The magnetic detection of magnetically labelled disease biomarkers from a sample of bodily fluid presents an interesting architecture for a disease diagnostic device. Many advantages are offered over conventional optical labelling and detection techniques, including reduced background signals and sample pre-processing requirements due to the lack of magnetically responsive material in biological samples, as well as enhanced control over various steps of the assay protocol as magnetic labels may be actuated remotely via the application of magnetic fields.

In this thesis, Hall cross sensors are explored for the detection and enumeration of large numbers of magnetic particles for applications in magnetic immunoassays. The response of a Hall cross to a magnetic particle is directly proportional to the stray field of that particle averaged over the active area of the cross. As the active area size increases relative to the particle size, due to the solenoidal nature of the particle’s stray field, this average tends towards zero. As a result, large area Hall sensors suffer from low single particle signals, limiting device resolution. In addition, the Hall response is found to be highly inhomogeneous as a function of particle position, limiting the certainty with which particles may be counted. Hall cross sensors for which the active area size match the particle size produce a much larger response and can be used detect the binary presence or absence of a particle, thus resolving issues with both signal strength and homogeneity. However, dense arrays of individually contactable sensors must be fabricated to detect meaningful numbers of particles, limiting their usefulness.

This work focuses on the optimisation of large area Hall cross sensors towards the goal of counting large numbers of magnetic particles simultaneously with improved resolution and measurement uncertainty. It is hypothesised that the inclusion of perforations within the active area of such devices could make the Hall cross insensitive at locations where the stray field components of a landing magnetic particle reduce the overall Hall signal, thus enhancing the response. This concept is explored using COMSOL simulations and it is found that when an array of perforations is added to the active area of a Hall cross and particles land at certain subsets of positions relative to these perforations, both the magnitude of the Hall response and the homogeneity of the response with position are vastly improved. Experimental prototype devices are fabricated from GaAs/AlGaAs at which a 2DEG has formed and the response of perforated devices to arrays of magnetic disks is measured at room temperature, with the aim of demonstrating the same improvement. Good agreement between computational and experimental results is found for the perforated devices, while for the equivalent continuous devices, the measured response suggests that the fabrication of magnetic particles directly on top of the GaAs/AlGaAs resulted in the local, partial depletion of the underlying 2DEG. As such, these devices behave in a manner consistent with having partially formed perforations.