Effect of sub-freezing temperatures on a PEM fuel cell performance, startup and fuel cell components

The cold-start behavior and the effect of sub-zero temperatures on fuel cell performance were studied using a 25-cm² proton exchange membrane fuel cell (PEMFC). The fuel cell system was housed in an environmental chamber that allowed the system to be subjected to temperatures ranging from sub-freezing to those encountered during normal operation. Fuel cell cold-start was investigated under a wide range of operating conditions. The cold-start measurements showed that the cell was capable of starting operation at −5 °C without irreversible performance loss when the cell was initially dry. The fuel cell was also able to operate at low environmental temperatures, down to −15 °C. However, irreversible performance losses were found if the cell cathode temperature fell below −5 °C during operation. Freezing of the water generated by fuel cell operation damaged fuel cell internal components. Several low temperature failure cases were investigated in PEM fuel cells that underwent sub-zero start and operation from −20 °C. Cell components were removed from the fuel cells and analyzed with scanning electron microscopy (SEM). Significant damage to the membrane electrode assembly (MEA) and backing layer was observed in these components after operation below −5 °C. Catalyst layer delamination from both the membrane and the gas diffusion layer (GDL) was observed, as were cracks in the membrane, leading to hydrogen crossover. The membrane surface became rough and cracked and pinhole formation was observed in the membrane after operation at sub-zero temperatures. Some minor damage was observed to the backing layer coating Teflon and binder structure due to ice formation during operation.     

Three-dimensional non-isothermal model development of high temperature PEM Fuel Cells

A three-dimensional non-isothermal mathematical model is developed in a triple mixed serpentine flow multichannel domain for a high temperature PEM Fuel Cell having a phosphoric acid doped PBI membrane as electrolyte and an active area of 25 cm² within Comsol Multiphysics. The inlet temperatures of cathode and anode reactants are taken as 438 K. Model predicts pressure, and temperature distribution along the channels and membrane current density distribution over the membrane electrodes. The model results are obtained at two different operation voltages, 0.45 V and 0.60 V. Resulting average current densities are respectively 0.313 A cm^?2 and 0.224 A cm^?2. The non-isothermal model results are compared to isothermal model results from a previous study and various other single channel non-isothermal model results available in the literature. The pressure drop at cathode compartment is predicted to be 6500 Pa, whereas it is found to be 6400 Pa for the isothermal model. The temperature difference within the system is found to be 0.18 K for the operation voltage of 0.6 V, whereas this value increases to 0.31 K for the operation voltage of 0.45 V. The temperature difference isocontours are illustrated for the whole cell. Considering changes in temperature, one can employ isothermal operation assumption for this system as an approximation and simplification for the governing equations, since the variation in the temperature within the cell is less than 1 K. It should be emphasized that multichannel model predictions are more realistic compared to single channel models. The model developed here can be extended to larger electrode active area and different multichannel configurations.     

Dynamic behaviors of PEM fuel cells under load changes

In this study, a one-dimensional isothermal single-phase transient model considering the finite-rate water absorption/desorption of membrane was established to study the dynamic behaviors of polymer electrolyte membrane (PEM) fuel cells under different cathode inlet humidity conditions in the presence of voltage step changes. Both the overshoot and undershoot phenomena were observed. Moreover, the distributions of water inside the electrolyte and the influence of that on the response current density of fuel cells were analyzed. When voltage stepped up/down, the water content in anode generally increased/decreased, and the water content in cathode is reversed. If the cathode intake is fully humidified, the water vapor in cathode is always over-saturated causing the change of ionic resistance is determined by that of the water content in anode. If the cathode intake is partially humidified, the change of ionic resistance could maintain within a small range owing to the change of water content in anode can be balanced by that of the water content in cathode.     

A PEM fuel cell model for cold-start simulations

In this paper, a transient multiphase multi-dimensional PEM fuel cell model has been developed in the mixed-domain framework for elucidating the fundamental physics of fuel cell cold start. Cold-start operations of a PEM fuel cell at a subfreezing boundary temperature of −20 °C under both constant current and constant cell voltage conditions have been numerically examined. Numerical results indicate that the water vapor concentration inside the cathode gas channel affects ice formation in the cathode catalyst layer and thus the cold-start process of the fuel cell. This conclusion demonstrates that high gas flow rates in the cathode gas channel could increase fuel cell cold-start time and benefit the cold-start process. It is shown that the membrane plays a significant role during the cold-start process of a PEM fuel cell by absorbing the product water and becoming hydrated. The time evolutions of ice formation, current density and water content distributions during fuel cell cold-start processes have also been discussed in detail.     

