- Email: [email protected]
Praseodymium doped ceria: Model mixed ionic electronic conductor with coupled electrical, optical, mechanical and chemical properties
Solid State Ionics 225 (2012) 194–197
Contents lists available at SciVerse ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Praseodymium doped ceria: Model mixed ionic electronic conductor with coupled electrical, optical, mechanical and chemical properties H.L. Tuller a,⁎, S.R. Bishop a, 1, D. Chen a, Y. Kuru a, b, 2, J.-J. Kim a, T.S. Stefanik a, 3 a b
Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
a r t i c l e
i n f o
Article history: Received 11 September 2011 Received in revised form 29 January 2012 Accepted 9 February 2012 Available online 27 March 2012 Keywords: Mixed conductor SOFC Cathode Chemical expansion Optical absorption Nonstoichiometry
a b s t r a c t Praseodymium doped ceria, PrxCe1 − xO2 − δ (PCO), exhibits many interesting and unusual properties including a pO2 dependent ionic conductivity, an anomalously large thermal expansion coefficient in air, and a significant electronic conductivity component at elevated pO2. These unusual features are discussed in terms of the variable valent nature of Pr at elevated pO2, its position within the ceria band gap and the creation of an impurity band supporting small polaron transport. Implications for use of PCO as a ceria solid electrolyte compatible cathode, stress induced chemical expansion and optically detected redox kinetics are analyzed and discussed. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The fluorite structured oxide, cerium dioxide (CeO2) or ceria, is of continuing interest in the solid oxide fuel cell (SOFC) field given its exceptionally high ionic conductivity when doped with lower valent cations such as Gd or Sm [1,2] and as the oxygen storage material in automotive three way catalysts, given its extensive oxygen nonstoichiometry [3]. This latter feature limits the use of acceptor doped ceria as a solid electrolyte at elevated temperatures, since the loss of oxygen at the anode side of the fuel cell is accompanied by the introduction of electronic conductivity as Ce 4+ is reduced to Ce 3+. A less well known feature is the interesting coupling between oxygen loss and lattice dilation. If not properly accounted for, it can lead to stress and then to mechanical failure [4]. In this article, a different variant of ceria is introduced, praseodymium doped ceria, PrxCe1 − xO2 − δ (PCO), which exhibits many interesting and unusual properties including (1) an oxygen partial pressure (pO2) dependent ionic conductivity, (2) an anomalously large thermal expansion coefficient in air, and (3) a significant electronic conductivity component not found in either the O 2p derived valence band or the Ce 4f derived
⁎ Corresponding author. Tel.: + 1 617 253 6890. E-mail address: [email protected] (H.L. Tuller). 1 Current address: International Institute for Carbon Neutral Energy Research (WPI-I2CNER), Kyushu University, Nishi-ku Fukuoka 819-0395, Japan. 2 Present address: Department of Materials Science and Engineering, Akdeniz University, Dumlupinar Bulvari Kampus, Antalya, 07058, Turkey. 3 Current address: Nanocerox, Ann Arbor, MI 48108, USA. 0167-2738/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2012.02.029
conduction band. These unusual features come about as a result of the variable valent nature of Pr at elevated pO2 (e.g. air). This makes it particularly convenient to study the above phenomena at readily accessible pO2 and temperatures and, furthermore, makes this ceria variant of interest as a ceria solid electrolyte compatible cathode material. 4 2. Atmosphere dependent ionic conductivity The ionic conductivity in most conventional solid oxide electrolytes such as yttria stabilized zirconia (YSZ), gadolinia doped ceria (GDC) or double acceptor doped lanthanum gallate (LSGM) results from the positively charged oxygen vacancies formed in response to the addition of the negatively charged acceptor impurities. The concentration of vacancies is fixed, as is the charge and concentration of the lower valent ions. For example, in GDC h
i •• ′ Gd Ce ¼ 2 VO
ð1Þ
The oxygen vacancy concentration formed in this manner is far in excess of those formed due to the reduction reaction:
••
′
OO ↔VO þ 2e þ 1=2O2 ðg Þ
ð2Þ
4 This paper is dedicated to Professor John Kilner of Imperial College, London, on the occasion of his 65th birthday and is most appropriate given his long standing interest in the ceria materials system and his contributions to its technological development.
