Solid oxide fuel cells (SOFCs) have been considered for almost a century. Yet, despite their promising high efficiency, commercial cells on a broad scale are not available. Operation at elevated temperatures introduces material stability and sealing problems, while lowering the temperature results in excessively high impedance in bulk diffusion and surface reactions. Many attempts have been made to lower the SOFC membrane and electrode impedance. We discuss the different solutions examined, which include changing dense membrane compositions by doping to enhance ionic conductivity; considering mixed ionic-electronic conductors as membranes; simplifying the cell construction by allowing porous membranes and applying a mixture of fuel and air to both electrodes; making the membrane very thin; considering highly porous membranes that take advantage of ion conductivity in a chemisorbed layer on the pore surfaces; and manipulating the catalytic properties of the cell electrodes.

Experimental results of the defect concentrations in highly reduced, monomolecular surface layer of ceria-based oxides, are discussed. The data is XPS vs. oxygen pressure relations of reduced Pr0.1Ce0.9O2.-x (PCO), CeO2-x and Sm0.2Ce0.8O1.9-x (SDC). In PCO the analysis predicts that the surface layer is negatively charged and the concentration of Pr3+ ions is higher than in the bulk. A double layer exists between the surface layer and the bulk. In CeO2 and SDC the concentration of Ce3+ ions and oxygen vacancies in the surface is higher than in the bulk. The surface is neutral. The analysis predicts that the surface layer is metallic, i.e. the electrons on Ce3+ are delocalized and not localized small polarons. The bulk is a semiconductor. The Ce3+ ions are randomly distributed on the Ce sublattice and not in the boundary of oxygen vacancies. The latter are doubly ionized vacancies VO center dot center dot despite the presence of a high concentration of quasi free electrons. It is also predicted that the surface of ceria or SDC has a phase diagram of temperature vs. oxygen vacancy concentration, at T > 450 degrees C, like that of ceria bulk or SDC bulk, respectively. Further, the phase diagram of SDC bulk (and surface) is like that of ceria bulk shifted to a higher oxygen vacancy concentration and doping has no other effect at elevated temperature. Both in ceria and SDC the difference between the surface and the corresponding bulk is only in the oxygen pressure at which a level of reduction is reached, with the surface more easily being reduced than the bulk.

The monomolecular surface layer of acceptor doped CeO2 may become neutral and metallic or charged and semiconducting. This is revealed in the theoretical analysis of the oxygen pressure dependence of the surface defects concentration in acceptor doped ceria with two different dopant types and operated under different oxygen pressures. Recently published experimental data for highly reduced Sm0.2Ce0.8O1.9−x (SDC) containing a fixed valence dopant Sm3+ are very different from those published for Pr0.1Ce0.9O2−x (PCO) with the variable valence dopant Pr4+/Pr3+ being reduced under milder conditions. The theoretical analysis of these experimental results fits very well the experimental results of SDC and PCO. It leads to the following predictions: the highly reduced surface of SDC is metallic and neutral, the metallic surface electron density of state is gs = 0.9 × 1038 J−1·m−2 (1.4 × 1015 eV−1·cm−2), the electron effective mass is meff,s = 3.3me, and the phase diagram of the reduced surface has the α (fcc) structure as in the bulk. In PCO a double layer is predicted to be formed between the surface and the bulk with the surface being negatively charged and semiconducting. The surface of PCO maintains high Pr3+ defect concentration as well as relative high oxygen vacancy concentration at oxygen pressures higher than in the bulk. The reasons for the difference between a metallic and semiconducting surface layer of acceptor doped CeO2 are reviewed, as well as the key theoretical considerations applied in coping with this problem. For that we make use of the experimental data and theoretical analysis available for acceptor doped ceria.

