Evolving Beyond the Thermal Age of Separation Processes: Membranes Can Lead the Way William J. Koros School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332 DOI 10.1002/aic.10330 Published online in Wiley InterScience (www.interscience.wiley.com). Introduction I n 2003, the U.S. comprised 4.8% of the world’s population and was responsible for a disproportionate 25% of the world’s energy consumption (DOE, 2004). The industrial sector was responsible for 33% of our energy consumption, followed by transportation (27%), residential (22%) and com- mercial (18%) areas. Within the industrial sector some pro- cesses, such as raw material refining and polymer production intrinsically require high-temperatures to occur economically. Discounting these special cases, a significant fraction of indus- trial processes are currently carried out using thermally-driven methods, simply because installed capital investments provide inertia against change. Electricity generation and separation devices based on non- thermal processes are especially attractive avenues to avoid thermodynamically imposed efficiency limitations on heat uti- lization. Membrane technology, important for nonthermal sep- aration devices, will be the focus of this article. Opportunities offered by linking separation devices and fuel cells will also be considered. Replacing energy-inefficient separation processes requires confronting both materials and processing challenges to more broadly extend benefits available from first generation membranes. Indeed, despite many advantages, membranes have only recently emerged as a realistic platform for use in large scale processes. Since membrane technology is based on deceptively simple fundamentals, early work in this field over- looked the need to integrate four critical capabilities that are discussed later. The realization that membranes require treat- ment as a cross-disciplinary specialty area to allow this inte- gration, has enabled movement of the technology from the laboratory into commercial reality. It is critical to maintain this perspective in order to position membranes to economically handle aggressive feed streams that must be treated to signif- icantly improve the efficiency of global energy use in separa- tions. This improvement is especially important as the world population expands and emerging economies develop. This article also discusses large scale examples where over an order of magnitude reduction in energy use have been achieved by replacing thermally driven approaches with mem- brane processes. As such, this article should be of interest not only to potential creators of membrane devices, but also to those seeking to introduce energy-saving technology tools. Membrane Technology-Four Essential Elements (1) Development of high-efficiency modules with large amounts of area per volume was a necessary first step for the emergence of membranes in large-scale separations. The num- bers are impressive: hollow fiber modules can contain 10,000 m 2 /m 3 of module, which is over 100 times larger than early plate and frame units (Baker, 2004). Such high-efficiency mod- ules provide the needed volumetric productivity to maintain compact system sizes for large scale applications with huge membrane area requirements. (2) Creation of advanced materials with tunable capabilities to separate molecularly similar components has been a second key factor in the emergence of membranes as a broadly appli- cable technology platform. Gels, rigid thermally stable poly- mers, amorphous carbons, ceramics, zeolites, and metals pro- vide a rich array of choices for forming functional high surface area units to perform separations (Kesting, 1985; Pinnau and Freeman, 1999; Pixton and Paul, 1994; Buxbaum, 1993; Langer and Peppas, 1993; Nair and Tsapatsis, 2003, Akin and Lin, 2002). These materials have applications running the gamut from processing of simple gases to complex biorelated feeds. (3) Development of sophisticated capability to control mi- croscopic transport phenomena by tailoring of morphology at multiple levels within a membrane cross-section has been a less obvious third factor in the emergence of membranes. For instance, in the thickness dimension of Figure 1, a submi- crometer ultrathin selective skin region is supported atop low- resistance transition and microporous substrate layers (Carruth- ers et al., 2003). Within such an ultrathin top layer, additional structure with molecularly-selective and reactive features can exist. As is the case in Figure 1, the scale of these critical features are truly molecular in nature and too fine to be imaged, even with the highest resolution microscopy. The detailed functional elements present in a membrane vary greatly depending on the W. J. Koros e-mail address is william.koros@chbe.gatech.edu. © 2004 American Institute of Chemical Engineers Perspective 2326 AIChE Journal October 2004 Vol. 50, No. 10