A High-Sensitivity, Low-Noise, and Low-Hysteresis Tunneling Magnetoresistance Sensor Based on Structural Optimization of Magnetic Tunnel Junctions

Accurate measurement of magnetic fields holds immense significance across various disciplines, such as IC circuit measurement, geological exploration, and aerospace. The sensitivity and noise parameters of magnetic field sensors play a vital role in detecting minute fluctuations in magnetic fields. However, the current detection capability of tunneling magnetoresistance (TMR) is insufficient to meet the requirements for weak magnetic field measurement. This study investigates the impact of structural and fabrication parameters on the performance of TMR sensors. We fabricated series-connected TMR sensors with varying long-axis lengths of the elliptical cross-section and adjusted their performance by modifying annealing magnetic fields and magnetic field bias along the easy axis. The results demonstrate that TMR sensitivity decreases with increasing long-axis length, increases initially and then decreases with an annealing magnetic field, and decreases with a higher bias magnetic field along the easy axis. The voltage noise level of TMR sensors decreases as the long-axis length increases. Notably, the detection capability of TMR sensors exhibits a non-monotonic dependence on long-axis length. Moreover, we optimized the hysteresis of TMR sensors by applying a magnetic field bias along the easy axis. When the bias along the easy axis reached 16 Oe or -40 Oe, the hysteresis level was reduced to below 0.5 Oe. After encapsulating the TMR devices into a full Wheatstone bridge structure, we achieved a detection capability of 17 nT/Hz@1Hz. This study highlights that the detection capability of TMR devices is jointly influenced by fabrication parameters. By optimizing parameter configuration, this work provides theoretical guidance for further enhancing the performance of TMR devices in magnetic field measurements.