Effects of Co Doping on the Electrochemical Performance of Double

Article pubs.acs.org/JPCC

Effects of Co Doping on the Electrochemical Performance of Double Perovskite Oxide Sr2MgMoO6−δ as an Anode Material for Solid Oxide Fuel Cells Zhixiang Xie,† Hailei Zhao,*,†,‡ Zhihong Du,† Ting Chen,† Ning Chen,†,‡ Xiaotong Liu,† and Stephen J. Skinner§ †

School of Materials Science and Engineering, University of Science and Technology Beijing, Beijing 100083, China Beijing Key Lab of New Energy Material and Technology, Beijing 100083, China § Department of Materials, Imperial College London, Prince Consort Road, London SW7 2BP, United Kingdom ‡

ABSTRACT: Double-perovskite materials of composition Sr2Mg1−xCoxMoO6−δ (SMCMO, x = 0 to 0.7) were evaluated as potential SOFC anode materials. Their lattice structures, electrical and ionic conductivity, thermal expansion coefficient (TEC), and electrochemical performance were investigated as a function of Co content. Co doping was found to increase the TEC of the Sr2MgMoO6−δ material; however, the TEC was within the range of the commonly used La0.8Sr0.2Ga0.8Mg0.2O3‑δ (LSGM) electrolyte. SMCMO also showed good chemical compatibility with the LSGM electrolyte at temperatures below 1300 °C. Both the electronic and ionic conductivity increased with increasing Co doping. To investigate the effect of Co doping on the conduction properties of SMCMO, we performed first-principle calculations. From these results, the weak Co−O bond is considered to be responsible for the enhanced ionic conductivity of SMCMO materials. The substitution of Co was also found to increase the sinterability of SMCMO, resulting in a decrease in the polarization resistance of the SMMO electrode. Single-cell tests indicated the potential ability of the Co-doped SMMO to be used as SOFC anodes.

1. INTRODUCTION The fuel cell is a well-known device that directly and efficiently converts chemical energy to electrical energy. Of the various fuel cell types, the solid-oxide fuel cell (SOFC) is considered to have the greatest potential for large-scale power generation; however, one of the difficulties faced in the improvement of the current state-of-the-art SOFCs is related to the need for a suitable anode that can operate not only with pure H2 gas but also with low-cost and more abundant fuels such as natural gas.1−7 Although the conventional anode material, Ni-YSZ, exhibits excellent catalytic activity and current collection, it suffers from carbon deposition and sulfur poisoning when using hydrocarbon as fuels. It also shows low tolerance to redox cycles, which make this anode vulnerable to exposure to high oxygen partial pressure environments.8−13 In recent years, mixed ionic and electronic conductors have been investigated as potential anode materials considering their high resistance to sulfur poisoning and carbon deposition, high thermochemical stability, and extended triple-phase boundary (TPB). They are principally chromite-based and titanate-based perovskites. Doped SrTiO3 was reported to have high electronic and ionic conductivity and showed good resistance © 2012 American Chemical Society

to carbon deposition. However, the low catalytic activity and the poor redox stability limited the cell performance.14−17 Although chromite-based anode (La0.75Sr0.25Cr0.5Mn0.5O3−δ) exhibited good performance when operating in pure H2 and CH4 at high temperatures, fuels containing H2S would deteriorate its operating performance.18−20 Recent ly , th e B-sit e-ordered do uble-perovskite Sr2MgMoO6−δ (SMMO) reported by Huang et al. has attracted great interest as a candidate anode material for solid oxide fuel cells (SOFCs).21,22 This material was reported with suitable chemical stability and good resistance against carbon deposition and sulfur poisoning.23−25 Unfortunately, it showed low electrical conductivity and poor electrocatalytic activity when using Ag or other non-platinum current collector, both of which lead to large anode polarization. Substituting A and/or B sites with alkali earth and transition-metal elements, respectively, offers the possibility of improving the electronic or ionic conductivity of Sr2MgMoO6−δ as well as the catalytic Received: December 27, 2011 Revised: April 6, 2012 Published: April 10, 2012 9734