Recent experimental results of the dependence on oxygen partial pressure, P(O2), of the small polarons concentration in the monomolecular surface layer of doped ceria, are interpreted theoretically. Two different, experimental conditions are examined. First, the highly reduced surface of Sm0.2Ce0.8O1.9 – δ (SDC) is treated and the [Ce^3 + ]_s-P(O_2) relations are evaluated. The experimental data are explained by a metallic surface layer, approximately neutral, i.e. with negligible transfer of charge to the bulk. The surface is metallic as the outer electrons on Ce3+ ions, which are small polarons when under low concentration, lose their localization under high concentration. Second, the surface of Pr0.1Ce0.9O2 – δ (PCO) is examined. The experimental data of [^3 + ]_s-P(O_2) relations are explained by a semiconducting, negatively charged, monomolecular surface layer forming a double layer with a positive bulk. In both SDC and PCO the surface defect concentration dependence on P(O2) is different from that of the bulk.

The dependence on oxygen pressure, P(O-2), of the Pr3+ concentration, [Pr3+]s, in the surface monomolecular layer of Pr0.1Ce0.9O2-delta, is analyzed. The surface considered, as a core, is negatively charged due to excess Pr3+ ions over oxygen vacancies, VO, while in the bulk the diffuse charge is positive due to excess oxygen vacancies over Pr3+ ions. The [Pr3+]s - P(O-2) relations are derived for the full range of reduction of Pr4+ to Pr3+. Under low P(O-2), [Pr3+]s reaches the limit of the total concentration of Pr ions. The high P(O-2) asymptotic relation is, [Pr3+](s) proportional to P(O-2)(-1/8). This is different from that in the neutral bulk, [Pr3+](b) proportional to P(O-2)(-1/6). For calculating the [Pr3+]s P(O-2) relations the space charge distribution in the bulk need not be fully evaluated.

The concentrations of the point defects, Ce3+ and oxygen vacancy, in the highly reduced, monomolecular surface layer of Sm doped ceria, are analyzed applying a method in which the surface layer and the rest of the bulk are treated as two separate, neutral, phases, then brought into contact and the point defect concentrations modified if charge transfer takes place. The theory yields a very weak dependence of the surface negative charge concentration, [Ce3+], on the oxygen partial pressure, P(O-2), being at most logarithmic [Ce3+]alpha - ln (P(O-2)), becoming even weaker for very low P(O-2). This is in excellent agreement with experimental results reported by Chueh et al. (2012) for reduced Sm0.2Ce0.8O1.9-x (SDC). The analysis treats the surface with a high density of small polarons as a two-dimensional, neutral, metallic layer in which the small polarons lose their localization. The bulk of the oxide is a semiconductor. The surface defect band generated by the cations Ce4+ and Ce3+ has a density of states g(s) = 1.4x10(15)eV(-1)cm(-2) and the effective mass of the electrons is m(eff) = 3.3m(e). The reduced surface exhibits the f.c.c., alpha phase, known for bulk ceria and bulk doped ceria. There is no significant difference between the a phase composition range in the surface layer and in the bulk of SDC. There is a significant difference in the oxygen partial pressure under which the alpha phase in the surface and in the bulk is generated. The concentration of the dopant Sm is predicted to be 40% higher in the surface layer than in the bulk in agreement with experiment.

The dependence of point defect concentrations in the surface layer (the outer most layer of the bulk) of binary oxides, on oxygen partial pressure, P(O2), and acceptor concentration, Ab, is evaluated. With one calibration point the defect concentration, at constant temperature, under any P(O2) and Ab can be determined. The method of calculation is demonstrated on binary oxides which exhibit oxygen vacancies, electrons and holes. A relation between the dependence on P(O2) and Ab of concentrations of point defects in the surface layer, and the dependence of the concentrations of the corresponding defects deep in the neutral bulk, is presented. Two types of surface layers are distinguished according to their electron states. One that forms a band which is partially populated by electrons, leading to a two dimensional metallic surface and Fermi level pinning (FLP) and a second, insulating one and no FLP. Four characteristic cases are identified: FLP or the absence of FLP in the surface layer combined with the presence or absence of significant chemisorption. It is shown that the dependence on P(O2) and Ab of the defect concentrations in the surface layer may be the same as that of the corresponding defects deep in the neutral bulk. The difference is only in the magnitude. The conditions for that to take place are an insulating surface and low concentration of chemisorbed particles. The latter conditions are quite common, as the surface of oxides is mostly not metallic, oxygen chemisorption is low at elevated temperature and for acceptor doped, p-type oxides, chemisorption is also low at low temperature. Different relations are derived under other conditions. In particular, for metallic surface, the oxygen vacancy concentration in the surface layer follows only one dependence, P(O2)−1/2, irrespective of the dependence on P(O2) and Ab of the bulk vacancy concentration and irrespective of chemisorption. The electrical potential distribution is derived for the four cases.

