Hydrogen permeation in a palladium membrane tube: Impacts of outlet and vacuum degree

Palladium (Pd) membranes are a crucial device for separating hydrogen and are usually operated at normal pressure on the permeate side with a single outlet. Instead of these common operating conditions, the difference between using a double outlet and a single outlet is studied. Four different vacuum degrees (15–60 kPa) are applied on the permeate side, and the results are compared with the non-vacuum operations. Situations under the vacuum and the effects of temperatures (300–400 °C) on H₂ permeation are discussed. Finally, the influences of different feed gas mixtures (H₂/N_2, H₂/CO_2, and H₂/CO) on the Pd membrane performance are investigated. The results show that there is no difference in H₂ permeation impact the single outlet and the double outlet on the permeate side. When a vacuum is imposed on the permeate side, the H₂ permeation rate and H₂ recovery are efficiently intensified, that is, when the pressure difference is 9 atm, they increase from 73.21 to 84.51% and from 0.0035378 to 0.0040808 mol?s^?1, respectively. Moreover, the H₂ recovery can be improved to up to 68.44% under a vacuum degree of 60 kPa. At a given Reynolds number, an increase in temperature increases the H₂ permeation rate but lowers its recovery, stemming from more H₂ in the feed gas. This study also investigates the feed gas of H₂/N₂ under a vacuum to provide a useful insight into H₂ production and separation from ammonia, and the results are compared with two different feed gases of H₂/CO₂ and H₂/CO mixtures. The results suggest that the impurities (i.e., N₂, CO₂, and CO) have a negative influence on the Pd membrane, which causes the H₂ permeation rate to decrease, and the effect of N₂ is the least significant compared to the other two.     

Hydrogen permeation enhancement in a Pd membrane tube system under various vacuum degrees

Hydrogen purification using palladium (Pd) membrane technology has been seen as a potential solution for producing pure hydrogen form hydrogen-rich gas. Compared to traditional practices of operating the permeate side of the membrane at atmospheric pressure, in this study, a vacuum is applied. The effects of various vacuum degrees applied to the permeate side of the Pd membrane are investigated and compared to the results under normal operation without a vacuum. The feed gas used for experiments consists of a mixture of hydrogen (70 vol%) and nitrogen (30 vol%). Three membrane operating temperatures (320, 350, and 380 °C), four pressure differences (2, 3, 4, and 5 atm) across the membrane, and four vacuum degrees (−15, −30, −45, and −53 kPa) applied to the permeate side are considered. For the three operating temperatures, the best improvements in the performance of hydrogen permeation are at 320 and 350 °C when a −53 kPa vacuum is applied, resulting in 79.4% and 79.1% improvements, respectively, compared to normal operations. Increasing temperatures leads to an increase in H₂ permeation both with and without a vacuum; however, best performances of H₂ permeation are observed in cases without a vacuum.     

Optimization analysis of hydrogen separation from an H2/CO2 gas mixture via a palladium membrane with a vacuum using response surface methodology

This study uses a palladium membrane to separate hydrogen from an H₂/CO₂ (90/10 vol%) gas mixture. Three different operating parameters of temperature (320–380 °C), total pressure difference (2–3.5 atm), and vacuum degree (15–49 kPa) on hydrogen are taken into account, and the experiments are designed utilizing a central composite design (CCD). Analysis of variance (ANOVA) is also used to analyze the importance and suitability of the operating factors. Both the H₂ flux and CO₂ (impurity) concentration on the permeate side are the targets in this study. The ANOVA results indicate that the influences of the three factors on the H₂ flux follow the order of vacuum degree, temperature, and total pressure difference. However, for CO₂ transport across the membrane, the parameters rank as total pressure difference > vacuum degree > temperature. The predictions of the maximum H₂ flux and minimum CO₂ concentration by the response surface methodology are close to those by experiments. The maximum H₂ flux is 0.2163 mol s^?1 m^?2, occurring at 380 °C, 3.5 atm total pressure difference, and 49 kPa vacuum degree. Meanwhile, the minimum CO₂ concentration in the permeate stream is t 643.58 ppm with the operations of 320 °C, 2 atm total pressure difference, and 15 kPa vacuum degree. The operation with a vacuum can significantly intensify H₂ permeation, but it also facilitates CO₂ diffusion across the Pd membrane. Therefore, a compromise between the H₂ flux and the impurity in the treated gas should be taken into account, depending on the requirement of the gas product.     