dx.doi.org/10.1021/jp212505c | J. Phys. Chem. C 2012, 116, 9734−9743

The Journal of Physical Chemistry C

Article

coefficient measurement. The bulk densities of sintered samples were measured by Archimedes method. The phase composition and crystal structure of the samples were determined by X-ray diffraction (XRD, Rigaku D/max-A X-ray diffractometer) using Cu Kα radiation. To investigate the influence of Co doping on the lattice structure of the samples, some of the cell parameters were refined by the full profile Rietveld analysis using the Fullprof program.40 A LEO-1450 scanning electron microscope (SEM) was used to observe the microstructure of the sintered pellets. Thermal expansion coefficients (TECs) were measured using a dilatometer (NETZSCH DIL 402C) with an Al2O3 reference. The measurements were performed in air with a flow rate of 50 mL min−1. A heating rate of 5 °C min−1 was applied during the measurements. The electrical conductivity of all samples was measured in forming gas (5% H2/Ar) with the standard four-terminal dc method in the temperature range of 100−900 °C, as previously described in detail.41 The sintered pellets were used for ionic conductivity measurements. The ionic conductivities were determined by the electron blocking electrode method.42 The test was operated in 5% H2/Ar atmosphere in the temperature range of 500−800 °C. The dense LSGM electrolytes were also prepared from powders synthesized by the citric acid−nitrate method, which were pressed into disks with a diameter of 16 mm and then sintered in air at 1450 °C for 6 h. To examine the chemical compatibility of SMCMO with LSGM electrolytes, the 1250 °C-prepared Sr2Mg0.5Co0.5MoO6−δ, which already shows a pure double-perovskite structure and will be used as the anode powder in cell evaluation, was mixed with LSGM in the weight ratio of 1:1 and then pressed and calcined at 1300 °C for 10 h in air. The calcined pellet was crushed and examined by XRD to clarify the phase purity. Impedance measurements were performed on symmetrical cells of SMCMO/LSGM/SMCMO at temperatures between 700 and 850 °C in 5% H2/Ar. The measurements were recorded on a Solartron 1260 frequency response analyzer in combination with a Solartron 1287 potentiostat over the frequency range 0.01 to 106 Hz with a perturbation amplitude of 5 mV, which was controlled by the Zplot electrochemical impedance software. For fabricating symmetrical cells, the dense LSGM pellets had a thickness of 600 μm. The electrode layers were produced by mixing SMCMO powders with αterpineol and ethylene cellulose using starch as a pore former to create a slurry. The mass percentage of the pore former was 30%. The slurry was symmetrically screen printed onto both surfaces of the electrolyte, followed by calcination at 1300 °C for 2 h. Single fuel cells were prepared using the electrolytesupported technique with 400 μm thick dense LSGM as the electrolyte, SMCMO as the anode, and SmBaCo2O5+δ as the cathode. Anode and cathode slurries were obtained by mixing each prepared SMCMO and SmBaCo2O5+δ powder with αterpineol and ethylene cellulose. The anode slurry was screen printed onto one side of the sintered LSGM electrolyte and fired at 1300 °C for 2 h in air; then, SmBaCo2O5+δ was screen printed on the other side of the cell and calcined at 1000 °C for 2 h in air. After that, Ag paste was attached to the electrode surface as the current collector, which was heated to 600 °C for 30 min in air. The area of electrodes was 0.502 cm2. The electrolyte-supported cell was sealed using a ceramic-based material (Cerama-bond 552-VFG, Aremco).

properties. Doping La on the Sr-site has been reported to improve the electrocatalytic properties for fuel oxidation26 but deteriorates the structural stability in high oxygen partial pressure.27 Sm-doped SMMO (Sr1.6Sm0.4MgMoO6−δ) displayed excellent electrochemical performance while maintaining structural stability.28 Doping Mn on the Mg-site could increase the electrical conductivity, but unfortunately, Mn-doped SMMO was found to be sensitive to Po2, with the electrical conductivity rapidly decreasing with increasing oxygen partial pressure.22 The electrical conductivity of Fe-substituted SMMO (Sr2Mg0.75Fe0.25MoO6−δ) increased, but again the material was unstable under oxidizing conditions due to phase segregation above 600 °C.23,27 Vasala et al. reported that doping W and Nb on the Mo site of SMMO depressed the electrical conductivity, but a small amount of Nb could increase the ionic conductivity.29 Most recently, it is reported that SMMO exhibited electrocatalytic activity toward fuels only when using Pt paste as a current collector.30 It is well known that the 3d-block transition metals, such as Co, Ni, and Fe, normally exhibit good catalytic activity for the oxidation of H2 and CH4. More recently, Sr2CoMoO6−δ (SCMO), Sr2NiMoO6−δ, A2FeMoO6−δ (A = Ca, Sr, Ba), and Sr2Fe1.5Mo0.5O6−δ as anode materials containing a 3d-block transition metal were reported. The results showed that Sr2CoMoO6 anodes exhibited high cell power density with H2 and CH4 as fuels; however, it was unstable under reducing atmosphere.31−36 The poor stability under reducing environments is due to the weak bond between cobalt and oxygen. The structural stability of perovskite oxides is mainly determined by the stability of octahedral BO6. A stronger B−O bond usually leads to a more stable perovskite structure.37−39 In this work, we consider the advantages of both SMMO and SCMO in electrocatalytic and structural stability and develop a new anode material of Co-doped SMMO (Sr2Mg1−xCoxMoO6−δ, SMCMO). Co was introduced into the Mg site of Sr2MgMoO6−δ to improve the conductivity and the electrocatalytic acitivity with the premise of not deteriorating the structural stability. The effects of Co doping on the lattice structure, microstructure, thermal expansion, and electronic and ionic conductivities were systematically investigated. In addition, the electrode polarization was analyzed using symmetrical cells, whilst the single-cell electrochemical performance of SMCMO anodes on LSGM (La0.8Sr0.2Ga0.8Mg0.2O2.8) electrolyte was examined.

2. EXPERIMENTAL SECTION Single-phase Sr2Mg1−xCoxMoO6−δ (x = 0 to 0.9) powders were prepared via a nitrate−citrate route with citric acid monohydrate as the chelating agent and Sr(NO3)2, Mg(NO3)2·6H2O, Co(NO3)3·6H2O, and (NH4)6Mo7O24·4H2O as the starting materials (all A.R. grade). The molar ratio of the total amount of metal ions to the amount of citric acid monohydrate was fixed at 1:2. The resultant solution was dehydrated and slowly heated until self-combustion of the gel occurred; then, the asobtained black precursors were ground and fired in air at 900 °C for 10 h to eliminate the organic matter. The prepared powders were then ground slightly to destroy agglomerates and pressed into bars (40 mm × 7 mm × 3 mm) and pellets (diameter = 15 mm) by uniaxial pressing (ca.120 MPa). To obtain dense samples, we sintered the resulting pellets and bars at 1450 °C in air for 10 h (hereafter, noted as sintered sample) and then reduced them at 800 °C in 5% H2/Ar for 24 h (noted as reduced sample) for conductivity and thermal expansion 9735



Article

Figure 1. XRD patterns of Sr2Mg1−xCoxMoO6−δ (a) sintered at 1450 °C for 10 h in air and (b) sintered at 1450 °C for 10 h in air and then reduced at 800 °C in 5% H2/Ar for 24 h.

Figure 2. Rietveld refinement profiles of Sr2Mg1−xCoxMoO6−δ after being sintered at 1450 °C in air then reduced at 800 °C in 5% H2/Ar: (a) x = 0; (b) x = 0.5. Solid circles show the observed data points and the solid line represents the calculated diffraction pattern.

