Speaker
Description
Biomedical applications typically require very low noise magnetic sensors, since the magnetic field to be measured are usually very small. Traditionally SQUIDs have been the only sensors able to reach noise levels low enough to be employed in this field. However, the necessity of cryogenic temperature limits the applications of these sensors to fields where extremely low noise is required. Meanwhile, during the last lustra the noise of other sensors have been strongly reduced until becoming competitive in fields like magnetocardiography where tens of pT in a bandwidth from 1 Hz to few tens of Hz is required. Among them orthogonal fluxgate have shown remarkable results achieving noise well below 1 $\mathrm{pT}/\sqrt{Hz}$ in the bandwidth of interest [1]. One of the advantages of orthogonal fluxgates is the high spatial resolution which, together with the low cost allowed to create a matrix of sensors to map the magnetic field produced by a human heart [2].
In this contribution I will present how it was possible to achieve these results. Starting from the requirements for magnetocardiography (MCG) diagnostic and in particular the necessity to measure the T wave on a large number of points, I will present the methods used to achieve low noise magnetometer with high spatial for MCG. These include the development of a composition for amorphous Co-rich magnetic microwires which retain low magnetostriction when annealed to increase their circular anisotropy; this is obtained by decreasing the amount of iron so that the negative magnetostriction becomes close to zero after annealing. I will present the dependence of the noise on the geometry of the sensor and in particular how to increase the spatial resolution in the axis of the sensor without causing an excessive increase of noise. Unexpectedly, even using multiple shorter coils on the same core does not significantly increase the noise.
A very important aspect of biomagnetic fields is that they rapidly decay in space, therefore one my be lead to believe that, for instance, just a few cm far from the chest the MCG signal is vanishing and not measurable. From this point of view we show how the ferromagnetic core of the fluxgate acts as a flux concentrator which carries the signal even at several centimeters far from the chest making it measurable. An important focus will be given to the electronic which creates low noise excitation current for the sensors, as well as amplify and demodulate the signal from the pick-up coil and how this was designed to minimized the noise of the magnetometer. In particular, I show how a proper design of the preamplifier can reduce the current absorbed by the coil and return higher sensitivity contributing to decrease the noise floor of the sensor.
Another important aspect of the implementation of orthogonal fluxgates for real-life applications like MCG is the manufacturing process of the sensor itself. Traditionally fluxgates are known to have a low noise only on few samples over a large batch of produced sensors. Since applications like MCG require a large number of sensors with low cost it would not be feasible to use this technology unless a very high yield ratio. I will show how we managed to achieve a yield larger than 95 $\%$ on batch of serially produced sensor, proving that the technology is ready for mass implementation. Finally I will also discuss the limitations of this technology with special regard to offset stability with temperature and how, however, this limitations can be mitigated when not directly corrected.
