A.S. Nowick

High-temperature protonic conductors with perovskite-related structures

Abstract The various perovskite-structured oxides found to be high-temperature protonic conductors are reviewed. They fall into two groups: the simple or ABO3 types and the complex (or mixed) types of formula A2B′B″O6 and A3B′B2″O9. Some microstructural features observed by TEM and electron diffraction are pointed out. A simple “first-order” model is then given to serve as a basis for consideration of various fundamental aspects of these protonic conductors. Three fundamental aspects are considered: the nature of proton incorporation by exposure to water vapor at high temperatures, the environment in which protons are located in the lattice, and the nature of proton migration. Proton incorporation is studied by water uptake experiments and the proton environment by spectroscopic techniques. Proton migration is studied by conductivity and isotope-effect experiments, which show that migration occurs by simple hopping of protons from one O2− ion to the next. However, the classical absolute rate theory (ART) requires modification because of the low efficiency of energy transfer between an excited proton and the cage of heavy ions that surrounds it.

Oxygen-ion conductivity and defect interactions in yttria-doped ceria☆

Conductivity, σ, of the oxygen-ion conductivity solid electrolyte system CeO2:Y2O3 is studied in detail as a function of temperature and dopant concentration, c0, from 0.05 to 40 mole% Y2O3. It is shown that the well-known maximum in σ versus c0 is composed of two distinct ranges. In the dilute range (<1%Y2O3), an oxygen vacancy is associated with one Y3+ ion to form a charged (incompletely compensated) pair. Here σ increases, due to a linear decrease in association enthalpy HA with c130 resulting from electrostatic interactions. In the high-concentration range, on the other hand, σ decreases and HA increases due to the development of deep vacancy traps.

Ionic conductivity of CeO2 with trivalent dopants of different ionic radii

Abstract The ionic conductivity, σ, of dilute CeO 2 :M 2 O 3 solid solutions has been investigated for four different dopants M 3+ , viz., La 3+ , Gd 3+ , Y 3+ and Sc 3+ . Conductivity does not vary monotonically with dopant radius but is a maximum for Gd 3+ doping. The association energy, H A , of the MV O pairs (V O = oxygen-ion vacancy) varies in an inverse manner to σ. For Sc 3+ doped samples, σ is ≈10 3 times lower than for the other dopants. Double doping experiments also show that Sc 3+ acts as a scavenger for vacancies. After considering other models, it is concluded that the behavior of Sc can only be explained by the formation of ScV O pairs of very high association energy.

The “grain-boundary effect” in doped ceria solid electrolytes

Abstract The ac electrical behavior of sintered polycrystalline CeO 2 :CaO and CeO 2 :Y 2 O 3 solid electrolytes is studied. Complex impedance plots show the presence of an extra arc due to the presence of relatively high-resistivity material along the grain boundaries. This “grain-boundary effect” is greatest for dilute solid solutions and “pure” samples and becomes vanishingly small for dopant concentrations ≳ 15 mole%. The effect can also be reduced by shortening the sintering time. The activation enthalpy for the grain-boundary conductivity is substantially higher than that of the lattice conductivity, especially at low dopant levels. This fact eliminates the Bauerle constriction model. The results suggest that the effective dopant concentration near the grain boundaries may be substantially reduced from the bulk value.

The incorporation and migration of protons in Nd-doped BaCeO3

Abstract A detailed study of protonic conductivity in Nd-doped BaCeO 3 is made using water-uptake and galvanic-cell measurements to supplement the conductivity results. Samples were pretreated at 900°C at different P H 2 O atmospheres and the conductivity measured with a frozen-in proton concentration. Isotope effect (D 2 O versus H 2 O) measurements were also made. The results show that interstitial protons enter the lattice as compensators for the dopant, replacing oxygen-ion vacancies, and migrate by hopping between adjacent O 2- ions with a low migration energy of 0.54 eV. Comparison with other acceptor-doped perovskite oxides is made in order to better understand the origin of the high protonic conductivity in this material.

Fast high-temperature proton transport in nonstoichiometric mixed perovskites

Two different nonstoichiometric mixed perovskites: Sr2(Sc1+)xNb1−xO6−σ (with x=0.05 and 0.1) and Ba3(Ca1.18Nb1.82)O9−σ, are shown to become good high-temperature protonic conductors upon exposure to H2O (or D2O) atmospheres. Conductivity is studied in the temperature range of 300 ∼550 K where the proton concentration is frozen in. The activation energies EH for proton migration in these compounds are 0.62 and 0.54 eV, respectively. The conductivities fall in the same range as those for M3+-doped SrCeO3 and BaCeO3, but the present materials do not show electronic conduction after treatment in highly reducing atmospheres as do the cerates. It is therefore concluded that these compounds deserve serious consideration as fast high-temperature protonic conductors.