Fuel cell tests were carried out using humidified H2 as fuel and air as oxidant operating at 800 °C. The flow rates of fuel and oxidant gases were all controlled at 30 mL min−1 by mass flow controllers. Before testing, the cells were exposed to pure H2 for 4 h at 800 °C to reduce the anode materials. The measurement was started after the open circuit voltage stabilized as a constant to obtain a sufficiently stabilized system. All measurements were taken by holding at each point for least 10 min when no significant changes in current and voltage were observed. In this work, the DOS of SMMO with and without Co doping was calculated using the CASTEP module, part of the Materials Studio software.43 All calculations here were carried out by using the plane-wave pseudopotential method within density functional theory (DFT) as implemented in CASTEP. All calculations were performed within the generalized gradient approximation (GGA) and parametrized by Perdew−Burke− Ernzerhof formula (PBE) for exchange correlation energy.44 The cutoff energy of plane wave was set at 380 eV. The maximum root-mean-square convergence error of the total energy was set as 1.0 × 10−6 eV·atom−1.

Because the mixed conducting oxide used as an anode should be stable under SOFC operating environments, all of the samples were reduced to 800 °C in 5%H2/Ar for 24 h. As shown in Figure 1b, when the doping amount of Co was at x ≤ 0.7, the samples were found to be phase-pure, whereas impurities of Co3O4, SrO, MgO, and SrMoO4 were observed in the sample with x = 0.9. The structural stability of perovskite oxides is primarily rested with the stability of octahedral BO6. The B−O bond energy is one of the main factors that influences the stability of perovskite oxides. The generation of impurities in the sample with x = 0.9 under reducing environments is mainly due to the weak bond energy between cobalt and oxygen.37 More Co doping is obviously unfavorable for the structural stability of SMMO. To understand the detailed crystal structure of SMCMO, samples with x = 0 and 0.5 were selected to perform Rietveld refinement for the crystal structure using a triclinic I-1 model. This space group was utilized for Sr2MgMoO6−δ by BernuyLopez et al.25 The results of Rietveld refinement for SMCMO samples with x = 0 and 0.5 are given in Figure 2 and Table 1. The profile R value (Rp), weighted profile R value (Rwp), and χ2 value of the refined structure parameters indicate that the refinement results are acceptable. The Rietveld refinement results indicate that the Co doping increases the lattice volume, which may be caused by the introduction of relatively large Co ions compared to Mg ions or by the further generation of oxygen vacancies. Either of them could lead to the deduction that the Co ions mainly adopt the

3. RESULTS AND DISCUSSION 3.1. Phase and Microstructure Development. 3.1.1. Phase Structure. Figure 1a shows the XRD patterns of Sr2Mg1−xCoxMoO6−δ (x = 0 to 0.9) bars sintered at 1450 °C for 10 h in air. All samples can be indexed as a single doubleperovskite phase with no detectable impurity peaks.45,46 9736



Article

that the porosity decreases and the grain size increases with increasing Co-doping amount, indicating that Co doping improves the sinterability of Sr2MgMoO6−δ. 3.2. Characteristics of Conductivities. 3.2.1. Electrical Conductivity. Figure 4 shows the total electrical conductivity as a function of temperature for the Sr2Mg1−xCoxMoO6−δ (x = 0 to 0.7) samples in forming gas between 50 and 850 °C. The electrical conductivity increased with increasing Co content, mainly ascribable to the electron delocalization correlated to the stronger covalent characteristics of the Co−O bond48−50 compared with the Mg−O bond. The little charge-transfer gap between the 3d band of Co ions and the 2p band of O ions makes the electron jump much easier. Besides, the weaker Co− O bond compared with Mg−O bond will cause more lattice oxygen loss at high temperature, which also made a positive contribution to the electronic and ionic conductivity, as expressed in eq 2. The electrical conductivity of each composition except for x = 0 increased with increasing temperature through a peak, then decreased. The temperature of the electrical conductivity maximum point shifted from approximately 750 to 300 °C as the Co content increased from x = 0.3 to 0.7, which was similar to other Co-based perovskite oxides.38,51

Table 1. Rietveld Refinement Results for Sr2Mg1−xCoxMoO6−δ (x = 0, 0.5) Samples: Lattice Parameters X

0

0.5

space group

I-1

I-1

a (Å) b (Å) c (Å) α (deg) β (deg) γ (deg) V (Å3) Rp (%) Rwp (%) χ2

5.57524(15) 5.57253(15) 7.92324(10) 90.015(1) 89.961(2) 90.013(5) 246.160(7) 9.78 13.5 2.82

5.57432(21) 5.57268(20) 7.93888(16) 89.966(2) 90.012(5) 90.006(8) 246.612(4) 9.12 12.8 3.16

+2 valence state. In double-perovskite Sr2Mg0.5Co0.5MoO6−δ, magnesium has a fixed valence state of +2, whereas the cobalt ion may take multiple valence states of +2 or +3. The ionic radius for six-coordinated Mg2+, Co2+ (HS), Co2+ (LS), Co3+ (HS), and Co3+ (LS) is 0.72, 0.745, 0.65, 0.61, and 0.545 Å, respectively.47 Only Co2+ ion has a larger size than Mg2+ ion. from the viewpoint of defect chemistry, the substitution of Co3+ for Mg2+ will suppress the generation of oxygen vacancies in SMCMO, as described in eq 1. Accordingly, the increased lattice volume by Co doping suggests that the majority of Co ions should take +2 valence state in SMCMO.

