SUPERCRITICAL CARBON DIOXIDE AS A UNIQUE REACTION MEDIUM
Supercritical fluids are becoming increasingly important in industry partly in response to the
adverse environmental impact of solvent use and disposal. Carbon dioxide has received special
attention as a result of its easily accessible supercritical point (31 °C, 75.8 bar). Supercritical carbon
dioxide (SC-CO2) is a desirable replacement for organic solvents because it is inexpensive, non-toxic,
non-flammable, and exhibits ease of recycling and disposal.1 These properties make it an especially
suitable solvent for large-scale industrial synthesis. In fact, two small-scale plants using SC-CO2, one
owned by Thomas Swan & Co. and the other by DuPont, have been in operation for several years.
Thomas Swan & Co. has very recently completed a large plant for the use of supercritical carbon fluids
in industrial scale synthesis. In addition, DuPont has allocated $40 million for the construction of a
plant for the production of fluoropolymers that is expected to be fully functional by 2006.
At the First International Symposium on Supercritical Fluid in 1998, Dr. Val Krukonis, an expert
in the field and founder of Phasex corporation, stated, “There’s no point in doing something in
supercritical fluid just because it’s neat. Using the fluids must have some real advantage.”2 In
deference to this sound advice, this review will be focused on reactions in which the outcome either
cannot be obtained using traditional organic solvents or is influenced to a great extent by the unique
The first part of this review will focus on how the physical properties of SC-CO2 can affect the
reaction rates and product distributions of several organic reactions. The second part will provide some
examples in which SC-CO2 is used both as a solvent and reactant.
PROPERTIES OF SUPERCRITICAL CO2
The boundaries between the solid, liquid and gas phases are shown in a typical phase diagram
(Figure 1). The supercritical fluid state occurs above the critical temperature when increasing the
pressure no longer causes a phase change to liquid. Unlike the gas, solid, and liquid phases; the
supercritical fluid phase of CO2 near the critical point is inhomogeneous.3 Low and high-density regions
exist in equilibrium throughout the medium. The loss of entropy when a molecule of CO2 moves from a
low density to high-density region is balanced by the increase in favorable intermolecular interactions.
As a consequence, SC-CO2 is easily compressed near its critical point; and many bulk properties such as
solvent density, dielectric constant, and solubility parameter change dramatically with small changes in
pressure. Thus, SC-CO2 is a tunable solvent which can be adjusted to accommodate a wide variety of
Figure 1. Representative phase diagram for CO2
Many studies have reported the effects of solvent heterogeneity in SC-CO2. It is believed that
when solutes are placed in SC-CO2, they are surrounded by the more dense regions of the supercritical
fluid. This effect is called “clustering,” and may result in enhanced solvent cage effects. Recent studies
by Tanko and coworkers have probed the solvent cage effects of both geminate and diffusive caged
The photolytic cleavage of dicumyl ketone 1 leads to the formation of various products (Scheme
1). Hydrogen abstraction within solvent cage 2 yields 3 and 4. By contrast, escape from the solvent
cage 2 gives radical 5. This radical can associate to form the diffusion radical cage 7 which leads to the
disproportionation products of 8 and 3, resulting from hydrogen abstraction or to 9 by dimerization. A
comparison of the quantities of products obtained provides information on the rates of escape and rates
of hydrogen abstraction occuring in geminate solvent cage 2. The ratio of products (8 + 9):4 reveals the
relative rate of escape (kesc) from radical cage 2 versus the rate of hydrogen abstraction within the
solvent cage(kH). In addition, the ratio of kdim/kdisp is directly related to the ratio of products 9/8. The
relative rates of dimerization versus disproportionation are a sensitive probe of local solvent density
because there are more geometric constraints for the dimerization reaction than for the
disproportionation. Since the local density affects the ease of rotation of the solute, kdim/kdisp is expected
to decrease with increasing local solvent density.