Rapid self-start of polymer electrolyte fuel cell stacks from subfreezing temperatures

Polymer electrolyte fuel cell (PEFC) systems for light-duty vehicles must be able to start unassisted and rapidly from temperatures below −20 °C. Managing buildup of ice within the porous cathode catalyst and electrode structure is the key to self-starting a PEFC stack from subfreezing temperatures. The stack temperature must be raised above the melting point of ice before the ice completely covers the cathode catalyst and shuts down the electrochemical reaction. For rapid and robust self-start it is desirable to operate the stack near the short-circuit conditions. This mode of operation maximizes hydrogen utilization, favors production of waste heat that is absorbed by the stack, and delays complete loss of effective electrochemical surface area by causing a large fraction of the ice to form in the gas diffusion layer rather than in the cathode catalyst layer. Preheating the feed gases, using the power generated to electrically heat the stack, and operating pressures have only small effect on the ability to self-start or the startup time. In subfreezing weather, the stack shut-down protocol should include flowing ambient air through the hot cathode passages to vaporize liquid water remaining in the cathode catalyst. Self-start is faster and more robust if the bipolar plates are made from metal rather than graphite.     

Diagnosis of PEM fuel cell stack dynamic behaviors

In this study, the steady-state performance and dynamic behavior of a commercial 10-cell Proton Exchange Membrane (PEM) fuel cell stack was experimentally investigated using a self-developed PEM fuel cell test stand. The start-up characteristics of the stack to different current loads and dynamic responses after current step-up to an elevated load were investigated. The stack voltage was observed to experience oscillation at air excess coefficient of 2 due to the flooding/recovery cycle of part of the cells. In order to correlate the stack voltage with the pressure drop across the cathode/anode, fast Fourier transform was performed. Dominant frequency of pressure drop signal was obtained to indicate the water behavior in cathode/anode, thereby predicting the stack voltage change. Such relationship between frequency of pressure drop and stack voltage was found and summarized. This provides an innovative approach to utilize frequency of pressure drop signal as a diagnostic tool for PEM fuel cell stack dynamic behaviors.     

Membrane degradation in PEM fuel cells: From experimental results to semi-empirical degradation laws

Chemical membrane degradation is one of the main lifetime-limiting factors for proton exchange membrane fuel cells. In this paper, the degradation of state-of-the-art catalyst coated membranes with an ePTFE reinforcement layers is studied under various operating conditions to quantify the impact of cell potential, pressure and humidity on degradation rate. The membrane electrode assembly has been monitored using various experimental techniques, including electrochemical techniques, post-mortem analyses and fluoride release measurements of exhaust water. Semi-empirical laws are proposed to quantitatively model the observed impact of cell potential and cathode oxygen partial pressure on fluoride release rate and membrane thinning. For the low humidity cases, the observed membrane thinning, the hole formation and the open circuit voltage drop are correctly captured by the proposed model by considering an initial default in the membrane where the membrane degradation, which is highly non-linear, will accelerate. Regarding humidity, a higher value does not reduce the degradation rate but seems to induce a more homogeneous repartition of the degradation preventing the formation of local pinholes and therefore extending the fuel cell lifetime.     

Numerical investigation of transient responses of a PEM fuel cell using a two-phase non-isothermal mixed-domain model

In this paper, a transient two-phase non-isothermal PEM fuel cell model has been developed based on the previously established two-phase mixed-domain approach. This model is capable of solving two-phase flow and heat transfer processes simultaneously and has been applied herein for two-dimensional time-accurate simulations to fully examine the effects of liquid water transport and heat transfer phenomena on the transient responses of a PEM fuel cell undergoing a step change of cell voltage, with and without condensation/evaporation interfaces. The present numerical results show that under isothermal two-phase conditions, the presence of liquid water in the porous materials increases the current density over-shoot and under-shoot, while under the non-isothermal two-phase conditions, the heat transfer process significantly increases the transient response time. The present studies also indicate that proper consideration of the liquid droplet coverage at the GDL/GC interface results in the increased liquid saturation values inside the porous materials and consequently the drastically increased over-shoot and under-shoot of the current density. In fact, the transient characteristics of the interfacial liquid droplet coverage could exert influences on not only the magnitude but also the time of the transient response process.     

Detaching behaviors of catalyst layers applied in PEM fuel cells by off-line accelerated test

The detaching behavior of catalyst layers in membrane electrode assembly (MEA) for PEM fuel cells could affect the lifetime of both catalyst layers and membranes. However, this issue is always neglected. Therefore, the study of detaching behavior of catalyst layer is very conducive to investigate the failure mechanism of fuel cells. The detaching of catalyst layers was simulated by dipping membrane electrode assemblies (MEAs) into H₂O₂ solution with or without Fe²⁺. We observed the presence of detaching of catalyst layers and found the varied detaching behaviors with different accelerated testing solutions: a layered-type detaching behavior is shown for the catalyst layer treated with 30% H₂O₂ solution, whereas a crack-like detaching behavior in the case of 30% H₂O₂ solution with Fe²⁺ species (or Fenton's test). At the same time, the layered-type detaching of catalyst layers has a higher detaching rate than the crack-like detaching. These detaching behaviors should have an inherent link to degradation of recast-ionomer (Nafion) films in catalyst layers. In addition, the effect of detaching behaviors of catalyst layers on the lifetime of fuel cells has been studied by hydrogen crossover measurement, and shows that, for the crack-like detaching, the membrane has a shorter lifetime than that for the layered detaching.     

Effects of inlet humidification on PEM fuel cell dynamic behaviors

The dynamic behaviors of a proton exchange membrane (PEM) fuel cell have been studied both experimentally and numerically. The objective of this paper is to investigate the effects of cathode inlet humidification on PEM fuel cell load change operations and the fuel cell performance during a simulated start‐up process. The PEM fuel cell was found to respond quickly and reproducibly to load changes. It was also found that an increase in the cathode inlet humidification significantly influences the start‐up performance of a PEM fuel cell. The cathode inlet relative humidity (RH) under 30% significantly dropped the cell dynamic performance. Extensive numerical simulations, with the transient processes of load jump and gradual changes considered, were performed to characterize dynamic responses of a singe‐channel PEM fuel cell under different inlet humidification levels. The results showed that the response time for a fuel cell to reach steady state depends on water accumulation in the membrane, which is consistent with the experimental results. Copyright © 2010 John Wiley & Sons, Ltd.     

Numerical simulation of mass and charge transfer for a PEM fuel cell

A three-dimensional, steady state, single phase model is developed to study the mass and charge transfer within a proton exchange membrane (PEM) fuel cell. A single set of conservation equations is used for all PEM fuel cell layers and the governing equations are solved numerically using a finite-volume-based computational fluid dynamics technique. The numerical results for the flow field, species transport and phase potential are presented for two designs, namely a PEM fuel cell with conventional and interdigitated flow fields for the reactant supply.     

Performance analysis of a new designed PEM fuel cell

In this paper, a new design for the flow channels is presented, and a parametric study of the proton exchange membrane (PEM) fuel cell is conducted in order to investigate the effect of the new flow channels, as well as different operating parameters, on the efficiency and energy output of the cell. Design parameters are selected based on studies presented in the literature to build a physical and practical model. With the new design of the flow channels, it is noticed that the cell efficiency increases from 33.8% to 47.7% if the temperature of the cell is increased. The power output of the cell increases from 2.6 to 282.5 W when the cell temperature and the current density are increased. Moreover, decrease in the efficiency of the cell ranges from 45.5% to 28.4% with the increase in the current density and membrane thickness. Based on the analytical model, design parameters were selected to manufacture a fuel cell that has a power output of 175 W and an efficiency of 35% running at 353 K and 3 bar, with an effective membrane area of 450 cm². Experiments are conducted to investigate the effect of newly designed flow channels on pressure distribution. It is found that when hydrogen is supplied from both inlets, pressure across the channels become symmetric and, therefore increasing the power output. This study reveals that, with the proper choice of design parameters, a PEM fuel cell is an attractive economical, efficient, and environmental solution when compared with conventional systems of power generation such as gas turbines. Copyright © 2011 John Wiley & Sons, Ltd.     

