Portland, Ore. -- For a soldier in the field, a slight hand tremor, tic of the eye, sudden sore throat or whiff of a noxious odor could urgently put his battery-powered portable lab to work. Such microfluidic labs-on-a-chip could let soldiers test their own blood for exposure to many toxins simultaneously and in a matter of minutes. At least, that was the goal of a professor at the Massachusetts Institute of Technology when he landed a U.S. Army contract to pursue work on the tiny tool.

Although commercially available "gene chips" can test blood for thousands of toxins simultaneously, it takes hours to diffuse a blood sample across the whole array and then scan for fluorescence. Laboratory high-voltage power supplies could speed the operation with the pumping action of an electric field, but a battery-powered device could not supply that much juice. Enter an MIT mathematician with friends in the engineering school and funding from the Army's Institute for Soldier Nanotechnologies.

"The Army decided to take a chance on a mathematician who wanted to do some experiments," said MIT's Martin Bazant. Those experiments, performed in the lab of professor Todd Thorsen, showed that "the micropump can generate enormous electric fields--a volt per 5 microns in our current device, but potentially as high as a volt per 100 nanometers--while using a very low supply voltage, about 1 volt, and very low current too--on the order of milliwatts," Bazant said.

Bazant has founded ICEO Technologies Inc. with his postdoctoral associate, Jeremy Levitan. ICEO will commercialize the micropump technology.

Today, commercial test arrays from companies like Affymetrix Inc. can take hours to a day to make a detection. "Our micropump can greatly accelerate the speed with which you get results," Bazant said. "Without the velocity we create to quickly pass samples over all the detectors, you would have to wait for diffusion to find each target molecule, and there might be a thousand different targets in the bottom of a channel."

Speeding up this process with electric fields requires hundreds of volts--too much for a soldier's battery-powered kit. Bazant dropped the requirement to a single volt using a combination of photolithographic electrodes, mechanical engineering of the steps on the electrodes' surface and a distributed model of pumping liquids. Electrophoresis and electro-osmosis form the basis--using electric fields to move liquids down lab-on-a-chip channels that are 10 microns wide and deep. Thousands of "detector" molecules are affixed to the channel's bottom beforehand. Blood cells are about 8 µm in diameter, making them a tight fit, but the 10-µm channels keep them all in single file. After pumping blood cells into all the channels, any spots that test positive will glow fluorescent, identifying the toxin by its location.

"By integrating an active micropump into the chip, it can pass samples over the chip's surface much more quickly, and that not only speeds up detection, but also makes it more sensitive," said Bazant.

To test blood, Bazant dilutes it about 10 times before turning on the micropumps patterned above each channel. "Our technique is especially good for mixing samples with a reagent, because we can create very nice velocities. But there is a bit of a trade-off, because we cannot create the strong pressure gradients you need for very thick liquids or to unclog a channel," he said.

In the lab, electric fields speed mixing with a high-voltage power supply in which the electrodes are located at either end of a millimeter-long channel. MIT has enabled lower voltages by putting the electrodes closer together, a space of 5 µm. With photolithography, MIT patterned a continuous string of electrodes the length of each 10-µm-wide channel on its chip, using platinum and gold interdigitated stepped microelectrodes and a glass substrate pretreated with oxide. "We use photolithography to build up on a glass substrate, with a bonding layer such as titanium oxide, on top of which we lay down a gold layer," Bazant said. "Then we build the steps by electrodeposition using a kind of mold we make with photolithography."

The key to Bazant's mathematical design was replacing dc with alternating current. If dc was used, each volt step, every 5 µm, would result in a 200-V difference between each end of a millimeter-long channel. With ac that requirement drops to just 1-V direct current, by virtue of the interdigitated microelectrodes-- like two combs facing each other with interlaced teeth. The power supply alternates the voltages to each comb to create the ac electrical field down the length of a channel.