Taking the Transistors Out of Computers

Making a better computer chip today means cramming more transistors onto a silicon wafer. But future computers may not even need transistors, which use electricity to store or relay strings of 1's or 0's as data. Toward that goal, researchers describe in tomorrow's Science the first logic gate that does not rely on electrical current, but instead on the quantum nature of electrons.

The trouble with transistors is that they communicate data through electrical wires, which radiate heat that internal fans must wick off to prevent damage to other computer components. "Early versions of the Pentium chip produced as much heat per unit area as an electric range," says Gregory Snider, an electrical engineer at the University of Notre Dame in Indiana. Today's chips work cooler--but a computer without heat-generating wires altogether would run even more efficiently. "Our idea is to make a complete break with transistors," Snider says.

In 1991, Craig Lent and Wolfgang Porod of Notre Dame proposed that transistors could be replaced someday by "quantum-dot cells." Now Snider's team is the first to make four of these cells work as a logic gate. The researchers fashioned micrometer-sized aluminum squares linked by aluminum oxide junctions, through which electrons can "tunnel" from one square to another. Each quantum dot cell stores a "1" if two electrons are in the northeast and southwest corners, and a "0" if they are in the other corners. Such cells, they found, could communicate with each other--at temperatures just above absolute zero--through changes in the electric fields set up by the extra electrons. The communication is a direct order: Three cells surrounding one will dictate whether that cell carries a "1" or a "0." For instance, when two out of three adjoining cells bear a 1, the central cell is forced to match it.

Experts say this working quantum logic gate is impressive, but caution that a wireless computer could be at least 20 years away. "There's no reason this can't work, but it's going to be a long process," says nanophysicist John Tucker of the University of Illinois, Urbana-Champaign. The challenge will be to shrink the size of these quantum dots to the size of individual molecules and to get them to work at room temperature.

Posted in Math