Performance recovery of electrolysis cells

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 1/10 B1002 Performance recovery of electrolysis cells Martin Höglinger* (1), Julius Rauh (1), Joshua Eder (1), Oliver Mogg (1), Stefan Kartusch (1), Rebekka Köll (1), Marie Macherhammer (1), Alexander Trattner (1,2) (1) HyCentA Research GmbH, Graz/Austria; (2) Institute of Thermodynamics and Sustainable Propulsion Systems; Graz University of Technology, Graz/Austria; *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Proton Exchange Membrane (PEM) electrolysis is a key technology for sustainable hydrogen production, offering high efficiency and compact design. However, despite rapid advancements, performance degradation remains a critical challenge with increasing cell voltages over time. These degradation effects impact long-term operation, but some of these losses are reversible and can be recovered through specific operation conditions. Addressing this issue is crucial for enhancing the durability and efficiency of PEM electrolyzers. This paper presents an investigation into operational conditions that can reset or mitigate reversible degradation, based on extensive experimental studies on stack level. We propose a systematic approach for conditioning procedures designed to stabilize electrochemical performance and minimize performance variability across multiple operational cycles. Various conditioning methodologies were assessed for their reproducibility and influence on reversible degradation mechanisms. The findings provide valuable insights into the mechanisms driving reversible degradation and demonstrate effective strategies for mitigating these effects. By implementing optimized conditioning procedures, we demonstrate improvements in cell performance recovery and long-term stability. These insights are not only critical for real-world PEM electrolyzer applications, where consistent performance is required, but also contribute to the development of standardized, reproducible testing protocols that enhance the reliability of long-term performance evaluations in laboratory and industrial settings. Figure 1: Cell voltage increase over time and after different operation modes EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 2/10 Introduction Electrochemical water splitting for hydrogen production represents a critical technology for the renewable energy transition, offering a pathway to convert surplus renewable electricity into chemical energy storage through green hydrogen generation. However, the widespread deployment of electrolysis technologies, particularly proton exchange membrane water electrolysis (PEMWE), faces significant durability challenges that directly impact the economic viability and operational reliability of these systems. [1,2,3] The degradation phenomena in electrolysis systems are complex and multifaceted, encompassing both irreversible and reversible performance losses that occur across different temporal scales and under varying operational conditions. While irreversible degradation mechanisms such as catalyst dissolution, membrane thinning, and permanent structural changes have been extensively studied, the understanding and mitigation of reversible degradation phenomena have emerged as equally critical for optimizing electrolyzer lifetime and performance. [4,5] Reversible degradation in electrolysis refers to performance losses that can be recovered through specific operational procedures or operating conditions without requiring component replacement. These phenomena are particularly significant because they represent a substantial portion of the total voltage degradation observed during electrolyzer operation, and their mitigation can extend system lifetime while maintaining efficiency. [4,6] Recent research has identified several key mechanisms contributing to reversible degradation across different electrolysis technologies. In PEMWE systems, catalyst surface oxidation emerges as a primary reversible degradation mechanism, where iridium catalysts undergo oxidation under high electrode potentials, leading to reduced catalytic activity that can be recovered through reductive treatments or low voltage operations [4,6,7]. Membrane contamination represents another critical reversible degradation pathway, where metal ion impurities from system components or feed water accumulate in the proton exchange membrane, disrupting ionic conductivity and increasing ohmic resistance. This contamination can be partially or fully reversed through chemical cleaning procedures, such as nitric acid treatment, which effectively removes accumulated cations and restores membrane functionality [1,3,8]. The temporal characteristics of reversible degradation vary significantly depending on the underlying mechanism and operational conditions. While catalyst surface oxidation can be recovered within minutes to hours through appropriate operation, membrane contamination may require more extended recovery periods ranging from hours to days. Understanding these temporal scales is crucial for developing effective recovery strategies and optimizing operational protocols to minimize degradation impact. [1-4] Current density and voltage cycling emerge as critical