Volume 77, Issue 3

Solubilities in supercritical fluids are of great importance in a wide variety of applications. These applications include: production of controlled drug delivery systems, powder processing, pollution prevention and remediation, methods for spraying paints and coatings, precipitation/crystallization processes, bioseparations, and food processing. Examples of some of these applications will be given. Supercritical fluid processes make use of the significant variation of solubility with pressure when the solvent is a supercritical fluid. This permits the accomplishment of many processes that would be otherwise very difficult to achieve. To fully understand the benefits of such processes, one must understand the phase behavior of such systems. The types of phase behavior that can be observed will be discussed, along with the ways that this behavior can be used to modify solubilities in such fluids. There have been many methods proposed for the experimental determination of solubilities in supercritical fluids. These will be reviewed in general, and recommendations for preferred methods given. Solubilities in supercritical fluids can be correlated and predicted by making use of thermodynamic calculations involving equations of state, as well as by other methods. The various methods will be reviewed, and recommendations given for various applications.

The extended universal quasichemical (UNIQUAC) model is a thermodynamic model for solutions containing electrolytes and nonelectrolytes. The model is a Gibbs excess function consisting of a Debye–Hückel term and a standard UNIQUAC term. The model only requires binary ion-specific interaction parameters. A unique choice of standard states makes the model able to reproduce solid–liquid, vapor–liquid, and liquid–liquid phase equilibria as well as thermal properties of electrolyte solutions using one set of parameters.

This report presents the systematic study on the solubilities of 1-alkyl-3-methylimidazolium hexafluorophosphate [e, or bmim][PF 6 ], 1-alkyl-3-methylimidazolium methylsulfate [almim][CH 3 SO 4 ], 1-hexyloxymethyl-3-methylimidazolium ionic liquids (ILs) [C 6 H 13 OCH 2 mim][BF 4 ], or [C 6 H 13 OCH 2 mim][(CF 3 SO 2 ) 2 N] in aliphatic hydrocarbons (heptane, octane), cyclohydrocarbons (cyclopentane, cyclohexane) and aromatic hydrocarbons (benzene, toluene, ethylbenzene, o -xylene, m -xylene, p -xylene), and in alcohols (methanol, ethanol, propan-1-ol, propan-2-ol, butan-1-ol, butan-2-ol, tert -butyl alcohol, and 3-methylbutan-1-ol) as well as of 1-alkyl-3-methyl-imidazolium chloride [C 4 , C 10 , or C 12 mim][Cl] in alcohols. The solubilities have been measured by a dynamic method from 290 K to the melting point of IL or to the boiling point of the solvent. The solubility of [emim][PF 6 ] and [bmim][PF 6 ] in aromatic hydrocarbons decreases with an increase of the molecular weight of the solvent. The difference on the solubility in o -, m -, and p -xylene is not significant. The solubility of [emim][PF 6 ] in alcohols decreases with an increase of the molecular weight of the solvent and is higher in secondary alcohols than in primary alcohols. In every case, with the exception of methanol, the mutual liquid–liquid equilibrium was observed. The shape of the equilibrium curve is similar for [emim][PF 6 ] in every alcohol. The observations of upper critical solution temperatures were limited by the boiling temperature of the solvent. For example, for [emim][PF 6 ], solubilities in methanol and ethanol are higher than that in aromatic hydrocarbons. The miscibility gap for C 3 alcohols is higher than that in benzene, but comparable with the solubility in toluene; solubility in 3-methyl-butan-1-ol is very similar to ethylbenzene and o -, m -, and p -xylene. The data were correlated by means of the UNIQUAC and modified nonrandom two-liquid (NRTL) equations utilizing parameters derived from the solid–liquid equilibrium (SLE). The root-mean-square deviations of the solubility temperatures for all calculated data depend on the particular system and the equation used. The solubilities of [C 4 , C 10 , or C 12 mim][Cl] and alkyl-(2-hydroxyethyl)-dimethyl-ammonium ILs in octan-1-ol and water have been measured and used to calculate the octan-1-ol/water partition coefficients as a function of temperature and alkyl substituent. Experimental partition coefficients (log P ) are negative for imidazoles, 1,3-dialkylimidazolium chlorides, and investigated ammonium salts. Good knowledge of ILs permits the elimination of volatile organic compounds and makes chemistry cleaner and greener.

