The present paper deals with important differences in the distribution of free volume in high and low free volume polymers as indicated by a joint investigation utilizing molecular modeling and positron annihilation lifetime studies. The main focus of this paper is on the molecular modeling approach. The polymers in question are the ultrahigh free volume polymer poly(1-(trimethylsilyl)-1propyne) (PTMSP) and two polystyrene derivatives containing Si and F. Extended equilibration procedures were necessary to obtain reasonable packing models for the polymers. The transition state Gusev-Suter Monte Carlo method was utilized to prove a reasonable agreement between simulated and measured diffusivity and solubility values for the model structures. The free volume distribution was analyzed for the validated packing models and compared with respective positron annihilation data. In both cases, a bimodal distribution of free volume was observed for PTMSP while the polystyrene derivatives as other conventional glassy polymers showed a more or less monomodal behavior. Good qualitative agreement is demonstrated between size distributions of free volume elements in these polymers obtained via molecular computer modeling and experiments using positron annihilation technique.

The present paper deals with a molecular modeling-based characterization of the distribution of free volume in a number of different stiff chain glassy polymers including ultrahigh free volume and conventional materials. The free volume distribution was analyzed for the validated packing models and compared with respective positron annihilation data (if available). It will be shown that the molecular modeling approach permits a more detailed insight into free volume distributions. There the observed distributions reach from more or less symmetric monomodal (with maximum probability at free volume element radii of about 3 Å) via monomodal distributions with extended tails toward larger radii up to distinctly bimodal distributions with a tendency toward a continuous nanoporous free volume phase besides "conventional", free volume organized in isolated holes of radii up to a few angstroms. The described sequence roughly corresponds to the trend of measured permeabilities for small molecules in the investigated materials.

No abstract

Well-equilibrated molecular-packing models have been produced for 10 different polyimides. The Gusev-Suter transition-state theory was used to calculate gas solubilities and diffusion coefficients for nitrogen, oxygen, methane, and carbon dioxide. Good agreement with experiment (factors 1-4) was found, except for CO 2. The difficulties in this comparison were discussed. A significant improvement from former results could be assessed for the predicted O2/N2 selectivity values. The simulated models allow an accurate determination of structural parameters, either as a single parameter, like the fractional free volume, or as size-distribution function of free-volume elements accessible for a certain penetrant. The 2,2′-bis(3,4-dicarboxy-phenyl) hexafluoropropane polyimides with the highest oxygen permeability (50-130 Barrer) show a wider size distribution with an additional peak or shoulder at larger radii (>5-6 Å) than conventional polyimides. A constitutive structural element seems to be the o-methyl groups in the aromatic diamine moiety.

The aim of this work is to predict the adsorption of pure-component and binary mixtures of methane and carbon dioxide in a specific activated carbon, A35/4, using grand canonical Monte Carlo (GCMC) simulation. Methane is modeled as a one-center Lennard-Jones (LJ) fluid and carbon dioxide as a twocenter LJ plus point quadrupole fluid. Experimental adsorption data for the system have been obtained with a new flow desorption apparatus. The pore size distribution (PSD) for the carbon was determined from both of the experimental CH4 and CO2 isotherms at 293 K. To extract numerically the PSD, GCMCsimulated isotherms for both pure components in slit-shaped pores ranging from 5.7 to 72.2 Å were used. Using only pure experimental CO2 isotherm data, it was not possible to determine a PSD that allowed a reasonable prediction of the pure methane adsorption. However, with both experimental data sets for the pure components, it was possible to derive a PSD that allowed both experimental pure-component isotherms to be fitted. With this PSD and the simulated adsorption densities in single pores, it was possible to predict in good agreement with experiment (i) the adsorption of binary mixtures of CO2 and CH4 and (ii) the adsorption of both pure components at higher temperatures. However, the model was unable to reproduce precisely the experimental pressure dependence of the CO2 selectivity.

Grand canonical Monte Carlo (GCMC) simulations of binary Lennard-Jones mixtures in the zeolite silicalite have been used to predict the adsorption of CH4 and CF4 mixtures as a function of gas phase composition, total pressure, and temperature. For single components and mixtures, predictions of adsorption isotherms and isosteric heats are in good agreement with experiment at room temperature. Within the experimental pressure range of 0 to 17 bar, the mixtures are well described by the ideal adsorbed solution (IAS) theory. For very high loading, deviations from IAS theory appear. The configurations generated in the simulation were used to calculate sorbate−zeolite interaction energy distributions for different types of siting locations within the zeolite pores. These distributions display a pore shape related energetic heterogeneity in different regions of silicalite. Near saturation at a total loading of 12 molecules per unit cell, the shape of the observed energy distribution is relatively independent of the composition in the pore. Nevertheless, the energetic heterogeneity is responsible for a mild segregation in the adsorbed mixtures, with methane adsorbed preferentially in the silicalite zigzag channels and CF4 preferentially in the straight channels.

No abstract

Adsorption equilibria of argon, nitrogen, and methane on the 13X molecular sieve and the AS activated carbon were measured at five temperatures over a wide pressure range from 0.1 to 20 MPa using a microbalance. From experimental adsorption isotherms, which are excess functions, the absolute isotherms were calculated using equations of state for a real gas phase. Grand canonical Monte Carlo simulations and density functional theory calculations were carried out in order to explain specific features of the resulting isotherms. Different thermodynamic functions evaluated from the excess and absolute adsorption isotherms were analyzed over a wide pressure range at the average temperature of the range studied.