Temperature and Current Density distributions in a 100 cm2 PEM Fuel Cell: Effects of flow field designs

Temperature and Current Density distributions in a 100 cm 2 PEM Fuel Cell: Effects of flow field designs G.M. Cabello Gonz´ alez a,* , Baltasar Toharias b , Felipe Rosa a,c , J.J. Guerra a , Alfredo Iranzo a,c a Departamento de Ingeniería Energ´ etica, Grupo de Termotecnia. Escuela T´ ecnica Superior de Ingeniería. Universidad de Sevilla. Camino de los Descubrimientos s/n. 41092, Sevilla, Spain b Asociaci´ on de Investigaci´ on y Cooperaci´ on Industrial de Andalucía – AICIA. Grupo Termotecnia, Camino de los Descubrimientos s/n. 41092, Sevilla, Spain c ENGREEN, Laboratory of Engineering for Energy and Environmental Sustainability, Universidad de Sevilla, Spain HIGHLIGHTS •Comparative CDM study of three 100 cm 2 PEM cell flow field configurations. •As Current demand rises, temperature and current distributions become more uneven. •Tapered design enhances operation via uniform temperature and current distribution. ARTICLE INFO Keywords: PEM fuel cell Temperature distribution Current density mapping Dynamic load cycle ABSTRACT Electro-thermal mapping provides valuable insights into the performance evaluation of polymer electrolyte membrane fuel cells (PEMFCs) by depicting the spatial distribution of current density and temperature. In this study, electro-thermal maps were generated for three different designs of 100 cm 2 PEMFC flow fields (conventional serpentine with two different channel depths, and serpentine-tapered). The performance of each design was characterized by analyzing the surface (in-plane) distributions of current density and temperature at different cell voltages. At elevated current densities, a linear increase in the non-uniformity of temperature and current density distribution is observed. The central region of the bipolar plate exhibits higher temperatures, whereas the region with high current densities is situated near the hydrogen inlet, gradually diminishing as the hydrogen depletes towards the outlet. Results show that, in general, the tapered flow field design exhibits better performance with a more homogeneous temperature and current distribution throughout the entire active area. This behavior can be attributed to better water management and gas diffusion towards the electrode due to the acceleration and pressure increase of the reactant fuel gas along the narrowing channel. Novel insights were identified by applying the Current Distribution Mapping (CDM) technique for analyzing current density and temperature in-plane distributions under dynamic load conditions, comparing the different channel depths or tapered designs during the dynamic operation of the cell. During dynamic tests, temperature increased rapidly for increasing loads but the decrease was more slowly when load was lowered, leading to an overall gradual temperature rise and less homogeneous distribution at higher currents, while the current distribution adjusted almost instantly with constant standard deviation during both load increases and decreases. 1. Introduction In the pursuit of mitigating climate change, the energy sector stands at the forefront of decarbonization efforts [1] since, globally, it is the largest emitter of global greenhouse gas emissions [2]. Governments worldwide are getting involved, implementing policies and incentives to transition away from fossil fuels towards cleaner alternatives, limiting the adverse impacts of climate change. However, achieving decarbonization goals necessitates not only renewable energy sources like wind and solar, but also versatile energy vectors capable of storing and * Corresponding author. E-mail address: [email protected] (G.M.C. Gonz´ alez). Contents lists available at ScienceDirect Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour https://doi.org/10.1016/j.jpowsour.2025.237625 Received 14 February 2025; Received in revised form 13 May 2025; Accepted 7 June 2025 Journal of Power Sources 652 (2025) 237625 Available online 23 June 2025 0378-7753/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). transporting energy efficiently [3]. In this regard, hydrogen emerges as a potential energy vector candidate due to its high energy density and, over all, its versatility, that lies in hydrogen’s ability to be used in multiple applications, such as fueling vehicles, providing heat for industrial processes, and serving as a feedstock for chemical production [4]. As an energy carrier, hydrogen can be produced through electrolysis using renewable electricity, thereby offering a pathway to integrate excess renewable energy and store it for later use [5]. Hydrogen can be utilized directly in fuel cells to generate electricity with water vapor as the only byproduct, offering a clean alternative to combustion engines. Proton Exchange Membrane (PEM) fuel cells have gained significant interest as power sources for a wide range of applications, including automobiles and