Design of RIS with 1-bit unit cell

As a reflective RIS, it should ensure reflective efficiency when designing the unit cell structure. In order to realize the reconfigurable function, it is necessary to provide a stable 180° phase difference between ON/OFF states. In this 1-bit metamaterial design, the switching diode element is selected to adjust the reflection phase difference, so this element needs to be designed at the unit’s peak value of the surface current. The proposed 1-bit RIS-based unit cell is shown in Fig. 1. In order to achieve a good reflection effect and prevent electromagnetic energy leakage, an RIS-based antenna generally consists of hundreds of unit cells. It leads to an increase in the number and complexity of biasing lines. Therefore, considering the size of the unit cell is too small for the half wavelength at the center frequency of 6.8 GHz (λ ₀ = 44 mm), each unit cell consists of 2 × 2 identical meta-atoms as shown in Fig. 1(a). It can avoid the complexity of circuit control and expensive manufacturing costs, although the accuracy of the reflection beam is slightly reduced. At the same time, the grating lobes can be avoided by increasing the unit interval.

Figure 1.

The configuration of the unit cell: (a) unit cell, (b) meta-atom on the side view, and (c) meta-atom on the top view.

As shown in Fig. 1(b), the meta-atom consists of two symmetrical rectangular patches (w ₁ × l ₁) loaded on the top side and bias lines on the opposite side. The patches and bias lines are printed on two identical substrates (FR-4, ε _r = 4.3, tan δ = 0.025, h ₁ = 2.5 mm, h ₂ = 0.5 mm) separated by a copper ground plane. One of the patches is connected with the bias lines by a metalized via hole. The thickness of the copper is 0.03 mm and the conductivity is σ = 5.96 × 10⁷ S/m. Moreover, the other patch is connected with the ground plane by another metalized via. The radius of all the via holes is 0.5 mm.

Figure 1(c) shows that the PIN diode connecting the two symmetrical metal patches is located in the middle of the y-axis. In the top layer, the metallized patches have four slots, which are used to enhance resonance. Moreover, two metallized patches are connected to the ground plane and the bias lines through two vias, respectively. So, the voltage between the PIN diode can be controlled by the bias lines.

The control variable method is used to optimize each dimension of the meta-atom by the parameter sweep function in Computer Simulation Technology (CST) Studio Suite. Scattering characteristics of meta-atoms in the PIN diode ON/OFF state with different sizes are shown in Fig. 2. It can be seen that the lengths l ₁ and w ₁ of the metal patches affect the working frequency band. When the target frequency is selected at 6.425–7.125 GHz, the final lengths of the patches are l ₁ = 6.2 mm and w ₁ = 13 mm. Other parameters in the element are optimized by similar methods. The geometrical parameters of the meta-atom are given as l = 15 mm, l ₂ = 3.65 mm, l ₃ = 1 mm, l ₄ = 0.1 mm, w = 15 mm, w ₂ = 3 mm.

Figure 2.

Scattering parameters of meta-atoms with the PIN diode ON/OFF state. For l ₁, (a) phase difference and (b) reflection amplitude. For w ₁, (c) phase difference and (d) reflection amplitude.

Considering various factors, SKYWORKS’s PIN diode, SMP1340-040LF, is selected as the key component of realizing the reconfigurable functions because of its broad frequency band and low insertion loss. According to the datasheet provided by SKYWORKS, a diode can be modeled as an RLC equivalent circuit, as shown in Fig. 3. According to the element scattering parameter file provided by SKYWORKS, the specific equivalent parameters of its ON/OFF states in the working frequency are obtained by using the parameter fitting method in Advanced Design System. In numerical simulations, the ON/OFF states of the PIN diode are represented by effective circuits displayed in Fig. 3. In the ON state, the real equivalent circuit is a series RLC circuit with a capacitance of C _ON = 1.6 × 10⁻¹³ F, an inductance of L _ON = 0.45 × 10⁻⁹ H, and a resistance of R _ON = 10 Ω. In the OFF state, it is a series RL circuit with an inductance of L _OFF = 0.45 × 10⁻⁹ H and a resistance of R _OFF = 1 Ω.

Figure 3.

The equivalent circuit of the PIN diode: (a) ON state and (b) OFF state.

The proposed meta-atom and unit cell are analyzed using periodic boundary by the full-wave simulation CST Microwave Studio. The unit cell conditions are set in the y- and x-directions. The perfect electric (E _t = 0) is set as the full ground in -z. The excitation source is the plane wave, which propagates in the z-direction, and the frequency-domain solver with a tetrahedral mesh is utilized for simulation. Figure 4(a) shows that the reflection phase difference of both the unit cell and the meta-atom between the ON/OFF state is 180° (±20°), and the reflection amplitude is more than 0.75 at the working frequency band. These indicate that the proposed element meets the requirement of RIS.

