Nanowire Project

Investigator: Ifan Hughes
Co-Investigator: Charles Adams
Post-Doctoral Research Associate: Kevin Weatherill
PhD Student: Adam West

EPSRC Grants: EP/F025459/1 and EP/F024886/1

Please select a section to read more about this project:

Mobile Atom Traps:

Schematic of a trapping potential based on domain wall fringing fields

The long term goal of the nanowire project is to create atom traps based on nanomagnetic domain walls. The resulting fringing fields have already been used to confine objects that are strong-field-seekers, e.g. quantum dot nanostructures. We hope to harness the fields to trap individual weak-field-seeking atoms.

This architecture has a number of attractive properties:

  • Scalability - can microfabricate millions of domain wall sites
  • Large field gradients - leads to extremely tight traps
  • Reconfigurability - magnetisation structure can be changed dynamically
  • Low noise - no currents flow, smooth field from point-like source, low conductivity materials eliminate Johnson noise

A trapping potential is formed very simply by biasing the fringing fields produced by a single domain wall. This procedure produces a magnetic field well approximated by a 3D quadrupole. The trap frequency associated with such a potential is expected to be of the order of 2π × 0.1-1 MHz.

There are a number of challenges associated with producing such traps. Notably, it will be difficult to load them, and we must take care to circumvent Majorana losses.

Scheme for producing a domain-wall based atom trap
The fringing field has a roughly 1/r2 form

A time-averaged potential scheme: external fields rotate the potential

A number of schemes have been successfully used to circumvent spin-flip losses in atomic traps:
  • Ioffe-type traps - QUIC, baseball etc.
  • RF-dressing
  • Time-averaged potentials (above right)
  • Blue-detuned plugs
Unfortunately, none of these is feasible in the regime of extremely tight traps. Whilst the use of a time-averaged potential is theoretically possible, the technical requirements associated with producing the required fields at frequencies of the order of 2π × 100 -1 MHz are huge.

Schematic of a piezoelectrically-actuated time-averaged potential

To overcome the problem of spin-flips we propose a novel type of time-averaged potential, which relies on the mechanical oscillation of the field source. This is achieved through the use of piezoelectric actuators. This has a number of advantages over conventional time-averaged potentials based on the addition of oscillating magnetic fields, particularly for very tight potentials:

  • More adiabatic potentials can be produced
  • Deeper traps can be produced
  • Technically easier - draw very little current

By the displacement of a domain wall by only 300 nm we can emulate the potential created via conventional time-averaging techniques. This has been observed to be a very easy task to achieve, even at frequencies above 2π × 1 MHz.

We also anticipate that the same technique will facilitate the production of toroidal traps (below right).

Once the atoms are loaded into traps, they can then be shuttled around. As in racetrack memory applications, domain walls can be deterministically moved around networks of nanowires. The realisation of such a scheme would provide a system with all the hallmarks of a quantum information processing architecture. The schematic shown below left illustrates how atoms working as qubits may be interacted via their close passage.

Nanowire networks could form the basis of a QIP scheme
for atoms

Toroidal potential formed via piezoelectric actuation of a domain wall