Cold-starts of a polymer electrolyte fuel cell (PEFC), isothermally maintained at various subfreezing temperatures, were visualized using high resolution dynamic in-plane neutron imaging. The results obtained aim to bring new knowledge about the water accumulation mechanisms leading to the voltage drop usually observed in isothermal mode after a given working time, also called voltage failure. In particular, the data presented should be useful for comparison to simulation predictions provided by modeling studies. As main result, water in a condensed phase was observed to accumulate not only in the membrane-electrode-assembly (MEA), but also in the cathode gas diffusion layer (GDL) at −15 • C and even in the cathode gas channels at −10 • C. Moreover, approximately 400 cold-starts were realized without neutron imaging and revealed stochastic distributions of working times. The presence of water in super-cooled state is discussed and finally retained as only valid explanation of the results obtained. Additionally, the sudden freezing of super-cooled water is thought to cause the rapid water accumulation observed in the MEA during the voltage failure.

Systematic variations of current density and asymmetric relative humidity are realized on differential Polymer Electrolyte Fuel Cells (PEFCs). The water distribution is visualized and quantified by high resolution in-plane imaging, and its effect on performance is discussed. Two cells are investigated: one using paper type gas diffusion layers (GDLs) and one with cloth type GDLs. The novel output of this work is an extensive database of local measurements obtained by the differential cell approach and an asymmetric variation of relative humidity. This should be particularly valuable for the validation of modeling studies. The negative impact of water accumulation on the performance is clearly stronger for the paper type GDLs than for the cloth type GDLs. On the contrary, the performance of the cell with cloth GDLs is low in dry conditions. The accumulation of water in the channel region of the cathode GDL has a crucial impact on performance. The anode side is observed to play an important role for water removal. The presence of a maximal liquid water saturation in the GDL for a wide range of current densities is observed and discussed. In-plane water distribution profiles are presented for the channel and rib regions for all conditions.

No abstract

A novel experimental method for the measurement of mass transport losses in a polymer electrolyte fuel cell (PEFC) was developed, based on the comparison of cell voltage during operation with air, helox, (79% He, 21% O 2 ) or pure O 2 . Using only short periods (2 seconds) of helox or pure O 2 operation, the disturbing artifacts (dry out or change of catalyst surface oxidation state) associated with continuous operation with these gases were avoided. High resolution neutron imaging was conducted simultaneously in order to study the relation between mass transport loss and liquid water accumulation. For the studied design, comparisons in steady state and dynamic experiments indicated a larger impact of water accumulation in the gas diffusion layers (GDL) under the flow channels, or in the portion of the channels near the GDL surface. Increased diffusion losses were observed, as expected, under high humidity conditions, but also at very low humidity conditions in absence of any liquid water. The latter was explained by a redistribution of current toward the rib region, having longer diffusion paths.Polymer Electrolyte Fuel Cells (PEFCs) are energy converters, bearing attractive characteristics for automotive applications in terms of efficiency and absence of pollutant emissions. The dimension of an automobile fuel cell system is primarily determined by the required peak power. In consequence, reaching the highest possible power density allows a reduction of the system size, weight, and cost. At high current density operation, losses introduced by the limitations in the diffusive transport of oxygen through the gas diffusion layers (GDLs) can become significant. The presence of liquid water in the gas diffusion layers reduces the effective diffusivity of such media and results in increased losses.The visualization of liquid water in operating PEFCs using different methods has been increasingly reported. A recent review on in situ visualization methods can be found in a publication by Tsushima et al. 1 Besides its high contrast for liquid water, neutron imaging 2-20 offers the advantages of good transparency of usual fuel cell materials (including metals), and of a negligible impact of the non-ionizing radiation on cell operation, making it an excellent non-invasive method. Water visualization in fuel cells has been a strong driving force for addressing the issue of the limited spatial resolution of neutron imaging. [19][20][21][22][23] The approach used at the Paul Scherrer Institut, based on optimized optical setup and scintillator screens, 24 combined with specific improvements for fuel cell imaging, 20,21 allowed reaching the combination of an effective spatial resolution of 20 μm with exposure times of 10 seconds. Thus, the measurement of liquid water content in different layers of a fuel cell is possible, not only in steady state, but also in transient operation. In order to assess the impact of liquid water on cell performance, a method for directly measuring diffusive transport losses is of high interest. For...

By means of advanced diagnostic tools -neutron imaging, pulsed gas method (helox or O 2 pulses during air operation) and electrochemical impedance spectroscopy (EIS) -three different flow fields are characterized in a differential (small-size cell operated at high stoichiometry) Polymer Electrolyte Fuel Cell (PEFC) to emulate the local behavior of a full-size cell. The channelrib structures are designed so as to obtain 2D, 1D and 0D gradients of O 2 concentration across the gas diffusion layer (GDL). The important bulk-diffusion losses (evaluated with pulsed helox) of the "2D" cell compared to the "1D" cell indicate the effect of lateral oxygen diffusion in the GDL. Non-bulk diffusion losses (Knudsen and/or thin film diffusion, evaluated with pulsed O 2 ) also play a non-negligible role on the overall diffusive losses. In particular, their increase in dry conditions for the "1D" and the "0D" cells might be a consequence of drying effects in the electrode. EIS spectra can be interpreted consistently with the results of the pulsed gas method: the increase of arc diameter can be correlated with the overall mass transport losses including both bulk and non-bulk diffusion losses.In order to increase the market attractiveness, the cost of the Polymer Electrolyte Fuel Cell (PEFC) must be reduced. This can be achieved by lowering the cost of the components and by enhancing the performance. The so-called flow field plate (also referred to as bipolar plate) is an essential component of the fuel cell used for the reactant distribution, the water removal, and the electrical and thermal collection. The cost of the flow field must be minimized (material, manufacturing and coating) and the influence on performance must be improved. In this context, the flow field local structure (meaning here the shape and the size of the channels and ribs) is a crucial set of parameters. Under constraint of low manufacturing costs, it must ensure a homogeneous reactant distribution and a good water management, in terms of efficient water removal and sufficient membrane hydration. Numerous studies were realized in the last decade to investigate the effect of the flow field pattern on performance (see review papers 1,2 ). The local distribution of parameters over the active area was modeled 1 or measured in segmented cells. 3,4 The distribution and dynamics of water droplets in the channels were visualized for instance by optical or neutron imaging, 5-12 to test different fuel cell designs 7-11 or to search correlations with the local current density on segmented cells. 5,6,12 The majority of these studies were realized on full-size cells supplied by gases at technical stoichiometry (<2). Since local operation parameters are not controlled, the local impact of different flow field patterns on the performance cannot be directly tested. In addition, the local influence of the water content on the performance is difficult to understand in such cells. Indeed, liquid water droplets in the channels can impact the gas flow distribution at the full-size...