Scale-Up of Molecular Sieve Silica Membranes for Reformate Purification M. Duke, V. Rudolph, G. Q. (Max) Lu, and J. C. Diniz da Costa The Nanomaterials Centre, School of Engineering, The University of Queensland, Brisbane 4072, Australia DOI 10.1002/aic.10200 Published online in Wiley InterScience (www.interscience.wiley.com). Keywords: molecular sieve silica, nanotechnology, reformate, scale-up, fuel cells, membrane reactors Introduction Molecular sieve silica (MSS) membranes are a developing technology involving nanotechnology techniques. These mem- branes demonstrate excellent properties for permeation of hy- drogen molecules with a very high selectivity, as presented in Table 1 (de Vos and Verweij, 1998; Diniz de Costa et al., 2002; Kusakabe et al., 1999; Tsai et al., 2000), making them ideal for applications requiring separation of hydrogen from other gases. The permeation of hydrogen is now approaching the limit of 1  10 -6 mol m -2 s -1 Pa -1 , equivalent to 0.7 kg of hydrogen per square meter of membrane area per hour, and at 1 atmo- sphere differential pressure across the membrane. MSS mem- branes can operate at temperatures up to 500°C and up to 20 atmospheres. However, there are limited reports in the litera- ture regarding scale-up of silica-based membranes. To cite but one example, Koukou et al. (1999) scaled-up alpha-alumina tubes (length: 20 cm; outer diameter: 1.4 cm) with silica- derived films for gas separation. They tested their membranes using a feed-gas mixture of 62% H 2 and 38% CH 4 for pressure differences ranging from 5 to 20 bar, resulting in best H 2 /CH 4 selectivities of 5. Potential applications of the MSS membranes include gas separation, such as extracting H 2 from a gas stream in a petrochemical refinery process; filtration of contaminants (such as CO) from H 2 supply for proton-exchange membrane (PEM) fuel cells; and chemical production using membrane reactors to carry out dehydrogenation. MSS membranes are envisioned to be integrated into PEM fuel cell systems, which is a technology being developed for motor vehicles (50 –100 kW), residential (2–10 kW) and com- mercial (250 –500 kW) power generation, as well as small/ portable generators and battery replacement (ECW, 2000). PEM fuel cells usually require a pure H 2 source for operation. However, for most of these applications, tank storage of H 2 at extremely high pressure (up to 35 MPa) presents serious con- cerns with respect to safety. Hydrogen is thus typically ob- tained by reforming a hydrocarbon fuel (methanol or gasoline) onboard. In this reaction, the product termed as “reformate” consists of H 2 , CO, CO 2 , and water. CO is detrimental for the fuel cell operation because of catalyst poisoning and severe fuel cell degradation. Thus, CO levels must be reduced to 10 parts per million (ppm) for standard anodes or 100 ppm for CO-tolerant anodes in fuel cell technology (Hasegawa et al., 2002; Sotawa et al., 2002). Metal membranes based on palla- dium and its alloys can provide high H 2 purification and various research groups have reported their use in fuel cell systems (Koros and Mahajan, 2000; Ledjeff-Hey et al., 1998; Lin and Rei, 2000). However, palladium membranes are a major economic impediment in fuel cell technology, particu- larly because they cost in the order of US$1,250 to US$2,000 per square meter of membrane area (Illgen et al., 2001). Thus, there is a major advantage of using silica-derived membranes because such materials are cheap and readily available. Given that the cost effectiveness of technologies such as fuel cells and membrane reactors is paramount to making these technologies competitive against traditional technologies (Ogden et al., 2001), MSS membranes can potentially deliver cost savings over use of metal membranes. Much of the current MSS membrane research has been limited to laboratory testing. Operating systems in fuel cells, membrane reactors, and gas separation undergo cycles of start-up to shutdown with varying compositions and heating rates. The development of inorganic membranes displaying outstanding chemical and thermal resistance is required to open new fields of industrial applications (Jonquieres et al., 2002). Although laboratory testing has shown the potential of MSS membranes, there is a need to scale-up to test how the tech- nology can perform under as close as possible to real applica- tion conditions. Thus, in this work we scaled-up MSS mem- brane tubes for fuel cell reformate purification. Of particular Correspondence concerning this article should be addressed to J. C. Diniz da Costa at joedac@cheque.uq.edu.au. © 2004 American Institute of Chemical Engineers 2630 AIChE Journal October 2004 Vol. 50, No. 10