Ostwald solubility coefficients, as log L, for solutes in water and ethanol have been combined to give log PEtOH for partition between the two pure solvents. Sixty-four such values have been correlated through our solvation equation, the coefficients of which lead to the conclusion that ethanol and water solvents are equally strong hydrogen-bond bases, but that ethanol is much weaker as a hydrogen-bond acid. A slightly different solvation equation has been used to correlate 68 values of log LEtOH; the coefficients in this equation yield the same conclusions as to the hydrogen-bond acidity and basicity of bulk ethanol. In addition, an analysis of the various terms in the log LEtOH correlation equation allows the elucidation of the various chemical factors that govern the solubility of gaseous solutes in ethanol solvent at 298 K.Key words: solubility, partition, hydrogen-bonding, ethanol, water.

Values of the Ostwald solubility coefficient of gases and vapours in methanol solvent, L-MeOH, at 298 K have been determined for 23 solutes by an indirect method in which experimental partition coefficients between methanol and hexadecane were combined with literature data on Ostwald solubility coefficients in hexadecane. Another 70 L-MeOH values were obtained from literature data and the total 93 values were correlated by the Abraham equation to give the regression, where n = 93, r(2) = 0.9952, sd = 0.13 and F = 3681.log L-MeOH = -0.004 - 0.215 R-2 + 1.173 pi(2)(H) + 3.701 Sigma alpha(2)(H) + 1.432 Sigma beta(2)(H) + 0.769 log L-16The solute descriptors In eqn. (i) are: R-2 an excess molar refraction, pi(2)(H) the dipolarity/polarisability, Sigma alpha(2)(H) the overall hydrogen-bond acidity Sigma beta(2)(H) the overall hydrogen-bond basicity and log L-16, where L-16 is the Ostwald solubility coefficient on hexadecane at 298 K, The number of data points, or solutes, is n, the correlation coefficient is r, the standard deviation is sd and Fis the F-statistic, Just as for the case of water solvent, solute dipolarity/polarisability, hydrogen-bond acidity and hydrogen-bond basicity all lead to an increase in log L, although methanol is much less acidic than water. However, contrary to the solubility of vapours in water, the log L-16 descriptor now also leads to an increase in log L, Explanations for the different behaviour of water and methanol are given.An analysis of log P values for the transfer of solutes from water to methanol also shows that bulk methanol is as strong a hydrogen-bond base as bulk water but is a much weaker hydrogen-bond acid.

Gas-liquid partition coefficients, K, have been obtained for 20-70 solute analytes on 14 candidate phases for chemical microsensors at 298 K and on three of the phases at higher temperatures. The phases can then be characterized through the equationlog K = c + rR(2) + s pi(2)(H) + a Sigma alpha(2)(H) + b Sigma beta(2)(H) + IlogL(16)where log K relates to a series of solutes on the same phase. The explanatory variables are solute parameters, R(2) an excess molar refraction, pi(2)(H) the solute dipolarity-polarizability, Sigma alpha(2)(H) and Sigma beta(2)(H) the solute hydrogen bond acidity and basicity and log L(16) where L(16) is the K-value on hexadecane. The coefficients in the above equation then characterize the particular phase, the most important being s the phase dipolarity-polarizability, a the phase basicity, b the phase acidity and I a constant that reflects a combination of cavity effects and general dispersion interactions and is related to the ability of the phase to distinguish between homologues. Derivation of the constants for the various phases provides a quantitative method for the analysis of the selectivity of phases for particular solute analytes and a term-by-term investigation of log K values shows exactly the solubility interactions that lead to sorption of a solute by a phase and hence to the analytical determination of the solute through chemical microsensors.

A general linear solvation energy equation has been used to analyze published partition coefficients in the systems water-octanol (613 solutes), water-hexadecane (370 solutes), water-alkane (200 solutes), and water-cyclohexane (170 solutes). The descriptors used in the equation areR2, an excess molar refraction; π2H, the solute dipolarity/polarizability; ∑α2Hand ∑β2H, the effective solute hydrogen-bond acidity and basicity; and VX, the characteristic volume of McGowan. It is shown that the water-octanol partition coefficient is dominated by solute hydrogen-bond basicity, which favors water, and by solute size, which favors octanol, but solute excess molar refraction and dipolarity/polarizability are also significant. For the water-alkane partition coefficients, the same factors are at work, together with solute hydrogen-bond acidity as a major influence that favors water. An analysis of 288 ΔlogPvalues shows that solute hydrogen-bond acidity is the major factor but that solute hydrogen-bond basicity and, to a lesser extent, solute dipolarity/polarizability and size are also significant factors that influence the ΔlogPparameter.

