# Quantum leap

By Charles Seife CRYPTOGRAPHERS can hardly wait for quantum computers to become a reality, as a suitably large system will be able to crack all popular encryption schemes. But one major hurdle is the fact that it is impossible to make perfect copies of information stored in quantum computers. Now two research teams have taken major steps towards overcoming this problem: one by working out how well a computer can copy quantum bits; the other by creating a quantum computer that corrects mistakes made in the copying process. Ordinary computers store information in two-state bits, namely zero or one. But quantum computers use quantum bits – qubits – that can be a combination of zeros and ones. The bizarre nature of the quantum world enables these computers to do things that an ordinary computer finds impossible, like answering 100 “true or false” questions in one go. But one thing that you can’t do with qubits is make perfect copies, because measuring them destroys the information they represent. “We learnt, some time ago, that you cannot clone a copy of an unknown quantum state and keep the original,” says Artur Ekert, a physicist at the University of Oxford. This might seem to wreck the idea of making a quantum computer, but imperfect copies might prove adequate. In a forthcoming edition of Physical Review Letters, Ekert and his colleagues calculate just how imperfect cloned qubits are. “If a cloner takes, say, one qubit and makes two, the fidelity is at most five-sixths,” says Ekert. This means there’s a one-in-six chance that the copier introduces an error – which might be acceptable. In the current issue of the same journal (vol 81, p 2152), Raymond Laflamme, a physicist at Los Alamos National Laboratory in New Mexico, describes a primitive quantum computer that preserves information even when qubits are error prone. In their computer were three atomic qubits that all stored the same information. Even if one of the three was changed, a majority vote of the three qubits restored the original information – without measurements that would destroy the system. More on these topics: