Search

Solid Oxide Cell (SOC) technology is considered as an efficient electrolysis technology to produce hydrogen at large scale. It can also operate in fuel cell mode using different fuels (carbon-based or non-carbon based like ammonia), and in reversible mode. Though proofs-of concept have been achieved at different relevant scales for those operating modes, some R&D works still need to be performed to improve performance, durability and cost. Improved and upscaled cells and stacks need to be developed, with a methodology combining multiscale and multiphysics modelling, electrochemical characterization in relevant conditions and advanced post-test analysis. Their integration into modules made of several stacks is also a stepping stone in order to reach multi-MW electrolysers. CEA is working on the whole value chain of SOC technology, from cell development and optimisation to module design and operation through stack upscaling. Recent achievements on those aspects will be presented.

La1-xSrxCoy-1FeyO3-δ (LSCF) is a widely used oxygen electrode in Solid Oxide Cells (SOCs). Despite its high electrochemical activity combined with a good mixed ionic and electronic conductivity, it is prone to a phase decomposition upon operation decreasing the global cell performance. However, the underlying mechanisms for the reaction mechanism and the degradation are still not precisely understood and require investigations at atomic scale. Therefore, Density Functional Theory (DFT) calculations are used to better unravel the mechanisms occurring in the electrode. In particular, the interaction of the dioxygen gas with the AO-terminated (100) surface is studied. In this context, the oxygen reduction reaction is decomposed in a series of elementary steps including oxygen vacancy formation, adsorption of the dioxygen molecule on the bare surface and into a surface vacancy accompanied with the formation of intermediate oxygen species, followed by the dissociation and incorporation into the bulk. The calculations have been carried out at two different Cobalt contents (y=0.125 and y=0.25).

Solid oxide cells (SOCs) have attracted a growing attention thanks to their high efficiency and ability to operate in both electrolysis (SOEC) and fuel cell (SOFC) modes. Despite its great potential, the current SOC technology faces significant cell degradation during long-term operation. The degradation phenomena are still not well understood as they involve complex and intricate processes arising at different length scales. Additionally, it has been shown that the degradation rate under SOEC operation is generally higher than in SOFC mode. To address this issue, durability tests were performed in SOEC mode at 750, 800 and 850 °C for 2000 h at -1 A cm-2 in 10/90 vol.% H2/H2O and dry air. The electrochemical performance of the cells was assessed using electrochemical impedance spectroscopy. In addition, advanced characterization techniques with nanometer and atomic resolution have been used to study material degradation after long-term testing.

A long-term test of 2000 h has been carried out on a typical solid oxide cell in electrolysis mode at −1 A.cm−2 and 750 °C. The 3D reconstructions of the pristine and aged cermet have revealed a strong Ni depletion at the electrolyte interface. To explain this result, an electrochemical and phase-field model has been developed to simulate the Ni migration in Ni/YSZ electrode. For this purpose, a mechanism has been proposed that takes into account the impact of polarization on the Ni/YSZ wettability. In this approach, it assumes that the Ni/YSZ interfacial energy is changed by the concentration of oxygen vacancies in the electrochemical double layer. Thanks to the model, the Ni migration has been computed in the same condition than the experiment and complemented by a simulation in reverse condition in SOFC mode. In good agreement with the experiment, the simulations have revealed a strong Ni depletion at the electrolyte interface after operation under electrolysis current. On the contrary, a negligible Ni redistribution with a very slight Ni enrichment has been predicted at the electrolyte interface after SOFC operation. These results tend to prove the relevance of the mechanism.

Detailed electrochemical and 3D microstructural characterization of Ni/YSZ electrodes confirms that Ni migration also occurs during CO2 electrolysis, most likely driven by the gradient in the Ni–YSZ interfacial energy along the current direction.

International audience; A novel particle-based model is proposed to generate synthetic yet representative 3D microstructures of typical SOC electrodes. The model steps can be related to the real electrode manufacturing routes, classically via powders processing, making it a practical tool for electrodes design optimization. The representativeness of the synthetic microstructures is checked on several two-phase (LSCF, LSC) and three-phase (Ni-YSZ) electrodes reconstructed by synchrotron X-ray and FIB-SEM tomography. The validation shows a very good agreement between the real and synthetic media in terms of metric, topological and physical properties. Furthermore, the model is adapted to simulate the microstructure of a typical Ni-YSZ current collecting layer by taking into account a bimodal pore-size-distribution. In this objective, the macro-pores resulting from the burning-off of specific pore-formers are morphologically separated in the reconstruction from the micro-porosity network. Finally, the geometrical features of the macro-pores are meticulously characterized and successfully emulated by using parametric ellipsoids.