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197
Therefore, the oxygen vacancy concentration is virtually independent of pO2, except for the most extremely reducing conditions at elevated temperatures. The same is true for PCO at intermediate pO2 at total which the total Pr concentration [PrCe ] is reduced and trivalent, and thus acts like a standard acceptor dopant, like fixed valent Gd or Sm. Like Eq. (1), this fixes the oxygen vacancy concentration as described by: h
i •• ′ Pr Ce ≈2 VO
ð3Þ
thereby leading to predominantly ionic conductivity independent of pO2. However, as pO2 is increased, the Pr begins to oxidize to the 4 + oxidation state, which is isovalent with the Ce host ion, thereby leading to a corresponding decrease in oxygen vacancies. At high pO2, the electrical conductivity and the thermogravimetric behavior, as demonstrated below, are found to be consistent with the following reaction.
′
••
2PrCe þ OO ↔2Pr Ce þ VO þ 1=2O2 ðg Þ
ð4Þ
with a corresponding mass action law given by h
i2 Pr′ Ce V••O 1=2 −ΔH : 2 P O2 ¼ K ∘ exp kT PrCe OO
ð5Þ
For sufficiently high pO2, most of the Pr takes on the + 4 state and so x h total i PrCe ≅ PrCe
ð6Þ
which gives the following expression for the oxygen vacancy concentration: 1 •• h total i =3 =3 =3 −ΔH −1=6 VO ¼ PrCe P O2 OO K o exp 3kT 2
1
ð7Þ
This predicts that the oxygen vacancy concentration, and therefore the ionic conductivity, should exhibit a power law dependence on pO2 with a −1/6 slope in high pO2. This transition from pO2 independence at intermediate pO2 to a decreasing oxygen vacancy concentration with increasing pO2 is illustrated in Fig. 1, which shows the dependence of the oxygen nonstoichiometry (δ) as a function of pO2 at 650 °C.
Fig. 1. Non-stoichiometry for three compositions of PrxCe1 − xO2 − δ. Experimentally derived values (symbols) for x = 0.2 are from [16] and x = 0.1 are from [6]. Solid lines are from a defect equilibrium model developed in [6].
195
3. Mixed ionic and electronic conductivity under oxidizing conditions It is interesting to note that the reduction reaction described by Eq. (4) contains no explicit electrons, in contrast to the reduction reaction described by Eq. (2). This follows from the fact that while the electrons released by the reduction reaction in Eq. (2) end up in the conduction band of ceria (increasing the concentration of Ce 3+), the electrons released by the reduction reaction in Eq. (4) instead end up localized and trapped on the Pr ion (increasing the concentration of Pr 3+), whose energy level lies within the ceria band gap, as illustrated in Fig. 2. At low concentrations of Pr, these levels are nearly discrete and thus no measurable conduction along these levels is expected or observed. As the Pr concentration increases, wave function overlap between adjacent Pr ions leads to broadening of the discrete levels into impurity bands with potential to support carrier hopping from one atom to the next. This small polaron hopping, however, also requires that the Pr be mixed valent, given that electrons on Pr3+ sites can only hop to an adjacent Pr site if it is empty, e.g. in the Pr4+ state. Thus the small polaron contribution to the electronic conductivity is proportional to the product x ′ ][PrCe of [PrCe ] leading, ideally, to a maximum as a function of pO2 at the point where the concentrations of the two valence states are equal. Indeed, at high pO2 the conductivity data for PrxCe1 − xO2 − δ for x =0.2 in Fig. 3 illustrates this feature, while PCO with low concentrations of Pr do not. In intermediate pO2, PCO with x =0.2 has a lower pO2 independent ionic conductivity than PCO with x =0.1, consistent with other observations showing that defect association at high dopant acceptor levels leads to corresponding reductions in ionic conductivity [5]. With models