operational parameters influencing reversible degradation rates. High current densities accelerate various degradation mechanisms while simultaneously affecting their reversibility, creating complex optimization challenges for system operators [1,3,10-12]. Dynamic operation, characteristic of renewable energy integration, introduces additional complexity as frequent load changes can either accelerate or mitigate certain degradation mechanisms depending on the specific operational protocols employed. [12-14] The distinction between reversible and irreversible degradation has profound implications for electrolyzer design, operation, and maintenance strategies. Recognition of reversible degradation mechanisms enables the development of recovery procedures that can be integrated into regular operational protocols, but also into conditioning procedures for electrolysis development testing for reproduceable and reliable results. Furthermore, understanding these mechanisms facilitates the optimization of operational parameters to minimize degradation while maximizing recovery potential. This work aims to provide an experimental analysis of reversible degradation mechanisms in electrolysis cells and stacks. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 3/10 1. Scientific Approach The effect of reversible degradation, has remained poorly understood despite its significance for accurate degradation assessments. The distinction between reversible and irreversible performance changes is crucial for proper characterization and comparison of cell performance for material development, as well as degradation and lifetime prediction. This study investigates the phenomenon with different materials and cells, as well as an extended view from single cell to stack setups. However, the analysis in this paper is based on different cell and stack setups, hence materials and design variations occur as described in the chapter [Experiments]. As a first step the presence of short-term effects are investigated on various commercially available membranes, which can be measured in the range of below ten hours. The second step is to investigate the effect during long-term testing at constant current and different load profiles. The operation profiles range from operation at constant conditions, current cycling at different levels and frequent start-stop operation. Current cycling includes profiles suggested with cycling from low to high current densities but also include shut downs. Figure 2 illustrates a schematic view of four distinct Accelerated Stress Test (AST) profiles, labeled AST 1 through AST 4. The y-axis represents current density (A·cm⁻²), while the x-axis indicates time in seconds for a single representative cycle. AST 1 (blue trace) features a square-wave current profile oscillating between 1.5 and 2.0 A·cm⁻², with a cycle period of approximately 20 seconds. This pattern reflects a moderate amplitude cycling condition, representative of dynamic load conditions typically encountered during regular operation. AST 2 (orange trace) displays a similar square-wave profile but at lower current densities, alternating between 0 and ~0.5 A·cm⁻². The lower current density and identical cycle period suggest a mild cycling condition, possibly simulating partial load or idle phases. AST 3 (gray trace) comprises a stepped current profile, beginning with a constant high current density of 3.0 A·cm⁻², followed by a sustained intermediate level 0.3 A·cm⁻², then a return to 0 A·cm⁻². The stepped structure mimics start-up, load, and shutdown phases within a single cycle, simulating more complex operational scenarios. AST 4 (yellow trace) consists of a high-amplitude square-wave profile alternating between 0 and 3.0 A·cm⁻². The waveform is initiated later in the time axis for clarity in this schematic, but in practice, it is applied as a cyclic stress over extended durations. The high current density and sharp transitions represent an aggressive load cycle, suitable for evaluating material durability under severe operational conditions. These profiles are intended to be repetitively applied over multiple hours forming the basis of long-term durability and degradation testing protocols for electrolyzers. However; the diversity of the forms allows the investigation of reversible degradation mechanisms effecting the cells. Figure 2: AST profiles used for this analysis EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 4/10 2. Experiments All shown experiments serve for deepening the understanding of the degradation phenomena are conducted on different testbenches and stack dimensions. Each testbench was equipped with key features including an anode loop with thermal management and continuous water purification. Additionally, temperatures, pressures and water conductivity are measured at the stack inand outlet. The cathode loop is responsible for pressure regulation and features a discontinuous water separation, recorded measurements are the hydrogen pressure and temperature at the stack outlet. A more detailed description of the layout used for the testbenches can be found in [5]. The Units Under Test (UUT) range from single cells with an active area of 25 cm² up to stack level with 16 cells and active areas up to 1500 cm² with different commercially available membranes. Extensive results shown for constant current and AST measurements are conducted with a stack from the EU Project Recyalyse [15]. The presented stack tests are from the second generation build in the course of the project with 14 cells and an active area of 75 cm². The main goal for this stack was the reduction of catalyst loading at the anode side, therefore an IrRu alloy with low loading was used as described in [16]. 3. Results In Figure 3 measurements for different membranes are summarized with operation at 60 °C, different pressure levels and elevated current densities. Due to simplicity the average cell voltage is shown as reversible degradation effects all cells in similar manner. The fluctuations in cell voltage occur due to the temperature and pressure control of the testbench. For each line a short comparison after a shutdown protocol is shown, unfortunately the shut-down procedures are different and range from several minutes to hours. Therefore, the cell voltage at the same operation conditions is shown as indicating factor for reversible degradation. For all shown curves a lower voltage after restarting the electrochemical operation can be seen. The effect of reversible degradation is therefore relevant for various membrane types and cell and stack designs as suggested by different results in literature. [2,4,6] EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 5/10 Figure 3: Operation at elevated current densities for different cells and stacks Testing data of longer operation at constant current is only available for a low number of stacks due to resource limitations. However, one example was the stack used in the EU funded project Recycalyse, which was operated for a period of 200 hours at 1 A∙cm-2 at a pressure level of 5 bar and water inlet temperature of 58 °C. As shown in Figure 4, the cell voltages under constant current conditions increase over the first 50 hours before reaching nearly steady state. This behaviour is congruent to the formation of an oxidation layer as reported in [4], in the case of this study it saturates after approximately 50 hours. The shut down recovers the oxide layer explaining the lower voltage afterwards. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 6/10 Figure 4: Constant operation of a stack at 1 A∙cm-2 - first test After about 200 hours the stack is shut-down for one day and restarted at the same conditions. Figure 5 shows the following operation at 1 A∙cm-2 for about 140 hours. Again, cell voltages show a steeper increase at the beginning until about 30 hours, with a stabilization and slower increase from 50 hours until the end of the load point. Between the 1 A∙cm-2 and 1.5 A∙cm-2 a polarization curve was executed with 0.04 A∙cm-2 as lowest point. However, no complete shut-down occurred and at the higher current density of 1.5 A∙cm-2 no additional voltage increase can be seen. This indicates that operation at lower current densities is insufficient to restore the initial performance. After operating another 50 hours at 1.5 A∙cm-2 the stack is shut down again for approx. one week due to testbench maintenance and then restarted at 1 A∙cm-2. Figure 5: Constant current operation of the stack - second test EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 7/10 After restarting, the cell voltage is lower again as a comparison in Figure 6 shows in a zoomed display for the second test. Figure 6: Comparison of the cell voltages Figure 7 shows the comparison of the first 50 hours for both measurements at same conditions and constant current operation. Both tests show similar behaviors as a voltage increase occurs for the first operation hours except for cell 1, which has a higher cell voltage in phase 1. The average cell voltage shows a similar behaviour as each cell, however, the increase in test 1 cannot be seen as clearly as figure 8 shows. Figure 7: Comparison of the cell voltages for the first 50 hours of two constant current test Comparing the constant current operation with the AST profiles shown as described in Figure 2 indicate that phases without electrochemical operation have an influence on the cell voltage. As all ASTs are based on current cycling only some include a phase at 0 A·cm2. In order to compare the cycles, the average cell voltage at the same current density is depicted over the whole test. Figure 9 shows the results for AST 1 and AST 2, with a slight increase for the average cell voltage at same current density for AST 1. AST 2, AST 3 and AST 4 as seen in Figure 9 and Figure 10 do not show this increase for the average cell voltage. This suggests a reset of the reversible degradation can be achieved by short interrupts to 0 A·cm-2 in electrochemical operation. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 8/10 Figure 9: Comparison of the average cell voltage from AST 1 and AST 2 Figure 10: Average cell voltages for AST 3 and AST 4 Further analysis is required to fully understand the reversible degradation phenomenon, investigate different recovery strategies and to gain deeper insights into its long-term effects on PEM electrolysis. Since testing at the stack level is resource-intensive, additional longterm experiments are planned at the single-cell level, which has proven to be sufficient for this purpose, as indicated by the short-term comparison results. The experimental investigation into reversible degradation mechanisms in PEMWE stacks reveals critical insights into operation strategies. The application of AST demonstrates that reversible voltage losses can be reset through certain operational protocols. Specifically, interruptions in electrochemical operation, such as shutdowns or current density cycling, trigger recovery phenomena by potentially mitigating catalyst surface oxidation and probably gas bubble accumulation. This aligns with prior studies showing that voltage drops below 1.5 V reactivate iridium-based catalysts by reducing surface oxides, thereby restoring anode kinetics [4,6]. The observed voltage recovery after shutdown periods (e.g., 1–7 days) further corroborates the temporal dependency of reversible degradation mechanisms, where longer interruptions enable more complete restoration of catalyst activity. The current study’s focus on short-term reversible mechanisms (≤200 hours) leaves open questions about long-term degradation synergies. For instance, repeated recovery cycles may exacerbate interfacial delamination between catalyst layers and membranes, compounding irreversible losses over time. The findings highlight a promising potential for developing effective recovery strategies to mitigate reversible degradation in PEM electrolysis stacks. By targeting the mechanisms underlying performance losses, such strategies could significantly enhance long-term EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476282 B1002 Page 9/10 operational stability and efficiency. Continued investigation, particularly at the single-cell level, offers a valuable and resource-efficient path toward optimizing recovery protocols, ultimately contributing to improved performance and durability of PEM electrolysis systems. References [1] Wallnöfer-Ogris, E, Grimmer, I, Ranz, M, Höglinger, M, Kartusch, S, Rauh, J, Macherhammer, MG, Grabner, B & Trattner, A 2024, 'A review on understanding and identifying degradation mechanisms in PEM water electrolysis cells: Insights for stack application, development, and research', International Journal of Hydrogen Energy, vol. 65, pp. 381-397. https://doi.org/10.1016/j.ijhydene.2024.04.017 [2] Blair SJ, Wrubel JA, Parrish C, Parimuha MR, Foster J, Pylypenko S et al. Reversible Losses in Proton Exchange Membrane Water Electrolysis. J. Electrochem. Soc. 2025;172(3):34517. https://doi.org/10.1149/1945-7111/adc23f. [3] Tjarks, G., Gibelhaus, A., Lanzerath, F., Müller, M., Bardow, A., & Stolten, D. (2024). A review on understanding and identifying degradation mechanisms in PEM water electrolysis stacks. International Journal of Hydrogen Energy, 65, 381-397. [4] Krenz T, Rex A, Helmers L, Trinke P, Bensmann B, Hanke-Rauschenbach R. Reversible Degradation Phenomenon in PEMWE Cells: An Experimental and Modeling Study. J. Electrochem. Soc. 2024;171(12):124501. https://doi.org/10.1149/1945-7111/ad96e4. [5] Höglinger M, Kartusch S, Eder J, Grabner B, Macherhammer M, Trattner A. Advanced testing methods for proton exchange membrane electrolysis stacks. International Journal of Hydrogen Energy 2024;77:598–611. https://doi.org/10.1016/j.ijhydene.2024.06.118. [6] Suermann M, Bensmann B, Hanke-Rauschenbach R. Degradation of Proton Exchange Membrane (PEM) Water Electrolysis Cells: Looking Beyond the Cell Voltage Increase. J. Electrochem. Soc. 2019;166(10):F645-F652. https://doi.org/10.1149/2.1451910jes. [7] Feng, Q., Yuan, X. Z., Liu, G., Wei, B., Zhang, Z., Li, H., & Wang, H. (2017). A review of proton exchange membrane water electrolysis on degradation mechanisms and mitigation strategies. Journal of Power Sources, 366, 33-55. [8] Park, J., Lee, S., Kim, T., Park, S., You, N. H., Kang, S. Y., ... & Kim, H. T. (2023). Performance recovery of proton exchange membrane electrolyzer cells after artificial soft-water electrolysis. International Journal of Hydrogen Energy, 48(100), 3988239893. [9] García-Valverde, R., Espinosa, N., & Urbina, A. (2012). Simple PEM water electrolyser model and experimental validation. International Journal of Hydrogen Energy, 37(2), 1927-1938. [10] Rauh JF, Höglinger M, Karan B, Steiner J, Macherhammer M-G, Trattner A. Need for Speed: Fast Activation of PEM Electrolyzers for End-of-Line Testing. Meet. Abstr. 2024;MA2024-02(46):3218. https://doi.org/10.1149/MA2024-02463218mtgabs. [11] Trinke, P., Bensmann, B., & Hanke-Rauschenbach, R. (2018). Current density effect on hydrogen peroxide formation and mitigation in PEM water electrolysis. Journal of The Electrochemical Society, 165(7), F502. [12] Alkhaldi, S., Aziz, M., Amrite, A. et al. Parametric study of PEM water electrolyzer performance. J Appl Electrochem 55, 327–343 (2025). https://doi.org/10.1007/s10800-024-02187-9 [13] Rakousky, C., Reimer, U., Wippermann, K., Carmo, M., Lueke, W., & Stolten, D. (2017). An analysis of degradation phenomena in polymer electrolyte membrane water electrolysis. Journal of Power Sources, 326, 120-128. [14] Rex A, Krenz T, Trinke P, Helmers L, Bensmann B, Hanke-Rauschenbach R. Reversible Performance Dynamics in PEM Water Electrolysis: An Electrochemical