Engineering magnetic properties of garnet films and designing CoFeB-based magneto-ionic synapses for spin-based computing

The increasing demand for efficient data processing and storage in the digital era has driven spintronics-based computing technologies to the forefront of research. Magnetic materials with engineered properties hold great promise for energy-efficient devices, offering solutions for nextgeneration computing paradigms like neuromorphic computing and reservoir computing. This thesis focuses on exploring materials and device concepts for energy-efficient spintronics-based computing technologies. The first part of this thesis investigates bismuth-substituted yttrium iron garnet (Bi-YIG) films, which combine low magnetic damping and robust perpendicular magnetic anisotropy (PMA). The structural and magnetic properties of these films, including magnetic damping, ferromagnetic resonance (FMR) linewidth, and PMA, are systematically studied for varying thicknesses. Despite their excellent magnetic properties, broad FMR linewidths caused by magnetic inhomogeneities hinder spin wave propagation in these films, highlighting the need for further improvement in film growth. The second part of this thesis builds on these findings by optimizing the growth parameters of Bi-YIG films, leading to reduced FMR linewidths. Additionally, helium ion (He+) irradiation is employed as a post-growth process to modulate the magnetic properties of the Bi-YIG films, offering a versatile approach to further tailor the properties of Bi-YIG films for spin-based technologies. The irradiation process causes a continuous reduction in PMA and a spin reorientation from out-of-plane to in-plane magnetization at higher ion fluences, with minimal impact on key magnetic parameters, such as damping and saturation magnetization. The Dzyaloshinskii-Moriya interaction (DMI), which is crucial for stabilizing chiral spin textures such as Néel domain walls and skyrmions, is unchanged after irradiation. The combination of reduced PMA and constant DMI strength highlights the interfacial origin of DMI in these films. In the final part of this thesis, a lithium-ion-based magneto-ionic device is developed to explore its potential for neuromorphic computing. The device, designed as a solid-state supercapacitor, utilizes voltage-driven Li+ ion migration to reversibly control skyrmion and stripe domain states in a perpendicularly magnetized CoFeB layer. Its nonlinear response and intrinsic memory enable the emulation of synaptic functions, such as potentiation and depression. These properties allow the device to function as a physical reservoir, demonstrated through a sine-square waveform classification task. This highlights its promise as an energy-efficient and scalable building block for spintronics-based computing architectures.