H2O2 formation and decomposition under PEMWE conditions

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 1/8 A0202 H2O2 formation and decomposition under PEMWE conditions Martin Ise*, Baran C. Erer, Elvira Fernandez Sanchis, Markus Ungerer Siemens Energy Global GmbH & Co. KG, SE TI SES PRM CD ECH, Schuckertstr. 2, 91058 Erlangen/Germany *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Under polymer electrolyte membrane water electrolysis (PEMWE) conditions, H2O2 is an unstable side product. As a reactive species, H2O2 can attack materials used in PEMWE systems. Especially, it is harmful for polymer electrolyte membranes in combination with Fenton active cations of metals like Fe. Moreover, some polymer and metal materials are unstable in H2O2 solutions. Therefore, it is important to understand the effects caused by H2O2 and possibilities for keeping its concentration low. In order to find a way for reducing the attack of H2O2 on PEMWE materials, decomposition mechanisms of H2O2 were analyzed. Experiments for catalyst enhanced H2O2 decomposition were conducted with Pt and Pd catalyst particles in H2O2 solutions. In agreement with literature data, faster initial decomposition of H2O2 was found for Pt. As an additional parameter, the effect of electrochemical potential on the decomposition rate will also be discussed. For effective removal of H2O2 out of the PEMWE process water loops, a catalyst bed containing a suitable amount of H2O2 decomposition catalyst is needed. In this direction, lab experiments with catalyst coated mesh materials and with a catalyst containing ion exchange resin (DuPontTM AmberTecTM UP4000Pd OH) were conducted. The results of flow through experiments with this ion exchange resin are shown in Fig. 1. Next steps are the optimization of the H2O2 decomposition catalyst bed design and stability analyses of the catalyst bed. Figure 1: H2O2 decomposition ratio as function of space velocity in catalyst bed filled with DuPontTM AmberTecTM UP4000Pd OH. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 2/8 Introduction Hydrogen peroxide (H2O2) can form as an undesired byproduct under PEM water electrolysis conditions, which can cause degradation of system components. It is widely considered the main contributor to chemical degradation of the membrane over time, in presence of Fenton active species [1]. Due to its corrosive properties, high concentrations of H2O2 in the process water could lead to increased metal dissolution rates from metal components of PEM water electrolyzer plants. Additionally, the ion exchangers that are present in the water loops are chemically less stable in H2O2 solutions. Overall, these factors cause a lifetime decrease of PEM water electrolyzer components in the long term. H2O2 and related radical species can be formed during cell operation which can then attack the chemical bonds of perfluoro sulfonic acid membranes leading to chemical degradation either by chain unzipping (by releasing 2 HFs while shortening the backbone by one CF2 unit) or by direct scission of side chain chemical bonds. The degree of this process is commonly measured by monitoring fluoride release rates with higher rates being associated with shorter membrane lifetime [2]. This in turn leads to protonic conductivity loss and membrane thinning. Similar phenomena of chemical bond cleavage also apply to the ion-exchange resins, resulting in a reduction of their lifespan. Significant H2O2 concentrations were measured especially in PEMWE cathode water loops. Fig. 2 shows the average H2O2 concentrations measured from the cathode water loops of 300 cm² PEM electrolyzer test stands operated by Siemens Energy during over 20 test runs conducted with atmospheric, 10 bar and 35 bar operating pressure, for membranes with and without gas recombination catalysts (GRC). Figure 2: Averaged measured H2O2 concentrations in laboratory test stands from the cathode side water loop at different operating pressures, including standard error bars. The H2O2 concentration in the process water is desired to be lower than 0.1 mg/l, however much higher concentrations were occasionally measured, as indicated by Figure 1. A significant increase in H2O2 concentrations with rising operating pressure is evident, with the average of the samples extracted at 35 bar being approximately 4 mg/l. Reaching elevated operating pressures is a desired condition for future generations of PEM water electrolysers, which highlights the importance of investigating ways to achieve H2O2 removal from the process water loops. This issue was aimed to be tackled by this study. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 3/8 1. Literature Study H2O2 is a commodity chemical which is extensively used as an oxidizing agent in various industries such as paper and pulp industry, and in catalytic oxidation processes such as propene oxide synthesis from propene [3-4]. Although in trace amounts, some H2O2 is present even in DI water, typically formed as result of the ultraviolet light treatment to remove organic impurities [5]. H2O2 presence is undesired in applications that require the use of highest purity water, such as the semiconductor manufacturing or PEMWE, due to its oxidative and corrosive properties. Yano et al. reported that even H2O2 concentrations of 10-40 µg/L accelerate copper dissolution rate in water significantly [5], which emphasizes that the effects of H2O2 can be strong even at extremely low concentrations. a) H2O2 formation H2O2 can be formed chemically in presence of H2 and O2 as reactants by the direct synthesis of hydrogen peroxide (DSHP) process. In PEMWE, this condition may arise due to gas crossover through the membrane. The DSHP reaction proceeds as following [3]: DSHP: H2 + O2 → H2O2 (∆G0298K = −120.4 kJ/mol) Pd based catalysts have been extensively studied and have been recognized as the most effective transition metal catalyst for the DSHP reaction by numerous studies, although considered prototypical due to the limited selectivity towards H2O2 caused by consecutive H2O2 decomposition and hydrogenation reactions [3-4]. Considering that Pt has similar properties to Pd regarding electronegativity and lattice parameters, it would also be expected to show catalytic activity towards DSHP which has been previously demonstrated, even though with less selectivity [6]. Typical state-of-the-art PEMWE cathode catalyst layers are Pt based, and the presence of trace Pt in water at the cathode side is possible due to dissolution or particle detachment. Pt is also incorporated as a GRC in PFSA membranes, particularly in stacks that operate at elevated pressures, to promote H2 and O2 recombination to H2O and limit H2 cross-over to the anode. This does not mean, however, that the DSHP reaction could be disregarded, which thermodynamically could also occur on the GRC. H2O2 can also be formed electrochemically, either by partial O2 reduction reaction (Partial ORR) or H2O oxidation under suitable conditions, which are all theoretically present in a PEMWE cell. Partial ORR and H2O oxidation reactions proceed as following [7]: Partial ORR: O2 + 2 H+ + 2 e− → H2O2 (E0 = 0.695 V) H2O Oxidation: 2 H2O → H2O2 + 2 H+ + 2 e− (E0 = 1.763 V) From the given equilibrium half cell potentials it follows that significant electrochemical H2O2 formation would be expected only in case of not well performing anodes operating at voltages above 1.763 V, while thermodynamically, such formation is always possible at the cathode. b) H2O2 decomposition H2O2 can be decomposed by several chemical reactions. The most widely known H2O2 decomposition mechanism, so called the Fenton reaction has been discovered over 100 years ago by H.J.H. Fenton and denotes the decomposition of H2O2 ions homogeneously catalysed by ferrous ions. It is also occurring with other Fenton active ions. The Fenton reaction generates highly reactive OH∙ radicals as products and is therefore widely utilized in applications such as wastewater treatment that require oxidative destruction of EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 4/8 contaminants. This reaction causes degradation of