The evolving spectral allocations in the mmWave bands across 30-100 GHz for 5G and beyond, and opportunities for spectrum sharing across licensed/unlicensed bands necessitate future front-ends to address multiple bands spread across the spectrum. The links at these frequencies are enabled by phased arrays exploiting beamforming to close the link budget. However, current mmWave phased arrays can only operate across a narrow range of frequencies due to 1. Narrow-band operation of transceiver circuits, in particular, power amplifiers (PAs). It is extremely challenging for a PA to achieve GHz channel bandwidths across multiple bands simultaneously over a 3:1 mmWave bandwidth while maintaining high efficiency and linearity. 2. Antenna spacing limitation of conventional uniform linear array geometry. Such arrays are fundamentally incapable of operating across a wide range of frequencies extending beyond an octave without spatial aliasing (grating lobes).
To break the bandwidth limitation and enable a frequency agile, multi-standard, low-cost array front-end with beamforming capability across 30-100 GHz, my work proposes 1. Multi-dimensional (device, architecture, and design methodology) techniques for mmWave broadband and high-efficiency PAs in Si and III-V process. Innovative ideas include stacked common base PA cells, mmWave harmonic engineering, Non-Foster impedance synthesis to enable high back-off efficiency over the wide frequency range, and deep learning-enabled PA inverse design methodology. Such broadband PAs also support concurrent multi-band operation capable of achieving higher data rates and quality of service. 2. Fundamental design and optimization strategies aiming toward a non-uniform sparse antenna array geometry, capable of operating across ultra-wide frequency ranges while overcoming the trade-offs between directivity, inter-element coupling, and grating lobe issues. 3. Design for the first time: an ultra-wideband (30-100 GHz) transmitter phased array system, with design highlights in broadband blocks such as phase shifter, on-chip passives, IQ upconverter, and antenna element. The single aperture size, low-cost prototype demonstrates the potential to redefine the dynamics of the next generation of multi-band wireless networks, enabling agility between the shared licensed and unlicensed bands.