Motivation and introduction

[Back to the project]

Desulfurized natural gas fuel, with the main component CH4, can be either "externally" reformed with water towards syngas CO + H2 in a separate reforming unit located upstream of the SOFC anode or reformed "internally" using a suitable catalytically active anode material. The relatively high operating temperatures of SOFC's allow to feed hydrocarbons directly to the anode without external reforming. By this route, the CO and H2 molecules are produced close to the anode catalyst sites where they finally become electro-oxidized toward H2O and CO2. Efficient internal reforming SOFC's are desirable because they represent less costly and complex systems, with the additional advantage that the waste heat released from the electrode can be used to drive the endothermal methane steam refoming reaction directly. Yet, a number of basic and technological problems has to be solved to obtain sufficiently stable and efficient electrodes, stimulating a variety of applied and fundamental research activities during the last years [1]. One of the obstacles being particularily interesting for heterogeneous catalysis research of reforming reactions, is the high propensity of commercial Ni cermet anodes for carbon deposit/carbon filament formation when directly exposed to a hydrocarbon-rich fuel gas [2-4]. The growing filaments have been shown to disperse the originally well-percolated Ni structures into "dust" [5] and to induce mechanical stress which may lead to electrode fracture [6]. These detrimental effects can be suppressed by steam co-feed, whereby reversal of methane-induced carburization is less demanding than removal of carbon induced by higher hydrocarbons, requiring higher water partial pressures. Therefore, and because of the fundamental importance of methane activation in basic catalysis research, methane will be our preferential carbon source, and methane reforming among our preferential test reactions.

 

Since the mechanism for carbon fiber formation on Ni involves deposition of carbon on the Ni surface followed by dissolution into the bulk, the subsequent precipitation of carbon e.g. as a fiber [4] is a complex process which can only be understood on the basis of a broad knowledge of the thermodynamic and kinetic preconditions governing the carbon bulk and surface concentrations and the kinetic rates of carbon dissolution/re-segregation and of reactive removal by water vapour. The related experiments, performed both on bulk metal samples and nanoparticles, represent the first major focus of our sub-project. Due to the high carbon bulk solubility and diffusion rate in Ni it is of paramount importance to take a look into the "third catalytic dimension", i.e. the subsurface- and bulk regions of both Ni nanoparticles and extended bulk samples, e.g. by various available diffraction- and spectroscopic techniques. Even superior is to study both surface- and bulk-related phenomena on bulk Ni and other metallic samples tsimultaneously by in-situ techniques under realistic catalytic conditions, preferentially by ambient-pressure in-situ XPS spectroscopy using tunable synchrotron radiation for probe depth variation. A thorough in-situ study of the "(re)active" Ni-carbon system should account both for the processes of carbon loading (in clean CH4 or steam reforming gas mixtures) and carbon depletion by reaction of carbon with steam toward CO and H2. In a series of in-situ XPS- and MB studies performed by our group, analogous carbon dissolution, re-segregation and reactivity phenomena were studied successfully both on Pd(111) and Pd foil samples under ambient-pressure reaction conditions [7-9].

 

The next logical step is to increase the model system complexity toward the real Ni/ZrO2 and Ni/YSZ systems, which exhibit a large interface between oxide and metal. This can be achieved by our directional model catalyst approach, which proved to be highly valuable for simulation of catalytic metal-oxide phase boundary effects [10]. Model phase boundaries will be either prepared by deposition of sub-monolayers of ZrO2, CeO2 or YSZ (the latter also via an additional EB-PVD route [11]) on top of metal substrates or vice versa by embedding Ni metal particles in a thin film of supporting ZrO2, CeO2 or YSZ oxide. Phase boundary effects can in principle be expected to influence methane reforming activity/selectivity, but also the efficiency of methane- and water activation and of carbon removal. Moreover, the efficiency of reductive activation of lattice oxygen at the Ni-YSZ interface is of paramount importance for the triple phase boundary electrooxidation of CO and H2 toward CO2 and H2O. Kinetic studies in our UHV-compatible high-pressure reaction cells will tackle the salient catalytic/kinetic effects induced by the presence of the phase boundary sites, allowing us to identify interesting phenomena to be eventually studied by advanced in-situ techniques.

 

Following a logical approach to control carburization and filament formation, but also to keep up good (electro)catalytic efficiency, our next research focus will be on bimetallic anode model systems, and on the basic principles counteracting carbon formation/dissolution and re-segregation. So far, anodes from materials that do not dissolve major amounts of C, such as conductive ceramics and metallic Cu (but also Ag and Au) have been tested for SOFC anode cermet applications with promising success [1]. Unfortunately, these materials exhibit different shortcomings: the ceramics lack of sufficient electronic conductivity, the copper metals (Cu, Ag, Au) exhibit very low thermal sintering stability and catalytic activity [12, 13]. On the other hand, metals such as Ni, Co, Pd and Ru, despite of their adverse carbon-dissolution properties [4, 14], are simply the better (electro)catalysts, and therefore interesting approaches toward bimetallic systems containing e.g. Cu in combination with e.g. Ni or Co have been reported in the more applied research literature [1], allowing for enhanced sintering stability (e.g. up to 1173 K for Cu-Co), better catalytic perfomance and less coking/filament formation than the pure metals. A bimetallic approach involving Au doped Ni single crystal model studies was highly successful for the optimization of industrial methane steam reforming [15].

