Characterization of electrode polarization losses in solid oxide fuel cells: Impedance spectroscopy involving spatially-limited electrode geometry

Journal of Physics and Chemistry of Solids 74 (2013) 496–503

Contents lists available at SciVerse ScienceDirect

Journal of Physics and Chemistry of Solids journal homepage: www.elsevier.com/locate/jpcs

Characterization of electrode polarization losses in solid oxide fuel cells: Impedance spectroscopy involving spatially-limited electrode geometry Byung-Kook Lee a, Seung-Muk Bae a, Jong-Ho Lee b, Jinha Hwang a,n a b

Department of Materials Science and Engineering, Hongik University, Seoul 121–791, Korea High-Temperature Energy Materials Center, Korea Institute of Science and Technology, Seoul 136–791, Korea

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 August 2012 Received in revised form 29 October 2012 Accepted 26 November 2012 Available online 5 December 2012

The concept of ‘‘spatially-limited geometry’’ in an ionic conductor is incorporated into AC two-probe impedance spectroscopy in order to investigate the electrode-related responses of solid oxide fuel cells (SOFCs). A semispherical cylinder-shaped ionic conductor made of yttria–stabilized zirconia (YSZ) is applied to the planar electrode, and the electrical and electrochemical losses in the electrolyte and electrode are characterized. According to our study, the spatial constriction of the contact interface amplifies the electrode-related impedances, and can more efficiently separate the bulk-related impedance from the electrode polarization that is very sensitive to the microstructural features of the electrode such as size and distribution of constituent phases. In this study, the resolved bulk resistances are analyzed using the concept of ‘‘spreading resistance,’’ whereas the resulting polarization losses in the electrolyte/electrode interface are analyzed in combination with an equivalent circuit model as a function of the resistances and constant phase elements. The present approach with spatially-limited contact impedance spectroscopy is fully explored for various electrode materials in order to confirm the validity of current methodology. & 2012 Elsevier Ltd. All rights reserved.

Keywords: A. Interfaces A. Oxides D. Electrical properties D. Electrochemical properties D. Transport properties

1. Introduction Over the last two decades, solid oxide fuel cells (SOFCs) have attracted widespread attention due to the increased petroleum price and various environmental issues associated with air pollution [1–5]. SOFCs are highly efficient electrochemical conversion systems which are usually operated at high temperatures near 1000 1C, in which oxygen and hydrogen are electrochemically reacted to generate electric power. In general, the high operating temperatures of SOFCs can introduce numerous undesired technical problems due to the chemical instability and mechanical discordance between dissimilar materials in the SOFC components, including the electrolytes, electrodes, interconnects, and sealing materials. Hence to overcome these problems, most research and development activities have focused on the development of intermediate-temperature solid oxide fuel cells IT-SOFCs that can be operated at a temperature range of 600–800 1C [2–5]. However, as the operating temperature is lowered, the polarization loss of a cell also greatly increases, mainly due to the electrode overpotential, especially at the cathode. In order to prevent significant electrodic polarization loss at lower temperature, electrode has to be carefully designed to keep their electrochemical activity as high as possible by maximizing the

n

Corresponding author. E-mail address: [email protected] (J. Hwang).

0022-3697/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jpcs.2012.11.018

number of electrode reaction sites as well as promoting the supplying of reactants or removal of products from the electrode reaction. Hence in order to optimize SOFC electrode, it should be thoroughly exploited in terms of its physicochemical properties and corresponding electrochemical functions. There have been substantial efforts in the research and development of high performance electrode. Nevertheless, still some unsolved issues are remained [6–10]. These issues are closely ascribed to the lower electrochemical efficiency of the electrodes, such as activation polarization loss or concentration polarization loss at the electrode. In general, most of the challenging issues are strongly dependent on the microstructural factors of the electrode, such as the volume fraction, interconnectivity, and size distribution of the constituent components, i.e., solid phases and pores. In particular, optimization of the triple phase boundary (TPB) configuration, which has been recognized as a critical issue in achieving high electrochemical conversion efficiencies for the electrode reaction, is strongly influenced by the constitutional forms of the solid electrolyte, electrode, and pore phase. In addition to the extensive research and development regarding the overall cell performance of SOFCs, there have been many studies intended to provide an exhaustive understanding of the correlation between the microstructure and its related electrodic functions, either at the anodes or cathodes [6–10]. Impedance spectroscopy has been an exceedingly powerful technique to separate the bulk responses from the grain boundary

