When Harvard built the Mark I computer in 1944 to make calculations for the Air Force, it took minutes to do tasks that now take microseconds. Now, Harvard professors are working on conductors so small that electrons sometimes get stuck and bounce back.

Their research focuses on the use of nanowires and nanotubes to construct tiny electronic circuits. These wires and tubes, which can only be seen with the most powerful of microscopes, are so small that just moving them around is a challenge.

"It is a very important step towards building greater devices," says Phaedon Avouris, the manager of Nanometer Scale Science and Technology at IBM's Watson Research Center.

In a discovery that could radically change the way computers are made, Hyman Professor of Chemistry Charles M. Lieber says he has constructed working logic circuits, interconnecting transistors, with nanowires only ten atoms across.

The computer on your desk, by contrast, has wiring in the chips about 400 atoms wide.

Although members of Lieber's research group refuse to discuss the details of how they constructed these tiny circuits, such a discovery could lead to the development of ultra-small computers that consume very little power.

In a pair of papers published in Science magazine earlier this year, Lieber and his group showed how exceedingly small structures can be fabricated and arranged into circuits . This work then led to the group's most recent discovery.

"I am probably more excited than since I first came to Harvard when I was naïve and young," says Lieber.

"We don't just have the material or just how to put them together. We have constructed simple logical devices," he says. "There are huge challenges to meet but that is part of the excitement."

According to Lieber, the problem with current microchip fabrication methods is that the smaller the scale becomes, the higher the cost. Liebers says a "different approach" is necessary to make even smaller chips possible.

The solution: nanowires.

"These are wires as thick as your DNA, ten atoms across, that can be made very long," says Lieber. "They are as small as have been made."

BF: Moore's law lives on

Moore's law, developed by Gordon Moore, one of the founders of the computer processor manufacturer Intel, states that the number of calculations a microchip can do doubles approximately every eighteen months.

The law, which has held true since 1965, is dependent upon continual improvements in manufacturing methods.

Current top-down methods of chip production use electron beam lithography, an expensive process that involves etching silicon into smaller and smaller pieces. Such methods will eventually hit a limit at which further miniaturization would entail extreme expense.

"The current microelectronic industry will hit a wall in the near future," says Assistant Professor of Chemistry Hongkun Park.

Instead of taking large structures and cutting them into smaller pieces, however, some scientists, however, are taking the opposite approach. Using a bottom-up approach, they are focusing on taking extremely small units and combining them into working machines

In November, Science ran a special issue on the present reality and the possibilities of nanotechnology, the building of complex structures atom by atom. Possible applications include quantum cryptography, exotic new materials and, of course, miniature computers.

According to Lieber, using the bottom-up approach to make logical devices involves steps where the parts of a microchip are first fabricated and then arranged to form complex logical devices.

BF: Vacuum Tubes to Nanotubes

Making the basic units of a microchip in nanoscale is challenging in itself. The Lieber group has used nanowires made of silicon with added chemical impurities that allow the material to conduct some electricity.

Another top candidate for the building blocks of extremely small computers are carbon tubes two hundred nanometers across, called nanotubes.

Park has done work defining the electrical properties of nanotubes. His recent research has focused on how the interactions of electrons moving through nanotubes can be predicted and explained with quantum mechanics.

These tubes could potentially be used as the primary conductive element in computers or other electronic devices.

But there are problems.

Although there are various ways to make them, nanotubes can be randomly generated in carbon soot. But no one has yet announced a way to make nanotubes to pre-set specifications.

"The problem with nanotubes is you can't control chirality", explain Yi Cui, graduate student researcher in the Lieber group.

Chirality, or the exact geometric arrangement of the atoms in the nanotube, effects how electrons move through the tube. Electrons moving along the tubes are sensitive to the chirality. Since chirality cannot be controlled, the electrons act in unpredictable ways, for example bouncing back the way they came.

Nanowires do not have such limitations. Approximately three nanometers across, nanowires can be synthesized by a technique called laser assisted catalytic growth.

These wires conduct electricity like normal wires but they can also be arranged so they act like mini-transistors, which are the smallest working unit of a computer.

The challenge is arranging these wires in a useful way. Manipulating a wire ten atoms across is itself a technical feat.

BF: When Tweezers Won't Cut It

The Lieber group has been able to control the arrangement of the wires on a surface using what the researchers call a "micro-fluidic channel."

Moving liquid in the groove of the channel aligns the ultra-small nanowires in the same direction.

"Take a river or fast moving stream; if you put logs in the river they line up," says Lieber.

Aligning nanowires is the first step in making electronic circuits that could be used in ultra-small computers.

In their February paper in Science, Lieber's group used nanowires to make basic electronic components. Now, he says, he has linked such units together-a discovery that provides a crucial first step towards building complex logical circuits, like microchips.

"It seems to me that we really have a way to take something from basic science and change how technologies as we know them are in society today, in a positive way," says Lieber.

"It's a very exciting time," he says. "I wouldn't be doing it if it were not exciting and fun," he says.