Ti-Al alloy discs with a diameter of 34 mm, a thickness of 2.5 mm, and an average pore size of 6 μm were used as the support for preparing the thin Pd composite membranes. The porous Ti-Al alloy discs were prepared using the method reported by He et al. (2007).

Prior to preparing the membranes, the supports were cleaned with acetone in an ultrasonic bath for 2 h to remove surface dirt and grease and were then dried overnight at 423 K. The ceramic membranes were prepared on Ti-Al alloy discs by sintering the Ti-Al alloy discs under a static air atmosphere at 1,023 K for 2 h in an electric furnace at the heating and cooling rate of 2 K/min. The procedure for the preparation of ceramic membranes has been described in detail in Zhang et al. (2015).

Electroless plating was used to deposit the Pd membranes on porous ceramic/Ti-Al alloy supports (Alique et al., 2016). Prior to the preparation of the palladium membranes, the surface of the Ti-Al alloy supports was activated and sensitized to seed Pd nuclei (Keuler et al., 2002). The deposition of Pd membranes was carried out at 313 K, and the plating bath was changed after every 60 min until dense Pd membranes with an ideal thickness were obtained. Deionized water was applied to clean the surface of the palladium membranes immediately after electroless plating to remove the impurities adsorbed on the surface and pores of the Pd layer. The composite membranes were then dried overnight at 423 K for 12 h (Zhang et al., 2012).

An electronic balance (Sartorius BS2202S, Germany) was used to monitor the weight of the ceramic membranes. XRD (Bruker D8 Advance diffractometer, with Cu- radiation) was used to examine the phase composition of the ceramic/Ti-Al alloy composite membranes. EDX (JSM, 5600, LV, Japan) was used to analyze the composition of the ceramic membrane surface. The surface and cross-section of the palladium and ceramic membranes were investigated by SEM (Quanta 200, Philip, USA). The gas permeation of the composite membranes at high temperatures was measured using a permeability apparatus. The ceramic/Ti-Al alloy composite membranes were placed in a stainless steel permeator. The permeate side was maintained at atmospheric pressure, while the retentate side was fed with H₂ or N₂. A back-pressure regulator was used to control the upstream pressure. A mass flow controller (Models D08-4D/ZM, Beijing Sevenstar Electronics Co., Ltd., Beijing, China) was used to measure the single gas flux once the temperature and pressure became constant.

The ceramic membranes were prepared on Ti-Al alloy supports by an in situ oxidation method at 1,023 K for 2 h in an air atmosphere followed by their heat treatment in an H₂ atmosphere for 20 h at 773, 873, and 973 K, respectively. The stability test results of the membranes are summarized in Table 1. From the table, it is clear that the weight of the ceramic/Ti-Al alloys did not change when the heat treatment was carried out at temperatures lower than 973 K. This indicates that the oxide layers were stable at temperatures lower than 973 K. As can be seen from Table 2, the change in the N₂ flux of the ceramic/Ti-Al alloy composite membranes after the heat treatment was negligible. This indicates that the heat treatment did not destroy the pore structure of the ceramic membranes and that the ceramic layers were stable at high temperatures.

The surface SEM images of the ceramic/Ti-Al alloys are shown in Figure 1. The oxidation layer was evident, and no surface damages were observed even after the heat treatment at 973 K for 20 h in an H₂ atmosphere. Hence, the ceramic layer prepared by in situ oxidation was stable in an H₂ atmosphere. This is consistent with the weight and N₂ flux measurement results.

In order to investigate the effect of heat treatment on the elemental composition of the surface of the Ti-Al alloy support, an EDX analysis of the membranes was carried out. As shown in Figure 2, oxygen was present on the support surface even after the heat treatment at 973 K in an H₂ atmosphere for 20 h, indicating that the ceramic layer prepared by in situ oxidation was stable at higher temperatures in an H₂ atmosphere.