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

features in electroceramics, such as ZnO varistors, BaTiO3 PTCRs, and magnetic ferrites [11–14]. Impedance spectroscopy allows the simultaneous monitoring of electrical and dielectric phenomena given an arc or semicircle in a Nyquist plot (an imaginary part of measured impedance (-Im(Z)) versus a real impedance part (Re(Z))). The conductivity and dielectric constants can be estimated simultaneously using an equivalent circuit model. The degree of depression in the impedance arcs is evaluated in order to find the homogeneity of the relevant mechanisms. Impedance spectroscopy is also useful for the study of the cement-based materials, and batteries, and to monitor the corrosion behavior of metals in aqueous environments [15–17]. Impedance spectroscopy is now becoming a main tool to characterize the electrolyte/electrode properties in fuel cells, especially SOFCs [18–20]. A detailed understanding of the rate-limiting steps in SOFC electrode reactions is highly desirable from an optimization point of view. In general electrode reaction comprises multiple electrochemical as well as physical or chemical transport steps [6–8,18–20]. It is important to find out the major limiting process among various reaction steps which are normally related to the microstructure of electrode and surroundings, such as contacting state and gas supplying condition. In that sense, impedance spectroscopy is very efficient to analyze this complex system. Until now, AC three-probe impedance spectroscopy has generally been applied to SOFCs in order to separate the electrical/ dielectric features of the bulk from the electrochemical response from the electrodes. In the sophisticated three-probe cell design, which guarantees parallel equi-potential lines across the solid electrolyte, it is possible to separate the electrochemical responses originating from the electrode as long as the impedance of the reference electrode is smaller than the inherent impedance of the frequency response analyzer [20,21]. However, if the appropriate configuration in three-point impedance spectroscopy is not guaranteed, the obtained electrode information can exhibit erroneous arc features in electrode-based responses due to the undesired distortion in equi-potential lines [22]. The electrode placement similar to the Luggin capillary probe is preferred in terms of the measurement accuracy in estimating the electrolyte and electrode parameters [23]. Recently, the spatially-limited contact AC two-probe configuration was introduced that is expected to substitute conventional AC three-probe impedance spectroscopy to resolve the electrical responses of one electrode from the bulk electrolyte by amplifying the resistive part of the relevant electrochemical responses associated with that electrode along with the bulk response. One of the main advantage of this spatially-limited contact electrode with AC two-probe impedance spectroscopy is that it does not suffer from the limited impedance magnitude of reference electrodes encountered in AC three-probe impedance spectroscopy [20,21], thereby enabling a proper investigation of the electrode responses in SOFCs with a two-probe configuration. The concept of this limited-contact method is very useful to separate the information from the electrolyte/electrode interface near the point-contact between the electrolyte and electrode, which is believed to be qualitatively equivalent to the information obtained from AC three-probe impedance spectroscopy. The spatially-limited contact geometry can be formed into either a semispherical YSZ probe onto the cathode materials or a semispherical electrode probe onto the electrolytes [18,19,24]. The latter case was reported by Baker et. al. 18 and Fleig [24], where a limited contact was constructed based on the electrode materials. However, since the cathode materials are normally porous and not so robust, the elaborate machining of cathode into the spherical shape is not straightforward, which may lead to mechanical failure. On the other hand, spatially-limited contact based on a robust electrolyte such as YSZ electrolyte is expected to give more reliable measurement setup.

497

In this study, we report on a unique characterization method for the electrode reactions of SOFCs via impedance spectroscopy combined with a spatially-limited contact AC two-probe configuration using a semispherical cylinder-shaped ionic conductor. With this spatially-limited configurations based on semispherical cylindershaped electrolyte, we can intentionally amplify the electrode and bulk impedances, expecting to obtaining a universal parameter which allows a straightforward comparison between the electrodes. In order to validate the current spatially-limited contact method, various materials were chosen for the testing electrode, Unlike a preliminary study [19] incorporating composite cathode materials, the current approach was applied to three types of cathode systems found in cathodes of SOFCs, i.e., electronically-conducting metallic, electronically-conducting ceramic, and mixed-conducting electrodes. The ramifications of ‘‘limited-contact’’ impedance spectroscopy are discussed for an exhaustive understanding of electrode-related polarizations in SOFCs.

