The surface and cross-sectional analyses were performed
with scanning electron microscopy (SEM, FEI Quanta-200)
and metallographic microscopy (LEICA, DM-4000 M). Before
cross-sectional analysis, the samples were pretreated on
a grinderepolisher (BUEHLER, Phoenix Beta) after an encapsulation
in epoxy resin. The permeation tests were carried
out with a H2/N2 single gas method [23e26], and the
membrane selectivity was defined as the ratio of hydrogen
flux versus the nitrogen flux under identical conditions
of temperature and pressure. Each time, the membrane was
assembled in a testing cell and sealed with graphite [27],
giving an effective membrane length of 50 mm. The pressure
on the permeate side was always ambient. In order to
avoid the hydrogen embrittlement damage, the palladium
membrane should not be exposed to hydrogen below 573 K [1].
Before testing, the membrane was heated at 2 K min1 up to
623 K under nitrogen, and it was then treated at this temperature
under the synthetic air for 2 h to burn out the organic
remnants. The fluxes of pure nitrogen and pure hydrogen
were measured as a function of their pressures on the
membrane retentate side. After test, the membrane was
thoroughly purged with nitrogen before cooling down to room
temperature.
3. Results