Frequency-domain analysis of hydrogen permeation across Pd77Ag23 metallic membranes

Hydrogen permeation across palladium-alloys membranes is an industrial process used for purification purposes. In state of the art systems, several tens of microns thick metallic membranes can be used and rate limitations generally come from atomic H diffusion. Cost considerations (for example for application in the automotive industry) require a reduction of the membrane thickness and operation at lower temperature. In the micron-thick range, surface contributions become increasingly rate-determining. To optimize permeation membranes, there is therefore a need to separately measure surface and bulk rate contributions to the overall permeation process. In this paper, pneumato-chemical impedance spectroscopy (PIS) is used to analyze the dynamics of hydrogen permeation across Pd₇₇Ag₂₃ membranes. Experimental pneumato-chemical transfer functions of the membrane are measured at different temperatures. Model impedances are calculated and fitted to the experimental ones, yielding microscopic rate parameter values such as surface resistance and hydrogen diffusion coefficient.     

Hydrogen sorption by Pd77Ag23 metallic membranes. Role of hydrogen content, temperature and sample microstructure

Permeation across metallic membranes is a process used in the industry for purifying hydrogen. In conventional technology, a few tens of micrometers thick metallic membranes made of palladium alloys are used in the 400-600 °C temperature range, using a driving force of several bars for enhanced kinetics. In stationary conditions of flow, the diffusion-controlled transport of atomic hydrogen across the membrane is usually rate-determining. When thin (sub-micron thick) membranes are used, surface rate contributions become more significant. To optimize permeation performances, there is therefore a need for separately measuring surface and bulk rate contributions. In this communication, we report on the kinetics of hydrogen permeation across Pd₇₇Ag₂₃ metallic membranes using pneumato-chemical impedance spectroscopy. The role of different operating parameters (temperature, surface state, membrane microstructure) on the kinetics of permeation is analyzed and discussed.     

Time and frequency domain analysis of hydrogen permeation across PdCu metallic membranes for hydrogen purification

Hydrogen permeation is a process used in the industry for purification purposes. Palladium alloys (PdAg and PdCu) are commonly used as membrane material. In this communication, we report on the kinetics of hydrogen permeation across Pd_0.47Cu_0.53 metallic membranes which can be used in catalytic crackers of biofuels. The permeation mechanism is a multi-step process including surface chemisorption of molecular hydrogen (upstream side of the membrane), hydrogen diffusion across bulk regions, hydrogen recombination (downstream side of the membrane) and evolution. The role of different operating parameters (temperature, surface state, sample microstructure) is analyzed and discussed using both time and frequency domain experiments. Experimental pneumato-chemical impedance diagrams show that there is no significant rate-limitation at surfaces, except at low temperatures close to room temperature. Diffusion-controlled transport of hydrogen across the membrane is rate-determining. However, the value of the hydrogen diffusion coefficient does not rise exponentially with operating temperature in the 40–400 °C temperature range under investigation, as expected for a thermally activated diffusion process. At temperatures as low as 300 °C, new rate-limitations appear. They can be attributed to recrystallization and/or phase transformation processes induced by temperature and the presence of hydrogen.     

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 dynamics across a palladium membrane in a varying pressure environment

Permeation dynamic of hydrogen through a palladium (Pd) membrane in an environment of varying pressure is investigated and analyzed experimentally. By monitoring the instantaneous pressure and mass transfer rate of hydrogen in the conducted system, the present study provides a comprehensive and precise measurement on the permeance of the membrane. It is found that a threshold of pressure difference between the both sides of the membrane for hydrogen permeation is exhibited. That is, when the driving force of the mass transfer is below the minimum pressure difference, hydrogen permeation will not occur. Accordingly, a modified equation accounting for the hydrogen permeation flux through the membrane is suggested. As a whole, the hydrogen permeation flux versus the pressure difference is characterized by a linear relationship, regardless of what the pressure exponent is. Nevertheless, the optimal pressure exponent is located between 0.5 and 0.7. A dimensionless time, the permeation number, is derived to describe the permeation process. The characteristic time of hydrogen permeation depends on the pressure exponent. The experiments reveal that the permeation number is around 7–13 for the hydrogen permeation flux in the system reaching the quasi-steady state.     

