Research and development programs worldwide have increasingly focused on developing and improving technologies for hydrogen production and separation. Hydrogen can be purified by pressure swing adsorption, cryogenic distillation, or membrane separation. Membrane-separation processes require far-less energy, and are capable of operating continuously.

For non-galvanically driven operation, the membrane material must conduct both protons and electrons to maintain local charge neutrality within the membrane. The most-developed hydrogen separation membranes are noble metal-based membranes such as palladium and palladium alloys. However, the expense and low availability of palladium makes palladium-based membranes prohibitively costly for large–scale commercial applications. Ceramic-based membranes represent an attractive alternative. To date, no ceramic material exhibits sufficient electronic and protonic conductivity to enable industrially relevant hydrogen fluxes. To increase hydrogen flux and improve performance, researchers have developed composite membranes, where one phase serves as the electronic conductor, while the other phase serves as the proton conductor.  Composite ceramics include BCY-YDC, where BaCe_0.8Y_0.2O_3-d (BCY20) provides high proton conduction and Ce_0.8Y_0.2O_2-d (YDC20) provides electronic conduction [1]. Other ceramic-ceramic composites are also being developed [2].

The composite ceramics face unique challenges. Spinodal decomposition of ceria was observed after operation of only a couple of days at 900 °C [1]. Additionally, the presence of oxide-ion conduction in the ceria phase favors water splitting on the sweep side of the membrane [2]. As a result, the measured hydrogen fluxes are the sum of the hydrogen generated by the water splitting and by the hydrogen transport through the membrane. This can reduce the driving force for hydrogen separation, decreasing membrane performance.

Replacing the electronic phase with a more-stable material with limited (or ideally non-existent) oxide-ion conduction is necessary to improve performance. Metals could be good candidates. Ceramic-metal composite materials with nickel have been successfully prepared [3], but it is well known that Ni catalyzes carbon formation in hydrocarbon-containing atmospheres. Copper is more stable than Ni and appears to be a more-promising material, though its low melting temperature makes it challenging for fabrication.

BCY20-Cu cermet membranes were prepared by infiltration of copper through a porous BCY20 backbone [4]. In this work we present the infiltration of Cu into a BCY20 backbone prepared by a novel freeze-drying tape-casting technique. This method has recently been developed to form hierarchically oriented channels/pores in SOFC electrodes [5]. Figures 1a and b are respectively secondary electron and back-scattered electron images of a freeze-dried, tape-casted BCY20 after lamination and sintering. Large columnar pores are observed. A polished cross-section of a similar sample after Cu infiltration is shown in Figure 1c. The fluxes of these dense BCY20-Cu ceramic-metal composites were measured and compared to state-of-the-art membranes, with favorable results obtained.

[1] W. Rosensteel, S. Ricote, N.P. Sullivan, accepted in Intern. J. Hydrogen Energy 2015

[2]  E. Rebollo, C. Mortalò, S. Escolástico, S. Boldrini, S.Barison, J. M. Serra, M. Fabrizio, accepted in Energy & Environmental Science 2015

[3] C. Zuo, T.H. Lee, S.E. Dorris, U. Balachandran, M. Liu, J. Power Sources 159 (2006) 1291–1295

[4] W. Rosensteel, Ph.D dissertation, Colorado School of Mines, 2015.

[5] Y. Chen, Y.  Zhang, J. Baker, P.  Majumdar, Z. Yang, M. Han, F. Chen, ACS Appl. Mater. Interfaces 6 (2014) 5130−5136