Investigation of the effect of graphitized carbon nanotube catalyst support for high temperature PEM fuel cells

In this study, it is aimed to investigate the graphitization effect on the performance of the multi walled carbon nanotube catalyst support for high temperature proton exchange membrane fuel cell (HT-PEMFC) application. Microwave synthesis method was selected to load Pt nanoparticles on both CNT materials. Prepared catalyst was analyzed thermal analysis (TGA), Transmission Electron Microscopy (TEM) and corrosion tests. TEM analysis proved that a distribution of Pt nanoparticles with a size range of 2.8–3.1 nm was loaded on the Pt/CNT and Pt/GCNT catalysts. Gas diffusion electrodes (GDE) were manufactured by an ultrasonic spray method with synthesized catalyst. Polybenzimidazole (PBI) membrane based Membrane Electrode Assembly (MEA) was prepared for observe the performance of the prepared catalysts. The synthesized catalysts were also tested in a HT-PEMFC environment with a 5 cm² active area at 160 °C without humidification. This study demonstrates the feasibility of using the microwave synthesis method as a fast and effective method for preparing high performance Pt/CNT and Pt/GCNT catalyst for HT-PEMFC. The HT-PEMFC performance evaluation shows current densities of 0.36 A/cm²0.30 A/cm² and 0.20 A/cm² for the MEAs prepared with Pt/GCNT, Pt/CNT and Pt/C catalysts @ 0.6 V operating voltage, respectively. AST (Accelerated Stress Test) analyzes of MEAs prepared with Pt/GCNT and Pt/CNT catalysts were also performed and compared with Pt/C catalyst. According to current density @ 0.6 V after 10,000 potential cycles, Pt/GCNT, Pt/CNT and Pt/C catalysts can retain 61%, 67% and 60% of their performance, respectively.     

Experimental and theoretical analysis of ionomer/carbon ratio effect on PEM fuel cell cold start operation

The effect of ionomer/carbon (I/C) ratio on proton exchange membrane (PEM) fuel cell cold start is investigated experimentally with theoretical water transport analysis. The scanning electron microscope (SEM) images show larger agglomerates and smaller effective reaction area by increasing the I/C ratio from 0.7 to 1.7. For normal operation, increasing the I/C ratio can improve the humidity tolerance, especially in the cathode. For cold start >?10 °C, a lower I/C ratio leads to better performance because the core reaction area is shifted towards the membrane, leading to more membrane water absorption and slower ice formation. For <?15 °C, the total water production is low and almost the same for the different I/C ratios because the ice formation takes place before effective membrane water absorption; and although the cathode catalyst layer (CL) and micro-porous layer (MPL) can provide sufficient space to store all the ice, higher I/C ratios (e.g. 1.2) still cause more ice formation in GDL and flow channel because the core reaction area becomes closer to GDL. The results show that the CL design has significant effect on the cold start performance, and there is a potential for further improvement.     

Cold pre-compression of membrane electrode assembly for PEM fuel cells

Direct compression from the land structure of bipolar plate in a PEM fuel cell is considered as an important factor for the higher performance under the land than under the channel areas. Therefore the objective of this study is to determine if a cold pre-compression treatment on the whole membrane electrode assembly (MEA) area may have a significant positive effect on the overall performance of the cell. Five different levels of cold pre-compression have been applied and the experimental results show that the overall performance of the cell first increases with the level of compression to a maximum, and then decreases. These results clearly show that cold pre-compression of the MEA can significantly enhance the performance of the entire cell and there exists an optimal level of compression. Results of electrochemical impedance spectroscopy (EIS) show that the cold pre-compression results in a significant reduction in charge transfer resistance, especially in the high current density region. Further study by the cyclic voltammetry (CV) shows that the electro-chemical area (ECA) is changed with the different cold pre-compressed MEAs and there exists an optimal compression that results in the maximum ECA.     

