Inert gas permeation through homopolymer membranes

New data are reported for the permeation of inert gases through polyethylene, polytetrafluoroethylene, and silicone and natural rubbers. Additional data are compiled from the literature. The relative solubilities of these gases are practically insensitive to chemical variations in the homopolymer. Hence variations in structure at the glass transition (T_g) and melting (T_m) temperatures that affect diffusion also unambiguously affect permcation. Consequently an equivalence results between permeation at a given temperature for different polymers and permeation at different temperatures for a given polymer. Although the diffusion coefficient changes continuously with temperature, the Arrhenius parameters D_o and E_d apparently change discontinuously at T_g and T_m. Their magnitudes and variations with atomic weight reach maxima at about T_g. These data indicate a dependence of the classical correlation between D_o and E_d on polymer properties. A perturbed diameter for the permeant, specific for each polymer, is proposed for correlating the D_o and E_d data. This correlation makes the changes observed at T_g and T_m more perceptible.     

Gas permeation through water-swollen hydrogel membranes

This work deals with water-swollen hydrogel membranes for potential CO₂ separation applications, with an emphasis on elucidating the role of water in the membrane for gas permeation. A series of hydrogel membranes with a wide range of water contents (0.9–10 g water/g polymer) were prepared from poly(vinyl alcohol), chitosan, carboxyl methyl cellulose, alginic acid and poly(vinylamine), and the permeation of CO₂, H₂, He and N₂ through the membranes at different pressures (200–800 kPa) was studied. The gas permeabilities through the dry dense membranes were measured as well to evaluate the resistance of the polymer matrix in the hydrogel membranes. It was shown that the gas permeability in water-swollen membrane is lower than the gas permeability in water, and the selectivity of the water-swollen membranes to a pair of gases is close to the ratios of their permeabilities in water. The permeability of the water-swollen membranes increases with an increase in the swelling degree of the membrane, and the membrane permeability tends to level off when the water content is sufficiently high. A resistance model was proposed to describe gas permeation through the hydrogel membranes, where the immobilized water retained in the polymer matrix was considered to form transport passageways for gas permeation through the membrane. It was shown that the permeability of hydrogel membranes was primarily determined by the water content in the membrane. The model predictions were consistent with the experimental data for various hydrogel membranes with a wide range of water contents (0.4–10 g water/g polymer).     

Gas permeation through hydrogels : I. Gel cellophane membranes

Permeabilities and diffusion coefficients of various gases, Ar, N₂, O₂, CO₂, CH₄, C₂H₄, C₃H₈ and C₄H₁₀, were measured for Water-swollen gel cellophane membranes. No dependence of permeabilities on gas pressure below 1 atm was found. It was observed that the permeability coefficients were not related linearly to the coefficients in bulk water. For the two states of water in the membrane, an analytical method is presented to estimate the diffusion coefficients and the solubilities in free water and non-freezing water. It was found that the diffusion coefficients in non-freezing water were lower than those in free water, and the solubilities in non-freezing water were higher than in free water for all gases studied except CO₂ and C₂H₄, which gave reverse results.     

Gas permeation through PSF-PI miscible blend membranes

The permeation rates of He, H₂, CO₂, N₂ and O₂, are reported for a series of miscible polysulfone-polyimide (PSF-PI) blend membranes synthesized in our laboratory. For gases which do not interact with the polymer matrix (such as He, H₂, N₂ and O₂), gas permeabilities in the miscible blends vary monotonically between those of the pure polymers and can be described by simple mixture equations. In the case of CO₂, which interacts with PI, blend permeabilities decrease somewhat, compared to pure PSF and PI. This, however, is accompanied by a two-fold improvement in the critical pressures of plasticization vs. polyimide. Permselectivities of CO₂/N₂ and H₂/CO₂ in the blends deviate from mixing theory predictions, in contrast to selectivities of gas pairs which do not interact with PI. Differential scanning calorimetry measurements of pure and PSF/PI blend membranes show one unique glass transition temperature, supporting the miscible character of the PSF/PI mixture. Optical micrographs of the blend membranes clearly indicate perfect homogenization and no phase separation. Frequency shifts and absorption intensity changes in the FTIR spectra of the blends, as compared with those of the pure polymers, indicate mixing at the molecular level. This compatibility in mixing PSF and PI, results essentially in a new blend polymer material, suitable for the preparation of gas separation membranes. Such membranes combine satisfactory gas permeation properties, reduced cost, advanced resistance to harsh chemical and temperature environments, and improved tolerance to plasticizing gases.     

