Solid oxide fuel cells (SOFCs) are electrochemical reactors that directly convert the chemical energy of a fuel gas into electrical energy in an environment-friendly way with high efficiency and fuel flexibility. Although the commonly used yttria stabilized zirconia (Ni/YSZ) cermet anodes in SOFCs show excellent electrocatalytic properties in H2 fuel, they suffer from carbon deposition and sulfur poisoning problems when hydrocarbon fuels are used [1-4]. Moreover, Ni agglomeration in the cermet anodes make them insufficiently stable for long term running. In recent years, significant efforts have been devoted to developing alternative anodes to alleviate carbon deposition, sulfur poisoning and volume instability problems encountered in Ni-based cermet anodes [5-9]. Recently, Ishihara et. al. [10] reported a composite of Ce0.6Mn0.3Fe0.1O2 and La0.6Sr0.4Fe0.9Mn0.1O3 as an anode for direct hydrocarbon SOFCs, which showed a high peak power density of 1 W/cm2 at 800 °C when propane and butane were used as fuel. Unfortunately, this oxide anode has low catalytic activity for methane due to the stable structure and strong C-H bond in methane. Nickel-based catalysts have been widely used for hydrocarbon reforming with high catalytic activity [11, 12]. It is possible to produce high performance SOFCs with nanostructured electrodes by a wet infiltration method [13, 14]. However, the agglomeration and enlargement of the nanoparticles inevitably cause performance degradation in long term operation. It is well known that nano-size phases can be grown controllably from perovskites by a judicious choice of composition [15]. Compared with the conventional tri-layer cell structure, symmetrical SOFCs have attracted increasing attention due to the improved thermo-mechanical compatibility of the electrolyte and electrodes, which reduces fabrication cost, and enha nces resistance to coking and sulfur poisoning [16, 17].
Herein, we report the rational design of a highly active Ni doped La0.6Sr0.4FeO3 electrode (LSFN) for symmetrical hydrocarbon fuel SOFCs. Homogeneously dispersed nano-sized Ni-Fe alloy particles were in situ segregated onto the LSFN surface during the fuel cell operation. The symmetrical cell with impregnated LSFN electrode showed good stability in long term testing.
La0.6Sr0.4Fe0.9Ni0.1O3-δ (LSFN) and La0.6Sr0.4FeO3-δ (LSF) were synthesized by a solid state reaction with commercial La2O3 (99.99%), SrCO3 (99%), Fe2O3 (98.5%) and NiO (99.99%). La2O3 was pre-calcined at 1000 °C for 2 h before use. Stoichiometric proportions of La2O3, SrCO3, Fe2O3 and NiO were mixed and ground thoroughly in ethanol until the ethanol was volatilized. The powder was calcined at 800 °C for 2 h in air. Then, it was ball milled again and calcined at 1300 °C for 4 h in air.
The size and morphology of the products were examined by scanning electron microscope (SEM, Hitachi-S4800) and transmission electron microscope (TEM). TEM images and X-ray energy-dispersive spectrometry (EDS) measurements were carried out on a JEOL ARM200F microscope operated at 200 kV. The composition of the samples was analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES, Thermo Electron Corporation).
The phases and purity of the samples were examined by X-ray powder diffraction (XRD) performed on a Bruker D8-Advance diffractometer with Cu Ka radiation (l = 1.54 ). Gas chromatography-mass spectrometer (GC-MS) analysis was carried out on a Shimadzu QP2010 Plus instrument.
Tri-layer porous|dense|porous YSZ-electrolyte monolithic cells were fabricated by laminating one dense YSZ tape and two porous YSZ tapes on both sides with subsequent co-firing at 1350 °C. The porous layers were filled with 45 wt% starch as the fugitive material, yielding porosities of 45% as measured using the Archimedes’ method. Symmetrical cells were fabricated by impregnating the LSFN or LSF precursor solutions onto both porous layers and calcined at 500 °C, and then at 800 °C. The YSZ skeleton infiltrated with the catalyst were weighed before and after each impregnation-calcination cycle to estimate the loading of the impregnated LSFN or LSF catalyst, which were expressed as the mass ratio of the impregnated oxides to the porous YSZ backbones.
Electrochemical characterization was measured with a Zahner IM6 electrochemical workstation. Silver inks were applied on the electrode surface and silver wires were used as the voltage and current leads. All fuels were wetted with 3% H2O.
