1. Introduction
Solid oxide fuel cells (SOFCs) have been considered one of the most promising energy conversion devices due to their high efficiency and fuel flexibility. The high temperature operation (600°C–900°C) of the SOFCs facilitates the direct usage of commercial hydrocarbon fuels, such as syn-gas, natural gas, coal-gas, bio-gas, pure hydrogen and so on [
1,
2]. The hydrocarbon fuels provide many physiochemical advantages and characteristics over pure hydrogen fuels, such as easy transportation/storage and high energy density [
3,
4]. As hydrogen fuels are obtained from the hydrocarbon source through the reforming process, the direct utilization of the hydrocarbon fuels minimizes the necessity of additional system equipment, such as external reformers and/or purifiers, which improves the overall system efficiency and lowers the cost. The operation of several hydrocarbon fuels, however, cause carbon coking and sulfur poisoning in the nickel and yttria-stabilized zirconia (Ni/YSZ) anode, which is widely employed as the anode material in the SOFCs. Although the Ni-based cermet is considered an effective anode material in the SOFCs due to its excellent electrical conductivity and catalytic activity, it can tolerate only ppm levels of sulfur concentration [
5,
6]. Carbon deposition in the Ni/YSZ anode is also one of critical problems encountered due to the hydrocarbon fuel operation. Carbon deposition in the triple phase boundary (TPB), where the electrochemical reaction occurs, decreases the SOFC performance by covering the TPB sites. In addition, the bulk carbon deposition in the anode blocks the fuel gas flow causing an increase in the diffusion resistance.
Numerous alternative anodes that do not suffer from carbon deposition and sulfur poisoning have been developed for utilization in practical hydrocarbon fuels. Several stringent properties are required for an alternative SOFC anode [
7,
8]; (i) sufficient electro-catalytic activity for fuel oxidation, (ii) good electrical conductivity with ionic and electronic pathway, (iii) physical and chemical stability under the operating conditions of the SOFCs, (iv) chemical compatibility with electrolytes at high temperatures, and (v) resistance to carbon deposition and sulfur poisoning.
Due to the high catalytic activity for carbon formation exhibited in Ni, the carbon deposition easily occurs in the Ni/YSZ anode. Therefore, replacing Ni with other metals such as Cu, Co, Pd, and Ru that show relatively low catalytic activity for carbon formation be one of the solutions avoiding carbon deposition. The properties of such metal-based anodes, however, are still insufficient for the practical utilization of the commercial hydrocarbon fuels. Most metals are easily sintered and agglomerated at high temperatures during long-term operations. Noble metals are beyond consideration, as they are considerably expensive. The intrinsic limitation of a metal as an alternative anode material led to the consideration of the oxide-based materials, because most of the oxides are immune to carbon deposition. The desirable oxides possess the mixed ionic and electronic conductive (MIEC) property. An electrochemical reaction of the MIEC oxides can occur on the anode surface/fuel phase boundary (2PB) leading to an extension of the entire anode surface. Therefore, doped perovskites with the MIEC property can be considered excellent candidates for the alternative anode materials. Perovskites with the ABO
3 structure enables the substitution of rare-earth and/or lanthanide cations into the A-site and transition metal cations into the B-site. The multivalent function of P
O2 exhibited by the transition metal cations leads to the delocalization of electrons, resulting in good electronic conductivity. In addition, the misvalenced cations produce oxygen ion vacancies via the B-site substitution resulting in improved ionic conductivity. Therefore, many efforts have been devoted to develop perovskites with the MIEC property and employ them as the SOFC anodes [
9–
16]. Previously, we reported on various alternative anode materials with perovskite structures, such as, Sr
1-xY
xTiO
3-δ [
17,
18], Sr
1-xY
xTi
1-yFe
yO
3-δ [
19,
20], Sr
1-xY
xTi
1-yNi
yO
3-δ [
21], and Sr
2NiMoO
6-δ [
22], which produce satisfactory results for methane fuels.
In the present study, we have analyzed the electrochemical characteristics of Sr2Ni1.8Mo0.2O6-δ (SNM) as an alternative anode for the H2 and CH4 fuels. In addition, the SNM anode surface is modified using samarium-doped ceria (SDC) to improve its electrochemical properties.
