The focus of this thesis is the development of novel high temperature membranes based on perovskite ceramics for the separation of oxygen from air. Growing electricity demands worldwide and concerns about anthropological climate change are driving the development of a variety of 'clean-coal' technologies (e.g. oxyfuel firing, coal gasification), which require a feed of oxygen enriched air or even pure oxygen. This is traditionally supplied via cryogenic distillation, which is reliable but energy intensive, necessitating research into alternative technologies. In this thesis, novel perovskite mixed ionic and electronic conduction (MIEC) membranes have been developed.
Perovskite membranes have several key benefits: the oxygen purity is very high, operation at high temperatures could aid heat integration, and they have been claimed to cut oxygen production costs by 35% or more. Perovskites are a class of MIEC materials with crystal structure ABO3, and have exhibited some of the highest fluxes reported in the literature. For instance, the Ba0.5Sr0.5Co0.8Fe0.2O3-δ (BSCF) membranes produce some of the highest oxygen fluxes, and hence was selected as the base material for this thesis. However, these membranes are typically unstable at lower temperatures (below ~850°C). In an attempt to stabilise BSCF and other related compounds, various cations have been investigated in the open literature including bismuth, titanium, tantalum, scandium, aluminium and yttrium. In this thesis, yttrium has been selected as a substituent, since it is well known for its stabilisation properties such as in YSZ (yttria stabilised zirconia) and in perovskite compounds more closely related to BSCF.
A series of BSCF-based compounds were prepared in which yttrium was partially substituted for iron in varying quantities (termed BSCFY), and fabricated into thick (1 mm) membranes. Significant enhancements in bulk diffusion and hence oxygen permeation were observed for low concentrations of yttrium, reaching a maximum at 2.5mol% yttrium on the B-site, which then decayed to zero with increasing yttrium substitution. The oxygen vacancy content was found to increase concomitantly with the oxygen flux, providing more sites for oxygen transport. However, it was clear that several competing mechanism were involved, including lattice expansion. Yttrium substitution was also found to enhance the flux and hence surface kinetics for thin-walled (~0.3 mm) hollow fibres, which was validated using electrical conductivity relaxation (ECR). Yttrium was more beneficial bulk diffusion, since the characteristic length was reduced from ~1 mm for BSCF to ~0.4 mm for BSCFY.
In an attempt to further understand the role of yttrium, the substitution of yttrium was performed on both the A- and B-site. Clear trends in conductivity, lattice parameter and oxygen vacancies implied that the yttrium was incorporated in the nominated site. However, oxygen fluxes appeared to be relatively independent of the nominated site, with yttrium substitutions in both sites giving similar results to those obtained for iron, obtaining a maximum flux at 2.5mol% substitution yttrium on either the A or B site. This was attributed to the competing factors of lattice expansion, oxygen vacancy formation as well as the effect of substituting a univalent cation in place of the divalent cations in the case of B-site substitutions.
Despite the high fluxes obtained for BSCFY, the stability of the material appeared to be reduced. High temperature x-ray diffractions suggested that the formation of a detrimental hexagonal phase was faster for BSCFY compared with BSCF, which was consistent with permeation results which showed a faster decay over time for BSCFY compared with BSCF, even though the initial fluxes were higher for BSCFY. However, a 1500 h permeation test revealed that while some degradation was observed, this was quite reversible after treatment at higher temperatures. Similar work for BSCF resulted in the membrane failure at 400 h, which was unexpected as yttrium doping in principle caused a faster formation of the hexagonal phase. Interestingly, the lattice expansion of BSCFY below 850°C is smaller than BSCF (inferred from oxygen losses), thus suggesting that yttrium was able to counteract the lattice expansion and consequent stresses which result in membrane failure.
Finally, to determine the potential for improvements to BSCFY materials, catalytic surface modification was performed. ECR was used to quantify the improvements to surface exchange kinetics due to the addition of a silver catalyst, while permeation measurements were used to validate improvements. This knowledge was used to develop a high-performance BSCFY hollow fibre through the attachment of a palladium catalyst to the surface, which resulted in very high oxygen fluxes of 14 ml cm-2 min-1 at 900°C, one of the highest fluxes reported in the literature under similar conditions, and significantly above the desired target of 10 ml cm-2 min-1.
Overall, this thesis explored the potential for flux improvement using low concentrations of yttrium in a well-known perovskite, BSCF, raising fundamental questions about how dopants are incorporated into the structure, accompanied by full characterisation to explain the physical effects of yttrium doping and transport phenomena. This work opens up possibilities for the enhancement of other MIEC materials for industrial application through the use of low concentration substitution.