The photolysis reaction was performed under various pressures
9 of SC-CO2 to probe the effects of
increasing viscosity and local solvent density on the various rates. Two interesting results were
obtained. Above 1800 psi both kesc/kH and kdim/kdisp diminish with increasing pressure of SC-CO2. This
decrease is anticipated because the increase in viscosity slows the rate of radical escape from the solvent
cage and slows the rotation of 5 within solvent cage 7. However, the ratios are about half of what would
be expected based on viscosity studies in conventional solvents. These results indicate that at pressures
above 107.5 bar, SC-CO2 displays enhanced solvent cage effects compared to organic solvents. In
contrast, kesc/kH and kdim/kdisp increases with increasing pressure below 107.5 bar. In fact, kecs/kh is
about the same at 77.5 bar and 588.3 bar. These results support the theory of a clustering effect where
the local solvent density is greater than that of the bulk fluid. Computer simulations suggest that the
cause of these effects is the favorable interaction of SC-CO2 with aromatic compounds through
polarization of one of the C==O bonds. These results show that SC-CO2 demonstrates special properties
near the critical point which are not found in traditional organic solvents. The effects of clustering and
increased cage effects can explain unexpected rates and product distributions found in some of the
EFFECTS OF SC-CO2 PROPERTIES ON ORGANIC REACTIONS Radical decomposition
Similar to the free radical studies described above, DeSimone and coworkers have extensively
studied the decomposition of free radical initiator 2,2’-azobis(isobutyronitrile) 10 (AIBN) in SC-CO2.5,6
The rate of decomposition of 10 to 11 was found to be slower in SC-CO2 than in traditional organic
solvents. For example, the rate of AIBN decomposition in SC-CO2 is 3.5 x 10-6s-1 and 8.4 x 10-6s-1 in
benzene at ambient temperature. The lower rate of decomposition in SC-CO2 is believed to result from
the lower dielectric constant of this medium compared to organic solvents. This is further supported by
the observation that the addition of small amounts of THF greatly enhances the rate of decomposition
Trapping studies using the radical trap galvinoxyl 14 show that AIBN has a higher efficiency
factor in SC-CO2 (0.83) than in benzene (0.53).5 The efficiency factor is a measure of the fraction of
radicals that propagate through the solution to those that either dimerize to 12 or participate in a
disproportionation reaction (Scheme 2). The higher efficiency factor is attributed to
the low viscosity of SC-CO2. Higher initiation efficiency and tunable rates of O
dissociation of AIBN have important implications in the field of polymer chemistry
in which AIBN is a commonly used free radical initiator.
Polymerizations in SC-CO2
The first homogeneous free radical polymerization in SC-CO2 was reported in 1992 by
DeSimone and coworkers.5 SC-CO2 is an ideal solvent for the polymerizations of fluoropolymers and
silicon-based polymers which display limited solubility in organic solvents. Moreover, although most
commercial polymers are not soluble in SC-CO2, they can be synthesized in biphasic dispersion and
emulsion polymerizations.7,8 The molecular weights and polymer properties obtained in SC-CO2 are
similar to those obtained by analogous polymerization methods in organic solvents. Therefore, the
incentives of using SC-CO2 as a solvent lie not in the polymerization reaction, but in the decreased cost
of polymer processing.9 Polymers synthesized in SC-CO2 can be isolated simply by depressurization of
the reaction vessel. The CO2 is easily collected and recycled, eliminating the cost of solvent disposal.
In addition, the costly and energy intensive drying procedure, typical in polymer manufacturing using
traditional solvents is greatly reduced.7 Moreover, due to the increased plasticity of polymers in SC-
CO2, residual monomer and catalysts are easily removed from the polymer matrix.9
SC-CO2 can also be used to incorporate monomers for the generation of polymer blends or other
small molecules for polymer modification.10,11,12 McCarthy and coworkers used the SC-CO2 as a carrier
of maleic anhydride to functionalize linear low density polyethylene (LLDPE) and poly(4-methyl-1-
pentene) (PMP) (Scheme 3). Maleation to the extent of 2.96-3.52 % by weight in LLDPE and 2.06-2.62
% by weight in PMP was reported. This is a significantly higher level of functionalization than is
available in commercial maleated polymers which have nearly undetectable levels of maleation.
Tuning the properties of SC-CO2 through changes in pressure can also affect the outcome of
olefin metathesis.13 At 130 bar, 8% ring closing metathesis product 16 was obtained from 15 and 68%
of the product was oligomer (Scheme 4). However, increasing the pressure to 200 bar increased the
percent of product 16 to 87% and decreased oligomer formation to about 2%. Isothermal increases in
pressure increase the density of SC-CO2, favoring improved yields of the intramolecular ring closing
metathesis product 16 versus intermolecular formation of oligomers. This observation at first seems
counterintuitive and is the opposite to the behavior observed in traditional organic solvents. Usually
increased pressure favors intermolecular reaction in order to decrease the number of moles in solution.