 

We thus plan to study models of "isolated" bimetallic systems (layered, alloyed etc.) at first, again focussing on C dissolution/re-segregation and H2O + C reaction. If such alloy systems are capable of largely blocking C bulk diffusion, some graphitic carbon deposits will likely form rather at the surface [3] but such deposits are by far not as dangerous as C-fibers. Moreover it has been shown that C surface deposits can be controlled through additional catalytic coatings, such as ceria, which efficiently catalyzes their oxidation by steam [3]. This opens up an additional perspective for model studies, since now interfaces between ceria and the respective metallic/bimetallic components become interesting targets for studies on metal-support interaction, beyond the also envisioned bimetal - ZrO2 and bimetal - YSZ systems. The maximum complexity is finally represented by triple component models composed of (bi)metal, ceria and YSZ/zirconia. Cu–ceria–YSZ composite anodes were found to exhibit good performance in utilizing different hydrocarbon fuels up to C6 (even without steam co-feed) already at ~700 °C [16], and bimetallic systems are likely to further improve anode efficiency, power density and thermal stability, since pure Cu is prone to sintering above 1073 K [12, 13] and a poor hydrocarbon activator. Although Cu is particularly efficient in methanol activation and the water gas shift reaction, the role of hydrocarbon activation is likely played by the ceria co-catalyst.

 

In a model approach using functional layers of impregnated ceria, Vohs et al. tried to disentangle the problems of electronic conduction and catalytic activity [1] (the opposite requirement applies to perovskite-type anode systems [17]) by preparing anodes composed of a thin impregnated ceria-YSZ or Pd-ceria-YSZ functional layer responsible for improved (electro)catalytic activity in between the YSZ electrolyte and a thicker, porous conduction layer. The presence of a noble metal catalyst such as Pd is especially useful for operating the cells in dry CH4 without water vapour. The Pd-doped ceria functional layer improved the attainable cell power density by almost a factor of 40, at anode impedances close to those obtained with clean H2, and open-circuit voltages even higher than with pre-reformed CO/H2 fuel. The only explanation is that methane activation is a highly efficient process at the involved metal-oxide and/or oxide-oxide interfaces, and we regard this important result as an ideal starting point for our model catalyst approach.

 

References

[1] M. D. Gross, J. M. Vohs, R. J. Gorte, J.; Mater. Chem. 17 (2007) 3071 and references therein

[2] S. McIntosh and R. J. Gorte; Chem. Rev., 104 (2004) 4845.

[3] T. Kim, G. Liu, M. Boaro, S.-I. Lee, J. M. Vohs, R. J. Gorte, O. H. Al-Madhi, B. O. Dabbousi; J. Power Sources, 155 (2006) 231.

[4] M. L. Toebes, J. H. Bitter, A. J. van Dillen, K. P. de Jong; Catal. Today, 76 (2002) 33

[5] C. H. Toh, P. R. Munroe, D. J. Young, K. Foger; Mater. High Temp., 20 (2003) 129

[6] H. Kim, C. Lu, W. L. Worrell, J. M. Vohs, R. J. Gorte, J.; Electrochem. Soc., 149 (2002) A247.

[7] H. Gabasch, K. Hayek, B. Klötzer, A. Knop-Gericke, R. Schlögl, J.; Phys. Chem. B 110 (2006) 4947-4952

[8] H. Gabasch, E. Kleimenov, D. Teschner, S. Zafeiratos, M. Hävecker, A. Knop-Gericke, R. Schlögl, D. Zemlyanov, B. Aszalos-Kiss, K. Hayek, B. Klötzer; J. Catal. 242 (2006) 340-348.

[9] H. Gabasch, A. Knop-Gericke, R. Schlögl, W. Unterberger, K. Hayek, B. Klötzer; Catal. Lett. 119 (2007) 191-198

[10] K. Hayek, M. Fuchs, B. Klötzer, W. Reichl, G. Rupprechter; Top. Catal. 13 (2000) 55

[11] G. Laukaitis, J. Dudonis, D. Milcius; Vacuum 81 (2007) 1288–1291

[12] S. Jung, C. Lu, H. He, K. Ahn, R. J. Gorte, J. M. Vohs; J. Power Sources 154 (2006) 42

[13] M. D. Gross, J. M. Vohs, R. J. Gorte; Electrochim. Acta 52 (2007) 1951

[14] R. T. K. Baker, J. J. Chuldzinski, Jr.; J. Phys. Chem., 90 (1986) 4734

[15] F. Besenbacher, I. Chorkendorff, B. S. Clausen, B. Hammer, A. M. Molenbroek, J. K. Norskov, I. Stensgaard; Science 279 (1998) 1913

[16] H. Kim, S. Park, J. M. Vohs, R. J. Gorte; J. Electrochem. Soc. 148 (2001) A693

[17] R. Mukundan, E. L. Brosha, F. H. Garzon; Electrochem. Solid-State Lett. 7 (2004) A5