Solid Oxide Cell (SOC) is a high-temperature electrochemical device that can be operated in both fuel and electrolysis modes. It is constituted of a dense electrolyte in Yttria Stabilized Zirconia (YSZ) sandwiched between two porous electrodes. The hydrogen electrode is classically made of a cermet composed of Nickel and YSZ (Ni-YSZ). Because of system failures in operation (e.g. a fuel shortage or air re-introduction during the system shutdown), the Nickel is liable to be re-oxidized inducing its volume expansion. The Ni swelling within the cermet leads to a mechanical damage in the YSZ backbone [1] resulting in a degradation of the overall cell performances [2]. Therefore, it is still required to enhance the hydrogen electrode robustness for a higher redox tolerance. For this purpose, it is essential to predict accurately the formation and distribution of micro-cracks in the YSZ network during the re-oxidation. Nevertheless, the modeling of the initiation and propagation of cracks in complex microstructures of porous ceramics is still a subject of investigation. Moreover, there is also a lack of data on fracture properties for the porous YSZ material. In this frame, the present study was focused to propose a relevant model for the initiation and the growth of cracks in porous ceramics in order to simulate the fracture induced by Ni re-oxidation. For this purpose, a coupled experimental and modeling approach has been adopted (Fig.1). For the experimental part, the mechanical characterizations have been performed on thin YSZ membranes of different porosities representative of SOC microstructures. The compression test has been chosen as it generates a quasi-uniform stress field in the tested samples. To this end, micro-pillars of few tens of micro-meters in section have been milled in the YSZ membrane using a xenon plasma focused ion beam. For each investigated porosity, the compressive fracture strength has been measured on several pillars. The tested samples have been carefully characterized to assess the damage in the microstructure. As expected, the compressive strength was found to decrease with increasing the porosity. Moreover, a transition has been observed from a brittle behaviour at low porosity towards a local damage at high porosity [3]. All these experimental data have been used to validate a numerical tool to simulate the fracture in porous microstructure. In this study, a model based on the Phase Field Method (PFM) has been developed. In this approach, the crack is modeled implicitly through a smooth scalar damage variable and its width is controlled by a length scale parameter. Therefore, this method is well adapted to solve the fracture problem in complex and heterogeneous materials since it is mesh independent. The implemented model corresponds to the one proposed by C. Miehe et al. [4] who proposed a staggered scheme for the numerical resolution. A special attention has been paid in this work to evaluate the model capability to accurately predict the crack initiation for porous electrode material. For this purpose, the results of the PFM simulations have been compared to the mixed criterion proposed by D. Leguillon [5]. The study has been conducted on notched samples with different opening angles and for macro-cracks blunted by different cavity sizes. It has been highlighted that the two approaches provide consistent predictions for the crack initiation. Moreover, it has been checked that the PFM model is able to reproduce accurately the expected decrease of the fracture toughness when increasing the material porosity. Once these studies completed, the PFM model has been used to simulate the micro-compression test using a multiscale approach. It has been found that the model is able to reproduce correctly the decrease of the measured compressive fracture strength with the sample porosity. Moreover, the distribution of cracks in the simulations has enabled analysing the transition from a brittle to a local damage detected during the experiments. The validated model has been applied to simulate the fracture induced by the Ni re-oxidation. The repartition and density of micro-cracks in the YSZ backbone has been computed during the Ni reoxidation and the results have been confronted with an available tomographic data set. Finally, the impact on the cermet redox tolerance of the Yttrium content in YSZ has been discussed thanks to the modelling. [1] Faes, A. et al. J. Power Sources 193, 55–64. [2] Laurencin, J. et al. J. Power Sources 192, 344–352. [3] Meille, S. Et al . J. Eur. Ceram. Soc. 32, 3959–3967. [4] Miehe, C. et al. Comput. Methods Appl. Mech. Eng. 199, 2765–2778. [5] Leguillon, D. J. Mech. – A/Solids 21, 61–72. Figure 1

International audience; The classical 1D continuum model of oxygen exchange in a porous mixed conductor (bulk path) is presented in terms of Linear Sweep Voltammetry (LSV) and Cyclic Voltammetry (CV). We show that, from a formal point of view, it corresponds to an insertion, diffusion and chemical reaction mechanism. Some general rules for LSV/CV curves are established. Using a dimensionless description, we evidence eight different limiting behaviors of the electrochemical model, which depend on four dimensionless parameters. For an almost full insertion level at equilibrium, these parameters can be reduced to only two, i.e. the dimensionless electrode thickness L and the dimensionless chemical rate constant Lambda. We show that the resulting zone diagram, plotted using the representation log L versus log Lambda, is useful to predict the possible sequences of model behaviors as a function of the potential sweep rate and the electrode thickness. This theoretical analysis is applied to a porous LSCF oxygen electrode under air at high temperature.

International audience; An elementary kinetic model was developed to predict the electrochemical response of porous LSCF and LSCF-CGO electrodes. The model was validated thanks to experiments performed on symmetrical cells using a three-electrode setup. After the model calibration on polarization curves, it has been shown that the model is able to simulate accurately the experimental impedance diagram at OCP and under polarization without additional fitting. Moreover, the evolution of the electrode polarization resistance with the oxygen partial pressure is well reproduced by the model. The electrodes reaction mechanism was thoroughly analyzed and it has been shown that the transition from the bulk path to the surface path depends on the temperature, the polarization and the oxygen partial pressure. The rate-determining steps for the LSCF electrode have been identified at OCP as function of the oxygen partial pressure. Finally, a sensitivity analysis has been performed to study the impact of LSCF demixing on the electrode performances. For a given decomposition, it has been highlighted that the surface passivation would be more impacting than the decrease of the ionic conductivity. Moreover, the impact of the LSCF decomposition would be more detrimental for the electrode performances evaluated in electrolysis mode.

International audience; Three long-term experiments have been performed in SOEC and SOFC modes at different operating temperatures. The durability tests confirm a higher degradation in electrolysis mode with respect to fuel cell operation. In addition, a larger increase of the ohmic resistance is observed for the cell operated at higher temperature in electrolysis mode. The oxygen electrodes of the pristine and tested cells have been characterized by synchrotron X-ray micro-diffraction and micro-fluorescence to assess the relation between the material destabilization and the formation of insulating phases due to interlayer diffusion. The analyses of the pristine cell confirm the presence after the electrode sintering of strontium zirconate and a Gd-rich inter-diffusional layer in the electrolyte just below the zirconates. Moreover, evolutions in the LSCF unit cell volume reveal strontium segregation after aging. The associated material destabilization is linked to the accumulation of SrZrO 3 at the barrier layer-inter-diffusional layer interface in operation and both phenomena are found to be thermally-activated and promoted in electrolysis mode. Finally, the crystallographic evolution of the inter-diffusional layer in electrolysis mode has been investigated by X-ray diffraction. A slight increase of the phase peaks intensity detected at the highest temperature is correlated to the largest formation of SrZrO 3 observed in this condition. Based on these preliminary results, it is proposed that the loss of Zr 4+ from the electrolyte due to the zirconates formation could facilitate the inter-diffusion of Gd, reducing the local ionic conductivity and thus significantly contributing to the largest increase in the ohmic resistance observed in this case.

