A SOFC working in an environmental transmission electron microscope - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

A SOFC working in an environmental transmission electron microscope

Q. Jeangros
  • Fonction : Auteur
M. Bugnet
  • Fonction : Auteur
C. Frantz
  • Fonction : Auteur
S. Diethelm
  • Fonction : Auteur
D. Montarino
  • Fonction : Auteur
E. Tyukalova
  • Fonction : Auteur
Y. Pivak
  • Fonction : Auteur
J. van Herle
  • Fonction : Orateur
A. Hessler-Wyser
  • Fonction : Auteur
M. Duchamp
  • Fonction : Auteur

Résumé

Owing to the recent advances in in situ and operando transmission electron microscopy (TEM), we have set up an experiment to operate a SOFC (solid oxide fuel cell) inside an environmental TEM to identify how the device microstructure controls its electrical behavior. An elementary anode-electrolyte-cathode sandwich was prepared by focused ion beam (FIB), and mounted on a heating and biasing microelectromechanical systems (MEMS) based specimen holder (DENSsolutions) inserted a FEI Titan ETEM. The sample is made of a standard SOFC: the cathode is based on Sr-doped La manganite (LSM), the electrolyte is yttria-stabilized zirconia (YSZ) and the anode is a NiO cermet; both electrodes were co-sintered with YSZ. For practical reasons, we have used a single-chamber configuration where the anode and cathode were exposed simultaneously to the oxidant (O2) and reducing (N2:H2 in the ratio 20:1) gases appropriately mixed during cycles at about 15 mbar up to 750°C. Thanks to a difference in the catalytic activity between the electrodes, O2 should reduce at the cathode, while H2 should oxidize at the anode. A transient regime could be obtained where a voltage difference was detected by the biasing system of the holder. The variation of current was correlated to the evolution of the gas fuel composition and the anode microstructure. The latter was followed by conventional and high-resolution imaging, diffraction work and EELS (electron energy-loss Spectroscopy) [1]. [1] The authors acknowledge METSA (www.metsa.fr), CLYM (www.clym.fr), FACTS and the start-up grant M4081924 at NTU, the INSTANT France-Singapore MERLION. Introduction As for many technologies, the macroscale performance of SOFC devices is often seen to decrease without a clear understanding of the microstructural origins of these losses. In the end, metrology limitations often prevent the clear identification of performance-loss processes, which may range in SOFC from the coarsening and redistribution of the nickel catalyst on the anode side (e.g., due to reduction-reoxidation cycles), cracks that may form due to thermomechanical effects, the poisoning of electrochemically active sites by impurities, the diffusion of species across interfaces, etc. This lack of understanding then hinders the development of device stabilization strategies. 1. Scientific Approach We present here a new methodology to establish structure-property links in a SOFC. We analyze the cathode-electrolyte-anode of a SOFC in situ and down to the atomic scale using environmental transmission electron microscopy (ETEM). More specifically, a thin lamella containing all the materials of interest is electrically connected to a source meter, exposed to both reducing and oxidizing gas atmospheres, and heated to 600 °C to evaluate how the open circuit voltage of the SOFC evolves with its microstructure in a single-chamber configuration. With this approach, we correlate the evolution of the open circuit voltage of the device to the O2-to-H2 ratio in the cell environment, and the chemical and structural state of its constituent materials. 