Solid-State Ionic Devices by Eric D. Wachsman and Enrico Traversa

olid-state ionic devices can be used to harness chemical energy to produce Selectricity in a fuel cell (SOFCs), convert one chemical species to another (both with and without application of an electric potential), separate one chemical species from another, or detect chemical species by producing an electric signal. SOFCs were described in a companion article in this issue, and in this article we focus on some of the other applications of solid-state ion conducting materials. Solid-State Electrolytic Devices

Numerous electrolytic devices are made possible by solid-state ion conducting materials, whereby an applied electric potential is used to drive specific chemical reactions. Examples include high temperature steam electrolysis, upgrading of hydrocarbons, and pollution control (e.g., NOx reduction).1 Among these, ceramic Fig. 1. Faradaic production of O2 from CO/CO2 gas mixtures in a ceramic oxygen oxygen generators (COG) are receiving generator (from Ref. 2). the greatest attention for medical and aerospace applications because they gas separation technology have been ionic transport is a thermally activated can readily produce a pure oxygen gas achieved using membranes, either process and high ionic conductivities, stream from ambient air. porous3 or dense (e.g., Pd alloys for ~1 S/cm, are achieved in the 600- Compact COGs are being developed H separation4), operating under 1000°C temperature range. While to provide a continuous supply of 2 a pressure (total or partial) gradient. these high temperatures create several oxygen-enriched air for people with Among these, ion transport membranes materials issues, they are compatible breathing disorders. Similarly, these (ITM) are receiving a lot of attention with typical industrial gas processing devices can enrich the breathing oxygen due to their unique properties and (e.g., catalytic oxidation and steam concentration for high altitude aircrafts. ability to integrate with other chemical reforming) temperatures resulting in a More recently this technology has processes. In fact, recent advances in more efficient thermal integration. sparked the interest of NASA for space ITM technology present the possibility This advantage in thermal inte- exploration. Because of the distance for a dramatic reduction in the cost of gration provides the opportunity for involved, if we are to travel to other producing pure O , as well as converting ITMs to be used in membrane reactors. planets in the future, we need to utilize 2 petroleum, coal, and even biomass- Membrane reactors can circumvent available planetary resources (in-situ derived feed stocks to H and other thermodynamic equilibrium limit- resource utilization, or ISRU) for life 2 “value added” hydrocarbons. Because of ations by in situ removal of product support and propellant production. An this potential, several major corporations species, thus permitting greater product example is NASA’s plan to send a manned are actively pursuing development of yields.5 Not only do membrane reactors mission to Mars. In order for this to be this technology. result in greater yields, but the product feasible this mission will require the ITMs are based on dense, high- is already purified removing the need utilization of the Mars CO atmosphere 2 temperature, ion-conducting ceramics, for downstream purification processes. to produce the propellant for the return typically perovskite oxides. In the The major functional components of trip and oxygen, both for breathing oxygen-ion conductors, an oxygen-ion these reactors are the reaction catalysts and the fuel oxidant. The technology moves from a filled (O x) to a vacant (e.g., oxidation, steam reforming, and envisioned to make this possible is based O (V ••) oxygen site through a fixed cation water gas shift) and the oxygen-ion on a COG converting CO to O and CO. O 2 2 lattice (M x). Therefore, the actual or proton transport (or O and H Experimental results demonstrating the M 2 2 charged solid-state ionic species involved separation, respectively) membranes, efficacy of this technology are shown in oxygen-ion transport are V ••. In the which are described below. by the Faradaic oxygen production from O proton-conducting oxides the proton is CO /CO gas mixtures2 in Fig. 1. This 2 associated with an oxygen site (OH •) Oxygen-ion transport membranes — technology also has potential for mine O and hops between adjacent oxygen Oxygen-ion transport membranes are safety, military, and emergency response sites. Because oxygen and hydrogen based on metal oxides (such as La applications. 