Improved Prediction of Vapor Pressure for Pure Liquids and Solids from the PR+COSMOSAC Equation of State

Abstract

Vapor pressure of pure substances is a crucial piece of information for many industrial applications. A recently developed PR+COSMOSAC equation of state (EOS) with its molecular interaction parameters determined from quantum mechanical calculations has been shown to provide reasonable prediction accuracy for vapor pressure of almost any chemical species without the issue of missing parameters. In this work, we introduce two modifications to improve its prediction accuracy on the vapor pressure. The overall deviation in pressure for 1124 pure liquids from the modified version is 138%, which is about 1/4 of that from the original model 553%. In particular, the accuracy for the vapor pressure near triple point shows major improvement, with the average error reduced from 1062% to 233%.The sublimation pressure can also be estimated providing that the melting temperature and enthalpy of fusion are available. The average deviation in sublimation pressure from the modified PR+COSMOSAC EOS for 1140 substances is 412%, which is only 1/3 of that from the original model (1249%). This model is capable of providing both the vapor pressure and sublimation pressure over a wide range of conditions (from the critical point to below the triple point). It is particularly useful when experimental data are not available.

Improvement to PR+COSMOSAC EOS for Predicting the Vapor Pressure of Nonelectrolyte Organic Solids and Liquids

“…56,57 When the QM/COSMO calculation results for substances considered are available, the calculation procedure of thermophysical properties for pure substances and vapor− liquid equilibria of binary mixtures are referred to the literature. 37,39,43 It should be noted that the heat of fusion and melting temperature are required for determining solid phase fugacity in sublimation pressure calculations.…”

“…47 Therefore, two modification functions to the dispersion contribution were introduced to correct for the systematic underestimation of molecular interactions, and the accuracy of PR+COSMOSAC EOS in predicting vapor pressures and sublimation pressures of pure substances and vapor−liquid equilibrium of binary mixtures was improved significantly. 43 However, the temperature dependence of the dispersion contribution becomes a bit complicated, and the accuracy in describing hydrogen-bonding compounds remains almost the same.…”

“…The accuracy of the revised PR+COSMOSAC EOS in predicting the vapor pressure, sublimation pressure, normal boiling temperature, and critical properties of pure substances and the vapor−liquid equilibrium of binary mixtures is investigated and compared to those from previous versions of the model. 39,43 2. THEORY 2.1.…”

Visual and Semantic Enrichment of Analytical Chemistry Literature Searches by Combining Text Mining and Computational Chemistry

“…These structures are used to derive a number of molecular descriptors. From these descriptors and available experimental data, machine learning is used to predict important physicochemical and biological properties, such as aqueous solubility, − melting point, , vapor pressure, , bioavailability, , developmental toxicity, , and receptor binding …”

“…These structures are used to derive a number of molecular descriptors. From these descriptors and available experimental data, machine learning is used to predict important physicochemical and biological properties, such as aqueous solubility, 11−13 melting point, 11,14 vapor pressure, 15,16 bioavailability, 17,18 developmental toxicity, 19,20 and receptor binding. 21 This note describes how text mining for chemical entities can be combined with prediction of their physicochemical properties to guide the selection of analytical methods, extract information on proteins and their ligands, or study the history of a particular subfield of chemistry in semantically enriched literature searches.…”