Studies with inorganic precipitate membranes. III. Consideration of energetics of electrolyte permeation through membranes

The permeability of various electrolytes through parchment-supported ferrocyanide membranes of manganese, cobalt, silver, and cadmium has been measured at 10, 15, 20, 25, and 30°C. The order of permeability at a given temperature was Cl^- > NO₃^- > CNS^- > CH₃COO^- > SO₄^2- for both monovalent and divalent cations. For any given anion, the cations followed the sequence NH₄⁺ > Li⁺ > Ba²⁺ > Ca²⁺ > Mg²⁺ > Al³⁺. This sequence has been correlated with the size of the hydrated ion. Further, the data have been considered from the standpoint of the theory of rate processes and the values for the entropy of activation (ΔS′) have been derived assuming an equilibrium distance of 3 Å in the membrane. The values of ΔS′ were all negative and decreased with increasing valence of the ions. This was interpreted to mean electrolyte permeation with partial immobilization in the membrane.     

Transient permeation of organic vapors through elastomeric membranes

The permeation of benzene and acetone vapors through sulfur-cured natural rubber was studied by the time-lag method. The experimental results were analyzed by a method suggested by Meares. The zero concentration diffusion coefficient D₀ was obtained by the early-time method. The Frisch time-lag equation was utilized to estimate both the solubility coefficient s and the additional parameter b required to define the concentration dependence of the diffusion coefficient: D(c) = D₀ exp {bc}. This form of concentration dependence was manifested by the corresponding permeability coefficient values. At low entering penetrant pressure, where the transport coefficients are constant, indirect evidence was obtained that D₀ is the mechanistically correct diffusion coefficient. The solubility coefficient values calculated for benzene vapor in natural rubber are in reasonable agreement with published equilibrium sorption data for a similar rubber compound. At higher entering penetrant pressures, average diffusion coefficients obtained at steady state tended to be larger than the corresponding average diffusion coefficients derived from the time lags. This same effect has been detected by other experimental approaches. Permeation experiments designed for this rapid method of analysis appear capable of yielding information consistent with that obtained by more time-consuming traditional methods.     

Surface processes in hydrogen permeation through metal membranes

The phenomenon of hydrogen permeation through metals is considered as a complex of interphase and diffusion processes. Basic principles of application of the hydrogen permeation technique to studies of various stages of hydrogen interaction with metals are formulated. An analysis is made of possible errors of the permeation technique when used to measure the characteristics of hydrogen diffusion through metals originating from the finite rate of the interphase processes. The permeation of nonequilibrium and corrosion-produced hydrogen through metal membranes, and the hydrogen permeability of multilayer systems are considered.     

The permeation of binary gas mixtures through support structures of composite membranes

The dusty gas model (DGM) is used to describe transport of binary gas mixtures through porous membrane supports to quantify the resistance towards permeation. The model equations account for three different transport mechanisms for the permeating components: conventional viscous pore flow, Knudsen diffusion, and binary diffusion. Experimental data obtained with the uncoated membrane supports are used to determine the morphological parameters needed in the DGM equations. Flat sheet and hollow fiber membrane supports are characterized by the permeation of a TCE/nitrogen vapor. The DGM shows an excellent fit to experimental data when the asymmetric structure of the membrane supports is taken into account, but the morphological parameters cannot necessarily be related to precise physical structure parameters such as pore size, porosity, and tortuosity. The DGM works well even when the membrane supports are modeled as a single homogenous structure. The membrane supports exhibit different resistances towards the various transport mechanisms that occur within the porous support and the resistances vary with process conditions so that support optimization is not straightforward. With the analysis presented in this paper and transport equations specific to the dense coating and module geometries, the influence of the support layer on gas or vapor separation can be quantified.     

Modeling permeation through anisotropic zeolite membranes with nanoscopic defects

We have modeled permeation through anisotropic zeolite membranes with nanoscopic defects that create shortcuts perpendicular to the transmembrane direction (x). We have found that the dimensionless ratio D_y/(k_dΔy) can be used to estimate whether the shortcuts contribute significantly to the overall flux. Here D_y is the diffusion coefficient for motion in the plane of the membrane, k_d is the rate of desorbing into defect voids, and Δy is the spacing between adjacent defects. For values of D_y/(k_dΔy)⪢1, we find that shortcuts increase the flux by significant amounts. The magnitude of the flux is increased as the imperfection spacing Δy is decreased. For small values of Δy, permeation through shortcuts becomes sorption-limited so that decreasing Δy further does not increase the flux through a single shortcut. However, as Δy is decreased, the concentration of shortcuts increases, thereby increasing the total contribution of the shortcuts to the flux. We have found regimes where increasing Δy or decreasing D_y decreases the overall flux, showing that permeation can be diffusion-limited by motion perpendicular to the transmembrane direction.     

