SCHAUMBURG, Ill. — A group of strange bedfellows-electronic engineers, biochemists and biologists from across the country-stands poised to spark a revolution in health care and to create a new industry in the process. Motorola Inc.'s Bio-Chip Systems Unit merges disciplines to serve medical diagnostics, building biochips and biochip systems microfabricated with traditional semiconductor manufacturing know-how.

"We're shrinking the entire chemistry lab onto a chip," said unit chief Nicolas Naclerio, a former assistant director of the Electronic Technology Office at the Defense Advanced Research Projects Agency.

The information the chips yield is expected to have far-reaching implications for health care in four or five years. It potentially opens the door to a whole new discipline-molecular diagnostics, or attacking disease at the molecular level.

The unit is building a pilot line in a Motorola facility in Phoenix that will produce bio-chip prototypes for Motorola's partners in 12 to 18 months. The market will take a bit longer to develop, Naclerio said.

The next leg of the business, he said, is to develop portable instruments that a doctor would drop the chips into and then read out results. "Long-term, we see such small port-able electronic instruments sitting next to the stethoscope and flashlights in the doctor's office," he added.

To pursue this work, Motorola is pairing its engineers with life scientists the company has recruited nationwide. The company has also formed joint ventures with or invested in several pioneers in molecular diagnostics, including Packard Instrument Co. (Meriden, Conn.); Orchid Biocomputer (Princeton, N.J.), which designs microfluidic platform technologies; and Genometrix (The Woodlands, Texas), a supplier of high-microarray technology for applied genomic, diagnostic and pharmaceutical development.

Motorola has also licensed technology from Argonne National Laboratories and from the Engelhardt Institute in Moscow, both part of the Human Genome Project.

Two chip types
Motorola will focus on bioarrays and microfluidic chips. In bioarray ICs numerous tiny probes store molecules such as DNA. They can carry out several different tests at once to identify genetic markers that correspond to a certain response to a new drug, for example. Naclerio likens them to memory chips because they store large numbers of molecular probes the way DRAMs store data.

Microfluidic chips contain valves and pipes that store fluid to synthesize a compound or do a simple preparation, a kind of "lab-on-a-chip," Naclerio said. He compares microfluidic chips to microprocessors because they are programmable and carry out multiple-step chemical processes the way MPUs perform multiple mathematical operations.

Motorola borrowed from its various technologies to make both types of chips. The idea is to abstract information from the finished chips in the same way that large laboratory instruments do the analysis today.

Naclerio sees synergy between life sciences and electronics. "It is tre-mendously exciting [work]," he said. "The more I learn about biology, the more synergies I see with electronics."

For instance, he said, Motorola can leverage processing for high-density pc boards, flat-panel displays, thick-film circuitry and microelectromechanical sensors into biochips. The end pilot line, he believes, will probably combine equipment and processes.

Results are already being seen in pattern-recognition software developed to diagnose yield problems in wafer fabs. The same types of analysis tools can be applied to understanding the body and how genes are expressed.

"Scientists look at gene expression where they look at the cell and see what parts of DNA are translated into proteins in the body," said Naclerio. "They monitor the gene expression level and then when they give a drug to a patient, they identify whether the gene expression level is moved in tandem or out of sync with certain expected patterns."

This exam is not unlike the statistical process used to understand wafer- fab yield problems. After all, Naclerio said, DNA is really a huge information repository. Molecules in the cell-chromosomes-store billions of characters. Machin-ery in the cell copies and translates that information. "Trying to understand all this is very information-technology-intense," said Naclerio.

Along with similarities, silicon chips and biochips have differences-mostly in feature sizes and the types of materials and process conditions that can be employed. Biochips aren't nearly as fine as semiconductors-features are tens and hundreds of microns in size, a far cry from 0.25- and 0.18-micron line widths. But such fine geometries would make a biochip too small for cells. Also, in bioarrays and microfluidics, reaction volumes would be too small.

The materials in biochips are biological, and that limits the types of processes that can be used. For instance, for silicon semiconductor processing, a "low" temperature would be 400°C. In biochips, low means below 100°C. Some tests rely on enzymes, which are sensitive to heat as well as the usual material interactions.

The unit is using some existing manufacturing gear while developing unique equipment of its own. The collaboration with partner Packard Instrument has resulted in a project to build a liquid-dispensing machine, much like a pick-and-place machine, to deposit small droplets of chemicals onto a chip.

"Biochips will need the diagnostic content, which we will get from our partners and pharmaceutical companies," he said. "We are providing the basic distribution platform-the biochip."