This work reports a durability study in galvanostatic mode at 700 °C for 900–960 h related to structural and microstructural evolution of LaPrNiO4+δ (LPNO) oxygen electrode. The performances of the electrode, deposited by electrostatic spray deposition (ESD) and screen printing (SP), are not changed when operated under electrolysis current whereas a degradation is observed in fuel cell mode. Small quantities of secondary phases (La,Pr)3Ni2O7-δ in the fine ESD microstructure and Pr6O11 on both ESD and SP layers are detected in the pristine electrode by synchrotron μ-X-ray diffraction and fluorescence. The destabilization of LPNO electrode, which slightly occurs in the ESD layer after thermal aging in air, is strongly accelerated under anodic and cathodic currents. Indeed, the formation of a new phase (La,Pr)4Ni3O10-δ on both ESD and SP layers occurs with the extension of (La,Pr)3Ni2O7-δ in the SP layer, and a significant increase in the amount Pr6O11 at high current density. Nevertheless, the electrode microstructure, determined by FIB-SEM tomography, remains stable under current whereas severe delamination is observed after aging in fuel cell mode. Based on these characterizations, the mechanisms of electrode aging have been discussed to explain the different behavior in fuel cell and electrolysis modes.

Solid Oxide Fuel Cells (SOFCs) are regarded as a promising technology for economic power generation due to their high efficiency and large fuel flexibility. Durability is a severe hurdle towards their deployment. The near future targets in respect to Degradation Rate (DR) are about 0.1% kh−1, which needs improved monitoring and diagnostics. This work aims at introducing a new approach based on Differential Analysis of the i-V curves, named DiVA. It operates with the Differential Resistance Rd and its evolution during long term testing. Two new performance indicators are introduced. Since derivatives are more sensitive to small deviations, the Differential Resistance Analysis (DRA) ensures increased sensitivity and information capability in respect to degradation monitoring and diagnostics, which is demonstrated on a small stack during thermal cycling conditions – before and after the first thermal cycle, on button cells tested up to 9000 h, as well as on button cells operating in fuel cell and in electrolysis mode. The results show that DRA is several times more sensitive in comparison with the classical DR evaluation based on registration of the voltage decrease at constant current.

International audience; An original 3D stochastic model, based on the truncated plurigaussian random fields, has been adapted to simulate the complex microstructure of SOC electrodes. The representativeness of the virtual microstructures has been checked on several synchrotron X-ray and FIB-SEM tomographic reconstructions obtained on typical LSCF, LSC and Ni-YSZ electrodes. The validation step has been carried out by comparing numbers of electrode morphological properties as well as the phase effective diffusivities. This analysis has shown that the synthetic media mimic accurately the complex microstructure of typical SOC electrodes. The model capability to simulate different types of promising electrode architectures has also been investigated. It has been shown that the model is able to generate virtual electrode prepared by infiltration resulting in a uniform and continuous thin layer covering a scaffold. With a local thresholding depending on the position, continuous graded electrodes can be also produced. Finally, the model offers the possibility to introduce different correlation lengths for each phase in order to control the local topology of the interfaces. All these cases illustrate the model flexibility to generate various SOC microstructures. This validated and flexible model can be used for further numerical microstructural optimizations to improve the SOC performances.

Solid Oxide Fuel Cells (SOFCs) are a promising technology that can provide efficient and clean energy production. The general barriers hindering their market entry are durability, i.e. resistance to aging, and costs. In parallel to the deeper insight into the different degradation sources and improved understanding of ageing factors and their interactions, work towards higher accuracy for the assessment and monitoring of real-world fuel cell ageing in necessary. The requirements for operational stability formulate the parameter “degradation rate” (DR). Most often long term durability tests are performed at constant current load and the decrease of the voltage is used for its definition. In this work a new approach based on analysis of the volt-ampere characteristics, named Differential Resistance Analysis (DRA), is presented. It operates with the differential resistance, i.e. with the derivative of the voltage in respect to the current (dU/dI = Rd) which is more sensitive to small deviations and thus increases the sensitivity of the analysis. Two performance indicators are derived (Rd, min and ∆U*) with differing selectivity: ∆U* is more sensitive to activation losses and Rd, min - to transport hindrances. The application of the DRA is demonstrated on examples from measurements in fuel cell and in reverse (fuel cell/electrolyzer) mode, as well as on modeling data. The results show that the method is at least 10 times more sensitive to DR evaluation in comparison with the classical approach.

A set of long-term tests (t ≥ 1000 h) have been carried out in fuel cell and electrolysis modes on typical Ni-YSZ//YSZ//LSCF-CGO cells. The degradation rates were found to be higher in electrolysis than in fuel cell operation. Post-test analyses have revealed that Sr diffusion and formation of SrZrO3 at YSZ/CGO interface occur mainly during electrolysis operation, whereas the process is very limited in fuel cell mode. As a consequence, LSCF destabilization is found to be not involved in the degradation of cell performances during fuel cell operation while it could explain the highest degradation rates recorded in electrolysis mode. An in-house multi-scale model has been used to interpret the role of the cell operating mode on the LSCF demixing mechanism. The simulations have shown that the electrolysis operation leads to a strong depletion of oxygen vacancies in LSCF material (while the fuel cell condition results in an increase in the concentration of oxygen vacancies). It has been proposed that the depletion in oxygen vacancies under electrolysis polarization could drive the Sr release from the structure, and in turn, could explain the experimental results. Based on this proposition, a possible mechanism for the LSCF destabilization and SrZrO3 formation is detailed.