× OO + Mo×Mo →

(2)

Figure 4b displays the Arrhenius plots of temperature dependence of electrical conductivity. According to the fitted slopes, the activation energy of electronic conduction as a function of Co content can be derived, as detailed in Figure 4b. The activation energy (Ea) of SMCMO for electron hopping decreased with increasing Co content, indicating that Co doping facilitates the conduction process of electrons. For mixed conductor, the electrical conductivity can be expressed as eq 3. The temperature term “T” in the preexponential factor and inside the exponential part plays opposite roles in the electrical conductivity. Therefore, there is

MgO

• × Co2O3 + V •• O ⎯⎯⎯⎯→ 2Co Mg + 3OO

1 ′ O2 (g ) + V •• O + 2Mo Mo 2

(1)

3.1.2. Microstructure. Dense samples with similar relative density are required to understand the electrical conduction behavior of SMCMO samples and to elucidate the effect of Co doping on the conductivity. Bulk density measurement indicates that the relative densities of the reduced samples with x = 0, 0.1, 0.3, 0.5, and 0.7 are 91.26, 94.42, 95.56, 95.62, and 95.85%, respectively. SEM observation (Figure 3) shows

Figure 3. SEM micrographs of Sr2Mg1−xCoxMoO6−δ sintered at 1450 °C in air and then reduced at 800 °C in 5%H2/Ar for 24 h: (a) x = 0; (b) x = 0. 1; (c) x = 0.5; and (d) x = 0.7. 9737



Article

Figure 4. (a) Electrical conductivity of Sr2Mg1−xCoxMoO6−δ as functions of Co-doping level and temperature. (b) Arrhenius plots of the electrical conductivity.

Figure 5. Partial density of states diagrams of (a) SMMO and (b) Sr2Mg0.5Co0.5MoO6−δ.

responsible for the decreasing Tmax′ with increasing Co-doping level. 3.2.2. Electronic Structure. To gain insight into the effect of Co doping on the electrical conduction behavior of SMCMO, first-principles calculations were carried out based on the DFT.54 The density of states (DOS) were calculated after structure optimization with the Material Studio software. The result is shown in Figure 5. The DOS of SMMO is obviously changed by Co doping. A donor level was formed on account of Co-doped SMMO, causing the Fermi level to move closer to the valence band. Therefore, a much smaller band gap formed in Co-doped SMMO, which is propitious to the electron conduction and thus decreases the activation energy of electron conduction. The DOS of d electrons of the B-site ions around the Fermi level increased remarkably on Co doping and overlapped with the oxygen 2p electrons, which may imply an increase in charge carrier concentration. Both cases are favorable for the enhancement of electronic conduction. 3.2.3. Ionic Conductivity. The ionic conductivity of Sr2Mg1−xCoxMoO6 (x = 0, 0.3, 0.7) as a function of temperature is shown in Figure 6. The ionic conductivity increased with Co doping level. For perovskite oxides, the ionic conductivity is primarily dependent on the concentration of oxygen vacancies and the mobility of oxygen ions. The Co doping may increase the oxygen vacancy concentration at high temperature due to the weak Co−O bond, as indicated by eq 2. The activation energy (Ea) for oxygen ion migration in Sr2Mg1−xCoxMoO6−δ decreased remarkably with increasing

a characteristic temperature (Tmax) at which the electrical conductivity maximum was expected to occur. The characteristic temperature depends on the activation energy, Ea, as presented in eq 4.38,52 σ=

⎛ E ⎞ A exp⎜ − a ⎟ ⎝ RT ⎠ T

Ea = kTmax

(3) (4)

where A is a constant, relating to the lattice structure characters, R presents the gas constant, and k is the Boltzmann constant. Although the decreased Ea with Co substitution will cause the characteristic temperature Tmax to decrease, the calculated Tmax with extremely high value compared with the practically observed one suggests that the temperature shift of the maximum electrical conductivity is resulted by the lattice structure change during heating process. With temperature increasing, the lattice may lose oxygen, which will result in the generation of oxygen vacancies and electrons. The latter can lead to an increase in electronic charge carrier concentration, whereas the former will damage the connection of −Mo6+−O−Mg(Co)−O−Mo5+− and thus affect the polaron type conduction of electrons. Furthermore, the increased oxygen vacancy concentration generated at high temperature can trap electrons,53 thus decreasing the effective charge carrier concentration. These result in the lower practical transition temperature (Tmax′) of conductivity compared with the calculated one. The weak Co−O bond should be 9738



Article

Figure 7. Thermal expansion curves for Sr2Mg1−xCoxMoO6−δ (x = 0, 0.5, 0.7).

Figure 6. Temperature dependence of ionic conductivity of Sr2Mg1−xCoxMoO6−δ (x = 0, 0.3, 0.7) measured in flowing forming gas saturated with room-temperature water.

LSGM, which is frequently utilized in SOFC.59,60 The TEC of solids usually increases with decreasing binding energy between the ions in the lattice.61 The weaker Co−O bond compared with the Mg−O bond is responsible for the higher TEC of Codoped sample. Additionally, the TEC of oxides is also related to the oxygen vacancy concentration because there is a reduction in the electrostatic attractive force between the cations and anions when more oxygen vacancies are generated.59,60,62 Because Co doping facilitates the lattice oxygen loss, it could also contribute to the increase in TEC. 3.4. Chemical Compatibility with Electrolytes. Interfacial reaction between anodes and electrolytes may produce insulating phase, which increases the interfacial polarization resistance. To evaluate the chemical compatibility of Co-doped SMMO with electrolytes, we heated a pellet containing 50 wt % Sr2Mg0.3Co0.7MoO6−δ and 50 wt % LSGM to 1300 °C for 10 h in air. The sintered samples were crushed and examined by XRD, and the results are shown in Figure 8. There were no new phases detected in the Sr2Mg0.3Co0.7MoO6−δ/LSGM mixture, indicating that the anode Sr2Mg0.3Co0.7MoO6−δ powders have a