2. Experiment In order to avoid the vulnerability in shaping the electrode into the pointed configuration, we devised the limited contact in the electrolyte by machining the dense sintered body of YSZ (Tosoh, TZ8Y) into a semispherical cylinder. One side of the YSZ specimen had a semispherical shape and the other side maintained a cylindrical form. The hemispherical YSZ probe was fabricated so that the diameter of the YSZ probe is approximately 8 mm and the lateral portion of the circular column is approximately 2 mm long. The upper portion of the electrolyte was connected to the platinum lead wires using platinum paste and platinum mesh, in order to guarantee close contact between the YSZ and the Pt wire electrode. The mechanical contact between the Pt mesh and the Pt-painted surface was treated at 800 1C for 1 h in an electric furnace. After the complete electroding of the semispherical YSZ electrolyte, the lower spherical portion was placed on the planar top portion of the electrode material. Pure electronic conducting gold (Au) was employed as a reference material in order to verify the current ‘‘limited-contact’’ impedance spectroscopy. LSM ((La0.7Sr0.3)0.95 MnO3) and LSCo (La0.6Sr0.4CoO3) which are the typical electronic conducting and mixed ionic–electronic conducting cathode materials, respectively, were also incorporated into current configuration for checking the viability of this methodology. Au was employed in the form of thin foil whereas the ceramic electrode was employed in the form of disk-shaped bulk. The synthesized LSM powder using a glycine nitrate process [19] and commercial LSCo (AGC SEIMI CHEM, Japan) powder were compacted into disks using uniaxial pressing and heat-treated in a furnace at 600 1C after binder removal at 400 1C. The pre-sintered specimen was placed into a furnace and then heated to 800 1C to create a complete contact between the electrode materials and the Pt-pasted counter electrode. As a final step, the specimens were contacted with the semispherical part of YSZ electrolyte and then heat-treated at high temperature around 1200 1C for 1 h, in order to mimic actual cathode fabrication conditions of SOFCs. The final configuration of our spatially-limited contact electrode for AC two-probe impedance spectroscopy was shown in Fig. 1. Impedance spectra were collected as a function of temperature and oxygen partial pressure using a frequency response analyzer (SI 1260, Solatron, Berkshire, UK). Since the electrolyte/electrode interface is sensitive to the oscillating amplitude, an oscillating voltage was fixed at 25 mV in order to minimize the effects induced by the high oscillating voltage. The impedance spectra were acquired in a logarithmic manner between 1 MHz and 0.01 Hz with 10 points per decade. Impedance measurements were collected between 600 1C and 800 1C in an electric furnace which

498

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

was controlled using a programmable temperature controller. The oxygen partial pressure was controlled between 1 and 10  4 atm using pure oxygen and a commercial mixed gas with oxygen and argon gas.

3. Results and discussion Fig. 2(a) and (b) compare the impedance spectra from Au electrode at 750 1C under various oxygen partial pressures. The temperature dependencies are also shown in Fig. 2(c) and (d). Those in Fig. 2 (b) and (d) represent the enlarged portion of the impedance spectra, which are more relevant for estimation of the ohmic

Pt wire Pt mesh

2 8mm

4

YSZ

resistance of the cell. As shown in Fig. 2, the incorporation of a limited contact as an electrode leads to much higher value in electrode-related resistances (in the range of approximately 2  102–2  103 O) compared to the symmetrical parallel electrodes using platinum pastes [25]. The corresponding impedance and capacitance are shown as a function of frequency in Fig. 3; the impedance increased with decreasing oxygen partial pressure, and the capacitance increased with increasing oxygen partial pressure, especially in the lower frequency part originating from the electrode polarization. In the ‘‘limited-contact’’ configuration, the capacitance was reduced and the impedance was increased, compared to the capacitance and impedance from the planar electrode-type impedance measurement (not shown here). The limited-contact electrode revealed the presence of two semi-circles with noticeable depressions in the Cole–Cole plot in Fig. 2. Based on the obtained impedance characteristics, an equivalent circuit can be constructed, as shown in Fig. 4. At high frequencies, the intercept on the real impedance axis reflects the total ohmic resistance that originates from the electrolyte and the lead wires. RHighfreq, ¼ RLeads þRElectrolyte

Electrode Pt wire Fig. 1. The ‘‘spatially-limited contact’’ design and configuration applied to AC twoprobe impedance spectroscopy.

In our measurement setup, the resistance of the lead wires was negligible compared to the total electrolyte/electrode resistance, which had been amplified through the spatially-limited contact. Furthermore, even though the current electrode configuration was different from that associated with the conventional pointcontact measurement [24], the current flow was expected to be concentrated near the spherical portion of the electrolyte as shown in Fig. 5. Hence the equi-potential lines were expected to

1.4 104

2000 1 atm 0.21 atm 0.1 atm 0.01 atm 0.001 atm 0.0001 atm

4

1 atm 0.21 atm 0.1 atm 0.01 atm 0.001 atm 0.0001 atm

1500

-Im(Z) [Ohm]

-Im(Z) [Ohm]

1.2 104 1 10

ð1Þ

8000 6000 4000

1000

500

2000 0

0

2000

4000

6000

8000

0

1 104 1.2 104 1.4 104

0

500

1500

2000

4000

1 105 600°C

600°C

3500

700°C

8 104

700°C

3000

-Im(Z) [ohm]

800°C

-Im(Z) [ohm]

1000

Re(Z) [Ohm]

Re(Z) [Ohm]

4

6 10

4

4 10

800°C

2500 2000 1500 1000

4

2 10

500 0

0

2 104

4 104

6 104

Re(Z) [Ohm]

8 104

1 105

0

0

500

1000 1500 2000 2500 3000 3500 4000

Re(Z) [Ohm]

Fig. 2. Typical impedance spectra obtained through a spatially-limited contact in an electrolyte. (Measurement temperature: 750 1C, Oscillating Voltage: 25 mV). (a) Oxygen partial pressure dependence, (b) enlargement of Fig. 2(a), (c) temperature dependence (1% oxygen partial pressure), and (d) enlargement of Fig. 2(c).

