Experimental Demonstration of Peripherally-Excited Antenna Arrays

The paper addresses the challenge of achieving highly directive, electronically scannable 2D pencil beams at millimeter-wave frequencies with reduced cost and complexity compared to traditional phased arrays. Conventional solutions (e.g., parabolic reflectors, active phased arrays) face trade-offs among size, weight, cost, and complexity. The research question is whether a Peripherally-Excited (PEX) antenna array—excited by Huygens-equivalent sources placed only along the periphery—can be practically realized to produce single and multiple independently steerable pencil beams with high efficiency while substantially reducing the number of active elements. The study motivates the PEX approach as a means to reduce the active-element count from scaling with aperture area to scaling with aperture perimeter, while maintaining high directivity and enabling broadside and tilted beam steering along predefined contours.

Prior work to reduce cost/complexity includes thinned/sparse arrays and overlapped/interleaved subarrays, which reduce active elements but often degrade directivity, increase sidelobes, limit scan range, or require complex feeds. Continuous Transverse Stubs (CTS) arrays can scan via mechanical rotation and operate as leaky-wave antennas (LWAs), with good H-Plane scanning but still rely on mechanical motion for full coverage. Switched-beam lens-fed LWAs (e.g., parabolic or Luneburg) eliminate phase shifters by port switching but provide only discrete beams, require careful beamwidth overlap, may suffer aberrations/spillover, and are not typically optimized for broadside; often only one side of the LWA is excited, limiting scan contours. The PEX concept stems from the Huygens’ Box, which uses peripherally placed Huygens sources to synthesize arbitrary plane waves in a cavity; prior demonstrations validated wave synthesis within closed cavities but lacked a practical radiating implementation integrating peripheral sources, PCB compatibility, and unit cells enabling broadside radiation. The present work closes these gaps by demonstrating a PCB-based PEX array with practical peripheral sources and a unit cell engineered to close the broadside bandgap.

Concept and theory: The PEX array is a Huygens’ Box (thin metallic cavity formed by parallel plates and via sidewalls) with periodic radiating slots in the top plate. Using Schelkunoff/Huygens equivalence, plane waves are synthesized within the cavity by effective magnetic surface currents along the periphery. Practical peripheral Huygens sources are realized as coaxial feeds placed near the sidewalls and backed by vias to form a current loop that emulates magnetic current; spacing to the wall is set >λ/4 (in-cavity) to avoid impedance cancellation and mismatch. Radiating unit cell: A PCB unit cell on Rogers RT/duroid 5880 (εr≈2.2, thickness 1.575 mm) uses cross-shaped radiating slots (for robust leakage and polarization purity without scan blindness) plus four semi-square slots engineered to achieve accidental degeneracy at the Γ-point (broadside) to close the common open bandgap that prohibits broadside LWA radiation. Eigenmode simulations (HFSS eigenmode) optimize slot dimensions (cross and square notch sizes) to coalesce 1st–4th modes at 13.1 GHz with balanced Q (leakage), verified via 2D dispersion relations/contours. Analytical beam pointing: Beam directions are predicted by phase-matching between the in-cavity plane wave (angle ψ) and free-space radiation via Floquet theory, yielding simplified relations at the Γ-point using an effective permittivity derived from the unit-cell period-to-wavelength ratio. Full-wave simulations: A 15×15 array (unit cell size ≈14.3 mm × 14.3 mm; board 245.8 mm × 245.8 mm × 1.575 mm) with via sidewalls (0.8 mm dia., 2 mm pitch) was modeled in HFSS Driven Modal including conductor/dielectric losses. There are 31 coax ports per side plus four dummy corner ports (terminated 50 Ω). Sides are independently excited with equal amplitude and progressive phase shift to set ψ; unexcited sides are 50 Ω terminated. Broadside and tilted beams are characterized; realized gain, radiation/aperture efficiency, bandwidths, X-pol, and SLL are extracted; inter-side S-parameters quantify reflection, coupling, and transmission. Feeding networks: For experiments, modular 1×16 phase-shifting PCB networks (separate from the array) provide per-port progressive phase shifts; SMP cables link networks to array periphery. Two such boards and a 1×2 splitter enable dual-side excitation for multibeam tests. Experiment: A 7×7 PEX prototype (131.4 mm × 131.4 mm × 1.575 mm) with 15 ports per side (plus dummy corners), fabricated using standard PCB processes, was measured in a planar near-field system (NSI; VNA N5244A). Calibration used a broad-band horn. Cables/feed-network losses were de-embedded. Due to the Az/El measurement coordinate system, comparisons focused on cardinal planes equivalent to θ–φ E/H-planes. Single-beam tests excited side A (beams scan mainly in El plane) or side B (scan mainly in Az plane); multibeam tests excited sides A and B simultaneously; in all cases, other sides were terminated in 50 Ω.

