The degradation of Pt in the cathode catalyst layer remains a formidable challenge for the longevity of polymer electrolyte fuel cells. In recent years, high surface area carbons have emerged as promising support materials for Pt in order to yield long lifetimes. Existing degradation models for catalyst layers usually assume homogeneity of composition, microstructure and conditions prevailing in catalyst layers that result in uniformly distributed rates of degradation processes. Our work presents an extension of a well-established physical-statistical model that includes dissolution, redeposition, coagulation, and detachment. For the first time, we examine the impact of structural heterogeneity as well as the role of local structural reconfiguration during operation on performance. Pt nanoparticles in proximity to ionomer are exposed to a significantly higher proton concentration and are therefore expected to experience higher rates of dissolution than particles not in proximity to ionomer, suggesting that two populations of nanoparticles should be distinguished. This distinction of two separate populations is implemented into our modified model. Over time, the microstructure surrounding these particles reconfigures, exerting a strong impact on ECSA loss and other degradation characteristics. Our parameter study reveals peculiar influences of kinetic rates and structural changes. This allows guidelines to be provided for the design of catalyst layer microstructures, including the distribution of ionomer therein.

Vanadium redox flow batteries (VRFBs) are a promising technology to capture and store energy from renewable sources, reducing the reliance on fossil fuels for energy generation. However, during the charging process, the parasitic hydrogen evolution reaction at the negative electrode affects the performance and durability of VFRBs. The evolution of hydrogen bubbles causes the loss of effective reaction area and blocks the transport of reactants. We employ the lattice Boltzmann method to investigate the two-phase flow transport in the negative electrode of VRFBs. Systematic parametric analyses reveal that increased gas production leads to uneven gas removal from the electrode, while an optimal flow rate can effectively remove bubbles and reduce external pumping energy. Additionally, increasing the compression ratio hinders gas removal but enhances electrode electrical conductivity. Overall, the present study provides valuable mechanistic insights into bubble generation at the negative electrode of VRFBs and offers a theoretical reference for designing and optimizing VRFBs.

Electrochemical impedance spectroscopy is widely used to probe transport and reaction processes in polymer electrolyte membrane fuel cell catalyst layers. However, the relationship between catalyst ink composition and catalyst-layer impedance remains poorly understood, despite its potential value for catalyst-layer design and degradation analysis. Here, a previously developed impedance model is coupled with a structure-based wetting model that predicts water-retention curves and ionomer-film connectivity from ink composition. The resulting framework is also used to assess possible origins of degradation-induced impedance changes. Using a semi-global fitting strategy, the model reproduces experimental spectra for different ionomer-to-carbon ratios. Increasing the ionomer-to-carbon ratio produces steeper water-retention curves and greater ionomer-film connectivity. The associated decrease in polarization resistance and contraction of the low-frequency impedance arc arise from the coupled effects of water uptake on oxygen transport and ionomer connectivity on proton transport. A single composition change is insufficient to explain the experimentally observed non-monotonic impedance evolution during accelerated stress testing. Instead, the degradation response likely reflects coupled changes in electrochemically active surface area, oxygen-reduction-reaction kinetics, pore structure, wettability, and transport properties. This work establishes a framework linking ink composition, catalyst-layer structure, and impedance, and demonstrates that degradation-induced impedance evolution requires a multi-parameter interpretation.

Prognostics and Health Management is crucial for the reliability and lifetime assessment of Polymer Electrolyte Fuel Cells (PEFCs). Here, we review the current advances on this topic, focusing mainly on key degradation mechanisms and methodologies such as physics-aware, data-driven, and hybrid modeling approaches. Key open challenges are analyzed, including the need for more accurate degradation modeling, effective management of multi-stack systems, and advancements in the currently underdeveloped action phase, in which diagnostic and prognostic insights are translated into real-time system responses, such as dynamic load derating, thermal-management adjustments, or automated maintenance triggers, to prevent failures and extend PEFC life. While notable strides have been made in recent years in diagnostics and remaining useful life estimation, it remains challenging to seamlessly integrate these insights into actionable strategies. Future directions highlight the need to address data scarcity and advance interdisciplinary research. Key focus areas include sensor integration, artificial intelligence, and digital twins. Additionally material innovations play a crucial role in bridging existing gaps. This work, therefore, intends to map the further development of Prognostics and Health Management systems toward ensuring the viability of PEFCs in practical applications.

