In recent years, scientists have been getting closer to developing labs-on-a-chip, devices designed to manipulate pico- or even femtoliters of fluid. The chips could improve drug discovery, medical diagnostics, and even genomic testing; experiments that now require pipette-handling robots to manipulate large quantities of chemi­­cals, for example, could be done on a tiny device. But the technology isn’t quite there yet. The chips are small, but moving chemicals around inside them requires external pumps and valves. The state of the art, then, is closer to a lab around a chip than on one.

But now Neil Gershenfeld, the director of MIT’s Center for Bits and Atoms, and graduate student Manu Prakash say they’ve figured out a way to get rid of those pumps and valves. They’ve designed a device, described in a recent issue of Science, that manipu­lates its liquid contents without external controls. And the technique is based entirely on the flow of bubbles.

The researchers discovered the technique by accident. Part of the center’s mission is to help bring computing to remote areas by designing processors that use simple materials and don’t require complicated power systems or electronics. Prakash and ­Gershenfeld were trying to develop a ­microfluidics-­based microprocessor, but “the bubbles kept interfering,” ­Gershenfeld says. “It eventually occurred to us that we should use them.”

They did, and found that the flow of the bubbles–where they go, how quickly they get there, whether they merge or remain separate–can be controlled through the design of the channels in the chip. For example, the chip can synchronize the movement of two bubbles flowing along parallel paths, guaranteeing a meeting when the two avenues converge. If one bubble is far ahead of the other, the surrounding liquid, obeying the laws of fluid dynamics, floods through channels between the parallel paths. The altered flow drags on the leading bubble and accelerates the other one, helping it catch up.

There are many potential applications of this work, Prakash says. The chip could serve as a microprocessor: intersections between the channels would act as logic gates, and the bubbles would stand in for electrons as the basic units of information. For drug discovery purposes, the bubbles would be replaced by drops of chemi­cals and reagents. Instead of being mixed on a lab bench, these materials would react within the chip, making the whole endeavor simpler. Prakash even envisions a bubble-based diagnostic device that could be used in remote communities to test patients’ blood or urine for signs of disease. In this case, the doctor wouldn’t have to send the samples to a lab–the lab would be right there, inside the chip.