Phased array antennas are a critical element of modern wireless communication and sensing systems, enabling rapid electronic beam steering without mechanically moving the antenna. However, there are inherent challenges and limitations associated with phased arrays antennas. Notably, a drop in the antenna gain is observed as the beam is steered away from broadside, a phenomenon known as scan loss. Dielectric lenses, and in particular gradient index (GRIN) lenses, are a promising solution to mitigate this effect, potentially allowing for drop-in improvements to existing antennas.

Recent advances in additive manufacturing enable the realization of dielectric media with a continuously varying refractive index, allowing for more freedom in the form factor compared to conventional homogeneous or multilayer lenses. This thesis investigates the design, optimization, fabrication, and experimental validation of an additively manufactured flat GRIN lens to enhance the scan coverage of a planar phased array antenna.

A workflow for the realization of GRIN media using single-material fused filament fabrication (FFF) is presented. Spatially varying effective permittivity is achieved through sub-wavelength dielectric crystals based on triply periodic minimal surface (TPMS) structures, enabling continuously varying GRIN profiles. The dielectric properties of printed polylactic acid (PLA) are characterized experimentally using a split post resonator setup (courtesy of IT’IS). With these measurements, it is shown that PLA is suitable for proof-of-concept GRIN lenses at frequencies below 30 GHz, despite moderate dielectric losses.

A novel parametrization of GRIN lenses is introduced, in which the refractive index distribution is represented as a Fourier series expansion, normalized to the achievable minimum and maximum refractive indices imposed by manufacturing constraints. This formulation is especially well suited for use with curved-ray geometrical optics, as the gradient of the refractive index can be computed analytically. Based on this parametrization, the GRIN lens design problem is formulated as a multi-objective inverse problem, targeting a model for the ray direction and phase at the lens aperture. The optimization problem is solved using particle swarm optimization, enabling efficient exploration of a large design space with relatively low computational cost.

The proposed methodology is validated through the optimization of a Luneburg-like lens. The optimizer recovers a GRIN profile close to the ideal Luneburg lens, validating the ray-tracing algorithm as well as the fitness function. This approach is then applied to the primary design case; a flat, scan-enhancing GRIN lens for use with a planar phased array antenna. The optimized lens is experimentally evaluated in the Delft University Chamber for Antenna Tests (DUCAT) using a TMYTEK BBox 5G phased array antenna, operating at 28GHz. Measurements demonstrate an increase of +-10 degrees in the scan coverage compared to a free-space reference configuration, confirming the potential of additively manufactured GRIN lenses as a practical tool for drop-in enhancements of phased array antennas.