stationary power systems [6]. This is due to their numerous advantages, such as fast start-up times, low noise, high power density, and high energy conversion efficiency [7]. Fuel cell technology development is requiring the support of strong testing capabilities and techniques, to accurately address performance and durability of the cells and stacks [8]. However, conventional fuel cell testing methods, like polarization curve or impedance spectroscopy measurement, provide an overall performance assessment but lack localized information along the active area. The local distributions of current density and temperature are key in the electrochemical reaction of fuel cells because it directly impacts the efficiency, performance, and durability of the cell, and thus, understanding their variations along the active area is vital for assessing local energy conversion performance and potential degradation issues [9]. If the current density is unevenly distributed, localized regions of high or low current density may develop. This can lead to uneven utilization of reactants, incomplete electrochemical reactions, high temperature gradients, and decreased overall efficiency together with accelerated degradation [10]. Non-uniform distribution of current density can also result in localized degradation of the electrode and membrane materials or catalyst degradation [11]. High current density regions may experience accelerated degradation due to increased electrochemical activity, leading to reduced durability and lifetime of the fuel cell stack. Non-uniform distribution of current density can also lead to uneven water distribution within the cell, causing flooding or drying out of certain regions. This can impair proton conductivity in the electrolyte membrane and hinder the electrochemical reactions, leading to decreased performance. In spite of all this, current evidence does not conclusively demonstrate that current density inhomogeneity directly accelerates performance loss or induces operation far from optimal conditions. The observed degradation trends do not show a consistent correlation with areas of higher or lower current density [12,13]. It is important to note that conventional characterization techniques will not be able to identify such local gradients and issues. It is well known that efficient heat dissipation is crucial for preventing temperature gradients within the fuel cell stack, which affects performance and durability [14]. Uniform distribution of current density helps ensuring more uniform heat generation and dissipation, contributing to stable operating conditions and prolonged stack life. The local distribution of current density depends on factors like the local concentration of reactants at the electrode surface, as well as the potential distribution in the cell, which is influenced by pressure, relative humidity, and temperature [15]. Strategies such as electrode design optimization, improved catalyst formulations, and advanced flow field designs are employed to enhance current distribution and maximize the overall performance of fuel cell systems [16,17]. Since conventional testing methods such as polarization curve measurement or electrochemical impedance spectroscopy lack of local information, Current Density Mapping (CDM) provides valuable insights into the performance of a PEM fuel cell helping to identify areas of high and low current density, which can be used to assess the overall efficiency and functionality of the fuel cell, and may lead to localized issues such as catalyst degradation, gas starvation or membrane damage or dehydration [18]. In order to gain further knowledge in local distributions, local temperature measurements and its correlation with local current density can been assessed introducing thermocouples inside the fuel cell. Wilkinson et al. [19] proved that local temperature measurements can effectively correlate with local current densities using micro-thermocouples, suggesting a potential indirect method for current mapping. However, improvements were needed in thermocouple robustness and optimal placement. Determination of current density distribution can be assessed by mean of a segmented cell where current is measured individually in each segment by using a Hall-effect sensor. Hwang et al. [20] employed a specially designed composite plate with 16 segmented current collectors to test various cathodic flow-field patterns. The study evaluated their effects on current distribution for various operating conditions and concluded that the serpentine flow field yields the most uniform distribution due to its superior mass transfer and water management capabilities. Despite the interesting conclusions obtained with segmented cells, this method is limited to measure the current density distribution of a single fuel cell or the one at the ends of a stack, while sensor plates inserted into fuel cells can measure any cell in the stack since the current passes through a printed circuit