Commercialising High Volume Fuel Cell Production

Fuel cell technology is now wellproven, through extensive laboratory and prototype testing. It is known to work, and to be effective and usable in many applications. Of course the technology must continue to develop, in order to deliver further power density and efficiency enhancements. However, it is imperative that the unit cost of fuel cells themselves should be reduced – ultimately to consumer commodity levels. Manufacturing capacity must also increase dramatically to meet anticipated world-wide demand.

Lower costs and higher production volumes are closely linked. Hence, fuel cell production methods must become faster and more highly automated. For this, the industry needs high-speed, accurate, repeatable, cost-effective equipment, capable of supporting high-yield processes that will deliver large numbers of high-quality fuel cells, quickly, and at low unit cost. (Figure 1)

Process Challenges
The dominant fuel cell chemistries of the moment are the Proton Exchange Membrane (PEM), solid oxide and direct methanol. Their constituent parts - particularly the membrane, electrodes and backing layers of a PEM fuel cell – require very thin deposits of materials such as perfluorocarbon sulfonates (PFSA) and platinum/carbon/ionomer catalysts. The physical and chemical properties of these layers are critical to achieving fuel cell action. Variations in the characteristics - such as changes of thickness or the existence of voids or pin holes in the PEM cell’s membrane, for example - can impair performance and lifetime and may permit catastrophic failure by permitting uncontrolled mixing of hydrogen with oxygen. Hence these materials must be deposited in controlled and repeatable quantities onto a substrate that is typically flexible and may also be porous.

Manufacturing Challenges
Ultimately, the cost per kW of energy produced is the deciding factor that will determine the success or otherwise of the fuel cell age. Hence, the major challenges facing this nascent industry are rooted in ensuring sufficient throughput and yield to minimise the cost per unit produced. Highspeed techniques for depositing liquid compounds of varying viscosity are required. Preferably these should be capable of producing many units simultaneously, within a short cycle time.

Suitable techniques should also be easily integrated into a highly optimised production sequence. This will allow fuel cell manufacturers to maximise production efficiency and minimise human intervention.

Another issue that can have an appreciable effect on the overall cost of each fuel cell produced is the volume of material wasted after each process. Some electrodes and electrolytes are formed using spray deposition techniques, but these tend to waste relatively large volumes of material as overspray. Vapour deposition promises lower levels of wastage, but the need to create deposits in specific patterns complicates this approach.

In addition, manufacturers need to be able to swap quickly between different combinations of deposit and substrates, to create batches of components for a set of fuel cells, and to support a number of fuel cell technologies on the same set of equipment.

Sounds Familiar…
The challenges outlined bear striking similarities to those faced by the thick film electronics industry of the early 1970s and – more recently - surface mount (SMT) assembly, as the office and consumer electronics revolutions took hold. The solutions adopted, such as precision screen printing of thick film polymer inks, or simultaneous deposition of large numbers of solder-paste deposits, was instrumental in allowing assembly to be outsourced to specialist electronic manufacturing services (EMS) businesses. This model has enabled year-on-year reductions in the cost of everything from domestic appliances to TVs to mobile phones, while the capabilities of those products has increased immeasurably over the same period. If the fuel cell community could replicate that level of success, the world may contemplate a much cleaner, more sustainable future.

Proven Solution
But there appears to be a fly in the ointment. The resolution required for effective production of fuel cell components is much finer than that for depositing solder paste onto pads for mounting electronic components to a PCB. The thickness of the deposits is also much thinner, and the characteristics of fuel cell materials are also very different from those of solder paste, although not so different from those of polymer thick film materials.

However, screen printing equipment and processes have now reached a very high level of capability and refinement. Semiconductor manufacturers themselves are successfully using the latest generations of equipment to apply finelymetered quantities of diverse materials at silicon chip level, as well as to the interconnects and packaging surrounding the chip.

Screen printing is now a proven solution to depositing chip-attach epoxies, placing solder balls of 0.3mm diameter to create the latest flip chip and ball grid array packages, and even applying low-viscosity materials such as thermal interface material (TIM), lid seal or transfer moulding. Screen printing is also being used for backside wafer coating - applying a very thin and uniform layer of partial-cure adhesive to an entire silicon wafer of up to 300mm diameter, containing many individual chips.

