|dc.description.abstract||The dawn of 5G mobile networks and other technologies like automotive radars, wireless LANs (WiGig) enables usage of millimeter wave frequencies in mass-consumer markets. In contrast to the existing sub 6 GHz applications, devices operating at 5G mm-wave bands are faced with unprecedented performance challenges to support wide bandwidths with high efficiency while keeping the costs low. Consequently, RF front ends, especially antennas are considered a major bottleneck to achieve the desired performance as existing printed circuit board based planar antennas e.g. microstrip antennas, have narrow bandwidth and suffer from higher conductor losses. These planar antenna technologies are however ubiquitous in existing, mass produced devices mainly due to their low cost and mature fabrication processes. Thus, an antenna technology competing to be a viable alternative to planar antennas at the mm-wave regime should not only meet the desired high performance standards, but also be cost effective and mass-producible. Dielectric resonator antennas (DRAs), mainly comprising of 3D ceramic blocks, are high efficiency, wide band radiators particularly suited for mm-wave frequencies, where the size of each block becomes small enough to be employed in compact wireless devices. However, the inherent 3D nature of these antennas is also a limiting factor in adopting them for mass produced wireless systems as existing fabrication techniques lack the ability to realize these antennas in large quantities cost effectively.
In this work, an approach suitable for fabricating millimeter-wave dielectric resonator antenna (DRA) and arrays is presented. The proposed concept addresses DRA fabrication challenges to make this technology a suitable replacement for planar antennas especially at the mm-wave frequencies for the 5G networks.
The proposed methodology involves fabricating precise cavities in polymer templates and filling them with composite dielectric materials to create a monolithic embedded DRA template layer. The excitation feed lines are fabricated on a separate substrate layer and the two layers are aligned and bonded together to form a monolithic antenna module. The approach of making enables not only precise realization of individual DRA elements but also ensures accurate placement of elements in the arrays. Deep X-ray lithography (DXRL) is used as an enabling technology to fabricate templates with precision and high accuracy. The templates are then used in electro-forming nickel molds, which are subsequently used in hot embossing of acrylic sheets for potentially low cost replication of DRA arrays. A detailed investigation into the impact of template material on embedded DRA performance is carried out through simulations and also scrutinized analytically. Scalability of the proposed fabrication approach is demonstrated by designing various prototype arrays. As a proof of concept, a 4 element rectangular DRA array operating at 60 GHz is realized. The performance is characterized through simulations and also experimentally verified. The 4-element template-based DRA array offers a wide impedance bandwidth (12% at 60 GHz from 56.5−63.5 GHz), high gain (10.5 dB at 60 GHz), high radiation efficiency (>80%) and broadside radiation characteristics with frequency stable patterns over the operating bandwidths. Other prototype arrays e.g. 2×4 (two 4-element arrays arranged in a parallel configuration and fed through a 1×2 divider), 1×8 (series 8-element array) and 1×24 (series 24-element array) have also been designed for 60 GHz operation and arranged on a single, common template fabricated by deep XRL. These arrays also demonstrate wide band, high efficiency and stable radiation patterns in the simulations and measured results, giving confidence in the feasibility of the proposed template-based DRA fabrication approach. Furthermore, to demonstrate the quick replication of the polymer templates, a nickel mold is electro-formed by using the 60 GHz XRL fabricated template as a master. The use of polymer-based materials provides opportunities for cost effective volume fabrication using molding techniques. For example, hot embossing can produce array templates in 10−15 minutes as compared to hours required in XRL exposure and development. The nickel mold is used to stamp the polymer sheets to demonstrate the potential of the proposed approach in mass producing mm-wave DRA arrays. Hot embossing process parameters are determined to achieve good transfer quality from mold to the stamped pieces. The replicated cavities in polymer templates prototypes are filled with dielectric composites, and demonstrate repeatable, wide band performance and excellent radiation characteristics. Importantly, to assess the fabrication quality for dimensional accuracy, the structures are examined after each stage and found to verify the highly precise and accurate fabrication potential of this process. Importantly, hot embossing significantly reduced the time and number of process steps to fabricate the templates, thus demonstrating the potential of this fabrication approach and the embossing technology to realize the DRA arrays in large numbers to serve the mass-produced device market.
In short, this work bridges an important gap in the field of dielectric resonator antennas, realizing potentially cost effective and mass producible DRA arrays while providing high performance at mm-wave frequencies.||