sensor that measures current distribution by means of a shunt. Besides, this kind of sensors can allow higher resolution in the measurement. In this sense, Geske et al. [18] presented a measurement system for mapping current density distribution utilizing multi-layer technology for integrating shunt resistors within a printed circuit board tailored to the fuel cell’s flow field channels. Results for a serpentine flow channels indicate minimal temperature influence during cell warm-up, with maximum local current production typically near the oxygen inlet, though load current induces varying levels of inhomogeneity, highlighting potential for optimizing a more uniform current density distribution profile. In this sense, Heuer et al. [21] developed a CDM sensor and inserted them between the two central cells of a serpentine flow channels stack. They found that the stoichiometry of the cathode was the parameter that influenced the most on the current density profile since it had a large impact on the humidity balance, with lower or unstable cell voltages observed when the current density profile was widely spread. These results are in line to those obtained by Peng et al. [22] for a commercial size stack. They found that current density distribution is more sensitive to the flow rate of air than that of hydrogen, and local gas starvation occurs when the hydrogen stoichiometric ratio is 1.1. Heuer et al. [21] long-term analysis revealed an increase in the spreading of the current density profile, potentially related to aging, underscoring the importance of CDM for early detection and reaction to operational changes compared to traditional cell voltage measurements. Belhadj et al. [9] reached the same conclusion. They found that fresh MEA revealed non-uniform distributions primarily caused by preferential gas distribution in certain channels, with little influence from cell current density. However, aging resulted in significantly uneven distributions, with evidence of flooding at high current density in the last section. On a higher scale, Yin et al. [23] combined Computational Fluid Dynamics (CFD) model with current mapping showing that, in counter-flow hydrogen and air operations, the cathode outlet has the lowest and the mid portions the highest local current values. They reported that current distribution uniformity was improved by increased air stoichiometric ratio and that cross-flow configurations enhanced temperature uniformity more effectively than parallel-flow ones. In recent times, sensors have allowed for an electro-thermal mapping, including information for both, temperature and current density distribution. Meyer et al. [24] created for the first time a localized electro-thermal performance map for a 60 cm 2 PEM fuel cell. The study found that at low current densities, the current density distribution was mainly influenced by the anode and hydrogen consumption gradient, with the highest temperature gradient along the air flow direction. On the other hand, at high current densities, temperature influenced the current density distribution. Also, when the sensor was coupled with a low-frequency impedance spectroscopy device, the limit of operation as G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 2 well as the optimal performance range could be found [25]. The cell voltages at the current with the lowest resistance were approximately 0.60 V, which is slightly below the usual operating range of 0.70-0.65 V for automotive applications. In contrast, the voltage at the current with the highest power density is about 0.50 V. This demonstrates that the current associated with the lowest resistance is a more appropriate single-value measure than the point of maximum power. Besides, when electro-thermal mapping was used in combination with water imaging, hydration and dehydration effects under varying load and flow conditions could be studied, finding that water mainly accumulates under cooling channels. Using all three techniques together provides a comprehensive understanding of water management in fuel cells. A complete study carried out by Bethapudi et al. [15] comparing a 25 cm 2 fractal flow field cathode with a conventional serpentine flow field shows the correlation between cell performance and current-temperature distribution. The fractal design significantly enhances performance, particularly at high current densities, due to higher cell temperatures. Electro-thermal mapping shows the fractal flow-field achieves a more uniform current distribution and mitigates flooding better than the serpentine design, resulting in 10–15 % higher segment currents. Heterogeneity of temperature along with current density was also assessed by Jiang et al. [26]. Using an electro-thermal mapping device in a serpentine-pattern flow field 100 cm 2 active area PEM fuel cell, the study measured this heterogeneous distribution during polarization curve tests at temperatures of 50 ◦C and 60 ◦C. They observed that higher