The XRD patterns of the ceramic/Ti-Al alloy membrane are shown in Figure 3. The characteristic peak keeps constant even after the heat treatment in an H₂ atmosphere for 20 h at 773, 873, and 973 K, indicating that the ceramic/Ti-Al alloy composite membranes were highly stable to high temperatures in an H₂ atmosphere. These results are consistent with the EDX results shown in Figure 2.

The results of gravimetric, gas flux, SEM, EDX, and XRD analyses show that the ceramic layer prepared by the in situ oxidation method is a promising inter-diffusion barrier for Pd/ceramic/Ti-Al alloy composite membranes.

The cross-sectional and surface SEM images of the Pd/ceramic/Ti-Al alloy composite membranes are shown in Figure 4. Figure 4A shows that the surface of the Pd membrane was not very smooth and homogeneous (mainly because of the high surface roughness of the ceramic/Ti-Al alloy support); however, no pinholes or cracks were observed. Although the images show smaller areas of the membranes, the gas leak test results further confirmed that the membranes had no defects at room temperature. Figure 4B shows that the thickness of the Pd membranes was uniform and was ~11–14 μm, which is consistent with the value determined by the gravimetric method (average thickness = ~ 13 μm).

Single gas permeation is an effective and simple method to measure the N₂ and H₂ permeation of palladium composite membranes. The pressure dependence of the hydrogen flux of the Pd composite membranes at various temperatures is shown in Figure 5. The H₂ permeation flux increased with an increase in the pressure difference (Ph0.5-Pl0.5). The temperature also had an obvious effect on the H₂ permeation flux. These phenomena can be caused by the solution-diffusion mechanism.

Figure 6 shows the H₂/N₂ selectivity of the Pd composite membranes at different pressure and temperature. The results show that the selectivity increased with an increase in the temperature from 673–773 K. The H₂/ N₂ selectivity was found to be within the range of 200–700 under the operating conditions. This can be attributed to the thin Pd layers and the high surface roughness of the support.

Table 3 shows a comparison of the membranes prepared in this study with those reported previously. The previously reported membranes had a thickness of ~20 μm and a hydrogen permeance coefficient of ~10⁻³-10⁻⁴ mol m⁻² s⁻¹ Pa^−0.5. On the other hand, the hydrogen permeance coefficient of the membranes prepared in this study was of the order of 10⁻³ mol m⁻² s⁻¹ Pa^−0.5. Hence, the selectivity of these membranes should be improved further.

Activation energy (E) is another critical parameter affecting the membrane permeability. The relationship between the hydrogen permeance of membranes and temperature can be shown by the Arrhenius equation as: Q = Q0exp(-Ea/RT) The Arrhenius relation between temperature and the hydrogen permeance is shown in Figure 7. The average activation energy for H₂ permeation of the palladium composite membranes was calculated to be 13.86 kJ·mol⁻¹ from 673–773 K, which is in agreement with the previously reported value. Table 4 lists the apparent activation energy of various membranes. These apparent activation energies were quite dispersed, due to various factors such as temperature, pressure, membrane structure, and other factors. Although the thicknesses of the Pd membranes were similar, different Pd membranes had different compositions, physical and chemical structures, and apparent activation energies.

The long-term stability of Pd composite membranes is crucial for practical applications. To investigate the effect of temperature variations on the H₂ permeation of the composite membranes, temperature cycling was performed. Figures 8–10 show the hydrogen flux, nitrogen flux, and H₂/N₂ selectivity of the Pd/ceramic/Ti-Al composite membranes operating in three heat cycles of 673 K → 723 K → 773 K → 723 K → 673 K(at ΔP = 0.1 MPa). Under the same operating conditions, the H₂ flux, the N₂ flux, and H₂/N₂ selectivity of the Pd/ceramic/Ti-Al composite membranes showed no significant changes during the three heat cycles from 673–773 K, indicating that the Pd/ceramic/Ti-Al composite membranes possessed good thermal stability. This also shows that the ceramic layers were stable and could effectively avoid the inter-diffusion of metals.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer YJ declared a shared affiliation, though no other collaboration, with one of the authors YF to the handling editor.