Diffusion coefficients of He, O2, N2, CO2, and CH4 at 300 K in four silicone polymers, namely, poly(dimethylsiloxane) (PDMS), poly(propylmethylsiloxane), poly((trifluoropropyl)methylsiloxane), and poly(phenylmethylsiloxane), have been estimated by molecular dynamics (MD) simulations. The estimated diffusion coefficients decrease with increasing size of the polymer side chains and of the penetrant molecules, as was also found experimentally. The estimated diffusion coefficient for He in PDMS is consistent with its experimental value. The values of the estimated diffusion coefficients for the other gas/silicone systems considered in this study are within ± 40−60% of the corresponding experimental values. The MD simulations revealed two types of motions of the penetrant molecules:  (1) “jumps” from one cavity in a silicone matrix to another, and (2) “oscillating motions” inside cavities. The lengths of the jumps are of the order of 8−15 Å, whereas the oscillating motions are of the order of ≤5 Å. The total timeframe for the execution of a jump is about 5 ps (5 × 10-12 s). The number of jumps and the average length of a jump of a penetrant molecule inside a silicone matrix decrease as the size of the molecule and of the polymer side chain increase. Some problems associated with the estimation of gas diffusion coefficients in polymers by MD simulations are also discussed.

SynopsisPermeability coefficients P for He, and C3H, in 12 different silicone polymer membranes were determined at 35.0"C and pressures up to 9 atm. Values of F for CO,, CH,, and C,H, were also determined a t 10.0 and 55.0"C. In addition, mean diffusion coefficients D and solubility coefficients S were obtained for CO,, CH,, and C3H, in 6 silicone polymers a t 10.0, 35.0, and 55.0"C. Substitution of increasingly bulkier functional groups in the side and backbone chains of silicone polymers results in a significant decrease in P for a given penetrant gas. This is due mainly to a decrease in E, whereas S decreases to a much lesser extent.Backbone substitutions appear to have a somewhat lesser effect in depressing than equivalent side-chain substitutions. The selectivity of a silicone membrane for a gas A relative to a gas €3, i.e., the permeability ratio &A)/&€%), may increase or decrease as a result of such substitutions, but only if the substituted groups are sufficiently bulky. The selectivity of the more highly permeable silicone membranes is controlled by the ratio S(A)/S(H). whereas the selectivity of the less permeable membranes depends on both the ratios D(A)/D(B) and S(A)/S(R). The permeability as well as the selectivity of one silicone membrane toward CO, were significantly enhanced by the substitution of a fluorine-containing side group that increased the solubility of CO, in that polymer.

The permeability of nine different polyimide membranes to H2, N2, O2, CH4, and CO2 has been determined at 35°C and at applied pressures of up to 9 atm. The dianhydride monomers used for the synthesis of the polymides were PMDA and 6FDA, whereas the diamine monomers were ODA, BDAF, and p‐PDA. The selectivities of the 6FDA polymides toward CO2 relative to CH4 are higher than those of the PMDA polyimides at comparable CO2 permeabilities. Both types of polyimides exhibit significantly higher CO2/CH4 selectivities than more common glassy polymers, such as cellulose acetate, polysulfone, and polycarbonate. The selectivities of the PMDA and 6FDA polyimides to O2 relative to N2 are of the same magnitude and generally higher than those of common glassy polymers with similar O2 permeabilities. The polymides are more permeable to N2 than to CH4, whereas the opposite is true for many other glassy polymers. Possible factors responsible for the above behavior, such as segmental mobility, mean interchain distance, and formation of charge transfer complexes, are examined. The relevance of the study to the development of more highly gas‐selective and permeable membranes for the separation of gas mixtures is also discussed.

No abstract

No abstract