 |
Laboratoire d'Energétique et de Mécanique Théorique et Appliquée |
 |
The 1 kW bench set at
INPL
allows testing fuel cells in real operating conditions, in close partnership with the manufacturers.
This is the case for instance of the studies carried out within the framework of the
SPACT80 and ARTEMIS national projects.
In a different context, the fuel cell research group
participates to the design of the fuel cell powering the vehicle of students of
ESSTIN participating in the
Shell Eco-Marathon.
As a natural extension of our
scientific activities, it was decided to design and implement a 1.5 kW
(2 kW in peak) fuel cell, intended to produce the electric energy of
the vehicle built by students of Nancy-University
to participate in the Shell Eco-Marathon
in the urban concept
category.
Consequently, special attention was given to the simplicity, compacity and efficiency of the fuel cell:
it keeps a 2D geometry for gas supply it is self-humidified. It is air-cooled but without internal cooling channel,
which requires a careful optimisation of the bipolar plates geometry and materials.
By varying the number of cells (of 130 cm²) of the stack, it is possible
to estimate scale effects regarding gas distribution, electrical peformances and thermal design.
Transparent fuel cells allow to observe water condensation in the gas channels. They can also be designed with a simplified geometry, for instance bidimensionnal in term of gas flows, which can be parallel in the plane of the membrane. The other direction of mass fluxes is perpendicular to the membrane. In the fuel cell presented below, gas distribution is made through five 30 cm straight channels of 1×1 mm² in which gases can flow in the same (co-flow) or opposite (counter-flow) directions. The plates are made of PMMA, which is transparent but non conductive. The Paxitech Membrane Electrode Assembly (MEA) is divided into twenty 1.4 cm segments, electrically connected at one electrode only. The 30×1 cm² membrane (Nafion 115 or Nafion 112) is common for all segments. The current collection is realized through perimetric gold wires: 2 for the connected segments and 40 (2 per segment) for the other electrode. The wires are tightened between the backing layer and the PMMA plate. They pass through 20 Hall effect sensors for current density measurements and are then connected to the same potential. The cell temperature is controlled by 4 water flows (2 per plate) parallel to the gas channels. However, since PMMA is not a good conductor of heat (l = 0.17 W/m/K), local variations in the temperature can be significant: they are measured thanks to 20 thermocouples (one for each segment) inserted into a channel rib, as close as possible to the backing layer (in practice, the remaining PMMA thickness is 0.1 mm). Finally, an additional observable is the location of liquid water appearance in the channels.
The main results obtained with the bidimensionnal transparent fuel cell are illustrated in the following figures,
which depict the time variations of the average current density and temperature as well as their variation along the channels.
The three graphs below show that the pressure drop, the mean current density and the mean temperature
vary in two stages following fuel cell start-up:
These two operating modes illustrate very well the importance of heat and water management in polymer membrane fuel cells.
The two graphs below show the importance of local variations in temperature and current density, which confirms the prediction of
the models
[Lottin et col., 2008].
The location of highest current density correspond to the appearance of liquid water:
upstream, the membrane in less hydrated whereas downstream, the presence of liquid water in the channels
(and probably on the electrodes) restrains the access of hydrogen and oxygen towards the active layers.
Once again, a correlation between temperature and current density is observed but this time,
there is a correspondence between their maxima: the higher the temperature, the higher the local current density.
These results are published in Journal of Power Sources [Maranzana et col., 2008]. Currently, we are developping the next generation of instrumented fuel cells, which show performances equivalent to those of commercial products.
The design of the fuel cell presented above implies that there is only one gas channel per electrode. Thus, it makes it possible to study the effects of water condensation, since droplets tend to plug the channels before being removed by the gas streams. The recordings of the current intensity collected by all of the 18 channel ribs show clearly that the accumulation of liquid within the channels can unbalance the operation of the cell by reducing strongly the current inensity downstream. The sequence depicted below lasts for 6 secondes. The effects of liquid water become observable at t=2.2s and increase progressively until t=4.1s. Then, the droplet is evacuated and the current intensity profile goes back to its initial shape at t=4.6s. The fuel cell was operated at constant (total) current, so that one can observe a decrease in voltage while the channel is plugged.
Finally, electrochemical impedance spectroscopy is used to emphasize heterogeneities in behaviour between the various segments of the instrumented fuel cell: depending on the operating parameters, dry conditions can be observed at the beginning of the gas channels -following the gas flows- while the last segments may suffer from an exces of liquid water and/or from low reactant concentration.
The need to build fuel cells in LEMTA appeared with NMR imaging for which it is necessary to dispose of a small fuel cell adapted to the size of the probes (a few cm²). Magnetic Resonance Imaging (MRI) of a whole fuel cell is difficult to achieve because it requires a PEMFC free of ferromagnetic species and with a minimum of paramagnetic materials; moreover, further complications arise because of the electric conductance of most of fuel cell components (bipolar plates, backing layers, electrodes). This PEMFC must operate in conditions as close as possible to those prevailling in practice (especially in term of current density, but also in terms of gas hydration and stoichiometry). The first tests were conducted with fuel cells using carbon bipolar plates and Membrane Electrodes Assemblies (MEA) manufactured by Paxitech with Nafion 112.
Unfortunately, the carbon fuel cells do not allow to obtain satisfying images because the carbon bipolar plates shield the RF signal emitted by the hydrogen atoms. It is possible to obtain much more convincing images by using fuel cells without bipolar plate but with their MEA inserted between plates made of non-conductive materials. The gas channels are machined in these plates. In this case, it is very beneficial to use a transparent material (like PMMA) that makes it possible to visualize the presence (or the absence) of liquid water in the channels. Such kind of assembly are suitable only for fuel cells of small surface with which the electric current can be collected thanks to a gold wire applied to the periphery of the backing layers. Click here for a presentation of this work, published in Journal of Hydrogen Energy and Comptes Rendus Chimie.
Small fuel cells of a few cm² are also used for testing various protecting materials that can be applied on metallic bipolar plates (in stainless steel or aluminium). The protective layers are deposited by PVD (Physical Vapour Deposition) within the framework of a project carried out in common with the Laboratory of Chemical Engineering Sciences (LSGC) and the Laboratory of Surface Engineering (LSGS) and funded by Institut National Polytechnique de Lorraine (INPL). The performances and ageing of fuel cells using these PVD coated bipolar plates are studied and characterised in the laboratory.