Hydrogen permeation and recovery from H2–N2 gas mixtures by Pd membranes with high permeance

Hydrogen separation from H₂–N₂ gas mixtures by means of high-permeance Pd membranes is an appropriate route to gain pure hydrogen for fuel cell applications. To figure out the mass transfer phenomena of H₂ in membrane tubes, H₂ permeation and recovery characteristics of two high-permeance Pd membranes are investigated. Four important factors influencing H₂ permeation, namely, the H₂ pressure difference, H₂ concentration, the flow rate at the exit of the retentate side, and membrane temperature, are taken into account. The experimental results suggest that decreasing H₂ concentration, flow rate, and temperature reduce the permeances of the membranes and H₂ recovery, even though the H₂ pressure difference is identical. The dimensionless permeance, a permeance ratio between H₂–N₂ gas mixture and pure H₂ as feed gases, is used to evaluate the extent of concentration polarization. Within the investigated ranges of the four factors, the dimensionless permeances of the two membranes are in the ranges of 0.022–0.206 and 0.042–0.359, respectively, revealing that the concentration polarization diminishes the permeance of the membranes down to the level within two orders of magnitude. Nevertheless, over 46% of H₂ is recovered.     

Optimization of hydrogen enrichment via palladium membrane in vacuum environments using Taguchi method and normalized regression analysis

In this study, the separation of hydrogen from gas mixtures using a palladium membrane coupled with a vacuum environment on the permeate side was studied experimentally. The gas mixtures composed of H₂, N₂, and CO₂ were used as the feed. Hydrogen permeation fluxes were measured with membrane operating temperature in the range of 320–380 °C, pressures on the retentate side in the range of 2–5 atm, and vacuum pressures on the permeate side in the range of 15–51 kPa. The Taguchi method was used to design the operating conditions for the experiments based on an orthogonal array. Using the measured H₂ permeation fluxes from the Taguchi approach, the stepwise regression analysis was also employed for establishing the prediction models of H₂ permeation flux, followed by the analysis of variance (ANOVA) to identify the significance and suitability of operating conditions. Based on both the Taguchi approach and ANOVA, the H₂ permeation flux was mostly affected by the gas mixture composition, followed by the retentate side pressure, the vacuum degree, and the membrane temperature. The predicted optimal operating conditions were the gas mixture with 75% H₂ and 25% N₂, the membrane temperature of 320 °C, the retentate side pressure of 5 atm, and the vacuum degree of 51 kPa. Under these conditions, the H₂ permeation flux was 0.185 mol s^?1 m^?2. A second-order normalized regression model with a relative error of less than 7% was obtained based on the measured H₂ permeation flux.     

Thermal stability and effect of typical water gas shift reactant composition on H2 permeability through a Pd-YSZ-PSS composite membrane

This work comprises a study of hydrogen separation with a composite Pd-YSZ-PSS membrane from mixtures of H₂, N₂, CO and CO₂, typical of a water gas shift reactor. The Pd layer is extended over a tubular porous stainless steel support (PSS) with an intermediate layer of yttria-stabilized-zirconia (YSZ). YSZ and Pd layers were incorporated over the PSS using Atmospheric Plasma Spraying and Electroless Plating techniques, respectively. The Pd and YSZ thickness values are 13.8 and 100 μm, respectively, and the Pd layer is fully dense. Permeation measurements with pure, binary and ternary gases at different temperatures (350–450 °C), trans-membrane pressures (0–2.5 bar) and gas composition have been carried out. Moreover, thermal stability of the membrane was also checked by repeating permeation measurements after several cycles of heating and cooling the system. Membrane hydrogen permeances were calculated using Sieverts' law, obtaining values in the range of 4·10⁻⁵–4·10⁻⁴ mol m⁻² s⁻¹ Pa^−0.5. The activation energy of the permeance was also calculated using Arrhenius' equation, obtaining a value of 16.4 kJ/mol. In spite of hydrogen selectivity being 100% for all experiments, the hydrogen permeability was affected by the composition of feed gas. Thus, a significant depletion in H₂ permeate flux was observed when other gases were in the mixture, especially CO, being also more or less significant depending on gas composition.     

