The flow field structures in fuel cells have to meet several requirements, such as homogeneous fluid distribution, product water removal, good current transport, good under-rib convection and high conductivity. Each factor can considerably influence the performance of fuel cells. However, this work, as mentioned previously, is focused on the effects of flow field structure, cell conductivity and stack assembly on the performance of micro PEMFC.
Yu et al. [2] successfully fabricated a miniature fuel cell on silicon wafers using micro-electronic fabrication techniques, which included photolithography, dry and wet etching, chemical and physical vapor deposition. They found that sputtering different thicknesses of Cu/Au (including 0.5μmAu, 0.2μm Au plated 1.4μm Cu, and 0.9μm Au plated 1.5μm Cu) composite layer on the top of the silicon wafer as a current collector can reduce the resistance, indicating that the cell performance is improved by increasing the thickness of the composite layer on the silicon wafer.
Meyers and Maynard [3] used MEMS techniques to fabricate micro flow channels of PEM fuel cell on silicon substrate. Their designs consisted of bipolar design, similar to the standard cell sandwich commonly used in PEMFC system, and monolithic design, essentially an
5
“unfolded” fuel cell with the anode and cathode on the same substrate (see Fig. 1.4). By making comparison, they found the former shows a better performance than the latter as illustrated in Fig. 1-5. The research also suggested that many system-level issues, like thermal management, air movement, fuel delivery, humidification control, water management, power load management and system integration, must be considered in order to achieve higher performance.
Spiegel et al. [4] used deep reactive-ion etching process to fabricate the serpentine microflow fields in the siliconwafer with the fuel cell channel widths and depths ranging from 20 to 200μm to compare with the conventional ones using traditional CNC machining processes with flow field channel dimensions of 500 and 1000μm. With the same MEA and percentage of active area (channel to rib ratio of 1:1), and under the same test conditions, the fuel cells could have the best performance with the 20-μm of width, depth and rib because such dimensions allowed increase of channel velocity, rapid diffusion and homogenous reactant distribution along the flow channel. Therefore, smaller flow channel dimensions (>100μm) appear to be promising for future micro fuel cell technologies.
Kim et al. [5] used photolithography, anisotropic wet etching, anodic bonding and physical vapor deposition to manufacture a miniaturized PEM fuel cell with silicon separators. A 400μm × 230μm flow channel was made with KOH wet etching on the front side of a silicon separator, and a 550nm gold current collector and 350nm TiNx thin film heater were respectively formed on the front side and the opposite side by PVD. Two separators were assembled with the membrane electrode assembly having
6
a 4cm2 active area for the single cell. With pure hydrogen and oxygen under atmosphere pressure without humidification, the performance of the single fuel cell could lead to 203mW/cm2 at 0.6V at room temperature.
Xiao et al. [6] fabricated a silicon/glass based micro fuel cell system by micromachining technique. The anode and cathode catalyst layers were formed by directly sputtering platinum on the ICP-etched (Inductively Coupled Plasma) high-aspect-ratio columns on silicon substrate. Integrated gold-based micro current collectors were patterned on the silicon and glass surfaces. The high-aspect-ratio columns effectively increased the catalyst surface area, and the micro pillars grown by their etching process further improved the cell performance.
Hsieh et al. [7] developed a novel design and microfabrication for a micro proton exchange membranes with a cross section area of 5cm2 and thickness of about 800μm. It consisted of PMMA flow field plate with narrow and deep channel made by microsystem technology, platinum sputtering deposited on MEA, and an ultra thin copper layer again sputtering deposited on PMMA flow field plate used as a current collector.
For the single cell, a reliable power output was obtained.
Hsieh et al. [8] also developed a SU-8 photoresist microfabrication process for micro proton exchange membrane fuel cell flow structures for both anode and cathode flow field plates with a cross section area of 5 cm2 and thickness of about 750μm. The new design for flow field plates would have SU-8 used not only a photoresist but also as a microstructure material. Their fabrication could make a low-cost and high-mass production of small flat single fuel cell with an acceptable power density
7
of about 30mW/cm2 at 0.35V.