The accurate interpretation of measured electrochemical impedance spectroscopy (EIS) data is crucial for understanding degradation mechanisms in solid oxide cells (SOCs) [1]. EIS is a powerful tool for measuring SOCs in operando, but the analyses of EIS data requires physical models to decipher and interpret the underlying cell processes [2]. In this contribution, two connected physical models are presented. The first part focuses on a model of degradation processes in SOC. The key feature of the model that will be presented is the hierarchical approach linking local degradation at the particle level with performance effects on the electrode level. In the second part, a model for the electrochemical impedance spectrum of the anode active layer and the anode support layer will be presented. The underlying white box approach allows the interpretation of cause and effects involved in EIS experiments. In combination, the two models connect materials, component and cell properties, with the operating conditions of a SOC to the outcome of EIS experiments. The whole modeling framework allows fast computation on a normal laptop. In the first part, the degradation model will be described in more detail. The degradation model focusses on the fuel electrode, since the most severe degradation contribution originates in it [3]. The degradation framework, illustrated in Figure 1, comprises three interdependent structural levels. On the particle level , degradation takes place by Ostwald ripening, coagulation or poisoning of particles [4, 5, 6, 7, 8]. The rate of these particle degradation processes is influenced by the local reaction environment. First and foremost, the local electrode potential drives Ostwald ripening. On the electrode level, the ensemble of solid particles within the electrode is modeled. A statistical electrode is assumed and the local potential and current are calculated. The statistical electrode is built up in three steps with essential parameters passed from one to the next. We start with spherical particles describing the electrode. Two particle radius distributions (PRDs) – for metal and ceramic particles – capture the structural evolution. While ceramic particles remain static, the metal particles change with time under the influence of local conditions during cell operation and dependent on electrode properties [4, 5, 6, 7, 8]. In the “percolation theory” step (see Fig. 1), the electrode’s statistical microstructure is calculated from the PRDs. Assuming that the electrode is described by a porous medium, we can use percolation theory to extract the essential microstructural properties, such as the triple phase boundary length, and electronic and ionic conductivity [9, 10, 11, 12, 13]. These parameters provide a bridge between the structural evolution and the macroscopic performance of the electrode. In the performance model, we assume that the electrode consists of two interpenetrating homogeneous phases. Hence the electrode can be described by the phase’s conductivity and the reaction area between them. With porous electrode theory, spatially dependent variables, such as phase potentials and current distribution, are derived [14]. These local variables is fed back to the degradation processes, focusing on potential-driven Ostwald ripening as the dominant mechanism. This feedback loop allows for the iterative, self-consistent calculation of degradation in the fuel electrode. At the device level, the model computes the performance and health indicators like remaining useful life as functions of operational conditions over the lifetime of the cell. Additionally, diagnostic information like polarization curves and EIS spectra can be analyzed [15]. The description of the numerical model for electrochemical impedance spectra follows [15]. It is assumed that the SOC is isothermal and the pressure gradient across the cell is negligible. For the anode active layer, reaction processes are captured by the Butler-Volmer-equation and charge conservation is assumed. The transport of hydrogen through the anode support layer is described using Fick’s law of diffusion. The set of ordinary differential equations (ODEs) in the frequency domain, obtained for this system, can be solved numerically as a boundary value problem. Solving the ODE system for a range of perturbation frequencies leads to the electrochemical impedance spectrum. Our research has shown that the Warburg element alone is insufficient to describe the transport-related impedance contribution [15]. This contribution depends not only on the support layer's parameters but also on the electrochemical activity and capacitance of the adjacent active layer. The transport impedance of the support layer always depends on the adjacent electrode, especially on the electrode’s capacitance [16]. In case of a planar electrode, which was the adjacent electrode in case of Warburg’s experiment, the capacitance is three orders of magnitude smaller than the capacitance of a porous electrode [16]. In an SOC, the support layer is next to a porous electrode, and thus its transport impedance is not sufficiently described by the Warburg impedance [16]. This knowledge provides new access to understand impedance data as well as may help unravel degradation mechanisms in the electrode. From fitting, the impedance model can reveal effective parameters of the electrode like the effective ion conductivity and hydrogen diffusivity. When connecting this impedance model on the right loop in Figure 1 with the degradation model in the left loop, we found that the growth rate of metal particles increases with local potential, leading to faster growth near the dense electrolyte layer. This reduces the active surface area and the reaction rate in the active layer. Hence the anode active layer impedance increases. The model for the electrochemical cell impedance [15] shows that this local electrode degradation is connected to an increase of the transport impedance in the adjacent anode support layer. Furthermore, over time particle growth results in a loss of structural connectivity. When the metal particle network collapses, affected regions of the electrode become inactive and contribute only as ion-conducting domains. Over time, this leads to a shift the most active zones from the region near the electrolyte to region near the porous layer, increasing the ohmic resistance obtained from the impedance spectra. The feedback loop between the particle level and electrode-level properties allows the model to calculate lifetime, performance loss, and impedance spectra.We introduced a white-box modeling approach that captures the impact of local structural changes in the metal phase on the performance and impedance of a SOC fuel electrode, including its support. Two main observations become clear when performing a simulated degradation experiment: local degradation, originating at the dense electrolyte, affects hydrogen transport impedance in the anode support layer and the anode active layer shifts due to connection losses, increasing the ohmic resistance of the SOC. References [1] C. Mänken, et.al., J. Electrochem. Soc. 171 (2024) pp. 064503. [2] D. D. MacDonald, Electrochim. Acta 51 (2006) pp. 1376-1388. [3] M. B. Mogensen, et al., Fuel Cells 21 (5)(2021) pp. 415. [4] R. M. Darling, J. P. Meyers, J. Electrochem. Soc . 150 (2003) pp. A1523. [5] P. Urchaga, et al., Electrochim. Acta 176 (2015) pp. 1500-1510. [6] A. Kregar, T. Katrašnik, Open Physics 17 (1) (2019) pp. 779. [7] I. Lifshitz, V. Slyozov, J. Phys. Chem. Solids 19 (1961) pp. 35. [8] C. Wagner, Z. Elektrochem . 65 (1961) pp. 581. [9] P. Costamagna, et al ., Electrochim. Acta 43 (1998) pp. 375. [10] D. Bouvard, F.F. Lange , Acta. Metal. Mater. 39 (12) (1991) pp. 3083. [11] D. Stauffer, A. Aharony, Introduction to Percolation Theory: Second Edition, 1992, Taylor & Francis. [12] D. Chen,et al., J. Power Sources 191 (2) (2009) pp. 240. [13] S. H. Chan, Z. T. Xia, J. Electrochem. Soc . 148 (4) (2001) pp. A388. [14] M. Eikerling, A. A. Kornyshev, J. Electroanal. Chem . 453 (1998) pp. 89. [15] M. Knappe, A. Kulikovsky, J. Electroanal. Chem. 975 (2024) pp. 118773. [16] A. Kulikovsky, Electrochemistry Communications 84 (2017) pp. 28-31. Figure 1