Effect of Pd overlayer and mixed gases on hydrogen permeation of Pd/Nb30Hf35Co35/Pd composite membranes

Effect of Pd overlayer and mixed gases on hydrogen permeation of Pd/Nb₃₀Hf₃₅Co₃₅/Pd composite membranes was investigated. The diameter of Pd particle increases with increasing sputtering power. With this change, the membrane shows a signification reduction in hydrogen permeability/or flux, but its durability and stability increases significantly, which can be mainly attributed to a decrease in hydrogen solubility coefficient. In addition, H₂S impurity in mixed gases can greatly degrade membrane performance, especially the hydrogen permeability, whereas the Ar impurity content has less effect in the temperature range of 523–673 K. Lowering of permeability caused by the change of gas purity can be attributed to a decrease in hydrogen solubility, which is closely related to the stronger adsorption of H₂S molecules to the Pd overlayer of the membrane. Thus, it is concluded that aside from the optimum design for composition of Nb-based hydrogen permeable alloy to improve their permeability, the control of Pd overlayer film on membrane surface and gas purity in the feed gas is important.     

Hydrogen permeation properties of Pd-coated V89.8Cr10Y0.2 alloy membrane using WGS reaction gases

The influence of co-existing gases on the hydrogen permeation was studied through a Pd-coated V_89.8Cr₁₀Y_0.2 alloy membrane. Preliminary hydrogen permeation experiments have been confirmed that hydrogen flux was 6.26 ml/min/cm² for a Pd-coated V_89.8Cr₁₀Y_0.2 alloy membrane (thick: 0.5 mm) using pure hydrogen as feed gas. Also, the hydrogen permeation flux decreased with decrease of hydrogen partial pressure at constant pressure when H₂/CO₂ and H₂/CO₂/H₂S mixture applied as feed gas respectively and permeation fluxes were satisfied with Sievert's law in different feed conditions. It was found from XRD and SEM results after permeation test that the Pd-coated V_89.8Cr₁₀Y_0.2 alloy membrane had good stability and durability for various mixture feeding conditions.     

Comparative study on the numerical simulation of hydrogen separation through palladium and palladium–copper membranes

The hydrogen (H₂) diffusion through palladium (Pd) and Pd–copper (Cu) membranes was numerically investigated by developing a two-dimensional computational fluid dynamics model for predicting the performance of H₂ separation. The momentum and mass transport phenomena in the laminar flow conditions were solved at different operating conditions in a vertical cylindrical-type reactor. The effect of feed-gap distance, H₂ concentration, and reactor heating temperature on the H₂ permeation processes were simulated and compared for both Pd-based membranes. The concentration, velocity, and convective and diffusion mass transfer flux distributions were analyzed using the designed model. The H₂ concentration was proportional to the feed-gap distance/cross-sectional area. The smaller the feed-gap distance, the greater the probability of a H₂ molecule being adsorbed by the membrane surface and the ionization energy increasing, leading to further H₂ dissociation through the Pd-based membranes. It was found that the diffusion flux of all feed concentrations was substantially decreased 50 s after the start of the permeation process. Moreover, the diffusion flux of the Pd–Cu40% membrane was relatively larger than that of the pure Pd membrane under the same operating conditions. The distributions of the convective flux, diffusion mass transfer flux, and concentration of the Pd–Cu40% membrane were substantially increased up to 350 °C, then fell to a lower value at higher temperatures. The simulation results were validated with the experimental results, with analysis indicating a good agreement with the experimental results under the same operating conditions. It can be concluded that the simulation modeling for Pd-based membranes was able to predict the optimum operating conditions at high H₂ diffusion rates.     

Catalytic reforming of natural gas for hydrogen production in a pilot fluidized-bed membrane reactor: Mapping of operating and feed conditions

A fluidized-bed membrane reformer was operated in two independent laboratories to map various operating conditions, to investigate the effects of changing the composition of the natural gas feed stream and to verify earlier experimental trials. Two feed natural gases were tested, containing either 95.5 or 90.1 mol% of methane (3.6 or 9.9 mol% of other gaseous higher hydrocarbons). Experimental tests investigated the influence of total membrane area, reactor pressure, permeate pressure and natural gas feed rates. A permeate-H₂-to reactor natural gas feed molar ratio >2.3 was achieved with six two-sided membrane panels under steam reforming conditions and a pressure differential across the membranes of 785 kPa. The total cumulative reforming time reached 395 h, while hydrogen purity exceeded 99.99% during all tests.     

