(Ln = Gd, La, Tb, Pr, Eu, Er, Yb, Nd) and Effect of ... - ACS Publications

Article pubs.acs.org/JPCC

Structural−Transport Properties Relationships on Ce1−xLnxO2−δ System (Ln = Gd, La, Tb, Pr, Eu, Er, Yb, Nd) and Effect of Cobalt Addition María Balaguer, Cecilia Solís, and José M. Serra* Instituto de Tecnología Química (Universidad Politécnica de Valencia − Consejo Superior de Investigaciones Científicas), Av. Naranjos s/n, E-46022 Valencia, Spain S Supporting Information *

ABSTRACT: A large series of doped cerias have been prepared by the coprecipitation method combined with impregnation and completely characterized in order to have an overall understanding of the structural, oxygen vacancy concentration, and transport properties relationships. Several lanthanides were incorporated in the fluorite structure, and the effects of the final sintering temperature (1073 and 1573 K) and the addition of cobalt oxide on the structural properties were studied. The chosen lanthanides (Gd, La, Tb, Pr, Eu, Er, Yb and Nd) included a large range of ionic radii and different metals exhibiting variable oxidation states under the typical operating conditions for these materials. The materials have been characterized by powder XRD, high-temperature XRD, microRaman spectroscopy, helium pycnometry, and dc conductivity. Transport properties were correlated with structural features induced by the different ionic radii and variable oxidation state of the dopants. The highest ionic conductivity was obtained for the less distorted cells (Gd- and Nd-doped ceria) which represent the optimum balance between Coulomb interactions, steric effects, and vacancy distribution. The lowest Ea value was found for materials with long cell parameters.

1. INTRODUCTION Depending on the characteristics introduced by the dopant, doped ceria has been suggested and studied for a number of applications, e.g., electrolyte1,2 and anode3,4 in solid oxide fuel cells (SOFCs), mixed-conductive membranes for oxygen separation5,6 and partial oxidation of hydrocarbons,7,8 oxygen sensors,9 oxygen storage materials, and catalysts10,11 for industrial relevant processes.12,13 Doped cerias are especially attractive for these applications since they combine high oxygen-ion mobility, redox catalytic properties, chemical compatibility with water and carbon dioxide at high temperatures, and sufficient resistance to reduction under relatively low oxygen partial pressures. Pure CeO2 is a fluorite structured oxide in which the cations have 8-fold coordination. At low oxygen partial pressures, Ce4+ cations are easily reduced to the Ce3+ valence state in order to charge compensate for oxygen vacancy formation while the fluorite structure is retained. This material exhibits n-type electronic conductivity in reducing environments due to the formation of small polarons, which migrate by a thermally activated hopping.14 The ceria fluorite phase is very tolerant to such reduction and to the dissolution of other oxides which introduce oxide vacancies and thus extend the electrolytic region (area of predominant ionic conductivity) over that of ceria. Nevertheless, both acceptor doping and reduced cation formation result in a strain termed chemical expansion,14 which produces stresses or constrictions in the structure that affect the © 2012 American Chemical Society

conductivity and possibly result in mechanical failure of the solid state device.15,16 Actually, the highest ionic conductivities are reported for doping cations which produce the lowest size mismatch with respect to Ce4+, i.e., Gd3+ and Sm3+.17,18 Special attention has been paid to doping with lanthanide oxides, which are soluble in ceria up to values around 40%.19,20 Trivalent rare earths promote the ionic conductivity. Sm provides high ionic conductivity together with high resistance to reduction, followed by Gd;18 that is why they are the most commonly used dopants as electrolyte in SOFC.15,17,21 Doping with La increases the lattice constant of the ceria phase favoring the mobility of the ions, but the material becomes quite reducible instead.10,18,22 Dy has been proposed on ceria systems as yttrium codopant23 and Er has been used as dopant in the catalytic ceria−zirconia system.24 The addition of elements with two valence states has been less studied. They increase the ionic conductivity, but some of these materials develop mixed ionic electronic conductivity, i.e., ceria doped with Tb and Pr. The rise in electronic conductivity and the lattice expansion undergone at low partial pressures persuade the use of these materials as SOFC electrolyte. However, at high oxygen partial pressures ionic and p-type electronic conductivity are available in some materials and make them suitable for O 2 Received: December 2, 2011 Revised: March 7, 2012 Published: March 23, 2012 7975

