Freeze/thaw cycles in PEMFC: Experimentally observed degradation and theoretical mechanisms

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 1/9 B1803 Freeze/thaw cycles in PEMFC: Experimentally observed degradation and theoretical mechanisms Robin Maubert* (1), Guillaume Serre (1), Eric Pinton (1), Marine Cornet (2), Yann Bultel (2) (1) Univ. Grenoble Alpes, CEA, LITEN DEHT, 38000, Grenoble/France; (2) Univ. Grenoble Alpes, Univ. Savoie-Mont Blanc, CNRS, Grenoble INP, LEPMI, 38000 Grenoble/France; *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Durability is a crucial factor for Proton Exchange Membrane Fuel Cells (PEMFCs), notably in heavy-duty mobility applications. Life time of a PEMFC is particularly reduced when it is exposed to negative temperatures, notably during storage [1] or start-up [2]. In order to improve frost resistance, it is essential to understand and model the degradation mechanisms induced by the formation and melting of ice in the cell so that one can simulate various shutdown strategies in cold environment and determine the best one. This work aims at developing a model of the impact of freeze/thaw cycles on PEMFC degradations. We first examine the impact of freeze-thaw cycles on a PEMFC, where the cell temperature cycles below and above 0°C, causing the remaining water inside the cell after a stop to freeze and thaw. These phase transitions can induce structural changes in the catalytic layer, GDL, and their interfaces, affecting active surface area, and gas diffusion, ultimately impacting performance. To investigate further, we conduct electrochemical tests (cyclic voltammetry, polarization curve, EIS) and scanning electron microscopy. Performance losses may be influenced by factors such as the number of cycles, temperature variation speed, and minimum freezing temperature. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 2/9 Introduction In 2022, heavy transport (buses and trucks) accounted for 25% of greenhouse gas emissions in the transport sector [3]. The use of hydrogen fuel cells is often presented as a solution (in competition with battery-powered vehicles) to decarbonize this part of the transport sector [4]. However, hydrogen fuel cells use limited resources such as platinum, which is difficult to recycle when used in the form of nanoparticles [5], and their manufacture produces PFAS, which represent a major risk to the environment and human health [6]. It is therefore necessary to make the manufacture of a hydrogen fuel cell as cost-effective as possible, by aiming for the longest possible service life. To this end, it is necessary to anticipate and reduce all potential sources of degradation, in order to design cells and strategies that maximize their lifespan. In this context, we're looking to predict the ageing process undergone by heavy-duty vehicles equipped with hydrogen fuel cells. When the vehicle is stationary and the outside temperature is negative (typically in the case of a bus parked in a parking lot on a winter's night), the hydrogen fuel cell that equips it will be subjected to negative temperatures, likely to freeze part of the water contained in the membrane, the catalytic layers, the GDLs and the channels. To define mitigation strategies, we need to study and model how this phenomenon degrades the cell. What compromise needs to be made in the shutdown strategy to minimize both the degradation caused by severe drying out and that caused by water freezing? What degradation and drop in performance can be expected over the lifetime of a hydrogen fuel cell exposed to freezing outdoor conditions? To answer these questions, we also need to be able to model the degradation caused by cold start-up of hydrogen fuel cells, so that we can then realistically simulate cold start-ups using different protocols and determine the best ones. The difficulty is that a start-up involves several complex physical mechanisms in a short space of time: heat and mass transfer between the different compartments of the cell, and degradation linked to corrosion of the carbon support [7], which we do not intend to study here. In order to limit our study to a single object, cold-induced degradation, it is more practical to study separately the different stages occurring during a cold start. For this reason, we present here the impact of freeze/thaw cycles on a fuel cell, carried out after a single start-up at room temperature. 1. Scientific Approach The aim of this study is to understand the mechanisms by which freeze/thaw cycles of the water contained in a cell change the structure of the cell components, and then to quantify the impact of these structural changes on performances in order to develop a model describing these degradations. This model will then be added to the Mephysto code [8], which simulates the operation and ageing of hydrogen fuel cells. This will consist in a module giving the evolution of variables of interest (such as potential) at each freeze/thaw cycle, thus reproducing the performance decreases observed experimentally. Mechanism of degradation by freeze/thaw cycles According to the literature on freeze/thaw cycles in a porous, rigid, rock-like medium, degradation of the medium is not caused by the density change of water during freezing [9]. Instead, two other physical mechanisms explain how the cooling of liquid water in a porous medium can lead to its degradation: 1. Water does not entirely turn to ice: a thin layer of liquid water, a few nanometers thick, remains on the ice surface [10]. