The session's first paper was a case in point. Presented by Stefan Jung, it described how researchers at Infineon Technologies Laboratory for Emerging Technologies (Munich, Germany) are embedding electronic circuitry ah, seamlessly, into conventional apparel fabric. The technique replaces some warp threads with coated copper wires. The coating is pierced with a laser at points where encapsulated ICs are embedded in the weave. Applications from communications to entertainment to biotelemetry could then be literally put on like a piece of clothing. Thermoelectric piles composed of large arrays of thermocouples would use the temperature difference between the body surface and the ambient air to power the system.

The theme of ambient temperature as a power source was picked up by the following paper, delivered by Takakuni Douseki from NTT Microsystems Integration Labs (Kanagawa, Japan). By using a Bi-Te compound semiconductor material to bridge between two thermal masses, Douseki showed it is possible to generate a voltage on the order of 1V from a modest 5°C temperature difference. Since in this case the direction of the temperature gradient — and hence the polarity of the voltage — are not predictable, the design team created a dual-polarity switched-capacitor dc-to-dc converter to work with the thermoelectric device.

A more aggressive approach was presented by Martin Schmidt of the Microsystems Technology Labs at MIT (Cambridge, Mass). The Cambridge laboratory has been working with high-temperature MEMS to generate electric power from the oxidation of fuels. Schmidt described two different projects along these lines. The first was a MEMS catalytic reactor that oxidized, and eventually ignited, a gaseous-phase hydrocarbon fuel. The heat from this reaction was applied to one side of a bank of thermoelectric devices, generating power. The process also generated considerable waste heat, a discharge of burning gas and abysmal efficiency, Schmidt said. But that still left it an attractive alternative to awaiting battery improvements.

Tiny turbine

The second approach, which is not actually working yet, was considerably more ambitious. Schmidt described a MEMS gas turbine, the entire assembly of which is small enough to rest on a postage stamp — though in operation it would quickly burn its way through. As is often the case in MEMS, the operating parameters are rather mind-bending. A fuel-air mixture is compressed by turbine blades and driven into an ignition chamber, which in turn drives the high-pressure turbine. The ignited gasses reach a temperature above the melting point of silicon, but the turbine structure is protected by its own dynamics. The rotational speed of the tiny air-bearing turbine is a staggering one million revolutions per minute — a fact which is presenting some design problems. If a slightly out-of-balance turbine brushes the side of its case, it disintegrates.

The turbine has been successfully spun at this speed in a cold experiment, but the device has not yet been tested with fuel and ignition, Schmidt said. It is envisioned that the turbine could be used either for propulsion or to drive an electrostatic generator.

Another adventure in MEMS fabrication was reported for an entirely different purpose. Koenraad Van Schuylenbergh of the Palo Alto Research Center (Palo Alto, Calif.) described the development of self-assembling coils fabricated on the surface of conventional ICs. The loops of the coil are formed by sputter-deposited molybdenum-chromium film that is pre-stressed and pre-formed. When the film fingers are released from the substrate, they magically curl upward, meeting well above the surface of the IC, to form a helical coil. The coil rests on one side on the surface of the chip, with its axis parallel to the surface. Such coils can be used as solenoids or as high-Q inductors, and have proved robust enough to survive injection-mould packaging.

Late in the session, Mitsuo Usami, a researcher at Hitachi Ltd. (Tokyo), described yet another entirely different idea: LSI powder. This catchy name was coined by Hitachi to describe an elegantly conceived self-contained single-chip RFID device that occupies only 0.3 x 0.3 mm of silicon. The chips are designed to be encapsulated within their antenna so as to be virtually invisible. The antenna can then be inserted into virtually any product to provide identification at ranges up to 300 mm.