WILL quantum computers ever be as easy to build as desktop PCs? This week, Japanese researchers are more optimistic than ever, after revealing a chip that can encode two quantum bits (qubits) of information. While other promising quantum computers have involved complex liquid or 鈥渋on trap鈥� systems, these qubits are encoded in components that can be built into a silicon chip using standard techniques.
Other teams have been working on chips that contain qubits (New Scientist, 30 November 2002, p 21) but, crucially, this is the first time that anyone has managed to entangle a pair of them. Entanglement is key to unlocking quantum computing power because qubits linked in this way share information, allowing lots of calculations to be carried out at once. The new chip is 鈥渁 pretty big step forward鈥� says Martin Plenio, an expert in quantum computing at Imperial College, London.
In all computers, information is encoded as strings of 1s and 0s. In classical computers a bit stores either a 1 or a 0. Quantum computers are potentially far more powerful because a qubit can represent both at once. This means that an entangled pair of qubits can store four combinations (00, 01, 10 and 11) simultaneously. A two-qubit quantum computer will outperform a two-bit classical computer by carrying out four calculations at the same time.
Advertisement
Shen Tsai at the NEC Fundamental Research Laboratories in Ibaraki, Japan, and his colleagues have built a chip containing two squares of aluminium, called 鈥淐ooper pair boxes鈥�, each about 0.1 micrometres across. Each box contains a few pairs of electrons, bound together in a state called a Cooper pair, and is connected to a reservoir of other electrons. When an electric field is applied across the box, an extra Cooper pair is pulled out of the reservoir and into the box. This changes the qubit鈥檚 state from 0 to 1. Removing a pair changes the state back. The quantum nature of the boxes allows them to exist in both states at once.
A capacitor between the two boxes stops any electrons from hopping from one to the other, but they are close enough to feel each other鈥檚 influence. This allows them to become entangled. Tsai and his team were able to prove that this actually happened by measuring the oscillating charge within each box that entanglement produces (Nature, vol 421, p 823).
The group鈥檚 next step will be to build logic gates into the circuits to allow the qubits to be used for computation. They hope to have a simple gate working this year. Then, says Dimitri Averin at the State University of New York at Stony Brook, who is a theorist on the project, they will be able to solve simple problems.
But two qubits will not be enough to do serious number-crunching. Only if the design can be scaled up to tens of qubits will it be really useful. Averin thinks half a dozen qubits are within easy reach, putting the NEC team within sight of the record of seven qubits set in 2001 by a 鈥渓iquid NMR鈥� quantum computer (New Scientist, 8 June 2002, p 24). The liquid NMR machine is the most sophisticated quantum computer built so far, but it is unwieldy and may never work with more than about 15 qubits. Tsai says his chip-based quantum computer has the potential to be much bigger.
He still has a long way to go before creating the ultimate PC, however. Cooper pairs are only created when aluminium superconducts, so the chip has to be cooled to almost absolute zero.