This dissertation investigates approaches to enhance the performance, especially the sensitivity and signal to noise ratio of magnetoelectric sensors, which exploits the magnetoelectric coupling in magnetostrictive and piezoelectric laminate composites. A magnetic sensor is a system or device that can measure the magnitude of a magnetic field or each of its vector components. Usually the techniques encompass many aspects of physics and electronics. The common technologies used for magnetic field sensing include induction coil sensors, fluxgate, SQUID (superconducting quantum interference device), Hall effect, giant magnetoresistance, magnetostrictive/piezoelectric composites, and MEMS (microelectromechanical systems)-based magnetic sensors. Magnetic sensors have found a broad range of applications for many decades. For example, ultra sensitive magnetic sensors are able to detect tiny magnetic fields produced outside the brain by the neuronal currents which can be used for diagnostic application. Measuring the brain's magnetic field is extremely challenging because they are so weak, have strengths of 0.1–1 pT and thus requiring magnetic sensors with sub-picotesla sensitivity. In fact, to date, these measurements can only performed with the most sensitive magnetic sensors, i.e., SQUID. However, such detectors need expensive and cumbersome cryogenics to operate. Additionally, the thermal insulation of the sensors prevents them from being placed very closed to the tissues under study, thereby preventing high-resolution measurement capability. All of these severely limit their broad usage and proliferation for biomedical imaging, diagnosis, and research.

A novel ultra-sensitive magnetic sensor capable of operating at room temperature is investigated in this thesis. Magnetoelectric effect is a material phenomenon featuring the interchange between the magnetic and electric energies or signals. The large ME effect observed in ME composites, especially the ME laminates consisting of magnetostrictive and piezoelectric components shows a promise to make novel ultra-sensitive magnetic sensors capable of operating at room temperature. To achieve such a high sensitivity (∼pT level), piezoelectric sensors are materialized through ME composite laminates, provided piezo-sensors are among the most sensitive while being passive devices at the same time. To further improve the sensitivity and reduce the [special characters omitted] noise level, several approaches are used such as magnetic flux concentration effect, which is a function of the Metglas sheet aspect ratio, and resonance enhancement. Taking advantage of this effect, the ME voltage coefficient α _ME=21.46 V/cm·Oe for Metglas 2605SA1/PVDF laminates and α_ME=46.7 V/cm·Oe for Metglas 2605CO/PVDF laminates. The resonance response of Metglas/PZT laminates in FF (Free-Free), FC (Free-Clamped), and CC (Clamped-Clamped) modes are also investigated. α_ME=301.6 V/cm·Oe and the corresponding SNR=4×10⁷ [special characters omitted]/Oe are achieved for FC mode at resonance frequencies. In addition to this, testing setups were built to characterize the magnetic sensors. LABVIEW codes were also developed to automatize the measurements and consequently get accurate results.

Then two commonly used integration methods, i.e., hybrid method and system in package (SIP), are discussed. Then the intrinsic noise analysis including dielectric loss noise, which dominates the intrinsic noise sources, and magnetostrictive noise is introduced. A charge mode readout circuit is made for hybrid method and a voltage mode readout circuit is made for SIP method. For sensors, since SNR is very important since it determines the minimum signal it can detect, the SNR of each configuration is discussed in detail. For charge mode circuit, by taking advantage of the multilayer PVDF configuration, SNR=7.2×10 ⁵ [special characters omitted]/Oe is achieved at non-resonance frequencies and SNR=2×10 ⁷ [special characters omitted]/Oe is achieved at resonance frequencies. For voltage mode circuit, a constant SNR=3×10³ [special characters omitted]/Oe is achieved at non-resonance frequencies. Both of the advantages and disadvantages of each method are also discussed.

Piezoelectric single crystal PMN-PT with optimum orientation and cut direction is developed to increase the ME coefficient α _ME and reduce the intrinsic dielectric loss noise, consequently to improve the SNR of the ME sensors. For Metlgas/PMN-PT laminates, SNR=3.9×10 ⁶ [special characters omitted]/Oe is achieved at non-resonance frequencies and SNR=7.3×10 ⁸ [special characters omitted]/Oe is achieved at resonance frequencies.