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

105

499

10-4 1atm 0.21atm 0.1atm 0.01atm 0.001atm 0.0001atm

Capacitance [F]

10

|Z| [Ohm]

104

1atm 0.21atm 0.1atm 0.01atm 0.001atm 0.0001atm

-5

1000

10-6 10-7 10-8 10-9 10-10

100 0.1

1

10

100

104

1000

105

106

10-11 0.1

1

1.5 10-5

Capacitance [F]

10

100

1000

104

105

106

Frequency [Hz]

Frequency [Hz]

1atm 0.21atm 0.1atm 0.01atm 0.001atm 0.0001atm

1 10-5

5 10-6

0.1

1

10

100

Frequency [Hz] Fig. 3. Bode plot of (a) impedance and (b) capacitance of the data in Fig. 2(a). (c) Enlargement of the low-frequency portion of Fig. 3(b).

R0

R1

R2

CPE1

CPE2

Fig. 4. An equivalent circuit model for the spatially-limited contact impedance spectra.

be qualitatively similar to those of the conventional point-contact measurement, in which a point-contacted electrode is placed on a planar electrolyte [26]. Accordingly, the resistance of the electrolyte was a summation of the bulk and spreading resistances as described by the following equation, RElectrolyte ¼ RBulk þRSpreading

ð2Þ

where the spreading resistance was the additional resistance induced by the localized current flow around a point-contacted electrode, as described by the following equation: RSpreading ¼

1 4sr

ð3Þ

In Eq.(3) RSpreading is a kind of contact resistance, s is the conductivity of the electrolyte, and r is the radius of the contact area. The spatially-limited contact leads to a much smaller radius in Eq. (3) compared to the apparent disk radius, so that RSpreading becomes much larger than RBulk. Then Eq. (2) can be simplified into the following equation: RElectrolyte  RSpreading

ð4Þ

By using the conductivity of the YSZ electrolyte and the obtained contact resistance from the impedance measurement, the contact radius could be calculated, which led to the circumference of the limited-contact electrode that is equivalent to the length of the triple phase boundary (TPB) of pure electronic conducting electrode like Au electrode. From the estimated resistance of 582 O from the high-frequency intercept in Fig. 2(d) and the measured conductivity of 0.015 (O cm)  1 of the polycrystalline YSZ at 700 1C, the contact radius was calculated using Eq. (1) to be approximately 300 mm. The micrograph in Fig. 6 shows the deformed electrode where the interface was formed between the YSZ electrolyte and the Au electrode. The circumference of the contacted region was measured to be around 600–900 mm, which is in reasonable order of magnitude agreement with the numerical estimation from Eq. (1). The spatially-constricted shape of the electrolyte was also effective in determining the polarization resistance originating from the electrode. In Fig. 2(b), the intercept at a high-frequency on the real impedance axis was approximated by Eq. (1), whereas the remaining part of the resistance was assigned to the electrode impedance as shown in the following equation, Z Lowfreq ¼ Z Planar þ Z Limited

ð5Þ

where ZPlanar originates from the parallel part, i.e., the bottom electrode in Fig. 1, and ZLimited is from the interface adjacent to the top electrode/electrolyte interface. The schematic diagram of the electrode configuration led to the conclusion that the impedance of the planar electrode was much smaller than ZLimited due to the

500

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

Electrode

YSZ

Electrode

Electrode

Fig. 5. Schematic equi-potential line diagrams (a) in a parallel symmetrical electrode configuration and (b) in an asymmetric electrode configuration involving spatial constriction at the electrode/electrolyte interface.

500 µm µ

Fig. 6. Microphotograph of the spatially-limited contact between the YSZ and the Au. The deformed gold electrode shows the close contact between electrolyte and electrode.

limited-contact electrode and the significantly different effective area ratios of the two electrodes. Thus, the low-frequency impedance was simplified into the following equation, Z Lowfreq ¼ Z Planar þ Z Limited  Z Limited

ð6Þ

The resulting low-frequency impedance spectrum originated solely from the geometrically-limited portion, i.e., the contact part between the cathode and the semispherical electrolyte. The low-frequency part of the impedance spectra was analyzed in terms of two resistors and two constant phase elements (CPEs) as shown in Fig. 4. The incorporation of constant phase elements in this study reflected the depression of the impedance arcs [27]. The constant phase elements are empirical formulations, expressed mathematically as ZðCPEÞ ¼ AðjwÞn