Ceramic high-temperature fuel cells and electrolysers are efficient energy conversion systems for electrical power generation and hydrogen production. Their core components are constituted by a stack of electroactive Solid Oxide Cells (SOCs) in which the electrochemical reactions take place. Thanks to their flexibility, the same stack can be alternatively operated in both fuel cell and electrolysis modes. However, the insufficient durability of SOCs still constitutes a major limitation for the technology. The present study addresses this issue and aims to bring some new insights on the effect of the Solid Oxide Fuel Cell (SOFC) versus Solid Oxide Electrolysis Cell (SOEC) operating modes on degradation of a typical Ni-YSZ//YSZ//CGO//LSCF-CGO cell. The electrochemical degradations are generally attributed to several underlying phenomena such as electrode microstructural evolution, material chemical decomposition or electroactive sites poisoning by contaminants. Among them, it is generally considered that Ni agglomeration in the Ni-YSZ cermet and Lanthanum Strontium Cobalt Ferrite (LSCF) material destabilization are two prevalent mechanisms involved in the cell performance deterioration. Therefore, these two mechanisms have been specifically investigated by a coupled approach of long-term testing in both SOFC and SOEC modes (1000 ≤ t (h) ≤9000) and post-test characterizations. The experimental results have then been analyzed in the frame of an in-house multi-scale model with the purpose to interpret them and to quantify the effect of material ageing on cell performances. The extent of Ni agglomeration has been characterized by three-dimensional electrode reconstructions obtained by X-ray nano-holotomography at European Synchrotron Radiation Facility (ESRF) (on the new Nano-Imaging beamline ID16A-NI). The new set-up and protocol enable the reconstructions of valuable 3D volumes (Fig. 1) with a large field of view (~50 µm) along with a high spatial resolution (~50 nm). The electrode morphological properties, which have been measured on the 3D volumes, have revealed a substantial Ni coarsening over time at 850°C and 750°C, whereas no Ni depletion was detected at the electrolyte interface. The increase of the Ni particle size is found to induce a decrease in both (i) the density of Triple Phase Boundary (TPBs) lines and (ii) the interfacial surface area between Ni and gas. Moreover, it was found that the Ni/YSZ interfacial surface area does not evolve during the experiments. This statement indicates that the ceramic backbone in the cermet prevents a massive Ni agglomeration at the SOFC/SOEC operating temperature. The compilation of all experimental data have allowed fitting the parameters of a physically-based law for Ni coarsening that was introduced in the modelling framework. The simulations have revealed that Ni agglomeration explains around 20-25% of the electrochemical degradation at 850°C after 1000 hrs of operation. However, the electrode microstructural evolution is found not to be affected by the cell polarizations. Therefore, the mechanism cannot explain the higher degradation rates recorded in electrolysis mode compared to the fuel cell ones. To explain the impact of the operating modes (SOFC or SOEC) on the degradation rates, several post-test analyses (i.e. Scanning Electron Microscopy, Transmission Electron Microscopy, X-ray µfluorescence and µdiffraction techniques) have been employed to investigate the phase reactivity in the region of the CGO barrier layer. The characterizations have revealed that Sr diffusion across the barrier layer and formation of SrZrO3 secondary phase occur mainly during electrolysis operation, whereas the process is very limited in fuel cell mode (Fig. 2). As a consequence, LSCF destabilization is found not to be involved in the degradation of cell performances during fuel cell operation while it could explain the highest degradation rates recorded in electrolysis mode. The post-test analyses have also revealed a diffusion and an accumulation of Co in the region of the barrier layer which is concomitant with the formation of SrZrO3. The formation of these Co-rich segregates in contact with SrZrO3 grains have been identified as cobalt-ferrite type compound. The in-house multi-scale model has been used to interpret the role of the cell operating mode on the LSCF destabilization mechanism. The cell polarization curves and the local quantities within the O2 electrode have been computed in both fuel cell and electrolysis modes. The simulations have shown that the electrolysis operation leads to a strong depletion of oxygen vacancies in the LSCF material. It has been proposed that the depletion in oxygen vacancies under electrolysis polarization could drive the Sr release from the structure, and in turn, could explain the experimental results. Based on this proposition, a possible mechanism for the LSCF destabilization and SrZrO3 formation has been detailed. Figure 1

International audience; Two types of solid oxide cells with different Ni-YSZ cermet microstructures have been aged in electrolysis and fuel cell modes for operating times ranging from 1000 to 15000 hours. The pristine and aged cermets have been reconstructed by synchrotron X-ray holotomography. Nickel agglomeration has been observed in the bulk of the operated samples inducing a significant loss of triple phase boundary lengths. The inspection of the microstructural properties has confirmed the stabilizing role of YSZ on Ni coarsening. Furthermore, the gradients of properties quantified at the electrolyte interface have revealed a depletion of Ni only in the electrochemically active region of the electrode. The process is strongly promoted for a coarse cermet microstructure when operated under electrolysis current. The evolution of the microstructural properties has been implemented in an in-house multiscale model. The simulations have shown that the loss of performance is dominated by the depletion of Ni in case of a coarse microstructure. Thanks to the computations, it has been shown that the Ni depletion is controlled by the cathodic overpotential. To explain this dependency, it has been proposed that the accumulation of oxygen vacancies in the double layer could deteriorate the Ni/YSZ interface and trigger the Ni depletion.

Chapter 11: Cell, Stack, and System Modeling and Simulation; International audience; As it plays a key role on the performances and durability of solid oxide cells, the electrode microstructure has to be finely characterized and modelled. In this frame, a plurigaussian random field model and an original sphere packing algorithm have been developed to simulate a large variety of electrode microstructures. The two stochastic geometrical models have been validated on various electrode reconstructions obtained by synchrotron X-ray holotomography including the microstructures of thin functional layers and a thick current collector. Afterwards, semi-analytical microstructural correlations have been proposed and validated on a large dataset of representative synthetic microstructures. These equations allow establishing the links between the microstructural properties controlling the performances (Triple Phase Boundary lengths density and specific surface areas) to the ‘basic’ geometric attributes of the electrode (composition, porosity and particle size distribution). These correlations have been used to model the loss of active TPBl due to the Ni agglomeration.

