Prediction of solubility curves and melting properties of organic and pharmaceutical compounds

(Solid+liquid) solubility of organic compounds in organic solvents – Correlation and extrapolation

“…0 < x < 1 and 0 < a s < 1. This means that ln γeq ≥ ln a s [3]. A 1/T-dependence in ln γeq will, for negative deviation from Raoult's law, result in activity coefficients at very low temperatures decreasing rapidly, causing artefacts of extrapolated mole fractions exceeding unity.…”

“…This is done in the interest of further reducing the number of parameters, and to make the model more robust by eliminating the inconvenient necessity of making sure that C converges to ≥ 2. This three-parameter model is compared to i) the semi-empirical regression model, equation (13) with parameters from tables 1 and 2, as well as two empirical regression equations: ii) equation (17), a common, simple three-parameter equation, and iii) equation (18), a more flexible four-parameter equation [3]. Figure 8.…”

“…It is also possible, in theory, to obtain the melting temperature and the associated enthalpy and entropy of fusion through extrapolation of solubility data. As shown by Nordström and Rasmuson [3], through an evaluation of 15 different functional forms of regression equations based on 41 organic compounds, the accuracy that can be expected when extrapolating to higher temperatures is in general fairly low, and depends on the functional form of the regression equation used as well as on the temperature range of extrapolation. In general, while simple polynomial functions will often be quite sufficient for interpolation purposes they are not able to capture the full complexity of the temperature dependence of a solubility curve all the way to the melting point, and conversely, higher order polynomials tend to become too erratic to allow any kind of extrapolation.…”

Thermodynamics of risperidone and solubility in pure organic solvents

“…7, with the possible exception of cumene. A number of regression equations have been proposed for temperature correlation of solubility data, many of which were evaluated by Nordström and Rasmuson [11]. For risperidone, all of these provide good fits to the experimental solubility data in terms of average coefficients of determination, which exceeds 0.97 for all models including a linear equation in T -1 .…”

“…Such extrapolations are frequently done on the basis of assuming linear van't Hoff plots [13][14][15][16]. It has been shown [11] that a linear model will generally lead to poor accuracy, but with a proper choice of model extrapolation of van't Hoff curves to ln xeq = 0 can produce a reasonable estimate of the melting properties of the pure solute. At ln xeq = 0 the intersection of the extrapolated van't Hoff curve with the x-axis (Fig.…”