dx.doi.org/10.1021/jp211594d | J. Phys. Chem. C 2012, 116, 7975−7982

The Journal of Physical Chemistry C

Article

membranes25,26 and SOFC cathodes. Specifically, praseodymium has been employed as dopant for gravimetric gas sensors because it increases the level of nonstoichiometry in ceria under oxidizing conditions,27 but it has a doping limit close to 30% due to the high and nonlinear thermal expansion behavior and large chemical expansion upon reduction.28 Doping with rare earth elements does not generally affect the sintering properties, with the exception of Yb, Nd, and La,29 whose pellets or probes remain fragile and crumble on the edges, despite they may give densities greater than 85−90% of theoretical density. Many metal oxides have been proposed to improve the density such as MnO2, Mn2O3, Ga2O3, Co3O4, CoO, and Fe2O3.30,31 They act as sintering aid because of their melting point, which is lower than that of ceria−lanthanide and thus promotes the atomic diffusion throughout the particle surface under high temperatures causing the sintering at lower temperatures. Cobalt oxide has been demonstrated to be an attractive sintering aid since it has high volatility and is able to reduce the sintering temperature more effectively than other metal oxides.32 The addition of lanthanides induces the formation of oxygen vacancies, which are mostly free at low dopant content. At higher dopant levels defect associations due to Coulombic attractions might appear.17 Thus, we have chosen a dopant content of 10%. Trivalent lanthanides as Gd, La, Nd, and Er have been added. Pr, Tb, Eu, and Yb are elements with two possible valence states. Cobalt has been added to the same series of doped ceria in order to check whether it increases the mixed conductivity of these materials as previously reported for Gd-, Pr-, and Tb-doped ceria.33−35 Although there are many reported works on Ln-doped ceria, this work presents a systematic and exhaustive study of both structure and transport properties of the Ln-substituted ceria compounds, including multivalent dopants and the effect of cobalt addition. Previous reported works of a series of Ln are based on theoretical studies or on literature conductivity values36 which usually are quite scattered. Here the materials have been systematically prepared and characterized by powder XRD, micro-Raman spectroscopy, helium pycnometry, and dc conductivity. Specifically, Raman spectroscopy is a widely used technique for structural analysis of ceria and rare-earth-doped ceria due to its sensitivity to any structural changes (cell parameter and particle size) as well as to the oxygen vacancy concentration.37,38

To identify the crystalline phase and determine the crystallite size and the lattice parameters of the samples, the powders were characterized by X-ray diffraction (XRD). The measurements were carried out by a PANalytical Cubix fast diffractometer, using Cu Kα1 radiation (λ = 1.5406 Å) and an X’Celerator detector in Bragg−Brentano geometry. XRD patterns were recorded in the 2θ range from 0° to 90° and analyzed using X’Pert Highscore Plus software. Lattice parameter (a) was extracted by fitting (422) reflection with Gaussian function.39 The lattice parameters for the (hkl) plane were calculated using the equation for a cubic system. The studies of the thermal behavior of the material were performed by high-temperature XRD (HT-XRD) in an Anton Paar XRK-900 reaction chamber attached to the diffractometer, either in a dry air atmosphere or in helium up to 1173 K. Density features of the materials were calculated from mass and volume values, obtained by a helium pycnometer, and comparing them with the theoretical values given by eq 1, valid for a cubic fluorite structure, which has 4 lanthanide and 8 oxygen atoms (and corrected stoichiometry for doped samples): ρT =

∑n xn[4AWn + 8AWO] [CV][NA]

(1)

where n represents the number of different elements forming the material, x is the molar fraction of the element in the material, AW is the atomic weight of the n-lanthanide and the oxygen, CV is the volume of the unit cell, and NA is Avogadro’s number. Raman spectra were measured with a Renishaw inVia Raman spectrometer equipped with a Leica DMLM microscope and a 514 nm Ar+ ion laser as an excitation source. A ×50 objective of 8 mm optical length was used to focus the depolarized laser beam on a spot of about 3 μm in diameter. The Raman scattering was collected with a CCD array detector. Thermogravimetric (TG) measurements were performed on a Metler-Toledo StarE balance under dry air flow in the range 300−1273 K following a heating ramp of 5 K/min. Rectangular probes (4 × 0.4 × 0.2 cm3) of the powders fired at 1073 K were uniaxially pressed at 125 MPa during 1 min and subsequently sintered for 5 h at 1573 K in air. Electrical conductivity measurements were conducted by standard fourpoint dc technique on the sintered rectangular bars using silver wire and paste for contacting. The measurements were carried out in a temperature range from 673 to 1073 K by cooling down (1 K/min) in constant different O 2 -containing atmospheres (Linde calibrated gas mixtures checked by an YSZ oxygen sensor). The constant current was supplied by a programmable current source (Keithley 2601), and the voltage drop through the sample was detected by a multimeter (Keithley 3706). The conductivity measurements are thermally activated and were analyzed on the basis of Arrhenius behavior, σ(T) = (A/T) exp(−Ea/kT). The activation energy Ea was extracted from the slope of the graphs. Once the highest temperature (1073 K) was reached, the samples were stabilized for 2 h in order to warrantee the high-temperature reduction state corresponding to the specific pO2.