Simulations (15×15 array at 13.1 GHz): • Broadside beam (ψ=0°, excite side A in-phase): realized gain ≈25.1 dB; aperture efficiency ≈84.6%; radiation efficiency ≈77.2%. • Electronic steering by changing ψ produces tilted beams up to ±33.5° (limited by grating-lobe onset) with realized gain ≈23.7–25.1 dB across ψ = ±20°. • Instantaneous bandwidths at broadside: 1 dB ≈2.5%, 3 dB ≈4.5%; for tilted cases, 1 dB ≈2.4–2.6%, 3 dB ≈4.2–4.4%. • Polarization purity and sidelobes: low cross-pol (E-plane often >20 dB, at broadside reported very high), sidelobe levels ~11–14 dB depending on plane. • Power flow example (side A excitation at broadside): ≈8.0% reflected to A, ≈1.3% coupled to B, ≈1.3% to D, ≈44.4% transmitted to opposite side C; highlights traveling-wave/termination trade-offs. • Analytical beam-pointing predictions (Floquet-based) align closely with full-wave results; scanning occurs along predefined contours set by unit-cell dispersion and Floquet modes within the unimodal (no grating lobe) region. • Multibeam simulations: Two and three simultaneous beams (e.g., ψ = 0° and 285° on sides A/B; ψ = −10° and 280°; plus ψ = 105° on side C) with independent steering and minimal mutual degradation due to orthogonal beam polarizations from orthogonal sides. Measurements (7×7 prototype): • Single-beam operation: Good agreement with simulations in beam directions and shapes. Measured peak directivity ≈19.8–22.1 dB (smaller aperture than simulated case explains lower gain). Maximum tilted scan ≈±33.5° (ψ≈±20°) before grating-lobe risk. • Bandwidths: Side A average 1 dB ≈2.1%, 3 dB ≈7.8%; side B average 1 dB ≈3.3%, 3 dB ≈8.2%. Slightly wider than simulations due to smaller aperture (wider beam) and phase-shifter dispersion. • Peak gain discrepancies (~2–3 dB) relative to simulation arise from amplitude variation versus phase in the practical phase shifters and measurement environment reflections. • Multibeam operation: Two simultaneous beams (broadside + tilted and tilted + tilted) were measured with independent steering and negligible mutual coupling, attributed to orthogonal beam polarizations from sides A and B. General: • The engineered unit cell closes the Γ-point bandgap at 13.1 GHz via accidental degeneracy, enabling efficient broadside radiation. • Active-element count reduction: excitation only along the aperture perimeter (e.g., 31 ports per side in 15×15 case; 15 per side in 7×7 prototype) excites hundreds of passive radiating cells, demonstrating the PEX concept’s scaling advantage.

The experimental results validate that a PCB-based PEX array can synthesize in-cavity plane waves using peripherally placed Huygens-equivalent sources and radiate them through a periodic leaky-wave unit cell to form narrow, electronically steerable pencil beams. The findings confirm the central hypothesis: PEX arrays can substantially reduce the number of active elements (perimeter scaling) yet achieve high directivity and efficient broadside/tilted beam steering along predefined contours. Agreement between analytical predictions, full-wave simulations, and measurements supports the physical model and design methodology. The orthogonal polarization of beams generated from orthogonal sides minimizes mutual coupling, enabling independent multibeam and potential duplex operation from a shared aperture. Noted trade-offs include: (i) limited scan directions (predefined contours and grating-lobe constraints set by unit-cell period), (ii) traveling-wave power management between radiation, termination, and ohmic losses affecting aperture/radiation efficiencies, and (iii) slight frequency-dependent beam squint inherent to LWAs defining instantaneous bandwidth. The work highlights paths to extend scan coverage beyond contours via mechanical rotation or by globally tuning the PEX cavity’s effective permittivity (e.g., ferroelectric substrates, artificial dielectrics), and discusses scalability to millimeter-wave frequencies with alternative feed implementations to mitigate connector pitch limits.

This work presents the first practical realization and experimental demonstration of a Peripherally-Excited (PEX) antenna array. A specially engineered slot unit cell achieves accidental degeneracy and closes the broadside bandgap at 13.1 GHz, enabling efficient broadside and tilted radiation. Simulations of a 15×15 array show high realized gain (≈25 dB), high aperture efficiency (≈85%), and acceptable bandwidths; measurements on a 7×7 prototype confirm electronically steered single and multiple pencil beams, with low mutual coupling between beams from orthogonal sides. The PEX approach offers substantial active-element-count reduction by exciting only the aperture perimeter while using passive radiators in the interior, and supports independent multibeam/duplex operation within a shared aperture. Future work includes extending scan range (mechanical rotation, tunable-permittivity fillings), optimizing feed networks (phase shifters with flatter amplitude response and digital control), beamforming with sidelobe control via peripheral weighting, tapered leakage profiles for combined high aperture and radiation efficiencies, scaling to millimeter-wave (and potentially fully metallic/artificial-dielectric implementations), and exploring higher-frequency-compatible feed architectures.

• Beam-pointing directions are constrained to predefined contours governed by unit-cell dispersion and Floquet conditions; full 2D coverage is not intrinsic. • Maximum tilt is limited by the grating-lobe onset set by unit-cell period (e.g., θc≈37° for d=14.3 mm at 13.1 GHz); reducing period increases scan range at the cost of fabrication complexity. • Traveling-wave operation entails a trade-off among aperture efficiency, radiation efficiency, and termination (resistive) losses; without tapering, increasing aperture size can reduce aperture efficiency. • Frequency scanning (beam squint) limits instantaneous bandwidth (few percent for 1 dB criterion). • Practical phase shifters used in the prototype exhibit amplitude variation with phase and frequency dispersion, causing gain discrepancies versus ideal simulations. • Measurement imperfections (environment reflections) introduced small deviations. • Scaling to sub-THz with the same phase-shifter approach is challenging; connector pitch and losses at high frequencies require alternative feed solutions.