available to predict the magnitudes of both ionic and electronic conductivities as functions of temperature, pO2, and x in PCO [6,7], it becomes possible to map out the ionic transference number, tion, as a function of these parameters, as illustrated in Fig. 4. Transference numbers close to 0.5 are optimum for supporting rapid redox kinetics given that the oxygen chemical diffusivity is maximum at this point [8]. Since Pr serves to introduce electronic conductivity at higher pO2, this leads to ionic transference numbers less than unity, a desirable feature for cathode materials as discussed below. Interestingly, at very low pO2, tion is observed to reach a maximum at x = 0.1. This comes about because the oxygen vacancy concentration is higher than for x = 0.01, while the enthalpy of ion migration at x = 0.1 is lower than for x = 0.2. 4. PCO as a cathode material The MIEC behavior of PCO at high pO2 presents an opportunity for its implementation in SOFC cathodes, where mixed conductivity is known to enhance oxygen surface exchange kinetics [9]. Impedance measurements performed on thin film PCO electrodes deposited onto yttriastabilized zirconia with varying geometries, and reported elsewhere [10], demonstrate that the electrode performance, dominated by oxygen surface exchange kinetics, compares favorably with that of several high performance cathode materials (e.g. LSF) as illustrated in Fig. 5. This demonstrates the potential of using PCO as a compatible cathode material to ceria based electrolytes, given its identical fluorite structure, good lattice match, and similar chemistry.
Fig. 2. Schematic of Pr impurity band located within the gap between the conduction (C.B.) and valence (V.B.) bands of ceria defined in [6].
196
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197
Fig. 3. Electrical conductivity for three compositions of PrxCe1 − xO2 − δ. Symbols are experimentally measured values from [7] and solid lines are from a defect transport model developed in [6].
Fig. 5. Area specific resistance (Rsurf) of three PrxCe1 − xO2 − δ films with x = 0.01, 0.10, and 0.20 from [10]. Rsurf of the more highly doped PCO films are seen to have values comparable to those of LSF [17].
5. Monitor redox kinetics by optical means As discussed in a submitted accompanying paper in this volume [11], the redox state of PCO can be monitored optically. This comes about since the Pr 4+ ion absorbs in the visible while the Pr 3+ ion does not. Consequently, oxidized PCO appears red while reduced PCO appears transparent. This can, among other things, allow one to monitor redox kinetics in situ as illustrated in Fig. 6. In this series of experiments, the addition of a thin Au film on the surface of PCO is observed to increase the oxidation kinetics. These and related studies are reported elsewhere [12]. 6. Thermo-chemical expansion As mentioned in the introduction, ceria is found to dilate upon the introduction of oxygen vacancies [13]. Fig. 7 shows the temperature induced expansion of PCO in air as measured by high temperature x-ray diffraction and bulk dilatometry. Below 500 °C, all samples show nearly identical thermal expansion behavior. With increasing temperature, however, PCO with large Pr concentrations, exhibits a strong deviation from the nominally linear thermal expansion, with the effective coefficient of thermal expansion becoming greater than two times that found at lower temperatures. The large increase in expansion rate can be correlated with the onset of reduction of the Pr and the corresponding generation of oxygen vacancies [14]. A coefficient of chemical expansion (αC), analogous to the thermal expansion coefficient, can be derived from a known change in the concentration of oxygen vacancies by εC ¼ α C Δδ
Fig. 4. Ionic transference number of PCO calculated at 650 °C.