PFSA membranes by radical attacks and should be avoided in PEMWE cells. The overall reaction proceeds as Fenton Reaction: H2O2 + Fe2+ → Fe3+ + OH− + OH∙ Moreover, chemical H2O2 decomposition is possible by two reaction equations which do not produce OH∙ radicals, namely the H2O2 disproportionation and hydrogenation reactions [3]. Whether radicals are generated as intermediates depends on the intermediate reactions taking place. Disproportionation: H2O2 → H2O + 1/2 O2 (∆G0298K = −116.7 kJ/mol) Hydrogenation: H2O2 + H2 → 2 H2O (∆G0298K = −353.8 kJ/mol) High H2O2 disproportionation rates are especially found for platinum group metals. Specific first order reaction rate constants reported for platinum, iridium and palladium are summarized in Table 1 [9-10]. Due to differences in the experimental conditions the results are not identical, but the same order of catalyst efficiency (Pt > Ir > Pd) was found in the two given references. Table 1: First order specific rate constants of platinum, iridium and palladium for H2O2 disproportionation reaction reported in the literature Catalyst k1 (s-1cm-2) [9] k1 (s-1cm-2) [10] Platinum 3.23 x 10-7 6.4 x 10-6 Iridium 1.03 x 10-7 1.2 x 10-6 Palladium 3.67 x 10-8 3.4 x 10-7 Another possible way of H2O2 removal is its electrochemical decomposition. This could occur at the electrolysis cell electrodes or possibly by introducing an additional electrochemical cell/stack in the process water loops. This process would involve the utilization of the reverse reactions of the partial ORR and H2O oxidation reactions described above as electrochemical H2O2 formation mechanisms with the same equilibrium potentials [7]: H2O2 Oxidation: H2O2 → O2 + 2 H+ + 2 e− (E0 = 0.695 V) H2O2 Reduction: H2O2 + 2 H+ + 2 e− → 2 H2O (E0 = 1.763 V) Considering the concentration of H2O2, the equilibrium potentials for the equations above are only valid for standard conditions (1 mol/l H2O2). Lowering the H2O2 concentration results in an increase of the H2O2 oxidation potential resp. a decrease of the H2O2 reduction potential of approximately 30 mV per order of magnitude [7]. For a concentration of 1 µmol/l H2O2 (34 µg/l), this results in an equilibrium potential of 0.875 V for the H2O2 oxidation and an equilibrium potential of 1.583 V for the H2O2 reduction. Under PEMWE operation conditions, H2O2 oxidation is always expected at the anode and H2O2 reduction is always expected at the cathode. As suggested by Liu et al. [11] an electrochemical cell fed with 0.5 mol/l H2O2 (in water with 0.5 mol/l H2SO4) at the anode and operating at approximately 1 V could utilize the EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 5/8 H2O2 oxidation reaction at the anode and the hydrogen evolution reaction at the cathode with more than 40% faradaic efficiency. However, as was shown by Katsounaros et al. [7] the decomposition of H2O2 with an initial concentration of 2∙10-3 mol/l on a Pt electrode was almost independent of the applied anodic potential in the range of 0.5 - 1.3 V. Therefore, the chemical H2O2 decomposition reaction appeared to be the dominant decomposition pathway. For this reason, we focus on chemical H2O2 decomposition reactions in the following chapters. 2. Experimental Analyses a) Measurements of H2O2 decomposition by catalyst powders in solution In order to measure the H2O2 decomposition efficiencies of available catalyst powders, palladium black (Heraeus, Type 100-11) and platinum black (Heraeus, Type 600-02) powders were used as catalysts for the kinetic evaluation with the specific surface areas of 25.8 m2 /g and 23.5 m2 /g respectively. 1 l of H2O2 solution with an aimed initial concentration of 17 mg/l was prepared in ultrapure water and the initial concentration was verified. 