Gas permeation through hydrogels : II. Poly(vinylalcohol-co-itaconic acid) membranes

Permeability and diffusion coefficients of O₂, He, CO₂ and C₄H₆ were measured in water,swollen poly(vinylalcohol-co-itaconic acid) membranes having various water contents from 0.48 to 0.83. The permeability coefficients of CO₂ and C₄H₆ were found to depend on the upstream pressure, while the permeability coefficients of O₂ and He were independent of the pressure. With decreasing pressure the permeability coefficients of CO₂ and C₄H₆ increased, and the pressure dependence became larger with decreasing water content of the membranes. A parallel permeation model based on the two states of water in the water-swollen membranes could be applied successfully to CO₂ and C₄H₆.     

Water and methanol permeation through short-side-chain perfluorosulphonic acid ionomeric membranes

The water and methanol transport into a short-side-chain perfluorosulphonic acid ionomeric (PFSI) membrane suitable for application in proton exchange membrane fuel cells (PEMFC), namely Hyflon^® Ion, was studied between 35 and 65 °C. In particular, the permeabilities of pure water, pure methanol and their mixtures at different temperatures were measured through pervaporation experiments, at various values of feed composition. Due to the presence of mutual interactions between permeants as well as among penetrants and polymeric matrix, the composition of the feed solution affects the membrane permeability in a way which cannot be predicted on the basis of permeability data of the pure liquid components alone. It has been found in particular that the presence of the water in the mixture enhances the methanol permeability, due to the positive effects of matrix plasticization and favourable energetic interactions. In turn, by considering water permeability data in the presence of a poorly permeating component such as glycol, it can be concluded that also water permeation is enhanced by the presence of methanol, although to a lower extent.     

Water—glycerol permeation through styrene-grafted and sulphonated PTFE membranes

The permeability and selectivity of styrene-grafted and sulphonated PTFE membranes towards binary mixtures of water and glycerol were studied.The swelling of the membrane shows maxima at an intermediate concentration of the mixture and increases with increasing sulphonation. The permeating flow, ø, increases with the amount of sulphonic groups in the membrane, while the selectivity, α, only changes slightly with sulphonation. Both ø and α decrease with increasing glycerol concentration in the feed mixture. A strong increase of both α and ø values with temperature was observed, mainly due to increased diffusion of water through the membrane.     

Induction time in metal permeation processes through supported liquid membranes

The origin of the induction time observed in the permeation process of cadmium species through trilaurylammonium chloride in triethylbenzene supported liquid membranes is discussed. A model for the non-steady state transference process, where aqueous film diffusion coupled to an interfacial chemical reaction are the main rate determining processes, was developed. By comparison with experimental transference data, the rate constant of the interfacial reaction between cadmium chloride aqueous complexes and the membrane carrier, trilaurylammonium chloride, was evaluated. The time evolution of the concentration profiles through the aqueous diffusion film is also described.     

Temperature dependence of H permeation through Pd and Pd alloy membranes

H permeabilities (normalized fluxes), have been measured through Pd and some Pd alloy membranes at a series of constant upstream H(2) pressures with the downstream pressure being ~0 in the temperature range from 393 to 573 K. From these data, activation energies for H permeation, E(P), have been determined. Conditions of constant upstream p(H(2)) are of most interest since most determinations of E(P) in the literature have employed this boundary condition. Permeabilities have also been measured at a series of constant upstream H concentrations with the downstream concentration being ~0 and, under these conditions, the slopes of the Arrhenius plots give activation energies equivalent to those for H diffusion. It is shown here that under constant upstream p(H(2)) conditions, nonideality of the H leads to nonlinear Arrhenius plots of P for Pd and especially for some Pd alloy membranes where the H(2) solubilities are significant even at moderate p(H(2)). For example, the permeabilities of a Pd(0.77)Ag(0.23) alloy membrane and a Pd(0.94)Y(0.06) alloy membrane are found to be nearly independent of temperature (423 to 523 K) in the range of constant upstream pressures from 16.1 to 81 kPa.     