International audience; Rare earth nickelate, LaPrNiO4+δ (LPNO), is one of the most suitable oxygen electrode materials for Intermediate Temperature Solid Oxide Cells (IT-SOC). By using Electrostatic Spray Deposition (ESD) and Screen Printing (SP), various architectural designs of electrodes are fabricated on Ce0.9Gd0.1O2-δ (GDC) electrolyte. Three electrode configurations are investigated based on a single Active Functional Layer (AFL), a double layer of the same composition which is an AFL topped by a current collecting layer (CCL) and a triple layer which is mainly the double layer with an LPNO/GDC composite interface. For each configuration, the electrode response and properties are analyzed by using Electrochemical Impedance Spectroscopy (EIS), X-ray diffraction and Scanning Electron Microscopy (SEM). The electrochemical properties are discussed as a function of the electrode design, microstructure, and composition with the presence of a higher order nickelate such as (La,Pr)3Ni2O7-δ. The electrochemical performance is enhanced thanks to better adhesion of the electrode onto the electrolyte, the presence of an LPNO/GDC composite interface, a more homogeneous distribution of fine grains, and a flatter electrode surface. A significant decrease in the polarization resistance (Rpol) is measured from 0.72, 0.50, and down to 0.20 Ω cm2 at 600 °C from single to triple layer, respectively.

International audience; The reaction mechanism and the impact of microstructure on performances for a porous LSCF‐CGO composite used as O2 electrode in SOCs have been investigated. For this purpose, an integrated approach coupling (i) electrochemical testing, (ii) advanced 3D characterizations, and (iii) modeling was proposed. In this frame, a symmetrical cell was tested in a three‐electrode setup and the microstructure of the LSCF‐CGO working electrode was reconstructed by FIB‐SEM tomography. The experimental polarization curves and the impedance diagrams along with the extracted microstructural parameters were used to validate an in‐house microscale electrochemical model. This model considers two parallel reaction pathways with (i) an oxidation/reduction at TPBls (surface path) and (ii) an oxygen transfer at the gas/LSCF interface (bulk path). It was shown that the LSCF‐CGO electrode reaction mechanism is mostly controlled by the charge transfer at TPBls whatever the polarization. Finally, the impact of electrode composition, porosity and particles size on the cell polarization resistance was investigated by using the electrochemical model in combination with a large dataset of synthetic microstructures generated by a geometrical stochastic model. The effect of each microstructural parameter on the electrode performance was analyzed in order to provide useful guidelines for the cell manufacturing.

International audience; La0.6Sr0.4Co0.2Fe0.8O3-δ (LSCF) electrode deposited on Gadolinium Doped Ceria (GDC) has been characterized by cyclic voltammetry and impedance spectroscopy in the temperature range of 300 to 700 °C. We demonstrate that the LSCF microstructure has a strong influence on the shape of the voltammograms and on the variation of the series resistance, Rs, measured on the impedance diagrams as a function of the dc bias. LSCF was deposited with two different microstructures, either porous layer (~10 μm thick) by screen-printing (SP) or almost dense layer (~2 μm thick) by Electrostatic Spray Deposition (ESD). For the denser film, one cathodic peak and one reverse anodic peak were evidenced under air at 300 and 500°C. The Rs value was also found to increase with the dc cathodic bias. On the contrary, no peak was observed with the porous film under the same conditions and Rs was independent on the applied potential from 300 to 700°C. Decreasing the oxygen partial pressure allowed the peaks to be evidenced and Rs to vary. The results are discussed in terms of oxygen exchange rate at the LSCF/gas interface, which depends on the LSCF specific surface area and the oxygen partial pressure.

The durability and efficiency of Solid Oxide Fuel Cells or Electrolysis Cells (SOFC/SOEC) depend on the reaction mechanisms taking place within the electrodes. These mechanisms combine a series of elementary steps including mass and charge transfers along with chemical and electrochemical reactions. These processes are still not precisely understood as they involve complex physical phenomena occurring at the atomic scale. Therefore, it is still required to carry out some fundamental studies to better unravel the basic mechanisms involved in the material properties such as the electronic or ionic conductivity, etc. For such basic investigations, the ab-initio modeling is nowadays recognized as a powerful and relevant method. In this context, this work aims to provide deeper insights into the mechanisms at atomic scale by using Density Functional Theory (DFT) calculations. La1-xSrxCoyFe1-yO3-δ (LSCF) and LSCF-based materials are Mixed Ionic-Electronic Conductors, which are commonly used as oxygen electrode for the SOFCs/SOECs [1]. LSCF is a perovskite oxide that contains two transition metals, i.e. cobalt and iron. Owing to the complexity of this system, our work was first devoted to the parent material La1-xSrxFeO3 (LSF), which was here studied by DFT+U calculations. A special focus was paid to finely describe the localization of holes and the formation of oxygen vacancies as a function of the oxygen partial pressure. Electronic structure calculations were carried out at different strontium contents x (x=0.0, x~0.1, x~0.4, x~0.6). Progressive changes during Sr substitution were evaluated in the Density Of State (DOS). The probability densities of the gap-state wave functions were visualized in order to describe the spatial extension of the localized electron holes. The oxygen vacancy concentration, which is related to the oxygen diffusion coefficient, was calculated and coupled to a defect model in order to give at any temperature and partial oxygen pressure the defect concentration. This work will be extended to the typical LSCF material to provide some essential insights into the chemistry of this imperfect solid. [1] S. P. Jiang, International Journal of Hydrogen Energy, 2019, 44, 7448 – 7493

Due to the mixed ionic electronic conductive properties of the Lanthanum Strontium Cobalt Ferrite (LSCF) La0.6Sr0.4Co0.2Fe0.8O3−δ compound, it is of great scientific and technological interest. Especially in the Solid Oxide Fuel Cell (SOFC) technology, this compound has receiving great attention as a cathode material. However, its chemical reactivity with the Yttria-stabilized Zirconia (YSZ) electrolyte still remains one of the main challenges, which demands a comprehension in the μm and sub-μm range. In order to address the reactivity issues locally in the micrometre scale range, 2D and 3D X-ray μ-diffraction and μ-fluorescence analysis have been performed on a pristine LSCF cathode layer. The cathode was deposited on a dense YSZ electrolyte substrate spaced by a thin Gadolinium doped Ceria Oxide (CGO) barrier layer in between LSCF and YSZ to limit the reactivity. The present approach offers a larger field of view in comparison to electron microscopy techniques. The method can provide a more representative information and may offer some insights on the reactivity distribution along the interfaces. The formation of micro SrZrO3 inclusions in LSCF layer is then indubitably identified, as well as in the CGO/YSZ interface.

