The key idea is to form an array of tunable quantum dots, each capable of trapping a single electron. The electron spin offers the most robust way to encode a qubit, because a strongly confined spin-1/2 particle does not couple strongly to electric fields. Loss and DiVincenzo (Phys. Rev. A 57, 120, 1998) proposed methods to perform both single-qubit and two-qubit gate operations in such a dot array; single-spin rotations are realized by electron spin resonance, whereas two-qubit logic relies on gate control of the interdot exchange interaction. This two-qubit logic can be extremely fast, on the sub-nanosecond timescale, as demonstrated by Petta et al. (Science 309, 2180, 2005). The inability to perform single-spin rotations on a comparably fast timescale has so far limited the progress of this approach.

One solution is to utilize a semiconductor material in which conduction-band electrons experience a relatively large spin-orbit interaction (SOI). The SOI provides a pathway to coherently couple applied electric fields to the spin. It causes an electron spin in the presence of an ac electric field to "see" an effective ac magnetic field that can be much stronger than externally applied ac magnetic fields achievable in practice.

Researchers are trying to implement these ideas experimentally by building quantum dot devices using InAs nanowires. InAs has a SOI that is 1–2 orders of magnitude larger than in GaAs, the standard workhorse material for tunable quantum dots. InAs nanowires also possess a number of electronic properties that make them well suited to realizing tunable quantum dots.

A prototype 5-gate, double-dot device was fabricated using electron-beam lithography techniques applied to a nanowire of length 1.5 µm and diameter ~60 nm. The ability of the ~40 nm wide gate electrodes to locally deplete carriers at low temperatures has been demonstrated, and further characterization is in progress.

In the report, the authors discuss strategies for efficient readout of spin states, and estimate spin coherence times in the presence of nuclear hyperfine coupling (abundant 115In and 75As spins) to be in the order of 3 µs, long enough to allow 102–103 gate operations. The prospects are good for using such a device as a quantum computing testbed in the near future.