Studies on the permeation of aromatic hydrocarbons through liquid surfactant membranes

The effect of operating parameters on the batch scale permeation of hydrocarbons from benzene—heptane mixtures and a straight run naphtha through liquid membrane is reported. The thirteen operating parameters studied include: mixing intensity, surfactant concentration, treat ratios, contact times, temperature and additives. The variations observed in the two key properties of selectivity and aromatic recoveries as well as in product compositions with change in operating parameter is discussed. Surfactant concentration contact time during permeation, type and concentration of additive used appear to exert a marked effect on the enrichment obtained. The careful optimization of operating parameters give selectivities as high as 50 and aromatic recoveries of 75% in one stage at 30°C. Comparison of data with batch liquid—liquid extraction data from extraction of similar feed mixtures with the most widely used solvent, sulpholane, under typical industrial conditions, has shown that selectivities and aromatic recoveries in liquid membrane permeation (LMP) are much higher. Batch scale LMP experiments with straight run naphtha as feed show that under optimum conditions of membrane stability and operating parameters the dearomatization of naphtha from an initial aromatic level of 22 vol.% to 10.5 vol.% is possible in one stage at 30°C with a raffinate yield of 63%. The results obtained on benzene—heptane model mixture compare fairly well with those obtained on naphtha feed.     

Modeling permeation of volatile organic molecules through reverse osmosis spiral-wound membranes

Reverse osmosis is an interesting process to eliminate organic solutes from distillery condensates before recycling them into the fermentation step. However, organic solutes transport phenomena through reverse osmosis membranes are specific. Rejection and sorption of five compounds were studied on a brackish water membrane. Acetic acid and 2,3-butanediol were not sorbed on the membrane while furfural and 2-phenylethanol presented strong sorption following the Langmuir pattern. These sorption effects coupled with solute molecular weight (MW) led to low rejections of acetic acid and furfural (30–60%) and high rejections of 2,3-butanediol and 2-phenylethanol (80–98%). With intermediate sorption and MW, butyric acid showed rejections between 70 and 80%. A modified solution-diffusion model was developed to take into account the sorption pattern and predict the concentration profile along the membrane on the retentate and permeate sides. Equilibrium properties were determined experimentally while transport properties were identified with data obtained from a synthetic condensate. This model was validated for various operating conditions with the synthetic and the industrial condensates. It was then used to simulate the influence of the recovery rate on the retentate and permeate concentrations. It showed the behavior differences between solutes with a linear sorption and solutes with a saturating sorption.     

Frictional coefficients and activation energies for permeation through cellulose acetate membranes

Analysis of previous experimental results, referring to water permeation in cellulose acetate membranes, is presented by using the frictional model proposed by Kedem and Katchalsky. The frictional coefficients decrease with temperature. The conclusions about the transport mechanisms responsible for the flow resemble those obtained in our previous papers. The activation energies for permeation have been calculated; they are of the same order of magnitude as those reported by other authors. The relation between frictional coefficients and activation energies is considered.     

Studies on phenol permeation through supported liquid membranes containing functionalized polyorganosiloxanes

In this study, the functionalized, linear, hydrophobic fluid organosiloxane polymers, namely, methylhydrosiloxane–dimethylsiloxane copolymers supported on a polypropylene microporous flat sheet membrane (Celgard 2502 and 2402) have been tested as supported liquid membranes (SLMs) for phenol recovery from aqueous phases into a 0.1 M NaOH phase. The functionalized polymers include, Me₃SiO[MeSi(OR)O]_x[Me₂SiO]_ySiMe₃ (containing x = 15–18, 25–35 and 50–55 mol% of R, where R is –(CH₂)_nNMe₂ (n = 3 or 4 or 6) or –(CH₂)₂OEt pendent organofunctional groups. The functionalities, R, tested were derived from the commercially available 3-dimethylamino-1-propanol and 2-ethoxyethanol as well as newly synthesized 4-dimethylamino-1-butanol and 6-dimethylamino-1-hexanol which have been made for the purpose of this study.

The study showed that phenol permeation expressed as permeate flux through the membranes increases with the larger number of carbon spacers in the alkyl chain of the aminoalcohol pendent, larger porosity of the polypropylene support films, higher mol% of the methylhydrosiloxane portion functionalized and faster flow rates of both the feed and the receiving phases. Phenol permeation was enhanced significantly when the mol% of the methylhydrosiloxane portion was 50–55 or 25–35 with 6-dimethylamino-1-hexanol functionality supported on Celgard 2502.