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 3/9 2. In liquid water at negative temperature, there is a critical radius: this is the minimum size an ice crystal must reach in order to grow. Below this radius, the crystal melts spontaneously; above it, it can grow by solidifying the water in contact with it. This critical radius decreases with temperature according to a -1/T law, where T is the temperature in degrees Celsius [11]. At negative temperature, inside a pore with a radius greater than the critical radius and partially filled with water, the freezing process occurs making the ice grow. At the beginning, the ice is formed using the available local liquid water in the pore. Not all the liquid can be frozen since a very thin layer of liquid water (few numbers of water molecules) always remains at the ice surface. The film thickness remains constant and thus, the ice growth is possible only if the film is fed by incoming liquid from the neighbouring pores. They may contain liquid water if their radii are less than the critical one. This phenomenon is known as cryosuction, which can cause smaller pores to collapse due to a local pressure drop. Another consequence of water displacement due to the liquid film is that it continuously draws water into pores with a radius larger than the critical radius. This increases the pressure at those locations and can lead to pore expansion. These two mechanisms explain why large pores tend to expand, while smaller ones may close or collapse, a phenomenon that has been experimentally observed in hydrogen fuel cell catalyst layers [12]." Factors affecting freeze-thaw cycles Research into food preservation has shown that temperature descent kinetics have a direct influence on the size of ice crystals forming in permeable porous media. At least in some cases [13] [14], the faster the descent, the smaller the crystals and the smaller the quantity of water displaced. It is therefore expected that rapid kinetics, represent a minimum level of degradation. Lower is the minimum temperature reached, higher is the degradation rate since more pores can be frozen (the critical radius being lower). The longer the time spent at negative temperature the worse are the degradations [9]. The higher the moisture content, the greater the structural changes [15]. For this reason, different freeze-thaw cycles can cause very different degradations. It is therefore expected that the degradation caused by freeze-thaw cycles on cell will depend on all these physical parameters. This dependence on the physical parameters of freeze/thaw cycles is not studied in the literature dealing with the impact of freeze/thaw cycles on PEMFCs, which may explain the differences in results from one article to another, in addition to the fact that the components used in the cell vary from one article to another. From literature, it is known that freeze-thaw cycles could induce an increase in porosity in the GDL [16], a decrease in Electrochemical Surface Area [1], an increase in Ohmic resistance [12], delamination between the CL and MPL layers and cracks appearing on the surface of the CCL [2]. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 4/9 Action plan Figure 1 Experimentation protocol In order to isolate the impact of freezing from the impact of ”classical” operation, the study consists in submitting a small differential cell to freeze/thaw cycles, such as described in figure 1, with the aim of assessing the effects of low temperatures on a stopped fuel cell that contains produced water in its different components, Electrochemical characterizations are carried out before and after freeze/thaw cycles to monitor changes in variables of interest such as oxygen transport resistance, electrochemically active platinum surface area (ECSA), and ionic resistance. In addition, post-mortem observations using a scanning electron microscope (SEM) are used to analyse structural changes, notably porosity evolution, as well as the appearance of cracks, the detachment of material fragments or delamination between the different layers. The final objective is to correlate these post-mortem observations with changes in electrochemical parameters, in order to develop a mechanistic or semi-empirical model of PEMFC degradation. 