Search

The development routes to improve the performances of SOFCs are globally well identified and lead to shift the characteristic i/v curve towards higher potentials and current densities, an effect that is consistent with a lowering of the three overpotentials (activation, resistance and concentration). An increase of the Triple Phase Boundary (TPB) density at the electrodes, a reduction of the materials thickness, the development of materials with improved performances, the optimization of electrodes microstructure all serve as guidelines that aim to this goal. Taking as an example the specific design of battery electrodes, we have explored how the corrugation of electrode/electrolyte interfaces at a mesoscopic scale in a SOFC could improve the electrochemical performances, from theoretical and experimental points of view. First, an electrochemical model was established and implemented in the COMSOL Multiphysics software, taking into account masses and charges conservation, gas transport and electrochemical reaction kinetics. The model results demonstrated that the presence of a periodic pattern (parallelepipeds, pyramids, ellipsoids) at the electrolyte/electrode interfaces, along with an electrolyte the thickness of which is significantly smaller than the dimensions of the pattern, could lead to a strong increase of the exchange surface, hence to higher exchange currents and cell performances. To that extent, it is theoretically shown that the patterns must have concave and convex singularities so as to confine the cathode material on the cathode side and the anode material on the anode side, hence the active layers on both sides, so that a much larger number of TPB are solicited and involved in the chemical reactions, reducing then the activation overpotential, as compared with a flat surface. It is also shown that the geometrical feature can be chosen so that it minimizes the concentration overpotential. In addition, the modeling provided us with key information regarding the characteristics of the patterns (shape, width and depth, distance between them, …) that should lead to a large gain in exchange current throughout the interface. For example, an increase of the exchange current of about 60% was calculated for a parallelepipedic pattern with 70 μm width and 100 μm depth, each geometric feature separated from the other by a few hundred microns. With the use of laboratory standard ceramic processes (tape casting, screen printing, bar coating, cold pressing, cold stamping), we have implemented such mesoscopic architectures on green self supported anodes (YSZ + Ni) on top of which a thin layer of electrolyte (YSZ) was deposited. As anticipated the electrical testing and impedance spectroscopy results show that such corrugation at anode/electrolyte interfaces improves significantly the electrochemical performances of the cell even though the sintering tends to alter and blunt the geometrical features of the pattern. Based on the electrochemical performances and modeling, the geometry of the pattern and its evolution during sintering is discussed in terms of activation and concentration overpotentials.

A new concept of a high temperature fuel cell based on a porous dual-conducting (H+ and O2-) central membrane was successfully developed as part of a European project (“IDEAL-Cell”, FET-Energy/FP7, 2008-2011). This additional layer ensures the connectivity between both SOFC cathode/electrolyte and PCFC electrolyte/anode “water-free” compartments so that the system operates with three independent chambers. During the development of this concept, it was discovered that BCY15, used so far as a protonic conductor, exhibited a very significant anion conduction at operating temperature under oxygen atmosphere, opening up the route for a simplified and more efficient configuration; since then, BCY15 is used in all the constituent layers to lead to the so-called monolithic concept. Therefore the porous dual-conducting central membrane is made of a single BCY15 phase instead of a composite mixture of a proton conductor (i.e. BCY15) and an anion conductor (i.e. YDC15). This CM is thus sandwiched between 2 BCY15 electrolytes with a BCY15 + Ni anode on one side and a LSCF48 + BCY15 cathode on the other side. Recent works proved that electrical performances of the monolithic configuration at 600°C are higher than those measured for standard SOFCs and PCFCs at equivalent geometry. The present work aims at presenting the first results of the manufacturing and electrochemical testing of a two-cells short stack specifically designed with 3 independent chambers in which central membrane (CM) supported monolithic cells are integrated. The supported-CMs were fabricated by using BCY15 and a polymer PMMA as a pore former agent. The CMs were sintered at 1400°C for 5 hours under static air leading to a porosity content of around 50 ± 3 vol. %. The electrolytes were then deposited onto the CMs by tape casting from an ethanol-based suspension containing BCY15 and 1 wt. % of ZnO, and were sintered at 1350°C for 3 hours. Anode ink containing 60 vol. % BCY15, 40 vol. % NiO and 10 wt. % of graphite was deposited on one electrolyte and sintered at 1350°C for 3 hours. Cathode ink containing 50 vol. % BCY15, 50 vol. % LSCF48 and 10 wt. % of graphite was deposited on the other electrolyte and sintered at 1100°C for 3 hours to avoid delamination and creeping. This processing sequence leads to flat cells having a total thickness lower than 1 mm. 2 cells were integrated in a specifically designed short stack made of INOX 440C, and sealed with Thermoculite 866 LS. The electrochemical performances of the short-stack have been measured in real operating conditions via a dedicated 3-chamber set-up named Real Life Tester (RLT), and compared to those of single cells operated under the same conditions. Results are discussed in terms of cells geometry and microstructure, and stack design improvements are proposed.