The solubility of 408 gaseous compounds in water at 298 K has been correlated through eqn. (i), where the solubility is expressed as the Ostwald solubility coefficient, L(w), and the solute explanatory variables are R2 an excess molar refraction, pi2H the dipolarity/polarizability, SIGMAalpha2H and SIGMAbeta2H the effective hydrogen-bond acidity and basicity, and V(x) the McGowan characteristic volume. A similar equation using the log L16 parameter instead of V(x) can also be used; L16 is the Ostwald solubility coefficient on hexadecane at 298 K. log L(w) = -0.994 + 0.577R2 + 2.549 pi2H + 3.813SIGMAalpha2H + 4.841SIGMAbeta2H - 0.869 V(x) (i) n = 408 rho = 0.9976 sd = 0.151 F = 16810 The main factors leading to increased solubility are solute pi2H, SIGMAalpha2H and SIGMAbeta2H values; conversely, the corresponding properties of water are dipolarity/polarizability, hydrogen-bond basicity and hydrogen-bond acidity. Solute size plays a minor role, and slightly decreases solubility, contrary to observations on all non-aqueous solvents. It is shown that this peculiar behaviour of water is due to (a) a greater increase in the unfavourable cavity effect with increase in solute size, for solvent water, and (b) a smaller increase in the favourable general dispersion interaction with size, for solvent water.A new method for the determination of log L(w) values is put forward, using the relationship L(w) = L16/P where L16 is as above, and P is either the water-hexadecane partition coefficient or the water-alkane partition coefficient. For 14 solutes using the former P-value, agreement with values calculated through eqn. (i) is 0.08 log units on average and for 45 solutes using the latter P-value, the corresponding agreement is 0.15 log units, with log L(w) values ranging up to 8 log units.

The general solvation equation

The sorption of vapors by fluoropolyol, poly(epichlorohydrin), and poly(isobutylene) is examined by gas-liquid chromatography (GLC), and these results are compared with the responses of surface acoustic wave (SAW) vapor sensors coated with the same polymers. The sensor responses exceed those which can be attributed to gravimetric effects, indicating that the SAW devices are responding to some other change in the coating properties. A model is developed to estimate the effect of polymer swelling on SAW sensor responses. The model is based on the use of partition coefficients determined by GLC as an independent measure of polymer mass loading, and polymer thermal expansion on SAW surfaces as a measure of volume change effects which is independent of mass loading effects. Both experimental comparisons and the model indicate that swelling effects can be ca. 4 times greater than mass-loading effects. The likely mechanism by which swelling influences the SAW sensor response is via reductions in the modulus of the polymer overlayer.

Previously reported results on twenty-two gaseous compounds with soybean oil as the stationary gas chromatographic phase have been used to characterize soybean oil in terms of dipolarity/polarizability, hydrogen-bond basicity and lipophilicity. The solubility of these gases in soybean oil has been factored into components that show exactly the compound-soybean oil interactions that favor solubility. The same equation used to obtain this information also can be used to predict the gas chromatographic specific retention volume and then the weight-fraction activity coefficient for numerous other compounds on soybean oil, thus leading to predictions of the solubility behavior of these compounds as bulk liquids with soybean oil.

The new solvation equation: log L = c + rR2 + s-pi-2H + alpha-alpha-2H + b-beta-2H + l log L16 has been applied to the solubility of 43 gaseous probes on each of nine hydrocarbon polymers using the data of Munk et al.. In this equation, L is the gas-liquid partition coefficient of a series of probes on a given polymer, and the explanatory variables are solute properties as follows: R2 is an excess molar refraction, pi-2H is the probe dipolarity-polarizability, alpha-2H and beta-2H are the probe hydrogen-bond acidity and basicity, and L16 is the gas-liquid partition coefficient of the probe on hexadecane at 25-degrees-C. Each of the nine equations, one for each polymer, had correlation coefficients of around 0.999 and standard derivations of around 0.025 log units. The solubility of the gaseous probes, as log L values, as well as the polymer-probe interaction parameter-chi calculated by Munk, have been analysed in terms of particular polymer-probe interactions.

Two solvation energy models, developed independently for the characterization of the solvent properties of gas-liquid chromatographic stationary phases, are compared using accurately determined gas-liquid partition coefficients for 30 test solutes on 25 stationary phases at a common reference temperature. Remarkably good agreement between the regression model of Abraham and the free energy model proposed by Poole is demonstrated for the contributions to retention characterized by cavity formation, nonpolar interactions and polar interactions. Exceptional behavior for the liquid organic salts is explained by the differences between the experimental and predicted gas-liquid partition coefficients for the n-alkanes used in the two models. In addition, both models conclusively demonstrate that at the measurement temperature, 121.4°C, none of the stationary phases behave as significant hydrogen-bond acid solvents, contrary to commonly held beliefs.