So the accuracy, resolution, repeatability and materials handling capabilities of screen printing are known to be sufficient to offer an effective solution to producing fuel cell components. What is more, mature, reliable, highly automated platforms with conveyorised input and output enabling fast cycle times and easy integration into high-volume automated assembly lines are proven and readily available. These machines use highspeed linear motors and accurate, repeatable position encoders, to achieve high positional accuracy and repeatability, at high rates of throughput with few stoppages and very low human intervention.

Screens and Stencils
One of the most critical elements of a screen printing process is the screen or stencil itself. Its properties define the size, thickness and shape of the deposit. In a metal stencil, the characteristics of the apertures are critical. In an emulsion screen, the pitch and thickness of the mesh, and the properties of the emulsion, have the heaviest influence over results. The choice individual requirements of the application determine whether a metal stencil or emulsion screen will produce the best results. When depositing small volumes of material at close spacing, such as precision solder paste deposits for fine-pitch ICs, a metal stencil is optimal. For coating the backside of a semiconductor wafer, on the other hand, an emulsion screen may be preferable.

Both technologies can offer a solution in fuel cell manufacture. For example, each is capable of depositing polymer materials of thickness 50 µm or lower, within ±12.5 µm total thickness variation. This is suitable for producing the ionomeric membrane for a PEM fuel cell, for example, which is typically between 50-175 µm thick. The platinum/carbon/ionomer catalyst layers, 5-50 µm thick, as well as the wet-proofed porous carbon paper backing layers typically 100-300 µm, are also candidates for high-speed, precision manufacturing using emulsion screen or metal stencil technologies.

One of the key advantages of this approach, besides outright speed, is the fact that very little material is wasted compared to, for example, spray or vapour-phase deposition. The doping materials for fuel cell components, particularly the platinum catalyst, are expensive, and wastage will have a significant impact on unit cost when production goes commercial.

Other characteristics of the deposit are also easier to control when using a printing process. When depositing materials using, for example, a dispenser or syringe-type arrangement, controlling the shape of the deposit is a complex challenge. Uniform thickness is also difficult to achieve.

The printing approach, which forces the material onto the substrate through apertures in the screen or stencil using a squeegee or enclosed printing head, creates a deposit of more uniform thickness.

Also, since the size and shape of the deposit is defined by the size and shape of the aperture, complex shapes can be formed quickly and easily, and shapes are extremely repeatable. Deposits can also be located very close together if required; DEK’s emulsion screens are capable of printing deposits separated by only 50 µm.

Enclosed head printing
As well as enhancements to resolution and dimensional stability of screen printing, delivered by the latest screen and stencil technologies, the arrival of enclosed head printing in 1997 has provided a significant enabling technology for screen printing processes addressing fine-pitch SMT and semiconductor applications. DEK has developed its ProFlow enclosed head printing system to produce a number of head variants including low volume heads for expensive materials such as via-fill materials or low-alpha solder pastes (Figure 2). The DirEKt Ball Placement™ head, for placing solder balls to create ball grid array packages, was also introduced recently for semiconductor packaging applications. ProFlow also allows screen printing of other materials such as polymer thick film conductive inks, which can be difficult to control under production conditions.

First Fuel Cell Successes
Several successful processes have already been implemented, hosted on the DEK Europa screen printing platform and leveraging precision emulsion screens combined with specially developed tooling for flexible substrates. A new enclosed printing head for fuel cell applications has also been successfully adopted. This leverages DEK’s ProFlow technology, storing the compound for deposition in a pressurised tank, and permitting a finely metered volume to be transferred through the screen onto the substrate below at uniform thickness and with no voids or pin-holes.

Screen printing has further economic advantages, in addition to high throughput and automation. The equipment can easily be reconfigured to produce a variety of components, simply by changing to a different material, fitting the appropriate screen, and loading new process settings either directly or via a network. A changeover can be completed in a matter of minutes.

Conclusion
Screen printing presents attractive advantages as a high-throughput, highly-automated, inline solution for creating fuel cell components. Suitable platforms for hosting such processes are already in production.

Suitable screen and stencil technologies are in place, and enclosed head printing enables control of the diverse constituent materials for manufacturing fuel cell electrodes and membranes in each of the dominant technologies entering use in the near future.