Improved hydrogen permeation through thin Pd/Al2O3 composite membranes with graphene oxide as intermediate layer

Here we proposed the decreasing in the roughness of asymmetric alumina (Al₂O₃) hollow fibers by the deposition of a thin graphene oxide (GO) layer. GO coated substrates were then used for palladium (Pd) depositions and the composite membranes were evaluated for hydrogen permeation and hydrogen/nitrogen selectivity. Dip coating of alumina substrates for 45, 75 and 120 s under vacuum reduced the surface mean roughness from 112.6 to 94.0, 87.1 and 62.9 nm, respectively. However, the thicker GO layer (deposited for 120 s) caused membrane peel off from the substrate after Pd deposition. A single Pd layer was properly deposited on the GO coated substrates for 45 s with superior hydrogen permeance of 24 × 10⁻⁷ mol s⁻¹m⁻² Pa⁻¹ at 450 °C and infinite hydrogen/nitrogen selectivity. Activation energy for hydrogen permeation through the Al₂O₃/GO/Pd composite membrane was of 43 kJ mol⁻¹, evidencing predominance of surface rate-limiting mechanisms in hydrogen transport through the submicron-thick Pd membrane.     

Hydrogen permeation studies of composite supported alumina-carbon molecular sieves membranes: Separation of diluted hydrogen from mixtures with methane

One alternative for the storage and transport of hydrogen is blending a low amount of hydrogen (up to 15 or 20%) into existing natural gas grids. When demanded, hydrogen can be then separated, close to the end users using membranes. In this work, composite alumina carbon molecular sieves membranes (Al-CMSM) supported on tubular porous alumina have been prepared and characterized. Single gas permeation studies showed that the H₂/CH₄ separation properties at 30 °C are well above the Robeson limit of polymeric membranes. H₂ permeation studies of the H₂–CH₄ mixture gases, containing 5–20% of H₂ show that the H₂ purity depends on the H₂ content in the feed and the operating temperature. In the best scenario investigated in this work, for samples containing 10% of H₂ with an inlet pressure of 7.5 bar and permeated pressure of 0.01 bar at 30 °C, the H₂ purity obtained was 99.4%.     

Effect of PBI-HFA surface treatments on Pd/PBI-HFA composite gas separation membranes

For pure hydrogen separation, palladium was deposited on surface-treated polybenzimidazole (PBI-HFA, 4,4′-(hexafluroisopropylidene)bis(benzoic acid)) via the vacuum electroless plating technique (VELP). Since the hydrophobic characteristics of the polymer surface restrict strong adhesion of Pd on it and cause the peel-off of Pd film, various surface treatments have been employed. To increase the number of Pd anchoring sites on the PBI-HFA surface, mechanical abrasion (polishing) was applied, and to increase the hydrophilicity of the PBI-HFA surface, wet-chemical and O₂ plasma treatment (dry etching) were used. The thickness and effective permeating area of the deposited Pd films on the PBI-HFA membranes were estimated to be in the range of 160–340 nm and 8.3 cm², respectively. Among the tested membranes, membranes with Pd layers deposited on O₂ plasma treated PBI-HFA surfaces had the most uniform microstructure and the least number of defects compared to the other membranes. Gas permeation experiments were performed as a function of temperature and pressure. The gases used in the permeation measurements were H₂, N₂, CO₂, and CO (99.9% purity). A Pd-O₂30 m membrane, fabricated by O₂ plasma surface treatment during 30 min, exhibited superior gas separation performance (H₂ permeability of 275.5 Barrer), and proved to be impermeable to carbon monoxide. Enhancement of H₂ permselectivity of Pd/PBI-HFA composite membrane treated by O₂ plasma shows promising hydrogen separation membrane.     