ð8Þ
where εC is the chemical expansion induced by a change in oxygen stoichiometry (Δδ). The value of αC = 0.084 derived from these studies for x = 0.1 is in good agreement with values obtained for both Gd and undoped ceria under much more highly reducing conditions [13,15]. 7. Summary The PrxCe1 − xO2 − δ system has been shown to exhibit a number of interesting features related to the ability of the Pr ion to reduce under relatively oxidizing conditions. This included the observation of a pO2-dependent ionic conductivity, mixed ionic electronic conductivity due to electron hopping within the Pr impurity band, enhanced oxygen exchange, absorption in the visible tied to electron excitation from the valence to the Pr band, and a dramatically increased effective thermal expansion coefficient due to vacancy induced dilation upon the reduction of PCO. These features demonstrate that PCO serves as a model system exhibiting interesting chemo-mechanical and related electrochemical and optical properties with potential applications in solid oxide fuel cells, permeation membranes, and oxygen storage materials. Acknowledgments This research is being funded, in part, by the Basic Energy Sciences, Department of Energy under award DE SC0002633 and the Division of Materials Research, National Science Foundation under the Material World Network (DMR-0908627) in collaboration with Prof. Moos, Universität Bayreuth. J. J. Kim thanks The Kwanjeong Educational Foundation and both he and Y. Kuru thank the MIT Energy Initiative for partial fellowship support. SRB recognizes partial support from
Fig. 6. Time dependent, 600 °C in-situ optical absorption measured at a wavelength of 532 nm for a film of PCO deposited onto a sapphire substrate exposed to oxygen after being equilibrated in 0.1% CO/CO2. Details of optical absorption in PCO are discussed in [11].
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197 [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
Fig. 7. Expansion as a function of temperature in air, showing large high temperature increases of expansion from oxygen vacancy formation with resulting chemical expansion. High temperature x-ray diffraction (HTXRD) data are from [18] and x = 0.2 and x = 0.1 dilatometry are from [15] and [14], respectively. The thermo-chemical expansion model is derived in [14].
[13] [14] [15] [16] [17] [18]
2
WPI-I CNER, supported by the World Premier International Research Initiative (WPI), MEXT, Japan. References [1] H. Inaba, H. Tagawa, Solid State Ionics 83 (1996) 1–16. [2] B.C.H. Steele, Solid State Ionics 129 (2000) 95–110.
197
R.J. Gorte, AICHE J. 56 (2010) 1126–1135. A. Atkinson, T.M.G.M. Ramos, Solid State Ionics 129 (2000) 259–269. J.A. Kilner, Chem. Lett. 37 (2008) 1012–1015. S.R. Bishop, T. Stefanik, H.L. Tuller, Phys. Chem. Chem. Phys. 13 (2011) 10165–10173. T.S. Stefanik, Electrical Properties and Defect Structure of Praseodymium–Cerium Oxide Solid Solutions, MIT Ph.D. Thesis (2004). J. Maier, Physical Chemistry of Ionic Materials, 1st ed. John Wiley and Sons, Inc., Hoboken, NJ, 2004. S.B. Adler, Chem. Rev. 104 (2004) 4791–4843. D. Chen, S.R. Bishop, H.L. Tuller, J. Electroceram. 28 (2012) 62–69. J.-J. Kim, S.R. Bishop, N. Thompson, Y. Kuru, H.L. Tuller, Solid State Ionics (in press). S.R. Bishop, J.-J. Kim, N. Thompson, D. Chen, Y. Kuru, T. Stefanik, ECS Trans. 35 (2011). S.R. Bishop, K.L. Duncan, E.D. Wachsman, Electrochim. Acta 54 (2009) 1436–1443. S.R. Bishop, H.L. Tuller, Y. Kuru, B. Yildiz, J. Eur. Ceram. Soc. 31 (2011) 2351–2356. C. Chatzichristodoulou, P.V. Hendriksen, A. Hagen, J. Electrochem. Soc. 157 (2010) B299–B307. C. Chatzichristodoulou, P.V. Hendriksen, J. Electrochem. Soc. 157 (2010) B481–B489. F.S. Baumann, J. Fleig, G. Cristiani, B. Stuhlhofer, H. Habermeier, J. Maier, J. Electrochem. Soc. 154 (2007) B931–B941. Y. Kuru, S.R. Bishop, J.-J. Kim, B. Yildiz, H.L. Tuller, Solid State Ionics. 193 (2011) 1–4.
