Density Considerations
Thermodynamic Properties
Diffusivity/Viscosity Considerations
Transport Properties
Biological Properties
Solvating Strength
General Process Concepts
Modified Supercritical Fluids

As indicated previously, an exceedingly small number of viable fluids can be used in analytical SFE. As stated previously, those that might possess higher sol­vating power than CO 2 have higher Tc and Pc values. Table 10 lists the solubility parameters calculated from the Lee and Kessler31 equation of state (EOS) for a number of pure components under supercritical conditions (i.e., TR = 1.02; PR = 2). The solubility parameter is defined as the square root of the cohesive energy density which is the energy content (E, cal/mol) of the fluid per unit molar vol­ume (v, cm3/mol) relative to its ideal gas state. It is obvious from this list that only ammonia has a higher solvating parameter than CO 2. Shoenmakers32 indi­cated that the solubility parameter is a better indication of overall polarity of a solvent than the dipole moment or dielectric constant. He further notes that there is no correlation between the dipole moment of a solvent and its solubility para­meter under SF conditions. Consider that 5 = 5.6, 5.7, 8.9, 9.3, and 13.5 for dichloromethane, ethyl acetate, methanol, ammonia, and water, whereas the re­spective dipole moments (in debye) are 1.8, 1.9, 1.7, 1.5, and 1.8.

Even at high densities, CO 2 has limited ability to dissolve polar molecules.

However, the characteristics of SF CO 2 can be enhanced by the addition of mis-cible, polar compounds to the primary fluid. The second component is referred to as a modifier, cosolvent, or entrainer. These substances, which are liquids un­der ambient conditions, have been used extensively to alter SF CO 2's solvating properties. Table 11 lists the most widely used modifiers and their selected physicochemical properties. Methanol is currently the most common modifier for SF CO 2. Mixed fluid solvent systems are complicated by their phase behav­ior. In addition, Tc and Pc are altered by the incorporation of modifier into SF CO 2. The Tc of different CO 2-modifier mixtures lies between those of the two pure components.

Table 10 Solubility Parameters of Selected Fluids at Tr= 1.02and/>R = 2

Table 11 Frequently Used Modifiers


The Pc value usually shows a maximum at intermediate com­position between that of either pure CO 2 or the pure liquid modifier. The critical constants of the mixture can be approximated as the arithmetic mean of the criti­cal temperatures and pressures of the two components (e.g., CO 2 and modifier) as follows:

T C + X CO2 T c(CO2) + Xm Tc(m)

P c = X CO 2 P c(CO 2) + Xm Pc(m)

where X CO2 and Xm are the mole fractions of CO 2 and modifier respectively.33 . More elaborate treatments for calculating the Tc and Pc of mixtures are the method of Cheuh and Prausnitz34 for Te, Kreglewski and Kay's for Pc 35 or the Peng-Robinson equation of state.36

While the incorporation of modifier enhances the fluid's solvating power in many instances, the effects on mass transfer may also be altered, especially if the modifier interacts with the solute. Diffusion coefficients for acridine, phenan-threne, and benzoic acid have been measured in pure and methanol-modified (3.5 mol %) CO 2 as a function of temperature (Table 12).37 Acridine and benzoic acid showed considerably lower diffusion coefficients in the modified fluid which indicated that methanol associates (e.g., hydrogen bonding) with the ana-lytes. In the presence of modifier, phenanthrene, being neither acidic or basic, "173 bar. *3.5 mol %.

showed no change in diffusion coefficient relative to 100% CO 2 within experi­mental error.

Table 12 Diffusion Coefficients for Acridine, Phenanthrene, and Benzoic Acid (10' D12, m2/s)

Not all polar liquids are highly miscible with SFs. It is known, for example, that water is slightly soluble (< 0.1% w/w) in CO 2. Attempts to create higher wa­ter concentrations in SF CO 2 produced two phases. Even dissolution of small quantities of water in methanol or isopropyl alcohol phase produced multiple phases when the water-alcohol mixture was added to CO 2.

It has been reported the solubility of methanol, acetonitrile, and CHC1 3 modifiers in CO 2 at various temperatures (-4 to 50°C) and pressures (15-170 arm).38 For a methanol/CO 2 mixture to be homogeneous between -4 and 29°C, the methanol concentration must be kept below 12 mol %. It was noted that above 90 atm, the increase of methanol solubility on both pressure and tem­perature was significant, as shown in Figure 11. However, the solubility of ace­tonitrile in CO 2 was increased only slightly with increased pressure and tempera­ture. To ensure homogeneity of the mixture at pressures up to 170 atm, acetonitrile concentration should be kept below 1.6 mol %. Note that both tem­perature and pressure had a definite influence on the solubility of CHC1 3 in CO 2. A CHCI 3 concentration of 1.4 mol % was reported to be necessary for obtaining a homogeneous CHC1 3/CO 2 mixture. The work of A. D. King39 determined that at low pressures methanol was miscible in CO 2 to a much greater extent than in N 2, Ar, or CH 4. A greater understanding of the solubility of modifiers such as methanol, acetonitrile, and chloroform in SF CO 2 at various densities and tem­peratures is greatly needed. According to Francis,40 all three liquids (CH 3OH, CH 3CN, and CHCI 3) are mutually miscible with liquid CO 2 at room temperature.