ð7Þ

where A is a constant, w is the angular frequency, and n is the measure of arc depression which has a value less than 1. Although the meaning of a constant phase element is not exactly the same as that of a capacitance, the characteristics are quite similar. If n is equal to 1, i.e., there is no depression in the impedance arc, the constant phase element is a capacitor: the constant A in Eq. (7), is equivalent to the ideal capacitance, C. The resolved resistances from the theoretical fitting of each impedance arc are plotted in Fig. 7 as a function of the oxygen partial pressure range from 1 to 10  4 atm. As shown in Fig. 7, the resistance

estimated from the high-frequency intercept, R0 was independent of the oxygen partial pressure, since it originated primarily from the bulk portion of the purely ionic conducting YSZ electrolyte. The remaining impedance spectra, which represent the electrode polarizations, were analyzed based on the combination of two resistors (R1 and R2) and two constant phase elements (CPE1 and CPE2) as proposed in Fig. 4. According to the results in Fig. 7, R1 and CPE1 are almost constant over the oxygen partial pressure range under investigation, whereas R2 and CPE2 vary with respect to the oxygen partial pressure. The depression of the arc due to R1 and CPE1 was significant: n1 ranged between 0.46 and 0.69, while the depression of the arc due to R2 and CPE2 was relatively small: n2 was between 0.85 and 0.96. Given the oxygen partial pressure dependencies, the R1 value was attributed to the charge transfer reaction resistance at the TPB of the electrode, whereas R2 was attributed to the surface diffusion resistance, which normally shows oxygen partial pressure dependence with a log exponent of 1/4 [28]. The high-frequency resistance, R0, is attributed to the bulk (electrolyte) resistance whose magnitude is related to current spreading at the geometricallyconstricted interfaces between YSZ electrolytes and cathodes. The magnitude is much larger than that measured in the parallel electrode configuration. Generally, the cathode reactions can be described in terms of adsorption/desorption of oxygen molecules, ionization of adsorbed oxygen ions, transport to triple phase boundaries, and charge transfer at TPBs in electronic conducting cathodes, as shown in Table 1 [28–30]. In mixed-conducting cathodes, the charge transfer takes place through the incorporation of the interfaces involving electronic carriers. From the current oxygen partial pressure dependence of polarization losses obtained in Au electrodes, the lowfrequency resistance is attributed to the surface diffusion to TPB sites with the exponent of 1/4: the polarization loss of R2 increases systematically with decreasing oxygen partial pressure. The intermediate-frequency electrode resistance, R1 is dominated by charge transfer at triple phase boundaries since the resistance is independent of oxygen partial pressures. The temperature dependence of the Ohmic conductance (the reciprocal of the Ohmic resistance, R0) can be used to calculate the activation energy of the ionic conduction in the YSZ electrolyte. According to Fig. 8, the activation energy was estimated to be 0.94 eV, which is very similar to the reported value for YSZ, which is approximately 0.95 eV [31]. The activation energies of R1 (the intermediate-frequency portion of the impedance spectra) and R2 (the low-frequency portion of the impedance spectra)

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

Table 1 Various reaction models for oxygen reduction at SOFC cathodes.

105 R0

o 750 C

R1

Reaction step

Resistance [Ohm]

R2

1000

100 10-5

0.001

0.01

0.1

Model for mixed-conducting cathode 1 O2 ðgÞ-2Oad 2 Oad þ e-O ad  3 O ad -OINTERFACE  4 OINTERFACE þ e-O2 INTERFACE

1

0.0001

1 750oC

X  O2 INTERFACE þ V O -OO

5 0.9

0.8

m of Pm O2

10-5

0.7

CPE(2)

0.0001

0.001

0.6

0.01

0.1

1

0.5

PO2 [atm]

– 3/8 1/4 – 0

1/R0 1/R1 1/R2 E(R0)=0.94eV

-7 -8

E(R1)=1.07eV

-9

were found to be 1.07 and 0.93 eV, respectively, even though there was some degree of scattering in the data with regard to the temperature. In addition, the current approach allowed the estimation of a new materials constant, i.e., the electrode resistance per unit length. This newly-defined value could be employed to describe the electrical characteristics of the electrode, which were controlled by the microstructure of the electrode. As explained before, the high-frequency impedance was dominated by the electrolyte adjacent to the semispherical limited contact, whereas the low-frequency impedance was controlled predominately by the electrode polarization associated with the geometricallylimited contact. In other words, the resulting impedance can be interpreted as the total summation of the bulk and electrode compartments near the semispherical contact. Then the impedance spectrum provides a resistance-based ratio constructed of the high-frequency and low-frequency resistances, RElectrode/Bulk, ð8Þ

The above ratio allows a relative comparison of polarization behaviors, which are dependent upon the composition and microstructure of the electrodes, specifically both cathodes and

E(R2)=0.93eV

-11 -12 0.0009

RElectrodeðSpreadingÞ RBulkðSpedaingÞ

Adsorption/Dissociation Ionization of oxygen Surface and/or bulf diffusion Ionization of oxygen at interface Charge transfer at interface

-10

Fig. 7. Representation of the circuit parameters associated with equivalent circuit models. (a) Resistances as a function of oxygen partial pressure, and (b) constant phase elements as a function of oxygen partial pressure.