Contents lists available at SciVerse ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Praseodymium doped ceria: Model mixed ionic electronic conductor with coupled electrical, optical, mechanical and chemical properties H.L. Tuller a,⁎, S.R. Bishop a, 1, D. Chen a, Y. Kuru a, b, 2, J.-J. Kim a, T.S. Stefanik a, 3 a b
Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
a r t i c l e
i n f o
Article history: Received 11 September 2011 Received in revised form 29 January 2012 Accepted 9 February 2012 Available online 27 March 2012 Keywords: Mixed conductor SOFC Cathode Chemical expansion Optical absorption Nonstoichiometry
a b s t r a c t Praseodymium doped ceria, PrxCe1 − xO2 − δ (PCO), exhibits many interesting and unusual properties including a pO2 dependent ionic conductivity, an anomalously large thermal expansion coefficient in air, and a significant electronic conductivity component at elevated pO2. These unusual features are discussed in terms of the variable valent nature of Pr at elevated pO2, its position within the ceria band gap and the creation of an impurity band supporting small polaron transport. Implications for use of PCO as a ceria solid electrolyte compatible cathode, stress induced chemical expansion and optically detected redox kinetics are analyzed and discussed. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The fluorite structured oxide, cerium dioxide (CeO2) or ceria, is of continuing interest in the solid oxide fuel cell (SOFC) field given its exceptionally high ionic conductivity when doped with lower valent cations such as Gd or Sm [1,2] and as the oxygen storage material in automotive three way catalysts, given its extensive oxygen nonstoichiometry [3]. This latter feature limits the use of acceptor doped ceria as a solid electrolyte at elevated temperatures, since the loss of oxygen at the anode side of the fuel cell is accompanied by the introduction of electronic conductivity as Ce 4+ is reduced to Ce 3+. A less well known feature is the interesting coupling between oxygen loss and lattice dilation. If not properly accounted for, it can lead to stress and then to mechanical failure [4]. In this article, a different variant of ceria is introduced, praseodymium doped ceria, PrxCe1 − xO2 − δ (PCO), which exhibits many interesting and unusual properties including (1) an oxygen partial pressure (pO2) dependent ionic conductivity, (2) an anomalously large thermal expansion coefficient in air, and (3) a significant electronic conductivity component not found in either the O 2p derived valence band or the Ce 4f derived
⁎ Corresponding author. Tel.: + 1 617 253 6890. E-mail address: [email protected] (H.L. Tuller). 1 Current address: International Institute for Carbon Neutral Energy Research (WPI-I2CNER), Kyushu University, Nishi-ku Fukuoka 819-0395, Japan. 2 Present address: Department of Materials Science and Engineering, Akdeniz University, Dumlupinar Bulvari Kampus, Antalya, 07058, Turkey. 3 Current address: Nanocerox, Ann Arbor, MI 48108, USA. 0167-2738/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2012.02.029
conduction band. These unusual features come about as a result of the variable valent nature of Pr at elevated pO2 (e.g. air). This makes it particularly convenient to study the above phenomena at readily accessible pO2 and temperatures and, furthermore, makes this ceria variant of interest as a ceria solid electrolyte compatible cathode material. 4 2. Atmosphere dependent ionic conductivity The ionic conductivity in most conventional solid oxide electrolytes such as yttria stabilized zirconia (YSZ), gadolinia doped ceria (GDC) or double acceptor doped lanthanum gallate (LSGM) results from the positively charged oxygen vacancies formed in response to the addition of the negatively charged acceptor impurities. The concentration of vacancies is fixed, as is the charge and concentration of the lower valent ions. For example, in GDC h
i •• ′ Gd Ce ¼ 2 VO
ð1Þ
The oxygen vacancy concentration formed in this manner is far in excess of those formed due to the reduction reaction:
••
′
OO ↔VO þ 2e þ 1=2O2 ðg Þ
ð2Þ
4 This paper is dedicated to Professor John Kilner of Imperial College, London, on the occasion of his 65th birthday and is most appropriate given his long standing interest in the ceria materials system and his contributions to its technological development.