temperatures and current densities increased heterogeneity and offered a new way to convert power-current curves into power-heterogeneity curves. This way, it can be ensured that the cell operates with similar output performance but reduced current density heterogeneity within a specific range, enhancing its lifespan. Also, Su´ arez et al. [27] used this technology in a 50 cm 2 parallel serpentine ElectroChem Inc fuel cell in order to test different configurations and assess the cell response during a standardized driving cycle. They found that inverse hydrogen flow configuration outperformed the other configurations in terms of polarization and power curves. Local current density distribution presented significant heterogeneities showing an inverse bell-shaped distribution across all configurations while, in contrast, local temperature distributions were highly homogeneous. Nevertheless, to the best of the authors’ knowledge, there have been no prior studies examining the local current density and temperature distributions during dynamic load cycling tests comparing different designs of fuel cells. This is a significant gap in the literature, particularly since the spatial and temporal variations in these distributions during load changes could create challenging gradients for the cell’s operation. This study aims to address this gap by providing novel insights into fuel cell behavior under dynamic conditions, specifically addressing how and why different flow field designs affect current density and temperature in-plane distributions, which can be also of interest for PEM electrolyzers. In particular, the novelty of this work is the use of the Current Distribution Mapping (CDM) technique for analyzing current density and temperature in-plane distributions comparing different channel depths or tapered designs during dynamic conditions. The analysis was conducted over three distinct single-cell 100 cm 2 flow field configurations: a parallel serpentine design with two different channel depths, and a tapered design. In addition to these comparative insights, this study also examines the time-varying behavior of current density and temperature distributions during load changes, following a well-established dynamic load protocol. The findings from this analysis provide a further understanding on how flow field design impacts the cell behavior. 2. Experimental facility and methodology 2.1. Fuel cell design The experimental work was conducted using a specialized PEM fuel cell station with capacity to test single cells and short stacks up to 500 W. The testing environment included a reactant gas handling unit equipped with mass flow controllers, humidifiers, back pressure regulators, an electronic load, and a temperature control system where the heat provided by two heating pads is regulated, as well as a set of external fans for cooling purposes. Throughout the experiments, data on current density and temperature distribution inside the fuel cell were collected at 1 Hz acquisition frequency using a CDM sensor (S++ current scan shunt, Germany) placed between the cathode current collector and the cathode bipolar plate, with an appropriate sealing. The active area of 100 cm 2 is, this way, divided into electrically isolated segments in a matrix of 14x14 for current and 7x7 for temperature measurement. The cell hardware used in the experiments was the Pro-RD from Pragma Industries (France), where state-of-the art 7-layer MEAs from IRD Fuel Cells (Denmark) were used (8 μ m fuel cell membrane, 0,1 mg/cm 2 in anode and 0.3 mg/cm 2 in cathode, with Gas Diffusion Layer (GDL) in anode of 135 μ m @1 MPa and GDL in cathode of 146 μ m @1 MPa. All the experiments were carried out using a single cell. three different graphite plates (6.0 mm thick) were tested for the cathode side, with an active area of 100 cm 2 . The designs in all cases were based on a 7-channel parallel serpentine, with the only difference being the depth of the channels (Fig. 1). The first design was a conventional serpentine design with a channel depth of 1 mm, whereas in the second the depth was decreased down to 0.5 mm. The third design was also a 7-channel parallel serpentine with a tapered design where the depth was reduced progressively from 1 mm to 0.5 mm along the channel path (Fig. 1a). The width of the channel and the rib remained constant for the three designs, being 0.8 and 1.0 mm respectively. It is important to note that the first design was used at the anode side for all the experiment, and only the bipolar plate for the cathode side was changed. Reactant gases enter the plate from the top left/right corner and circulate countercurrent in the horizontal sections of the channels and concurrent in the short vertical parts, leaving the plate through the outlet located at the opposite corner (Fig. 1c). The assembly of the cell was done using twelve bolts and nuts placed in the edges of the plates, with a tightening torque of 9.0 Nm, to prevent leaks and promote uniform current collection. 