Properties of the supercritical fluid

Although the first step in optimizing an SFE method should be selecting the extractant, in practice, CO, is almost invariably the choice as alternative extractants are only occasionally employed. 

Experimentally, the performance of an SFE process relies on two essential properties of the supercritical fluid, namely its density, which is dictated by its pressure and temperature and its chemical nature (polarity). As shown below, however, other factors such as the presence of a chemical modifier and the volume of extracting fluid are also influential. 

The solvent power of a supercritical fluid is a function of its density. No other separation technique allows it to be altered in such a simple manner as by changing the physical conditions. By way of example. Fig. 7.5A shows the density-pressure isotherms for carbon dioxide at a variable temperature and Fig. 7.5B illustrates the influence of the CO, density on the extraction efficiency for rra 5-/3-carotene when all experimental variables except pressure are kept constant [24]. As can be seen, the recovery changed virtually linearly with the fluid density throughout the studied range. The effect is similar with other types of analytes such as fatty acids in red seaweed [25]. 

At a constant temperature, the extraction of non-polar analytes is usually favoured by a low pressure, whereas that of polar analytes requires an increased one. This behaviour is used for class-selective extractions, an example of which is the sequential extraction of air particulates with supercritical CO,. Alkanes are extracted at 75 bar (45°C), whereas PAHs remain unextracted until the pressure is raised to 300 bar [26]. In some cases, an increase in pressure increases not only the solubility of the analyte but also its difiusivity. Such is the case with polymers, which absorb large amounts of COj under supercritical conditions, thereby swelling and facilitating diffusion of the solute. 

Flowever influential it may be, the solubility of the analyte in the supercritical fluid is not the sole variable to be considered. In fact, it is necessary to overcome matrix-analyte interactions, which are very strong in some samples. The fluid polarity is a key factor under these conditions as diffusion in the solid matrix does not seem to be the extraction rate-determining step otherwise, there should be no change in the recovery rate on 

---

The first use of supercritical fluid extraction (SFE) as an extraction technique was reported by Zosel [379]. Since then there have been many reports on the use of SFE to extract PCBs, phenols, PAHs, and other organic compounds from particulate matter, soils and sediments [362, 363, 380-389]. The attraction of SFE as an extraction technique is directly related to the unique properties of the supercritical fluid [390]. Supercritical fluids, which have been used, have low viscosities, high diffusion coefficients, and low flammabilities, which are all clearly superior to the organic solvents normally used. Carbon dioxide (C02, [362,363]) is the most common supercritical fluid used for SFE, since it is inexpensive and has a low critical temperature (31.3 °C) and pressure (72.2 bar). Other less commonly used fluids include nitrous oxide (N20), ammonia, fluoro-form, methane, pentane, methanol, ethanol, sulfur hexafluoride (SF6), and dichlorofluoromethane [362, 363, 391]. Most of these fluids are clearly less attractive as solvents in terms of toxicity or as environmentally benign chemicals. Commercial SFE systems are available, but some workers have also made inexpensive modular systems [390]. 

The physical properties of the supercritical fluid differ from those of the bulk liquid. One of the most notable changes is the lower dielectric constant of polar solvents such as water which allows the accumulation of low-polarity solutes at this interface. This explains the crucial role of the hydro-phobicity of solutes during reactions in the solution. Thermolysis as well as radical abstraction reactions occur in this region. A temperature of approximately 800 K was determined for the interfacial region surrounding the... 

A number of other important potential applications of a micellar phase in supercritical fluids may utilize the unique properties of the supercritical fluid phase. For instance, polar catalyst or enzymes could be molecularly dispersed in a nonpolar gas phase via micelles, opening a new class of gas phase reactions. Because diffusivities of reactants or products are high in the supercritical fluid continuous phase, high transport rates to and from active sites in the catalyst-containing micelle may increase reaction rates for those reactions which are diffusion limited. 

Unfortunately, most polymers are insoluble in supercritical C02, and hence extraction from the ionic liquid by this method is difficult. However, if C02-soluble polymers were synthesized (for example, fluoropolymers and polysi-loxanes), then this method has the potential to be a very useful approach. Moreover, supercritical fluid-swollen ionic liquids offer a new solvent system that combines the viscosity-lowering properties of the supercritical fluid with the good solubilizing properties of the ionic liquid and may be a hybrid exotic solvent of the future. 

