Investigating PEMFC performance at 95°C

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 1/11 A1104 Investigating PEMFC performance at 95°C Guillaume Etienne*, Aouatef Ouhammi, Assma El Kaddouri, Julia Mainka, Jérôme Dillet, Olivier Lottin Université de Lorraine, CNRS, LEMTA, F-54000, Nancy/France; *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Increasing the operating temperature of proton exchange membrane fuel cells (PEMFC) is an important step in accelerating their large-scale deployment, especially in heavy-duty transport. However, higher materials degradation rates are expected and water management issues may be more pronounced [1]. For these reasons, it is necessary to study the impact of elevated temperature on the behavior durability of PEMFC membrane electrode assemblies (MEA). First, performances were tested by varying temperature, relative humidity (RH), gas stoichiometry and absolute pressure. With commercial MEA, higher performances were observed with increasing pressure (1 to 3 bar), relative humidity (between 15 to 50% at 95°C) and stoichiometry (Sair between 2 and 3 at 3 bar). However, polarization curves demonstrated lower performances when the temperature was increased in a cell configuration with parallel channels. In addition, polarization curves showed instabilities at high current density, at 95°C and low RH (i.e. 20-30%), with a lower current density limit than at 70°C, where no steep voltage decrease was observed. According to the literature, two mechanisms are possibly responsible for these phenomena: an increase in oxygen dilution due to the excessive water production on the cathode side leading to a voltage drop [2] and/or a severe dehydration of the membrane electrode assembly due to both the electro-osmosis effect [3],[4] and evaporation because of the heat produced by the reaction [5]. To clarify, we performed electrochemical impedance spectroscopy (EIS) to study the high frequency resistance, as an indication of the membrane and the electrodes ionomer hydration. Rhf measurements demonstrate that there is an optimal operating point for all RH at 95°C, most probably due to the increase of heat production and more pronounced electro-osmosis at high current density, leading to MEA dehydration. Conversely, there is no such significant MEA dehydration at 70°C and Rhf remains stable. Additional impedance spectra over the whole frequency range and stable operating points at medium and high current densities are also measured. These results underline the large behavior difference between usual and higher operating temperature. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 2/11 Introduction The large-scale use of proton exchange membrane fuel cells (PEMFC) for typical transport applications such as trains, heavy vehicles, boats etc requires the development of compact systems. One of the ways in which one can accelerate the integration of PEMFC in transport is to increase the operating temperature of the cell. Indeed, this results in a reduction in the size of the heat exchangers, thereby enhancing the system’s compactness. Furthermore, it facilitates operation with impure hydrogen due to enhanced catalytic activity, thereby mitigating the poisoning of the Pt particles by CO, CO2 and other compounds. The result is a more compact system that is better suited to transport applications. The use of impure hydrogen is beneficial in terms of cost-effectiveness, as it is cheaper. PEMFC currently operates at low temperature (LT-PEMFC), i.e. below 80°C but several works have examined the potential benefits and drawbacks of higher operating temperature. PEMFC system with higher operating temperature are currently divided in two categories: high-temperature PEMFC (HT-PEMFC), which operate above 120°C and intermediate-temperature PEMFC (IT-PEMFC), which operate between 80 and 120°C. HTPEMFC are interesting because of their high tolerance to impurities [6],[7], as well as the high facilitation in water and heat management [8],[9],[10]. However, durability issues have been observed [9],[11],[12] and materials development is still required to improve the system. On the contrary, targeting intermediate operating temperature allows the system to earn the benefits of HT-PEMFC such as heat management and tolerance to hydrogen impurities, while keeping well-known materials used in low-temperature fuel cell (LTPEMFC) operating at temperature around 70-80°C, with for example LSC-PFSA (Long Side Chain-PerfluoroSulfonic Acid) membrane. Indeed, it is essential to use these materials at temperatures below 100°C to prevent hydration issues, which can deteriorate the membrane and electrodes properties, including the proton conductivity, and consequently compromise the PEMFC operation. While the benefits to the overall system are well known, to the best of our knowledge, the impact of operating temperature on performance, between 70 and a higher temperature below 100°C, is not clearly identified. A higher operating temperature is expected to prevent flooding and enhance water management. However, this modification is likely to induce a different water hydration state for the electrodes and membrane, which may have a negative impact on performance. To address the question of the impact of operating temperature on PEMFC performance, a study was conducted on the difference in PEMFC performance between usual operation at 70°C and 95°C on a commercial membraneelectrode assembly (MEA) made of state-of-art materials. First, we varied external parameters such as gas stoichiometry and relative humidity (RH). We then proceeded to measure the high frequency resistance, which is representative of the hydration state of the MEA, along the entire polarization curve under various gas RH and temperature conditions to assess the impact of operating conditions at several current density setpoints. 