But within a certain range, increasing the pressure of SC-CO2 at a constant volume forces more
molecules of CO2 between solute molecules thus, mimicking the effects of increased dilution in
SC-CO2 AS A REACTANT
In addition to acting as a solvent with unique physical properties, SC-CO2 proves to have
synthetic utility a variety of reactions. Furstner and coworkers report the use of SC-CO2 as a labile
protecting group for secondary amines whereas Noyori et al. use SC-CO2 as a C1 building block.13,14,15
CO2 as a protecting group
The volatility of carbon dioxide allows it to act as a temporary in situ protecting group for
secondary amines. Recent reports by Furstner et al. show the metathesis of 17 to give 18 in 74% yield
in SC-CO2 without protection of the secondary amine (Scheme 5).13 This is a convenient solution to one
of the few limitations of the metathesis reaction; secondary amines poison ruthenium metathesis
catalysts in traditional organic solvents.16 In SC-CO2 protection is not necessary because the carbon
dioxide itself acts as a temporary protecting group for the amine. When the reaction vessel is vented, the
carbamic acid 20 reverts spontaneously to the amine 19 without the need for an additional deprotection SC-CO2 as a carbon source
Noyori and coworkers use SC-CO2 as a carbon source in the formation of formic acid, methyl
formate, and dimethyl formamide.14,15 Noyori et al. report the homogeneous hydrogenation of SC-CO2
to give formic acid (Equation 1). Additives such as water, methanol and DMSO all accelerate the rate of
reaction as long as only one phase is observed in the reaction vessel. Yields, product distributions, and
rates of reaction are all sensitive to temperature, the addition of co-solvents, and H2 pressure. The yields
and rates observed are also very sensitive to the phase behavior of SC-CO2.
Under optimized conditions, formic acid was produced at turnover frequencies exceeding 4000
h-1 (turnover frequency = moles product/moles catalyst per hour) (Equation 2). Methyl formate was
produced with turnover number (TON) of 3500. The TON, which is a measure of moles of product/mol
catalyst, is one order of magnitude greater than any pervious reported result at any temperature. DMF
was produced at a turnover frequency of 8000 h-1 and a TON of 420 000. This result is two orders of
magnitude greater than any previously published TON.
The high degree of solubility of H2 gas in SC-CO2 is one factor contributing to the improved
results of these reactions in this medium. Only in super critical fluids can the catalyst and all of the
reagents be dissolved in the same phase. A weaker coordination sphere surrounding the catalyst may
also contribute to an increase in the reaction rate. Furthermore, it is believed that the lifetime of the
catalyst is longer in SC-CO2 than in organic solvents. The increased rate paired with the increased
lifetime of the catalyst lead to significantly higher product yields.
Many unique properties of SC-CO2 make it a useful solvent for a wide variety of reactions.
Unlike traditional organic solvents, SC-CO2 is a tunable solvent with the density, dielectric constant, and
viscosity dependant on pressure. The utility of SC-CO2 as a reactant has also been demonstrated in a
Brennecke, J. F.; Chateauneuf, J. E. Chem. Rev.1999, 99, 433.
Darr, J. A.; Poliafoff, M. Chem. Rev.1999, 99, 495.
Tucker, S.C. Chem. Rev., 1999, 99, 391.
Tanko, M. J.; Pacut, R. J. Am. Chem. Soc.2001, 123, 5703.
DeSimone, J. M.; Guan, Z.; Elsbernd, C. S. Science, 1992, 257, 945.
Guan, Z.; Combes, J. R.; Menceloglu, Y. Z.; DeSimone, J. M. Macromolecules, 1993, 26, 2663.
Kendall, J. L.; Canelas, D. A.; young, J., L.; DeSimone, J. M. Chem. Rev.1999, 99, 543.
Herk, A. M; Manders, B. G. Macromolecules, 1997, 30, 4780.
Krukonis, V. Polymer News, 1985, 11, 7.
Watkins, J. J.; McCarthy, T. J. Macromolecules, 1995, 28, 4067.
Watkins, J. J.; McCarthy, T. J. Macromolecules, 1994, 27, 4845.
Hayes, H. J.; McCarthy, T. J. Polym. Prep.1999, 426.
Fürstner, A.; Ackermann, L.; Beck, K.; Hori, H.; Koch, D.; Langemann, K.; Liebel, M.; Six, C.;Leitmer, W. J. Am. Chem. Soc.2001, 123, 9000.
Jessop, P. G.; Ikariya, T. Noyori, R. Nature, 1994, 368, 231.
Jessop, P. G.; Hsiao, Y.; Ikariya, T.; Noyori, R.; J.Am. Chem. Soc.1996, 118, 344.
Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res.2001, 34, 18.
Cellular and Molecular Neurobiology [cemn] Cellular and Molecular Neurobiology, Vol. 23, No. 2, April 2003 ( C 2003) Rapid Communication A Novel Method of Eliminating Non-Neuronal Proliferating Cells From Cultures of Mouse Dorsal Root Ganglia Parker L. Andersen,1 J. Ronald Doucette,2 and Adil J. Nazarali1,3 Received November 5, 2002; accepted December 18, 2002 1. We hypothes
Persons using assistive technology may not be able to fully access information in this file. For assistance, e-mail [email protected] Include the Web site and filename in your message. 4. SCREENING BASELINE VISIT 4.1 INTRODUCTION Consistent with the philosophy and design of a large study, screening activities performed during the Baseline Visits have been kept streamlined. Screening a