Proton exchange membrane fuel cells hold great promise as sustainable and efficient energy converters. However, further advances in cell performance hinge on a comprehensive understanding of the causal relations between structure, properties, and electrochemical responses. This study elucidates oxygen transport limitations within the cathode catalyst layer of the cell by investigating the dependence of the pressure-independent mass transport resistance upon temperature, relative humidity, and electrochemically active platinum surface area per geometrical electrode area, i.e., roughness factor, employing a novel transient limiting current procedure. Utilizing a transmission-line-model, we simulate the pressure-independent mass transport resistance in microporous layer and catalyst layer, deriving a model that can be straightforwardly implemented into macrohomogeneous fuel cell models. It reproduces well the dependence of mass transport resistance upon temperature and relative humidity for both a low and a high surface area carbon support. Through accelerated-stress-test-induced ageing, the platinum surface area is modified to validate the model, demonstrating excellent agreement. The insights from this investigation contribute to an improved understanding of oxygen transport limitations and can thus guide the design of proton exchange membrane fuel cells.

Solid Oxide Cell (SOC) technology is currently experiencing a high level of interest, as the capabilities and the potential of the technology are well aligned with the global efforts to achieve zero emissions. Forschungszentrum Jülich has been heavily involved in the SOC research for more than three decades and has become a cornerstone of the global SOC research community. Selected historical highlights of this research include Jülich’s contributions to the development of the fuel electrode-supported design with material innovations and the advancement of the design for scalability in cells and stacks [1, 2], the development of the Crofer 22 APU alloy in close cooperation with industrial partners [3, 4] and the long-term operation of a short stack in fuel cell mode for more than 100,000 hours [5]. Today, the SOC research at Jülich is organized in five institutes and is characterized by a multidisciplinary approach to solving current fundamental and applied challenges in order to further develop the technology for various application areas. This comprehensive approach covers all aspects from materials to systems, including synthesis, fabrication, modeling, testing, demonstration, and post-test analysis. This contribution provides a general overview of these activities addressing current research topics and recent advances. Sustainability & cost reduction Making SOC technology more economically attractive is an important goal, which can be achieved in different ways. One approach is to use cheaper steel grades, e.g. DIN1.4509 or DIN1.4016 that were not originally optimized for SOC application, but whose performance could be improved by coatings retaining Cr-evaporation. As part of a European project, long-term oxidation studies were carried out on coated steels [6], revealing excellent adhesion and microstructure stability of the coating systems. In addition, new coatings based on Mn-Co-Fe or Mn-Cu-Fe-spinels applied by electrophoresis (EPD) have been developed. EPD could be a more sustainable alternative to plasma spraying, especially suitable for thin-film cassette-type interconnects. Scale-up to real stack sizes and comparison with wet powder spraying [7] are currently underway. Another way to reduce costs is to minimize efforts to purify reactants (water, air, etc.). For this case the effect of pollutants such as Cl or S on the microstructure and properties of the cell components has been studied [8], revealing important degradation mechanisms. Cell degradation studies can be complemented with recently developed FIB/SEM and X-ray computed tomography [9]. Furthermore, approaches for the recycling of SOC metallic constituents have been developed with the aim of producing different Cr- and Ni-containing stainless steel grades, supported by thermodynamic modeling [10]. A recycling strategy for operated stacks and especially also for the cell fraction has been implemented in a large German funded R&D project [11, 12]. The developed cell recycling route starts with a reoxidation step of the metallic Ni, the acidic leaching of the air electrode and the remaining contact layer material. The resulting fraction then consists of NiO, 8YSZ and remnants from GDC. This material mixture could then be post-processed (milling) and re-dispersed into a tape casting slurry for the fuel support up to amounts of 50% of recyclate. Full cells were fabricated and showed similar performance to the state-of-the-art fuel-electrode supported cells. The leached fraction could be separated into La-phase and a residual phase. The La could be re-processed. The ongoing work focuses on the remaining leached fraction. Cell development & modeling Extensive studies of long-term SOC operation and degradation at Jülich show that a ceramic material substitution is necessary to achieve a long lifetime of the SOC technology. The efforts include the development of proton conducting ceramic cells [13] and new electrode materials for oxygen conducting ceramics [14]. Finally, the driving force is to lower the application temperature to reduce the impact of degradation mechanisms. Since Ni-YSZ cermets have shown high degradation rates in steam electrolysis due to Ni migration, there is a strong focus on replacing Ni-YSZ in fuel electrode-supported cells. To enable a Ni-GDC cermet electrode, a three-layer electrolyte (GDC-YSZ-GDC) has been developed using a combination of screen-printing and magnetron sputtering. Excellent cell/stack performance can be achieved by avoiding interdiffusion between YSZ and GDC at the electrode/electrolyte interface, but cell processing needs to be optimized. Other efforts to replace Ni-YSZ include the development of an all-ceramic electrode made of SrTi 0.5 Fe 0.5 O 3- d (STF), and the development of perovskite oxides with exsolved Ni particles. To gain a fundamental understanding of the degradation of electrode structures, a hierarchical model was developed that relates changes at the level of electrode particles to the evolution of the electrode structure and resulting material properties, and ultimately to the overall lifetime performance. In the fuel electrode, it was found that the limited ion conduction leads to a locally enhanced degradation rate close to the electrolyte side, until the breakdown of the percolating nickel particle network and thus of the electron conductivity is reached, resulting in a movement of the degradation zone deeper into the electrode. This creates a moving degradation front at the microstructural level, which leaves a fingerprint in the electrochemical impedance spectra. Overall, the model can be easily modified and extended (e.g. by including