Summary
This experimental project concerns driving the dynamics of frustrated nanomagnets, such as artificial spin ices, using spin torques to study their response to a drive with a different symmetry to a magnetic field and to study their miniaturisability for applications in fields such as neuromorphic computing.
Frustration is the inability of a physical system to simultaneously satisfy competing constraints. It occurs across physics and beyond, but is a particularly important topic in magnetism, a field in which (relatively) simple systems can be represented by toy statistical mechanical models that can then be extended into other fields model phenomena as diverse as forest fires and financial networks.
The study of frustration in magnetism has recently been given a new lease of life since artificial frustrated systems can now be built and studied using nanotechnology: in the case of magnetism, this is done by constructing arrays of magnetic nanoelements arranged in patterns where their magnetostatic interactions are frustrated.
The advantages of this approach is that it is possible to build experimental realisations of models that nature does not provide crystal structures for, with every parameter in the model tunable by adjusting the element size, shape, and spacing. Moreover, the microstates of these artificial statistical mechanical systems can be inspected in detail using advanced magnetic microscopy methods, including time-resolved imaging presenting new ways of studying the underlying models, with potential applications in predicting their behaviour under different conditions.
However, manipulating the magnetic states of these systems has to date always been done by applying external magnetic fields. This requires bulky external magnets and can consume a great deal of energy. We therefore propose to use spin-torques to drive the dynamics of these nanostructures using electrical currents. These spin torques can have different symmetries to magnetic fields, opening up the prospect of new forms of dynamics. The fact that we will have an all-electrical system also means that it can be miniaturised into chip form, suitable for low energy applications in areas such as unconventional and neuromorphic computing.