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197
Therefore, the oxygen vacancy concentration is virtually independent of pO2, except for the most extremely reducing conditions at elevated temperatures. The same is true for PCO at intermediate pO2 at total which the total Pr concentration [PrCe ] is reduced and trivalent, and thus acts like a standard acceptor dopant, like fixed valent Gd or Sm. Like Eq. (1), this fixes the oxygen vacancy concentration as described by: h
i •• ′ Pr Ce ≈2 VO
ð3Þ
thereby leading to predominantly ionic conductivity independent of pO2. However, as pO2 is increased, the Pr begins to oxidize to the 4 + oxidation state, which is isovalent with the Ce host ion, thereby leading to a corresponding decrease in oxygen vacancies. At high pO2, the electrical conductivity and the thermogravimetric behavior, as demonstrated below, are found to be consistent with the following reaction.
′
••
2PrCe þ OO ↔2Pr Ce þ VO þ 1=2O2 ðg Þ
ð4Þ
with a corresponding mass action law given by h
i2 Pr′ Ce V••O 1=2 −ΔH : 2 P O2 ¼ K ∘ exp kT PrCe OO
ð5Þ
For sufficiently high pO2, most of the Pr takes on the + 4 state and so x h total i PrCe ≅ PrCe
ð6Þ
which gives the following expression for the oxygen vacancy concentration: 1 •• h total i =3 =3 =3 −ΔH −1=6 VO ¼ PrCe P O2 OO K o exp 3kT 2
1
ð7Þ
This predicts that the oxygen vacancy concentration, and therefore the ionic conductivity, should exhibit a power law dependence on pO2 with a −1/6 slope in high pO2. This transition from pO2 independence at intermediate pO2 to a decreasing oxygen vacancy concentration with increasing pO2 is illustrated in Fig. 1, which shows the dependence of the oxygen nonstoichiometry (δ) as a function of pO2 at 650 °C.
Fig. 1. Non-stoichiometry for three compositions of PrxCe1 − xO2 − δ. Experimentally derived values (symbols) for x = 0.2 are from [16] and x = 0.1 are from [6]. Solid lines are from a defect equilibrium model developed in [6].
195
3. Mixed ionic and electronic conductivity under oxidizing conditions It is interesting to note that the reduction reaction described by Eq. (4) contains no explicit electrons, in contrast to the reduction reaction described by Eq. (2). This follows from the fact that while the electrons released by the reduction reaction in Eq. (2) end up in the conduction band of ceria (increasing the concentration of Ce 3+), the electrons released by the reduction reaction in Eq. (4) instead end up localized and trapped on the Pr ion (increasing the concentration of Pr 3+), whose energy level lies within the ceria band gap, as illustrated in Fig. 2. At low concentrations of Pr, these levels are nearly discrete and thus no measurable conduction along these levels is expected or observed. As the Pr concentration increases, wave function overlap between adjacent Pr ions leads to broadening of the discrete levels into impurity bands with potential to support carrier hopping from one atom to the next. This small polaron hopping, however, also requires that the Pr be mixed valent, given that electrons on Pr3+ sites can only hop to an adjacent Pr site if it is empty, e.g. in the Pr4+ state. Thus the small polaron contribution to the electronic conductivity is proportional to the product x ′ ][PrCe of [PrCe ] leading, ideally, to a maximum as a function of pO2 at the point where the concentrations of the two valence states are equal. Indeed, at high pO2 the conductivity data for PrxCe1 − xO2 − δ for x =0.2 in Fig. 3 illustrates this feature, while PCO with low concentrations of Pr do not. In intermediate pO2, PCO with x =0.2 has a lower pO2 independent ionic conductivity than PCO with x =0.1, consistent with other observations showing that defect association at high dopant acceptor levels leads to corresponding reductions in ionic conductivity [5]. With models