In this paper the potential of time-of-flight secondary ion mass spectroscopy combined with focused ion beam technology to characterize the composition of a solid oxide fuel cell (SOFC) in three-dimension is demonstrated. The very high sensitivity of this method allows even very small amounts of elements/compounds to be detected and localized. Therefore, interlayer diffusion of elements between porous electrodes and presence of pollutants can be analyzed with a spatial resolution of the order of 100 nm. However, proper element recognition and mass interference still remain important issues. Here, we present a complete elemental analysis of the SOFC as well as techniques that help to validate the reliability of obtained results. A discussion on origins of probable artifacts is provided.

This work aims at investigating through a modeling approach the difference in thermal and electrochemical responses between high temperature H 2 O electrolysis and H 2 O+CO 2 co-electrolysis in solid oxide cells. The study has been conducted by considering a typical planar stack configuration with cathode supported cells. The influence of the local temperature on the polarization curve is discussed. Operating maps are simulated for both electrolysis modes depending on cell voltage and inlet gas flow rate, covering a complete range of gas conversion rates. The optimum domains of operating conditions combining high performances and reasonable temperature elevations are identified. In overall, higher performances are found in steam electrolysis. Indeed, the co-electrolysis process is found to be strongly limited by mass transport through the thick cathode. However, co-electrolysis exhibits an easier thermal management. Finally, the composition of the syngas produced by co-electrolysis is found to be highly flexible through adjustments of the operating parameters.

This paper shows how X-ray computed nanotomography (CNT) can be correlated with focused ion beam time-of-flight secondary ion mass spectrometry (FIB-TOF-SIMS) tomography on the same sample to investigate both the morphological and elemental structure. This methodology is applicable to relatively large specimens with dimensions of several tens of microns whilst maintaining a high spatial resolution of the order of 100 nm. However, combining X-ray CNT and FIB-TOF-SIMS tomography requires innovative sample preparation protocols to allow both experiments to be conducted on exactly the same sample without chemically or structurally modifying the sample between measurements. Moreover, dedicated algorithms have been developed for effective data fusion that is biased with nine degrees of freedom. This methodology has been tested using a porous and heterogeneous solid oxide fuel cell (SOFC) that has features varying in size by three orders of magnitude - from hundreds of nanometre large pores and grains to tens of micron wide functional layers.

The longer lifetimes of solid oxide cells (SOC) – obtained and still required, both in fuel cell and electrolysis operation – exacerbate the need for accelerated stress tests, since neither reliable degradation models are currently available nor validated methodologies to evaluate expected stack durability in a practical timeframe. The project AD ASTRA ( “HArnessing Degradation mechanisms to prescribe Accelerated Stress Tests for the Realization of SOC lifetime prediction Algorithms”) started operations to tackle this on January 1st 2019, funded by the European Fuel Cells and Hydrogen Joint Undertaking. Generally, a great number of operational parameters influence SOC degradation during lifetime, each contributing according to characteristic times and intensities, often in convoluted or even contrasting fashion. Aggravation of test input parameters can thus lead to non-representative accelerated degradation of a component where concurring effects are differently influenced, leading to a biased end state of the component investigate. To systematically address, harness and accelerate the selected failure modes, a matrix was set up of degradation mechanisms, test items, test procedures and characterization methods, that was translated to protocols in such a way that consistent measurement data are generated, helping subsequent analysis and correlation of phenomena and parameters by the project’s modelling and simulation activities. The testing approach consists of a dual-focus campaign targeting: • macroscopic stack testing procedures, where short stacks will be operated in situ, in aggravated conditions with the objective of identifying test settings that will stress the stack in a representative way but minimizing testing time and resource expenditure • specific component ageing tests, to reproduce (ex-situ) the degraded condition of critical stack components or interfaces in a faster time, based on analysis of the components extracted from stacks that have been field-tested. A radical innovation of the project consists in the integration of these two methods: components specifically aged ex situ assembled into otherwise “new” stacks for in-situ testing, so that their effect on stack performance can be easily isolated from other components that in real-world conditions would also degrade. These tests allow to observe from the start (t=0) the end-of-service stack degradation, radically reducing in-situ testing time and providing powerful means to validate models simulating the long-term effects of component degradation in a stack.

In order to finally and systematically address the growing need for accelerated stress tests, given the longer lifetimes of solid oxide cells – both in fuel cell and electrolysis operation – the Fuel Cells and Hydrogen Joint Undertaking has launched an international initiative to overcome this epic challenge. The overall objective of the project that was awarded the task is the development of protocols that allow quantitative identification and prediction of critical degradation mechanisms, correlating them with overall performance variables in selected stack components (fuel electrode, oxygen electrode and interconnect). These will build firstly on the analysis of numerous field-tested samples of SOC stacks provided by the industrial partners, followed by applying existing and developing improved testing and modelling methods based on ex-situ component ageing and aggravated stack testing.