Semiconducting-ionic membranes (SIMs) have exhibited significant superiority to replace the conventional ionic electrolytes in solid oxide fuel cells (SOFCs). One interesting phenomenon is that the SIMs can successfully avoid the underlying short-circuiting issue and power losses while bringing significantly enhanced power output. It is crucial to understand the physics in such devices as they show distinct electrochemical processes with conventional fuel cells. We first presented experimental studies of a SIM fuel cell based on a composite of semiconductor LiCo0.8Fe0.2O2 (LCF) and ionic conductor Sm-doped CeO2 (SDC), which achieved a remarkable power density of 1150 mW cm(-2) at 550 degrees C along with a high open circuit voltage (OCV) of 1.04 V. Then, for the first time we used a physical model via combining a semiconductor-ionic contact junction with a rectifying layer which blocks the electron leakage to describe such unique SIM device and excellent performance. Current and power are the most important characteristics for the device, by introducing the rectifying layer we described the SIM physical nature and new device process. This work presented a new view on advanced SIM SOFC science and technology from physics. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Determining the ionic mobility in solids is often challenging due to inherently low ionic conductivities, typically requiring measurements at elevated temperatures, with high impedance analyzers and/or extended resistivity transients. Moreover, in many solids, the ionic conductivity is masked by dominant electronic conductivity, characterized by much higher carrier mobility. By focusing our measurements on nano-dimensioned thin films, we are able to overcome such limitations. First, measurement time and temperature can be reduced, the latter to near ambient conditions, in selected materials, due to the considerably faster response of thin films compared to bulk materials to applied electrical stimulus. Second, the effect of the redistribution of ionic defects on the nano-scale, at relatively short time, has a major impact on the total conductivity. The proposed method relies on measuring the nonlinear I-V response of thin films to linear voltage sweeps of high amplitude (relative to the thermal voltage) at different sweep rates. This differs from the impedance spectroscopy technique that analyzes the small signal response of systems. The method also differs from the well-known technique of cyclic voltammetry for deriving ion diffusion coefficients, as it does not require redox reactions to take place during voltage sweeps, thereby making it more generally applicable. In this work, we present a novel and simple method for determining the ionic mobility from the measured I-V relations. It is determined from the position of a peak in the I-V relations and the sweep rate. The mobility so derived is compared with the one obtained by fitting the I-V relations based on solving the drift-diffusion equations for ions and electrons as published recently. Though the near ambient temperature ionic mobility was found to be seven orders of magnitude lower than the corresponding electronic one, it could nevertheless be well deconvoluted and characterized by both methods. The experimental results are compared with previous measurements on bulk and powder samples of molybdenum trioxide.

The I-V curve of forming, switching between SET and RESET and memory are evaluated theoretically. The key features of the model are: a) drift diffusion of ions and of electrons, b) stoichiometry changes by exchange of oxygen with the ambient, c) temperature changes with current, d) Mott transitions in forming the filament and e) switching by changes in tunneling through a thin insulating layer. Forming, SET and RESET are induced by stoichiometry changes. Both voltage ramp and voltage pulses are considered. For pulses the interplay between pulse duration and pulse voltage is evaluated.

The use of isotope exchange was recently shown to have a potential in determining continuous series of elementary steps in surface reactions. This was demonstrated by analyzing oxidation reactions on oxides. The method is based on equilibrating the oxide in 16O2 then exposing it to 18O2 and determining the dependence of the rate of 18O16O molecule desorption on oxygen partial pressure and on optional acceptor doping of the oxide. The isotope 18O2 can be applied as a step function or as a pulse. We here discuss an implicit assumption made concerning the reaction that governs the isotope desorption, discuss the integrated desorption measured after the termination of a pulse of 18O2, extend the discussion to isotopes other than oxygen and in particular show how to cope with the mass effect when using deuterium and hydrogen and finally discuss the relevance of isotope exchange experiments, where chemical equilibrium prevails, to analyzing electrocatalysis and heterogeneous catalysis in which significant gradients in the defect concentrations may arise.