available to predict the magnitudes of both ionic and electronic conductivities as functions of temperature, pO2, and x in PCO [6,7], it becomes possible to map out the ionic transference number, tion, as a function of these parameters, as illustrated in Fig. 4. Transference numbers close to 0.5 are optimum for supporting rapid redox kinetics given that the oxygen chemical diffusivity is maximum at this point [8]. Since Pr serves to introduce electronic conductivity at higher pO2, this leads to ionic transference numbers less than unity, a desirable feature for cathode materials as discussed below. Interestingly, at very low pO2, tion is observed to reach a maximum at x = 0.1. This comes about because the oxygen vacancy concentration is higher than for x = 0.01, while the enthalpy of ion migration at x = 0.1 is lower than for x = 0.2. 4. PCO as a cathode material The MIEC behavior of PCO at high pO2 presents an opportunity for its implementation in SOFC cathodes, where mixed conductivity is known to enhance oxygen surface exchange kinetics [9]. Impedance measurements performed on thin film PCO electrodes deposited onto yttriastabilized zirconia with varying geometries, and reported elsewhere [10], demonstrate that the electrode performance, dominated by oxygen surface exchange kinetics, compares favorably with that of several high performance cathode materials (e.g. LSF) as illustrated in Fig. 5. This demonstrates the potential of using PCO as a compatible cathode material to ceria based electrolytes, given its identical fluorite structure, good lattice match, and similar chemistry.
Fig. 2. Schematic of Pr impurity band located within the gap between the conduction (C.B.) and valence (V.B.) bands of ceria defined in [6].
196
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197
Fig. 3. Electrical conductivity for three compositions of PrxCe1 − xO2 − δ. Symbols are experimentally measured values from [7] and solid lines are from a defect transport model developed in [6].
Fig. 5. Area specific resistance (Rsurf) of three PrxCe1 − xO2 − δ films with x = 0.01, 0.10, and 0.20 from [10]. Rsurf of the more highly doped PCO films are seen to have values comparable to those of LSF [17].
5. Monitor redox kinetics by optical means As discussed in a submitted accompanying paper in this volume [11], the redox state of PCO can be monitored optically. This comes about since the Pr 4+ ion absorbs in the visible while the Pr 3+ ion does not. Consequently, oxidized PCO appears red while reduced PCO appears transparent. This can, among other things, allow one to monitor redox kinetics in situ as illustrated in Fig. 6. In this series of experiments, the addition of a thin Au film on the surface of PCO is observed to increase the oxidation kinetics. These and related studies are reported elsewhere [12]. 6. Thermo-chemical expansion As mentioned in the introduction, ceria is found to dilate upon the introduction of oxygen vacancies [13]. Fig. 7 shows the temperature induced expansion of PCO in air as measured by high temperature x-ray diffraction and bulk dilatometry. Below 500 °C, all samples show nearly identical thermal expansion behavior. With increasing temperature, however, PCO with large Pr concentrations, exhibits a strong deviation from the nominally linear thermal expansion, with the effective coefficient of thermal expansion becoming greater than two times that found at lower temperatures. The large increase in expansion rate can be correlated with the onset of reduction of the Pr and the corresponding generation of oxygen vacancies [14]. A coefficient of chemical expansion (αC), analogous to the thermal expansion coefficient, can be derived from a known change in the concentration of oxygen vacancies by εC ¼ α C Δδ
Fig. 4. Ionic transference number of PCO calculated at 650 °C.