2.2. Experimental tests Polarization and power curves, along with driving tests (New European Driving Cycle dynamic load tests), were obtained for each design under two distinct operating conditions (Table 1), following the protocols established by the EU Commission Joint Research Centre (JRC) [28]. The naming system for the tests is as follows: PAA_TBB_aCCHRDD_cEEHRFFair. In this nomenclature, AA represents the outlet gauge pressure in bar with a decimal, BB denotes the operating temperature in Celsius, CC indicates the anode stoichiometry with a decimal, DD signifies the anode relative humidity in percentage, EE stands for the cathode stoichiometry with a decimal, and FF indicates the cathode relative humidity in percentage. The word "air" signifies that the experiment was conducted using air instead of pure oxygen in the cathode. Concurrently with the polarization curve and driving test, the electro-thermal mapping sensor was employed to record temperature and current density data, with an acquisition frequency of 1 Hz. Prior to obtaining the I-V polarization curves, the membrane was activated operating the fuel cell stack for 24 h, at P05_T65_a13HR60_ c25HR60air and constant current density of 0.5 A/cm 2 . After activation, the cell was set to Open Circuit Voltage (OCV) for 120 s. Measurements were taken in galvanostatic mode, starting from OCV with fixed current density steps as per indication of the testing protocol [28]. For low current densities (up to 0.08 A/cm 2 ), measurements were taken every 120 s, and for higher current densities (0.1 A/cm 2 to maximum), every 300 s. The test concluded when cell’s voltage dropped below 0.3 V. Data G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 3 were acquired at a rate of one sample per second, where for each current the voltage average of the last 30 samples was used for post-processing. Three repetitions for each experiment were carried out to verify the reproducibility of the obtained results. The tests for the Fuel Cell Dynamic Load Cycle (FC-DLC) were carried out according to the well-defined testing protocol [28], where 1 test cycle of 1200 s represents a driving distance of about 11 km over a span of 20 min. 3. Results and discussion 3.1. Polarization performance The polarization curves for the three designs at two different operating conditions are compared in Fig. 2a while power vs current density curves are shown in Fig. 2b. As shown, the curves obtained under P10_T70_a15HR60_ c30HR55air operating conditions (continuous line in Fig. 2) exhibit better performance at any current density. This result is consistent with previous studies conducted on a 7-cell stack with a similar design (bipolar plates used were the same as in the first design) [29] where the effects of changes in different operating conditions were studied and such operating condition was identified as the optimal one for the stack. For the 1.0 mm thickness channels, when reaching certain current densities (above 0.6 A/cm 2 at P05_T65_a13HR60_ c25HR60air and 0.8 A/cm 2 at P10_T70_a15HR60_ c30HR55air), the voltage drops clearly indicating a performance loss, while the 0.5 mm and the tapered design will operate at least until 1.0 A/cm 2 . The tapered design presents Fig. 1. a) Plate designs with different channel depth. From left to right: 1.0 mm, 0.5 mm, and tapered (1.0–0.5 mm). b) CDM sensor used in the experiments (S++, Germany). c) Flow configuration for the cell used in the tests. Front view (left) and perspective (right). Table 1 Operating conditions defined in the experimental tests. Case Pg (bar) T (◦C) RHa (%) RHc (%) λ a (−) λ c (−) Oxidant P10_T70_a15HR60_ c30HR55air 1.0 70 60 55 1.5 3.0 Air P05_T65_a13HR60_ c25HR60air 0.5 65 60 60 1.3 2.5 Air G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 4 the highest performance for almost all the current density range, reaching a maximum power value of 45 W at 1 A/cm 2 (Fig. 2). The superior performance of the tapered design can be attributed to the progressive acceleration of the reactant fuel gas in the tapered channel as well as to the air pressure increase along the channel path, where oxygen is then forced into the GDL enhancing the electrochemical reaction. The progressive gas acceleration in the channel actually increases also the water removal capability and thus improves cell performance. These results are in line with those reported in literature, where the positive tapered slope is proven to perform better, especially at high current densities [30]. While it is inevitable that the use of tapered channels incurs a pressure-loss penalty, this loss increase is generally quite low (around 0.018 bar reported for a tapered design with a taper ratio of 0.1) [31]. The three designs tested exhibited a pressure drop below 0.1 bar, which is within the precision limits of the pressure sensors installed in the test bench. 