Factors affecting the performance of a thin film magnetoresistive vector magnetometer

Abstract

Thin nickel-iron films with uniaxial anisotropy are vacuum-deposited onto glass. A bridge-connected group of conductive meander patterns is formed by photo-lithography; the current flow in each pattern is aligned at 45° to the anisotropy axis and the magnetoresistors are subjected to bias fields perpendicular to this axis. External fields cause magnetic vector rotation, producing magnetoresistive changes which unbalance the bridge. Poles along the edges of each meander pattern strip generate a demagnetisation field. Its magnitude and direction can be calculated from the effective pole density, which changes with the film's direction of magnetisation. The field is dependent on the geometry of the conducting strips. Calculation of the equilibrium magnetisation direction is based on the Stoner-Wohlfarth model, modified by the effects of the demagnetisation field. There is close correlation between theoretical and experimental results, particularly at higher bias levels where most of the film exists as a single domain. Vector magnetometers must have low sensitivity to orthogonal fields. Careful choice of pattern geometry and optimised bias fields in theory allow a zero orthogonal response, a prediction confirmed by experiment. The magnetometer has a linear response and the system noise is less than 1γ per Hz1/2.