Ni migration). Since its computational time is low, it could be used as a concomitant analyzing tool during the operation of the SOC. Another challenge that is being addressed is the impact of different fuel sources (ammonia, biogas and their impurities) [15] on the functional properties and lifetime of fuel cells. Similarly, different types of components (e.g. sodium chloride) in water sources can affect the application of solid oxide electrolyzers. As access to high purity water is an issue and electrolysis should not contribute to further depletion of drinking water sources, the development of wastewater or saltwater electrolysis is an important sustainability goal for the hydrogen economy. Stack technology & characterization One of the 20-layer stacks assembled with the prospect of being used in the rSOC system showed a short circuit after the initial joining process and cell reduction. To avoid the disposal of the stack, dismantling was carried out level by level for six repeating units until the damaged layer could be removed. A green foil of the glass-based composite sealant was placed on the residuals of the broken joint and a second joining process was performed against a new top plate. After this repair process, the stack operation could be started with promising results. In the field of electrochemical stack characterization, the research focuses on the development of innovative measurement and analysis techniques. Fiber optic sensors are used for precise and compact temperature measurements under highly dynamic SOC operating conditions. A combined approach using electrochemical impedance spectroscopy (EIS) and total harmonic distortion provides detailed insight into the performance of a co-electrolysis stack. A novel data-driven methodology was developed using 2,600 EIS measurements from SOC stacks operated in various modes for over 47,000 hours. This method allows reconstruction of the EIS from sparse frequency sampling [16]. Long-term degradation effects are studied in a multi-stack configuration consisting of six sub-stacks operated under co-electrolysis conditions, revealing the effect of operating time with a common history of all samples. Additional degradation analysis focuses on one stack under steam electrolysis at reduced temperatures, with variations in current density and feed gas composition over four 1,000-hour phases. Modeling efforts include a CFD-based sulfur poisoning analysis of the co-electrolysis and a predictive performance evaluation method coupling phase field modeling with CFD. At the system level, the rSOC system design in the 10/40 kW power class demonstrated reliable operation for over 11,500 h at temperature, with ongoing optimization of control strategies for cyclic operation and realistic load profiles. A digital twin of the integrated module of the rSOC system, developed using OpenFOAM, was validated and supports fast, accurate characterization. Post-test analysis Post-test analyses provide critical insight into failure modes and degradation processes, including electrical behavior, material interactions, and operational influences. Failure mechanisms such as short circuits, leakage, and external factors are characterized alongside degradation phenomena such as chromium poisoning and sealant degradation. Key operating parameters such as temperature, current density, and fuel composition, are evaluated for their impact on performance and material stability. An SOC-stack autopsy methodology has been developed that demonstrates the disassembly of a module for the subsequent post-test analysis. The Jülich long-term test in fuel cell operation, which lasted about 10 years, was investigated immediately after the end of the test [17, 18]. The results showed relatively few changes, interactions or damage considering the long operating time. One of the main conclusions was that the interface between the LSCF air electrode and the GDC barrier layer was somewhat changed. Secondary phase formation was observed, leading to tiny nanocrystals and partial incorporation of Cr into the LSCF grains. The secondary formed crystals were also found in the pores of the GDC layer. Additional advanced characterization tools such as Raman spectroscopy and µ-Laue diffraction revealed similar results compared to the SEM characterizations. However, one simple question remained unanswered. None of the techniques applied could verify or falsify whether the LSCF perovskite was still a perovskite or had transformed into another crystal structure after such a long time. Thus, additional high-resolution TEM investigations were performed. Finally, it was proven that the entire air-electrode volume, from the interface to the GDC to the bulk layer, is still a perovskite. This result proves the chemical stability of the perovskite structure. Acknowledgement The authors would like to thank their colleagues at Forschungszentrum Jülich GmbH for their great support and the Helmholtz Association and the German Federal Ministry of Education and Research (BMBF) for funding these activities within the framework of the SOC Degradation 2 (FKZ 03SF0621A), ReNaRe/H2Giga (FKZ 03HY111J), ElChFest/H2Giga (FKZ 03SF0641A), iNEW2.0 (FKZ 03SF0627A), MacGyver (FKZ 03SF0785A), PRELUDE, DryHy (FKZ 03SF0716A) and PHOENIX (FKZ 03SF0775A) projects and the European Union under the NOUVEAU project (GA 101058784). References Stöver, D., et al., ECS Proceedings Volumes, 1999. 1999-19 (1): p. 812. Menzler, N.H., et al., Fuel Cells, 2013. 14 (1): p. 96-106. Hojda, R., et al., Production-capable materials concept for high-temperature fuel cells; Grossserientaugliches Werkstoffkonzept fuer Hochtemperatur-Brennstoffzellen. 2003. VDM Crofer 22 APU Material Data Sheet No. 4146 March 2022 . 2022 [cited 2025 18.01.2025]; Available from: https://www.vdm-metals.com/fileadmin/user_upload/Downloads/Data_Sheets/Data_Sheet_VDM_Crofer_22_APU.pdf . Fang, Q., et al., Journal of The Electrochemical Society, 2019. 166 (16): p. F1320. NOUVEAU Newsletter / issue n°3 / May 2024 . 2025 [cited 2025 24.01.2025]; Available from: https://www.nouveau-project.eu/wp-content/uploads/2024/05/NOUVEAU-Newsletter-M20.pdf . Wolff, M., et al., Journal of Power Sources, 2024. 592 : p. 233931. Kirillov, I., et al., Investigating the Impact of Chlorides on LSCF Cathode Degradation in Solid Oxide Fuel Cells (SOFCs) at Elevated Temperatures . 2025. Wehner, L., et al., ECS Meeting Abstracts, 2023. MA2023-01 (54): p. 34. Lastam, J., et al. Metals, 2024. 14 , DOI: 10.3390/met14040406. Sarner, S., et al., Green Chemistry, 2024. Sarner, S., et al., Advanced Energy Materials, 2022. 12 (35): p. 2201805. Schley, L., et al., Energy Advances, 2024. 3 (4): p. 861-873. Uecker, J., et al., Electrochimica Acta, 2023. 452 : p. 142320. Kumar, A., et al., ChemElectroChem, 2024. 11 (15): p. e202300845. Mänken, C., et al., Journal of Power Sources, 2025. 628 . Dellen, C., et al., ECS Transactions, 2023. 111 (6): p. 1845. Menzler, N.H., et al., Journal of Power Sources, 2020. 478 : p. 228770.

