The strong demand for the advancing of Moore’s law on device size scaling down has accelerated the miniaturization of passive devices. Among these important electronic components, inductors are facing challenges because the inductance value, which is strongly dependent on the coil number for the air core inductor case, will be sacrificed when the size is shrinking. Adding magnetic core is one of the solutions due to its enhancement of inductance density but it will also add complexity to the fabrication process, and the core loss induced by the eddy current at high frequency is another drawback. In this report, the output of this research will be presented, which has three parts. In the first part, the CoZrTaB thin films are sputtered on different substrates and characterized comprehensively. The laminated CoZrTaB thin films have been also investigated, showing low coercivity and anisotropy field on both Si and polyimide substrates. Also, the different process conditions that could affect the magnetic properties are investigated. In the second part, Ansys Maxwell software is used to optimize the lamination profile and the magnetic core inductor structure. The measured M-H loop is imported to improve the simulation accuracy. In the third part, a novel method to fabricate the magnetic core inductors on flexible substrates is proposed. The sandwich magnetic core inductor is fabricated and assembled with flipchip bonder. The measurement result shows that this single-turn magnetic core inductor can achieve up to 24% inductance enhancement and quality factor of 7.42. The super low DC resistance (< 60 mΩ) proves that it is a good candidate to act as the passive component in the power delivery module and the use of polyimide-based substrate extends its compatibility to more packaging form factors.

This work is an investigation into the information provided by the concurrent use of in situ reflection high energy electron diffraction (RHEED) and reflection electron energy loss spectroscopy (REELS). The two analytical methods were employed during growth of metal, semiconductor and superconductor thin films by molecular beam epitaxy (MBE). Surface sensitivity of the REELS spectrometer was found to be less than 1 nm for 20 KeV electrons incident at a 2 degree angle to an atomically flat film surface, agreeing with the standard electron escape depth data when adjusted incident angle. Film surface topography was found to strongly influence the REELS spectra and this was correlated with in situ RHEED patterns and ex situ analysis by comparison with atomic force microscopy (AFM). It was observed in all the experimental results that from very smooth films the plasmon peak maxima did not fall at the predicted surface plasmon values but at slightly higher energies, even for nearly atomically flat films. This suggested the REELS plasmon loss spectra are always a combination of surface and bulk plasmon losses. The resulting summation of these two types of losses shifted the peak to below the bulk plasmon value but held its minimum to a higher energy than the pure surface plasmon value. Curve fitting supported this conclusion.