1. Experimental procedure Materials & methods Characterizations were performed using a straight-parallel channels single cell. The bipolar plates include 50 channels, with a depth of 0.15 mm at the anode side and 0.3 mm at the cathode side. They have been crafted from 316L stainless steel, which has EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 3/11 undergone a gold coating process. All external parameters were controlled, such as cell temperature, stoichiometries, pressure and relative humidity of gases. Hydrogen and air were introduced in a counter flow configuration with a constant compression rate of 25% applied to the commercial MEA. The commercial MEA are composed of: • Anode: Pt/HSAC 50 wt% (Tanaka TKK TEC10E50E) - 0.1 mg/cm² • Cathode: Pt/HSAC 50 wt% (Tanaka TKK TEC10E50E) - 0.3 mg/cm² • Nafion D2020CS as the electrode ionomer for both anode and cathode side with an ionomer/carbon ratio equal to 0.8 • Gore788.12 membrane • Freudenberg H14CX653 GDL Each MEA was submitted to a conditioning step at 70°C and RH equal to 50% to enhance its performance. This preliminary step lasts approximately two and a half hours. During this time, potentiostatic steps at Open Circuit Voltage (OCV), 0.6 and 0.3 V were applied while hydrogen and stoichiometry were adjusted respectively to 1.5 and 3. Except for the hydrogen flowrate, all external parameters (T, P, Stair, RH) were modified to assess their impact on MEA performance. Identical experiments were carried out on a cell with serpentine channels. However, given the amount of results, only the results for the cell with straight parallel channels are presented, bearing in mind that the trends between the two geometries are identical. Following analysis of the results, reference conditions were established in line with the optimal performances (3 bar, Stair=3). Variations in RH and cell temperature were then introduced to assess the impact of temperature on MEA humidification. Polarization curves and electrochemical impedance spectroscopy (EIS) Performance mapping characterizations were measured using polarization curves and impedance spectra in galvanostatic mode. Polarization curves were measured after 30 min at a constant current density of 1 A/cm², to avoid any discrepancies caused by a different starting point. The process consists of 100 s steps at constant current density: the current density increases until a voltage of 0.3 V is reached, after which the value of the current density decreases until the OCV is reached. The average voltage is evaluated for each current density step, taking into account only the last 50 s. Following this, polarization curves and impedance spectra were estimated in potentiostatic mode to study the behaviour of MEA near instability points, particularly at high current densities. In a similar manner to the protocol in galvanostatic mode, a 30-minute step at a constant current density of 1.5 A/cm² was first carried out to ensure that any differing states of hydration of the MEA prior to measurement were avoided. The next step in the process consisted in applying a 30-minute step at constant voltage. At this point, the hydrogen and air flow rates were calculated according to the current measurement. This allows us to obtain stoichiometries of 1.25/3. Finally, impedance spectroscopy in potentiostatic mode was performed at the same voltage and gas flow rates. The overall procedure was repeated at different voltages to obtain the polarization curve and also to evaluate the evolution of Rhf as a function of current density. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 4/11 Accelerated Stress Test (AST) Following the performance mapping and Rhf measurements, an accelerated stress test (AST) was carried out, based on previous works from our group [13],[14],[15]. The elementary sequence of the cycle is 60 s in duration and is repeated during 24h. Its effects can be categorized into 3 parts: • For a duration of 43 s, cycling in current density is applied, corresponding to cycling between OCV (3 s) and 0.55 V (2 s) at the beginning of life of the MEA. The process of cycling in potential can result in a reduction in the electrochemical surface area (ECSA), equivalent to the ageing of the catalyst layer (CL) [16],[17]. • Over a period of 17 s, OCV is applied in order to induce chemical ageing of the ionomer in the membrane and in the electrodes [18],[19]. • Mechanical aging is achieved through the application of MEA hydration during the load cycling and MEA dehydration during the OCV stage. It is equivalent