Co amount. This is attributable to the weaker Co−O bond compared with Mg−O bond, which ensures that removal of the oxygen ions around Co is more facile. The average metal− oxygen bond energy (ABE) can be expressed by eq 5, where ΔHAmOn and ΔHBm′On′ are the formation heats of AmOn and Bm′On′ oxides, respectively; ΔHA and ΔHB stand for the sublimation heats of A and B metals, respectively; and DO2 presents the dissociation energy of oxygen.55 The ABE of Mg− O and Co−O is calculated by using eq 4, which is 166.4 and 152.4 kJ/mol, respectively. ABE =

1 ⎛⎜ ΔH A mOn − mΔHA − 12m ⎝ 1 ⎜⎛ + ΔHBm′On′ − m′ΔHB 6m′ ⎝

⎞ n DO2⎟ ⎠ 2 ⎞ n′ − DO2⎟ ⎠ 2

(5)

A further factor contributing to the decreased Ea is the increased lattice parameter for samples with high Co-doping level as more space can be utilized for oxygen ion migration in the lattice. The weak Co−O bond allows more oxygen vacancies to generate in Co-rich samples at high temperature, which will contribute significantly to the oxygen ion conduction and thus increase significantly the ionic conductivity, as evidenced in Figure 6. 3.3. Thermal Expansion Behavior. The thermal expansion compatibility of an anode material with the electrolyte is an important property for device development. Poor TEC matching between electrode and electrolyte may induce exfoliation of the electrodes under fabricating and operating processes and hence shorten the lifetime of the SOFCs.56 The TEC of SMMO, Sr2Mg0.5Co0.5MoO6−δ, and Sr2Mg0.3Co0.7MoO6−δ was measured in the temperature range of 50−1300 °C. The thermal expansion curves are shown in Figure 7. The average TEC of the Sr2Mg0.5Co0.5MoO6−δ and Sr2Mg0.3Co0.7MoO6−δ samples are 13.3 × 10−6 and 13.9 × 10−6 K−1, respectively, between 50 and 1300 °C, which are slightly larger than that of the SMMO sample (12.9 × 10−6 K−1). In general, perovskite materials with large Co content exhibit high TECs,57,58 but in comparison with other cobalt-containing oxides, the TEC of samples Sr2Mg0.5Co0.5MoO6−δ and Sr2Mg0.3Co0.7MoO6−δ was close to that of the electrolyte

Figure 8. XRD patterns of mixtures of Sr2Mg0.3Co0.7MoO6/LSGM powders after heat treatment at 1300 °C for 10 h in air. 9739



Article

electrolyte layers. (The XRD result is not shown here.) Both of them can cause the increase in polarization resistance. The cross-sectional SEM micrographs of a 1300 °C-fired symmetrical cell with Sr2Mg1−xCoxMoO6−δ anode and LSGM electrolyte are displayed in Figure 10. The electrolyte is fully dense, and all electrodes present high porosity and good adherence to the electrolyte, which ensured a low contact resistance between anode and electrolyte. 3.5.2. Electrode Porosity Effect. The impedance spectra of Sr2Mg0.7Co0.3MoO6−δ/LSGM/Sr2Mg0.7Co0.3MoO6−δ symmetrical cells with electrodes prepared with different amounts of pore former were collected in 5% H2/Ar to distinguish the kinetic processes of each arc. The results are shown in Figure 11. The electrode responses are composed of two arcs

good chemical compatibility with electrolyte LSGM in air below 1300 °C. However, other groups have reported that SMMO can react with LSGM to produce a minor impurity at temperatures above 1200 °C.63 This discrepancy may be due to the high preparation temperature of SMMO powders (1250 °C) in this work. High preparation temperature usually leads to low activity. The result demonstrated that Co-doped SMMO is suitable for use as an anode material for SOFCs. 3.5. Symmetrical Cell Performance. 3.5.1. Calcination Temperature Effect. To assess the influence of the calcination temperature on the microstructure of the anode, the adhesion of the anode with the electrolyte as well as the anode performance, we prepared symmetrical cells Sr2Mg0.5Co0.5MoO6−δ/LSGM/Sr2Mg0.5Co0.5MoO6−δ at different temperatures. Figure 9 shows the Arrhenius plot of ASR

Figure 11. Impedance spectra of the symmetrical cells Sr2Mg0.7Co0.3MoO6−δ/LSGM/Sr2Mg0.7Co0.3MoO6−δ measured under 5%H2/Ar flow under open circuit conditions at 800 °C. The electrodes were prepared by different amounts of pore former.

Figure 9. Area-specific polarization resistance (ASR) values for the symmetrical cells Sr2Mg0.5Co0.5MoO6−δ/LSGM/Sr2Mg0.5Co0.5MoO6−δ measured under 5% H2/Ar at different calcination temperatures.

including a high-frequency (HF) arc and a low-frequency (LF) arc. The HF arc increases but the LF arc decreases with increasing amount of pore former. High pore former content causes high porosity and thus high specific surface area in electrode. The overall electrochemical reaction at the SOFC anode involves many steps such as gas diffusion, surface adsorption/dissociation, and charge transfer. The high porosity is favorable for gas diffusion, whereas the high specific surface area is certainly beneficial to the charge transfer and gas absorption/desorption processes due to the increased active

measured on symmetrical Sr2Mg0.5Co0.5MoO6−δ/LSGM/ Sr2Mg0.5Co0.5MoO6−δ configuration calcined at different temperatures (1200−1350 °C). The cell fired at 1300 °C shows the lowest resistance, whereas the one fired at 1200 °C gives the highest. Lower firing temperature easily leads to poor adhesion between the Sr2Mg0.5Co0.5MoO6−δ anode and LSGM electrolyte layers, whereas excessive higher firing temperature (e.g., 1350 °C) results in a reaction between the electrode and

Figure 10. SEM images of the cross-section of the Sr2Mg1−xCoxMoO6−δ/LSGM interface for the symmetrical cells sintered at 1300 °C (a) x = 0 and (b) x = 0.5. 9740



Article

Figure 12. Impedance spectra of the symmetrical cells Sr2Mg1−xCoxMoO6−δ/LSGM/Sr2Mg1−xCoxMoO6−δ measured under 5%H2/Ar under open circuit conditions under different temperatures (a) 700, (b) 750, (c) 800, and (d) 850 °C. The ohmic resistance (RO) has been subtracted for a better comparison. The equivalent circuit used to fit the impedance data was shown in inset.

