Dynamic Operation of an AEM Electrolysis System

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17244123 A0806=A1811 Page 1/5 A0806=A1811 Dynamic Operation of an AEM Electrolysis System August Gladik* (1,3), Marc Riedel (1), Rüdiger-A. Eichel (2,3) (1) Robert Bosch GmbH, Renningen/Germany; (2) Fundamental Electrochemistry (IET-1), Institute of Energy Technology, Forschungszentrum Jülich GmbH, Jülich/Germany; (3) Chair of Materials and Processes of Electrochemical Energy Storage and Conversion, RWTH Aachen University, Aachen/Germany; *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Today, alkaline electrolysis (AEL) and proton exchange membrane (PEM) electrolysis represent the most relevant low temperature electrolysis technologies. However, anion exchange membrane (AEM) electrolysis is emerging as a promising technology that combines the advantages of the alkaline environment with the advantages of using an active membrane instead of a diaphragm. For instance, both the use of rare iridium in the anode and fluorine-containing membranes can be avoided with this technology. As the share of renewable energy in the electricity grid is increasing, the fluctuation of the electricity supply is getting more prominent. Consequently, the ability of an electrolyzer to adapt and meet these requirements of transient operation while maintaining a low degradation is of great importance for the viable success of hydrogen production in future energy systems. This study examines the experimental start-stop operation of an AEM electrolysis stack operated in a system environment. The measurements involve parameter variation and the optimization of the cold start procedure to decrease start-up duration while maintaining low degradation. Analyses with 150 cold starts per parameter set reveal that the degradation of the AEM electrolysis stack correlates with the steepness of the current slope during the voltage ramp-up. By applying a gentle start procedure, a degradation of 10 µV start-1 could be observed, while applying a substantially faster current increase results in degradation rates of 26 µV start-1 leading to a significantly decreased lifetime of the electrolyzer. The primary objective is to develop an in-depth understanding of the capabilities and limitations of an AEM electrolysis system regarding the entire transient operation including start and stop procedures. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17244123 A0806=A1811 Page 2/5 Introduction The global transition towards sustainable energy sources intensifies the focus on efficient hydrogen production technologies. Green hydrogen, produced via water electrolysis, is a vital component in decarbonizing various sectors, including industry, energy storage and transport applications. Among the different electrolysis technologies, anion exchange membrane water electrolysis (AEMWE) emerges as a promising method due to its advantages, such as operating in alkaline conditions and eliminating the need for rare and costly iridium. AEMWE employs a hydroxide ion-conducting membrane, allowing for differential pressure operation and enhancing resilience to pressure fluctuations. This capability is particularly advantageous in applications with renewable energy sources, such as wind and solar. The intermittent nature of these energy sources necessitates electrolyzers that can operate in transient conditions, responding to fluctuating power inputs [1]. Cold start behavior, defined as the performance of an electrolyzer starting from ambient temperature and zero voltage, is an important aspect of transient operation. It is thus critical for the effective integration of AEMWE systems into renewable energy grids. Understanding the impact of various starting parameters - such as voltage ramp-up rates, target voltages, and heating strategies - on the cold start process and subsequent degradation is essential for optimizing system performance and longevity. This study investigates these factors through a series of experiments designed to quantify their effects on the cold start behavior and degradation of an AEMWE stack. 1. Experimental 1.1. Test Bench The experiments are conducted using a custom-built AEMWE system test bench, which includes a stack of 19 cells, each with an active area of approximately 60 cm². This results in a stack power output of approximately 2.2 kW. The anion exchange membrane utilized is 100 µm thick and incorporates quaternary ammonium groups for ionic conductivity. The anode is utilizing a Co₃O₄-based catalyst, while the cathode employs a platinum catalyst. The system is designed to operate under ambient pressure on the anode side and at 30 bar on the cathode side. On the anode side, potassium hydroxide (KOH) electrolyte is circulated by a centrifugal pump at a flow rate of 2.5 mL cm⁻² min⁻¹ and a temperature of 50 °C. The test bench is equipped with various sensors and control systems to monitor and adjust operational parameters, including temperature, pressure, and current density. The setup also includes a data acquisition system to record real-time measurements of each sensor. Electrochemical impedance spectroscopy (EIS) of up to 17 individual cells is possible. 1.2. Start Procedure The start procedure is divided into four distinct phases: Conditioning, voltage ramp-up, temperature-driven current increase, and temperature increase. Conditioning phase: Initially, the electrolyte pump and electric heater are activated for 90 seconds to ensure proper hydration of the membrane. This step is crucial to create constant conditions prior to each starting process for high quality measurements but can potentially be shortened or omitted for industrial applications. Voltage ramp-up: Following conditioning, the voltage is rapidly increased to the reversible cell voltage of 1.23 V within 0-5 seconds. After reaching the reversible voltage, the voltage is gradually ramped up to the target voltage, which varies between 1.7 V and 2.1 V, over a specified ramp-up duration (ranging from 12 to 240 seconds). EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17244123 A0806=A1811 Page 3/5 Temperature-driven current increase: Once the target voltage is reached, the system operates in a potentiostatic mode, allowing the current to increase to the targeted current density of 0.85 A cm⁻² due to the temperature rise in the system. Temperature increase: After achieving the target current density, the electric heater is turned off, and the system continues to operate in a galvanostatic mode. The temperature is allowed to rise to the nominal operating temperature of 50 °C, at which point the electrolyzer is considered to be in steady-state operation. 1.3. Degradation Analysis Degradation analyses are performed through repetitive start-stop cycles, with each cycle consisting of a start procedure (see section 1.2) followed by 10 minutes of full load operation. The current is then reduced linearly to zero, and the AEMWE stack is cooled to ambient temperature. The stack voltage is intentionally discharged to 0 V to simulate a cold start after extended downtime. EIS is employed to monitor changes in cell resistance throughout the degradation analysis. EIS measurements are conducted at various points during the start-stop cycles to assess the impact of operational parameters on the electrochemical performance of the AEMWE stack. The data obtained from EIS are analyzed using equivalent circuit models as described by Gladik et al. [2] to quantify the resistive and capacitive components of the cell, providing insights into the degradation phenomena. 2. Results and Discussion 2.1. Cold Start Behavior Cold starts are performed as described in section 1.2 to determine the influence of ramp-up duration, target voltage, and heating strategy on the start duration. The results indicate that the parameters influence the cold start behavior of the AEMWE system in varying degrees. A faster ramp-up duration (12 s) results in a reduced start duration of approximately 454 s, compared to 606 s for a slower ramp-up (240 s). However, faster ramp-ups also lead to increased voltage overshoots of single cells inside the stack and the effect on the start duration is limited. The target voltage during the start procedure plays a crucial role. Increasing the target voltage from 1.7 V to 1.9 V reduces the start duration by 74 %, demonstrating that higher voltages facilitate quicker transitions to full load operation. However, higher voltages are generally associated with increased degradation rates [3]. Heating with an external heater reduces the start-up duration. However, the temperature increase in the system remains the most sluggish process during the starting process and the installed power of the external heater is generally limited by economic considerations. The study shows that the energy requirement for the heater is always higher than the energy saving due to the higher operating temperature of the electrolyzer. The comparatively slow temperature increase of the system can ultimately only be avoided by selecting a higher target voltage. 2.2. Degradation Analysis As part of the degradation analyses, the target voltage is varied between 1.9 V and 2.1 V and the ramp-up duration between 12 s and 120 s in order to analyze degradation effects caused by fast starting processes. For each parameter set, 150 cold starts are performed (Fig. 1). The increase in cell voltage is shown as degradation per start, with the lowest degradation of 10 µV start-1 occurring at the lower target voltage (1.9 V) and the longer ramp-up duration (120 s). A target voltage of 2.1 V and a ramp-up duration of 120 s results in a degradation of 11 µV start-1. A target voltage of 1.9 V and a ramp-up duration of 12 s results in a degradation rate of 15 µV start-1. The highest degradation of 26 µV start-1 occurs at a higher target voltage (2.1 V) and shorter ramp-up duration (12 s). It can be seen that EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17244123 A0806=A1811 Page 4/5 both a higher target voltage and a shorter ramp-up duration lead to a higher degradation rate. A more in-depth analysis of the data reveals a dependence of the degradation on the maximum increase in current during the starting process [2]. These findings suggest that while higher target voltages can accelerate the startup process, they also contribute to increased