RElectrode=Bulk ¼

1/4 – 0

-6

ln(1/R) [Ohm-1]

-5

10

– 3/8

Charge transfer at TPB

-5

n(1) n(2)

CPE(1)

– 3/8 1/4 –

-4

n

Constant Phase Element [F]

Reaction type

TPB TPB X  O2 TPB þ V O -OO

5 0.0001

PO2 [atm]

10

Reaction equation

Model for electronic conducting cathode 1 O2 ðgÞ-2Oad Adsorption/dissociation 2 Oad þ e-O Ionization of oxygen ad  Surface diffusion 3 O ad -OTPB X X 4 Charge transfer at TPB O TPB þ e þ V O -OO Model for composite cathode 1 O2 ðgÞ-2Oad Adsorption/Dissociation 2 Oad þ e-O Ionization of oxygen at ad adsorption site   Surface diffusion 3 Oad -OTPB 4 Ionization of oxygen at TPB O þ e-O2

104

-6

501

0.00095

0.001

0.00105

0.0011

0.00115

1/T [K-1] Fig. 8. Temperature dependence of various resistances due to the electrolyte and electrodes in the Au/YSZ/Pt system. (R0: the resistance due to the electrolyte and the lead, R1: the resistance in the intermediate frequency range and R2: the resistance in the lowest frequency range).

anodes with regard to the electrolyte being used, which in this work was YSZ. In order to check the viability of current methodology, limited contact impedance spectroscopy was applied to a various cathode materials, including electronically-conducting and/or mixedconducting oxide electrodes. LSM represented the electronicallyconducting electrode whereas LSC did the mixed ionic–electronic conducting electrode. As like for the previous characterization of Au electrode, the temperature dependence of the intercepts from impedance spectra of each electrode material would lead to the discrete calculation of the bulk behavior originating from the bulk YSZ around the geometrically-limited contact regions. Simultaneously, the low-frequency Cole plot would allow us to determine the resistances and constant phase elements of the corresponding electrodes. According to the results in Fig. 9, the impedance spectra show dissimilar characteristics with respect to the cathode materials even though the high-frequency intercept on the real impedance axis was still kept reflecting the contribution of

502

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

Table 2 Summary of activation energies and a ratio of the electrode polarization (Rt: RElectrode(Spreading)) to the bulk electrolyte resistance (R0:RBulk(Spreading)) calculated with the Eq. (8).

-1200 Au

800oC

LSM

-Im(Z) [Ohm]

-1000

LSCo

-800

Au LSM

-600

Ea (R0)

Ea(Rt)

Rt/R0 (800 1C)

Rt/R0 (700 1C)

Rt/R0 (600 1C)

0.87 eV 0.79 eV

1.50 eV 1.56 eV

8.72 2.55

10.66 5.07

45.00 15.18

50 Au

-400

LSM

(=R /R )

40

LSCo

0

0

400

800

1200

1600

2000

2400

30

20

R

Re(Z) [Ohm]

Electrode/Bulk

t

0

-200

0

10

Im(Z) [Ohm]

-100

0 550

-200

600

650

700

750

800

850

o

T [ C] -300

Fig. 10. Ratio of the electrode polarization to the bulk electrolyte calculated with the Eq. (8).

-400 Au LSM LSCo Pt [23]

-500 -600

0.01

0.1

1

10

100

1000

104

105

106

Frequency [Hz] Fig. 9. (a) Impedance spectra for a various electrode materials in an oxidizing atmosphere at 800 1C. (b) Impedance spectra in frequency domain.

the ionic conduction in YSZ. Unlike the high-frequency feature, the low-frequency impedance plots were not identical, due to variations in relative contribution of electrode material to the electrode reaction at the triple phase boundaries. Impedance spectra in frequency domain (Fig. 9 (b)) showed that each electrode material had different spectral features with different time constants which indicated different electrode reaction mechanism were working for each electrode material. As shown in Fig. 9 (b), the impedance spectrum of Au electrode showed similar behavior with that of pointed Pt electrode [25] which can be fitted well with the equivalent circuit model in Fig. 4. However the other ceramic materials, LSM and LSC showed quite different and complex characteristics which seem to be analyzed with more complex equivalent circuit model. However further detailed mechanistic study is beyond the scope of this paper. Hence for the simplicity of comparison, we only separated the ohmic contribution from the overall electrode impedance obtained at three different temperatures and calculated the aforementioned new materials constant, i.e., the electrode resistance per unit length which has been defined as the resistance-based ratio constructed of the high-frequency and low-frequency resistances, RElectrode/Bulk in Eq. (8). The resulting data are listed in Table 2 and plotted in Fig. 10. As shown in Fig. 10, among the electrodes tested in our experiment, the LSC materials showed the lowest