A mixed electronic and protonic conducting hydrogen separation membrane with asymmetric structure

In this work, highly doped ceria with lanthanum, La_0.5Ce_0.5O_2−δ (LDC), are developed as hydrogen separation membrane material. LDC presents a mixed electronic and protonic conductivity in reducing atmosphere and good stability in moist CO₂ environment. LDC separation membranes with asymmetrical structure are fabricated by a cost-saving co-pressing method, using NiO + LDC + corn starch mixture as substrate and LDC as top membrane layer. Hydrogen permeation properties are systemically studied, including the influence of operating temperature, hydrogen partial pressure in feed stream and water vapor in both sides of the membrane on hydrogen permeating fluxes. Hydrogen permeability increases as the increasing of temperature and hydrogen partial pressure in feed gas. Using 20% H₂/N₂ (with 3% of H₂O) as feed gas and dry high purity argon as sweep gas, an acceptable flux of 2.6 × 10⁻⁸ mol cm⁻² s⁻¹ is achieved at 900 °C. The existing of water in both sides of membrane has significant effect on hydrogen permeation and the corresponding reasons are analyzed and discussed.     

A sorption rate hypothesis for the increase in H2 permeability of palladium-silver (Pd-Ag) membranes caused by air oxidation

Hydrogen permeation measurements were performed at 300 °C for 25-μm cold-rolled Pd-Ag 25 wt% membranes before and after air oxidation at the same temperature as permeation. The air oxidation resulted in enhanced H₂ permeation through the membrane, as well as a roughening of the surface with the formation of surface grains and defects. The protruding grains can be leveled off by exposure to H₂ but the surface defects cannot. These microstructure changes are only on the membrane surfaces and do not create transmembrane defects that would allow permeation for gas species other than H₂. The H₂ permeability of the oxidized membrane increased by 25-90% compared to that of the as-received film at the same permeation condition, and the membranes retained perfect H₂ selectivity over N₂. The percent improvement of H₂ permeability decreases with increasing H₂ feed pressure. A new sorption kinetics hypothesis is proposed to elucidate the increase in H₂ permeability of Pd-Ag membranes caused by oxidation. H₂ solubility and sorption rate results were presented to test the new hypothesis. It is found that air oxidation does not change the H₂ solubility in Pd-Ag membranes, but enhances the H₂ sorption kinetics significantly. The extent of kinetics enhancement also decreases with increasing H₂ pressures. The much faster sorption equilibrium implies higher effective H₂ diffusivity at the Pd-Ag membrane surface for the oxidized sample and a higher transfer rate of atomic hydrogen from surface/sub-surface to the membrane bulk that contributes to the increase of H₂ permeability observed in experiments.     

Effect of UV irradiation on PC membrane and use of Pd nanoparticles with/without PVP for H2 selectivity enhancement over CO2 and N2 gases

The hydrogen-based economy is one of the possible approaches toward to eliminate the problem of global warming, which are increases because of the gathering of greenhouse gases. Palladium (Pd) is well-known material having a strong affinity to the hydrogen absorbing property and thus appropriate material to embed in the membrane for the improvement of selective permeation of hydrogen gas. In present work, we have functionalized polycarbonate (PC) membranes with the help of UV irradiation to embed the Pd nanoparticles in pores as well as on the surface of the PC membrane. Use of Pd Nanoparticles is helpful to enhance the H₂ selectivity over other gases (CO₂, N_2, etc.). Also, the UV based modification of membrane increases the attachment of Pd Nanoparticles. Further to enhance the Pd nanoparticles attachment, we used PVP binder with Pd nanoparticles solution. Gas permeability measurements of functionalized PC membranes have been carried out, and better selectivity of hydrogen has been found in the functionalized and Pd nanoparticle binded membrane. PC membrane with 48 h UV irradiated and Pd NPs with PVP have been found to have maximum selectivity and permeability for H₂ gas. All the samples being characterized by scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), Raman spectroscopy and UV–Vis spectroscopy for their morphological and structural investigation.     

Silica-zirconia membrane supported on modified alumina for hydrogen production in steam methane reforming unit

A hydrogen-selective and hydrothermally stable membrane composed of silica-zirconia layer deposited on a modified alumina sub layer was successfully prepared. The composite membrane was used for hydrogen purification from synthesis gas in steam methane reforming process. Silica-zirconia layer was synthesized using the CVD method at 923K and atmospheric pressure while the alumina base layer was prepared via sol-gel procedure. DLS, XRD, SEM/EDX, and BET characterization techniques were used to prove that γ-alumina base and silica-zirconia layer are fabricated successfully. Using the composite membrane, hydrogen selectivity toward other gases has improved significantly. Moreover, H₂/CH₄, H₂/CO, and H₂/CO₂ selectivity have been increased from 700, 350 and 70 in 5 h CVD synthesized membrane to 1600, 750 and 570 for 12 h CVD synthesized membrane respectively. The synthesized Silica-Zirconia membrane successfully altered gases permeability tendency order from H₂ > CH₄ > CO₂ > CO to H₂ > CO₂ > CO > CH₄ which leads to better separation of the product from methane feed. Finally, hydrothermal stability test demonstrated that permeability loss in silica-zirconia CVD coated membrane for H₂ is 45.7% after 48 h, while for silica CVD coated on the modified alumina approaches 92.5%.     