The fuel cell design can be simplified and the energy density and fuel utilization improved while the production cost lowered if a fuel cell is operated on a mixture of fuel and air. This type of fuel cell, denoted as mixed reactant fuel cell, imposes stringent catalytic constraints on the materials of the cell in particular the electrodes but not only on them. The catalytic requirement are first of all selectivity for certain reactions to take place while others to be suppressed. Enhancing the rate of desired reactions is a later step. We here discuss possible ways to achieve the required selectivity.

The width δ of the active electrode area along the triple phase boundary, TPB, is determined using a novel method based on ac impedance measurements between two interdigitated electrodes. The theoretical basis is presented and used to derive the width δ for gold electrodes applied onto the solid electrolyte (SE) Y2O3 stabilized ZrO2 (YSZ). Gold electrodes are used in the demonstration of the method due to their ion blocking nature and poor cathodic properties which requires that all or at least most of the electrochemical reaction steps take place on the free surface of the SE near the TPB. The measurements are done at relatively low temperatures of 250–300 °C. The width δ is determined to be 0.4, 0.3 and 0.2 μm at 250, 275 and 300 °C, respectively, decreasing with increasing temperature. In addition to the width δ, the ionic conductivity of YSZ at those low temperatures is determined as 2.2 × 10− 7, 7.2 × 10− 7 and 2.3 × 10− 6 S/cm at 250, 275 and 300 °C, respectively, with an activation energy of 1.27 eV.

Objective There is no method to date for determining continuous series of elementary steps in an electrode reaction as well as in oxidation and hydration reactions and also in heterogeneous catalysis. The overall reaction, and in many cases also intermediate steps and species, can be identified but not a whole continuous series of elementary steps. We suggest a novel method for determining series of elementary steps in surface reactions. It is based on a new way of applying and interpreting the isotope exchange technique.1,2 The method is applicable to the identification of series of elementary steps in electrocatalysis, heterogeneous catalysis and oxidation and hydration reactions. An outstanding example of a problem for which no answer was found so far is the role of the oxygen vacancy in the cathodic reaction in a solid-oxide-fuel-cell (SOFC). Is the vacancy required for the early step of dissociation of an oxygen molecule or is it involved in the cathodic reaction only at a later step after the molecule is already dissociated? New Results We concentrate on the cathodic reaction in SOFCs. The novel method then requires that the isotope to be exchanged is 18O2. The identification is then based on the relation between the rate, r(16O18O), of the production of the mixed molecules 16O18O, the oxygen partial pressure, P(O2) and the concentration of acceptor doping, [A] (if any) in the oxide on which the reaction takes place. We evaluate the theoretical relations for ten plausible continuous reaction series at a cathode in an SOFC and show that all have different relation of the rate r(16O18O) vs. P(O2 and [A] and can thus be identified once the relation is determined experimentally. Here are two examples of r(16O18O) vs. P(O2) and [A] relations calculated. The first reaction is: O2(gas) + Vad x +VO,s ●● → O­ad \ + OO,s x +3h●(1) where we have used the Kröger-Vink notation of point defects, Vad x is a neutral vacant adsorption site, VO,s ●● a doubly charged oxygen vacancy in the surface layer s, O­ad \ a singly charged, adsorbed oxygen ion, OO,s x an oxygen ion in the lattice of the surface layer and h● a hole in the valence band. The oxygen molecule dissociates by interaction with an oxygen vacancy in the surface layer. I.e. the oxygen vacancy is imperative for the dissociation process. The rate of evaporation of the mixed molecules 16O18O when the oxide saturated with 16O is suddenly exposed to 18O2 is, in the short time approximation, r(16O18O) α P(O2)7/4[A]5/2(2) On the other hand in the following reaction the molecule dissociates on the surface without interaction with an oxygen vacancy: O2(gas) + 2Vad x → 2O­ad \ + 2h●(3) In this case, r(16O18O) α P(O2)7/4[A]1/2(4) Changing the source that provides electronic charges to the oxygen from the valence band to the conduction band changes the rate relation. Thus instead of reaction (1) let us consider the following reaction, O2(gas) + Vad x +VO,s ●● + 3e\ → O­ad \ + OO,s x(5) In this case, r(16O18O) α P(O2)-1/2[A]1/4(6) Significance The method is of fundamental significance. The r(16O18O) vs. P(O2) and [A] relations are specific to the reactions shown. This can be demonstrated for two consecutive reactions, as well.1,2 The latter allows to identify a whole series of elementary steps which include dissociation of the oxygen molecule and which ends by a slow step. The fast elementary steps are combined into a first fast reaction which is followed by a slow step as the second reaction. The corresponding r(16O18O) vs. P(O2) and [A] relations can serve to identify all the elementary steps involved in the two reactions. Acknowledgement This research was supported by Israel Science Foundation, ISF, under grant No. 699/11. Reference I. Riess, Solid State Ionics, 280, 51 (2015). I. Riess, Solid State Ionics, (in press).

