August 1999, Vol. 16, No. 7 MIT's quantum computer leap Researchers create computer that simulates quantum system Seventeen years ago physicist Richard Feynman speculated that a quantum computer might be better at simulating quantum mechanics than a traditional computer. This June researchers at the Massachusetts Institute of Technology (Cambridge, Mass.; web.mit.edu) have succeeded in programming a prototype quantum computer to do what Feynman proposed. MIT's quantum computer can only count to four, but it has demonstrated quantum simulation for the first time. Research such as this may one day result in commercial hardware capable of computations beyond the reach of any conceivable conventional computer. David G. Cory, associate professor of nuclear engineering at MIT, Raymond Laflamme of Los Alamos National Laboratory and colleagues have come up with a general scheme for quantum simulation that would work on any quantum computer. In a paper in the June 28th issue of Physical Review Letters, the researchers illustrate the scheme on a liquid state nuclear magnetic resonance (NMR) quantum computer developed at MIT. "What we've had this quantum computer demonstrate — the dynamics of harmonic and anharmonic oscillators — is very simple. A first-year quantum mechanics student could do it on paper. But this is probably the first reachable application of information processing on a quantum system," said Ching-Hua Tseng, an MIT postdoctoral associate on the nuclear engineering research team and co-author of the paper. The other authors are Shyamal Somaroo and Timothy F. Havel of Harvard Medical School. Cory's research group, and Neil A. Gershenfeld and colleagues in MIT's Artificial Intelligence Laboratory, with Isaac Chuang at IBM, independently helped develop the quantum computer. The simulation scheme they used borrows heavily from average Hamiltonian theory, an NMR formalism introduced by John S. Waugh, Institute professor emeritus of chemistry at MIT. Unlimited power At the moment, quantum computers don't possess the calculating power of a pocket calculator. But, the potential is there for a computer that would far surpass a regular computer in power and efficiency. Because quantum mechanics allows a quantum computer's components to represent many states simultaneously, it should be able to perform many computations simultaneously, making it an enormously powerful tool. A quantum computer may be able to quickly solve problems involving weather prediction and fluid flow — problems so big they couldn't be stored in a conventional computer's memory. Also, quantum computers could potentially do these calculations in seconds instead of years. In proposing how a quantum computer would work, Feynman suggested that ones and zeroes (representing the "on" or "off" states of the electronics of a conventional computer) could be represented by independent quantum states. According to quantum mechanics, the protons in a molecule act as small bar magnets with only two possible quantum states. The spins align in a magnetic field to be either up or down. This is the equivalent of the conventional computer's "on" and "off" states. After data is stored in this manner, the computer must also incorporate logical gate operations, or commands, to perform calculations. A single bit of information in the quantum realm is known as a qubit. A qubit can represent one, zero or the two states at once. It is possible to create a command that causes a controlled interaction among two or more qubits, producing a coherent change in the state of one qubit that is contingent on the state of another. This is the basis of quantum computation. Cory, Tseng and their colleagues are among a small handful of researchers using NMR to experiment with qubits. NMR techniques are widely used to study the structure of molecules and to image internal structures within the human body. NMR allows scientists to manipulate the atomic spins of nuclei by applying an electromagnetic pulse to molecules diluted in a liquid. Sending a pulse for a specific amount of time generates a known signal. The signal is amplified by the molecules acting in parallel. Controlling the spin of the liquid molecules performs the gate operation. It then arrives at an answer. In this case, the computer simulates a quantum mechanical system by carefully forcing its qubits to evolve like another quantum system. Getting atoms to cooperate The main obstacle to constructing a practical quantum computer is control, the difficulty of engineering the quantum states required, the phenomenon of decoherence (the tendency for these quantum states to lose their coherence properties through interactions with the environment), and the quantum measurements required to read out the result of a quantum computation. In the face of all these challenges, researchers are unsure whether a quantum computer can be efficiently scaled up beyond a prototype. But even primitive quantum computer research helps scientists better understand underlying physical principles of quantum mechanics and, in turn, use what is physically possible to inspire new avenues in computer science. Simulating the behavior of quantum mechanics is a useful research tool because "if you can calculate the behavior of a system you can't normally calculate, you can find out more about its behavior and structure and do research on that," Tseng said. This work is supported by the Defense Advanced Research Projects Agency and the Department of Energy. Web Site © Copyright 1999 by Lionheart Publishing, Inc. All rights reserved. Lionheart Publishing, Inc. 2555 Cumberland Parkway, Suite 299, Atlanta, GA 30339 USA Phone: 770-431-0867 | Fax: 770-432-6969 E-mail: lpi@lionhrtpub.com Web: www.lionheartpub.com Web Design by Premier Web Designs E-mail: lionwebmaster@preweb.com