2. Experimental
The Sr2Ni1.8Mo0.2O6-δ (SNM) powder was prepared using the sol-gel method. Strontium nitrate (Sr(NO3)3·H2O, Aldrich), nickel nitrate hexahydrate (Ni(NO3)2·6H2O, Aldrich), and ammonium heptamalybdate tetrahydrate ((NH4)6Mo7O24·4H2O, Aldrich) were dissolved in ethylene diamine tetra acetic acid (EDTA), which is a chelating agent. The quantities of these compounds were determined using stoichiometric calculations. The sol-gel mixture was slowly evaporated on a hot plate equipped with a magnetic stirrer at 150°C. The viscous gel was swelled and ignited in a subsequent heating process at 400°C. After removing most of the organic compounds, it was calcined at 800°C for 10 h to form a perovskite structure. The synthesized crystal structures were analyzed using an X-ray diffractometer (XRD, Rigaku, RINT-5200). After grinding thoroughly, the SNM powder was blended with a binder, a solvent and some additives in a ball mill. The prepared SNM slurry was tape-casted on an electrolyte. In addition, the samarium-doped ceria ((CeO2)0.8(Sm2O3)0.2, SDC) sol was prepared to modify the SNM anode. Samarium(III) nitrate hexahydrate (Sm(NO3)3)6H2O, 99.9%, Aldrich) was diluted in deionized water with cerium(IV) oxide (CeO2, 30–50 nm particles dispersion, Aldrich). The SNM anode was dip-coated in the prepared SDC sol and calcined at 700°C. The cathode side was masked completely during the anode coating process to minimize the cathode effects owing to the SDC modification. The coating process was repeated five times to ensure that the entire anode pore wall was coated. The microstructure and composition of the SNM anode were analyzed using a scanning electron microscope and an energy-dispersive X-ray spectroscopy (FE-SEM, EDS, Hitachi, S-4200, Japan).
The yttria-stabilized zirconia (YSZ, (ZrO2)0.92 (Y2O3)0.08, Fuelcellmaterials) electrolyte was prepared by uniaxial dry pressing and sintered at 1400°C for 10 h. The YSZ electrolyte with button-type was 25 mm in diameter and 1.0 mm in thickness. The prepared SNM slurry was tape-casted on the YSZ electrolyte and sintered at 1400°C for 5 h. The SNM anode was 0.95 cm2 in area and 50 μm in thickness. Lanthanum strontium manganite (LSM, (La0.8Sr0.2)0.95MnO3-δ, Fuelcellmaterials) was used as the cathode material. The prepared LSM slurry was tape-casted on the YSZ electrolyte and co-fired at 1100°C for 10 h.
To investigate the electrochemical performance, the disk-type cell was mounted between two double-layered alumina tubes and sealed with Pyrex glass, as illustrated in
Fig. 1. A perforated Pt plate and Ag paste were used as the current collectors. For the anode, humidified hydrogen (~3 vol% H
2O) and humidified methane (~3 vol% H
2O) were used as the fuel gases, which were supplied at a flow rate of 200 ml/min. For the cathode, air was supplied at a rate of 300 ml/min as an oxidant gas. The electrochemical characteristics were analyzed using an impedance analysis device (SP-150, Biologic Science Instrument). The impedance spectra were recorded in the frequency range of 10
−2–10
6 Hz with an exciting voltage of 10 mV to ensure a linear response. The impedance analyses were performed at 750°C–850°C, where the Nyquist plot indicated an equilibrium state.
3. Results and Discussion
The double perovskite with A2B1-xB’xO6-δ comprises BO6/2 and B’O6/2 corner-shared octahedra. An aliovalent substitution of the A lattice sites and/or B lattice sites can improve the physical and electrochemical properties of SNM rendering it suitable for utilization as an alternative SOFC anode. The divalent cation in the A sites enables various oxygen state combinations in the two B cations. In Sr2Ni1-xMoxO6-δ, the Ni2+ cations in the B sites alter the Mo6+ cations to Mo5+ by introducing oxygen vacancies. The mechanism of the oxygen vacancy formation can be provided as follows:
At high temperature, the oxygen vacancies may be formed via
reaction (1). The released electrons pair may react with Mo
6+ to reduce the Mo
5+ state via
reaction (2). The Goldschmidt tolerance factor of Sr
2Ni
1.8Mo
0.2O
6-δ is 0.962. It is defined as follows:
where
rA,
rB, and
rO represent the Shannon’s radii [
23] of the A-site cations, B-site cations and the oxygen ion. For the B-site cations, r
B is calculated as an average of the radii of Mo
5+ (0.61 Å) and Ni
2+ (0.69 Å) with six coordination numbers. The bigger radii of Ni
2+ than that of Mo5+ Ni-O results to shorter band length of Ni-O than that of Mo-O. This leads to a reduction of the bonding angle of Ni-O-Mo from 180° and a decline in the tolerance factor to less than 1 in Sr
2Ni
1.8Mo
0.2O
6-δ.