Metal/oxide/metal thin devices may exhibit hysteresis and negative differential resistance (NDR) under time-varying voltage at low temperatures that strongly depend on the frequency of the applied voltage. Herein, we demonstrated and analyzed this in Au/MoO3-delta/Au devices, tested at 55-80 degrees C. Reduced MoO3-delta is a mixed ionic-electronic conductor (MIEC) that conducts electrons and oxygen vacancies. Hysteresis and NDR disappear at high scan rates of the applied voltage when the ion motion is practically frozen and at low scan rates when the response to voltage cycles is quasi-static. Contrary to cyclic voltammetry in electrochemistry, peaks that appear and end with NDR are not because of a redox reaction but a result of the dynamics of the ionic motion. A low rate of exchange of oxygen with the ambient is detected during prolonged measurements. The anodic reaction is found to be faster than the cathodic reaction, and the oxide is reduced under (anti) symmetric voltage cycles. Upon fitting a theory previously reported by our group, the electron mobility and activation energy, oxygen vacancy mobility and activation energy, as well as oxygen exchange current density of the electrodes, of the device were obtained at relatively low temperatures.

Mechanisms proposed for redox-based memristors are reviewed. Emphasis is given on MIM (metal/insulator/metal) devices of the type MOM where the insulator is an oxide. The oxide conducts oxygen via oxygen vacancies. MOM devices in which the insulator conducts intercalated cations are analogous to the ones with mobile oxygen vacancies. Switching, memory and short term hysteresis are three independent phenomena governed by different mechanisms. A necessary condition for memory is presented. Electroforming, filament formation and alteration and I-V curve crossing are discussed. A new mechanism for unipolar switching is suggested. The metal electrodes are sorted into four types according to the nature of their oxygen transfer. The effect of humidity in the ambient is discussed.

SrTiO3 (STO) is a model system for studying oxide electronic devices. This work examines the electronic transport through a heterostructure comprising an acceptor (Fe)-doped STO layer on a donor (Nb)-doped STO substrate. This is done by fitting the steady-state current–voltage (I–V) curve measured at ambient temperature to numerical solutions of the drift-diffusion equations of itinerant (i.e., free) electrons and holes (band conduction) and localized electrons hopping through defect states within the bandgap (hopping conduction). The analysis shows that at reverse bias and small forward bias the current is carried mostly by hopping electrons that give rise to unexpectedly high currents. At forward bias above 1 V, most of the current is due to electron transport through the conduction band. The transition from hopping to band conduction occurs when the quasi Fermi levels of electrons and holes depart from the energy level of the defect states through which the electrons hop. These observations shed new light on the transport properties of STO-based devices at ambient temperatures wherein ionic defects such as oxygen vacancies are immobile and holes are trapped, for the most part.

The conclusions reached by a diverse group of scientists who attended an intense 2-day workshop on hybrid organic-inorganic perovskites are presented, including their thoughts on the most burning fundamental and practical questions regarding this unique class of materials, and their suggestions on various approaches to resolve these issues.