Researchers May Be On Verge Of Upsetting Known Computer World

Researchers at Australia's University of New South Wales and at Purdue have achieved what seems like it should be an episode of "Star Trek," but it is real as they have made a transistor from one atom. It's not the first time atomic-level transistors have been created, but those efforts were hit and miss. This new effort is repeatable and one-day may result in computers that will be very small and fast.

Calling this the groundwork for the quantum computer, it seems that we are entering a weird, weird world where normal rules just don't seem to apply. Normally, computers - and there are some very small computers on the relatively near horizon - make decisions based on digital 1s and 0s or on and off. Quantum computers, on the other hand, offer the seeming ability to make many simultaneous decisions. Instead of using normal data bits, the quantum computers envisioned using this new atom-based transistor data called qubits and rely on quantum mechanics and since quantum mechanics are not nearly as predictable, yet, as other forms of computing, qubits could represent multiple data values simultaneously.

These computers would be small and fast enough to undermine even today's fastest computers on which ecommerce and data privacy rely. It is that big a breakthrough as as the discovery of Moore's Law over 50 years ago which predicted that silicon technology would shrink in size as the number of transistors doubled every 18 months. It has worked out and continues to work today. Gerhard Klemich, a professor of engineering and computer technology at Purdue and team leader there, said the single-atom computer proves that "Moore's Law can be scaled toward the atomic" level "in silicon." (As if to prove him right, Intel recently announced it was building nanoscale wires an atom thick on an array of atoms four atoms wide, still based on silicon technology.)

The research teams, using a device called a tunneling scanning microscope looked at a single silicon crystal and scraped out a patch in the center. Then, using phospine gas, they planted on phosphorus atom in the hole they had scraped and they then sealed the one-atom transistor in several layers of silicon atoms for stability.

Michelle Simmons, group leader of the South Wales research team, called the "device perfect. This is the first time anyone has shown control over a single atom in a substrate with this level of accuracy."

The researchers are not done in their work, though, as the next steps involve making this new transistor into an array that can handle work and then activating its switching function, using standard industry techniques. Of course, things are not likely to change overnight because the first experiments in atom-level transistors actually happened a decade ago.

Still, the development the researchers in Australia and at Purdue have made is monumental because no one before them had been able to place the atom exactly where it was needed. It took the development of the tools available now - the scanning tunneling microscope and supercold to make things happen. Indeed, some have overlooked the supercold part of the research but it is needed.

Intel's Mike Mayberry, vice president in charge of the company's component research group, cautions the use of cold is important and stressed the need for caution in moving ahead. "It's good science, but it's complicated," he emphasized. His group, though, has been moving ahead on the miniaturization front a working array that is 22 atoms wide, the smallest and quickest in the industry.

Meantime, IBM's own research team at its Almaden Research Center has used similar techniques today to develop a working 12-atom array from which they can retrieve and store 1s and 0s. In other words, they have achieved a way to make a 12-atom array work as a computer which is a breakthrough in itself.

That pales a bit when you consider last January's breakthrough by the Purdue and South Wales teams. Though not widely known, they developed single-atom silicon nanowires that fed four-atom arrays.

So, the miniaturization race is continuing and observers have noted there will be real breakthroughs coming in drug development and biology once these computers are running.

This whole issue proves the logic of a statement by the famed Dr. Richard Feymann who predicted more than 50 years ago that there was still "plenty of room at the bottom." It looks like the world is finding that out and no one knows where it will lead. Maybe we will end up with button-sized supercomputers handling routine items like talking to multiple satellites to several people using pin-sized cameras for videochats and this wouldn't even make the computer break into a sweat. Will it happen, we'll see in about 10 years.