Modeling, design and fabrication of substrate-embedded inductors with high inductance density and low DC resistance for integrated voltage regulators

Abstract

There is an increasing need for voltage regulators to be integrated closer to active devices such as CPUs and GPUs. These integrated voltage regulators (IVRs) provide numerous performance benefits including higher efficiency, lower parasitics, and increased functionality while miniaturizing the overall system size. However, passive components i.e. inductors, generally occupy the largest volume in power distribution networks (PDNs). Therefore, realizing high-density inductors with ultra-thin form-factors is the main bottleneck to enable highly miniaturized heterogeneous integration of IVRs. An approach that can design cores and topologies with ultra-high inductance density without increasing the real-estate by using low-cost integration processes is required to address the challenges of developing inductors for IVRs. Metal-polymer composites (MPCs)-based interleaved substrate-embedded toroid inductors can meet all the criteria. MPCs as cores for inductor packages have high permeability, high resistivity, low eddy current losses and high frequency stability. An embedded interleaved toroid inductor topology using MPC cores can provide 50% more inductance, 33% higher Q-factor for the same DC resistance compared to an embedded solenoid topology. The combination of better material properties of MPCs, with an efficient toroid topology provides very high inductance densities at smaller inductor sizes, while maintaining low losses and DC resistance. The proposed work aims to improve upon current material approaches for inductor fabrication by providing better density, reduced thickness and DC resistance in a package-integrated format. This work aims to provide a basic understanding of the performance of a single-inductor using embedded toroid approach and the properties that govern its electrical behavior which can be further be scaled to coupled/tapped inductors in next-generation systems.