1- are transported through lattice sites ySryCo1-xMxO3 perovskite-type oxides) (vs. porous membrane diffusion), the that exhibit both ionic and electronic Ion Transport Membranes and gas separation selectivity is essentially (mixed) conductivity. For O2 separation Membrane Reactors infinite. This is a dramatic improvement these oxides are typically p-type in selectivity over porous gas separation semiconducting oxides so that a high • Gas separation and purification is membranes that employ size exclusion degree of electron-hole (h ) conduction one of the most important industrial (the difference in size between gas is obtained under oxidizing conditions. chemical processes. Major advances in molecules relative to the pore size) as the Because of their significant electronic mechanism for separation.3 In addition, conductivity, these mixed ionic-

The Electrochemical Society Interface • Winter 2007 37 9 Wachsman and Traversa pumps, electrolyzers, and gas sensors. will tend to oxidize to H2O. A water gas (continued from previous page) BaCe1-xMxO3-type protonic conductors shift catalyst, with the introduction of have sufficient ionic conductivity to additional H2O, in the later part of the obtain comparable flux rates to the reactor could push the conversion to electronic conductors (MIEC) have oxygen-ion conductors. However, they higher H2 yields. an internal electrical short and the have insufficient electronic conductivity. ionic species will selectively permeate The electronic conductivity is Exhaust Sensors through a dense film of the material necessary to balance the transport of under a partial pressure gradient. charge through the material. Further, Undoubtedly, oxygen sensors (λ The potential permeation flux rates it is desirable that the electronic of these materials are extremely high. probes) are the most widespread and conductivity is n-type (e’ conduction) successfully implemented solid state For example, based on the results of so that electronic conductivity is 6 ionic device today (Fig. 3). Comprised Teraoka, et al. and assuming the flux obtained under low P conditions. is bulk diffusion limited, calculated O2 primarily of an oxygen-ion conducting If comparable electronic conduction electrolyte (yttria-stabilized zirconia) O2 flux rates through a 50-µm-thick can be obtained with the BaCe xMxO3- with a Pt electrode on the exhaust membrane of La0.6Sr0.4Co0.8Cu0.2O3 type protonic conductors, they would at 600°C are 22400 L (STP) h-1•m-2 (sensing) side and another Pt electrode be excellent H2 permeation membrane on the air (reference) side, it generates of membrane surface area under an materials, equivalent to palladium alloy 0.21 atm P -gradient. The governing a voltage via the Nernst equation due O2 films, but with the potential advantages equation under these conditions is the to the difference in PO2 between the of higher temperature operation (for two sides. Wagner Equation (described in The thermal integration) and lower cost. Chalkboard article in this issue). By monitoring the exhaust PO2, these Because of the dramatic potential sensors have been used to control the Because of the potential high O2 flux of these materials in H2 production, air-to-fuel ratio (λ) near stoichiometric of these materials under a simple air PO - 2 several groups including Argonne combustion, of essentially every gradient, major industrial gas suppliers National Lab, Georgia Tech, and are developing ITMs to produce pure automobile produced during the last the University of Florida, have been 30 years. While initially deployed O2 from air. By siting one of these ITM developing mixed protonic-electronic units next to a hydrocarbon processing to work with the three-way catalytic membrane materials. Most of the earlier converter to reduce air pollution, they unit, dramatic improvements in the work focused on developing composite conversion efficiency is achieved because also produce the additional benefit of membranes