The importance of biomolecules is well recognized nowadays, owing to their application in many industrial processes, particularly in the food, pharmaceutical, and cosmetic industries. Amino acids, carbohydrates, and proteins are the three types of biomolecules considered in this review. Solubilities of several amino acids and sugars have been measured in water and, more recently, in mixed solvents, both with or without salts. Experimental data on partition of proteins in aqueous two-phase systems (ATPSs), with polymers or salts, have also been reported in the literature. The experimental techniques used are briefly discussed. Regarding modeling, the complexity of these solutions is very high. Sugars form strong hydrogen bonds with one another and with water or other solvents like alcohols. Liquid solutions with amino acids or proteins present not only short-range interaction forces, but also long-range forces, owing to the appearance of charged species. The capabilities of different molecular models and group-contribution-based methods are shown.

The protonation constants of two series of tetraazamacrocyclic ligands with acetate and methylphosphonate pendant arms, as well as their stability constants with Cu 2+ , La 3+ , Sm 3+ , and Ho 3+ , were determined. All the values were determined in aqueous solution at 298.0 K and 0.10 mol dm –3 in N(CH 3 ) 4 NO 3 . In the first series, the effect of both types of pendant arms was observed by appending them in the same macrocyclic backbone, a 14-membered tetraazamacrocycle containing pyridine (ac 3 py14, p 2 py14, and p 3 py14). In the second series, two effects were taken into account, the increase of the cavity size of the macrocycle, from 12- to 14-membered, and the appending of acetate (dota, trita, and teta) or methylphosphonate (dotp, tritp, and tetp) arms. The ligands containing methylphosphonate arms have higher thermodynamic stability compared to the corresponding ones with acetate arms, especially in the series of compounds containing pyridine, even upon correction of the different basicity values of the ligands. On the other hand, the ligands with smaller macrocyclic cavity size, namely, dota and dotp, exhibit the largest values of stability constants. In contrast, ac 3 py14 presents low stability constants with lanthanides. An interpretation of these features based on the known adopted arrangement of dota and teta when free or coordinated with lanthanides is evaluated.

The thermodynamic stability and transformation of pharmaceutical polymorphs was investigated, especially those of the thiazole derivative pharmaceutical, 2-(3-cyano-4-isobutyloxyphenyl)-4-methyl-5-thiazolecarboxylic acid (BPT). The influence of methanol compositions in solvents and temperature on the solubility and the transformation behavior of BPT was clarified. The transformation behavior was explained by the chemical potential difference between the stable and metastable forms. It was shown that a specific solute–solvent interaction contributes to the preferential nucleation and growth of the stable or metastable forms and influences the transformation behaviors. The solubility of BPT of the solvated crystals is much more influenced by the solvent compositions than the true polymorphs. The solubility ratio of the solvated crystals depends on the solvent composition, whereas the solubility ratio of the true polymorphs is considered to be independent of the solvents. The crystallization behavior was also investigated. The transformation rate after crystallization appeared to depend on the initial concentration of BPT and the addition rate of the antisolvent. The cause of this phenomenon was presumed to be a slight inclusion of the stable form in the metastable form.

The newly developed VW- pVT technique for simultaneously measuring the solubility of a gas in a polymer and the concomitant change of volume of the polymer has been used to estimate the solubilities of N 2 , CO 2 , and two hydrofluorocarbons, HFC-134a and HFC-152a, in polystyrene (PS). In conjunction with these solubility data, Chow’s model has been used to calculate the change of the glass-transition temperature of the polymer resulting from gas sorption. The performance of this model for predicting the glass-transition temperature shift is then discussed.