RElectrode/Bulk value which is in reasonable agreement with other reports presenting the better electrochemical property of MIECbased cathode than pure electronic or electron conducting prevailing materials such as Au and LSM [2,32,33]. In summary, the current characterization was made using a simplified two electrode configuration. Unlike conventional twoand three-electrode impedance spectroscopy, the current ‘‘spatially-limited contact’’ two-probe impedance spectroscopy allowed the simultaneous analysis of the bulk and the adjoining electrode contributions. Generally, three-point impedance spectroscopy is meaningful as long as the contact impedance is much smaller than the impedance of the instrument. However, the current methodology puts no restriction on the instrument impedance, in terms of electrodes and electrolytes. Because of the efficient characterization of the electrode materials, numerous electrodes can be searched and optimized with regard to a specific electrolyte, irrespective of electronic, mixed-conducting and composite electrodes. Nonetheless, there are still some technical issues in the methodological aspects. For example, the current approach is highly sensitive to any change in the local surroundings at the interfacial region regarding the contact status and microstructure. Hence, it is essential to secure an improved contact state in order to prevent undesired effects on the impedance spectra which may result in unreasonable interpretations. Further experiments and explanations will be provided in an upcoming paper in order to justify the applicability of our ‘‘spatially-limited contact’’ electrode geometry.

4. Conclusions The geometrical constriction was fabricated between the electrode and the electrolyte using a semispherical electrolyte

B.-K. Lee et al. / Journal of Physics and Chemistry of Solids 74 (2013) 496–503

and two-point electrode impedance spectroscopy in order to separate the interfacial responses encountered in solid oxide fuel cells. This configuration allows quantitative determination of the polarization due to a single electrode with regard to the given electrolyte. The information was analyzed in terms of an equivalent circuit model composed of three resistors and two constant phase elements: the contact resistance of high-frequency was allocated to the spreading resistance between the electrolyte and electrode. Two parallel circuits (R.CPE) can be properly separated in our limited-contact impedance spectroscopy due to the difference in the time constant, which was not the case for a planar electrode, in which the impedance arcs were overlapped and thus were difficult to resolve in order to discern the essential features of electrode reactions.

Acknowledgment This research was supported by the Seoul R&BD Program (CS070157) and by a grant from the Fundamental R&D Program for Core Technology of Materials funded by the Ministry of Knowledge Economy, Republic of Korea. References [1] B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies, Nature 414 (2001) 345–352. [2] Z. Shao, W. Zhou, Z. Zhu, Advanced synthesis of materials for intermediatetemperature solid oxide fuel cells, Prog. Mater. Sci. 57 (2012) 804–874. [3] E.-O. Oh, C.-M. Whang, Y.-R. Lee, S.-Y. Park, D. Hari Prasad, K.J. Yoon, J.W. Son, J.-H. Lee, H.-W. Lee, Extremely thin bilayer electrolyte for solid oxide fuel cells (SOFCs) fabricated by chemical solution deposition (CSD), Adv. Mater. 24 (2012) 3373–3377. [4] S. Tao, J.T.S. Irvine, Discovery and characterization of novel oxide anodes for solid oxide fuel cells, Chem. Rec. 4 (2004) 83–95. [5] D.J.L. Brett, A. Atkinson, N.P. Brandon, S.J. Skinnerd, Intermediate temperature solid oxide fuel cells, Chem. Soc. Rev. 37 (2008) 1568–1578. [6] S.B. Adler, Factors governing oxygen reduction in solid oxide fuel cell cathodes, Chem. Rev. 104 (2004) 4791–4843. [7] M.J. Jørgensenz, M. Mogensen, Impedance of solid oxide fuel cell LSMO/YSZ composite cathodes, J. Electrochem. Soc. 148 (2001) A433–A442. [8] M. Mogensen, S. Skaarupb, Kinetic and geometric aspects of solid oxide fuel cell electrodes, Solid State Ionics 86-88 (1996) 1151–1160. [9] J.R. Wilson, W. Kobsiriphat, R. Mendoza, H.-Y. Chen, J.M. Hiller, D.J. Miller, K. Thornton, P.W. Voorhees, S.B. Adler, S.A. Barnett, Three-dimensional reconstruction of a solid-oxide fuel-cell anode, Nat. Mater. 5 (2006) 541–544. [10] K.-R. Lee, Y.S. Pyo, B.S. So, S.M. Kim, B.K. Lee, J.H. Hwang, J. Kim, J.-H. Lee, H.-W. Lee., Interpretation of the interconnected microstructure of an NiO-YSZ anode composite for solid oxide fuel cells via impedance spectroscopy, J. Power Sources 158 (2006) 45-51. [11] A.R. West, M. Andres-Verges, Impedance and modulus spectroscopy of ZnO varistors, J. Electroceram. 1 (1997) 125–132.