Table 2. Fitting Results of Symmetrical Cells Sr2Mg1−xCoxMoO6−δ/LSGM/Sr2Mg1−xCoxMoO6−δ As a Function of Temperature in 5%H2/Ar 700 °C

composition x x x x x x

= = = = =

0 0.1 0.3 0.5 0.7

750 °C

800 °C

850 °C

RHF (Ω cm2)

RLF (Ω cm2)

RT (Ω cm2)

RHF (Ω cm2)

RLF (Ω cm2)

RT (Ω cm2)

RHF (Ω cm2)

RLF (Ω cm2)

RT (Ω cm2)

RHF (Ω cm2)

RLF (Ω cm2)

RT (Ω cm2)

3.28 1.61 0.67 0.46 0.32

5.01 2.84 2.04 1.77 1.31

8.29 4.45 2.71 2.23 1.63

1.80 0.85 0.43 0.28 0.19

3.22 1.63 1.06 0.85 0.73

5.02 2.48 1.49 1.13 0.92

0.53 0.49 0.13 0.11 0.07

1.29 0.77 0.79 0.54 0.42

1.82 1.26 0.92 0.65 0.49

0.29 0.32 0.09 0.03 0.015

1.24 0.76 0.53 0.46 0.295

1.53 1.08 0.62 0.49 0.31

fitted values were separated as the RHF (resistance at high frequency), R LF (resistance at low frequency), and R t (resistance of total arc), which are shown in Table 2. The ohmic resistance of the electrolyte was subtracted from the impedance spectra to have a comparison of Co-doping impact on the electrode polarization as a function of temperature. The resistance of the total arc (Rt) decreases with increasing temperature for all samples, indicating the thermal activation feature of electrode reactions. Not only the Rt but also the RHF and RLF exhibit an obvious decreasing tendency with Co content in SMCMO anodes, as shown in Table 2. The Nyquist plots for these anodes seem to change from two arcs to only a single arc. The discussion in Section 3.5.2 shows that the HF arc corresponds to oxygen ion transfer through electrode/ electrolyte interface and particle boundaries in the whole electrode, whereas the LF arcs are related to the charge transfer, gas adsorption/desorption, and gas diffusion processes. The decreased RHF is most probably due to the good sinterability of SMCMO anode with increasing Co content, which ensures

sites. High porosity in the electrode will result in poor contact with the electrolyte and fewer electrode particle/particle pathways, leading to high oxygen-ion transfer resistance between electrolyte and electrode and high resistance of finite-length oxygen-ion diffusion through the electrode. The results shown in Figure 11 demonstrated that the HF arc should correspond to the oxygen ion transfer process on the interface of electrode/electrolyte and between electrode particles, and the LF arc represents the catalytic charge-transfer process taking place on the electrode surface and gas-transportrelated process. This supposition is also supported by the result reported in literature.64−67 3.5.3. Co Content Effect. The AC impedance measurements were carried out on a series of symmetrical cells Sr2Mg1−xCoxMoO6−δ/LSGM/Sr2Mg1−xCoxMoO6−δ (x = 0, 0.1, 0.3, 0.5, and 0.7) at different temperatures to investigate the effect of Co on the kinetic process of SMCMO electrodes. Typical Nyquist plots and the equivalent circuits of the impedance spectra of these anodes are shown in Figure 12. The 9741



Article

resistance of SMCMO/LSGM/SMCMO symmetrical cells. The charge-transfer and gas absorption/desorption process is the rate-limiting step for the electrode reaction of SMCMO anode. Co doping can notably enhance the charge-transfer process of SMMO anode and thereby decrease the polarization resistance of electrode. Single-cell testing indicated that the electrocatalytic activity of SMMO electrode is remarkably improved with Co doping. By the optimization of the cell configuration, especially the electrode microstructure and the electrolyte thickness, the cell performance can be further improved. SMCMO is potential anode material for SOFCs.

good contact between electrode/electrolyte and electrode particles. The decreasing RLF suggests that Co doping in SMCMO can promote the charge-transfer process and gas adsorption/desorption processes. The RLF is much higher than the RHF for all investigated samples at any tested temperatures. This means that the charge transfer and gas adsorption/ desorption are step-limiting processes in SMCMO anodes. 3.6. Single-Cell Performance. Cell performance evaluation of SMMO and Co-doped SMMO was carried out in an electrolyte supported cell. Figure 13 presents the current−

■

AUTHOR INFORMATION

Corresponding Author

*Tel/Fax: +86-10-82376837. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was kindly supported by the National Nature Science Foundation of China (no. 20973021), Nature Science Foundation of Beijing (no. 2102031), and 863 Program of National High Technology Research Development Project of China (no. 2006AA11A189).

■

Figure 13. Cell voltage and power density as a function of current density for SMCMO with different Co contents at 800 °C under humidified pure hydrogen.