We report a physics–based model for the electrochemical impedance of a PEM fuel cell cathode. The model takes into account the transient behavior of oxygen and proton transport in the cathode catalyst layer caused by a variation of the liquid saturation with cell current. Transients of the catalyst layer oxygen diffusivity result in a second capacitive arc in the Nyquist spectrum, while proton conductivity transients lead to the formation of an inductive loop. In the range of capillary pressures in which the liquid saturation in the catalyst layer is independent of the capillary pressures, the loop does not form. A stability analysis of a reduced system of equations reveals that the static limit of inductive loop is unstable with respect to spatial perturbations, implying that the post–oscillatory steady state is unattainable. Possible scenarios of instability development are discussed.

A recent physics-based model for liquid and gaseous water transport in the cathode catalyst layer (CCL) is incorporated into our 1d + 1d model for the PEM fuel cell impedance. The model includes parametric dependencies of the CCL oxygen diffusivity and proton conductivity on the liquid saturation. Fitting of the 1d + 1d model to experimental impedance spectra of a PEM fuel cell reveals two intriguing effects. Contrary to common belief, the liquid water saturation in the CCL is nearly independent of cell current density due to the growing liquid pressure gradient that drives liquid water removal from the CCL. Further, the "dry"oxygen diffusivity of the catalyst layer increases with cell current density. Apparently, at small current density, electrochemical conversion proceeds primarily in narrow pores, where the Knudsen oxygen diffusivity is low. With growing current density, larger and better connected pores with higher oxygen diffusivity dominate in the current conversion, leading to increase in effective oxygen diffusivity observed in impedance spectroscopy data.