Keyur et al. [9] firstly used polydimethylsiloxane (PDMS) as the substrate to fabricate the miniature hydrogen-air PEM fuel cell. In this work PDMS microreactors were fabricated by micromolding on the silicon mold made by photolithography and ICP etching process. Though the performances were still not competitive with standard ones, it did show the feasibility of producing micro PEM fuel cells by using the cheaper base substrates like PDMS in order to minimize their cost.
Song et al. [10] developed a simple and rapid method to fabricate planar PEM fuel cells in PDMS. They patterned a perfluorinated ion-exchange resin, such as a Nafion resin, on a glass substrate using a reversibly bonded PDMS microchannel to generate an ion-selective membrane between the fuel-cell electrodes.
Siu and Chiao [11] used PDMS as the material of gasket and electrode in microbial fuel cell. They used MEMS technology, such as etching and evaporation process, to form the microchannel pattern on the silicon wafer, and molding PDMS on the wafer. Compared with the recent silicon micromachined microbial fuel cell, their result showed the better performance in the average power density and average current density.
Different from the traditional "banded" way, Lee et al. [12] used the
"flip-flop" method on the fuel cell stacks. They applied a variety of etching and deposition techniques adopted from microfabrication on glass and silicon substrates in two-cell and four-cell configurations. The current collection was achieved by pattering metal films on an insulating substrate. They found that film thickness is the dominant factor compared
8
to other parameters related to channel topology. Though this new idea can simplify the interconnect design, the fuel may become difficult to supply.
Zhang et al. [13] fabricated a 6-cell PEMFC stack combined with a small hydrogen storage canister. Each cell was made by sandwiching a membrane-electrode-assembly (MEA) between two flow field plates fabricated by a classical MEMS wet etching method using silicon wafer as the original material. The plates were made electrically conductive by sputtering a Ti/Pt/Au composite metal layer on their surfaces. The 6-cells lay in the same plane with a fuel buffer/distributor as their support, which was fabricated by the MEMS silicon–glass bonding technology. The performance obtained a peak power of 0.9W at 250mA/cm2.
Chen et al. [14] designed the 10-cell planar array stack (6cm×6cm×
0.9cm) and compared the performance of the parallel connected one to that of the serial connected one. They found that the parallel connected stack is better because it will not be affected by certain cell with worse performance, and both performances can be enhanced by force convection. They also found the uniformity of each individual cell in serial connected stack that using force convection is better than that using natural convection.
Kim et al. [15] used thin flexible printed circuit board (FPCB) as a current collector in order to reduce an air-breathing monopolar stack's volume. They also designed different patterns of the air-breathing holes on the cathode to find the effect of varying the geometry and opening ratios on stack performance. They found that in cathode the circular-hole pattern with opening ratio of 38% has the best performance, but also
9
causes cathode flooding and unstable output problems. And the rectangular cathode opening pattern with 65% opening ratio was chosen for high performance and voltage stability.
Cha et al. [16] observed the scaling effects of various flow channels in fuel cells with gaseous hydrogen/air reactants from macro feature size (>500μm) to micro one (<100μm). They found that scaling behavior is quite complicated due to highly non-linear convection both in the flow channels and porous electrode. Considering the model predictions, flooding issues, and pressure drop losses, the performance of interdigitated channels decreases as the feature size decreases. Therefore, a good compromise may be found between the reduced pressure drop and reduced performance at intermediate feature sizes depending on operation requirements.
Noponen et al. [17] introduced a measurement system for mapping of current distribution into a free-breathing polymer electrolyte membrane fuel cell. The result showed that the operating temperature has a significant influence on the fuel cell performance. Furthermore, at low cell temperatures the limiting factor is inadequate free convection, and at high temperatures it is drying. They also showed that under some condition this type of fuel cell has fairly homogenous current distribution and therefore does not need any auxiliary pumps or fans to increase the airflow.
Sun et al. [18] applied the technique of current density distribution measurement gasket to measure local current distribution in a PEM fuel cell with serpentine flow field at various humidification temperatures.