2. EXPERIMENTAL SECTION Lanthanide-doped ceria compounds have been prepared by coprecipitation method (Figure S1, Supporting Information) in order to synthesize powders of nanometric size. A clear solution of commercial lanthanide nitrates mixture in distilled water was prepared at 323 K. (NH4)2CO3 solution in a 1:1.5 molar ratio was dropped into the cation solution to achieve the total precipitation of the mixed cations. The resulting precursor powders were dried at 373 K after filtration and rinsing with water. The cobalt addition (when required) was done over the dried precursor powder by incipient wetness impregnation. Calculated 2% molar of Co(NO3)2·6H2O was dissolved in distilled water (volume according with the pore volume) and mixed with the powder. Finally, each powder was calcined for 5 h in air atmosphere at 1073 K to decompose the residual nitrates and carbonates and to favor the formation of the fluorite phase.

3. RESULTS AND DISCUSSION Thermogravimetric analysis of precipitate powder (Figure S2, Supporting Information) proves that the mass is stabilized at 873 K, which indicates that all the nitrates and (hydroxy)7976



Article

(+3/+4) would need an apparent ionic radii between those corresponding to +3 and +4 oxidation states in order to preserve the general trend. Thus, it is possible to estimate the amount of the trivalent and tetravalent cations, which gives rise to 40% and 62% of trivalent cations for Pr and Tb, respectively. Co addition does not cause the formation of any impurity phase (Figure S3, Supporting Information) neither produces any remarkable change among the cell parameter range of accuracy and maintain same tendency. Several studies about Co addition in cerias attempted to elucidate the location of Co atoms. Co has been reported to be soluble in CGO less than 0.5 mol % and diffuse into the grain boundary to form a film of roughly 0.5 nm thickness together with Gd or clustering particles.41 For Pr-doped ceria, the addition of Co at 2 mol % results in the apparent dissolution of cobalt while, when 5 mol % is added, traces of Co oxides were observed by XRD, which appear segregated along the grain boundaries as well.34 On the contrary, Co causes shrinkage in the cell parameter of Tbdoped ceria and is soluble up to 1.3% as previously reported.35 Regardless of the location, Co enhances the sinterability28 and growth of the grain size as inferred from the sharpening of XRD peaks (Figure S3, Supporting Information) and SEM images shown in Figure 2 (0.3 ± 0.1 μm and 1.3 ± 0.3 μm on average for samples without and with Co, respectively).

carbonates disappear completely at this temperature. A slight mass loss is observed at high temperatures, which is ascribed to oxygen release from the ceria lattice. The same behavior is observed for other samples, as expected from the same chemistry of rare earth salts. Therefore, calcination at 1073 K ensures that there will not remain any phase related to the precursor salts. Samples were also sintered at 1573 K afterward to carry out electrochemical experiments where dense specimens are required. Experimental density values summarized in Table 1 are very close to the expected theoretically by eq 1 for Table 1. Theoretical and Experimental Density of Doped Cerias materials

ρtheor (±0.001), g/cm3

ρexp (±0.25), g/cm3

%

CeO2−δ Ce0.9Pr0.1O2−δ Ce0.9Yb0.1O2−δ Ce0.9Eu0.1O2−‑δ Ce0.9Tb0.1O2−δ Ce0.9Gd0.1O2−‑δ Ce0.9La0.1O2−δ Ce0.9Er0.1O2−δ

7.209 7.197 7.345 7.190 7.304 7.245 7.061 7.314

7.11 6.63 7.29 6.76 7.25 7.046 7.05 7.12

98.6 92.1 99.3 94.0 99.3 97.3 99.8 97.4

most of the doped compositions. Pr-doped ceria shows the highest deviation with a 92% of the expected value, as indicated in the table, which is ascribed principally to closed porosity. However, we should take into account the statistical experimental error obtained in these measurements. 3.1. Structural Characterization. XRD Analysis. XRD patterns of Ce0.9Ln0.1O2−δ and Ce0.9Ln0.1O2−δ + Co (Ln = La, Gd, Eu, Pr, Tb, Er, Yb, Nd) show the expected cubic fluorite structure (Figure S3, Supporting Information). Diffraction peaks corresponding to any other lanthanide oxide or any precursor are not observed, which means that the dopants are fully incorporated into the lattice of ceria. Figure 1 depicts the cell parameter as a function of the trivalent (8-fold coordination) ionic radii given by Shannon40 for two different annealing temperatures and shows a linear correlation except for Pr and Tb. Despite the two possible oxidation states (+2/+3) of Eu and Yb, they remain trivalent in the fluorite structure. On the other hand, multivalent Pr and Tb

Figure 2. SEM images of Ce0.9Ln0.1O2−δ and Ce0.9Ln0.1O2−δ + Co calcined at 1573 K (Ln = Tb, Pr, Eu, and La).

Under real operating conditions these materials present different redox processes dependent on temperature and pO2. Therefore, the evolution of cell parameters associated with vacancy concentration is studied by means of high-temperature XRD (HT-XRD) measurements by heating the specimens in air up to 1173 K and cooling them down in helium. The fluorite structure is proved to be maintained in the whole temperature range, and cell parameters are plotted in Figure 3 for some selected compositions. Ln = La, Gd, and Eu show a linear dependency of cell parameter with temperature, which corresponds to thermal expansion coefficients (TEC) in air and in He, respectively, summarized in Table 2. On the other hand, Tb- and Pr-doped cerias show an increase in the slope at ∼723 and 823 K, respectively, when they are heated up in air. Accordingly, an additional chemical expansion coefficient (CEC) is calculated, which is assigned to the reduction of Ln4+ to Ln3+ and the associated release of molecular oxygen.