of the proton conducting improved fuel efficiency. The three- the N2 is removed prior to reaction of O2 oxide and a metal phase for electronic with the hydrocarbon feed stream. Due 10 way catalytic converter combined with transport; however, more recent work λ probe control of the air-fuel ratio, to the thermal compatibility of these has shown that by use of a multivalent membranes with, for example, catalytic has been the pioneering technological dopant cation, n-type electronic breakthrough to minimize pollutant oxidation processes, membrane reactors conduction can be achieved in a single are under development to integrate 11-13 emissions from automotive exhaust phase material. gases, simultaneously reducing the the partial oxidation of hydrocarbons, Ultimately these two MIEC such as natural gas, to syngas using concentrations of both oxidizing (CO membranes can be combined in series, and HCs) and reducing (NO ) exhaust an oxygen-ion transport membrane, as and integrated with partial oxidation, x shown in Fig. 2. The integrated process pollutants. steam reforming, and water gas shift However, regulations for vehicle further improves efficiency and reduces catalysts, also in series, to optimize cost. pollutant emission are becoming the thermodynamics of hydrogen pro- increasingly stringent worldwide, duction. The first membrane separates Proton transport membranes — A series and proposed limits cannot be met O2 from air (with infinite selectivity) using conventional three-way catalytic of perovskite-type oxides (e.g., BaCe1-x and reacts any of the hydrocarbons in converter and λ probe control. For MxO3, where M is a metal acceptor the feed to form CO and H , facilitated dopant) have been shown to have 2 instance, recent regulations require by an oxidation catalyst. The second On-Board Diagnostic (OBD) systems to high proton conductivity at elevated membrane separates the H (also with 7,8 2 control exhaust pollutants. The current temperatures. These materials are infinite selectivity) providing a pure H receiving considerable attention 2 OBD systems monitor the converter gas stream. As the process gas stream efficiency through the measurement because of their numerous applications continues through the reactor, any as H+ electrolytes in fuel cells, hydrogen of the oxygen content up-stream and hydrogen that has not been separated down-stream of the converter using two λ probes but does not directly monitor the actual pollutant concentrations.14 Syngas In contrast, emission control and fuel Natural Gas & Steam efficiency could be further improved CH4 H2 by deployment of sensors capable of Oxygen Oxygen H O CO directly detecting CO/HCs and NO in 2 vacancy x anion the exhaust gas stream, in addition to O2, to provide improved feedback for automotive engine control. ITM 2- o - Me CH4 + O CO + 2H2 + 2e A recent strategy to improve fuel mb ran efficiency (and thus reduce overall e 2- - O 2e emissions) is to operate the engine under lean-burn combustion. However, - 2- ½O2 + 2e o O the high oxygen concentration hinders the NOx reduction reaction in the catalytic converter, making necessary Metal the implementation of a transient NOx Air , Low Pressure Spent Air cation storage process. When the engine is operated in lean conditions CO and Nitrogen Oxygen HCs are rapidly oxidized and NOx is adsorbed on the catalyst surface. Then the engine operation is shifted to rich- Fig. 2. ITM Syngas technology being developed by Air Products.

38 The Electrochemical Society Interface • Winter 2007 microstructure of the oxide electrode, in terms of morphology and grain size, can significantly affect the sensor response,28 and therefore care should be paid to the electrode fabrication process. Most sensors have been tested in the laboratory, though some reports show tests performed in more practical conditions, such as in real engine exhaust environments.29,30 Solid-state sensors for CO, HCs, and NOx have been reviewed in recent papers.31-33 An example of recent literature results is that of a Pt/YSZ/ Nb2O5 sensor, using Nb2O5 and Pt electrodes on tape-cast YSZ.34 The EMF response of this sensor to different concentrations of propylene in air, in the temperature range 500-700°C, is shown in Fig. 4. As is typical for these types of sensors, increasing the temperature