A quantum leap in computing

A LABORATORY in Oxford houses one of the rarest machines on Earth: a quantum computer. It is just a prototype, but then there are perhaps only four others in the world - even the best can only, at present, count up to 15.

Quantum computers will be able to solve in minutes problems that would take the current generation of PCs thousands of years: they will model new drugs, sort enormous lists of data and create super-precise clocks that will humble the present atomic clocks.

Next century, says Andrew Steane, "every university and many companies doing research and development will own a quantum computer. That day might possibly arrive in my lifetime". Dr Steane, of the department of atomic and laser physics at Oxford University's Clarendon Laboratory, has submitted a paper to the science journal Nature that could make building the machines easier.

The problem is that their development requires techniques never before attained. A group led by Professor Rainer Blatt in Innsbruck caused a buzz among scientists recently by not only detecting single atoms (which can be done by a variety of methods), but also measuring the "quantum state" - the quantum machine's equivalent of the traditional binary computer code - for each atom.

At Oxford, Dr Paul Barton, who works in Dr Steane's group, carries out the laborious task of setting up and testing their quantum machine. That involves trapping four individual calcium atoms in a vacuum chamber, cooling them with lasers - "It's like slowing a bus down by firing millions of ping-pong balls at it; it's hard, but it works," says Dr Barton - then measuring various characteristics of the atoms. Having four atoms means the computer can count to 15. Preparing it to carry out a calculation can take months, although the calculation itself takes a second.

"We are still at a stage comparable to valves in old computers," says Dr Barton. "If somebody could build the equivalent of a transistor for a quantum computer, the picture could change dramatically."

The key to the quantum computer is the behaviour of matter in the subatomic world, where objects have a dual nature. Just as light can behave like both a particle and a wave (provided you do not try to see which of the two it "is"), so single atoms can exist in "superposed" states, where they may "spin" in two directions at once, as long as you do not try to measure that spin. In a quantum computer, the direction of that spin can represent the 1 and 0 "bits" of a normal computer; they are known therefore as quantum bits, or "qubits".

Standard PCs do no more than manipulate strings of 1s and 0s, which are then interpreted for us into recognisable outputs - a spreadsheet, letter or Lara Croft. Similarly, running a quantum computer would require a means of deciding how to pose the question in terms of qubits.

Once that has been done, the qubit particles of the quantum computer will calculate all the branches of the problem simultaneously, just as a wave can pass through many openings at once.

It is the same baffling phenomenon as in school experiments to see whether light is a wave or a particle: the result depends on which answer the design of the experiment favours.

The normal computer must work step by step through each collection of bits. But in the quantum world, it all happens at once.