The compositions of the samples were determined by ICP and EDS since perovskite oxides can form A-site and B-site non-stoichiometric phases. Table 1 shows the ICP and EDS results of the synthesized LSFN and LSF powder. In a H2-3%H2O atmosphere at 750 °C similar to the anode operating environment in SOFCs, it was observed that the LSFN sample showed a quite different morphology compared to that of the LSF sample. As shown in Fig. 1(a)-(c), many smaller nanoparticles with a size range of 10-20 nm had evenly segregated onto the surface of the LSFN particles after reducing for 2 h. However, this phenomenon was not observed in the case of the LSF sample (Fig. 1(d)-(f)). The phase purity and crystal structure of the products were examined by XRD. Fig. 1(g) shows the XRD patterns of the LSFN and LSF powder calcined at 1300 °C, which were indexed as a pure perovskite phase. From Fig. 1(g), it can be seen that LSNF and LSF still maintained the perovskite structure after they were calcined in the reducing atmosphere at 750 °C.
The morphology and structure of the La0.6Sr0.4Fe0.9Ni0.1O3-δ powder after calcination in a wet H2 atmosphere at 750 °C for 2 h were characterized by high resolution TEM. Fig. 2(a)-(c) are typical TEM images of the reduced La0.6Sr0.4Fe0.9Ni0.1O3-δ powder. It can be clearly seen that the surface of the big bulk particle was covered with uniformly dispersed smaller nanoparticles with an average diameter of ~20 nm. Clear lattice fringes were observed throughout the particle, as shown in the HRTEM image (Fig. 2(b)), which suggested the nanoparticles were single crystals. Fig. 2(d) shows an annular dark-field (ADF) scanning transmission electron microscope (STEM) image and the EDS elemental mapping of Ni, Fe, La, Sr and O of the reduced La0.6Sr0.4Fe0.9Ni0.1O3-δ powder. To further clarify the chemical composition of the segregated nanoparticles, the EDS line-scanning profile is shown in Fig. 2(e). This clearly demonstrated that the segregated nanoparticles were Ni-Fe alloys. Therefore, these nanoparticles should have a cubic structure, which was confirmed by the fast Fourier transform (FFT) pattern (inset in Fig. 2(c)) in which the lattice spacings of 0.208 and 0.129 nm observed in the HRTEM image corresponded to the (111) and (220) planes of the cubic structure, respectively. Ni-Fe alloy has a high catalytic activity for a carbon based fuel [18, 19]. As a SOFC anode, the catalytic activity of La0.6Sr0.4Fe0.9Ni0.1O3-δ would be significant improved by the in situ formed Ni-Fe nanoparticles. The XRD results indicated that the LSFN cathode decomposed into LaSrFeO4, La2O3, Fe and Ni at over 800 °C, as shown in Fig. 3. As far as structural stability is concerned, these results suggested that LSFN -based materials are not suitable for use as a SOFC anode above 800 °C.
YSZ has a lower ionic conductivity at a lower operating temperature [20]. Tri-layer porous dense porous monolithic cells with a porous micro/nanostructure have demonstrated excellent performance in reduced-temperature SOFCs [21, 22]. Furthermore, symmetrical SOFCs with a redox stable material as both the anode and cathode have many advantages, including the elimination of sulfur poisoning and coke formation by reversing the electrode as well as having enhanced cathode stability [9, 23, 24]. Therefore, to enable the LSFN electrode to work well at a lower temperature, we used a symmetrical LSFN-YSZ|YSZ|LSFN-YSZ monolithic cell configuration to evaluate the electrochemical performance. Fig. 4(a) presents a SEM image of a typical YSZ-electrolyte monolithic cell consisting of a dense electrolyte layer 15 μm thick and two porous layers. The porosity of the porous layers was 45% and they were 300 μm thick. The effective electrode area was 0.385 cm2 (0.7 cm diameter). Symmetrical single cells were fabricated by impregnating a LSFN or LSF precursor solution onto the porous layers of YSZ backbones with a subsequent heat treatment. Fig. 4(b) shows a higher magnification view of the impregnated electrode. After impregnating and calcining, the LSFN particles uniformly filled the pores and adhered to the YSZ backbone support. The average particle size of the electrode material was 200 nm. We also optimized the amount of LSFN electrode loading. The electrochemical performances of cells with different electrode loadings are shown in Fig. 4(c). For the cells with a low LSFN loading, the electrode material would not form a continuous conducting network. For the cell with 40 wt% LSFN loading, the effective triple-phase boundary (TPB) of anode was decreased due to the over coverage by the LSFN particles. The cell with 30 wt% loading showed the best performance. The infiltrated LSFN layer was porous but well internally connected at this loading. As shown in Fig. 4(d), there are only diffr action peaks from fluorite YSZ and perovskite LSFN in the XRD pattern of the single cell pellet with 30 wt% LSFN loading.
