A Dual-Polarized Reconfigurable Reflectarray With a Thin Liquid Crystal Layer and 2-D Beam Scanning

Peyman Aghabeyki, Yang Cai, Guangsheng Deng, Zheng‐Hua Tan, Shuai ZhangIEEE Transactions on Antennas and Propagation 2023 · DOI

This research paper from Aalborg University presents a significant step forward in the design of reconfigurable antennas for 5G and satellite communications. By rethinking how liquid crystals (LC) are integrated into antenna "unit cells," the authors have found a way to make these devices faster and more versatile without the usual performance trade-offs.

TL;DR

The authors developed a dual-polarized reconfigurable reflectarray that uses an ultra-thin liquid crystal layer to enable fast 2D beam steering at 28 GHz. By sandwiching the LC between two coupled patches, they achieved a significant reduction in response time while maintaining a wide phase-shift range and high efficiency.

What Problem Does This Paper Address?

Reconfigurable reflectarrays (RAs) are essential for high-gain, steerable antennas. Liquid crystals are an attractive tuning material because they are low-cost and leverage the massive existing infrastructure of the LCD industry. However, LC-based antennas face a major "speed limit": the response time of a liquid crystal is proportional to the square of its thickness ($d^2$).

To make an antenna switch beams quickly, you need a very thin LC layer. The problem is that in conventional designs, thinning the LC layer usually ruins the antenna's performance, leading to high signal loss and a narrow "phase shift" (the ability to steer the beam). Most existing LC antennas are either too slow (thick layers) or only support single-polarization and 1D scanning.

Methods

The researchers introduced a novel "sandwich" unit cell design:

  1. The Structure: A bottom square patch and a top cross-shaped patch are separated by a thin LC layer. These two patches are coupled through two slots for each polarization.
  2. The Mechanism: By applying a variable voltage between the patches, the permittivity (electrical property) of the LC changes. This shifts the resonant frequency and the phase of the reflected wave.
  3. Dual-Polarization: The symmetrical cross-patch design allows the antenna to handle two orthogonal signal polarizations simultaneously.
  4. Local Phase Mapping: Recognizing that manual fabrication is imperfect, the authors proposed a measurement technique that divides the array into 12 regions. They mapped the specific phase-voltage curve for each region to compensate for physical non-uniformities during the beam-steering process.

Datasets

As an experimental hardware paper, it does not use traditional software datasets. Instead, the authors:

  • Used CST Microwave Studio for electromagnetic simulations.
  • Validated their design against a comprehensive comparison table of prior state-of-the-art LC reflectarrays (Refs [6], [8], [11], [12], [18]).
  • Performed experimental measurements using a Vector Network Analyzer (VNA) in an anechoic chamber.

Key Insights and Ideas

  • The Superstrate Shift: Instead of using LC as the main substrate (which requires thickness for performance), they used it as a tunable superstrate between two radiating elements. This "aha moment" allows the LC to be incredibly thin ($3.73 \times 10^{-4} \lambda$) while still providing a 270° phase range.
  • Scaling the Speed: By reducing the LC thickness from the typical 100+ µm down to 4 µm, the theoretical response time drops from seconds to milliseconds.
  • Error Compensation: The idea of "locally investigating" the phase behavior of different parts of the array is a clever workaround for the challenges of lab-scale manual assembly, ensuring the antenna can still steer accurately even if it isn't physically perfect.

Key Results

  • Ultra-Thin Layers: Successfully simulated designs with 30 µm and 4 µm LC layers.
  • Efficiency: Achieved a simulated aperture efficiency (AE) of 25.3% (for 30 µm) and 22.5% (for 4 µm).
  • Experimental Proof: A manually fabricated 10×10 prototype (30 µm layer) achieved a realized gain of 14.6 dBi and an AE of 12.5% at 28.5 GHz.
  • Steering Range: Demonstrated 2D beam steering of ±45° for both polarizations.
  • Switching Speed: The 30 µm prototype had a steering time of ~3 seconds; the 4 µm version is estimated to reach ~30 ms, making it viable for real-world 5G applications.

Limitations and Future Work

  • Fabrication Precision: The gap between simulated (25%) and measured (12.5%) efficiency was largely due to manual assembly errors and the use of thin copper layers (600 nm) necessitated by lab constraints. The authors note that industrial LCD manufacturing processes would significantly improve these results.
  • Phase Range: The current design provides a 270° phase shift. While sufficient for steering, reaching a full 360° would further reduce side-lobe levels and improve gain.
  • Material Loss: Future work involves using lower-loss LC mixtures (like GT7) which could potentially push the aperture efficiency above 30%.

Summary updated

Abstract

A novel approach for dual-polarized liquid crystal (LC) unit cell design is introduced, allowing significant thickness reduction of an LC layer without compromising its loss and phase shift. For each unit cell, LC is used for a tunable superstrate between a cross patch above and a square patch below. The square and cross patches are coupled through two slots for each polarization. By changing the voltage applied to the square and cross patches, the LC in-between varies the resonant frequency and phase of the unit cell to realize 2-D beam steering. <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10\times 10$ </tex-math></inline-formula> dual-polarized reflectarray (RA) with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30\,\mu \text{m}$ </tex-math></inline-formula> ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.0028 \lambda$ </tex-math></inline-formula> ) and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4\,\mu \text{m}$ </tex-math></inline-formula> ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$3.73\times 10 ^{-4}\lambda$ </tex-math></inline-formula> ) thick LC layers are simulated in the 28 GHz band, where the aperture efficiencies (AE) are 25.3% and 22.5%, respectively. As a proof of concept, the dual-polarized RA with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30\,\mu \text{m}$ </tex-math></inline-formula> thick LC is fabricated manually and measured at 28.5 GHz. A measurement technique is proposed to obtain the phase curves of the RA and reduce the effect of fabrication errors. With the imperfect manual fabrication, the RA still achieves the measured realized gain of 14.6 dBi corresponding to 12.5% AE. Besides, ±45° beam steering is demonstrated in the 2-D plane for both polarizations.
BibTeX
@article{aghabeyki2023dual,
  author = {Peyman Aghabeyki and Yang Cai and Guangsheng Deng and Zheng‐Hua Tan and Shuai Zhang},
  title = {{A Dual-Polarized Reconfigurable Reflectarray With a Thin Liquid Crystal Layer and 2-D Beam Scanning}},
  journal = {IEEE Transactions on Antennas and Propagation},
  year = {2023},
  doi = {10.1109/tap.2023.3240853},
}
Figure