Phys. Rev. X (in the press); preprint at http://arxiv.org/abs/1704.08837

A basic operation in quantum networks is to transform photonic, 'flying' qubits into stationary qubits in a matter system. One way to make the mapping efficient is to use large ensembles of atoms. These absorb photons with high probability and store the quantum information they carry in collective spin excitations. The information, however, ends up encoded in delocalized states, which precludes the sort of local manipulation needed for processing it further. Help may now come from Alexander Glaetzle and colleagues, who have proposed an elegant solution for converting delocalized matter qubits into localized ones.

They argue that the Hamiltonian of atomic arrays can be changed — for example, through suitable laser fields applied to an ensemble of Rydberg atoms — to allow it to act as a 'quantum spin lens'. Delocalized states are then coherently focused, ultimately to single atoms, in a process analogous to how optical lenses work. Glaetzle et al. present several variants of their lenses, including multifocal ones, where a single delocalized spin excitation is transformed into a state with entanglement between excitations at spatially separated focal points.