The utilization of water electrolyzers for hydrogen production represents one of the promising technologies in the global transition away from fossil fuels towards sustainable energy sources. Among the various types of electrolyzers, alkaline electrolyzers in the zero-gap configuration with different separator types, like diaphragms and anion conducting membranes, have garnered significant attention, among others, due to their potential to utilize non-precious metal catalysts such as iron and nickel for the electrodes. This characteristic holds promise for reducing costs and increasing accessibility to hydrogen production technologies, including the establishment of largescale stationary systems and hydrogen production powered by renewable energy. In order to accelerate the development process and to understand the important mechanisms in more detail within these cell types, numerical Computational Fluid Dynamics (CFD) simulations are conducted and combined with experimental cell testing at the Juelich Research Center. The studies focus on investigating the influence of variations in the porous transport anode-electrode, particularly nickel fiber fleece and nickel foam, along with their characteristic properties on the performance of the anion exchange membrane water electrolysis cells. This includes investigating porosity gradients, optional additional catalyst layers and their loadings, and the effects of changing types and thicknesses of porous transport electrodes. These studies are conducted at varying electrolyte concentrations and operating conditions. Using the open-source platform OpenFOAM® and the developed libraries of OpenFuelCell21, CFD simulations are performed at both the micro and meso scales to examine the aforementioned questions (see Fig. 1). The model incorporates all key transport phenomena, such as two-phase fluid flow, described via an Eulerian-Eulerian approach, heat and mass transfer, electrochemical reactions, species transfer, and charge transfer across the various functional layers of the cell. It allows for the analysis of electrochemical reactions by employing the Butler-Volmer equation to describe electrokinetics across various types of electrochemically active porous electrodes. For the multi-physical simulation of the detailed flow within the porous electrodes at micro scale, microtomography images are used as geometrical input. In addition, the microscopic data are subsequently utilized for characterizing the porous electrodes in the macro-homogeneous simulations performed at the meso-scale. Literature: Zhang, S. Hess, H. Marschall, U. Reimer, S. Beale, W. Lehnert, openFuelCell2: A new computational tool for fuel cells electrolyzers, and other electrochemical devices and processes, Comput. Phys. Commun. 298 (2024) 109092. Figures: Figure 1: Numerical simulation of the flow through different micro-porous electrodes (left) which results are used for the macro-homogeneous simulations of alkaline electrolysis cell in 2-D and 3-D (right). Figure 1

Fuel cells and electrolyzers are key components in the energy transformation from non-renewable fossil fuels, to technologies based on renewable energy, such as solar and wind, which utilize or synthesize hydrogen (and oxygen) rich fuels/feedstocks in place of hydrocarbons. Following a brief introduction, a description of modeling activities, past, present and future, developed by the authors over many years, are described. Electrochemical conversion cells may be considered as natural problems in combined heat and mass transfer with the electrochemical reactions functioning as the main driving force for the interconversion of chemical and electrical energy, and heat. Ionic and electronic charge transfer within the electrodes and electrolyte must be considered, and multi-phase flow (liquid-gas) is also frequently present. The authors have led development of a suite of object-oriented open source models for both polymer electrolyte membrane cells (PEMCs) and solid oxide cells (SOCs), the emphasis being on generic (rather than specific) algorithms. The implementation is based around the popular open-source finite-volume-based software library, OpenFOAM. This allows the user access to a flexible framework/library of C++ classes, thereby allowing for further model research, as well as development appropriate to practical real-life applications. The main features of the methodology are described and explored both, with the benefit of hindsight from past experiences, and foresight to the problems to be tackled in the future. In addition to the above cell-level macro-homogeneous analysis, substantial work is being devoted to micro-scale analysis; for example, detailed studies of flow phenomena within the porous transport layers and electrodes of electrochemical cells as well as construction of digital twins of physical morphologies. These are necessary, not only for basic qualitative understanding and control of the underlying 3-D processes, but also to provide quantitative closure parameters for quantifying the dynamics of macro-scale equations (multi-scale models). The state-of-the art of CFD-based modeling in electrochemical science is presented with examples of PEMC (and SOC) technologies together with a discussion of the important issues and limitations of the present generation of models as well as examples of success stories. Some of the issues for concern include degradation and evolution of material properties, code stability and numerical issues, suitability of present-generation existing two-phase models (mixture, Euler-Euler, volume of fluid methods, etc.) to electrochemical applications, and computational issues such as parallel efficiency. In spite of numerous unexpected problems encountered over the years, great progress was, is, and will continue to be made in this quest for mathematical descriptions and prototypes of complex electrochemical processes and products, based on physical modeling. Artificial neural network techniques have found applications in tackling scientific and industrial challenges involving the concept of digital twins. In the domain of electrochemical cells, data-driven models or physically informed data-driven methods have been shown to provide valuable insight into issues that were traditionally explored using physical models. Areas for investigation include; material properties, performance prediction, and control strategies. By integrating physical/mathematical models with data-driven counterparts, a more comprehensive suite of methodologies should become capable to address ever more complex issues in the field.

Efforts in design and optimization of catalyst layers for polymer electrolyte fuel cells hinge on mathematical models that link electrode composition and microstructure with effective physico-chemical properties. A pivotal property of these layers and the focus of this work is the proton conductivity, which is largely determined by the morphology of the ionomer. However, available relations between catalyst layer composition and proton conductivity are often adopted from general theories for random heterogeneous media and ignore specific features of the microstructure, e.g., agglomerates, film-like structures, or the hierarchical porous network. To establish a comprehensive understanding of the peculiar structure-property relations, we generated synthetic volumetric images of the catalyst layer microstructure. In a mesoscopic volume element, we modeled the electrolyte phase and calculated the proton conductivity using numerical tools. Varying the ionomer morphology in terms of ionomer film coverage and thickness revealed two limiting cases: the ionomer can either form a thin film with high coverage on the catalyst agglomerates; or the ionomer exists as voluminous chunks that connect across the inter-agglomerate space. Both cases were modeled analytically, adapting relations from percolation theory. Based on the simulated data, a novel relation is proposed, which links the catalyst layer microstructure to the proton conductivity over a wide range of morphologies. The presented analytical approach is a versatile tool for the interpretation of experimental trends and it provides valuable guidance for catalyst layer design. The proposed model was used to analyze the formation of the catalyst layer microstructure during the ink stage. A parameter study of the initial ionomer film thickness and the ionomer dispersion parameter revealed that the ionomer morphology should be tweaked towards well-defined films with high coverage of catalyst agglomerates. These implications match current efforts in the experimental literature and they may thus provide direction in electrode materials research for polymer electrolyte fuel cells.