REFERENCES

(1) Sun, C. W.; Stimming, U. J. Power Sources 2007, 171, 247−260. (2) Steele, B. C. H.; Heinzel, A. Nature 2001, 414, 345−352. (3) Singh, P.; Minh, N. Q. Int. J. Appl. Ceram. Technol. 2004, 1, 5−15. (4) Brandon, N. P.; Skinner, S.; Steele, B. C. H. Annu. Rev. Mater. Res 2003, 33, 183−213. (5) Singhal, S. C. Solid State Ionics 2002, 152−153, 405−410. (6) Haile, S. M. Acta Mater 2003, 51, 5981−6000. (7) Minh, N. Q. Solid State Ionics 2004, 174, 271−277. (8) Atkinson, A.; Barnett, S.; Gorte, R. J.; Irvine, J. T. S.; Mcevoy, A. J.; Mogensen, M.; Singhal, S. C.; Vohs. J. Nat. Mater 2004, 3, 17−27. (9) Mclntosh, S.; Gorte, R. J. Chem. Rev. 2004, 104, 4845−4866. (10) Jiang, S. P.; Chan, S. H. J. Mater. Sci. 2004, 39, 4405−4439. (11) Fergus, J. W. Solid State Ionics 2006, 177, 1529−1541. (12) Tao, S.; Irvine, J. T. S. Chem. Rec. 2004, 4, 83−95. (13) Goodenough, J. B.; Huang, Y. H. J. Power Sources 2007, 173, 1− 10. (14) Miller, D. N.; Irvine, J. T. S. J. Power Sources 2011, 196, 7323− 7327. (15) Fagg, D. P.; Kharton, V. V.; Kovalevsky, A. V.; Viskup, A. P.; Naumovich, E. N.; Frade, J. R. J. Eur. Ceram. Soc. 2001, 21, 1831− 1835. (16) Hui, S. Q.; Petric, A. J. Eur. Ceram. Soc. 2002, 22, 1673−1681. (17) Vincent, A.; Luo, J. L.; Chuang, K. T.; Sanger, A. R. J. Power Sources 2010, 195, 769−774. (18) Tao, S. W.; Irvine, J. T. S. Nat. Mater. 2003, 2, 320−323. (19) Peña-Martínez, J.; Marrero-López, D.; Ruiz-Morales, J. C.; Savaniu, C.; Núnez, P.; Irvine, J. T. S. Chem. Mater. 2006, 18, 1001− 1006. (20) Lu, X. C.; Zhu, J. H. Solid State Ionics 2007, 178, 1467−1475. (21) Huang, Y. H.; Dass, R. I.; Denyszyn, J. C.; Goodenough, J. B. J. Electrochem. Soc. 2006, 153, A1266−A1272. (22) Huang, Y. H.; Dass, R. I.; Xing, Z. L.; Goodenough, J. B. Science 2006, 312, 254−257. (23) Vasala, S.; Lehtimäki, M.; Huang, Y. H.; Yamauchi, H.; Goodenough, J. B.; Karppinen, M. J. Solid State Chem. 2010, 183, 1007−1012. (24) Marrero-López, D.; Peña-Martínez, J.; Ruiz-Morales, J. C.; Pérez-Coll, D.; Aranda, M. A. G.; Núñez, P. Mater. Res. Bull. 2008, 43, 2441−2450.

voltage (I−V) and the corresponding power density (I−P) curves for a SCMO/LSGM/SmBaCo2O5+δ cell operating under pure H2 at 800 °C. The power density of the single-cell increased with increasing cobalt content in SMCMO. The increased electronic and ionic conductivities and the decreased polarization resistance for the electrode reactions should be responsible for the improved anode performance of Co-doped SMMO. The highest power density, obtained for a cell with SMCMO (x = 0.7) anode is 120 mW cm−2. After electrode microstructure optimization and electrolyte thickness reduction, the cell performance will be further improved. On the basis of these observations, the Sr2Mg0.3Co0.7MoO6−δ composition is a potential anode material for SOFCs.

4. CONCLUSIONS Double perovskite oxides of composition Sr2Mg1−xCoxMoO6−δ were synthesized by a simple citrate method. The solid solution limit of Co in Sr2Mg1−xCoxMoO6−δ after 1450 °C sintering in air and 800 °C reducing in 5%H2/Ar is x ≈ 0.9. The Co ions mainly take +2 valence in SMCMO. The substitution of Co for Mg does not significantly change the TEC of SMMO, Sr2Mg0.3Co0.7MoO6−δ having a TEC close to that of LSGM electrolyte in the temperature range of 50−1300 °C. SMCMO also shows a good chemical compatibility with LSGM electrolyte below 1300 °C. The electrical and ionic conductivity of Sr2MgMoO6−δ was remarkably enhanced by Co doping in the Mg site. DOS calculations revealed that Co doping considerably decreases the forbidden gap of SMMO and thus makes the electron hopping much easier. The weak Co−O bond is responsible for the increased ionic conductivity, which leads to more oxygen vacancies and lower activation energy for oxygen ion migration. Firing temperature has strong impact on the polarization 9742



Article

(61) Ruffa, A. R. J. Mater. Sci. 1980, 15, 2258−2267. (62) Hayashi, H.; Kanoh, M.; Quan, C. J.; Inaba, H.; Wang, S R.; Dokiya, M.; Tagawa, H. Solid State Ionics 2000, 132, 227−233. (63) Marrero-López, D.; Peña-Martínez, J.; Ruiz-Morales, J. C.; Gabás, M.; Núñez, P.; Aranda, M. A. G.; Ramos-Barrado, J. R. Solid State Ionics 2010, 180, 1672−1682. (64) Vernoux, P.; Guillodo, M.; Fouletier, J.; Hammou, A. Solid State Ionics 2000, 135, 425−431. (65) Lv, H.; Tu, H. Y.; Zhao, B. Y.; Wu, Y. J.; Hu, K. A. Solid State Ionics 2007, 177, 3467−3472. (66) Jiang, S. P.; Badwal, S. P. S Solid State Ionics 1999, 123, 209− 224. (67) Serra, J. M.; Vert, V. B.; Büchler, O.; Meulenberg, W. A.; Buchkremer, H. P. Chem. Mater. 2008, 20, 3867−3875.

