Group IV and V metals of the periodic table that have bcc structure such as titanium, vanadium, and niobium have higher hydrogen permeabilities than palladium but have surface oxides that prevent hydrogen absorption. Therefore, these metals require a thin film (1 micron or less) of catalyst on their surfaces such as palladium. Generally, Group IV and V metals have shown embrittlement upon exposure or cycling in hydrogen. Besides embrittlement, another problem experienced by metal composite membranes is metallic interdiffusion between the substrate foil and the palladium overcoat. However, alloys of these metals may possess some of the qualities desired in a metal membrane including high hydrogen permeability and relatively low cost.

Overall, more than 250 different alloys were investigated. Alloys were screened for potential use as hydrogen separating membranes by Charpy impact testing of striker bars cut from ingots of each alloy prepared by arc melting. Membrane discs were cut from alloys that passed the brittleness test by thinly slicing an ingot using wire electrode discharge machining, and sand blasting to remove bulk oxide. The discs (between 0.3 and 0.5 mm thick) were ion-milled, and 100-nm-thick palladium or palladium-alloy coatings were applied to both sides in situ via electron-beam-evaporation physical vapor deposition. Membranes were tested for hydrogen permeability over a range of temperatures (up to 700˚C), pressures (up to 0.3 MPa), and durations (> 500 h). Membrane resilience was often assessed by conducting permeability testing in hydrogen at 400˚C and then decreasing the temperature, noting the onset of hydrogen embrittlement as evidenced by crack formation in the membrane.

Several palladium coated titanium and nickel based alloys were tested for hydrogen permeability but these materials exhibited little or no hydrogen flux or broke when exposed to hydrogen. Niobium based materials were permeable to hydrogen but were often too fragile for practical usage. Vanadium based alloys also commonly suffered from embrittlement but were the most ductile of the alloys tested, usually failing in hydrogen at temperatures around 250˚C after initial hydrogen permeability testing at 400˚C. Most of the vanadium alloys displayed hydrogen permeabilities comparable to pure palladium. The influence of adding various elements to vanadium on the hydrogen permeability, resistance to hydrogen embrittlement, and effect on hydrogen flux stability with time was explored. For example, titanium increases hydrogen permeability but does not appear to lessen embrittlement and may promote atomic diffusion to the surface. Palladium clearly increased ductility, for example, a V-10Pd (at.%) membrane withstood hydrogen permeation testing at temperatures as low as 150˚C.

Furthermore, it is essential that the hydrogen dissociating surface catalyst is stable at the temperatures that metal membranes are expected to operate at (up to 450˚C) and lasts for the membrane lifetime (ideally > 100,000 h). One approach to the problem of metallic interdiffusion between the catalytic coating on the membrane surface and the constituents of the membrane itself is to use palladium alloys. The effect of palladium and palladium-alloy coatings on metallic interdiffusion and the hydrogen permeability of Group V membranes will be described. Interdiffusion between the surface coatings and membranes was characterized using Rutherford backscattering spectrometry (RBS) and Auger electron spectroscopy (AES).