Figure 4.

Reflection parameters of the meta-atom and the unit cell: (a) phase and (b) amplitude.

RIS-based antenna

Figure 5 shows the proposed RIS-based antenna prototype. Based on the proposed 2 × 2 meta-atoms, the size of a 10 × 10 RIS prototype is 300 mm × 300 mm × 3.1 mm. Referring to related research about RIS in wireless communication [Reference Tang, Chen, Dai, Zeng, Zhao, Jin, Cheng and Cui4,Reference Tang, Li, Dai, Jin, Zeng, Cheng and Cui5], the horn antenna is chosen as feed in this design. It has higher directivity and gain than other antennas. Besides, it also has a better time-domain characteristic such as lower tailing, which reduces signal noise interference.

Figure 5.

The prototype of RIS-based antenna.

In traditional RIS designs, the object is placed in the far-field region of the feed to obtain better radiation performance. The aperture efficiency or gain can be optimized by properly choosing the feed location to balance the spillover efficiency and illumination efficiency [Reference Yu, Yang, Elsherbeni, Huang and Rahmat-Samii26]. The ratio of the feed distance and the maximum length of RIS increases, with the directivity of the feed antenna being better. High aperture efficiency is able to be acquired with a narrow beamwidth feed but at the cost of a large reflectarray antenna profile [Reference Dahri, Jamaluddin, Seman, Abbasi, Sallehuddin, Ashyap and Kamarudin27]. When the feed distance is shortened, the aperture efficiency of the feed antenna decreases and the beam width increases. When the feed distance is too short, it is necessary to correct the phase of the feed radiation. As the feed distance increases, the risk of radiation leakage increases and requires improved feed antenna performance. The ratio of the feed distance and the maximum length of RIS ranges between 0.7 and 1.0, which can provide a good aperture efficiency with a manageable reflectarray antenna profile. Therefore, this design chooses d = 300 mm as the distance between RIS and the feed antenna, in which the ratio is 1.

Suppose that the incident wave reaches the ith unit and radiates again to produce a reflected main beam in the direction of (θ ₀, φ ₀). According to the theory of array synthesis, the phase shift of the reflective unit can be expressed as

(1)\begin{equation}\phi \left( {{x_i},{y_i}} \right) = - {k_0}\left( {{x_i}\,\sin {\theta _0}\cos {\varphi _0} + {y_i}\,\sin {\theta _0}\sin {\varphi _0}} \right).\end{equation}

Here, k ₀ is the propagation constant of electromagnetic waves in a vacuum and (x_i, y_i) is the coordinate of the ith unit.

The compensation phase quantity ϕ _R(x_i, y_i) of the ith unit can be calculated as

(2)\begin{equation}{\phi _{\textrm{R}}}\left( {{x_i},{y_i}} \right) = {{2{{\pi }}c} \over f}\left( {{{\textrm{d}}_i} - \left( {{x_i}\,\cos {\varphi _0} + {y_i}\,\sin {\varphi _0}} \right)\sin {\theta _0}} \right),\end{equation}

where d_i is the distance between the feeder phase center and the reflection unit, c is the speed of light, and f is the corresponding operating frequency or the center frequency of the band.

Then, the compensation phase of each unit can be calculated through the above phase compensation formula. According to the above formula, the corresponding compensation phase of the preset reflection angle is calculated in MATLAB. Then, the corresponding ON/OFF state of each unit cell is set in CST. The full-wave electromagnetic simulation is completed in the end. In order to reasonably distribute the ON and OFF states to the unit phase, the unit cell is set to the OFF state when the compensation phase is [0°, 180°]. The unit cell is set to the ON state when the compensation phase is [180°, 360°]. In this way, the direction of the reflected beam is controlled. Similar to binary coding, the ON state is considered as “1” and the OFF state as “0.” Based on this mechanism, a 1-bit unit state distribution of 10 × 10 unit cells is obtained by using the feed position d and the target deflection angle.

Figures 6–8 show beams with a range of target deflection angle with their corresponding continuous and discrete phase distributions. The proposed RIS-based antenna can achieve ±40° beam scanning with a gain tolerance of 3 dB. The radiation patterns are basically coincident in different frequencies and the deviation of the main lobe angle is low, indicating that the prototype has stable performance in the full working frequency band. It will not be easily affected by frequency changes. It can be seen from Fig. 9 that the maximum gain is 18.7 dBi, which indicates that the proposed RIS-based antenna has a higher gain.