Figure 1. Lattice parameter of doped ceria as a function of the trivalent ionic radius of the dopant cation (and tetravalent for Pr and Tb) in 8fold coordination. 7977



Article

Figure 3. HT-XRD heating up in air (left side) and cooling down in He (right side) of Ce0.9Ln0.1O2−δ. In the insets, the calculated cell parameters (a973 K) at 973 K in air (left side) and in He (right side) are plotted as a function of the trivalent ionic radii (and tetravalent for Pr and Tb).

Raman Spectroscopy Analysis. Cerium oxide has a cubic fluorite structure and belongs to the Oh5 (Fm3̅m) space group. The F2g Raman-active mode frequency (at 465 cm−1) can be understood as a symmetric breathing mode of the O atoms around each cation (Ce−O8 unit). Therefore, this unit should be very sensitive to any disorder in the oxygen sublattice due to induced nonstoichiometry.42 Normalized Raman spectra are plotted in Figure 4a. There are also reported additional peaks close to the F2g mode, mainly a shoulder at ∼570 cm−1, that have been attributed to the formation of oxygen vacancies (see inset of Figure 4a). The higher relative intensity observed for Pr and Tb of the ∼570 cm−1 shoulder is ascribed to the higher optical adsorption for these colored compounds.37 There is also a broad band at 600 cm−1 as consequence of either an ordering effect in the oxygen vacancies to produce confinement of the phonons within an inhomogeneous distribution of domains or structural distortions of the O sublattice.43 The influence of the Ln doping in the structure can be observed in Figure 4b (top), where the Raman shift of the F2g mode is depicted as a function of the of the experimental cell parameter, as determined by XRD. Raman mode shifts down as the cell parameter increases and only Pr doped CeO2 does not follow the linear relationship. The general trend agrees with expected Raman shift due to cell volume changes. The frequency shift (Δω) produced by a change in the lattice parameter (Δa) can be written in terms of the Grüneisen parameter:

Table 2. Thermal Expansion Coefficients Calculated in Air and in He for Different Compositions Ln La Eu Gd Pr Tb

TEC in air (K−1) 11.9 11.9 11.9 11.3 12.1

× × × × ×

10−6 10−6 10−6 10−6 10−6

TEC in He (K−1) 12.2 12.2 11.4 12.3 12.3

× × × × ×

10−6 10−6 10−6 10−6 10−6

CEC in air (K−1)

11.51 × 10−6 5.77 × 10−6

CEC is not observed for Eu-doped ceria, which confirms that there is not reduction of the cation either with the temperature or lower pO2 (He). When Pr- and Tb-doped samples are cooled down in He, the material cannot be reoxidized due to the lack of molecular oxygen necessary for reoxidation, and only thermal expansion can be observed for the material in the reduced state. On the basis of the HT-XRD cell parameters and the ionic radii of the different oxidation states, it is possible to estimate the amount of trivalent cation for Pr and Tb by comparing with those theoretical cell parameters of totally reduced and oxidized samples. The insets of Figure 3 plot cell parameters in air (left) and He (right) at 973 K as a function of the ionic radii (trivalent and tetravalent when required). From room temperature (RT) to 973 K in air the estimated amount of trivalent cation is increased from 40% to 53% for Pr and from 62% to 78% for Tb. After heating the materials up to 1173 K in air only 9.4% of Tb4+ remains and 16% of Pr4+ is still present. By changing the atmosphere to He, both materials are completely reduced and trivalent cations remain fixed during the cooling down in He to RT. The relative amount of Pr and Tb cations in their oxidation states proved by HT-XRD is related to changes in the oxygen vacancy concentration whose importance will be discussed latter in the transport measurements.

Δω = −3γω0Δa /a0

(2)

where ω0 and a0 are the Raman frequency and the lattice parameter for CeO2, respectively, and γ ≡ (B/ω)dω/dP is the Grüneisen constant (1.24 for CeO237). Other contributions, apart from volume, can be analyzed if the Grüneisen shifts are taken into account. Figure 4b (bottom) 7978