2. Experiments We investigated the following cell architecture in the ETEM: a LSM/YSZ cathode, a 2-µm-thick YSZ electrolyte, and a NiO/YSZ anode precursor. A TEM lamella featuring the full cathode-electrolyte-anode cell was extracted and contacted to a MEMS heating and biasing chip using a FIB/scanning electron microscopy (SEM) workstation, as shown in Fig. 1a. A STEM high-angle annular dark-field image (HAADF) image of the as-prepared thin lamella is shown in Fig. 1b. Elemental maps obtained by STEM energy dispersive X-ray spectroscopy (EDX) are displayed in Fig. 1c, highlighting the different phases present in the initial stack. The YSZ electrolyte is dense with grains of about 1 µm, while the LSM/YSZ layer is porous to ensure a gas permeation to the triple-phase boundaries (LSM-YSZ-porosity on the cathode side). On the other hand, the NiO/YSZ anode precursor side is dense in its as-sintered state. The as-sintered NiO phase was first reduced to Ni. For that purpose, 10 to 15 mbar of H2 was introduced in the ETEM, a pressure close to the maximal pressure allowed by the ETEM chamber. The temperature was raised up to 750ºC to trigger the reduction of NiO to Ni. To then simulate single-chamber operation conditions (no separation between H2 and O2 is possible in the ETEM), the temperature was lowered to 600ºC (to reduce thermomechanical stresses on the thin lamella). The forming gas flow was then set to 3 ml/min, before introducing an additional flow of O2 of ~0.1 ml/min, leading to an increasing O2-to-H2 ratio in the ETEM, a flow that was then stopped at some point during the experiments to return to a reducing atmosphere. We then analyzed the variation in voltage between the cathode and the anode as a function of this varying O2-to-H2 ratio and the chemistry and microstructure of the materials, focusing on the Ni catalyst on the anode side. The O2 to H2 ratio was measured using the currents measured by a residual gas analyzer (RGA), which is located at the exhaust of the ETEM chamber. To account for the time needed for the gas to flow from the sample location to the exhaust, a time of 3 minutes was subtracted to the time tag of the RGA to allow the comparison of microstructural and voltage data to RGA currents. 3. Results As observed in previous ETEM studies on NiO/YSZ anodes [1], the reduction reaction of NiO becomes visible through the creation of pores within the grains, with pores forming preferentially at the interfaces with YSZ due to a coarsening of the Ni phase at these temperatures. As also observed in previous experiments [2], [3], the NiO reduction kinetics inside the ETEM is slow. The reaction rate is first controlled by the nucleation of the first Ni seeds. The presence of H2O around the reaction sites then likely slows the reaction at high NiO-to-Ni conversion fractions. We then investigated the impact of a varying O2-to-H2 ratio and monitored the cell voltage in relation to the morphology and chemistry of the Ni catalyst. Fig. 2 shows a typical example of the data retrieved from the ETEM experiments. Fig. 2a shows a comparison of the evolution of the intensity measured on the CCD camera (integrated over the area shown in Fig. 2b, the surface of an initially metallic Ni grain) and the voltage measured between the two electrodes as a function of the H2-to-O2 ratio (from the RGA O2 current as the flow of H2 remained constant, with and without the time correction). A first small yet measurable voltage increase (in the sub-mV range) is observed after 380 s of experiment, which coincides with the presence of both O2 and H2 in the ETEM chamber and with Ni still in its metallic state (Fig. 3c). Indeed, the dense Ni grain morphology remains identical between Fig. 3b and c. As the O2-to-H2 ratio increases further after 400 s of experiment, a NiO scale starts to form on the metallic Ni grain (Fig. 3d). The Ni grain is now covered by a NiO scale that expands outwards [3]. The TEM image intensity within the region that was previously a void now decreases as NiO formed there (arrowhead in Fig. 3d). This oxidation coincides with the decrease of the voltage measured between the two electrodes. From this first ramp in O2-to-H2 ratio, it hence appears that the small increase in voltage correlates with the presence of metallic Ni on the anode side. Once the O2 flow is stopped and the O2-to-H2 ratio starts to decrease after 1500 s, the intensity at the location of the NiO scale starts to decrease further (see arrow(heads) in Fig. 3a and e). From high-resolution TEM images, this loss in intensity results from the growth of new Ni domains directly on the NiO scale as the partial pressure of H2 increases. The voltage increases in these conditions, which would be consistent with the presence of Ni on the outer