The importance of ion solvation in determining the properties of electrolyte solutions in aqueous–organic solvent mixtures is discussed. Solubility measurements are shown to be particularly useful for determining the Gibbs energies of transfer of ions between solvents, which reflect differences in the overall solvation of the ions in different solvent mixtures. Solubility measurements can also be used to determine the other thermodynamic parameters of transfer, but such quantities are usually better obtained by more direct methods. The inadequacy of current theories of ion solvation to quantitatively account for the thermodynamics of ion transfer is discussed by reference to measurements on some simple model systems. Although donor/acceptor interactions can explain many of the observed effects between pure solvents, the situation is more complex in aqueous–organic mixtures because selective solvation and even solvent–solvent interactions may become significant. This is illustrated by consideration of ion transfer from water to water + t -butanol solutions, where spectacular effects are observed in the enthalpies and entropies and especially in the heat capacities and volumes.

For the calculation of aqueous solubility of electrolytes, thermodynamic data from different sources are required. This can lead to errors if the data are inconsistent. This study reveals that some often-used semi-empirical equations for CO 2 solubility and other equilibria are inconsistent with CODATA (Committee on Data for Science and Technology) key thermodynamic data. The equations were recalculated to make them consistent with CODATA. Combining standard thermodynamic data of NaCl with the Pitzer model leads to significant deviations from the accepted experimental value at 25 °C. Owing to the inadequacy of the Pitzer model at molalities exceeding 6 mol kg –1 , this model leads to poor predictions of the HCl vapor pressure in equilibrium with highly concentrated aqueous HCl solutions. A long-standing inconsistency problem is related to the solubility of calcium carbonate. The main problem is disagreement on the existence of the CaHCO 3 + ion pair. It is shown that the inconsistency largely disappears if it is assumed that the CaHCO 3 + exists at low ionic strength, but becomes less stable at higher ionic strength.

If a true thermodynamic equilibrium with a well-known solid is expected to establish, chemical equilibrium thermodynamics allows estimation of the maximum concentration of a given radionuclide in a specified pore fluid of an underground repository. However, in the course of the review process for the Nagra/PSI Chemical Thermodynamic Data Base 01/01, important cases of insufficient chemical knowledge were identified, leading to gaps in the database. First, experimental data for the ThO 2 –H 2 O and UO 2 –H 2 O systems cannot be interpreted by a unique set of thermodynamic constants. There, a pragmatic approach was chosen by including parameters in the database that are not thermodynamic constants in a strict sense, but that reproduced relevant experimental observations. Second, potentially important thermodynamic constants are missing because of insufficient experimental data. Estimations of these missing constants led to problem-specific database extensions. Especially constants for ternary mixed carbonato-hydroxo complexes of tetravalent actinides have been estimated by the “backdoor approach”, i.e., by adjusting thermodynamic constants to maximum feasible values that are still consistent with all available experimental solubility data.

Numerous factors control the chemistry of metal ions in natural aqueous systems. These include the presence of a host of complexing agents, speciation among a variety of competing ligands, buffering due to precipitated solid phases, and the relative influences of kinetic versus thermodynamic phenomena. Natural aqueous systems are inevitably complicated, but simple, inorganic models can be constructed to characterize a variety of different environments. While available data sets for these models are large, it is clear that much further research is required in order to develop more sophisticated models. Some of the fields of research that need to be addressed are outlined, and some of the constraints on such research are briefly discussed.

Two approaches are followed to understand how molecular interactions influence the macroscopic properties of solutions: (1) experiment, through the determination of gas solubility, and (2) computer simulation, used to evaluate microscopic properties (structural and energetic). Examples of application of these approaches are considered in order to explain the properties of solutions containing fluorinated fluids or ionic liquids. The molecular structures and interactions are described by force fields built from ab initio quantum chemical calculations. These models allow the determination of free energies from computer simulations by using appropriate energy routes provided by statistical mechanics. The macroscopic properties related to the process of dissolution of several gases are interpreted in terms of the molecular structure of the solutions and of the solute–solvent interactions.