Article

the general trend. This shift up in frequency in the F2g mode can be directly assigned to the increasing disorder produced by the oxygen vacancies.37 It can be observed that La-doped material presents the highest shift, while the behavior of the Prdoped compound remains unexplained in this work. Tb-doped material, which presents both Tb3+ and Tb4+ as inferred from HT-XRD analysis, follows the general trend since the effect of the mixed oxidation state is already integrated in the experimental cell parameter. The defect concentration may be estimated from Raman intensity by using the spatial correlation length,44 although only calculations for pure CeO2 and CeO2 + Co were related to oxygen vacancy concentration (Figure S4). Raman demonstrates that the introduction of Ln3+ dopants produces a disorder due to the oxygen vacancies that increases with the cell parameter. This generated disorder is reflected in transport properties as it is discussed in the next section. 3.2. Transport Properties. Transport features are analyzed in Figure 5, where only a representative sample of each kind of dopant is depicted for the sake of simplicity: CeO2 as a reference, Eu(+2/+3), Tb, Pr(+3/+4), and Er(+3). The conductivity results for the rest of samples are available in the Supporting Information (Figure S5). Figure 5a presents the Arrhenius plot for the conductivity in air of the selected materials, which shows that any dopant improves the total conductivity of pure ceria due to the generation of intrinsic vacancies. Trivalent dopants may enhance the ionic conductivity better than those in a mixed valence as Tb or Pr, as they provoke a fixed number of oxygen vacancies at this pO2. The oxygen release and consequent reduction of the dopant caused by the rising temperature provokes a change in the Arrhenius slope (apparent activation energy) for Tb- and Prdoped ceria in air. Actually, the temperature of change in the slope matches that observed by HT-XRD analysis. Figure 5b shows the conductivity dependence on the pO2 for the selected materials with and without Co taken at 973 K. The studied range of pO2 includes from oxidant (0.21−1 atm) to moderately reducing atmospheres (10−5−0.21 atm), where the Ce4+ does not reduce to Ce3+ as it has been demonstrated that this reaction initiates at oxygen partial pressures lower than 10−10 atm.45−48 Except for Tb and Pr, all the materials show a mainly ionic conductivity behavior, which is independent of pO2.35 Tb- and Pr-doped ceria present a dependency on

Figure 4. (a) Raman spectra of the Ce0.9Ln0.1O2−δ (Ln = La, Gd, Eu, Pr, Tb, Yb, Er) (sintered at 1573 K) from 400 to 750 cm−1 and (b) F2g Raman shift as a function of the cell parameter for Ce0.9Ln0.1O2−δ and Ce0.9Ln0.1O2−δ + Co calcined at 1573 and 1073 K (top) and after subtracting Grüneisen shifts (bottom). The inset shows a magnification of the band related to the oxygen vacancy formation.

shows corrected Raman frequencies, after subtracting Grüneisen shifts, as a function of the cell parameter. Now, the Raman shift increases with the cell parameter except for Pr, again out of

Figure 5. Arrhenius plot of the total conductivity in air of CeO2 and Tb, Pr, Eu, and Er doped ceria (a) and total conductivity at 973 K of the Co-free and Co-containing CeO2 and Tb, Pr, Eu, and Er doped ceria as a function of oxygen partial pressure (b). 7979



Article

pO 2 −1/6 attributed to the dopant reduction as pO 2 decreases.35,49 Therefore, conductivity is mainly ionic in these oxidizing conditions, and the multiple oxidation state of some dopants makes the ionic conductivity pO2 dependent (region not observed in single-oxidation-state acceptor doped systems as Er and Eu). In general, the addition of cobalt to doped ceria improves the total conductivity, and it is especially evident for Tb and Pr, which present higher improvements. The conductivity enhancement is ascribed to the creation of electronic paths through the grain boundary, where (percolating) nanosized Co-rich phases have been reported to be preferentially located.28 Moreover, Pr and Tb with Co addition show enhanced total conductivity due to (a) the reducibility of the material at lower temperature that allows the oxygen vacancy concentration to be higher at the same temperature with respect to the Co-free samples, (b) the electronic conductivity promotion, and (c) the higher sample density and grain packing due to high sintering activity of Cocontaining ceria observed by SEM.28,32,35,41 3.3. Relationship between Structural and Transport Properties. In the following section the correlation between structural and transport properties of these series of lanthanidedoped ceria will be discussed. It has been widely reported that a dopant minimizing the internal strain will maximize the ionic conductivity as it has been traditionally observed for Gd- or Sm-doped ceria. In addition, especial attention will be paid to the influence that the structure induces over the activation energy (Ea). Fundamental physics behind the ionic conductivity determine that Ea is equal to the sum of the vacancy formation energy (Ef) and the migration barrier (Em) for pure ceria. The introduction of aliovalent dopants produces intrinsic oxygen vacancies whose vacancy-dopant association energy (Eass) influences vacancy mobility, and hence Ea can be understood as the sum of Em and Eass for bulk diffusion in doped ceria. Quantum mechanical calculations on Sm-doped ceria have revealed Em values around 0.46 eV, when there are no trivalent ions around a vacancy, while Em values increase up to ∼0.5 eV50,51 when neighboring dopant cations are present, since it is less favorable for the oxygen atom to move through a point close to large trivalent ions than to move to smaller Ce4+ ions. On the other hand, Eass depends on the distribution and concentration of the trivalent ions. In the present work the concentration of trivalent ions is the same (10%) for all dopants except for Tb and Pr that have a variable concentration depending on temperature and pO2. Thus, the vacancy-dopant distribution will be essential in the Eass performance. The experimental Ea values derived from the conductivity measurements corresponding to the different samples are represented in Figure 6 (top) as a function of the cell parameter at 973 K in air. Ea has been calculated over the full temperature range for samples with trivalent dopants. For Tb- and Pr-doped ceria, the plotted Ea was the one in the 1073−923 K range that includes the temperature at which the cell parameter was measured. Ea decreases with increasing cell parameter of the materials following a trend proportional to 1/a2 and tend to reach a fix value of 0.73 eV. The observed tendency can be easily related to Coulombic interactions and steric effects due to the strong dependency of the displacement distance upon dopant type and relative dopant-vacancy positions.52 The lowest Ea, obtained for samples with the largest cell parameters, are related to materials with an optimal balance between elastic (lattice distortions)

