in mole fraction (ALD-y) in predicting 99 solid solute solubility in supercritical carbon dioxide (a total of 2450 data points with a temperature range of 298 K–473 K and a pressure range of 45 bar–500 bar) was found to be 0.79 (or about 510%). This indicates that the PR + COSMOSAC EOS is able to reasonably estimate solubility order of magnitude without any adjustable model parameter and provide useful solubility information, especially when no experimental solubility data is available.

中文翻译：

---

使用 PR+COSMOSAC EOS 预测超临界二氧化碳中固溶质溶解度的第一性原理

摘要 固体溶质在超临界二氧化碳中的溶解度知识对于食品和制药行业相关工艺的设计和优化至关重要。我们系统地检查了第一性原理热力学模型 PR + COSMOSAC EOS 在预测超临界二氧化碳中固溶质溶解度方面的准确性，并与群贡献模型 PSRK EOS 的准确性进行了比较。所研究化合物的分子结构、熔化热和熔化温度是溶解度计算所需的唯一信息。所有其他信息都可以从量子力学溶剂化计算中获得。预测超临界二氧化碳中 99 个固溶质溶解度的摩尔分数平均对数偏差 (ALD-y)（总共 2450 个数据点，温度范围为 298 K–473 K，压力范围为 45 bar–500 bar）发现为 0.79（或约 510%）。这表明 PR + COSMOSAC EOS 能够在没有任何可调整模型参数的情况下合理估计溶解度数量级，并提供有用的溶解度信息，尤其是在没有实验溶解度数据可用时。