3.2. Cell temperature and current density distribution analysis The temperature and current distribution maps obtained with the CDM sensor at low, medium, and high currents (0.2, 0.4, and 0.7 A/cm 2 , respectively) for the three bipolar plate channel designs are shown in Fig. 3. Maps represent the local current and temperature measured by the sensor (placed between the cathode bipolar plate and the cathode current collector). In Fig. 3 a), the temperature value is displayed on the z-axis, while on the x and y axes, the active area is represented (with the segment number as units, 7x7 for temperature and 14x14 for current). The air inlet is located at coordinates (1,1) (blue arrow), while the hydrogen inlet is positioned at the corner (1,7) (red arrow). In this manner, the gases circulate countercurrent through the serpentine, which is aligned with its long edge parallel to the y-axis, allowing the air to exit at the corner (7,7) and the hydrogen at (7,1), as depicted by the inset in Fig. 3 a). In all the designs, the temperature map presents an inverted bowl figure, with higher temperature values at the central segment. This is clearly due to the fact that, as the cell is not internally cooled by a refrigerant circuit but rather by external fans, the heat generated is evacuated only from the cell external sides from the center outward, creating a radial temperature gradient with highest temperatures at the center. It can be noted that the average temperature for all tests is slightly higher than intended, as the thermocouple monitoring temperature is inserted into the bipolar plate but outside the active area. This offset increases as the current density increases, so it is worth noting that the conditions initially intended to be tested are not precisely the ones occurring inside the cell (which indicates as well that the differences in the temperature within the cell hardware are relevant). It can be observed that temperature is the parameter most affected by the design of the channels. In all cases, it is observed that as the current increases the temperature rises and its distribution becomes less homogeneous. A linear increase of the temperature standard deviation with current density was observed (depicted in Fig. 4). In the case of the 1.0 mm serpentine, at P05_T65_a13HR60_c25HR60air, this design exhibits the highest temperature value across the entire range of current density. Moreover, the increase in temperature with current density is very slight at low current values but increases significantly at high current values due to the higher heat generation and poor heat management of the externally air-cooled cell operation. The maximum differences in temperature and current density in the measurement plane are depicted in Fig. 7 for the three designs at 0.1, 0.4 and 0.7 A/cm 2 . Temperature differences are in all cases below 4 ◦C. Also, the variation in current distribution according to the current Fig. 2. Polarization (a) and power (b) curves for the different designs. ● tapered serpentine design, ■ 0.5 mm depth serpentine design and ▴ 1 mm depth serpentine design. Dashed line P05_T65_a13HR60_ c25HR60air. Solid line P10_T70_a15HR60_ c30HR55air. G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 5 density for the three channel designs can be observed at P05_T65_a13HR60_c25HR60air in Fig. 4. As the current increases, the heterogeneity in current distribution within the cell also increases, as evidenced by the rise in standard deviation with current density. This effect is consistent with previously reported literature [22], which indicates that when the fuel cell is operated at lower total currents, the difference in local current densities is smaller, and the dispersion of current density distribution at higher currents is larger. This current dispersion is linked with high water production and accumulation, indicating that the tapered design, with its ramp-shaped channels in the Fig. 3. Temperature a) and Current b) distribution mapping at 0.2, 0.4 and 0.7 A/cm 2 for the three considered designs. Operating condition: P05_T65_a13HR60_c25HR60air. Air inlet/outlet locations marked with blue arrows, hydrogen intake/outtake locations marked with red arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 6 direction of flow, presents better reactant distribution and water management as the current heterogeneity is lower for this design at any current density, with the maximum heterogeneity observed in the 1.0 mm design. Maximum values of current density in all designs are present in the central part of the active area, which corresponds as well to the zone with higher temperature. The designs with 0.5 mm and 1.0 mm depth exhibit a more pronounced (higher values) in the area corresponding to the hydrogen inlet and air outlet. There is a decrease in current intensity as reactants are consumed towards the outlet of the cell, due to the progressive reduction of reactant concentration along the channel path together with the corresponding increase in the relative humidity. However, the tapered design does not exhibit this trend; instead, its distribution is more parallel to the (x,y) plane. This indicates that the progressive narrowing of the channel depth along the flow path effectively enhances the uniform transport of gases towards the electrode, promoting the electrochemical reaction and improving the overall performance of the cell. As the overall current increases, the maximum current and temperature difference increases for all configurations (Fig. 5). The tapered configuration generally shows lower current differences and generally exhibits lower or similar temperature differences, indicating that the tapered design may result in more uniform current and thermal distribution, especially at medium currents. 