Want to take in all the sights of London without wearing out your shoes? Or make all those sales visits in the shortest possible distance? Let some glowing helium gas do the walking. Andreas Manz and colleagues at the Imperial College of Science, Technology, and Medicine, London, and a Harvard University team led by George Whitesides are taking a crack at the classic "traveling salesman problem" (TSP) in an entirely new way: using a lab on a chip.
Finding the shortest route around a certain number of stops--whether they are tourist attractions, potential customers, or workstations in a factory--is relatively easy for a few stops. But as more stops are added, the complexity of the calculation increases exponentially until it becomes impossible to solve. Mathematicians and computer scientists have struggled with the problem for decades, but Manz and Whitesides describe a more mechanistic approach in the May edition of the journal Lab on a Chip.
The researchers' first step was to represent the problem graphically by etching a map of London onto a glass chip. They then covered the etched part of the chip with another piece of flat glass to create a network of pipes. They also fixed tiny electrodes to the chip so that they could apply a voltage to various locations. The researchers then filled the pipes with helium. Using the electrodes, they could then apply an electric voltage between two points on the chip. The electric field would then guide an electric discharge along the shortest route between the two points, making the helium glow like a fluorescent tube just along that route. At present, the method can be used to find the way out of a maze and the shortest route between two points, but the team members hope to develop it for more complex problems.
"It is very, very cool," says microengineer David Beebe of the University of Wisconsin, Madison. But whether glass chips can rival a digital computer remains to be seen. Beebe thinks racing a digital computer is pointless, but he adds: "I'll bet there are other applications for the analog computing concept demonstrated that none of us have thought of yet."