Effect of Grafted Lewis Base Groups on the Phase Behavior of Model Poly(dimethyl siloxanes) in CO 2 S. Kilic, S. Michalik, Y. Wang, J. K. Johnson, R. M. Enick, and E. J. Beckman* Department of Chemical Engineering, University of Pittsburgh, Pittsburgh, Pennsylvania 15261 The impact of various Lewis bases on the miscibility of siloxane polymers in CO 2 was investigated using both ab initio calculations and experimental phase behavior studies. A series of side- chain functional silicones were synthesized containing various Lewis bases in the side chain, and their phase behavior was compared in CO 2 at 295 K. Calculations showed that interactions between CO 2 and ethers (either a dialkyl ether or the ether oxygen in an ester group) should be as favorable as interactions between CO 2 and a carbonyl oxygen. Indeed, phase behavior results seemed to support this, as ether-functional silicones exhibited miscibility pressures as low or lower than acetate-functional analogues. Further, a keto-functional material was not nearly as CO 2 -philic as the acetate functional analogue. In general, the location of the phase boundary in CO 2 is governed by a balance between forces working to increase miscibility pressures, such as increased cohesive energy density of the polymer or factors suppressing the entropy of mixing, and those working to lower miscibility pressures, such as enhanced specific interactions with CO 2 and increased free volume or chain flexibility. Introduction The possibility for the use of carbon dioxide as a process solvent has been widely investigated because CO 2 is an environmentally benign, inexpensive, and abundant material. Solubility parameter studies using equation of state data once suggested that CO 2 pos- sesses the solvent power of short n-alkanes, 1 and it was hoped that CO 2 could be used to replace an array of environmentally unfriendly nonpolar organic solvents. Although CO 2 initially looked to be useful only for nonpolar materials, it was thought that polar materials could be brought into solution by adding conventional alkyl-functional surfactants to the mixture. However, early attempts to put these surfactants to use were hindered due to the poor solubility of the amphiphiles in CO 2 . The fact that these amphiphiles showed ad- equate solubility in short alkanes such as ethane and propane and were quite insoluble in CO 2 2 revealed a gap between theoretical models and experimental data for CO 2 solubility. Johnston and colleagues suggested polarizability/free volume as a better method of evaluat- ing solvent power, 3,4 and by this method CO 2 is seen to be a very poor solvent when compared to short n- alkanes. The solvent quality of CO 2 has also been investigated experimentally. Francis tested the phase behavior of more than 250 compounds in ternary systems contain- ing liquid CO 2 . 5 Hyatt presented an extensive study of phase behavior of more than 30 organic compounds in liquid CO 2 to attempt to draw comparisons between CO 2 and organic solvents. 6 Phase behavior studies showed that CO 2 is a reasonably good solvent for aldehydes, ketones, esters, and low alcohols, but higher alcohols (C > 10), aromatic alcohols, and polar compounds such as amides, ureas, and urethanes exhibit poor solubility in CO 2 . Hydroquinone and multihydroxy compounds were found to be insoluble in the aforementioned study. Heller and co-workers evaluated, for the first time, the miscibility of commercially available polymers with CO 2 in an attempt to find a polymer to control the mobility of CO 2 during “enhanced oil recovery” (EOR) opera- tions. 7 They reported that tacticity plays an important role in determining the miscibility of a polymer in CO 2 . For example, they found that although atactic poly- butene and poly(propylene oxide) are miscible with CO 2 , isotactic polymers are not. It was also found that the presence of aliphatic side chains reduces the miscibility pressures significantly, but on the other hand, the presence of aromatic groups in a polymer raises misci- bility pressures drastically. They also reported that the presence of amide, carbonate, ester, and hydroxyl groups in the polymer backbone imparts immiscibility to a polymer with CO 2 , while ester and ether groups in the side chain do not have a detrimental effect on miscibil- ity. A number of groups continued the search for materi- als that would be soluble in CO 2 at significantly lower pressures than similarly sized alkyl-functional equiva- lents, and it was found that some fluorinated materials were miscible with CO 2 at relatively low pressures. 8-12 Harrison et al. synthesized a hybrid alkyl/fluoroalkyl surfactant that dissolved in CO 2 and solubilized a significant amount of water. 13 Through fluorination, even CO 2 -insoluble hydrocarbon polymers could be rendered miscible with CO 2 . 14,15 In addition, dispersion polymerization of methyl methacrylate in CO 2 was supported by block polymers containing fluorinated acrylate monomers, 16 leading to the generation of mono- disperse, micrometer-sized spheres. Other developments using fluoro-functional amphiphiles followed, including emulsion polymerization, 17 protein extraction, 18,19 and heavy-metal extraction from soil and water. 20 Without question, perfluoropolyacrylates are the most CO 2 -philic polymers discovered to date. Their thermo- dynamic compatibility with CO 2 might be attributed to their low cohesive energy density and relatively low glass transition temperature. Kirby and McHugh have * To whom correspondence should be addressed. Tel.: (412) 624-9630. Fax: (412) 624-9639. E-mail: beckman@ engrng.pitt.edu. 6415 Ind. Eng. Chem. Res. 2003, 42, 6415-6424 10.1021/ie030288b CCC: $25.00 © 2003 American Chemical Society Published on Web 09/27/2003