Modeling of transient hydrogen permeation process across a palladium membrane

Transient mass transfer processes of hydrogen permeating through a Pd membrane are modeled to aid in predicting the hydrogen transport behavior. The model is established in terms of the quasi-steady time and the steady permeation rate. Meanwhile, four important parameters are considered; they are the permeation lag time, the initial permeation rate, the concave up period and the concave down period. A unit step function is embedded in the model to account for the effect of the hydrogen permeation lag at a lower pressure difference. Corresponding to the lower, the moderate and the higher pressure differences (i.e. 3, 5 and 8 atm), though the hydrogen permeation undergoes a three-stage, a two-stage and a one-stage processes, respectively, these processes can be predicted well by an arc tangential function. By introducing an adjusting parameter in the arc tangential function, there exists an optimal value of the adjusting parameter when the pressure difference is lower. In regard to the moderate and higher pressure differences, the predictions agree with experiments well if the adjusting parameter is sufficiently large. Physically, the unit step function is used to account for the controlling mechanisms of hydrogen diffusion toward the membrane and the spillover of the hydrogen across the membrane. The initial jump parameter represents the rapid response of the initial hydrogen permeation. The adjusting parameter can be used to describe the relative importance of the concave up and the concave down periods.     

Stability of pore-plated membranes for hydrogen production in fluidized-bed membrane reactors

Pd-based membranes prepared by pore-plating technique have been investigated for the first time under fluidization conditions. A palladium thickness around 20 μm was achieved onto an oxidized porous stainless steel support. The stability of the membranes has been assessed for more than 1300 h in gas separation mode (no catalyst) and other additional 200 h to continuous fluidization conditions. Permeances in the order of 5·10⁻⁷ mol s⁻¹ m⁻² Pa⁻¹ have been obtained for temperatures in a range between 375 and 500 °C. During fluidization, a small decrease in permeance is observed, as consequence of the increased external (bed-to-wall) mass transfer resistances. Moreover, water gas shift (WGS) reaction cases have been carried out in a fluidized bed membrane reactor. It has been confirmed that the selective H₂ separation through the membranes resulted in CO conversions beyond the thermodynamic equilibrium (of conventional systems), showing the benefits of membrane reactors in chemical conversions.     

Toward effective membranes for hydrogen separation: Multichannel composite palladium membranes

Composite palladium membranes can be used as a hydrogen separator because of their excellent permeability and permselectivity. The total membrane area in a hydrogen separator must be reasonably large for industrial use, and it is important that each membrane provides a large enough area. Such a demand can be well met by introducing multichannel composite membranes. In this work, a commercially available microporous ceramic filter with 19 channels was used as a membrane substrate, and the diameter of each channel was 4 mm. A uniform thin palladium layer was fabricated inside the narrow channels by using an electroless plating method, and the resulting membranes were highly permeable and selective. This membrane concept provides a high surface-to-volume ratio without causing significant pressure loss, making the hydrogen separator compact and capable. However, special attention should be paid to cleaning the membrane after electroless plating.     

Graphite coating on alumina substrate for the fabrication of hydrogen selective membranes

Development of composite membranes is a suitable alternative to improve the hydrogen flux through palladium membranes. The porous substrate should not represent a barrier to gas permeation, but the roughness of its surface should be sufficiently smooth for the deposition of a thin and defect-free metal layer. In this study, the performances of the modification of the outer surface of an asymmetric alumina hollow fibre substrate by the deposition of a graphite layer were evaluated. The roughness of the substrate outer surface was reduced from 120 to 37 nm after graphite coating. After graphite coating, the hydrogen permeance through the composite membrane produced with 2 Pd plating cycles was of 1.02 × 10^?3 mol s^?1 m^?2 kPa^?1 at 450 °C and with infinite H₂/N₂ selectivity. Similar hydrogen permeance was obtained with the composite membrane without graphite coating, also at infinite H₂/N₂ selectivity, but 3 Pd plating cycles were necessary. Thus, graphite coating on asymmetric alumina hollow fibres is a suitable alternative to reduce the required palladium amount to produce hydrogen selective membranes.     

Corrosion behavior analyses of metallic membranes in hydrogen iodide environment for iodine-sulfur thermochemical cycle of hydrogen production

HI decomposition in Iodine-Sulfur (IS) thermochemical process for hydrogen production is one of the critical steps, which suffers from low equilibrium conversion as well as highly corrosive environment. Corrosion-resistant metal membrane reactor is proposed to be a process intensification tool, which can enable efficient HI decomposition by enhancing the equilibrium conversion value. Here we report corrosion resistance studies on tantalum, niobium and palladium membranes, along with their comparative evaluation. Thin layer each of tantalum, palladium and niobium was coated on tubular alumina support of length 250 mm and 10 mm OD using DC sputter deposition technique. Small pieces of the coated tubes were subject to immersion coupon tests in HI-water environment (57 wt% HI in water) at a temperature of 125–130 °C under reflux environment, and simulated HI decomposition environment at 450 °C. The unexposed and exposed cut pieces were analyzed using scanning electron microscope (SEM), energy dispersive X-ray (EDX) and secondary ion mass spectrometer (SIMS). The extent of leaching of metal into liquid HI was quantified using inductively coupled plasma-mass spectrometer (ICP-MS). Findings confirmed that tantalum is the most resistant membrane material in HI environment (liquid and gas) followed by niobium and palladium.     

