Correlation of Boiling Points with Molecular Structure. 1. A Training Set of 298 Diverse Organics and a Test Set of 9 Simple Inorganics

Abstract

A quantitative structure−property relationship (QSPR) treatment of the normal boiling points was performed for a structurally wide variety of organic compounds using the CODESSA (comprehensive descriptors for structural and statistical analysis) technique. A highly significant two-parameter correlation (R 2 = 0.9544, s = 16.2 K) employs just two molecular parameters, a bulk cohesiveness descriptor, G I 1/3, and the area-weighted surface charge of the hydrogen-bonding donor atom(s) in the molecule. A more refined QSPR model (with R 2 = 0.9732 and s = 12.4 K) includes, in addition, the most negative atomic partial charge and the number of the chlorine atoms in the molecule. The four-parameter equation offers an average predicted error of 2.3% for a standard set of compounds with an average experimental error of 2.1%. The QSPR equations developed allow remarkably accurate predictions of the normal boiling points for a number of simple inorganic compounds, including water.

A General Treatment of Solubility. 3. Principal Component Analysis (PCA) of the Solubilities of Diverse Solutes in Diverse Solvents

“…The second descriptor is related to the hydrogen bonding (HDSA-2). 28 We also noticed similarities with our previously reported two-parameter QSPR models of vapor pressure, where the graVitation index oVer all bonded atoms (G I ) and the hydrogen-bonding donor charged surface area (HDCA-2) gives a linear correlation with R 2 ) 0.88. 29 Two-parameter QSPR models for liquid viscosity (log η) also included the same two descriptors (G I and HDCA-2) giving correlation coefficient R 2 ) 0.79 using 337 30 and R 2 ) 0.81 using 361 diverse organic molecules, respectively.…”

A knowledge‐based approach for screening chemical structures withinde novomolecular evolution

“…Seventy-seven descriptors in total are used which cover seven classes of 3D descriptor including charged partial surface area (CPSA) [31], gravitation indices of mass distribution [32,33], molecular length-to-breadth ratio, molecular distance edge (MDE) [34], moment of inertia and radius of gyration, geometrical shape coefficients of radius-diameter diagram [35,36], Weighted Holistic Invariant Molecular (WHIM) descriptors [8], and three classes of 2D descriptors involving topological polar surface area (TPSA) based on fragment contributions [37], the number of hydrogen bond acceptors and the number of hydrogen bond donors. Molecules are represented by the vectors of those 77 descriptors.…”

Chemical Structure Indices in In Silico Molecular Design

“…Some geometrical descriptors useful in QSAR studies are 3D-Balaban index (J3D) [67], 3D Petitjean shape index (PJI3) [68] and Gravitational indices [69].…”

“…Among these, J3D [67] characterizes geometric distance degrees, PJI3 [68] is the fraction of the difference of geometric diameter and radius to the geometric radius, and Gravitational indices [69] reflect the mass distribution in molecules. The information content of these and many other graph theoretical descriptors have been validated through correlation studies involving different additive-constitutive properties of hydrocarbons [70][71][72] and biological properties of several classes of compounds [73][74][75][76].…”