Figure 6. Ea (top) and total conductivity at 973 K in air (bottom) of the different doped materials versus cell parameter measured at 973 K. Cell parameters at 973 K in air have been calculated from HT-XRD or by using reported TEC values for the other different dopants.

and electronic defect interactions (redistribution of the electronic structure). When the cell parameter is shorter than that of pure ceria (which is approximately equal to Gd-doped material), the structure is shrunk and Ea follows Coulombic tendency. For longer cell parameters, since the structure is not changed (Gd doped) up to the largest cell parameter (La doped), the Coulombic limit is overpassed and Ea reaches a steady value defined by dopant concentration. Experimental data has been fitted following eq 3 A Ea = E0 + 2 a − B2

(3)

where a is the calculated cell parameter at 973 K, E0 is the limit where Coulomb interactions do not affect the total energy for sufficient large cell parameters, A is a constant that principally includes charges and electric constants, and B is the minimum cell parameter that allows oxygen diffusivity.52 Solid lines have been fitted to eq 3 with only samples with trivalent dopants and dashed fits include Pr and Tb. The deviation from general trend of these dopants, especially Pr, is due to the remaining and variable tetravalent cations. This fact confirms the importance of the interatomic distances that govern Ea behavior, as they present a cell parameter according to their mixed valence state (as discussed in Figures 1 and 2) and almost fall in the same trend as all the other trivalent dopants. In fact, Ea decreases for Pr- and Tb-doped samples at lower pO2 (more trivalent dopant, higher vacancy density, and thus longer cell parameter) while trivalent dopants maintain their Ea (not structural changes). Raman measurements also confirm the Ea behavior as they predicted the highest oxygen vacancy disorder, and thus the lowest Eass, for the longest cell parameters (Figure 1) and the exceptional behavior of Prdoped samples. By fitting to eq 3 a steady E0 value of 0.70 ± 0.08 eV (0.68 ± 0.16 eV considering Tb and Pr) for Co-free series and 0.77 ± 0.08 eV (0.76 ± 0.13 eV considering Tb and Pr) for Cocontaining samples have been obtained. The higher activation energy of Co-doped samples results from the location of Co along the grain boundary as mentioned above.41 It is worthy to note the opposite behavior of the Tb-doped sample, whose Ea 7980



Article

attached in Figure S2; XRD patterns represented in Figure S3 show the pure ceria phase for all Co-free and Co-containing doped cerias; Figure S4 plots the obtained Raman correlation length as a function of the cell parameter; additional conductivity results of all doped cerias are plotted in Figure S5. This material is available free of charge via the Internet at http://pubs.acs.org.

decreases when Co is added since it has not the grain boundary segregation and the Co is principally incorporated in the ceria solid phase.35 The minimum cell parameter that allows oxygen diffusivity, B, was found to be close to 5.4 ± 0.9 Å. As the accuracy of B parameter is not high enough, it can be found in the literature53 a minimum cell parameter of 5.3987 Å corresponding to Lu, the smallest rare earth dopant, which presents the minimum ionic conductivity. Furthermore, Ea ≈ E0 for materials from the nondistorted cell (Gd doped) and longer cell parameters. The influence of the structural parameters on the transport properties is also illustrated in Figure 6 (bottom), which represents the total conductivity of all the samples (with and without Co) measured in air at 973 K as a function of the cell parameter at 973 K in air. Predominant ionic conductivity is expected, and confirming reported studies, there is a maximum conductivity for Gd- and Nd-doped ceria due to the low association energy between the oxygen vacancies and the trivalent dopants and the lowest migration energy minimized to those materials with cell parameters close to pure ceria,14,53 which represents also the less distorted cell. Furthermore, it should be pointed out that electronic transport starts to contribute to the total transport for the Tb- and Pr-doped materials with Co addition35 which could explain the dispersion of the data for these materials.

■

Corresponding Author

*Fax 0034.963.877.809; e-mail [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Financial support by the Spanish Ministry for Science and Innovation (Project ENE2008-06302, ENE2011-24761 and Grant BES-2009-015835), Spanish Industry Ministry (Project IAP-560620-2008-17), and Instalaciones Inabensa S.A. is kindly acknowledged. Dr. José L. Jordá contributed to this work with HT-XRD diffraction experiments.