To manage the grid with renewable electrical sources and ensure their storage, hydrogen chains have to be considered. However, hydrogen must be pressurized before being used or stored. The first levels of compression are less energetic for liquids than for the hydrogen gas, therefore the electrolysis of steam directly under pressure has to be studied. On one hand, with the solid oxide technology, the Open Circuit Voltage (OCV) presents the drawback of increasing with pressure. On the other hand, the pressure might improve the reactive species supply to the electrochemical sites and reduce the related losses. Thus, what are the electrolysis performances under pressure? And is there an energetic gain at the system level to produce pressurized hydrogen? In this context, experimental and modelling approaches have been adopted in order to understand the basic underlying mechanisms of pressurized electrolysis operation. In this frame, SOEC performance tests were realized under pressure in the range of 1 to 10 bars on two types of single commercial cells. They were based on conventional materials and differentiated by their microstructure. Concerning the set-up, the test platform requirements were achieved by developing an innovative mechanical design associated to a specific steam management. Moreover, a previously proposed electrochemical model of solid oxide electrolysis cell was adapted to take into account the pressure effects. The model allows computing all the most important operation parameters such as local current densities, local molar fraction, overpotentials and also the cell microstructural parameters [1]. According to the complete I-V curves from 0 to -1.3 A/cm2 at 800°C with high steam conversion, it has been observed that OCVs respect the Nernst relation under pressure while the limit current appears at higher current values. This work shows that the hydrogen production can be increased thanks to a 10 bars operation at the autothermal voltage. Moreover, these trends have been analyzed with the model. The impact of pressure on the electrochemical performances is much more important in the range of 1 to 5 bars than above. In our experimental conditions, the diffusion improvement in the porous cathode alone can explain this behavior. Finally, the impact of each microstructural parameter on the cell performances with pressure is discussed. The research leading to these results has received funding from the European Union's Seventh Framework Programme (FP7/2007-2013) Fuel Cells and Hydrogen Joint Undertaking (FCH-JU-2013-1) under grant agreement 621173. [1] J. Laurencin, D. Kane, G. Delette, J. Deseure, F. Lefebvre-Joud, Modelling of solid oxide steam electrolyser : impact of the operating conditions on hydrogen production, 2011, Journal of Power Sources, 196, 2080-2093. Figure 1

This work highlights an experimental and modeling approach devoted to a better understanding of H2O and CO2 co-electrolysis mechanisms at 800 °C. A standard Cathode Supported Cell (CSC) was used in this study. Through numerical adjustments on experimental polarization curves, the cathode microstructural parameters and exchange current densities for H2O and CO2 reductions were determined and subsequently implemented in an in-house co-electrolysis model. Additionally, micro gas chromatography (μGC) analyses were performed in co-electrolysis operating mode for different cell polarizations (from i = 0 to i = −1.75 A cm−2). μGC analyses at Open Circuit Voltage (OCV) were used to validate the kinetic constants of the Water Gas Shift (WGS) reaction implemented in the model. Predictions of both co-electrolysis polarization curves and outlet gas compositions were then compared to the experimental measurements. The good agreement between simulated and experimental data proves the relevance of the macroscopic representation of electrochemical processes through a “surface ratio” that takes into account the H2O and CO2 electrolyzes competition. A sensitivity analysis was performed to ensure a better understanding of co-electrolysis mechanisms and further investigate the influence of the reverse WGS reaction over CO production.

International audience; The relevance of micro Laue diffraction tomography (mu-LT) to investigate heterogeneous polycrystalline materials has been studied. For this purpose, a multiphase solid oxide fuel cell (SOFC) electrode composite made of yttria-stabilized zirconia and nickel oxide phases, with grains of about a few micrometres in size, has been analyzed. In order to calibrate the Laue data and to test the technique's sensitivity limits, a monocrystalline germanium sample of about 8 x 4 mu m in cross-section size has also been studied through mu-LT. The SOFC and germanium Laue diffraction pattern analyses are compared and discussed. The indexing procedure has been successfully applied for the analysis of the germanium Laue data, and the depth-resolved two-dimensional cartographies of the full deviatoric strain tensor components were obtained. The development and application of an original geometrical approach to analyze the SOFC Laue data allowed the authors to resolve grains with sizes of about 3 mu m and to identify their individual Laue patterns; by indexing those Laue patterns, the crystalline phases and orientations of most of the grains identified through the geometrical approach could be resolved

International audience; Long-term experiments have been carried out to investigate the impact of Nickel (Ni) coarsening on the performance of Solid Oxide Cell. Durability tests have been performed with H2 electrode supported cells at 850 °C and 750 °C in fuel cell and electrolysis modes. Microstructural changes in the composite electrode of Nickel and Yttria Stabilized Zirconia (YSZ) have been characterized by synchrotron X-ray nanotomography. Analysis of the reconstructions have revealed that Ni coarsening induces a significant decrease of both the density of Triple Phase Boundary lengths (TPBls) and the Ni/gas specific surface area. However, the contact surface between Ni and YSZ is not changed upon operation, meaning the Ni sintering is inhibited by the YSZ backbone. Moreover, the Ni coarsening rate is independent of the electrode polarization. The evolution of TPBls in operation has been fitted by a phenomenological law implemented in an electrochemical model. Simulations have shown that microstructural changes in the H2 electrode explain 30% of the total degradation in fuel cell mode and 25% in electrolysis mode at 850 °C after 1000-2000 h. Moreover, it has been highlighted that the temperature at which the degradation is estimated after the durability experiment plays a major role on the result.

Microstructural evolution of a Solid Oxide Electrolyser Cell (SOEC) Ni–YSZ cermet cathode is investigated using three dimensional electrode characterisations. 3D reconstructions are obtained on a reference and two long-term tested cells, which were maintained at −0.5 and −0.8 A cm−2 for 1000 h at 800 °C. During the long term tests, air was fed at the anode and a mixture of 10% H2–90% H2O was fed at the cathode. In this framework, reconstructions have been obtained from synchrotron X-ray nano-tomography technique. Microstructural properties extracted from the 3D reconstructions exhibit an evolution during the tests. Triple Phase Boundary length is decreasing from 10.49 ± 1.18 μm−2 for the reference cell to 6.18 ± 0.6 μm−2 for the long term tested cell at −0.8 A cm−2. Evolutions of morphological parameters were introduced in an in-house multi-scale model to evaluate their impacts on the electrode degradation, and hence, on the global SOEC performance.