3.3. CDM study under dynamic conditions (load cycling tests) The three designs performed normally during the load cycling tests simulating driving conditions (carried out according to testing protocol [28]), responding swiftly to changes in the set point, with minimal voltage oscillations during periods of constant intensity. Fig. 6 shows the voltage variations of the cell for the three designs under the two operating conditions considered, during the complete load cycle. For both conditions, the 1.0 mm design exhibits a lower voltage compared to the other designs as it was the case during the performance polarization curve tests (section 3.1). CDM data was also collected during dynamic tests. For the sake of simplicity, only the CDM images of the tapered design test are shown for one of the conditions in Fig. 7, while Fig. 8 will gather the quantitative results of all three tests at the two operating conditions analyzed. Five representative points were selected. The first and second ones are set on a ramp with low current ascent and descent, respectively; the third is on a medium current ascent; the fourth is on a high current ascent; and, finally, the fifth is in a steady state at high current. Also, the test results can be visualized in video format as Supplementary Material. Supplementary video related to this article can be found at https:// doi.org/10.1016/j.jpowsour.2025.237625 During the test, it was observed that the temperature responds rapidly to positive current changes, increasing its value. However, when the current demand decreases, the temperature drops at a much slower rate than it rises due to the limitations in the heat transfer rate, and a steady state is not reached during periods of constant demand. As a result, the overall effect during the first 800 s of the test is a gradual increase in temperature at each point within the cell. In the second phase, at higher current demands, the temperature distribution experiences greater changes, with a less homogeneous distribution and the appearance of hot spots in the center of the cell, which extend towards the sides. On the other hand, the current distribution responds almost instantaneously to the demands of the driving cycle, quickly reaching a Fig. 4. Temperature a) and Current b) mean (M; bars) and standard deviation (SD; lines) at 0.1, 0.4 and 0.7 A/cm 2 for the three considered designs. Operating condition: P05_T65_a13HR60_ c25HR60air. Fig. 5. Maximum temperature a) and current b) difference at 0.1, 0.4 and 0.7 A/cm 2 for the three considered designs. Operating condition: P05_T65_a13HR60_ c25HR60air. G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 7 steady state. As in the previously shown cases, it can be observed that temperature increases as more current is demanded, and, in the same way, the homogeneity in the current distribution decreases at high density currents (Fig. 8). Also, the 0.5 mm design is always presenting the lowest average temperature while the tapered design is the one with the highest temperature except for low current density conditions. However, it is worth noting that if we compare the rise and fall (point 1 and point 2 in Fig. 7) at the same current density, although the average temperature remains almost constant, the standard deviation is higher in the case of the descent current, and thus a less homogeneous temperature distribution is featured for transient operation while the current is decreased. This phenomenon is observed at any current density, with temperature differences between the load increase and load decrease being more pronounced in the central area of the cell, where heat dissipation is more challenging. For example, at a current demand of 10 A, the standard deviation of the temperature distribution during the ramp-up phase is 0.26, while for the ramp-down phase it is 0.32. At a demand of 45 A, the standard deviation during the ramp-up phase increases to 0.53, which is higher than at 10 A, as expected based on previously described trends. For the ramp-down phase at 45 A, the standard deviation is 0.58. The same inverted bowl shape observed at the various points in the polarization curve is also evident during the dynamic tests. The nonhomogeneity of current distribution, on the other hand, follows a linear trend as current density increases. Of the three designs tested, the tapered design exhibits the most uniform current distribution across the entire area, regardless of the current density value or operating condition. Unlike the behavior observed with temperature, in the case of current distribution, the standard deviation remains constant during both increases and decreases. 