Fig. 5(a) and (b) show the cell voltage and power density as a function of the current density for the monolith cell with different electrodes and various fuels. From the comparison, it can be found that LSFN as a symmetrical electrode exhibited better catalytic activity for the oxidation of hydrocarbon fuels at 750 °C. The peak power densities of cells with impregnated LSFN electrodes reached 280, 230, and 400 mW/cm2 in H2, CH4 and C3H8, respectively. In C3H8, the open circuit voltage (OCV) of the cell with the LSFN electrodes was 1.18 V, which was much higher than that in H2 (~1.05 V), indicating that LSFN has excellent catalytic activity in C3H8. In CH4, the OCV of the cell was 1.00 V, which was higher than with a Ce(Mn, Fe)O2: La(Sr)Fe(Mn)O3 composite anode (~0.6 V at 800 °C) [10], while the OCV of the cell with impregnated LSF electrodes was much lower in CH4, indicating a lower catalytic activity for hydrocarbon oxidation. The peak power densities of the cell with impregnated LSF electrodes were only 220, 80, and 240 mW/cm2 in H2, CH4 and C3H8, respectively. As a symmetrical electrode, the LSFN has superior catalytic activity for CH4 and C3H8 oxidation relative to the perovskite Sr2Fe1.5Mo0.5O6-δ [23] and K2NiF4-type structured Pr0.8Sr1.2(Co, Fe)0.8Nb0.2O4+δ catalyst [24]. To further examine the stability of the cell impregnated with LSFN and LSF electrodes, the current densities of the cell were recorded as a function of time under a constant voltage of 0.7 V i n humidified C3H8. As shown in Fig. 5(c), it can be seen that the symmetrical cell with LSFN electrodes showed excellent stability in humidified propane fuel at 750 °C and only a slight power output degradation was observed after more than 36 operation. However, the current density of the cell with LSF electrodes fluctuated strongly in humidified C3H8. Moreover, a lot of tar-like substance was collected from the exhaust gas. However, the tar-like substance was not observed for the cell with the LSFN electrodes. The tar-like substance was analyzed by GC-MS (Table 2). There were more than 35 peaks in the GC-MS pattern. For simplicity, we only selected the peaks with an area more than 4%, which accounted for 61% of the total peak area. The MS results of the selected peaks are shown in Table 2, from which it can be seen that most of the products
contained a benzene ring. This result indicated that C3H8 was not completely oxidized with the LSF electrode. All these results showed that the Ni doped La0.6Sr0.4FeO3-δ has better catalytic activity for hydrocarbon oxidation than La0.6Sr0.4FeO3-δ. The enhancement of the catalytic activity of the LSFN was attributed to the high dispersion of the in situ segregated Ni-Fe nanoparticles.
Pure phase perovskite compounds La0.6Sr0.4Fe0.9Ni0.1O3-δ (LSFN) and La0.6Sr0.4FeO3-δ (LSF) were synthesized and studied as anodes for hydrocarbon fuel SOFCs. After reduction at 750 °C in H2-3%H2O, highly active Ni-Fe alloy nanoparticles were in situ segregated onto the surface of LSFN particles. Symmetrical electrodes for SOFC single cells were prepared by impregnating the precursor solution onto the porous layers of YSZ monolithic cells and a subsequent calcination. The peak power density of a symmetrical LSFN-YSZ|YSZ|LSFN-YSZ monolithic cell reached 400 and 230 mW/cm2 at 750 °C in humidified CH4 and C3H8 fuels, respectively. The cell showed excellent long term stability at 750 °C for more than 36 h without coking and tar formation. The results demonstrated that the Ni doped La0.6Sr0.4FeO3-δ is a promising alternative anode for low temperature SOFCs and as both electrode materials for symmetrical SOFCs. The extraordinarily high performance of the new electrode would provide a very promising strategy for designing a highly active electrode material for SOFCs and as catalysts for other applications.