In the frame of a FP7 project (FET-Energy “IDEAL-Cell”, 2008-2011), we have developed with a European consortium a new type of high temperature fuel cell based on a dual membrane with mixed proton and oxygen ions conductivity. This dual conductivity porous membrane, made of a proton conducting phase (BCY15) and of an oxygen conducting phase (YDC15) was sandwiched between a dense BCY15 electrolyte and a dense YDC15 second electrolyte, forming a tri-layer itself sandwiched by a cathode (LSCF48) and an anode (BCY15 + Ni). In this configuration, water is formed neither at the cathode nor at the anode, minimizing electrodes concentration overpotentials, but rather within the dual membrane which interconnected porosity ensures water evacuation. The IDEAL-Cell project i) proved the concept, ii) showed that this new fuel cell was performing better at 750°C than PCFCs and SOFCs having equivalent thickness, iii) showed that the concept was fully reversible with a high dynamic when shifting from the fuel cell regime to the electrolyzer regime, iv) showed that BCY15 was also an excellent oxygen ions conductor when fed with oxygen, and then demonstrated that the cell could be fabricated by using solely BCY15 (replacing YDC15 in the dual membrane and for the oxygen electrolyte), leading to a drastic simplification of the concept, hence of the shaping process. Today, this simplified cell is fabricated at the laboratory scale by a sequence of successive steps (1/ cold pressing, rolling and sintering of the porous dual membrane at 1350°C, 2/ deposition of both electrolytes by dip coating, and sintering at 1350°C, 3/ deposition of both cathode and anode by bar-coating, and sintering at 1150°C), which is hardly cost effective in view of stacking and further development. The present work proposes to demonstrate that co-sintering of the whole cell in a single step is possible, and to determine under what conditions of geometry, starting materials and sintering cycle a flat and stress-free cell in view of stacking, efficiency and durability can be produced via this simplified process. The research described in this paper was based on the thermomechanical modelling of the multilayer deformation occurring during sintering via a finite element numerical simulation, in which debinding, elastic and irreversible deformations, kinetics of grain growth and pores shrinkage are integrated. The thermomechanical parameters were obtained on a differential CTE measurement set-up, and the microstructural ones by image analysis on SEM images.

The EVOLVE cell, named after a running FP7 project, is an innovative SOFC concept that integrates advanced materials, providing multiple functionalities (electrochemical activity, ion or/and electron conduction, gas diffusion…), and an inventive anode current collector made of a NiCrAl preoxidized foam impregnated with a conductive perovskite. Therefore, this anode combines the beneficial characteristics of MSC and ASC technologies, and strong improvements in terms of reliability and durability (inter-diffusion, Ni coarsening and agglomeration, sulphur tolerance, mechanical robustness and chemical stability under redox cycling in temperature) are expected. In order to enhance the electrode efficiency, the 3D microstructure needs to be precisely described at the microscopic scale by a morphology model. An optimized microstructure (controlled pores size, morphology and connectivity, grains size distribution, 3D tortuosity and 3D percolation of phases, volume fraction of phases, final thickness of components …) is closely related to the starting parameters (nature and quantity of starting powder, binder, shaping aids…) and to the control of the shaping process parameters (sintering treatment and atmosphere). It should meet the best compromise between a good electrocatalytic activity and a low ohmic resistivity while ensuring a stisfactory long term thermomechanical stability on the life time of the device. Based on a mathematical morphology approach applied on symmetrical LST/CGO anode layers, the present work aims at showing how a morphology model can be establised and 3D microstructural data relevant for shaping and performances can be derived from back-scattered electrons (BSE) SEM observations. For each sample, a series of images were used as input information for the microstructural modelling. All the original images were first filtered to remove the noise, thresholded and finally a series of morphological operators (openings, surface openings, closings and reconstruction with markers) were applied to remove artefacts. A 3D pluri-Gaussian model has been generated from 2D images, and then computed to predict volume fraction of the 3 phases, 3D tortuosity, gas permeation and ionic conductivities. These results were used for the optimization of the shaping process that includes the integration of the metal foam.