4. Conclusions Electro-thermal mapping was used to determine the performance of Fig. 6. Driving test results for the three design at a) P05_T65_a13HR60_ c25HR60air and b) P10_T70_a15HR60_ c30HR55air. Fig. 7. CDM measurements during the driving test for the tapered design at P05_T65_a13HR60_ c25HR60air. G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 8 three different 100 cm 2 flow field designs in a PEM fuel cell. The performance of each design was assessed by analyzing the surface distribution of current and temperature at different cell voltages under stationary and dynamic operation. The study uncovered relationships between local current densities and local temperatures as well as the influence of the channel design in the cell performance. At high current densities, the non-uniformity in temperature and current density distribution increases linearly. The area of highest temperatures is located at the center of the cell due to the external air-cooling design, while the area of high current densities is observed near the hydrogen inlet and progressively decreases as the hydrogen is consumed along its path towards the outlet. The tapered design exhibits superior performance compared to the other designs, particularly at current densities above 0.5 A/cm 2 . Additionally, it demonstrates the least variation in both surface temperature distribution and current density under any operating condition. This demonstrates the superior capability of a progressively reduced passage area in enhancing gas transport towards the electrode, as well as an improved water management within the cell as progressive gas acceleration promotes water removal along the channels. Conversely, the 1.0 mm channel thickness design performs the worst, with a more pronounced voltage drop and higher mass transport polarization region, observed also during the dynamic load tests. When comparing the rise and fall at the same current density, the standard deviation of temperature is higher during the current decrease, leading to a less homogeneous temperature distribution in transient operation. For current distribution, the standard deviation remains constant during both increases and decreases. CRediT authorship contribution statement G.M. Cabello Gonz´ alez: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Investigation, Formal analysis, Data curation. Baltasar Toharias: Writing – review & editing, Visualization, Validation, Software, Investigation, Formal analysis, Data curation. Felipe Rosa: Supervision, Resources, Project administration, Funding acquisition, Conceptualization. J.J. Guerra: Supervision, Resources, Project administration, Funding acquisition. Alfredo Iranzo: Supervision, Resources, Project administration, Methodology, Investigation, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgement Grant TED2021-130706B-I00 funded by MCIN/AEI/10.13039/ 501100011033 and by European Union “NextGenerationEU”/PRTR, and grant PID2023-146745OB-I00 funded by MICIU/AEI/10.13039/ 501100011033 and by ERDF/EU. Experimental infrastructure funded by UNSE15-CE2962 and EQC-2018-004258-P by MCIN/AEI/10.13039/ 501100011033, co-funded with ERDF funds. Contribution from Baltasar Toharias Funded by CDTI, with the support from Ministerio de Ciencia e Innovaci´ on. Financiado por la Uni´ on EuropeaNext Generation EU (project H2ENRY CER-20231027). Authors thank Pablo Iranzo for the realization of the video in Supplementary Material. Data availability Data will be made available on request. References [1] H. Liu, I. Khan, A. Zakari, M. Alharthi, Roles of trilemma in the world energy sector and transition towards sustainable energy: a study of economic growth and the environment, Energy Policy 170 (2022) 113238, https://doi.org/10.1016/j. enpol.2022.113238. [2] H. Ritchie, Sector by sector: where do global greenhouse gas emissions come from? Our World Data (2020). Published online: https://ourworldindata.org/ghgemissions-by-sector. [3] E. Papadis, G. Tsatsaronis, Challenges in the decarbonization of the energy sector, Energy (Calg.) 205 (2020) 118025, https://doi.org/10.1016/j. energy.2020.118025. [4] D.A. Cullen, K.C. Neyerlin, R.K. Ahluwalia, R. Mukundan, K.L. More, R.L. Borup, et al., New roads and challenges for fuel cells in heavy-duty transportation, Nat. Energy 6 (2021) 462–474, https://doi.org/10.1038/s41560-021-00775-z. Fig. 8. (a) P05_T65_a13HR60_ c25HR60air (b) P10_T70_a15HR60_ c30HR55air. Points 1 to 5 correspond to those marked in the driving test in Fig. 7. G.M.C. Gonz´ alez et al. Journal of Power Sources 652 (2025) 237625 9