■

REFERENCES

(1) Sahibzada, M.; Steele, B. C. H.; Zheng, K.; Rudkin, R. A.; Metcalfe, I. S. Catal. Today 1997, 38, 459−466. (2) Serra, J. M.; Vert, V. B.; Betz, M.; Haanappel, V. A. C.; Meulenberg, W. A.; Tietz, F. J. Electrochem. Soc. 2008, 155, B207− B214. (3) Sun, C.; Stimming, U. J. Power Sources 2007, 171, 247−260. (4) Fergus, J. W. Solid State Ionics 2006, 177, 1529−1541. (5) Park, H. J.; Choi, G. M. J. Eur. Ceram. Soc. 2004, 24, 1313−1317. (6) Sunarso, J.; Baumann, S.; Serra, J. M.; Meulenberg, W. A.; Liu, S.; Lin, Y. S.; Diniz da Costa, J. C. J. Membr. Sci. 2008, 177, 13−41. (7) Fagg, D. P.; Kharton, V. V.; Shaula, A.; Marozau, I. P.; Frade, J. R. Solid State Ionics 2005, 176, 1723−1730. (8) Shuk, P.; Greenblatt, M. Solid State Ionics 1999, 116, 217−223. (9) Jasinski, P.; Suzuki, T.; Anderson, H. U. Sens. Actuators, B 2003, 95, 73. (10) Sugiura, M. Catal. Surv. Asia 2003, 7, 77−87. (11) Chueh, W. C.; Falter, C.; Abbott, M.; Scipio, D.; Furler, P.; Haile, S. M.; Steinfeld. Science 2010, 330, 1797−1801. (12) Grirrane, A.; Corma, A.; Garcia, H. Science 2008, 322, 1661− 1664. (13) Trovarelli, A. Catal. Rev. 1996, 38, 439−520. (14) Mogensen, M.; Sammes, N. M.; Tompsett, G. A. Solid State Ionics 2000, 129, 63−94. (15) Atkinson, A. Solid State Ionics 1997, 95, 249−258. (16) Krishnamurthy, R.; Sheldon, B. W. Acta Mater. 2004, 52, 1807− 1822. (17) Steele, B. C. H. Solid State Ionics 2000, 129, 95−110. (18) Yahiro, H.; Eguchi, K.; Arai, H. Solid State Ionics 1989, 36, 71− 75. (19) Etsell, T. H.; Flengas, S. N. Chem. Rev. 1970, 70, 339−376. (20) Kharton, V. V.; Figueiredo, F. M.; Navarro, L.; Naumovich, E. N.; Kovalevsky, A. V.; Yaremchenko, A. A.; Viskup, A. P.; Carneiro, A.; Marques, F. M. B.; Frade, J. R. J. Mater. Sci. 2001, 36, 1105−1117. (21) Serra, J. M.; Vert, V. B.; Büchler, O.; Meulenberg, W. A.; Buchkremer, H. P. Chem. Mater. 2008, 20, 3867−3875. (22) Mori, T.; Drennan, J.; Lee, J. H.; Li, J. G.; Ikegami, T. Solid State Ionics 2002, 154, 461−466. (23) Tadokoro, S. K.; Muccillo, E. N. S. J. Eur. Ceram. Soc. 2007, 27, 4261−4264. (24) Maschio, S.; Aneggi, E.; Trovarelli, A.; Sergo, V. Ceram. Int. 2008, 34, 1327−1333. (25) Luo, H.; Efimov, K.; Jiang, H.; Feldhoff, A.; Wang, H.; Caro, J. Angew. Chem., Int. Ed. 2011, 50, 759−763.

4. CONCLUSIONS This work summarizes a complete structural and conductivity study on different doped cerias. A large series of partially substituted cerias were prepared and characterized by XRD, micro-Raman spectroscopy, and conductivity measurements in order to have an overall understanding of the structural relationship with the transport properties. Several lanthanides were incorporated in the fluorite structure, and it was studied the effect of the final sintering temperature and the addition of cobalt on the structural properties. The chosen lanthanides (Gd, La, Tb, Pr, Eu, Er, Yb, and Nd) included a large range of ionic radii and different metals exhibiting variable oxidation state under the typical operating conditions for these materials. XRD and HT-XRD combined with Raman spectroscopy allowed studying the substitution effect and Co-doping on the final fluorite structure and estimating the vacancy disorder. Moreover, the amount of each dopant in each oxidation state was also estimated finding that Eu and Yb remain always trivalent but the amount of trivalent Pr and Tb depends on temperature and atmosphere. The lowest E a value was found for materials with nondistorted cells (Gd3+) and long cell parameters. This is ascribed to the decreasing Coulombic interactions together with lowest Eass due to the higher vacancy disorder predicted by Raman. Vacancy-dopant concentration and distribution affect the total conductivity, giving rise to the highest ionic conductivity for the nondistorted cells (Gd-doped ceria) which represent the optimum balance between Coulomb interactions, steric effects, and vacancy distribution.

