Using an electric field might be simpler. The Penn State researchers coat eight-micrometer-long and 300-nanometer-wide silica-coated nanowires with three different types of DNA. They etch three rows of shallow rectangular wells on a silicon chip; each well is designed to hold a single nanowire. Then they apply an electric field to the first row--the field's strength is designed to vary along the row--and deposit a solution of the first type of DNA-coated nanowire. "There's a very strong force that pulls the wires in the direction of the highest field strength," says Penn State chemistry professor Christine Keating, who led the work with Mayer.

One by one, the nanowires fall into the wells, going from the well exposed to the strongest field to that under the lowest field. The other two rows of wells are filled in the same way. "Assembling single wires has been very tricky," Javey says. "This is probably the best result I've seen in terms of assembling individual wires at discrete locations."

This is just a first step for the Penn State researchers, though. They deposit electrodes on their nanowire arrays, showing that an electrical connection to the wires is possible. But right now, they detect the DNA-coated wires optically using fluorescent-tagged target molecules. Keating says that to make a practical device, they need to assemble silicon nanowires and connect the wires to transistors on the silicon chip.

Mark Reed, a professor of electrical engineering and applied physics at Yale, has come up with a different way to make arrays of silicon-nanowire detectors, using an etching process similar to that employed to make integrated-circuit chips. Reed says that nearly all of the sensors etched on the chips work. The new electric-field technique, meanwhile, has a lower yield. About 70 percent of the sensors in the arrays worked; the other wells remained empty or got filled with multiple nanowires.

The new technique could also face competition from the nanosensing technology developed at Lieber's lab, which startup Vista Therapeutics is commercializing. Nevertheless, University of Pittsburgh chemistry professor Alexander Star says that the new method "is a really elegant and sophisticated way to assemble functional nanostructures."

Javey says that the process would need to work on smaller nanowires. Narrower nanowires are much more sensitive, but they're also more difficult to work with because they are flexible and break easily. "This is beautiful work," he says, "but [if it] can be scaled down to 10-nanometer wires . . . then that would make it even more exciting."