results in a more stable and faster response, but the amplitude of the EMF decreases. Under exposure to 1000 ppm of propylene in air, the EMF value of Pt/YSZ/Nb2O5 at 700°C was –110 mV, still a remarkably large value for that Fig. 3. Cross-section of an oxygen sensor. high of a temperature. In addition, the results for the Pt/YSZ/Nb2O5 sensor were found reproducible and stable: burn conditions for a few seconds reference air and the oxide electrode after several days of measurements, and the adsorbed NO is reacted with to the exhaust gas environment, x the same EMF results were obtained. the temporarily higher CO and HC while other groups are exposing both The cross sensitivity was checked by concentrations. After this catalyst electrodes to the atmosphere to be performing the measurements under regeneration, the engine is returned to detected. A large number of oxides exposure to different saturated and lean-burn operation. Proper operation are being tested either for reducing unsaturated hydrocarbons, CO and NO of this method requires the availability or for oxidizing gas detection. The 2 in the temperature range 500-700°C. At of a sensor to measure NO conversion selection of the oxide is guided by x all the investigated temperatures, the as well as slip from the spent catalyst.15 its sensitivity and capability to detect sensor responded best to propylene. For Development of solid-state sensors, very low concentration, selectivity all the gases, a linear dependence of similar to the λ probe, that can directly towards specific gasses, stability the EMF response was observed. These measure CO/HCs and NO , has been at high temperatures, and response x results indicate that this sensor is a an area of tremendous research time. It has been also found that the effort. However, in contrast to the λ probe these sensors are based on a non-Nernstian potential mechanism, typically with both electrodes in the same gas stream. They are comprised of the same YSZ electrolyte and a metallic electrode (e.g., Pt), but the other electrode is typically an oxide electrode and the difference in potential generated between the two electrodes results in a non-Nernstian potential that is sensitive to the gaseous species of interest (CO, HCs, or NOx). These sensors seem to be the most promising to meet all of the requirements needed for exhaust gas monitoring.16 From a manufacturing point of view, these sensors are only a modification of the λ probe, by substitution of one of the Pt electrodes with an oxide, and can therefore easily be implemented into mass production and practical use. Indeed, several groups worldwide including Kyushu University, Los Alamos National Lab, the University of Rome “Tor Vergata,” and the University of Florida,17-27 have been developing this type of sensors, both for NOx and CO/HC detection, though with slightly different strategies: namely, some of the groups are developing sensors where Fig. 4. EMF response versus time of Pt/YSZ/Nb2O5 sensor to different concentrations of propylene in the metallic electrode is exposed to air in the temperature range 500-700°C (from Ref. 34).

The Electrochemical Society Interface • Winter 2007 39 Wachsman and Traversa L. Brosha, Solid State Ionics, 136-137, separation membranes, and solid (continued from previous page) 633 (2000). state sensors. He is a Fellow of The 17. F. H. Garzon, R. Mukundan, R. Lujan, Electrochemical Society and an Editor and E. L. Brosha, Solid State Ionics, of Ionics. He may be reached at ewach@ potential candidate for monitoring and 175, 487 (2004). mse.ufl.edu. controlling HCs in automotive exhaust 18. A. Dutta, N. Kaabbuathong, M. gases. L. Grilli, E. Di Bartolomeo, and E. Enrico Traversa joined the University Traversa, J. Electrochem. Soc., 150, of Rome “Tor Vergata” in 1988 where Summary H33 (2003). he is currently Professor of Materials 19. E. Di Bartolomeo, M. L Grilli, and Science and Technology, Director of Tremendous progress has been made in E. Traversa, J. Electrochem. Soc., 151, the Doctorate Course in “Materials for understanding the defect chemistry and H133 (2004). Environment and Energy.” He is Sr. transport in solid-state ion conductors. 