A two-phase non-isothermal mixed-domain PEM fuel cell model and its application to two-dimensional simulations

In this paper, a two-phase non-isothermal PEM fuel cell model based on the previously developed mixed-domain PEM fuel cell model with a consistent treatment of water transport in MEA has been established using the traditional two-fluid method. This two-phase multi-dimensional PEM fuel cell model could fully incorporate both the anode and cathode sides, properly account for the various water phases, including water vapor, water in the membrane phase, and liquid water, and truly enable numerical investigations of water and thermal management issues with the existence of condensation/evaporation interfaces in a PEM fuel cell. This two-phase model has been applied in this paper in a two-dimensional configuration to determine the appropriate condensation and evaporation rate coefficients and conduct extensive numerical studies concerning the effects of the inlet humidity condition and temperature variation on liquid water distribution with or without a condensation/evaporation interface.     

Electrochemical characterization of sulfonated PEEK-WC membranes for PEM fuel cells

Sulfonated PEEK-WC polymer was obtained according to chloro-sulfonic acid procedure making possible the preparation of different membrane samples with a sulfonation degree from 48 to 90%. Dense membranes were prepared from casting solutions of S-PEEK-WC dissolved in DMF. Proton conductivity measurements were performed on such S-PEEK-WC membranes in a range of temperature from 30 to 120 °C and 100% of relative humidity, reaching 2.5 × 10⁻² S cm⁻¹ as the best value at 100 °C and with a sulfonation degree of 90%. In the meanwhile, the open circuit voltage of this S-PEEK-WC membrane (DS = 90%) varied from 0.963 V at 60 °C to 0.802 V at 100 °C, demonstrating that an increase of temperature negatively affects the membrane performance due to the mechanical properties degradation. In this work, a wide experimental campaign was carried out to investigate the electrochemical performances in terms of polarization curves, open circuit voltage and proton conductivity of S-PEEK-WC membranes as well as the fuel crossover and water uptake.     

Prediction performance of PEM fuel cells by gene expression programming

In the present study, gene expression programming has been utilized to evaluate the output voltage of different PEM fuel cells as the performance symbol of these structures. A total number of 843 data were collected from the literature, randomly divided into 682 and 161 sets, and then trained and tested, respectively by different models. The used data as input parameters were consisted of current density, fuel cell temperature, anode humidification temperature, cathode humidification temperature, operating pressures, fuel cell type, O₂ flow rate, air flow rate and active surface area of the PEM fuel cells. According to these input parameters, in the gene expression programming models, the voltage of each PEM fuel cell in different conditions was predicted. The training and testing results in the gene expression programming model have shown an acceptable potential for predicting voltage values of the PEM fuel cells in the considered range.     

Fast start-up of a diesel fuel processor for PEM fuel cells

Fuel cell systems based on liquid fuels are particularly suitable for auxiliary power generation due to the high energy density of the fuel and its easy storage. Together with industrial partners, Oel-Waerme-Institut is developing a 3 kW_el PEM fuel cell system based on diesel steam reforming to be applied as an APU for caravans and yachts. The start-up time of a fuel cell APU is of crucial importance since a buffer battery has to supply electric power until the system is ready to take over. Therefore, the start-up time directly affects the battery capacity and consequently the system size, weight, and cost.     

Numerical study of serpentine flow‐field cooling plates on PEM fuel cells performance

The effect of a cooling plate on a PEM fuel cell was studied by three‐dimensional CFD modeling. The cyclic cell and the single cell were compared for the evaluation of the influence of cooling plate. The cyclic cell consisted of a single cell and a two‐channel serpentine flow‐field coolant, which then repeats by using a cyclic boundary on both ends. The single cell was composed of an active area of 200 cm² and a 10‐channel serpentine flow field. The following sets of equations were used in the model: the conservation of electrical current, the mass conservation of gases species, the Navier–Stokes equation, the energy balance, and the water phase change model. Comparison of cyclic cell and single cell shows that the voltage of cyclic cell was reduced at high current densities because of the increased ohmic losses. This was caused by the combined effect of membrane dehydration and higher local temperature. However, the cyclic cell showed more uniform current density distribution than the single cell, and this is attributed to the use of cooling plate. Increasing the coolant flux enhanced the cell performance by reducing the ohmic loss. Copyright © 2011 John Wiley & Sons, Ltd.