ð8Þ
where εC is the chemical expansion induced by a change in oxygen stoichiometry (Δδ). The value of αC = 0.084 derived from these studies for x = 0.1 is in good agreement with values obtained for both Gd and undoped ceria under much more highly reducing conditions [13,15]. 7. Summary The PrxCe1 − xO2 − δ system has been shown to exhibit a number of interesting features related to the ability of the Pr ion to reduce under relatively oxidizing conditions. This included the observation of a pO2-dependent ionic conductivity, mixed ionic electronic conductivity due to electron hopping within the Pr impurity band, enhanced oxygen exchange, absorption in the visible tied to electron excitation from the valence to the Pr band, and a dramatically increased effective thermal expansion coefficient due to vacancy induced dilation upon the reduction of PCO. These features demonstrate that PCO serves as a model system exhibiting interesting chemo-mechanical and related electrochemical and optical properties with potential applications in solid oxide fuel cells, permeation membranes, and oxygen storage materials. Acknowledgments This research is being funded, in part, by the Basic Energy Sciences, Department of Energy under award DE SC0002633 and the Division of Materials Research, National Science Foundation under the Material World Network (DMR-0908627) in collaboration with Prof. Moos, Universität Bayreuth. J. J. Kim thanks The Kwanjeong Educational Foundation and both he and Y. Kuru thank the MIT Energy Initiative for partial fellowship support. SRB recognizes partial support from
Fig. 6. Time dependent, 600 °C in-situ optical absorption measured at a wavelength of 532 nm for a film of PCO deposited onto a sapphire substrate exposed to oxygen after being equilibrated in 0.1% CO/CO2. Details of optical absorption in PCO are discussed in [11].
H.L. Tuller et al. / Solid State Ionics 225 (2012) 194–197 [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
Fig. 7. Expansion as a function of temperature in air, showing large high temperature increases of expansion from oxygen vacancy formation with resulting chemical expansion. High temperature x-ray diffraction (HTXRD) data are from [18] and x = 0.2 and x = 0.1 dilatometry are from [15] and [14], respectively. The thermo-chemical expansion model is derived in [14].
[13] [14] [15] [16] [17] [18]
2
WPI-I CNER, supported by the World Premier International Research Initiative (WPI), MEXT, Japan. References [1] H. Inaba, H. Tagawa, Solid State Ionics 83 (1996) 1–16. [2] B.C.H. Steele, Solid State Ionics 129 (2000) 95–110.
197
R.J. Gorte, AICHE J. 56 (2010) 1126–1135. A. Atkinson, T.M.G.M. Ramos, Solid State Ionics 129 (2000) 259–269. J.A. Kilner, Chem. Lett. 37 (2008) 1012–1015. S.R. Bishop, T. Stefanik, H.L. Tuller, Phys. Chem. Chem. Phys. 13 (2011) 10165–10173. T.S. Stefanik, Electrical Properties and Defect Structure of Praseodymium–Cerium Oxide Solid Solutions, MIT Ph.D. Thesis (2004). J. Maier, Physical Chemistry of Ionic Materials, 1st ed. John Wiley and Sons, Inc., Hoboken, NJ, 2004. S.B. Adler, Chem. Rev. 104 (2004) 4791–4843. D. Chen, S.R. Bishop, H.L. Tuller, J. Electroceram. 28 (2012) 62–69. J.-J. Kim, S.R. Bishop, N. Thompson, Y. Kuru, H.L. Tuller, Solid State Ionics (in press). S.R. Bishop, J.-J. Kim, N. Thompson, D. Chen, Y. Kuru, T. Stefanik, ECS Trans. 35 (2011). S.R. Bishop, K.L. Duncan, E.D. Wachsman, Electrochim. Acta 54 (2009) 1436–1443. S.R. Bishop, H.L. Tuller, Y. Kuru, B. Yildiz, J. Eur. Ceram. Soc. 31 (2011) 2351–2356. C. Chatzichristodoulou, P.V. Hendriksen, A. Hagen, J. Electrochem. Soc. 157 (2010) B299–B307. C. Chatzichristodoulou, P.V. Hendriksen, J. Electrochem. Soc. 157 (2010) B481–B489. F.S. Baumann, J. Fleig, G. Cristiani, B. Stuhlhofer, H. Habermeier, J. Maier, J. Electrochem. Soc. 154 (2007) B931–B941. Y. Kuru, S.R. Bishop, J.-J. Kim, B. Yildiz, H.L. Tuller, Solid State Ionics. 193 (2011) 1–4.