Water plays a crucial role for the operation of polymer electrolyte fuel cells. The distribution and state of water in the membrane-electrode-assembly (MEA) not only impacts the performance of the cell, but also its degradation and freeze-start capability. It is therefore important to have a model that precisely predicts the distribution of water across the fuel cell for arbitrary inlet conditions. Unfortunately, water transport coefficients reported in the literature vary by orders of magnitude while operation conditions and other crucial properties of the cell are not precisely known. Furthermore, a large amount of data is simply outdated due to improvements in measurement techniques and setups as well as fuel cell components. In the present work, we have devised a test procedure for water transfer measurements with state of the art MEAs. In parallel, we have derived a semi-empirical model and parametrized it with measured data. The model allows predicting MEA water fluxes and profiles of temperature and concentration over a wide operation range, relevant for passenger as well as heavy duty automotive applications.

Alkaline water electrolyzers (AWE) have several advantages over other types of electrolyzers, including their high efficiency and especially their relatively low cost due to the usage of non-precious metal catalysts, such as nickel and iron, for the electrodes. Information about local quantities and physical phenomena such as the formation of gas bubbles, current densities, temperatures or local species concentrations within a running cell are important for their improvement. Multiphysical computational fluid dynamics (CFD) simulations of electrochemical components using detailed three-dimensional models can provide valuable insight on local behaviors and characteristics that are difficult or impossible to measure experimentally. This work extends the CFD library openFuelCell2 1, which has been implemented using the open-source platform OpenFOAM®, to simulate AWE cells. The model considers the major transport phenomena, including two-phase fluid flow, heat and mass transfer, electrochemical reactions, species transfer and charge transfer in the various functional regions of the cell. It employs an Eulerian-Eulerian approach to characterize the behavior of each phase comprising interphase mass transport, momentum exchange and heat transfer. Appropriate mapping functions are used to couple the physically distinct regions together. A Butler-Volmer equation characterizes the electrochemical reactions that are assumed to occur in electrodes of finite thickness. This model is used to simulate a single zero-gap AWE cell, depicted in Fig. 1, for different operating conditions such as varying temperatures and volumetric flow rates. The conducted studies provide insight into the local formation of the created gas phase (bubbles), the distribution of species within the gas and electrolyte and their impact towards the performance of the running cell. These numerically obtained results are compared to in-house available and gathered experimental data. Figure 1 demonstrates that the polarization curves obtained at various temperatures are in good agreement with the experimental data. Figure 1

Forschungszentrum Jülich has been involved in the research and development of SOCs for more than 30 years. In this work, selected highlights on material, cell, stack and system development are presented, whereas improving performance and understanding degradation phenomena were in the focus. On cell level, operation in steam, co-electrolysis as well as pure CO 2 electrolysis mode was researched. A hierarchical degradation model framework was developed that relates changes at the level of electrode particles to changes in electrode structure, resulting materials properties and overall lifetime-performance. On stack level, progress in clarification and optimization of performance and lifetime relevant processes was made. The role of contaminants, foremost silicon species and sulfur dioxide in feed gases, was investigated to support technical applications. On system level, an rSOC system with an output power of 10 kW in fuel cell mode and an input power of 40 kW in electrolysis mode was developed.

This work presents a model for the prediction and analysis of voltage losses in proton exchange membrane fuel cells arising from accelerated stress testing. It consists of two submodels. The first submodel uses a statistical physics-based population balance approach to describe the degradation of the catalyst active surface. It is combined with a performance submodel that allows incorporating the degradation of the catalyst activity. During testing, a dedicated diagnostic procedure is used to determine the cell performance and the cathode properties, like the electrochemical active surface area, during the stress tests. It was found that the change of the catalyst activity, described by Tafel slope and exchange current density, correlates with the change in active surface area. The model allows the description of catalyst surface reduction, changes of Tafel slope and exchange current density as well as voltage losses. We find that the voltage losses attributed to the loss of electrochemical active surface area are minor, while the dominant factor is the change of the Tafel-slope. Accordingly, this study shows that during PEM-FC cathode degradation studies the Tafel slope should be the most relevant metric. The model describes the experimental data with a standard deviation of 7.1 mV in a range of 0-2.0 A/cm2. The model is intended to be used as a building block for the prediction of performance losses of PEM fuel cells under drive cycle conditions.

