Frederic Mertins

Ph.D. thesis

Thesis title:

Perovskite-type ceramic membranes - Partial oxidation of methane in a catalytic membrane reactor

[thesis in pdf format]




Prof. Dr. Ing. D.H.A. Blank

Assistant promotor:

Dr. H.J.M. Bouwmeester


The application of mixed ionic and electronic conductors as oxygen separating membranes offer an attractive alternative for the production of synthesis gas from methane when compared with traditional reforming. Materials with the perovskite structure are the most promising candidates thanks to the ease of tuning their electro-chemical properties with appropriate doping. A number of issues have to be addressed before any industrial breakthrough. The features of interest when using these materials in membrane reactor technology are the magnitude of the oxygen flux (proportional to the oxygen vacancies concentration) and the thermodynamical stability.

Chapter 2, the ionic conductivity of a series of materials with composition La1-xSrxCoO3- with 0.1 x 0.7 was determined via oxygen permeation measurements as a function of the oxygen partial pressure. It was shown that the ionic conductivity in this series of materials varies almost linearly with the concentration of oxygen vacancies. For small concentration of strontium (x = 0.1) the expected values are in good agreement with the theory. At higher doping levels, local discrepancies are found. It is suggested that the presence of a stagnant layer on the permeate side of the membrane is the cause for the local disagreement between the experimental and the expected values. The influence of oxygen vacancies interactions is thought to be significant at lower temperatures.

Chapter 3, different materials with a perovskite composition were tested for syngas generation. The influence of a number of dopants on both the A- and B-site of the ABO3 perovskite structure was investigated. Substitution with barium and chromium were found to limit the magnitude of the oxygen flux. Some materials exhibit a high oxygen flux but the critical issue of their thermodynamic stability make them unsuitable for industrial applications. The best potential candidate found in this study, namely La0.7Sr0.3Fe0.7Ga0.3O3-, exhibits a lower oxygen flux than the best performing materials but its stability under the harsh conditions associated with CPO is better.

The stability of this material was investigated further in
Chapter 4. X-ray diffraction revealed a second phase after synthesis. This second phase seems to be unstable when placed in an oxygen chemical potential gradient as it disappears from the near surface on the low-PO2 side of the membrane after the CPO experiment. X-ray photoelectron spectroscopy combined with X-ray fluorescence showed the enrichment of strontium on the methane side of the membrane and scanning electron microscopy revealed the presence of a dusty-like layer on the surface exposed to the methane stream. An oxygen flux of 5.1 was obtained at 900 C for a membrane thickness of 0.5 mm with a pure methane feed. Evidence that the oxygen transport is partially controlled by the surface reactions was found.

Chapter 5, dense tubular membranes were produced by centrifugal casting of an aqueous suspension, containing powder particles of the mixed-conducting perovskite La0.5Sr0.5CoO3- and a dispersant. The resulting green bodies were dried and sintered to produce tubes with a maximum length of 12 cm, having a relative density higher than 92 %. The particle morphology, the amount of dispersant and its burnout appeared to be of influence to the quality of the final product. Oxygen permeation measurements were conducted in the temperature range 850-950 C in air/He gradients. Results are found to be consistent with data reported for flat membranes.

Chapter 6, a dense bi-layered membrane concept is proposed as solution to the problem of membrane stability under reducing conditions of operation. A dense thin film of La0.8Sr0.2FeO3- was successfully grown with pulse laser deposition (PLD) on a dense support of La0.3Sr0.7CoO3- and this system was chosen as a proof-of-principle. Under air/He gradients, the oxygen flux of the bi-layered membrane is half of that obtained with an unmodified La0.3Sr0.7CoO3- dense support. Under CPO conditions, the oxygen flux is almost equivalent to that of a 0.5 mm thick La0.8Sr0.2FeO3- membrane. This example illustrates the fact that by fine-tuning the compositions of the support and the thin film and the relative thickness of both layers, it is possible to protect a high-flux membrane, otherwise non-stable against the severe conditions encountered during syngas generation. It also proves that high oxygen fluxes are achievable.