50 mg of the respective catalyst powder was added in the solution to start the reaction. The solution was stirred at 1000 rpm on a magnetic stirrer throughout the reaction time, at room temperature. Water samples were extracted by filtering through 0.22 micrometer pipette filters. The total reaction time was 1 hour, and the volume of each sample extracted was 10 ml, equal to 1% of the total reaction volume. H2O2 concentrations in the samples were measured photometrically using the SupelcoTM SpectroquantTM H2O2 test (MerckTM) immediately after extraction to avoid measuring inaccurate data caused by further self-decomposition of H2O2 or possibly by remaining catalyst particles in the sample. The photometric test had a measurement window between 0.015 - 3 mg/l. To measure higher concentrations than 3 mg/l, respective samples were diluted before measurement and the measured concentrations were then adjusted by multiplying with the dilution ratio. b) Measurements of H2O2 decomposition in flow-through setup Further tests of H2O2 decomposition were conducted using a flow-through setup. These tests consisted of the evaluation of the H2O2 decomposition by using an ion exchange resin loaded with Pd and on the other hand, the H2O2 decomposition by using Pt coated mesh materials. A flow-through setup was constructed in the laboratory for the evaluation of catalysts at varying flow conditions. For the Pd-loaded anion exchange resin DuPontTM AmberTecTM UP4000Pd OH, a fritted glass tube with 10 mm inner diameter was used as a catalyst bed. The bed volume (BV) was 3.7 ml (47 mm bed height). H2O2 solutions were fed from the top of the bed, in a top-to-bottom flow configuration. A peristaltic pump was used to adjust the flowrate to the desired value. The solution had no contact with metal components that might also cause rapid H2O2 decomposition. A pressure gauge was installed to the inlet of the catalyst bed to evaluate the pressure drop. The setup was not in a loop configuration and separate borosilicate glass reservoirs were used as feed and waste. For analyzing the influence of temperature, the temperature of the H2O2 solution feed reservoir was adjusted by a heating plate. The setup was also used for tests with Pt coated mesh materials, these runs were conducted using a polymer cylinder with an inner diameter of 37 mm as the catalyst bed. The bed volume was 13 ml (12 mm bed height) in this case. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 6/8 3. Results a) Kinetic evaluation of H2O2 decomposition by catalyst powders in solution The measured H2O2 concentrations from the decomposition experiments with catalyst powders are shown in Fig. 3. Assuming first order reaction kinetics and considering the specific surface area of the catalyst, the reaction rates were determined to be kPt Black = 1.13∙10-7 s-1cm-2 and kPd Black = 3.79∙10-8 s-1cm-2. Compared to the values obtained by McKee [9], the rate for Pd is similar whereas the rate for Pt is almost by a factor of 3 lower (see Table 1). The higher reaction rate of Pt is in agreement with the literature data; however, it was found that for longer reaction times there was no further decrease of the H2O2 concentration in our experiments with Pt catalyst powder in solution, partly explainable by an effect of traces of Pt catalyst in the photometer test. Therefore, decomposition of H2O2 in concentrations below 3.5 mg/l must be further investigated for Pt containing solutions. Figure 3: Decrease of H2O2 concentration vs. time curves for Pt black and Pd black. b) Measurements of H2O2 decomposition in flow-through setup The results of H2O2 decomposition in the flow-through measurements with the Pd-loaded anion exchange resin DuPontTM AmberTecTM UP4000Pd OH and 1 mg/l initial H2O2 concentration at 60 °C are shown in Fig. 1. The plotted decomposition ratios show a significant effect of the applied space velocity. Compared to a remaining H2O2 concentration of 90 ppb at 2323 BV/h, the H2O2 concentration is strongly reduced to 40 ppb at 353 BV/h. Even higher H2O2 decomposition ratios could be achieved by lowering the space velocity. Further experiments showed that the influence of temperature was moderate (decrease from 91 % decomposition at 60 °C to 84 % at 25 °C for the highest space velocity 2323 BV/h). Additional test runs at 25 °C showed that in the initial H2O2 concentration range of 0.5-4 mg/l the results