503

[12] I.M. Hodge, M.D. Ingram, A.R. West, Impedance and modulus spectroscopy of polycrystalline solid electrolytes, J. Electroanal. Chem. 74 (1976) 125–143. [13] J.T.S. Irvine, D.C. Sinclair, A.R. West, Electroceramics: characterization by impedance Spectroscopy, Adv. Mater. 2 (1990) 132-138. [14] J.T.S. Irvine, A. Huanosta, R. Valenzuela, A.R. West, Electrical properties of polycrystalline nickel zinc ferrites, J. Am. Ceram. Soc. 73 (1990) 729–732. [15] B.J. Christensen, R.T. Coverdale, R.A. Olson, S.J. Ford, E.J. Garboczi, H.M. Jennings, T.O. Mason, Impedance spectroscopy of hydrating cementbased materials: measurement interpretation, and application, J. Am. Ceram. Soc. 77 (1994) 2789–2804. [16] R. Ruffo, S.S. Hong, C.K. Chan, R.A. Huggins, Y. Cui, Impedance analysis of silicon nanowire lithium ion battery anodes, J. Phys. Chem. C 113 (2009) 11390–11398. [17] D. Prasai, J.C. Tuberquia, R.R. Harl, G.K. Jennings, K.I. Bolotin, Graphene: corrosion-inhibiting coating, ACS Nano. 6 (2012) 1102–1108. [18] R. Baker, J. Guindet, M. Kleitz, Classification criteria for solid oxide fuel cell electrode materials, J. Electrochem. Soc. 144 (1997) 2427–2432. [19] B.-K. Lee, J.-Y. Lee, H.-Y. Jung, J.-H. Lee, J.-H. Hwang, Monitoring of the LSM/ YSZ interface in SOFCs using limited-contact geometry in impedance spectroscopy, Solid State Ionics 21-26 (2008) 955–959. [20] G. Heish, T.O. Mason, L.R. Pederson, Experimental limitations in impedance spectroscopy: Part II — electrode artifacts in three-point measurements on Pt/YSZ, Solid State Ionics 91 (1996) 203–212. [21] G. Heish, S.J. Ford, T.O. Mason, L.R. Pederson, Experimental limitations in impedance spectroscopy: Part I — simulation of reference electrode artifacts in three-point measurements, Solid State Ionics 91 (1996) 191–201. [22] G. Hsieh, T.O. Mason, E.J. Garboczi, L.R. Pederson, Experimental limitations in impedance spectroscopy: Part III. Effect of reference electrode geometry/ position, Solid State Ionics 96 (1997) 153–172. [23] M.J.L. Ostergard, M. Morgensen, AC impedance study of the oxygen reduction mechanism on La1  xSrxO3 jn solid oxide fuel cells, Electrochim. Acta 38 (1993) 2015–2020. [24] J. Fleig, Solid oxide fuel cell cathodes: Polarization mechanisms and modeling of the electrochemical performance, Ann. Rev. Mater. Res. 33 (2003) 361–382. [25] B.-K. Lee, Y.-H. Yu, B.-S. So, S.-M. Kim, J. Kim, H.-W. Lee, J.-H. Lee, J.-H. Hwang, Electrical characterization of the platinum/YSZ interfaces in SOFCs via microcontact impedance spectroscopy, J. Electroceram. 17 (2006) 4735–4739. [26] J. Newman, Resistance for flow of current to a disk, J. Electrochem. Soc. 113 (1966) 501–502. [27] E. Barsoukov, J.R. MacDonald, Impedance Spectroscopy: Theory, Experiment, and Applications, John Wiley & Sons, Inc., New York, 2005, pp. 27–128. [28] F.H. van Heuveln, H.J.M. Bouwmeester, Electrode properties of Sr-doped LaMnO3 on yttria-stabilized zirconia, J. Electrochem. Soc. 144 (1997) 134–140. [29] F.H. van Heuveln, F.P.F. van Berkel, J.P.P. Hujismans, Characterization of solid oxide fuel cell electrodes by impedance spectroscopy and I-V characteristics, Solid State Ionics 72 (1994) 240–247. [30] J.-C. Yun, J.-H. Lee, J. Kim, H.-W. Lee, B.-H. Kim, Effect of cathode porosity on the cathodic polarization behavior of mixed conducting La0.6Sr0.4Co0.2Fe0.8O3, J. Kor. Ceram. Soc. 42 (2005) 251–259. [31] H. Zhao, L. Huo, S. Gao, Electrochemical properties of LSM–CBO composite cathode, J. Power Sources 125 (2004) 149-154. [32] J.M. Ralph, C. Rossignol, R. Kumar, Cathode materials for reducedtemperature SOFCs, J. Electrochem. Soc. 150 (2003) A1518–A1522. [33] H.Y. Jung, W.-S. Kim, S.-H. Choi, H.-C. Kim, J. Kim, H.-W. Lee, J.-H. Lee, Effect of cathode current-collecting layer on unit-cell performance of anodesupported solid oxide fuel cells, J. Power Sources 155 (2) (2006) 145–151.