Alloying effects of Ru and W on hydrogen diffusivity during hydrogen permeation through Nb-based hydrogen permeable membranes

The hydrogen permeability have been measured for pure niobium and Nb-5 mol%X (X = Ru and W) alloys in order to investigate the alloying effects of ruthenium and tungsten on the hydrogen diffusivity during hydrogen permeation. The hydrogen diffusion coefficient during hydrogen permeation is estimated from a linear relationship between the normalized hydrogen flux, J·d, and the difference of hydrogen concentration, ΔC, between the inlet and the outlet sides of the membrane. It is found that the addition of ruthenium or tungsten into niobium increases the hydrogen diffusion coefficient during the hydrogen permeation. On the other hand, the activation energy for hydrogen diffusion in pure niobium under the practical permeation condition is much higher than the reported values measured for dilute hydrogen solid solutions. It is interesting that the activation energy for hydrogen diffusion is decreased by alloying of ruthenium or tungsten into niobium.     

Computational analysis of geometrical factors affecting experimental data extracted from hydrogen permeation tests: I - Consequences of trapping

Electrochemical permeation tests enable the experimental determination of the diffusion coefficient of a metal. To get a better understanding and a correction of experimental measures, we investigated the effects of hydrogen trapping on the diffusion of hydrogen through a metallic membrane by simulating a FEM model. The trap binding energy ΔE_T ranges from −0.1 to −0.32 eV, the density of traps ranges between 10⁻⁴ and 100 mol/m³, and the thickness of the membrane fluctuates from 100 μm to 1 mm. It appears that the effective diffusion coefficient extracted from desorption flux data of a single membrane is not influenced by its geometry and depends on both the density of trapped hydrogen and the trap binding energy such as the apparent diffusion coefficient implemented in the code. Thus we do not detect any scale effect. In the other hand, the effective subsurface concentration evaluation using usually Fick’s laws doesn’t correspond directly to hydrogen concentration in the membrane. Analytical equations to solve the problem to extract erroneous data (diffusion coefficient and hydrogen concentration) to the experimental measurements of the flux vs time curves have been proposed.     

An alternative and comprehensive approach to estimate trapped hydrogen in steels using electrochemical permeation tests

The present study critically discusses the apparent diffusivity of H during build-up and decay permeation transients in four microalloyed line-pipe steels using Devanathan-Stauchursky cell in galvanostatic-potentiostatic condition. Among the existing models, the Carvalho's solution has been found to give the best fit with the experimental build-up transient. The numerical solution of trapping and detrapping events of H during build-up transients is investigated by solving the McNabb-Foster model. The residual hydrogen was measured in the steels using LECO-DH603 hydrogen determinator after saturation H-charging. Further the Carvalho's exponent of permeation transient has been found to have good correlation with the H trapping-detrapping rates during the permeation test and also with the residual H content in the steel. Finally, empirical relationships have been proposed that can be used to estimate the residual H in the studied steels along with the trapping-detrapping parameters using the permeation test data.     

Preparation, testing and modelling of a hydrogen selective Pd/YSZ/SS composite membrane

A palladium selective tubular membrane has been prepared to separate and purify hydrogen. The membrane consists of a composite material, formed by different layers: a stainless steel support (thickness of 1.9 mm), an yttria-stabilized zirconia interphase (thickness of 50 μm) prepared by Atmospheric Plasma Spraying and a palladium layer (thickness of 27.7 μm) prepared by Electroless Plating. The permeation properties of the membrane have been tested at different operating conditions: retentate pressure (1-5 bar), temperature (350-450 °C) and hydrogen molar fraction of feed gas (0.7-1). At 400 °C, a permeability of 1.1 × 10⁻⁸ mol/(s m Pa^0.5) and a complete selectivity to hydrogen were obtained. The complete retention of nitrogen was maintained for all tested experiment conditions, with both single and mixtures of gases, ensuring 100% purity in the hydrogen permeate flux.A rigorous model considering all the resistances involved in the hydrogen transport has been applied for evaluating the relative importance of the different resistances, concluding that the transport through the palladium layer is the controlling one. In the same way, a model considering the axial variations of hydrogen concentration because of the cylindrical geometry of the experimental device has been applied to the fitting of the experimental data. The best fitting results have been obtained considering Sieverts’-law dependences of the permeation on the hydrogen partial pressure.     