20. E. Di Bartolomeo, N. Kaabbuathong, Vice-Chair of the ECS High Temperature These materials are attracting increasing M. L. Grilli and E. Traversa, Solid State Materials Division. Dr. Traversa’s attention for a variety of applications, Ionics, 171, 173 (2004). research interests are in nanostructured most notable among these being fuel 21. M. Nakatou and N. Miura, Sens. materials for environment, energy, and cells for power generation, production Actuators B, 120, 57 (2006). health, with special attention to fuel of chemicals, oxygen generation for life 22. N. Miura, T. Koga, M. Nakatou, cells and chemical sensors. He is author support systems, and high temperature P. Elumalai, and M. Hasei, J. or co-author of seven patents, more electrochemical sensors. Oxygen Electroceram., 17, 979 (2006). than 400 scientific papers (about 200 of sensors are already a widespread 23. C. O. Park and N. Miura, Sens. them published in refereed international commercial success and many of the Actuators B, 113, 316 (2006). journals), and has edited ten books other solid-state ionic devices hold 24. J. Yoo, F. M. Van Assche, and E. D. or special issues of journals. He was similar promise as their performance Wachsman, J. Electrochem. Soc., 153, elected in 2007 to the World Academy is improved and cost is brought down H115 (2006). of Ceramics. He may be reached at through further technical advances and 25. J. Yoo, H. Yoon, and E. D. Wachsman, [email protected]. mass production. J. Electrochem. Soc., 153, H217 (2006). References 26. J. Yoo, S. Chatterjee, F. M. Van Assche, and E. D. Wachsman, J. Electrochem. Soc., 154, J190 (2007). 1. E. D. Wachsman, P. Jayaweera, G. 27. L. Chevallier, E. Di Bartolomeo, M. Krishnan, and A. Sanjurjo, Solid State L. Grilli, M. Mainas, B. White, E. D. Ionics, 136-137, 775 (2000). Wachsman, and E. Traversa, Sens. 2. J. Y. Park and E. D. Wachsman, J. Actuators B, in press (2007). Electrochem. Soc., 152, A1654 (2005). 28. E. Di Bartolomeo, N. Kaabbuathong, 3. M. C. Duke, J. C. D. da Costa, D. D. A. D’Epifanio, M. L. Grilli, E. Do, P. G. Gray, and G. Q. Lu, Adv. Traversa, H. Aono, and Y. Sadaoka, J. Funct. Mater., 16, 1215 (2006). Eur. Ceram. Soc., 24, 1187 (2004). 4. S. Wieland, T. Melin, and A. Lamm, 29. E. Di Bartolomeo, M. L. Grilli, Chem. Eng. Sci., 57, 1571 (2002). N. Antonias, S. Cordiner, and E. 5. E. Kikuchi, Catal. Today, 56, 97 Traversa, Sensor Letters, 3, 22 (2005). (2000). 30. M. L. Grilli, E. Di Bartolomeo, A. 6. Y. Teraoka, T. Nobunaga, K. Okamoto, Lunardi, L. Chevallier, S. Cordiner, N. Miura, and N. Yamazoe, Solid State and E. Traversa, Sens. Actuators B, Ionics, 48, 207 (1991). 108, 319 (2005). 7. H. Iwahara, H. Uchida, K. Ono, and 31. S. Zhuiykov and N. Miura, Sens. K. Ogaki, J. Electrochem. Soc., 135, Actuators B, 121, 639 (2007). 529 (1988). 32. J. W. Fergus, Sens. Actuators B, 121, 8. N. Bonanos, B. Ellis, and M. N. 652 (2007). Mahmood, Solid State Ionics, 44, 305 33. J. W. Fergus, Sens. Actuators B, 122, (1991). 683 (2007). 9. H. Iwahara, Y. Asakura, K. Katahira, 34. L. Chevallier, E. Di Bartolomeo, M. L. and M. Tanaka, Solid State Ionics, 168, Grilli, and E. Traversa, Sens. Actuators 229 (2004). B, in press (2007). 10. G. Zhang, S. E. Dorris, U. Balachandran, and M. Liu, Solid State Ionics, 159, 121 (2003). About the Authors 11. S. J. Song, E. D. Wachsman, S. E. Dorris, and U. Balachandran, Solid Eric D. Wachsman is the Past Chair State Ionics, 164, 107 (2003). of the High Temperature Materials 12. S. J. Song, E. D. Wachsman, S. E. Division, Director of the Florida Dorris, and U. Balachandran, J. Institute for Sustainable Energy, Director Electrochem. Soc., 150, A1484 (2003). of the UF-DOE High Temperature 13. S. J. Song, E. D. Wachsman, S. E. Electrochemistry Center, and an UF- Dorris, and U. Balachandran, Solid Research Foundation Professor of State Ionics, 167, 99 (2004). Materials Science and Engineering at the 14. J. Riegel, H. Neumann, and H.-M. University of Florida. Dr. Wachsman’s Wiedenmann, Solid State Ionics, 152, research interests are in ionic and 783 (2002). electronic conducting ceramics, from 15. R. Moos, Int. J. Appl. Ceram. Technol., fundamental investigations of their 2, 401 (2005). transport properties and heterogeneous 16. F. H. Garzon, R. Mukundan, and E. electrocatalytic activity, to the development of moderate temperature solid oxide fuel cells (SOFC), gas

40 The Electrochemical Society Interface • Winter 2007