Fig. 2 shows the XRD patterns of the SNM powder at varying temperatures (400°C, 600°C, 800°C, and 1000°C). At 400°C, no specific peaks were detected indicating that the perovskite structure was not formed under 400°C. SNM demonstrates a single perovskite phase at 600°C and marginal crystallization at higher temperatures. The XRD patterns of YSZ, SNM and the SNM/YSZ mixture presented in
Fig. 3 were used to investigate the chemical compatibility of the SNM anode and YSZ electrolyte. To verify the by-product formation between the SNM anode and YSZ electrolyte during the synthesis and/or operating process, the mixture of SNM and YSZ were co-fired at 1400°C for 10 h after the wet ball-milling. The XRD patterns of the SNM/YSZ mixture were compared to those of YSZ and SNM. No apparent peaks from the reaction of the by-products were detected other than the SNM and the YSZ phases.
Figs. 4(a) and 4(b) show the scanning electron microscopy images of the SNM anode and the interlayer between the YSZ electrolyte and the SNM anode, respectively. The average pore size was 2–5 μm in diameter and the pore volume was around 40 vol%. The SNM anode possessed a thickness of 50 μm on the dense YSZ electrolyte. The interlayer by-product between the SNM anode and the YSZ electrolyte was not detected.
Fig. 5 shows the transmission electron microscopy images and EDS composition analysis of O, Ni, Sr and Mo in the SNM anode. Ni was exsolved and sintered in the perovskite structure in nanoparticle size in contrast to the uniform phase formation in Sr and Mo. This phenomenon is consistent with our previous research [
22]. J.T.S Irvine Group also reported on the exsolution of Ni in the Ni-doped perovskite structure under reducing conditions at 800°C–900°C [
24].
Figs. 6(a) and 6(b) show the electrochemical impedance spectroscopy (EIS) results of the SNM symmetric cell with the YSZ electrolyte in H
2 (~3 vol% of H
2O) and CH
4 (~3 vol% of H
2O) at 750°C–850°C, respectively. H
2 and CH
4 were humidified in a bubbler before entering the reactor. Nyquist plots with Z
real (Re Z′) versus Z
imaginary (Im Z″) as the functions of frequency (0.01–10
6 Hz) were obtained under the open circuit voltage (OCV) conditions. The high frequency intercepts of the arcs with Z
real axis were corrected to compare the electrode polarization resistance values of the anodes. The polarization resistances of the SNM anode were 40.9 Ωcm
2, 20.1 Ωcm
2, and 12.1 cm
2 at 750°C, 800°C and 850°C, respectively in H
2. The anode polarization resistance of the SNM anode was much higher than that of the Ni/YSZ anode, which was typically 1 Ωcm
2 at 750°C–850°C in H
2. The marginal intrinsic catalytic activity of SNM for H
2 oxidation can increase the polarization resistance. In CH
4, the anode polarization resistances increased to 75.8 Ωcm
2, 30.7 Ωcm
2, and 18.4 Ωcm
2 at 750°C, 800°C and 850°C, respectively. The diffusion of CH4 was slower than that of H2 leading to an increase in the polarization resistance of mass transport. In addition, the exsoluted Ni nanoparticles can provide sufficient active sites for the CH
4 pyrolysis to form the carbon depositions resulting in deactivating the electrochemical reaction in the SNM anode. The bulk carbon deposition on the Ni nanoparticles blocks the anode pore leading to an increase in the mass transfer resistance.
Figs. 7(a) and (b) show the IV-characteristics of the single cell with the SNM anode/YSZ electrolyte/LSM cathode in H
2 and CH
4 at 750°C, 800°C, and 850°C, respectively. The cell performance was obtained after stabilizing in the humidified H
2 or the humidified CH
4 for an hour. The maximum power densities of the H
2 fuel were 18.6 mW/cm
2, 30.3 mW/cm
2, and 39.4 mW/cm
2 at 750°C, 800°C, and 850°C, respectively. The maximum power densities of the CH
4 fuel were 9.6 mW/cm
2, 13.2 mW/cm
2, and 15.9 mW/cm
2 at 750°C, 800°C, and 850°C, respectively. The CH
4 fuel can be decomposed to CO and H
2 in the presence of H
2O through a steam methane reforming reaction. However, this formation would be limited, because the humidified CH
4 contains only ~3 vol% of water. CH
4 could be decomposed to CO and H
2 via an electrochemical oxidation, which is defined as follows:
A relatively low OCV and cell performance in CH4 could be due to the poor catalytic/electro-catalytic activity of the SNM anode. In addition, the Ni nanoparticles exsoluted from the SNM phase can accelerate the pyrolysis of CH4 so that it can be decomposed to carbon and hydrogen. The carbon deposition via the pyrolysis of CH4 can increase the anode polarization resistance leading to a decline in the cell performance. Therefore, the SDC surface modification of the SNM anode can be effective in improving the catalytic and/or electro-catalytic performance of the SNM anode surface resulting in an improved cell performance.
