Laser boosts liquid computer

By Eric Smalley, Technology Research News

The most advanced experimental quantum computers use the same technology medical magnetic resonance images use to make images of our insides.

MRIs make images of soft tissue by aligning the body's hydrogen atoms with a continuous magnetic field, and then using pulses from other magnets to knock the hydrogen atoms out of alignment in the areas being imaged. After each pulse, the atoms realign with the continuous magnetic field and in the process they emit radio waves whose frequencies are specific to different types of tissue.

Nuclear magnetic resonance (NMR) quantum computers make qubits, which represent the ones and zeros of computing, by flipping the orientations of atoms within the molecules of a liquid. However, it's extraordinarily difficult to read the spin flips of more than a half a dozen of these qubits, or quantum bits, at a time.

Although a growing number of researchers are calling NMR quantum computing a dead end, some who are pressing ahead with the technology have moved it a small step forward.

The seemingly insurmountable problem with NMR quantum computers is that as they gain more qubits the signals from the qubits get weaker and are eventually drowned out by the random noise of the system. The largest NMR quantum computer built to date consists of seven qubits. They're not likely to get much bigger using current designs, and it will probably take thousands of qubits to make a practical quantum computer.

A team of researchers at Stanford University and IBM Research, however, has found a way to strengthen the signals from a two-qubit NMR quantum computer.

Nuclear magnetic resonance devices manipulate atoms magnetically in order to detect them. When some types of atoms are placed in a strong external magnetic field, their nuclei emit radio signals. The nucleus of an atom behaves like a tiny magnet, a property called spin. In order for nuclear magnetic resonance to work, enough atoms have to be polarized, or have their spins aligned either with or opposite an external magnetic field, that the collective signal cuts through the noise produced by randomly oriented atoms.

The signal decreases as the number of qubits increases in NMR quantum computers, said Anne S. Verhulst, a researcher at Stanford University. But because the NMR signals are "proportional to the polarization of the nuclear spins, our technique increases the signals. Hence we can allow more qubits before the signals disappear in the noise," she said.

To do this, the researchers mixed rubidium vapor with xenon gas and aimed a laser into the mixture. The laser light was circularly polarized, so that its electric field rotated. The laser polarized, or aligned the electrons of the rubidium atoms, which in turn polarized the xenon. The researchers then separated out the xenon, froze it into a liquid and mixed it with liquid chloroform. The xenon polarized some of the carbon and hydrogen atoms in the chloroform, and those atoms served as the stronger qubits in the researchers' quantum computer.

Boosting the polarization this way increased the strength of the spin signals by a factor of 10, said Verhulst. In order to use the scheme on a practical level, the researchers will have to increase the strength of the signals by 1,000 times and show that it works on computers made of more than two qubits, she said.

Despite the success of the initial results, the odds are still heavily against NMR yielding practical quantum computers. "Even though NMR quantum computers are the only ones existing so far, the problems related to the scaling issue are really huge. If one wants to make it work with liquids, then either some very special molecules or NMR technique or some additional source of high polarization has to be found, or a combination of all these things. And all of those are not straightforward to discover," she said.

The research doesn't have to lead to practical quantum computers to be useful, said John M. Myers, a project scientist at Harvard University, who helped build a 5-qubit NMR quantum computer in 1999.

"What is worthwhile about this research is the advance of molecular control," said Myers. "The techniques of NMR quantum computing can help to determine structures of large molecules, such as proteins. That is something special about NMR, in contrast to other quantum information processing schemes," he said.

It will take at least 15 to 20 years before practical quantum computers can be built, said Verhulst. "A lot of the efforts involve… nanofabrication, building extremely sensitive probes, trying to manipulate single electrons. And all of those are interesting for technological evolution as a whole," she said.

Verhulst's research colleagues were Oskar Liivak and Mark H. Sherwood of IBM Research, Hans-Martin Vieth of the Free University of Berlin, and Isaac L. Chuang now at the Massachusetts Institute of Technology. They published the research in the October 8, 2001 issue of the journal Applied Physics Letters. The research was funded by the Defense Advanced Research Projects Agency (DARPA).

Timeline:   15-20 years
Funding:   Government
TRN Categories:   Quantum Computing
Story Type:   News
Related Elements:  Technical paper, "Non-Thermal Nuclear Magnetic Resonance Quantum Computing Using Hyperpolarized Xenon," Applied Physics Letters, October 8, 2001


October 24, 2001

Page One

DNA could crack code

Transistor sports molecule-thin layer

Molecule connects contacts

PC immortalizes ancient temple

Laser boosts liquid computer


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.