Enhancing the extraordinary magnetoresistance by variations in geometry and material properties

Magnetic sensors are a part of our everyday lives and constitute a critical component in modern-day electronics. Detecting tiny magnetic fields is an ever-increasing demand in applications ranging from information storage to navigation and biomagnetism. Conventional magnetic sensors have been pushed to the limit, yet they either lack a simple room temperature operation or the sensitivity required to detect ultraweak
magnetic fields. In 2000, the discovery of extraordinary magnetoresistance (EMR) laid the foundation for a new type of magnetoresistive magnetometers with the potential of combining room temperature operation and high sensitivity. EMR is a geometrical effect occurring in hybrid devices consisting of high-mobility materials (typically semiconductors) with a metal inclusion where the Lorentz force causes current deflection and magnetic field sensitivity. Yet, the research field of EMR remains relatively small and thus EMR sensors hold a great deal of untapped potential. Of particular interest is the high degree of sensitivity in EMR devices to geometric variations, which leaves ample room for sensor optimization through the many geometric variations possible.
A numerical model is necessary to navigate the vast landscape of possible geometric variations and the effects of material properties. In this thesis, I have developed an experimentally verified model using the finite element method (FEM) to first study the effects of inducing asymmetry by displacing the metallic inclusion in the EMR device. The asymmetric geometry is found to significantly boost the weak-field sensitivity (dR/dB) contrary to symmetric devices, where the sensitivity vanishes for B → 0. The asymmetric design further produced a large negative magnetoresistance as well as the possibility to change the magnetoresistive response from linear at B = 0 T to a step-like function. Changing the contact permutation enables switching between symmetric and asymmetric configurations on the same device which benefits from an enhanced weak-field sensitivity in the asymmetric configuration and a high MR at large fields in the symmetric configuration.
Second, I have used the model to investigate the effect of material properties to find trends that are universal across multiple EMR device geometries. This is done by comparing the numerical results across the three most common EMR device geometries and the asymmetric EMR device. The key criteria for selecting materials compatible with high magnetoresistance across all the studied geometries are a starting material with high mobility (≥ 20, 000 cm²/Vs), which can form a low contact resistance (≤ 10⁻⁴ Ωcm²) to a metal in addition to having a significant ratio (>500) between metal and semiconductor conductivity. I further find an inverse relationship between the semiconductor carrier density and the magnetic field sensitivity, making it essential to lower the carrier density through material choice, doping or gating.
The combination of an experimentally verified model and high-throughput experimental realization of EMR devices allows for a comprehensive investigation of geometric EMR effects. Here, 6-inch silicon-on-insulator wafers were used to realize batch productions of EMR geometries and establish a solid fabrication process. The lithography process showed great results with around 830 devices fabricated in parallel per wafer. However, consistent with the numerical modeling the contact resistance was found essential to realizing well-performing EMR devices and the metal deposition requires additional optimization to reduce the interface resistance beyond my attempts with lowering of silicon resistivity, native oxide removal and rapid thermal annealing. To lay out a future experimental direction, FEM modeling was used to screen the EMR performance of a range of high-mobility materials which pointed toward high-mobility graphene encapsulated by hexagonal boron nitride being the most promising candidate. The thesis ends with perspectives on interesting possibilities for future research including my preliminary results in some of the directions.