Figs. 8(a) and 8(b) show the microstructure cut-view images of the unmodified SNM anode and the SDC-modified SNM anode, respectively. SDC has been studied as an alternative electrolyte material for the SOFCs due to its high ionic conductivity. The MIEC property in the reducing conditions is also a beneficial characteristic for the alternative anode materials. The MIEC property exhibited by the SDC can provide electrochemical oxidation sites that are beyond the triple phase boundary (TPB). In addition, the SDC has been recognized as an oxygen storage and transfer material. The electrochemical oxidation of CH
4 and the deposited carbon can likely occur on the SDC surface. In our previous research, we reported that the SDC-modified cells demonstrated an improvement of 30–50% in the cell performance due the MIEC property and good ionic conductivity of the SDC. In the present study, the SNM anode was modified by the SDC sol-gel coating to minimize the carbon deposition and improve the cell performance. The SDC layer with 10~ 50 nm in thickness was formed on the SNM anode surface. The SDC was chemically compatible with the SNM anode. No noticeable peaks of by-products in the SDC-coated SNM anode were detected other than SNM and SDC phase shown in
Fig. 9. The SDC thin layer can provide the catalytic active sites for the chemical and electrochemical decomposition of CH
4. In addition, the deposited carbon can be oxidized to CO or CO
2 electrochemically due to the oxygen storage property of the SDC.
Figs. 10(a) and 10(b) show the EIS results of the SDC-modified SNM anode symmetric cell with the YSZ electrolyte at 750°C, 800°C, and 850°C in the humidified H
2 and humidified CH
4 fuel conditions, respectively. The anode polarization resistances were 3.98 Ωcm
2, 3.01 Ωcm
2, and 1.30 Ωcm
2 at 750°C, 800°C, and 850°C in H
2, respectively. The SDC thin layer can provide high oxygen ion conductivity and expanded electrochemical reaction sites beyond the TPB to the SNM anode, leading to a significant decrease in the anode polarization resistance. In addition, the anode polarization resistance decreased from 75.8 Ωcm
2, 30.7 Ωcm
2, and 18.4 Ωcm
2 to 5.96 Ωcm
2, 4.87 Ωcm
2, and 2.91 Ωcm
2 at 750°C, 800°C, and 850°C in CH
4, respectively. In addition to the MIEC characteristics exhibited by the SDC, its excellent catalytic property towards CH
4 decreased the anode polarization resistance of the SDC-modified SNM anode. The carbon deposited on the anode surface can oxidize to CO or CO
2 due to the oxygen storage and transfer properties of the SDC.
Fig. 11(a) and 11(b) show the IV-characteristics of the SDC-modified SNM anode in H
2 and CH
4 at 750°C, 800°C, and 850°C, respectively. The cell performance significantly improved in H
2 and CH
4 due to the excellent oxygen ion conductivity and the MIEC property exhibited by the SDC. The maximum power densities improved from 18.6 mW/cm
2, 30.3 mW/cm
2, and 39.4 mW/cm
2 to 51.8 mW/cm
2, 69.8 mW/cm
2, and 117.7 mW/cm
2 in the H
2 fuel at 750°C, 800°C, and 850°C, respectively. They improved from 9.6 mW/cm
2, 13.2 mW/cm
2, and 15.9 mW/cm
2 to 39.2 mW/cm
2, 49.7 mW/cm
2, and 66.6 mW/cm
2 in the CH
4 fuel at 750°C, 800°C, and 850°C, respectively. As the Ni particles were exsoluted from the SNM anode, CH
4 would not only decompose chemically and electrochemically to form H
2 and CO but also decompose to C and H
2 via the CH
4 pyrolysis. The carbon deposition, however, can be minimized on the SDC thin layer due to the SDC’s excellent oxygen ion conductivity. The SDC coating layer on the pore surface of the SNM anode would minimize the direct exposure of the CH
4 fuel with Ni particles by exsolution, leading to limiting the pyrolysis of CH
4 on the anode surface. The deposited carbon can likely oxidize to CO or CO
2 electrochemically. In addition, the electrochemical reaction site would expand to 2PB (gas/anode) beyond the TPB leading to an improved cell performance.