Quantum computers go digital

By Eric Smalley, Technology Research News

Some of the same properties that would make quantum computers phenomenally powerful are also properties that make it difficult to actually build them.

Problems that would take the fastest possible classical computer longer than the lifetime of the universe to solve would be hours-long exercises for large-scale quantum computers. Such machines would be able to rapidly search huge databases and would render today's encryption methods useless.

The key to quantum computers' potential is that quantum bits, the basic building blocks of quantum computing logic circuits, can represent a mix of 1 and 0 at the same time, allowing a string of qubits to represent every possible answer to a problem at the same time. This means a quantum computer could check every possible answer using a single set of operations. Classical computers, in contrast, check each answer one at a time.

But today's qubits are difficult to work with and prone to errors, and the faster they go the more errors they produce. One of the challenges of building a quantum computer is reducing errors. Researchers from the University of Wisconsin at Madison have eased the problem with a method that reduces error rates by two orders of magnitude.

Today's computers are digital, meaning they use signals that are either on or off to represent two states -- a 1 or a 0 -- and all computations are done using combinations of these binary numbers. One advantage of using just two states is the signals that represent those states don't have to be exact, they simply have to be clearly closer to 1 than 0 or vice versa.

Qubits are analog devices, meaning they produce variable, continuous signals rather than discrete on and off states. For example, a particle can be in one of two orientations, spin up and spin down, but also some mix of the two. The 1s and 0s of digital information are mapped to the spin up and spin down states, but quantum computations have to be precise to ensure that the given particle is actually in one of those two states. "Classical bits have only two states... quantum bits can be in between," said Robert Joynt, a physics professor at the University of Wisconsin at Madison.

A qubit continually rotates between 0 and 1, which makes it prone to errors, said Joynt. "A rotation of a qubit can, for example, fall a little bit short with only a very minor error in the input signal," he said.

The researchers' method makes quantum computing a pseudo-digital operation. "In our set-up, a definite rotation rate for the qubits is associated with a range of input signals. [This way] the input does not have to be exceedingly precise," said Joynt.

Easing the requirements for precision could go a long way toward making quantum computers viable. "The driving force [for the idea] was objections from experienced electrical engineers, particularly at IBM, who believed that quantum computing would not work... the because the specs for the driving electronics would be much too [demanding]," said Joynt.

The researchers are applying the pseudo-digital qubits to their ongoing efforts to build a solid-state quantum computer. Their design calls for thousands of individually-controlled electrons in a silicon chip. The chip would allow for careful control of the interactions between neighboring electrons so that the states of the electrons could be used to carry out computations. Some of the fundamental logic operations in quantum computers are carried out through the interactions of pairs of qubits.

The researchers added the pseudo-digital qubits concept to their design by having pairs of electrons slide past each other rather than crash into each other, said Joynt. When the electrons are well separated the interaction is off, representing a 0, and when they are within close range the interaction is on, representing a 1.

When the researchers simulated the technique, they found that it reduced operational error rates by more than two orders of magnitude, according to Joynt. The researchers' pseudo-digital qubits could be implemented in other types of quantum computers, he added.

The pseudo-digital approach is a good one, said Bruce Kane, a visiting associate research scientist at the University of Maryland. "My guess is that future quantum computers will use the pseudo-digital approach," he said. It remains to be seen whether the devices the researchers are building will work well, however, he said.

Quantum computing naturally has many similarities to analog rather than digital computing, said Kane. Because digital computers operate using just two states -- 1 and 0 -- inputs can always be rounded. This type of rounding, however, is impossible in quantum computing, he said. "It [is usually] necessary to control parameters very precisely to keep the computation on track," he said.

The researchers' method is an attempt to find systems that "pretty much automatically have only two interaction strengths," said Kane. No system can have exactly this behavior, so the method doesn't eliminate the problem of errors creeping into a quantum computation, but it can reduce the severity of the errors, he said.

The researchers have shown how to minimize the adverse effects of turning interactions on and off in quantum computing, said Seth Lloyd, a professor of mechanical engineering at the Massachusetts Institute of Technology. "Although I doubt that this exact architecture will prove to be the one that is used to construct large-scale quantum computers, it is exactly this sort of imaginative quantum-mechanical engineering that is required to solve the problems of large-scale quantum computation," he said.

One of the challenges in implementing the scheme in a real quantum computer is fabricating the tiny qubits precisely, said Joynt. "The real issue is fabrication of quite complicated nanostructures," he said.

The researchers are working on qubits made from two basic pieces -- a semiconductor sandwich structure "which is really a monster club sandwich," said Joynt; and a gate structure, which controls the state of a qubit so that it can represent a one or a zero.

The researchers have made progress on the semiconductor sandwich structure and are gearing up now to produce the gate structure, "which is quite complex," Joynt said.

The researchers are also working on a readout apparatus that will fit on the chip. Reading the quantum states of particles is tricky because quantum states are easily disturbed.

It will take a decade to develop simple demonstration models, and probably 20 years before the devices can be used in practical quantum computers, said Joynt.

Joynt's research colleagues were Mark Friesen and M. A. Eriksson. They published the research in the December 9, 2002 issue of Applied Physics Letters. The research was funded by the National Science Foundation (NSF) and the Army research office (ARO).

Timeline:   10-20 years
Funding:   Government
TRN Categories:  Physics; Quantum Computing and Communications
Story Type:   News
Related Elements:  Technical paper, "Pseudo-Digital Quantum Bits," Applied Physics Letters, December 9, 2002.


January 29/February 5, 2003

Page One

Data stored in live cells

Faster quantum crypto demoed

Bumpy surface stores data

Quantum computers go digital

Tiny hole guides atoms against tide


Research News Roundup
Research Watch blog

View from the High Ground Q&A
How It Works

RSS Feeds:
News  | Blog  | Books 

Ad links:
Buy an ad link


Ad links: Clear History

Buy an ad link

Home     Archive     Resources    Feeds     Offline Publications     Glossary
TRN Finder     Research Dir.    Events Dir.      Researchers     Bookshelf
   Contribute      Under Development     T-shirts etc.     Classifieds
Forum    Comments    Feedback     About TRN

© Copyright Technology Research News, LLC 2000-2006. All rights reserved.