Since the discovery of microemulsion phases in supercritical fluids in the mid-1980s [1] and their subsequent characterization [2-16], there has been much interest in exploiting the unusual properties of the supercritical fluid phase in applications of these systems. One such application is as a new type of solvent for chemical reactions. In the following sections, I discuss the properties of these systems for reactions, review the progress so far, and analyze the future potential. As a prelude to these discussions, I begin with a brief overview of what is known about the molecular structure of microemulsions in near-critical and supercritical fluids. The details of the primary and secondary molecular structures of various types of microemulsion phases can dramatically affect the reactivity in these systems. 

The goal of these modeling studies was to get a better understanding of these reactions and how they are influenced by the properties of the supercritical fluid. In view of the technical application it is an advantage to have such a model. 

Specific physicochemical properties of the supercritical fluids offer flexible alternatives to established processes like chemical vapor deposition (CVD), which is used in the preparation of high-quality metal and semiconductor thin films on solid surfaces. Watkins et al. [43] reported a method named chemical fluid deposition (CFD) for the deposition of CVD-quality platinum metal films on silicon wafers and polymer substrates. The process proceeds through hydrogenolysis of dimethyl-(cyclooctadiene)platinum(ll) at 353 K and 155 bar. 

The use of supercritical fluid extraction (SEE) as an extraction technique is related to the unique properties of the supercritical fluid. These fluids have a low viscosity, high diffusion coefficients, low toxicity, and low flammability, all clearly superior to the organic solvents used in SPE extraction. The most common fluid used is carbon dioxide. SEE extractions of sediment samples have shown recoveries of >95% for all the individual PCBs. The separation of PCDDs from PCBs and chlorinated benzenes is difficult because of their similar solubility. An interesting development is the use of fat retainers. Samples, mixed in different weight ratios with, e.g., silica/silver nitrate 10% or basic alumina, can be placed in 7 ml extraction cells. The analytes are recovered by elution with 1.5-1.8 ml of hexane. With the correct fat-silica ratios and SEE conditions, no additional cleanup procedure is necessary for GC with an electron-capture detector (ECD). One drawback of SEE may be that the methods developed are valid for a specific matrix, but as soon as, e.g., the fat content of a biota sample or the type of lipids changes, the method has to be adapted. SEE is relatively complicated compared to other extraction techniques. In addition, the cell volumes are small, which limits the sample intake, and, with that, the detection limits. Einally, some reliable types of SEE equipment have recently been withdrawn from the market. This will have a substantial negative effect on the use of SEE in the near future. 

The transport properties of the supercritical fluids fall somewhat in between the gas and the liquid and also depend on how removed one is from the critical point. Dense gasses have the solubilizing power of liquids and the mobility of gasses as depicted in Table 20.1.3. There are quite a few empirical correlations and theoretical models, which are primarily extensions of corresponding low-pressure liquid and gas counter parts. Similarly, some of the classical experimental methods can be used for measurement of transport properties of supercritical fluids. A rather brief overview of the methods applicable for supercritical fluids will be presented since specialized reviews in the area give a good account of the state of the art. " " For engineering purposes, one can use applicable property estimation methods available in flowsheet simulators such as ASPEN PLUS, PROll, G-PROMS, and CHEMCAD. These methods are discussed in a text classical in the field." ... 

The physico-chemical properties of the supercritical fluid being strongly dependent on pressure and temperature, means that these two parameters must be closely controlled all along the process, which must include 6 closed-loop regulation systems. As can be seen on the eluent cycle (Fig. 7-1), the PS-SFC process requires 3 or 4 different temperature levels and as many heating or cooling systems which must be added to the investment cost. 

Therefore, for the higher-pressure case we may adopt a numerical analysis, based on iterative integration around the loop of the momentum equation (since mass is also conserved) for varying loop power inputs, using the thermophysical properties of the supercritical fluid as a function of actual thermodynamic state. Thus the general flow variation with major loop parameters (elevations, losses etc.) follows Equation (4) but with a non-linear expansion coefficient. 

---