Recently, reflectarray antennas [1–10] with a planar metasurface have gained significant attention in modern wireless communication systems owing to their ability to combine the advantages of reflectors and phased arrays. By equipping the planar metasurface with unit cells characterized by high reflectance and wide reflection phase variation, reflectarray antennas can effectively manipulate the phases of incident waves to achieve the desired radiation patterns. Among their various applications, multi-beam reflectarray antennas are particularly valuable for satellite communications, radar systems, and 5G networks because they can simultaneously cover multiple directions without requiring complex feeding networks or active components [11–15]. For instance, in [12], researchers divided the reflectarray surface into multiple sub-arrays, each designed to radiate a beam independently in a specific direction. This approach enables multi-beam formation, allowing each sub-array to generate a distinct beam and steer the transmitted signal toward the desired direction. Multi-beam characteristics were also achieved using a multi-feed reconfigurable source [14]. Furthermore, a method for multi-beam reflectarray design was proposed in [15], where the aperture fields associated with each beam were superposed to form beams at the reflectarray aperture. However, in this superposition method, the reflection amplitude of each unit cell was fixed based on the feed position and unit cell location, which were independent of the beam direction, making it difficult to control the radiation intensity in a specific direction.

In this paper, we propose a method for achieving multi-beam reflection by structurally reconfiguring the unit cells of a metasurface, originally designed for single directional reflection, into either a linear alternating or a checkerboard arrangement. The reconfiguration introduces spatial variations in the reflection phase across the aperture, effectively forming a discrete phase gradient. According to the generalized laws of reflection, such a phase gradient leads to directional redistribution of the reflected energy, enabling multi-beam generation. The proposed approach can also be interpreted within the framework of coding metasurfaces, where binary unit cell states act as digital phase elements. By controlling the ratio of each state, the radiation intensity of each beam can be selectively modulated, thus offering a structured and flexible beam control strategy that does not require continuous phase tuning.

This paper is organized as follows. Section II describes the reflection response of the designed metasurface unit cell and the design procedure adopted for the metasurface to achieve reflection in a specific direction, Section III presents the effects of the unit cell ratio of the designed metasurface for both the linear alternating and checkerboard configurations, and Section IV presents the conclusions.

Fig. 1 presents the structure and reflection response of the designed metasurface unit cell. As shown in Fig. 1(a), a circular ring structure with a ground plane is employed as the unit cell to ensure polarization-independent characteristics. The geometrical parameters of the unit cell are as follows: p = 15 mm, w = 0.4 mm, and t = 3.2 mm. The substrate utilized is ZYF 450, with a relative permittivity of 4.5 and a loss tangent of 0.0035. Fig. 1(b) shows that the reflection magnitude remains high (> −0.2 dB) even as the ring radius varies from 0.5 mm to 7 mm, achieving a reflection phase range of 310° upon changing the ring radius. Notably, although the unit cell size was about 0.4λ₀ at the operating frequency, it satisfied the essential criteria of a metasurface, including subwavelength-scale periodicity, wide phase coverage, and planar configuration [16, 17].

The full-wave simulated reflection magnitude and phase of the unit cell with regard to the ring radius were computed using commercial ANSYS Electronics desktop software. Notably, the main lobe of the beam was formed when the phase shifts (φ) matched the propagation-induced phase delays, expressed as = βp × sinθ_f, where p denotes the distance between the unit cells and θ_f refers to the desired beamforming angle.

Fig. 2 depicts the reflection phases of the metasurface at beamforming angles of −30° and 60°. The reflection phases were calculated by subtracting the incident phase at the metasurface from the phase shift required to direct the reflected wave in a specific direction. The operation frequency was 8.2 GHz, and the reflectarray antenna was fed by a rectangular horn antenna with a gain of 9 dBi.

Aperture efficiency can be expressed as the product of taper efficiency and spillover efficiency. The optimal F/D ratio for maximizing aperture efficiency is determined by the gain of the feed antenna, regardless of the aperture shape. In this study, the horn antenna was positioned considering an F/D ratio of 0.65 to optimize aperture efficiency [18]. At a beamforming angle of 0°, an antenna gain of approximately 22.70 dBi was achieved, along with an aperture efficiency of about 45%. Notably, the overall dimension of the metasurface composed of 14 × 14 unit cells is 210 mm × 210 mm.