for the decomposition ratio varied only by a few % (for space velocities of 353 and 2323 BV/h). Short term accelerated aging tests with this resin in a solution with 100 mg/l H2O2 at 80 °C for 7 h did not show resin aging effects by optical microscopy. However, further testing is necessary to prove sufficient long-term stability of the analyzed material under PEMWE conditions. Further flow-through experiments were conducted with Pt coated mesh materials. Initial trials with these materials resulted in significantly lower H2O2 decomposition ratios compared to the Pd-loaded resin analysis described above. Further analyses including geometrical optimization and long-term testing are necessary in this direction. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 7/8 4. Conclusions The following conclusions were drawn from the conducted studies: • H2O2 concentrations build up in PEMWE operation, these should be reduced to avoid H2O2 attack to the used materials. • Faster H2O2 decomposition was found with Pt black than with Pd black in solution, however, decomposition rates with Pt below 3.5 H2O2 mg/l must be further investigated. • Over 90% H2O2 decomposition is possible with DuPontTM AmberTecTM UP4000Pd OH at 60 °C with flow rates up to 2323 BV/h. • Further analyses are necessary for showing effective applicability and long-term stability of catalyst beds for decomposition of H2O2. Acknowledgement The authors gratefully acknowledge the financial support by the German Federal Ministry of Education and Research (BMBF) within the H2Giga project DERIEL (grant number 03HY122A). References [1] Pierre Millet, Degradation Processes and Failure Mechanisms in PEM Water Electrolysers. In “PEM Electrolysis for Hydrogen Production - Principles and Applications”, Taylor & Francis Group, 2016. [2] Ahmet Kusoglu, Adam Z. Weber, New Insights into Perfluorinated Sulfonic-Acid Ionomers. Chem. Rev. 2017, 117 (3), 987–1104. [3] Anthony Plauck, Eric E. Stangland, James A. Dumesic, Manos Mavrikakis, Active Sites and Mechanisms for H2O2 Decomposition over Pd Catalysts. Proc. Natl. Acad. Sci. 2016, 113 (14), E1973–E1982. [4] David A. Crole, Simon J. Freakley, Jennifer K. Edwards, Graham J. Hutchings, Direct Synthesis of Hydrogen Peroxide in Water at Ambient Temperature, Proc. R. Soc. A 472: 20160156, 2016. [5] Daisaku Yano, Masami Murayama, Masao Takahashi, Hikaru Kobayashi, Koji Yamanaka, Inhibition of Copper Corrosion by Removal of H2O2 from CO2-Dissolved Water Using Palladium Catalysts, ECS Transactions, 58 (6) 151-158, 2013 [6] H. W. Lee, H. Nam, G.-H. Han, Y.-H. Cho, B. C. Yeo, M.-C. Kim, D. Kim, K.-Y. Lee, S.Y. Lee, S.S. Han, Solid-Solution Alloying of Immiscible Pt and Au Boosts Catalytic Performance for H2O2 Direct Synthesis. Acta Mater. 2021, 205, 116563 [7] Ioannis Katsounaros, Wolfgang B. Schneider, Josef C. Meier, Udo Benedikt, P. Ulrich Biedermann, Alexander A. Auer, Karl J.J. Mayrhofer, Hydrogen Peroxide Electro-chemistry on Platinum: Towards Understanding the Oxygen Reduction Reaction Mechanism. Phys. Chem. Chem. Phys. 2012, 14 (20), 7384–7391 [8] Yimin Lin, Junlian Qiao, Yuankui Sun, Hongyu Dong, The Profound Review of Fenton Process: What’s the next step?, J. Environ. Sci. 2025, 147, 114–130 [9] D. W. McKee, Catalytic Decomposition of Hydrogen Peroxide by Metals and Alloys of the Platinum Group. J. Catal. 1969, 14 (4), 355–364 [10] G. Bianchi, F. Mazza, T. Mussini, Catalytic Decomposition of Acid Hydrogen Peroxide Solutions on Platinum, Iridium, Palladium and Gold Surfaces. Electrochimica Acta 1962, 7 (4), 457–473 EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476215 A0202 Page 8/8 [11] C. Liu, R. Ding, J. Yang, S. Liu, L. Chen, Q. Xu, J. Li, X. Yin, Low-Voltage Hydrogen Peroxide Electrolyzer for Highly Efficient Power-to-Hydrogen Conversion. ACS Sustain. Chem. Eng. 2023, 11 (6), 2599–2606 Keywords: EFCF2025, H2, Low-Temp. Fuel Cells & Electrolysers, H2O2 formation, H2O2 decomposition, Pt and Pd catalysts, Pd-loaded ion exchange resin Remark: This work is licensed under Creative Commons Attribution 4.0 International