QUANTUM computers are taking their first steps towards becoming a practical technology, according to experts who gathered late last month in Nashville, Tennessee.
The meeting, organised by the US Advanced Research Development Activity (ARDA), was closed to the public. But delegates interviewed by 快猫短视频 say the high point was the first demonstration of how quantum computers based on ion traps could be scaled up to large machines.
Quantum computers promise to solve mathematical problems so tough that today鈥檚 supercomputers could never crack them. By exploiting the peculiarities of the quantum world, which allow a particle to be in more than one state at once, they can perform many calculations in parallel.
Advertisement
To build a quantum computer, you need a system that can store quantum bits, or 鈥渜ubits鈥� of information in its different quantum states. These qubits can be single atoms, photons, or even devices built from superconductors and semiconductors. Teams are trying all of these approaches, but so far most have concentrated on machines that can only process one or two qubits.
For example, Yasunobu Nakamura and colleagues from electronics firm NEC in Japan demonstrated earlier this year that they had 鈥渆ntangled鈥� two qubits stored in a superconductor (快猫短视频, 22 February, p 16). Entanglement links the information in different qubits, speeding up the computer鈥檚 calculations (see 鈥淚n a tangle鈥�). Two other teams claimed at the meeting that they had created entanglement using different superconducting devices. And rumours were flying that Nakamura and his team have now carried out a logic operation using two qubits 鈥� a paper is expected in the next few weeks. Since logic operations are the building blocks of programs, this would be a significant advance.
But ion traps, with a mini calculation already behind them, are far ahead of the pack. At the end of last year, Rainer Blatt at Innsbruck University in Austria ran a very simple program on a quantum computer made from a single ion trapped within electric and magnetic fields (快猫短视频, 30 November 2002, p 21). But more powerful computers will need many more ions. 鈥淲e can do basic things with one or two qubits, but for something of genuine interest we need 10 qubits, or even better, hundreds,鈥� says Andrew Steane, who works with ion traps at the University of Oxford.
For larger machines, the trappers will need to be able to move ions around inside the traps, so they can shuttle information from one part of a quantum computer to another. So when David Wineland from the National Institute of Standards and Technology in Boulder, Colorado, described experiments completed just days before the meeting that suggest this will be possible, delegates were thrilled. 鈥淚t鈥檚 really exciting progress,鈥� says Martin Plenio, a theorist from Imperial College London.
Wineland and his team trapped two entangled beryllium ions in a cage of strong electric and magnetic fields. Each ion holds two qubits, in its electrons and nucleus. 鈥淲e were able to separate the two ions from one trap and put them into separate traps,鈥� Wineland told 快猫短视频. When the team tried this before, the fragile link between the quantum states of the two ions was destroyed. This time, the ions remained entangled even after being dragged from 3 micrometres apart to 300.
鈥淚 don鈥檛 believe in breakthroughs 鈥� they are all incremental steps,鈥� Wineland says. 鈥淏ut the pieces are coming together for us.鈥�
In a tangle
Don鈥檛 worry if you have trouble grasping the concept of entanglement 鈥� the strange phenomenon where the quantum states of different particles are inextricably linked, no matter how far apart they are. Even the experts seem to be struggling.
At the ARDA meeting of eminent quantum physicists last month, a debate raged over what 鈥渆ntanglement鈥� actually is. The argument started when Frederick Wellstood of the University of Maryland and his team announced that they have seen evidence of entanglement between the quantum states of two superconducting devices. Creating entanglement is an essential step in building a working quantum computer.
But other scientists at the meeting argued that Wellstood鈥檚 experiments are not measuring entanglement itself, but a weaker kind of coupling between the two quantum states. 鈥淭hat is not the same thing at all,鈥� says Richard Hughes, a theorist at the Los Alamos National Laboratory in California.
鈥淒ifferent people are using the word entanglement in slightly different ways,鈥� admits Wellstood. Most physicists use entanglement to refer to a particular type of coupling, in which the quantum states of the two bits are as strongly linked as it is possible for them to be. What Wellstood and his team have shown is a weaker kind of connection. Entanglement 鈥渏ust isn鈥檛 an appropriate word for what they are doing鈥�, says Hughes.
鈥淭he subject is confusing, no doubt about that,鈥� says Wellstood. And he reckons the confusion runs deeper than simply the definition of the word. Nobody actually knows what kind of coupling you need for quantum computing, he points out. 鈥淭he feeling is that entanglement is necessary,鈥� agrees Hughes. 鈥淏ut nobody can quite put their finger on why.鈥