Apparatus and technique for assessing degradation of hydrogen compressor alloys during temperature cycling

Abstract

Compressors employing metal hydrides (MH) may operate on waste or solar heat and are mechanically simple in comparison to mechanical hydrogen compressors, requiring only standard cooling components.

In principle, MH compressors operate by absorbing hydrogen at a low pressure (ρl) at the lower working temperature (Tl) by forming the concentrated hydride phase (β) and desorbing at the higher working temperature (Th) and higher pressure (ρh) by forming the solid- solution phase (α), in accordance with the van 't Hoff relation specific for the metal–H2 system in use. By pairing alloys, selected based on their thermodynamic parameters, higher compression ratios can be achieved.

Microstructural changes in the host metal owing to hydrogen absorption (dislocations, vacancies, decrepitation, trapped hydrogen, amorphisation…), generally summed up as degradation, may lead to loss of accessible hydrogen capacity during repeated cycles of absorption and desorption. Because the b phase may be thermodynamically unstable, absorption-desorption cycling between Tl and Th in a compressor may exacerbate degradation compared to isothermal cycling.

Here we present an apparatus designed to execute temperature cycles on multiple samples simultaneously, each isolated under an isochoral constraint by a valve, along with a mathematical model to describe approximately the locus of the metal–H2 system in pressure–composition space. This model allows estimation of the amount of metal sample needed to cycle isochorally between (ρl, Tl, xβ) and (ρh, Th, xα), where xα/β corresponds to hydrogen concentrations in the pure α/β phase, such that every sample will cycle between the β and α phases as the temperature cycles between Tl and Th, as depicted opposite. The pressure increase
Δρ=ρh-ρl has contributions from hydrogen desorbed into the closed volume and the increase in the temperature of the free gas in that volume. The practical application of this model will be discussed.