(25) Bernuy-Lopez, C.; Allix, M.; Bridges, C. A.; Claridge, J. B.; Rosseinsky, M. J. Chem. Mater. 2007, 19, 1035−1043. (26) Ji, Y.; Huang, Y. H.; Ying, J. R.; Goodenough, J. B. Electrochem. Commun. 2007, 9, 1881−1885. (27) Marrero-López, D.; Peña-Martínez, J.; Ruiz-Morales, J. C.; Martín-Sedeño, M. C.; Núñez, P. J. Solid State Chem. 2009, 182, 1027−1034. (28) Zhang, L. L.; He, T. M. J. Power Sources 2011, 196, 8352−8359. (29) Vasala, S.; Lehtimäki, M.; Haw, S. C.; Chen, J. M.; Liu, R. S.; Yamauchi, H.; Karppinen, M. Solid State Ionics 2010, 181, 754−759. (30) Bi, Z. H.; Zhu, J. H. J. Electrochem. Soc. 2011, 158, B605−B613. (31) Huang, Y. H.; Liang, G.; Croft, M.; Lehtimäki, M.; Karppinen, M.; Goodenough, J. B. Chem. Mater. 2009, 21, 2319−2326. (32) Zhang, P.; Huang, Y. H.; Cheng, J. G.; Mao, Z. Q.; Goodenough, J. B. J. Power Sources 2011, 196, 1738−1743. (33) Zhang, L. L.; Zhou, Q. J.; He, Q.; He, T. M. J. Power Sources 2010, 195, 6356−6366. (34) Liu, Q.; Dong, X. H.; Xiao, G. L.; Zhao, F.; Chen, F. L. Adv. Mater. 2010, 22, 5478−5482. (35) Liu, Q.; Yang, C. H.; Dong, X. H.; Chen, F. L. Int. J. Hydrogen Energy 2010, 35, 10039−10044. (36) Wang, Z. M.; Tian, Y.; Li, Y. D. J. Power Sources 2011, 196, 6104−6109. (37) Zhu, X. F.; Wang, H. H.; Yang, W. S. Solid State Ionics 2006, 177, 2917−2921. (38) Tai, L. W.; Nasrallah, M. M.; Anderson, H. U.; Sparlin, D. M.; Sehlin, S. R. Solid State Ionics 1995, 76, 259−271. (39) Deng, Z. Q.; Zhang, G. G.; Liu, W.; Peng, D. K.; Chen, C. S. Solid State Ionics 2002, 152−153, 735−739. (40) Rodríguez-Carvajal, C. Physica B 1993, 192, 55. (41) Xie, Z. X.; Zhao, H. L.; Chen, T.; Zhou, X.; Du, Z. H. Int. J. Hydrogen Energy 2011, 36, 7257−7264. (42) Li, X.; Zhao, H. L.; Gao, F.; Zhu, Z. M.; Chen, N.; Shen, W. Solid State Ionics 2008, 179, 1588−1592. (43) Clark, S. J.; Segall, M. D.; Pickard, C. J.; Hasnip, P. J.; Probert, M. J.; Refson, K.; Payne, M. C. Z. Kristallogr 2005, 220, 567−570. (44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (45) Ivanov, S. A.; Eriksson, S.-G.; Tellgren, R.; Rundlöf, H.; Tseggai, M. Mater. Res. Bull. 2005, 40, 840−849. (46) Aguadero, A.; Alonso, J. A.; Martínez-Coronado, T.; MartínezLope, M. J.; Fernández-Díaz, M. T. J. Appl. Phys. 2011, 109, 034907. (47) Shannon, R. D. Acta Crystallogr. 1976, A32, 751−767. (48) Dasgupta, N.; Krishnamoorthy, R.; Jacob, K. T. Mater. Sci. Eng., B 2002, 90, 278−286. (49) Matsumoto, Y.; Yamada, S.; Nishida, T.; Sato, E. J. Electrochem. Soc. 1980, 127, 2360−2364. (50) Xia, T.; Lin, N.; Zhao, H.; Huo, L. H.; Wang, J. P.; Grenier, J. C. J. Power Sources 2009, 192, 291−296. (51) Li, S. Y.; Lü, Z.; Huang, X. Q.; Su, W. H. Solid State Ionics 2008, 178, 1853−1858. (52) Tai, L. W.; Nasrallah, M. M.; Anderson, H. U.; Sparlin, D. M.; Sehlin, S. R. Solid State Ionics 1995, 76, 273−283. (53) Sun, X. F.; Guo, R. S.; Li. J. Ceram. Int. 2008, 34, 219−223. (54) Segall, M. D.; Lindan, P. J. D.; Probert, M. J.; Pickard, C. J.; Hasnip, P. J.; Clark, S. J. J. Phys.: Condens. Matter 2002, 14, 2717− 2744. (55) Voorhoeve, R. J. H.; Remeika, J. P.; Trimble, L. E. Ann. N.Y. Acad. Sci. 1976, 272, 3−21. (56) Mori, M.; Sammes, N. M. Solid State Ionics 2002, 146, 301−312. (57) Kim, J. H.; Cassidy, M.; Irvine, J. T. S.; Bae, J. J. Electrochem. Soc. 2009, 156, B682−B689. (58) Zhou, W.; Sunarso, J.; Chen, Z. G.; Ge, L.; Motuzas, J.; Zou, J.; Wang, G. X.; Julbe, A.; Zhu, Z. H. Energy Environ. Sci 2011, 4, 872− 875. (59) Ullmann, H.; Trofimenko, N.; Tietz, F.; Stöver, D.; AhmadKhanlou, A. Solid State Ionics 2000, 138, 79−90. (60) Aguadero, A.; Pérez-Coll, D.; Calle, C. D. L.; Alonso, J. A.; Escudero, M. J.; Daza, L. J. Power Sources 2009, 192, 132−137.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on April 20, 2012, with an error to Figure 7. The corrected version was reposted on April 25, 2012.

9743


Effects of Co Doping on the Electrochemical Performance of Double

Recommend Documents