The pros and cons of miniaturisation

Fritz Prinz, Professor of Engineering at Stanford University reports on new downscaled manufacturing techniques.

As PCs, phones, and PDAs continue to shrink, miniaturisation plays an ever bigger role in our lives. Sub-micron manufacturing has brought us operating frequencies unimaginable a few years ago.

Now, miniaturisation is moving into other product areas, such as turbo-machinery and fuel cells. Here, though, while promising performance benefits, it can also exact a price.

Recently, ideas from the miniature world have been taken up enthusiastically by other disciplines. A good example is MEMS (Micro-Electronic Mechanical Systems). These systems draw on existing micro-electronic fabrication techniques to create miniature sensors and micro-actuators. One advantage of MEMS is that they can be arranged relatively easily in parallel to improve reliability. However, as size goes down, precision loss and frictional losses go up.

Against such a background, the implications of MEMS manufacture are obviously worth a closer look. Two areas in which MEMS and other precision fabrication techniques are already making their mark and which demonstrate these implications well are micro-engines and micro-fuel-cells.

Turbo Scaling implications

A mechanical system can be described by its mass, distance, device size, and time. To compare large systems with their smaller-scale counterparts some assumptions have to be made. Like, for example, potential energy density, and material density remaining constant.

Having set out these assumptions, a look can be taken at how downsizing affects a turbine. The net amount of energy density transferred to the shaft of a turbomachine under steady-state conditions can be written as a form of the energy (kinetic + potential) equation. But the potential energy density does not depend on the size of the turbomachine, which leads to the important result that the length and time (period) are proportional.

For example, if a gas turbine is made smaller it will spin faster assuming the energy density of the driving gas is the same. While this analysis is largely simplified, useful insights regarding size and power density can be derived.

The dynamic behaviour of turbomachines can be described by mathematical relations between the rotor diameter (D) and the rotating speed (N). Keeping the surface speed constant and using the proportional relationship between length and time, DN will be constant. This indicates that the power density of turbomachinery increases as its size goes down.

Further, it is known that as the maximum linear speed of a rotor blade tip is determined by the strength of the rotor material, keeping constant surface speeds for turbomachinery at different size scales is a good idea.

The increased power density of small mechanical systems makes massive parallel mechanical systems possible. These systems have a number of size and reliability advantages.

Downscaled Drawbacks

Traditional manufacturing processes can achieve relative accuracies of the order of 10-4 to 10-6. Relative accuracy is defined as the manufacturing process tolerance Dl divided by the characteristic part dimension l. This decreases as part dimensions shrink. Modern micro-fabrication methods like reactive ion etching (RIE) achieve a Dl/l of only 10-2 to 10-4 and manipulation of individual atoms or molecules with AFM probes may achieve values for Dl/l of the order of 0.5 x 10-2 at best.

Fuel Cell Scaling Implications

A fuel cell produces electrical current directly from chemical reactions. A basic assembly consists of an ion-conducting electrolyte membrane between two electrodes, backed by fuel and oxidant flow distributors. Ideally, the fuel cell would provide as much current as the external load requires, and the voltage would remain constant. A high-performance fuel cell is one that offers high peak-power and high current density with minimum loss of voltage. Losses arise mainly from activation limitations at low current density, ohmic losses at medium current density, and reactant transport limitations at high current density.

The point of interest for fuel cell miniaturisation is its effect on the overall power density of the device. There are two approaches to examining the scaling implications.

One immediate question is whether miniaturisation can geometrically benefit power density in terms of watts per unit area? Here, the main focus is on the surface-to-volume ratio for very small fuel cells. The second consideration is the extent to which smaller dimensions may offer benefits in terms of current-voltage performance.

Most fuel cells have flat, microporous electrodes separated by an electrolyte layer, a design well suited for continuous fabrication. The principle of non-planar interface layers can be extended to provide an improvement in surface-to-volume ratio Adding a height component may increase the overall volume of the device compared with a strictly planar design. However, increasing the height increases the volume by only one finite amount, whereas narrowing the feature width increases the surface area with no further impact on volume.

The extent to which area can be enhanced is restricted only by the minimum feature width required for functionality and manufacturability. Thus, there will be limiting design rules and critical dimensions, as exist for integrated circuit fabrication. Typical dimensions already common to polymer electrolyte fuel cells offer the opportunity for improvement. Critical dimensions are expected to decrease as miniature fuel cell technologies develop.

Flow resistance is an important miniaturisation parameter. In flow channels, pressure loss is inversely proportional to the ‘hydraulic diameter’, which varies with channel size. Gas diffusion is also important. In fuel cells the transport of reactants to the electrolyte interface is driven by diffusion of gas molecules through porous electrodes.

An advantage of micro-machined flow structures is that obstructed regions at shoulder locations can be reduced. This enhances the overall gas diffusion performance across the porous electrodes.

A further property of a fuel cell is electrical resistance. Low resistance is desirable for high efficiency in energy conversion. The electrical resistance R is proportional to the travel length L and inversely proportional to the cross-sectional area A.

A benefit of miniature design is that circuit interconnections between series cells are generally shorter than their large-scale counterparts. However, a size reduction uniform in all dimensions generally results in higher electrical resistance.

Fuel cells have two common configurations for series connection, monopolar and bipolar. The bipolar construction is almost universally adopted for larger-scale applications because of its simple construction. However, the monopolar layout is more compact, because one fuel chamber services two anodes, and one oxidant chamber services two cathodes. The monopolar design has a disadvantage for large-scale systems as the electrical current must flow laterally across electrodes; but in a miniature system these distances are much shorter. Hence, a smaller system offers the option of more compact design through use of the monopolar stacking arrangement.

Fuel Cell Fabrication

Deep silicon etching permits high geometric complexity at near-zero marginal cost, a stark contrast to conventional manufacturing processes like machining. Fundamental requirements for the electrode material include high surface area for catalyst support, high electrical conductivity, and uniform gas diffusion. The concept of a meso-scopic, 3D interface, however, adds the requirement of texture definition.

Several candidate approaches have been investigated to achieve this level of fabrication control. Initial studies included plasma spraying, co-electroplating with a sacrificial material, and malleable paste formulations.

Despite considerable progress, much remains to be done in the empirical area. Here, advances in manufacturing technologies will play a critical role in the construction of prototypes for functional testing.