Pre-activation of SBA-15 intermediate barriers with Pd nuclei to increase thermal and mechanical resistances of pore-plated Pd-membranes

Thermal and mechanical resistances of palladium composite membranes prepared by Electroless Pore-Plating (ELP-PP) and containing SBA-15 as intermediate layer were improved by doping the silica material with Pd nuclei before its incorporation on the composite membrane. Textural properties of synthesized SBA-15 materials (both raw and doped ones) were analyzed by XRD, N₂ adsorption-desorption at 77 K and TEM, while the main properties of the composite membrane were determined by SEM and gravimetric analyses. Moreover, membrane permeation tests were also carried out with pure gases, hydrogen and nitrogen, and binary mixtures of them at temperature of 400 °C and pressure driving forces in the range of 0.5–2.5 bar. The use of bare SBA-15 intermediate layer leads to the appearance of cracks on the Pd layer during permeation experiments at high temperature. In contrast, the use of Pd-doped SBA-15 particles avoids this problem, thus improving both thermal and mechanical resistances of the composite ELP-PP Pd-membrane. Following this preparation method, an estimated Pd thickness of 7.1 μm was obtained, reaching a hydrogen permeance of 3.81·10^?4 mol s^?1 m^?2 Pa^?0.5 and ensuring an ideal H₂/N₂ separation factor higher than 2550 at 400 °C.     

Formation of hydrogen bubbles in Pd-Ag membranes during H2 permeation

Palladium membranes used for hydrogen separation seemingly develop cavities filled with hydrogen, i.e. hydrogen bubbles, along the grain boundaries. These bubbles may represent initial stages of pinhole formation that lead to unselective leakage and compromise the long-term stability of the membranes. Alloying with Ag improves the permeability of Pd, but whether these H₂ bubbles form in Pd-Ag membranes remained unknown. In this work, the microstructure of a Pd₇₇Ag₂₃ membrane was characterized by electron microscopy after H₂ permeation testing for 50 days at 15 bar at temperatures up to 450 °C. The results show that Ag does not prevent bubbles from emerging along high-angle grain boundaries, but reduces the number of potential nucleation sites for cavity formation by supressing the development of dislocation networks when H-saturated Pd is cycled through the miscibility gap. Both magnetron-sputtered and electroless plated membranes are afflicted by H₂ bubbles, thus their formation seems determined by intrinsic properties of the material independent of the fabrication technique. The qualitative discussion enables to point directions for enhancement of membrane stability.     

Evaluation of two gas membrane modules for fermentative hydrogen separation

The ability of (dimethyl siloxane) (PDMS) and SAPO 34 membrane modules to separate a H₂/CO₂ gas mixture was investigated in a continuous permeation system in order to decide if they were suitable to be coupled to a biological hydrogen production process. Permeation studies were carried out at relatively low feed pressures ranging from 110 to 180 kPa. The separation ability of SAPO 34 membrane module appeared to be overestimated since the effect concentration polarization phenomena was not taken into consideration in the permeation parameter estimation. On the other hand, the PDMS membrane was the most suitable to separate the binary gas mixture. This membrane reached a maximum CO₂/H₂ separation selectivity of 6.1 at 120 kPa of feed pressure. The pressure dependence of CO₂ and H₂ permeability was not considerable and only an apparent slight decrease was observed for CO₂ and H₂. The mean values of permeability coefficients for CO₂ and H₂ were 3285 ± 160 and 569 ± 65 Barrer, respectively. The operational feed pressure found to be more adequate to operate initially the PDMS membrane module coupled to the fermentation system was 180 kPa, at 296 K. In these conditions it was possible to achieve an acceptable CO₂/H₂ separation selectivity of 5.8 and a sufficient recovery of the CO₂ in the permeate stream.