Unravelling the effects of surface modification pre-treatments on porous hastelloy X supports for H2 selective Pd-based membranes preparation with a statistical approach

Metal-based supports require specific pre-treatments to reach the surface quality needed for achieving thin Pd layers deposition without defects. Hastelloy X supports with high surface roughness and large pore diameter are acquired and pre-treated via polishing and chemical etching. They are then asymmetrically filled with α-Al2O3 of decreasing particle size (18 μm, 5 μm and 1.5 μm) and equipped with a γ-Al2O3 interdiffusion barrier. Each pre-treatment step is thoroughly characterized and elucidated with ANalysis Of VAriance (ANOVA) as statistical tool, introducing a hybrid observational-statistical approach to infer on a population of membrane supports.The analysis allowed to set average surface roughness (Ra) < 0.8 μm, average profile height (Rz) < 7 μm, in-pore leveling (Δ) < 6 μm, as targets for support pre-treatments reproducibility. The target average pore diameter after asymmetric filling was identified as 100–500 nm, while the target average pore diameter after interdiffusion barrier deposition was below 100 nm. The most effective particle size for in-pore leveling was identified as≤5 μm, and the most effective particle size for average pore diameter reduction as 18 μm.