Their experiments showed the following results: Whether the anode or
10
cathode, the performance or current density can reach the maximum when humidification temperature is equal to the cell temperature. Besides, when the anode and cathode humidification temperatures are equal, if the cell is highly under-humidified, then local current density starts very low and increases monotonously along the channel; if the cell is moderately under-humidified, then the local current density first increases, reaches a maximum, and then decreases along the channel; when the cell is well-hydrated or over-hydrated, the local current decreases monotonically along the channel.
Lee et al. [19] investigated the effect of changing the gas diffusion layer and bolt torque on the performance of a PEM fuel cell at fixed flow rates. The experiment results showed that brittle gas diffusion layer material has higher performance with a less bolt torque because the higher bolt torque may damage the diffusion layer. And when soft material is combined with the more brittle one, the same tendency can be observed. As a result, the bolt torque and the gas diffusion layer type are the important factors for performance of PEM fuel cell.
Fekrazad and Bergman [20] used a three-dimensional model of PEM fuel cell stack to predict the influence of nonuniform stack compression on thermal and electrical contact resistances at the BP-GDL. The result showed that the temperature distribution within the membrane is highly dependent on the clamping pressure distribution. Also, the application of nonuniform clamping pressure distributions can make thermal conditions within the stack to become more uniform. However, it has negligible impact on the fuel cell power output or maximum membrane temperature.
Lu and Reddy [21] combined the experimental and modeling
11
methods to investigate effects of different factors on the performance of the micro-PEM fuel cell (μ-PEMFC). Their results showed that among three types of resistance influenced by material, contact and thermal effects, contact resistance caused by the assembling mode of the μ -PEMFC contributes to 19.4% of the total inner resistance, which is much higher than the 2% of the others. Therefore, the designs of new flow field configurations and assembling modes are very important in improving the performance of μ-PEMFCs.
Lin et al. [22] investigated the gas permeability, bulk density, thickness and conductivity of two types of gas diffusion layer (OC14, NC14) as a function of the compressed thickness with an active area of 25 cm2 in a single PEMFC. They found that increasing compression of gas diffusion layer can produce high-quality contact, but excessive compression will damage the carbon fiber, reducing the gas permeability and contact resistance. These results concerning the balance between compression and performance provide vital information for the fabrication of stacks and support for industrial applications.
Chang et al. [23] used a special-designed test rig to measure the thickness, gas permeability, and porosity of a GDL sample under various clamping pressure conditions. Their results showed that at low clamping pressure levels (e.g.< 5 bar), increasing the clamping pressure reduces the interfacial resistance between the bipolar plate and the GDL that enhances the electrochemical performance of a PEM fuel cell. In contrast, at high clamping pressure levels (e.g. >10 bar), although increasing the clamping pressure can reduce the above ohmic resistance, it also narrows
12
down the diffusion path for mass transfer from gas channels to the catalyst layers. Comprising the above two effects does not promote the power density too much but reduces the mass-transfer limitation for high current density.
Zhou et al. [24] developed a model to analyze the effects of assembly pressure, operating temperature and humidity on PEM fuel cell stack deformation, contact resistance, overall performance and current distribution. The modeling results revealed that elevated temperature and humidity enlarge gas diffusion layer and membrane inhomogeneous deformation, increasing contact pressure and reducing contact resistance due to the swelling and material property change of the GDL and membrane. When an assembly pressure is applied, the fuel cell overall performance is improved by increasing temperature and humidity.
However, the stack would be more prone to degradation with significant variation of current distribution at elevated temperature and humidity.
Wen et al. [25] experimentally investigated the effects of various combinations of bolt configuration and clamping torque on the corresponding contact pressure distributions and performances of a single PEM fuel cell and a 10-cell stack. Their results showed that, for the single cell under the current experiment conditions, the larger mean contact pressure tends to yield the higher maximum power, regardless of the bolt configuration and the applied torque. The uniformity of the contact pressure distribution, the ohmic resistance and the mass transport limit current have highly linear correlations with the mean contact pressure.
However, in the case of the 10-cell stack, the effects of various combinations of bolt configuration and clamping torque on its