Maintaining a well-balanced water distribution is crucial for the operation of polymer electrolyte fuel cells. The water distribution depends to a large extent on the ease of liquid water transport in the diffusion media (DM) and in the flow field (FF) on the cathode side and, especially, on processes in boundary regions between these media. In this article, the droplet dynamics at the DM-FF interface are studied by a mechano-hydrodynamic pore model. The presented treatment allows droplet detachment to be rationalized, in dependence of pore radius and air flow velocity. Large pore radius and high air flow velocity favor droplet detachment during an earlier stage of growth, at which the contact line of the droplet remains pinned at the pore opening while the droplet volume expands during growth. We evaluate the trend of detachment time and detachment height of a droplet. Furthermore, we analyze the influence of the air flow velocity on the time-averaged liquid pressure at the pore exit. The physics governing the flooding behavior of the pore is discussed. The presented model-based analyses provide theoretical insights into material properties and operating conditions that improve water removal.

Water exerts a crucial influence on the performance of a polymer electrolyte fuel cell as both “catalyst activating agent” and “oxygen blocker”. Therefore, fine-tuning the water distribution is imperative for high performance. In this work, we present a water balance model to calculate the distribution of liquid water in cathode catalyst layer and diffusion media. The model incorporates the influence of the local liquid water saturation on the effective transport properties. Liquid water saturation is both a composition variable determining the effective properties and a variable that depends on the solution of the transport equations that use the effective properties. The model reveals the formation of a thin water layer in the diffusion medium adjacent to the catalyst layer at high current density. This interfacial water layer strongly impedes oxygen transport and reduces the oxygen concentration in the catalyst layer, which causes a drastic increase in the voltage loss at high current density that drastically reduces the cell performance. We elucidate the origin of the water layer, present parametric studies of this effect, and propose mitigation strategies. The fundamental understanding gained will aid the development of membrane electrode assemblies with tailored pore network properties to achieve vital improvements in performance.

The development, fabrication, and operation of electrochemical energy devices like fuel cells, batteries, or electrolyzers require powerful diagnostic techniques. Frequency response analysis methods deconvolute and quantify reaction and transport processes based on their dynamics, extract properties such as conductivity, diffusivity, permeability, and electrocatalytic activity, and assess the state-of-health, state-of-charge of a battery, mass activity of an electrocatalytic layer, or the presence of faults. Taking nonlinear information into account improves process identifiability, while using different input or output signals increases the sensitivity towards specific processes. Sensitivity analyses and design of experiments techniques are valuable tools to evaluate different frequency response techniques and help building optimized test protocols.

The defossilization of the energy sector requires the transfer of sustainable, carbon-neutral technologies and processes into application. Along with the development of a global hydrogen economy, technologies that generate, store, distribute and use hydrogen and derivatives are particularly relevant. Considerable potential in this sense is offered by the solid oxide cell (SOC), which can be operated as a fuel cell (SOFC), as an electrolysis cell (SOEC) and reversible (rSOC). Forschungszentrum Jülich has been involved in the research and development of SOCs for more than 30 years. In addition to material and cell development, stack and system development and understanding degradation effects are among the main topics today. Recently, an rSOC system with an output power of 10kW in fuel cell mode and input power of 40kW in electrolysis mode was developed. Four SOC stacks, separated and surrounded by a total of five heating plates plus an air preheater at one end and a fuel preheater at the other end, form the Integrated Module of the system; each stack has 20 layers with an active cell area of 19x19 cm². A compact and optimized design could be realized, which achieves a system efficiency of 63.3 % and 71.1 % in fuel cell mode and electrolysis mode, respectively. The system has already been tested in stationary operation modes. Current developments focus on the operating strategy, in particular on the temperature control of the stack in fuel cell mode and during the transient operation of the system. With a focus on the SOC stack, progress was made both in the area of actual stack development and in the area of clarification and optimization of performance and lifetime relevant processes. The role of contaminants, foremost silicon species and sulfur dioxide in feed gases, was investigated to support technical applications. Headway was also made in applying advanced measuring technology like fibre-optic sensors for temperature measurements in air channels. Degradation processes were investigated both experimentally and simulatively in fuel cells as well as in steam and co-electrolysis operation. On the one hand, machine learning approaches were pursued to analyze degradational patterns in SOC stacks, utilizing a specifically consolidated and curated set of long-term experiments and EIS measurements. On the other hand, a multiphysical stack model was developed that allows the relevant physical processes within the stack to be analyzed individually and coupled and thus to optimize the overall operation of the stack. In the area of the development and investigation of cells and materials, the performance of the SOC in the fuel cell mode as well as in the electrolysis mode was in the focus. In addition to operation in steam and co-electrolysis modes, operation in pure CO 2 electrolysis was also researched. On single cell level the degradation behavior in the different modes of electrolysis operation was investigated. Different alternative materials were examined both on the fuel side and on the air side as well. A hierarchical degradation model framework was developed that relates changes at the level of electrode particles to changes in electrode structure, resulting materials properties and overall lifetime-performance. Model-based diagnostic allows the extraction of model parameters from experimental data, model verification as well as identification and quantification of different degradation mechanisms. Overall, therefore, significant progress can be observed in the field of cell as well as in the field of stack and system development of SOCs in fuel cell, electrolysis and reversible operation at Forschungszentrum Jülich.