The hydrogen solubility and permeation in Pd77%Ag23% membranes have been determined as a function of temperature and membrane thickness. Unexpectedly, the solubility of hydrogen is found to systematically increase as the membrane thickness decreases from 11.2 to 2.2 µm. Topography studies by atomic force microscopy in conjunction with previously reported characterization suggest linkage of the hydrogen solubility to the density of grain boundaries. A higher average grain boundary density for thinner membranes results from the nucleation and growth proceeding during membrane fabrication by sputtering. For the membranes and conditions (no membrane pretreatment; 300-400°C; Δp H2 ≤ 2 bar) applied here, surface phenomena affect the hydrogen transport at thicknesses below ~5 µm.Determination of the solubility constants hence allowed extraction of the bulk diffusivity 2 parameters from the permeability measurements over the thicker membranes (6.7-11.2 µm), in good agreement with reported values obtained using other techniques.

Sputtered Pd77%Ag23% membranes of thickness 2.2–8.5 µm were subjected to a three-step heat treatment in air (HTA) to investigate the relation between thickness and the reported beneficial effects of HTA on hydrogen transport. The permeability experiments were complimented by volumetric hydrogen sorption measurements and atomic force microscopy (AFM) imaging in order to relate the observed effects to changes in hydrogen solubility and/or structure. The results show that the HTA—essentially an oxidation-reduction cycle—mainly affects the thinner membranes, with the hydrogen flux increasing stepwise upon HTA of each membrane side. The hydrogen solubility is found to remain constant upon HTA, and the change must therefore be attributed to improved transport kinetics. The HTA procedure appears to shift the transition from the surface to bulk-limited transport to lower thickness, roughly from ~5 to ≤2.2 µm under the conditions applied here. Although the surface topography results indicate that HTA influences the surface roughness and increases the effective membrane surface area, this cannot be the sole explanation for the observed hydrogen flux increase. This is because considerable surface roughening occurs during hydrogen permeation (no HTA) as well, but not accompanied by the same hydrogen flux enhancement. The latter effect is particularly pronounced for thinner membranes, implying that the structural changes may be dependent on the magnitude of the hydrogen flux.

No abstract