Dépôt numérique

Palladium-Copper-Gold Alloys for the Separation of Hydrogen Gas.

Honrado Guerreiro, Bruno Manuel (2015). Palladium-Copper-Gold Alloys for the Separation of Hydrogen Gas. Thèse. Québec, Université du Québec, Institut national de la recherche scientifique, Doctorat en sciences de l'énergie et des matériaux, 216 p.

Télécharger (6MB) | Prévisualisation


The symbols and special characters used in the original abstract could not be transcribed due to technical problems. Please use the PDF version to read the abstract.

The industrial applications of hydrogen gas have made this simple diatomic molecule an important worldwide commodity in the chemical, oil and even food sectors. Moreover, hydrogen gas is a promising energy carrier that aims the delivery of clean energy, bypassing the environmental problems created by carbon-based fuels. Hydrogen production is, however, still reliant on steam reforming of natural gas and coal gasification, despite the innumerous alternatives available. In order to introduce hydrogen in the energy market, hydrogen production costs need to be reduced, and more specifically, hydrogen purification needs to be simplified. In this regard, the use of dense palladium-based membranes for hydrogen purification are especially attractive, but their wide industrial application is impaired mostly by the high cost of palladium and by hydrogen sulfide poisoning. In the current work, the potential use of palladium-copper-gold alloys as membranes for the separation of hydrogen gas was tested. PdCuAu alloys were first prepared by pulsed electrochemical co-deposition on a titanium substrate from Pd(NO3)2, Cu(NO3)2 and Au(OH)3 in HNO3 0.35 M, over a wide composition range ([Pd] = 14-74 at.%, [Cu] = 2-82 at.%, [Au] = 0-66 at.%). All alloys had a single face centered cubic (fcc) phase and were less than 1000 nm thick. The deposits presented different morphologies that were dependent on the palladium content. For example, Pd14.3Cu82.1Au3.6 alloy was composed of spherical structures, while Pd74.3Cu18.4Au7.3 had a more needle-like structure. Phase transition from fcc to the more H2-permeable body-centered cubic (bcc) phase was achieved by a simple heat treatment of 4 hours under Ar at 400ºC. The aforementioned transition was found to be composition dependent, occurring for alloys with 29.5 ≤ Pd ≤ 45.8; 45.3 ≤ Cu ≤ 63.0; and 0.0 ≤ Au ≤ 17.4, with [Pd] + [Cu] + [Au] = 100. The hydrogen solubility of the fcc alloys was measured in NaOH 0.1M by an electrochemical method and was found to be maximum for pure Pd. Replacement of copper with gold resulted in a an increase in solubility for alloys with palladium content between 27 and 54 at.%. At the same composition, the bcc phase showed a substantial reduction in solubility when compared with the fcc phase. For example the hydrogen solubility of the fcc alloy Pd33Cu50Au17 was 10.2%, while that of the corresponding bcc alloy was 0.5%. Magnetron co-sputtering was also used for the production of fcc PdCuAu alloys on titanium substrates. The thickness of the films was below 500 nm. Phase transition from fcc to bcc upon annealing occurred in conditions analogous to the ones observed with the electrodeposited samples. Copper segregated at the surface of the alloy during heat treatment, as evidenced by Xray photoelectron spectroscopy and cyclic voltammetry. The hydrogen solubility of both fcc and bcc phases was investigated. In the case of the fcc phase, it was found that the hydrogen solubility, as measured by an electrochemical method in NaOH 0.1M, increased with the palladium content, being maximum for pure Pd. In general, annealing of the fcc phase resulted in lower solubility either because there was a change in the crystallographic structure to the bcc phase or because of the larger crystallites formed during heat treatment. Further studies with the bcc samples by electrochemical in situ X-ray diffraction in H2SO4 0.1M revealed that hydrogen solubility increased from 0.5 to 2.1% when palladium content increased from 40.3 to 45.6 at%, while keeping [Au]=3.5 at.%. However, no effect of gold was observed for alloys with approximate palladium concentration, in contrast to what was observed with the electrodeposited samples with fcc phase. The hydrogen solubility of both fcc and bcc phase varied linearly with the palladium content, and the increased observed was similar in both phases (0.286 and 0.295 % per [Pd] at.%, for fcc and bcc phase, respectively). This is an indication that the palladium content is the main factor governing solubility in each phase. Gas phase testing was performed on membranes prepared from PdCuAu ternary alloys and synthesized by mechanical alloying. Alloys with four different compositions were prepared, all with approximately the same palladium composition, and increased gold content, namely, Pd39Cu61, Pd41Cu56Au3, Pd40Cu43Au7, and Pd39Cu51Au10. NaCl was used as the process control agent at 2 wt%. The as-milled fcc alloys were first heat treated for 5 h under Ar 5% H2 at 400ºC, which promoted in general full transition to the bcc phase, in accordance with their composition. Membrane preparation from the bcc alloys involved a series of steps such as powder pressing into pellets, sintering, cold rolling, heat treatment and polishing. The final thickness of the membranes was between 277 and 327 nm. Hydrogen permeability was then measured at 464ºC between 30 and 60 psig H2 pressure. Under these conditions, hydrogen permeability was found to be composition dependent and was highest for the alloy with composition Pd40Cu43Au7 (2.1×10-8 mol•m-1•s-1•Pa-0.5) and lowest for the alloy Pd39Cu51Au10 (6.9×10-9 mol•m-1•s-1•Pa-0.5), while the hydrogen permeability of a reference 250 μm commercial foil was 1.2×10-8 mol•m-1•s-1•Pa-0.5. The ideal selectivity of the membranes was higher than 130. The PdCuAu ternary alloy membrane with the highest permeability was further tested at different temperatures and in the presence of hydrogen sulfide. Between 358 and 464ºC, the natural logarithm of the permeability followed an Arrhenius-like relationship with 1/T, resulting in an activation energy of 10.0 kJ•mol-1, which was lower that the value for a pure Pd reference membrane (21.9 kJ•mol-1). Ten minutes into the test with H2S, H2 flow decreased to zero, indicating the formation of a hydrogen impermeable layer on the surface. Consequently, gold did not provide any extra resistance to H2S poisoning under the tested conditions. H2 flow resumed with time, however, post-testing SEM analysis revealed that the formation of pores had taken place. XPS analysis confirmed the presence of multiple sulfurcontaining species, such as, Na2S, Na2S2O3, Na2SO3, Cu2S, CuS and CuSO4. Sulfur compounds with higher oxidation states were probably formed after H2S testing during storage of the membrane under air.

Type de document: Thèse
Directeur de mémoire/thèse: Guay, Daniel
Mots-clés libres: alliages de palladium; alliages de cuivre; alliages d’or; purification de l’hydrogène
Centre: Centre Énergie Matériaux Télécommunications
Date de dépôt: 26 avr. 2016 18:47
Dernière modification: 26 avr. 2016 18:47
URI: http://espace.inrs.ca/id/eprint/3376

Actions (Identification requise)

Modifier la notice Modifier la notice