■

AUTHOR INFORMATION

ASSOCIATED CONTENT

* Supporting Information S

A scheme of coprecipitation method used for sample preparation is depicted in Figure S1; thermogravimetric measurements of the formation of the oxides and total decomposition of nitrates and carbonates from the reagents is 7981



Article

(26) Lobera, M. P.; Serra, J. M.; Foghmoes, S. P.; Søgaard, M.; Kaiser, A. J. Membr. Sci. 2011, 385, 154−161. (27) Stefanik, T. S.; Tuller, H. L. J. Eur. Ceram. Soc. 2001, 21, 1967− 1970. (28) Fagg, D. P.; García-Martín, S.; Kharton, V. V.; Frade, J. R. Chem. Mater. 2009, 21, 381−391. (29) Balazs, G. B.; Glass, R. S. Solid State Ionics 1995, 76, 155−162. (30) Pikalova, E.Yu.; Demina, A. N.; Demin, A. K.; Murashkina, A. A.; Sopernikov, V. E.; Esina, N. O. Inorg. Mater. 2007, 43, 735−742. (31) Nicholas, J. D.; De Jonghe, L. C. Solid State Ionics 2007, 178, 1187−1194. (32) Jud, E.; Gauckler, L. J. J. Electroceram. 2005, 15, 159−166. (33) Perez-Coll, D.; Ruiz-Morales, J. C.; Marrero-Lopez, D.; Nunez, P.; Frade, J. R. J. Alloys Compd. 2009, 467, 533−538. (34) Fagg, D. P.; Perez-Coll, D.; Nunez, P.; Frade, J. R.; Shaula, A. L.; Yaremchenko, A. A.; Kharton, V. V. Solid State Ionics 2009, 179, 2372−2378. (35) Balaguer, M.; Solís, C.; Serra, J. M. Chem. Mater. 2011, 23, 2333−2343. (36) Yashima, M.; Takizawa, T. J. Phys. Chem. C 2010, 114, 2385− 2392. (37) McBride, J. R.; Hass, K. C.; Poindexter, B. D.; Weber, W. H. J. Appl. Phys. 1994, 76, 2435−2441. (38) Zhang, F.; Chan, S. W.; Spanier, J. E.; Apak, E.; Jin, Q.; Robinson, R. D.; Herman, I. P. Appl. Phys. Lett. 2002, 80, 127−129. (39) Kumar, A.; Babu, S.; Karakoti, A. S.; Schulte, A.; Seal, S. Langmuir 2009, 25, 10998−11007. (40) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751−767. (41) Jud, E.; Zhang, Z.; Sigle, W.; Gauckler, L. J. J. Electroceram. 2006, 16, 191−197. (42) Weber, W. H.; Hass, K. C.; McBride, J. R. Phys. Rev. B 1993, 48, 178−185. (43) Wang, X.; Hanson, J. C.; Liu, G.; Rodriguez, J. A.; Iglesias-Juez, A.; Fernandez-Garcia, M. J. Chem. Phys. 2004, 121, 5434−5444. (44) Patil, S.; Seal, S.; Guo, Y.; Schulte, A.; Norwood, J. Appl. Phys. Lett. 2006, 88 (1−3), 243110. (45) Fagg, D. P.; García-Martín, S.; Kharton, V. V.; Frade, J. R. Chem. Mater. 2009, 21, 381−391. (46) Fagg, D. P.; Marozau, I. P.; Shaula, A. L.; Kharton, V. V.; Frade, J. R. J. Solid State Chem. 2006, 179, 3347−3356. (47) Duncan, K. L.; Wang, Y.; Bishop, S. R.; Ebrahimi, F.; Wachsman, E. D. J. Appl. Phys. 2007, 101 (044906), 1−6. (48) Chatzichristodoulou, C.; Hendriksen, P. V.; Hagen, A. J. Electrochem. Soc. 2010, 157, B299−B307. (49) Bishop, S. R.; Stefanik, T. S.; Tuller, H. L. Phys. Chem. Chem. Phys. 2011, 13, 10165−10173. (50) Anderson, D.; Simak, S. I.; Skorodumova, N. V.; Abrikosov, I. A.; Johansson, B. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 3518−3521. (51) Ismail, A.; Hooper, J.; Giorgi, J. B.; Woo, T. Phys. Chem. Chem. Phys. 2011, 13, 6116−6124. (52) Frayret, C.; Villesuzanne, A.; Pouchard, M.; Mauvy, F.; Bassat, J. M.; Grenier, J. C. J. Phys. Chem. C 2010, 114, 19062−19076. (53) Omar, S.; Wachsman, E. D.; Jones, J. L.; Nino, J. C. J. Am. Ceram. Soc. 2009, 92, 2674−2681.

7982


(Ln = Gd, La, Tb, Pr, Eu, Er, Yb, Nd) and Effect of ... - ACS Publications

Recommend Documents