degradation, highlighting the need for a balanced approach in optimizing start procedures. As start-stop cycles are expected to increase degradation, the low degradation values observed during the degradation analyses show a high ability of AEMWE systems for intermittent operation [4]. Degradation targets for 2030 for AEMWE stacks are below 15 μV h−1 during stationary operation [5]. Fig. 1: Degradation analyses with 150 start-stop cycles of the AEMWE system and a variation of the target voltage Utarget and the ramp-up duration tramp-up during startup. Based on [2]. The EIS measurements provide further insights into the degradation mechanisms occurring within the AEMWE stack. The analysis of the Nyquist plots reveals distinct changes in the resistive components of the cells over the course of the degradation analyses. The highfrequency resistance (HFR), represented by R₀ in the equivalent circuit model, exhibits a decrease of approximately 1.5 mΩ cm² during the degradation analyses. This reduction in resistance may indicate improved ionic conductivity or reduced contact resistance within the cell. Conversely, the resistances associated with the electrode processes, R₁ and R₂, show an increase, suggesting that degradation is primarily occurring at the electrodes. Specifically, R₁, which corresponds to the hydrogen evolution reaction (HER), increases by 2.3 mΩ cm², while R₂, associated with the oxygen evolution reaction (OER), increases by 4.2 mΩ cm². This trend indicates that the degradation mechanisms are more pronounced at the anode, where the OER is kinetically more demanding. The resistance R₃ with the lowest frequency, which may represent mass transport and concentration gradient phenomena, decreases by an average of 2.4 mΩ cm². This decrease could be attributed to improved mass transport pathways within the electrode structure, potentially resulting from mechanical changes during the degradation process. The findings of this study highlight the complex interplay between operational parameters and the performance of AEMWE systems during transient operation. The ability to optimize EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17244123 A0806=A1811 Page 5/5 the cold start behavior while minimizing degradation is crucial for the successful integration of AEMWE systems into renewable energy applications. The results suggest that the target voltage has the most extensive impact to optimize the start duration. Optimizing the start procedure involves a careful balance between start duration and degradation. Additionally, the management of current slopes during startup is critical. Implementing a controlled current increase strategy, rather than a purely voltagecontrolled approach, may help mitigate degradation. 3. Conclusion and Outlook This study provides valuable insights into the startup behavior of an AEMWE system under various operational parameters. The findings indicate that both the ramp-up duration and target voltage significantly affect the start duration and degradation rates. While faster rampups and higher target voltages can reduce startup times, they also lead to increased degradation. However, it is shown that the current ramp has a significantly more detrimental influence on lifetime than just a high target voltage, which emphasizes the necessity for a careful optimization of the chosen parameters. Moreover, the analysis underscores the importance of managing the maximum current increase during startup to mitigate degradation while maintaining short start durations. The results suggest that AEMWE systems exhibit robustness against cyclic operation, with degradation rates remaining within industrially relevant limits (10 µV start-1 – 26 µV start-1) even after extensive start-stop cycles. Future work will focus on the shutdown procedures and dynamic power fluctuations as complementary aspects of intermittent operation. The sensitivity towards operating parameters during these processes and their effect on the degradation of the AEMWE system will be further investigated to ultimately be able to represent the holistic dynamic operation. References [1] V. Martinez Lopez, H. Ziar, J. Haverkort, M. Zeman, O. Isabella, Dynamic operation of water electrolyzers: A review for applications in photovoltaic systems integration, Renewable and Sustainable Energy Reviews, 2023 [2] A. Gladik, M. Riedel, R.-A. Eichel, Anion exchange membrane electrolysis at work – investigating impact of starting parameters and start-stop operation on cold start behavior and degradation, Journal of Power Sources, 2025 [3] S.M. Alia, S. Stariha, R.L. Borup, Electrolyzer durability at low catalyst loading and with dynamic operation, Journal of The Electrochemical Society, 2019 [4] A.K. Niaz, J.-Y. Park, H.-T. Lim, Operational parameters correlated with the long-term stability of anion exchange membrane water electrolyzers, International Journal of Hydrogen Energy, 2021 [5] Clean Hydrogen Joint Undertaking, Strategic research and innovation agenda 20212027, 2022 Keywords: EFCF2025, H2, LowTemp. Fuel Cells & Electrolysers, AEM, Start-stop, Degradation, Dynamic operation Remark: This work is licensed under Creative Commons Attribution 4.0 International