Magnetic sensors are used in a wide spectrum of applications, from keyboard switches to hard disk read heads to mineral exploration systems. However, most of the higher performance magnetic sensors used to date are made of materials that are non-IC process compatible. Capability to integrate magnetic sensors with circuits is of great interest because of the potentially better signal-to-noise ratio, lower cost, and higher reliability in an integrated sensor. Silicon would be the material of choice if one could successfully make high performance magnetic sensors out of it. It has been known that one of the impeding factors for achieving high performance silicon magnetic sensors is the relatively low silicon mobility. In this research, we begin by examining the ultimate sensitivity that ca...[ Read more ]

Magnetic sensors are used in a wide spectrum of applications, from keyboard switches to hard disk read heads to mineral exploration systems. However, most of the higher performance magnetic sensors used to date are made of materials that are non-IC process compatible. Capability to integrate magnetic sensors with circuits is of great interest because of the potentially better signal-to-noise ratio, lower cost, and higher reliability in an integrated sensor. Silicon would be the material of choice if one could successfully make high performance magnetic sensors out of it. It has been known that one of the impeding factors for achieving high performance silicon magnetic sensors is the relatively low silicon mobility. In this research, we begin by examining the ultimate sensitivity that can be achieved by a mobility-limited split-drain MAGFET (magnetic field effect transistor). We then circumvent the low mobility problem by designing two new magnetic sensors that use a combination of the carrier domain effect and positive feedback on an SOI(silicon on insulator) substrate.

We began the research with the split-drain MAGFET because it is probably the most MOS-compatible sensor. The device is made by splitting the drain of an ordinary MOS transistor into two. Our analytical model confirms that its sensitivity is directly related to the mobility. The sensor is found to have a maximum sensitivity of 5.8% per Tesla. It is also found that by applying asymmetrical biasing at the two drains, we can improve the sensitivity.

Silicon magnetic sensors's sensitivities can be greatly enhanced if both the carrier domain effects and the positive feedback are employed. In the past, attempts have been made to exploit these effects in bulk silicon technology. But, because of biasing requirements, those devices were not practical. Here, we solve the irksome biasing problems by taking advantage of the superior isolation offered by SOI. Correspondingly, we have developed two new magnetometers: the Lateral Carrier Domain Magnetometer (LCDM) and the Lateral Thyristive Magnetometer (LTM). The LCDM has a relative sensitivity of 12.5%/Tesla. At the same time, the LTM achieves a relative sensitivity of 210%/ Tesla at a total biasing current of 25 [muA], making it the most sensitive and lowest power silicon mag-netic sensor reported to date.