To verify the feasibility and performance of the proposed dual-beam reflectarray antenna, we simulated and measured its characteristics by considering both the linear alternating and checkerboard configurations. The beamforming angles were chosen to be −30° and 60°, and the unit cell ratios investigated were 1:1 and 1:2.

Fig. 3 shows the reflection phase of the linear alternating metasurface with a unit cell ratio of 1:1, as well as its full-wave simulated far-field radiation pattern. The open and closed circles denote the metasurface unit cells employed for beam formation at −30° and 60°, respectively. According to the simulation results, peak gains of 16.3 dBi and 14 dBi were attained by the antenna at −29° and 54°, respectively. Furthermore, the simulation results for the beamforming angle exhibited an error of about 10%.

Fig. 4 shows the reflection phase of the checkerboard metasurface with a unit cell ratio of 1:1, as well as its full-wave simulated far-field radiation pattern. Simulated peak gains of 16 dBi and 13.5 dBi were achieved by the antenna at −29° and 55°, respectively. Overall, the results of both simulations demonstrate comparable performance, thus confirming the antenna’s feasibility.

Notably, as the beam-forming angle increased, the maximum gain, which is one of the indicators of antenna performance, decreased due to cos(θ)-dependent aperture reduction. Therefore, to control the radiation intensity of the dual reflected beams, the unit cell ratio was changed to 1:2, and the designed antenna was fabricated and experimentally validated. Fig. 5 shows a photograph of the reflectarray antenna with a unit cell ratio of 1:2 for the checkerboard configuration. A horn antenna (HS-090-UBR-10A; 8.2–12.4 GHz) with a gain of 9 dBi at 8.2 GHz was employed, and a height-adjustable jig was fabricated using a 3D printer. Notably, the far-field radiation pattern was measured in an anechoic chamber at 5° intervals.

Fig. 6 illustrates the reflection phase of the linear alternating metasurface with a unit cell ratio of 1:2, along with its full-wave simulated and measured far-field radiation patterns. As indicated by the simulation results, peak gains of 13.4 dBi and 16.5 dBi were achieved by the antenna at −30° and 56°, respectively. Fig. 6(b) confirms that the measured results agree well with the full-wave simulated results, except for a slight difference between the two caused by the jig supporting the horn antenna.

Fig. 7 shows the reflection phase of the checkerboard metasurface with a unit cell ratio of 1:2, along with its full-wave simulated and measured far-field radiation patterns. The simulation results show that the peak gains of the antenna were 13.3 dBi and 16.2 dBi at −31° and 54°, respectively. Moreover, this case also exhibits good agreement between the simulated and measured results.

It is evident that in the case of a unit cell ratio of 1:2, both the chosen configurations exhibit a higher gain at 60° than at −30°, despite the large beamforming angle compared to the 1:1 unit cell ratio. Furthermore, the antenna performances achieved using the linear alternating and checkerboard configurations are comparable, demonstrating the feasibility of the proposed antenna. The results establish that the proposed approach ensures consistent performance, regardless of the spatial arrangement. This implies that the choice of configuration should be based on fabrication constraints, integration requirements, and system-level design considerations, rather than performance optimization.

In this paper, the multi-beam radiation intensity control of a reflectarray antenna through unit cell arrangement is proposed and analyzed. A metasurface was designed to achieve directional reflection, and multi-beam reflection was realized by rearranging unit cells in either a linear alternating or checkerboard configuration. Additionally, the radiation intensity of the reflected beams was effectively controlled by adjusting the number of unit cells. The experimental results demonstrated that the measured and simulated results are in good agreement, thus verifying the feasibility of the proposed design. Furthermore, it was confirmed that a higher unit cell ratio contributes to an enhanced reflection gain, even at large beamforming angles. The proposed method can therefore be extended to practical applications, such as reflector antenna systems and reconfigurable intelligent surfaces (RIS), where directional control and gain enhancement of reflected beams are essential. Moreover, the ability to reconfigure unit cell ratios provides a flexible framework for implementing beam steering mechanisms in next-generation wireless and sensing technologies.