Evaluation of hydrogen permeation through standalone thermally sprayed coatings of AISI 316L stainless steel

This research evaluates hydrogen permeation and its diffusion characteristics through standalone thermally sprayed coatings of AISI 316L stainless steel. The effects of various charging currents and other parameters on hydrogen diffusion coefficient were scrutinized using electrochemical hydrogen permeation tests. Hydrogen permeation through the thermally sprayed coatings displayed anomalous behavior such that a maximum pinnacle was observed in the permeation curves, attributed to heavily trapped hydrogen atoms in the delayed surface cracks. Therefore, new diffusion parameters were defined for modeling of the anomalous permeation curves. The fitted diffusion parameters were consistently identified, and hence, the model perfectly explained experimental data. The results showed that the increase in charging current caused fast activation and development of surface cracks. The measured diffusion coefficient of hydrogen in the stainless steel thermally sprayed coating was relatively high because the microstructure of the coating contained some ferritic phases and dense dendritic structure, which configure fast diffusion paths.     

Transient dynamic of hydrogen permeation through a palladium membrane

Transient dynamic of hydrogen permeation through a palladium membrane is studied in the present study. Three different pressure differences between the two sides of the membrane are considered; they are 3, 5 and 8 atm. The experimental results indicate that the variation in the hydrogen permeation process is notable at the selected pressure differences. When the pressure difference is relatively low (i.e. 3 atm), the hydrogen permeation process proceeds from a time-lag period, then to a concave up period and eventually to a concave down period. Therefore, the transient hydrogen permeation is characterized by a three-stage mass transfer process. When the pressure difference is increased to 5 atm, the time-lag period disappears, thereby evolving the three-stage mass transfer process into a two-stage one. However, the concave up period withers significantly. Once the pressure difference is as high as 8 atm, the transient hydrogen permeation is completely characterized by a concave down curve, yielding a single-stage mass transfer process. A quasi-steady state of hydrogen permeation is defined to evaluate the period of the transient mass transfer process. It suggests that, within the investigated conditions of operation, the time required for hydrogen permeation to reach the steady value is around or over 1 h. For the low pressure difference cases, the transient period is especially long, resulting from the time-lag characteristic. Once the hydrogen permeation is in the steady state, over 80% of hydrogen can be recovered from the membrane.     

Sensing mechanism of hydrogen gas sensor based on RF-sputtered ZnO thin films

The mechanism of hydrogen (H₂) gas sensing in the range of 200–1000 ppm of RF-sputtered ZnO films was studied. The I–V characteristics as a function of operating temperature proved the ohmic behaviour of the contacts to the sensor. The complex impedance spectrum (IS) of the ZnO films showed a single semicircle with shrinkage in the diameter as the temperature increased. The best fitting of these data proved that the device structure can be modelled as a single resistance-capacitance equivalent circuit. It was suggested that the conductivity mechanism in the ZnO sensor is controlled by surface reaction. The impedance spectrum also exhibited a decreased in semicircle radius as the hydrogen concentration was increased in the range from 200 ppm to 1000 ppm.     

Impact of surface oxide on hydrogen permeability of chromium membranes

Hydrogen permeation through pure and oxidised bulk chromium membranes was measured by the classical gas technique to get insight into oxide as a hydrogen permeation barrier (HPB). An additional palladium-coated reference chromium membrane was tested to avoid the influence of native Cr oxide. Key parameters for Cr permeability: P₀ = 3.23 × 10^?7 mol H₂/s/m/Pa^0.5 and E_a = 0.68 eV and Cr diffusivity D₀ = 9.0 × 10^?5 m²/s and E_a = 0.59 eV. In the sample preparation stage, a thin ～2 nm thick oxide was formed. Additional oxidation in pure oxygen at 400 °C increased the thickness from 20 to 50 nm. At this temperature, its efficiency as HPB was evaluated by comparing permeation rates to the reference chromium membrane. The highest permeation reduction factor of ～3900 corresponded to only a ～28 nm thick Cr oxide layer. Surface morphology and oxide thickness were investigated by SEM, while the thickness and type of chromium oxide by XPS.