surface of the NiO scale to enable the oxidation of H2. Furthermore, Ni L3/L2 EELS data is consistent with such a mixed NiO/Ni system: an intermediate L3/L2 ratio is measured in these conditions [2], [3]. After 2200 s, the NiO scale disappears, and the Ni islands present on the scale surface merge with the Ni grain (see arrowheads in Fig. 3f). In Fig. 3g, the NiO scale has completely disappeared, and Ni is back to its metallic state. The voltage starts to decrease as O2 is being removed from the ETEM chamber. The second voltage increase from ~1500 s to ~2300 s is broader than the first one. Overall, these oxidation-reduction cycles were repeated multiple times and similar trends were observed: at intermediate O2-to-H2 ratios and with Ni in its metallic state: a small yet distinct open circuit voltage formed between the two electrodes. This voltage difference is indicative of a difference in O2 partial pressure across the cell, which may indicate that the cell is electrochemically active (in open circuit conditions). Furthermore, ex situ experiments performed on the same materials but with a bulk sample confirmed this correlation between O2-to-H2 ratio and open circuit voltage of the cell. The voltage gain measured in the ETEM (0.1 mV) is however several orders of magnitude lower than that measured ex situ (0.8 V). This difference likely stems from the geometry of the TEM lamella (µm3 compared to mm3 for a bulk sample). The TEM sample features only few electrochemically active sites, hence giving rise to small gradient in O2 partial pressure across the cell. The exact role of the LSM cathode on the voltage is now being investigated. Its microstructure did not change in the conditions probed in the ETEM. To the best of our knowledge, these experiments are the first to capture in situ the microstructural evolution of a SOFC in direct relation to its electrical properties (in open circuit conditions). Our results reveal several insights regarding the impact of the morphology and oxidation state of the anode Ni catalyst on the electrical properties of the device. While the methodology should be refined, this work highlights the potential of in situ ETEM to provide structural information at the nanoscale for a wide range of solid-state energy conversion systems, such as batteries, supercapacitors, etc. References [1]Q. Jeangros et al., “In situ redox cycle of a nickel-YSZ fuel cell anode in an environmental transmission electron microscope,” Acta Mater., vol. 58, no. 14, pp. 4578–4589, 2010, doi: 10.1016/j.actamat.2010.04.019. [2]Q. Jeangros et al., “Reduction of nickel oxide particles by hydrogen studied in an environmental TEM,” J. Mater. Sci., vol. 48, no. 7, pp. 2893–2907, Apr. 2013, doi: 10.1007/s10853-012-7001-2. [3]Q. Jeangros et al., “Oxidation mechanism of nickel particles studied in an environmental transmission electron microscope,” Acta Mater., vol. 67, pp. 362–372, 2014, doi: 10.1016/j.actamat.2013.12.035. Fig. 1: a) SEM image of a biasing and annealing MEMS chip for in situ TEM. The anode and cathode of the SOFC lamella are electrically connected to the biasing electrodes of the MEMS. b) STEM HAADF micrograph of the electrically connected SOFC sample, and c) corresponding STEM EDX maps of the initial SOFC device acquired from the dashed area shown in b). Fig. 2: TEM image intensity integrated in the region dashed in b), the RGA O2 current (raw data, full line, and shifted forward by 180 seconds, dashed line), and voltage measured between the anode and cathode (raw and filtered data). b-f) Selection of TEM images of the edge of a Ni grain at the critical steps of the reoxidation and reduction process. Arrowheads highlight key morphological changes occurring at the surface of the Ni grain, as discussed in the text.
Fichier non déposé

Dates et versions

hal-04287868 , version 1 (15-11-2023)

Identifiants

  • HAL Id : hal-04287868 , version 1

Citer

Q. Jeangros, M. Bugnet, T. Epicier, C. Frantz, S. Diethelm, et al.. A SOFC working in an environmental transmission electron microscope. EFCF 2022: 15th European SOFC & SOE Forum, Jul 2022, Lucerne, Switzerland. ⟨hal-04287868⟩
6 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More