Fig. 11. Solubility of methanol in CO 2 as a function of pressure and temperature.


These results, however, were only for room temperature and involved a third component in the system which no doubt affected the measured solubility. Fur­thermore, the systems were not saturated with the liquid. The effect of a third component on solubility can readily be seen from the following data. When mixed with kerosene, CH 3OH is about 70% soluble in liquid CO 2, but when mixed with LiCI, CH 3OH solubility is closer to 20% and more like the results re­ported.

In the chemical engineering literature, clustering of polar modifier mole­cules around polar solute molecules in a large excess of a SF is widely accepted. Even in mixtures containing low-bulk concentrations of modifier, solutes may be surrounded by a solvation sphere containing a high local concentration of the modifier. Such clustering becomes more pronounced as the difference in polarity between the SF and the modifier is increased. The size of these clusters can reach values on the order of 100 solute molecules. This molecular attraction accounts for the high negative partial molar volume of a solute in a highly compressible SF solvent. It is reported that at high reduced pressures, where the flu­id is incompressible, the infinite dilution partial molar volume is slightly posi­tive. A very sharp negative dip in partial molar volume, however, was observed for naphthalene near the critical point of ethylene when using ethylene as a sol­vent (Fig. 12). As the partial molal volume of the solute begins to decrease dra matically, the average distance between molecules decreases and nonideal gas behavior begins to dominate the interactions between the solvent and solute, thus accounting for the enhancement in solubility.

Fig 12. Partial molar volume (of) vs. density for naphthalene in ethylene at 12"C (Tc = 9.3°C).

As the system temperature in­creased, the partial molar volume became less negative, which indicated that the strength of solute-solvent interactions and the size of solute-solvent clusters de­creased.

The solvent strength of mixtures of normal liquids with supercritical and near-critical fluids (e.g., CO 2, Freon-13, and Freon-23) has been measured us­ing Nile Red as a solvatochromic dye. Nile Red is soluble in several pure SFs and in polar mixtures of SFs and modifiers. It exhibits large shifts in the wave­length of its absorption maximum, thereby allowing subtle changes in solvent strength with fluid composition to be quantified. It is also stable in very strong acids. There is considerable debate about the value of solvatochromic solvent strength scales, but at least they provide an approximate correlation to solvent polarity. Transition energies for Nile Red were compared with results from the better known Reichardt's £t(26) and £,(30) dyes. The results obtained in this study are displayed in Figure 13. On either solvent strength scale [£(nr> or Et{i0)], the polarity of the pure fluid followed the order Freon-13 < CO 2 < Freon-23. The ad­dition of methanol as modifier appeared to shift the fluid solvent strength dra­matically. The larger the difference in solvent strength between the pure SF and modifier, the more dramatically the solvent strength of the mixture shifted. Thus, methanol-Freon-13 mixtures appeared to be more polar than methanol-CO 2 mixtures of similar composition on either solvent strength scale.

Fig. 13. Comparison of transition energies calculated from solvatochromic data. Wavelengths of maximum absorption of Nile Red and El(10) solvatochromic dyes for mixtures of methanol and three SFs.

Polar molecules were only sparingly soluble in Freon-13 and tended to be rejected by it. None of the dyes were soluble in pure Freon-13, nor could mixtures over 12% methanol in Freon-13 be produced between 25 and 80°C over a wide range of pressures.

Spectral shifts for a probe molecule, phenol blue, in SF CO 2 were reported earlier. Data obtained with SF CO 2 using the same probe with various amounts of methanol have now been reported (solvatochromic data). Considering a pressure of 100 bar where the absorption or transition energy (ET) of phenol blue in SF CO 2 is 54 kcal/mol, the red shift was increased more by the addition of 3.5 mol % methanol at constant pressure than by increasing the pressure of pure CO 2 to over 200 bar (Fig. 14). This red shift is also in excess of the theoretically pre­dicted shift if it is assumed that the mixture is random as opposed to containing clusters. With this phenol blue dye, the contribution of hydrogen bonding to sol­vent strength was incorporated.

In summary, solvating strength in the supercritical region is a direct function of density—which in turn is dependent on system pressure (at constant tempera­ture) or temperature (at constant pressure). Interestingly, there is no break in the continuity of solvent strength as the material goes from the near critical to the su­percritical region.

Fig. 14. The ET (transition energy) of phenol blue in a CO 2-methanol mixture (D = pure CO 2; A = CO 2-1% CH3OH; O = CO 2 - 3.5% CH3OH).

Thus it is possible to "fine tune" the solvating strength of the SF. This is not possible in the liquid state because of the noncompressibility of liquids. It is even now feasible, by adding small quantities of cosolvents to the SF, to custom design a SF for a specific application.

Designed and Programmed ELECSUS