High-temperature protonic conduction in mixed perovskite ceramics

Abstract Protonic conduction of both stoichiometric and nonstoichiometric mixed perovskite ceramics with the general formula A 2 ( B ′ 1+x B″ 1−x ) O 6−δ ( A = Sr 2+ , Ba 2+ ; B ′= Ga 3+ , Gd 3+ , Nd 3+ ; B ″= Nb 5+ , Ta 5+ ; x=0−0.2) has been investigated. The B-site occupancies of B and B can be ordered or disordered depending on the ionic radius of the A ion and the ionic radius difference of the B and B ions. Samples are treated in H2O vapor at 900°C to incorporate protons (or D2O vapor for deuterons) and the amount of water uptake is measured by the weight change. None of the stoichiometric samples (x=0) shows any proton incorporation. For nonstoichiometric samples the amount of water uptake varies from 5% to 50% of the maximum theoretical amount, depending on the material and thermal history. Protonic conductivity of the nonstoichiometric materials is measured by ac impedance analysis in a frozen-in condition following the pre-treatment in water vapor. The conductivity is found to be higher and the activation energy lower for the slowly cooled materials, with magnitudes comparable to that of Yb-doped SrCeO3. All of the nonstoichiometric samples show non-classical isotope effect of the conductivity, in the sense that the activation energy is slightly higher (by ∼0.04 eV) for deuteronic than for protonic conduction, similarly to results reported earlier for SrCeO3 and BaCeO3. Finally, the conductivity shows stage II behavior with activation energy equal to the proton migration enthalpy, but the number of free carriers is smaller than the total proton uptake.

Nature of the ac conductivity of ionically conducting crystals and glasses

Abstract Many disordered systems, both crystalline and glasses, show ‘universal dynamic response’, in which the ac conductivity as a function of frequency is given by σ(ω) = σ(0) + Aωs. Here σ(0) is the dc conductivity and the exponent s (⪕ 1) generally decreases with increasing temperature, T. For various ionically conducting crystals and glasses, we have obtained two ranges in which s = constant, independent of T: at high temperatures s = s0 (a value near 0.6) called the ‘Jonscher regime’, and at low temperatures s = 1, called the constant-loss regime. For the Jonscher regime, the parameter A is activated with activation energy closely related to that of σ(0). On the other hand, for the constant-loss regime (which extends down to cryogenic temperatures), the parameter A varies only slowly with T and is not activated. For the cases of borate and silicate glasses as well as heavily doped crystalline CaTiO3 and CeO2, careful analysis was carried out over the intermediate regime in which the effective s falls from 1.0 to s0. The results show that this regime can simply be described as a superposition of the Jonscher and constant-loss behaviors. Work by others on silicate glass over a much wider frequency range than that of the present study falls into the same pattern. Finally, a brief review is given of possible mechanisms of both the Jonscher and constant-loss types of behavior.

Some factors that determine proton conductivity in nonstoichiometric complex perovskites

Abstract Complex perovskites, of the types A2(B′B″)O6 and A3(B′B2″)O9, (in which A ions are always 2+, B′ ions are 3+ and 2+, respectively, and B″ ions are 5+), can become good high-temperature protonic conductors when they are made nonstoichiometric by increasing the concentration of the B′ ions at the expense of the B″ ions. After heating in water vapor, uptake of H2O occurs, and protons enter the structure. We consider the effects of several variables on the bulk ionic conductivity and activation energy for protonic conduction, including degree of nonstoichiometry, the effect of ordering on the B-sites, the effect of varying the radius of the B′ ion and the effect of changing the A ion from Sr2+ to Ba2+. The results show that disorder favors a higher conductivity, that the degree of order can change with nonstoichiometry, and that Ba compounds always have higher conductivity and lower activation energy than the corresponding Sr compounds.

X-ray and neutron diffraction study of intermediate phases in nonstoichiometric cerium dioxide

Powder samples of reduced ceria, CeO2−x, of known compositions in the range 0 < x < 0.3 have been examined by X-ray and neutron diffraction techniques in order to determine which intermediate phases belonging to the homologous series CenO2n−2 (with n = integer) truly exist. Through the appearance of superlattice lines in the neutron diffraction patterns, the existence of four distinct phases, corresponding to n = 7, 9, 10 and 11 was established. Aside from the phase Ce7O12, the structures of these phases cannot be accounted for with rhombohedral cells based on 〈111〉 vacancy strings, but indicate lower (monoclinic or triclinic) symmetry. The structure of Ce9O16 and Ce10O18 do not agree with structures proposed for the analogous PrnO2n−2 compounds.