The truncated solvation equation log SP = c + rR2 + l log L16 has been applied to numerous sets of gas-liquid chromatographic (GLC) data for alkylaromatic hydrocarbons on non-polar stationary phases. Here SP can be V(G) or can be the relative retention time, and the retention index I can in this context be used in place of log SP. R2 is the solute excess molar refraction, easily obtained from refractive index. A set of solutes of known log L16 is used to set up the equation and then values of log L16 can be back-calculated for other solutes; L16 is originally defined as the solute gas-liquid partition coefficient on hexadecane at 25-degrees-C. Through the above equation log L16 values were calculated for 190 solutes. Once log L16 is known, the reduced equation log SP = c + s-pi(2)H + l log L16 can be applied to GLC data on polar stationary phases, and the dipolarity/polarizability parameter pi(2)H obtained by back-calculation in a similar way. Values of pi(2)H for 120 solutes are listed. It is shown that n-alkyl substituents affect the pi(2)H value only slightly, but ortho substituents considerably increase pi(2)H, e.g., benzene (0.52), toluene (0.52), o-xylene (0.56), 1,2,3-trimethylbenzene (0.61), 1,2,3,4-tetramethylbenzene (0.65), pentamethylbenzene (0.66), hexamethylbenzene (0.72).

Two commercial samples of poly(methylphenylsiloxane) were characterised using our solvation equation, log L = c + rR2 + s-pi(H)(2) + a-alpha(H)(2) + b-beta(H)(2) + l log L16 where L is the gas-liquid partition coefficient for a series of solutes on a given stationary phase, and the explanatory variables are R2 an excess molar refraction, pi(H)(2) the solute dipolarity/polarisability, alpha(H)(2) and beta(H)(2) the solute hydrogen-bond acidity and basicity, and log L16 where L16 is the solute gas-liquid partition coefficient on hexadecane at 25-degrees-C. For both samples, a substantial b-constant was found, viz. 1.22 +/- 0.07 and 0.49 +/- 0.08 at 25-degrees-C, suggesting that they can act as hydrogen-bond acid (contrary to their chemical formulation). Examination of the bulk liquid stationary phases by IR showed the presence of OH groups and confirmed our analysis by the solvation equation. It is suggested that workers using the OV or SE series of siloxanes routinely check the bulk stationary phases by IR in order to assess the presence or absence of OH groups.

Two commercial samples of poly(methylphenylsiloxane) were characterised using our salvation equation, log L = c + rR2 + sπ2H + aα2H + bβ2H + l log L16 where L is the gas-liquid partition coefficient for a series of solutes on a given stationary phase, and the explanatory variables are R2 an excess molar refraction, π2H the solute dipolarity/polarisability, α2H and β2H the solute hydrogen-bond acidity and basicity, and log L16 where L16 is the solute gas-liquid partition coefficient on hexadecane at 25°C. For both samples, a substantial b- constant was found, viz. 1.22 ± 0.07 and 0.49 ± 0.08 at 25°C, suggesting that they can act as hydrogen-bond acid (contrary to their chemical formulation). Examination of the bulk liquid stationary phases by IR showed the presence of OH groups and confirmed our analysis by the solvation equation. It is suggested that workers using the OV or SE series of siloxanes routinely check the bulk stationary phases by IR in order to assess the presence or absence of OH groups.

The general solvation equation log K = c + rR2 + sπ2H + aα2H + bβ2H + l log L16 has been used to characterise 24 gas-liquid chromatographic stationary phases for which Poole and co-workers have determined log K values for a series of solutes at 121.4°C. The explanatory variables are R2, asolute excess molar refraction, π2H, the solute dipolarity, α2H and β2H, the solute hydrogen-bond acidity and basicity, and log L16, where L16 is the solute gas-liquid partition coefficient on hexadecane at 25°C. It is shown that the bβ2H term is not significant for any phase, and that the molten salts are all strongly dipolar and basic, with large s and a constants. A term-by-term analysis of the solvation equation yields a quantitative measure of the contribution to log K of various solute-stationary phase interactions, and leads to an understanding of how these interactions affect solute retention. The use of the characteristic constants c, r, s, a, b and l in the selection of stationary phases for particular separations is described.

AbstractChemInform is a weekly Abstracting Service, delivering concise information at a glance that was extracted from about 100 leading journals. To access a ChemInform Abstract of an article which was published elsewhere, please select a “Full Text” option. The original article is trackable via the “References” option.