Classical computers store and process information as "bits" that can have one of two logic states ("0" or "1"), but quantum computers work on the principle that a quantum particle (such as an electron or atomic nucleus) can be in two states at the same time – "spin up" or "spin down". These two spin states represent a logical "1") or a "0", so N such particles –or quantum bits (qubits) – could be combined or "entangled" to represent 2N values simultaneously. This would lead to the parallel processing of information on a massive scale not possible with conventional computers.

In practice, it is difficult to make even the simplest quantum computer, however, because these quantum states are fragile and are easily destroyed. They are also difficult to control. For a qubit to work, it should thus be well isolated from its environment to preserve its quantum properties, and prevent "decoherence". At the same time it should be robust enough so that its state can be read out and manipulated. The intrinsic magnetic moment of an atomic core, or nuclear spin, is a good qubit candidate in this respect because it fulfils all of these criteria.

There is a problem, however, in that the magnetic moment of a nuclear spin is 10 billion times smaller than the moment of one bit of a modern hard drive, and it is almost impossible to detect, let alone manipulate, such a tiny signal.

The hyperfine interaction

A team led by Stephen Lyon at Princeton studied a nuclear spin system made up of a dilute ensemble of phosphorus-31 and arsenic-75 nuclear spins in a silicon crystal. This donor nuclear spin system, as it is known, contains both nuclear spins and electron spins, and at low temperatures, each nuclear spin is naturally coupled to an electron bound by the donor charge through an interaction called the hyperfine interaction. This interaction can be thought of as an effective magnetic field produced at the nucleus by the electron.

Lyon and colleagues found that they could use oscillating electric fields to jostle the electron back and forth about the nucleus to modulate the hyperfine interaction. They produced these electric fields by applying an AC voltage to the central conductor of a 1D microwave Bragg grating. This structure picks up the weak magnetic signal from electron spins coupled to nuclei at microwave frequencies, but at the same time it allows a strong electric signal to affect the nuclear spins at radiofrequencies.

"We discovered that if the hyperfine interaction is modulated at the right frequency, it can be used to flip nuclear spins," explains team member Anthony Sigillito. "However, there was a technical complication to overcome first in that the hyperfine modulation alone cannot flip nuclear spins." This is because the hyperfine field is parallel to the electron spin quantization axis, which in turn is set by the external effective magnetic field.

Exploiting the spin-orbit Stark effect

The researchers got around this problem by exploiting the so-called spin-orbit Stark effect, which modulates the orbit of an electron about the nucleus in a way that tilts the direction of the hyperfine interaction away from the direction of the external magnetic field. "In this arrangement, the components of the hyperfine interaction are in the right direction for manipulating nuclear spins," says Sigillito.

Being able to control nuclei with electric fields was difficult until now and the new technique could be used to develop quantum computers that use nuclear spins to encode quantum information, say the researchers. "In future quantum processors, we envision arrays of donor nuclear spin qubits that are separated by just tens of nanometres," says Sigillito. "In these processors, we will need to address or control individual nuclear spins, but cross-talk will be a problem if we use RF magnetic fields, since they are difficult to confine on the nanoscale. Electric fields, however, are easier to confine, so as devices are scaled down, we would like to – or may have to – move to some all-electrical method of controlling nuclear spins."

The new work is also interesting from a fundamental physics standpoint, he tells, since the experiments probe a new mechanism for driving nuclear spins. "Indeed, it will now be possible for us to glean more information about the donor spins in silicon by studying the dynamics of these spins."

The researchers, reporting their work in Nature Nanotechnology doi:10.1038/nnano.2017.154, say that they will now be trying to increase the electric field response of the donor nuclear spins. "We might achieve this by either applying strain to the material to enhance electrically-driven NMR or by replacing silicon with other materials, such as germanium (which has already proved itself to be promising in previous experiments performed in our lab)," says Sigillito. bes measure electron flow in graphene (Apr 2017)