to humid/drying cycles, which in turn induce mechanical constraints. The AST was performed at a temperature of 95°C and 30% RH over a period of six days. Characterizations with polarization curves and impedance spectroscopy in potentiostatic mode were performed at two-day intervals. The purpose of this section is to determine whether electroosmotic or thermal effects are responsible for MEA dehydration. As the AST progresses, the current density at a given voltage will decrease, resulting in a lower electroosmotic flux. Following this, we will be able to ascertain whether the phenomenon is less significant, which would further substantiate the hypothesis that the electroosmosis flux is predominantly responsible for the increase in Rhf. 2. Results Impact of operating conditions on MEA performance Operating parameters such as cell temperature, gas RH, pressure and flowrate have been tested to observe their impact on PEMFC performance (错误!未找到引用源。). As mentioned before, identical tests have been carried out on a five-channel serpentine cell. Polarization curves show identical trends but lower performances in every case compared to the straight parallel cell design. T (°C) 70 95 RH (%) 50 80 30 50 P (bar) 1 2-3 1 2--3 1 2--3 1 2--3 Stair 2-2.5-3-4-5 2-2.5-3 2-2.5-3-4-5 2-2.5-3 2-2.5-3-4-5 2-2.5-3 2-2.5-3-4-5 2-2.5-3 Table 1: summary of experiments Note that hydrogen stoichiometry was always kept equal to 1.25. Also, we chose to only present a limited amount of polarization curves given the large amount of data. We can draw up trends in the impact of the parameters within the range tested: • The increase in air stoichiometry induces an increase in voltage for both temperatures at 50% RH. This enhancement in performance is evident across most of the polarization curve at 1 and 2 bar; however, at 3 bar, it is only apparent at high current densities. At 95°C/RH=30% and 3 bar, similar performances are observed EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 5/11 between the different air stoichiometries. Indeed, a high air flow rate will avoid gas starvation and will remove excess produced water at high current density. However, in dry conditions with high operating temperatures and low gas RH, a high air stoichiometry will cause the MEA to dry out, leading to a decrease in performance. • Increasing the gas pressure induces a significant increase in performance, primarily due to enhanced hydration of the MEA. This results in a lower high frequency resistance (Rhf), which corresponds to the hydration state of the membrane. At 95°C/30 and 50% RH, the Rhf measured at 1 A/cm² goes respectively from 168 to 66 and from 70 to 43 mΩ.cm² between 1 and 3 bar (Table 2). • The increase in gas RH improves performance due to enhanced membrane hydration, leading to a lower Rhf at equivalent gas pressure (see Table 2). Zhang et al. [20] obtained a similar trend with a decrease in membrane resistance from 700 to 146 mΩ.cm² at 120°C, between 25 and 100% RH. Note that excess gas RH can also have a negative impact on PEMFC performance by increasing oxygen transport resistance due to the greater amount of water [21]. • The increase in cell operating temperature results in lower performances. Although proton conductivity and catalyst activity are thermally stimulated, Rhf measurements demonstrate that a higher temperature lower the hydration of the membrane, especially at low RH (Table 2). Therefore, since proton conductivity of the membrane is highly related to water content, lower performances are observed, highlighting that MEA hydration plays a preponderant role in achieving optimal FC performances than, for example, to the gain in catalytic activity. Similar results have been reported by Jeon et al. [22] who observed the increase in Rhf at higher temperature and equivalent RH. Figure 1: polarization curves measures in galvanostatic mode: impact of temperature and air stoichiometry (a), impact of gas pressure and relative humidity (b) i (A/cm²) T (°C) RH=20% RH=30% RH=50% 1 bar 2 bar 3 bar 1 bar 2 bar 3 bar 1 bar 2 bar 3 bar 1 70 0,097 0,045 0,036 0,072 0,039 0,034 0,050 0,035 0,033 95 -- 0,177 0,094 0,168 0,096 0,066 0,070 0,054 0,043 Table 2: Rhf measurements with imposed current density at 1 A/cm² and 95°C As shown on the polarization curves at 95°C and low RH in Figure 1, a sudden drop in voltage is observed at high current densities, resulting in significant error bars. To gain a more in-depth understanding of this phenomenon, we conducted an exploratory survey with characterization tests at 95°C, with inlet RH ranging from 20% to 70% under reference conditions (3 bars StH2/air=1.25/3). The reference conditions were determined from our performance mapping. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 6/11 Study of MEA behaviour at 95°C with various gas RH From a theoretical perspective, voltage losses at high current densities can be attributed to a lack of reactant resulting from elevated levels of oxygen dilution in the water produced during the reaction. This is compounded by resistance in oxygen transport, further contributing to the loss of reactant. Therefore, increasing gas RH is intended to increase oxygen dilution, thus limiting FC performance at high current densities. At 70°C, a slight overall improvement in performance is observed across the entire polarization curve between 20 and 50% RH, attributable to enhance MEA hydration. However, this gain is considerably reduced between 30% and 50% RH at medium and high current densities, showing that the increase in MEA hydration is partly counterbalanced by the increase in oxygen dilution. As the temperature increases to 95°C, there is a notable enhancement in FC performance, which is accompanied by an increase in gas RH. This suggests that the sudden drop in voltage is unlikely to be attributable to oxygen dilution but rather to MEA hydration. Figure 2: polarization curves in reference conditions at various RH at 95°C (a) and 70°C (b) To verify the hypothesis that the voltage drop at high current density is related to MEA hydration, we performed impedance spectroscopy at 95°C, at various gas RH levels and at different current densities (错误!未找到引用源。). For all RH, an increase in Rhf with increasing current density was observed. Note that at 20% RH, it was not possible to reach the highest current density region in galvanostatic mode due to voltage instabilities. Figure 3: Rhf measurements at different current densities for different relative humidities While water production is higher at high current density, the MEA appears to dry out. Until 2 A/cm², the water produced appears to contribute to MEA hydration as Rhf values decrease with increasing i from 1 to 2 A/cm², particularly at low RH. To gain more insight into Rhf behaviour, we conducted an impedance spectroscopy study in potentiostatic mode. This enables us to accurately estimate MEA hydration within the range of 0.8 to 0.2 V, while ensuring the stability of the voltage. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 7/11 Typical polarization curves in potentiostatic mode at 70 and 95°C, 20 and 50% RH are shown in Figure 4. We observed a current density limit value, beyond which a decrease in voltage does not imply an increase but a decrease in current density for both temperatures. At 95°C, the trend is particularly pronounced when the relative humidity is reduced to 20%, whereas at 70°C, the same behaviour occurs at the same current density limit for all the relative humidities tested, showing that the impact of relative humidity is not significant compared with the 95°C case. In parallel, Rhf measurements are performed alongside the polarization curves. At 95°C for both RH, the current density limit corresponds to an increase in Rhf. It is important to note that low voltage and high current density result in high MEA drying, particularly at low RH levels. While the trend at 70°C may be comparable, there is not a similar increase in Rhf. Consequently, the fuel cell performance limitations observed at high current density at 70 and 95°C do not stem from the same origin. Figure 4: polarization curves (dot line) an Rhf measurements (solid line) at 95°C (a) and 70°C (b) As previously discussed, oxygen dilution is not a contributing factor to the voltage drop at 95°C and low RH, given the enhanced performance observed at higher gas RH levels. Given that the increase in Rhf occurs at a given current density (i.e. operating point), two origins can be hypothesised: • The electroosmosis effect, defined as the movement of water molecules hydrating protons during their transfer from the anode to the cathode [23],[24],[25], increases with current density increasing. Approaching the current density limit may result in the membrane drying out on the anode side [3],[4]. • Thermal effect may imply a water flux from the electrodes to the bipolar plates. Indeed, increasing current density results in an increase in thermal power produced and thus, MEA temperature. There is a temperature gradient between the MEA and the bipolar plate which can induce a high-water flux, similarly to the heat pipe effect where a saturated water vapor gradient induces membrane drying. Origin of voltage drop at 95°C To better discriminate between these two hypotheses, we undertook a six-day AST, with characterizations conducted every two days (Figure 5 (a)). As expected, a global decline in performance is observed with a loss of voltage across the overall polarization curve. It is interesting to note that the increase in Rhf is observed at lower current density as the AST reduced the FC performance. This means that there is not a precise current density value that induces the drying out of the membrane. In other words, the membrane is drying out at a lower rate of electroosmosis flux. Therefore, electroosmosis cannot be the main cause of the membrane drying. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 8/11 Figure 5: polarization curves and Rhf measurements during AST at 95°C / 30% RH (a) and in asymmetric gas RH conditions (b) As previously discussed, the electroosmosis flux is assumed to dry out the membrane mainly on the anode side, or from the anode side. To assess the impact of this phenomenon, we measured FC performance with asymmetric gas RH: 20% on the anode side and 50% on the cathode side (Figure5 (b)). Interestingly, there was no difference in performance between the symmetrical RH case and asymmetrical one. Reduce anode humidification by decreasing the RH on this side does not affect the polarization curve and Rhf behaviour, confirming that electroosmosis effect is not responsible for the sudden increase in Rhf and corresponding voltage drop at high current density. This means that heat production at the electrodes may be the main phenomenon governing the increase in Rhf at high current density and high temperature. 4. Conclusions Several characterizations have been carried out at varying cell temperatures, gas RH and pressures, and air stoichiometry levels. These were done to assess the impact of these parameters on performance at 95°C in comparison to the standard operating temperature of 70°C. Identical commercial MEAs made from state-of-the-art materials were used. First, the following conclusions can be drawn from the performance mapping: • The increase in gas pressure and RH induces higher performance with a global gain in voltage across the whole polarization curve. High frequency resistance measurements demonstrate that increasing gas pressure and RH results in a better or much better membrane hydration. • The increase in air stoichiometry has a beneficial impact, particularly at high current density with gas pressure of 3 bar and appropriate RH. Indeed, a higher air stoichiometry is needed in humid conditions for water management whereas a lower air stoichiometry is needed for dry conditions, especially at low RH and elevated temperature. • The increase in cell operating temperature leads to lower FC performance in every configuration, mainly due to higher Rhf, demonstrating a different hydration state of the membrane. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476227 A1104 Page 9/11 In addition, a voltage drop occurred in dry conditions, specifically at 95°C. Rhf measurements have been carried out across the polarization curve at 70 and 95°C, as well as at different gas RH levels. While voltage loss at 70°C at high current densities are related to oxygen dilution caused by the additional water produced, higher temperatures introduce a different loss mechanism, leading to significant membrane dehydration. A significant increase in Rhf has been observed at the same current density where a voltage drop occurs. Our results indicate that the electroosmosis effect is not the main cause of this phenomenon, heat production being most likely the main governing parameter. This conclusion is supported by RH measurements taken during AST, as well as measurements with asymmetrical RH. Thus, the phenomena limiting FC performances at high and low or usual temperatures are clearly different. References [1] C. Zhang, W. Zhou, L. Zhang, S. H. Chan, et Y. Wang, « An experimental study on anode water management in high temperature PEM fuel cell », Int. J. Hydrog. Energy, vol. 40, no 13, p. 4666‑4672, avr. 2015, doi: 10.1016/j.ijhydene.2015.02.037. [2] M. Boillot, C. Bonnet, N. Jatroudakis, P. Carre, S. Didierjean, et F. Lapicque, « Effect of Gas Dilution on PEM Fuel Cell Performance and Impedance Response », Fuel Cells, vol. 6, no 1, p. 31‑37, 2006, doi: 10.1002/fuce.200500101. [3] D. G. Sanchez, D. G. Diaz, R. Hiesgen, I. Wehl, et K. A. Friedrich, « Oscillations of PEM fuel cells at low cathode humidification », J. Electroanal. Chem., vol. 649, no 1, p. 219‑231, nov. 2010, doi: 10.1016/j.jelechem.2010.04.005. [4] X. Li, K. Han, et Y. Song, « Dynamic behaviors of PEM fuel cells under load changes », Int. J. Hydrog. Energy, vol. 45, no 39, p. 20312‑20320, août 2020, doi: 10.1016/j.ijhydene.2019.12.034. [5] S. G. Kandlikar et Z. Lu, « Thermal management issues in a PEMFC stack – A brief review of current status », Appl. Therm. Eng., vol. 29, no 7, p. 1276‑1280, mai 2009, doi: 10.1016/j.applthermaleng.2008.05.009. [6] Y. Devrim, A. Albostan, et H. Devrim, « Experimental investigation of CO tolerance in high temperature PEM fuel cells », Int. J. Hydrog. Energy, vol. 43, no 40, p. 18672‑18681, oct. 2018, doi: 10.1016/j.ijhydene.2018.05.085. [7] V. B. K.P, G. Varghese, T. V. Joseph, et P. Chippar, « Spatial analysis of CO poisoning in high temperature polymer electrolyte membrane fuel cells », Int. J. Hydrog. Energy, vol. 46, no 11, p. 8179‑8196, févr. 2021, doi: 10.1016/j.ijhydene.2020.12.001. [8] R. K. Abdul Rasheed, Q. Liao, Z. Caizhi, et S. H. Chan, « A review on modelling of high temperature proton exchange membrane fuel cells (HT-PEMFCs) », Int. J. Hydrog. Energy, vol. 42, no 5, p. 3142‑3165, févr. 2017, doi: 10.1016/j.ijhydene.2016.10.078. [9] Y. Nalbant, C. O. Colpan, et Y. Devrim, « Energy and exergy performance assessments of a high temperature-proton exchange membrane fuel cell based integrated cogeneration system », Int. J. Hydrog. Energy, vol. 45, no 5, p. 3584‑3594, janv. 2020, doi: 10.1016/j.ijhydene.2019.01.252. [10] E. Jannelli, M. Minutillo, et A. Perna, « Analyzing microcogeneration systems based on LT-PEMFC and HT-PEMFC by energy balances », Appl. Energy, vol. 108, p. 82‑91, août 2013, doi: 10.1016/j.apenergy.2013.02.067. [11] A. Suzuki et al., « Evaluation for sintering of electrocatalysts and its effect on voltage drops in high-temperature proton exchange membrane fuel cells (HT-PEMFC) », Int.