2. Experiments/Calculations/Simulations In this study, freeze/thaw cycles on PEMFC are performed between -20°C and +7°C. After the break-in phase and a first characterization, freeze/thaw cycles are repeated few hundred times before a last characterization. This protocol allows to track the main degradations induced by freeze/thaw cycles. Protocol: The cell is at 80°C, with pure hydrogen at the anode and air at the cathode. and the relative humidity of the inlet gases is set to 95% in order to ensure high moisture content, which EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 5/9 promotes significant degradation due to subsequent freezing. The reactive gases are injected with a very high over-stoichiometry. Conditioning The cell is at 0.6 V for 20 minutes, then undergoes 15 voltage cycles ranging from OCV for 30 seconds to 0.1 V for one minute. The scan rate is 100 mV/s. Characterization before degradation The cell undergoes: - A polarization curve with a voltage ranging from OCV to 0.3V at a scan rate of 2mV/s. - Three GEIS, at 180mA, 1800mA and 4000mA, for frequencies ranging from 100mHz to 100kHz. - Two polarization curves with low-oxygen air (0.5% and 1%) at a voltage ranging from OCV to 0.2V at a scanrate of 5mV/s. Air is replaced by nitrogen at the cathode to perform: - Cyclovoltammetry with a voltage ranging from 0.07V to 1.2V at a scanrate of 100mV/s. - LSV with a voltage ranging from 0.05V to 0.6V at a scanrate of 2mV/s. - A PEIS at 0.4V for frequencies ranging from 500mHz to 200kHz. Air is reintroduced at the cathode. - Voltage is imposed at 0.6V to ensure high cell humidity. - Reactive gases are expelled by injecting nitrogen at 95% relative humidity at the anode for two minutes, then at the cathode for two minutes. The goal is to remove the gases, not the water this is why this step is short. freeze-thaw degradation The cell is then disconnected from the bench, placed in a plastic bag, as shown in figure 2 to ensure a seal between the glycol and the AME. However, this configuration has one limitation: it reduces heat exchange, as the glycol can no longer circulate freely on the surface of the differential cell or in its cooling circuit. It therefore becomes necessary to extend temperature stages during freeze/thaw cycles. In practice, a 1-hour stage at -20°C is required to reach this temperature evenly, followed by a 1-hour stage at +10°C to guarantee a complete return to a positive temperature. Mini-conditioning and Characterization after degradation The cell is removed from the glycol and then, undergoes a mini break-in, identical to the initial one, but reduced to 5 voltage cycles instead of 15 in order to reduce platinum dissolution [17]. The cell is then submitted to the same electrochemical characterizations as before the freeze-thaw cycles. Post-mortem analysis After electrochemical characterization, the cell is dismantled for two types of post-mortem analysis: EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 6/9 • SEM observation: The AME and GDL assembly is embedded in resin, then polished. A thin layer of carbon (approx. 2 nm) is then deposited to enable observation under a scanning electron microscope. • Dynamic contact angle measurement: This analysis is carried out on the GDL as well as on the surface of the catalytic layer in contact with the MPL 3. Results Electrochemical analysis Figure 3 shows that the polarization curve changes sharply after 578 freeze/thaw cycles and relatively little after 75 cycles, suggesting that a large number of freeze/thaw cycles are required to observe degradation. Performance decreases occur mainly at high current densities, which would indicate that the degradation caused by freezing has an impact on gas diffusion in the cell. Figure 3 Evolution of polarization curves Impedance spectra Figure 4 and 5 show an increase in membrane resistance as well as the ionic or/and electronic resistances of the catalyst layers. We can also notice an increase in diffusion impedance, corresponding to a limitation of gas transport through the CL and GDL. Figure 2 Differential cell protected by a plastic bag during immersion in glycol for freeze/thaw cycles EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 7/9 Figure 4 Evolution de la PEIS avec les cycles gel/dégel Figure 5 Evolution de la GEIS avec les cycles gel/dégel ECSA The ECSA decreases by 35% after 578 cycles, giving us the qualitative result of [1], Experimentation limits: • The results obtained are not reproducible: imposing the same number of cycles twice does not produce the same degradation. • During these freeze/thaw cycles, the cell was not sealed against the glycol, which can reduce damage caused by freezing. A second test campaign is currently underway, with improved sealing of the cell. SEM comparison of a reference cathode and a cathode that has undergone 150 freeze/thaw cycles, as shown in Figure 6, reveals no significant structural changes. Statistical comparison of surface area and carbon volume also fails to differentiate the two MEAs. Figure 6 SEM observation of the evolution of a cathode subjected to 150 freeze/thaw cycles Reference cathode Cathode after 150 freeze/thaw cycles EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 8/9 SEM analysis limitations • During sample preparation, the sample is embedded in a resin that penetrates the pores of the structure. Under the scanning electron microscope (SEM), this resin is indistinguishable from the ionomer, making it impossible to differentiate the latter from the initially empty zones. As a result, only the carbon phases and platinum can be studied with this method. If freeze/thaw cycles mainly affect the ionomer, then it is consistent that this approach does not reveal structural degradation. • For the moment, only the cathode has been studied, and it is necessary to study the anode as well as the GDL in the near future. Conclusions The results obtained show that freeze-thaw cycles imposed to a single small cell after a short running time at normal temperature (80°C) reveal a progressive degradation of electrochemical performance mainly at high current density. This degradation is attributed to a deterioration in gas transport reduction, in ECSA reduction and an increase in ionic resistances, confirmed by impedance spectra. MEB observations on the in-situ and ex-situ aged MEA did not provide any information relatively to the active layer structure evolution. It could be the consequence of the observation technic that cannot distinguish pores from ionomer and thus, their deformations. Thanks This work was supported by a France 2030 government grant from Agence Nationale de la Recherche, reference 22-PEHY-0002. References [1] Deprez, M., De Kock, T., De Schutter, G. and Cnudde, V. 2020. A review on freezethaw action and weathering of rocks. Earth-Science Reviews. 203, (Apr. 2020), 103143. [2] Lee, S.-Y., Kim, H.-J., Cho, E., Lee, K.-S., Lim, T.-H., Hwang, I.C. and Jang, J.H. 2010. Performance degradation and microstructure changes in freeze–thaw cycling for PEMFC MEAs with various initial microstructures. International Journal of Hydrogen Energy. 35, 23 (Dec. 2010), 12888–12896. [3] Transition Écologique, M. de la 2025. Site officiel de l’environnement en France. [4] Ajanovic, A. and Haas, R. 2021. Prospects and impediments for hydrogen and fuel cell vehicles in the transport sector. International Journal of Hydrogen Energy. 46, 16 (Mar. 2021), 10049–10058. [5] Duclos, L. 2016. Vers l’éco-conception des piles à combustible : développement d’un procédé de recyclage des catalyseurs des systèmes de PEMFC à base de platine. [6] Abunada, Z., Alazaiza, M.Y.D. and Bashir, M.J.K. 2020. An Overview of Perand Polyfluoroalkyl Substances (PFAS) in the Environment: Source, Fate, Risk and Regulations. Water. 12, 12 (Dec. 2020), 3590. [7] Zhao, J., Tu, Z. and Chan, S.H. 2021. Carbon corrosion mechanism and mitigation strategies in a proton exchange membrane fuel cell (PEMFC): A review. Journal of Power Sources. 488, (Mar. 2021), 229434. [8] Schott, P. and Baurens, P. 2006. Fuel cell operation characterization using simulation. Journal of Power Sources. 156, 1 (May. 2006), 85–91. [9] Wang, R., Hu, Z., Li, Y., Wang, K. and Zhang, H. 2022. Review on the deterioration and approaches to enhance the durability of concrete in the freeze–thaw environment. Construction and Building Materials. 321, (Feb. 2022), 126371. [10] Su, Z., Tan, X., Chen, W., Jia, H. and Xu, F. 2022. A model of unfrozen water content in rock during freezing and thawing with experimental validation by nuclear magnetic EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476309 B1803 Page 9/9 resonance. Journal of Rock Mechanics and Geotechnical Engineering. 14, 5 (Oct. 2022), 1545–1555. [11] Zhou, H., Lin, X., Wang, M. and Huang, W. 2013. Calculation of crystal-melt interfacial free energies of fcc metals. Journal of Crystal Growth. 366, (Mar. 2013), 82–87. [12] Cho, E., Ko, J.-J., Ha, H.Y., Hong, S.-A., Lee, K.-Y., Lim, T.-W. and Oh, I.-H. 2003. Characteristics of the PEMFC Repetitively Brought to Temperatures below 0°C. Journal of The Electrochemical Society. 150, 12 (2003), A1667. [13] Yao, Y., Zhou, H., Liu, W., Li, C. and Wang, S. 2021. The Effect of Cooling Rate on the Microstructure and Macroscopic Properties of Rice Bran Wax Oleogels. Journal of Oleo Science. 70, 1 (2021), 135–143. [14] Petzold, G. and Aguilera, J.M. 2009. Ice Morphology: Fundamentals and Technological Applications in Foods. Food Biophysics. 4, 4 (Oct. 2009), 378–396. [15] Guo, J., Sun, W., Xu, Y., Lin, W. and Jing, W. 2022. Damage Mechanism and Modeling of Concrete in Freeze–Thaw Cycles: A Review. Buildings. 12, 9 (Aug. 2022), 1317. [16] Hou, J., Yu, H., Zhang, S., Sun, S., Wang, H., Yi, B. and Ming, P. 2006. Analysis of PEMFC freeze degradation at -20textdegreeC after gas purging. Journal of Power Sources. 162, 1 (Nov. 2006), 513–520. [17] Uchimura, M. and Kocha, S.S. 2007. The Impact of Cycle Profile on PEMFC Durability. ECS Transactions. 11, 1 (Sep. 2007), 1215–1226. Keywords: EFCF2025, H2, LowTemp. Fuel Cells & Electrolysers, Freeze thawn cycle, Ice, Heavy transport Remark: This work is licensed under Creative Commons Attribution 4.0 International