This work details an in-house 2D electrochemical and thermal model of solid oxide electrolysis cell (SOEC) dedicated to the simultaneous electrolysis of steam and carbon dioxide. The apparent exchange current densities for both H2O and CO2 electrochemical reductions have been fitted on single electrolysis polarization curves obtained on a standard cathode supported cell (CSC). By using these values, predictive simulations have been performed in a co-electrolysis mode. The results have been compared to experimental data obtained in the same conditions. A good agreement is found between the simulated points and the experimental polarization curves, meaning that the model stands to predict accurately the SOEC response in co-electrolysis mode. Simulations have been performed to investigate co-electrolysis mechanisms in terms of partial pressures, current densities, and overpotentials distributions along the cell. The relative influence of both water gas shift (WGS) reaction and CO2 electrolysis on the global CO production is also discussed.

SOFC electrochemical performances are strongly dependent on the micro and nanostructure of the electrodes. 3D image analysis is an essential tool to characterise these microstructures. However, the studied volumes have to be sufficiently large to be representative of the heterogenous medium. In this study, quantitative phase contrast X-ray nano-holotomography is performed on large volumes of SOFC anode consisting by a Ni/YSZ cermet. Both the electrode functional layer and the current collector of anode supported cell are characterised in 3D. The setup and data processing adaptation needed for such an absorbing materials are described. The X-ray nano-holotomography was performed with an isotropic voxel size of 25 × 25 × 25 nm3 and a field of view of about 50 μm. This technique is non-destructive and was used in local tomography mode. After filtering and thresholding, the segmentation of the 3D reconstructions allows precisely separating the network of each phase of the electrode (metallic, ceramic and porosity), giving representative features of the microstructure.

Microstructural evolution of a Solid Oxide Electrolyser Cell (SOEC) La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) anode is investigated using 3D reconstruction techniques both on a reference and a long-term tested cell. In this framework, reconstructions have been obtained from focused ion beam-scanning electron microscopy and synchrotron X-ray nano-tomography techniques. Microstructural properties extracted from the 3D reconstructions exhibit a slight evolution during the test. Synchrotron X-ray Absorption Spectroscopy (XAS) was also performed, and a modification of the cobalt oxidation state is found. LSCF conductivity decrease due to the change in Co mixed valence has been discussed. Evolutions of both morphological parameters and electrical conductivity were introduced in an in-house micro model to evaluate their impacts on the electrode degradation.

Thermo-mechanical properties of SOFC/SOEC electrodes have been computed by homogenization from three-dimensional reconstructions of their microstructure. The support and the functional layer of a Ni/YSZ supported cell and the oxygen electrode made of La 0.6 Sr 0.4 Co 0.2 Fe 0.8 O 3−δ (LSCF) have been studied. Values of the effective Young's modulus obtained for the thick porous Ni/YSZ support are in good agreement with the experimental data ( E = 35 GPa). Besides, for thin functional layers, the methodology supplies original values that are difficult to obtain by conventional experimental methods ( E = 105 GPa for the Ni/YSZ Functional Layer and E = 55 GPa for the LSCF electrode). Furthermore, it has been shown that the effective elastic parameters are influenced by the following morphological parameters: the porosity and the formation factor that could also be related to the manufacturing process. Theses morphological parameters have however negligible effect on the effective thermal expansion.

An in-house micro model has been built to describe the electrochemical mechanisms governing both H2 and O2 electrodes operating in SOEC mode. A special attention has been paid to take into account the microstructure properties of the ionic, electronic and gas phases as well as the processes occurring therein. A commercial LSM–YSZ symmetrical cell has been tested at 700, 750 and 850 °C in air. Simulations have been carried out to interpret the experimental data. It is suggested that the kinetic of O2 formation is controlled by a single charge transfer. A sensitivity analysis has been performed using the micro model to quantify the role of the microstructure in the electrode behaviour. Transport of oxygen ions in the functional layer has a strong impact on the cell response since it governs the delocalization of the electrochemical reactions. The density of TPB length is also a key parameter controlling the electrode efficiency. Evolutions of the microstructural parameters in operation have been associated to the degradation of the electrochemical performances. The decrease in TPB length due to Ni agglomeration has a moderate impact whereas the decrease in ionic conductivities of 8YSZ or LSFC could explain a large amount of the cell degradation.

Tomography is a standard and invaluable technique that covers a large range of length scales. It gives access to the inner morphology of specimens and to the three-dimensional (3D) distribution of physical quantities such as elemental composition, crystalline phases, oxidation state, or strain. These data are necessary to determine the effective properties of investigated heterogeneous media. However, each tomographic technique relies on severe sampling conditions and physical principles that require the sample to be adequately shaped. For that purpose, a wide range of sample preparation techniques is used, including mechanical machining, polishing, sawing, ion milling, or chemical techniques. Here, we focus on the basics of tomography that justify such advanced sample preparation, before reviewing and illustrating the main techniques. Performances and limits are highlighted, and we identify the best preparation technique for a particular tomographic scale and application. The targeted tomography techniques include hard X-ray micro- and nanotomography, electron nanotomography, and atom probe tomography. The article mainly focuses on hard condensed matter, including porous materials, alloys, and microelectronics applications, but also includes, to a lesser extent, biological considerations.

International audience; Experiments have been performed in pressurized co-electrolysis mode at 800 °C on a typical Ni-YSZ//YSZ//CGO-LSCF cell. The polarization curves and the composition of the produced syngas have been measured at 1 bar and 10 bar. It has been found that the cell performances are improved under pressure at 1.3 V. The gas analyses have revealed that the methane formation is only activated under polarization and pressure. These experimental results have been used to validate a model which encompasses a chemical and electrochemical description of the co-electrolyser combined with a mass transport module. It has been found that the model is able to predict accurately the polarization curves as well as the syngas compositions at the cell outlet. Once validated, the model has been used to analyze the operating mechanisms in pressurized co-electrolysis. The impact of pressure on the mass transfer, the electrochemical and chemical reactions has been discussed. The close interaction between the electrochemical and chemical reactions for the internal production of CH4 has been specifically highlighted. Finally, operating maps have been computed at 10 bar from 700 °C to 800 °C. These simulations have shown that formation of CH4 in the co-electrolyser remains limited at 700 °C.