Figure 6.

(a) Phase continuous distribution, (b) phase discrete distribution, and (c) simulation pattern at 15°.

Figure 7.

(a) Phase continuous distribution, (b) phase discrete distribution, and (c) simulation pattern at 30°.

Figure 8.

(a) Phase continuous distribution, (b) phase discrete distribution, and (c) simulation pattern at 40°.

Figure 9.

Simulated gains at different scan angles at 6.5 GHz, 6.8 GHz, and 7.1 GHz.

In order to study the influence of different phase compensation mechanisms on the reconfigurability of RIS-based antenna, the unit cell sets to the OFF state when the compensation phase is [−90°, 90°]. The unit cell sets to the ON state when the compensation phase is [90°, 270°]. Based on this mechanism, another 1-bit unit state distribution of 10 × 10 unit cells is obtained by using the feed position d and the target deflection angle. Taking the target angle of 30° as an example, the phase distribution and the performance of the radiation pattern are shown in Fig. 10. In this mechanism, the RIS-based antenna still can reach the target angle, which means that a stable phase shift is the key to reconfigurable.

Figure 10.

(a) Phase continuous distribution, (b) phase discrete distribution, and (c) simulation pattern at 30°.

As shown in Fig. 11, the two kinds of phase compensation mechanisms have only a little difference in the radiation pattern. To compare the detail directly, the other three performances are shown in Fig. 12. In Fig. 12(a), the amplitude of the main lobe of the latter mechanism is at least 0.6 dB higher than that of the former one. So it can get a better gain when the unit cell is distributed in the latter one. At the same time, the angular width in 3 dB is reduced by up to 4.5°, which means that the RIS-based antenna has better direction performance in the latter one.

Figure 11.

A comparison of the radiation pattern: (a) at 6.5 GHz, (b) at 6.8 GHz, and (c) at 7.1 GHz.

Figure 12.

A comparison of the radiation pattern: (a) the main lobe magnitude and (b) the angular width (3 dB).

Experimental results

The RIS-based antenna is simulated by the CST Microwave Studio. Meanwhile, in order to validate the simulated results, the proposed RIS has been fabricated using the standard printed circuit board technology. Then PIN diodes are soldered on top of the RIS, as shown in Fig. 13(a). Moreover, the bias line network is fabricated on the back of the RIS, as shown in Fig. 13(b).

Figure 13.

The photograph of the RIS: (a) top view and (b) back view.

Based on the previous research [Reference Wang, Zeng, Zhao, Guo and Liu28], the digital control board is designed to control the DC-biasing voltage on each PIN diode to achieve reconfigurability. In this design, the meta-atoms have the same voltage in the same unit cell. So, it only needs to control 100 unit cells to reconfigure 400 PIN diodes. The digital control board design is shown in Fig. 14. The main control chip uses the Single-Chip Microcomputer STM32F103RCT6. The 13 shift registers (74HC595D) are utilized to control all of the unit cells separately. Meanwhile, each LED is connected in series with each unit cell to present the unit cell’s status. A Bluetooth module is designed to facilitate the interaction between the digital control board and computers. In the experiment, the RIS is connected to this digital control board to control the phase distribution of RIS. According to the relevant parameters of the control circuit board, the power consumption of this metasurface is about 6.6 W. Referencing the relevant research, the RIS has a range in power consumption from 5.32 W [Reference Wan, Xiao, Zhang, Li, Eisenbeis, Wang, Huang, Liu, Zwick and Cui29], 8 W [Reference Gros, Popov, Odit, Lenets and Lerosey30] to 234 W [Reference Kamoda, Iwasaki, Tsumochi, Kuki and Hashimoto31]. In this design, unit combination can simplify the difficulty of control and reduce the number of shift registers and the requirement for high-performance chips on the control board. Although considering the size of the unit cell and the size of the antenna aperture, the number of PIN diodes increases. The power consumption in this design is still maintained at a low level.

Figure 14.

The photograph of digital control board.

Figure 15 shows the measured results with different reflection beam angles at 6.5 GHz, 6.8 GHz, and 7.1 GHz. The measurement results agree well with the simulation ones. The measurement results agree well with the simulation ones, which means the proposed RIS-based antenna can achieve ±40° beam scanning with a gain tolerance of 3 dB. This also means that a variety of beam distributions can be generated by changing the unit distribution of the two states of the RIS, which indicates that the RIS can quickly create a variety of reflection beams with different modes so as to significantly improve the imaging speed. Due to the impacts of the experimental environment and the extra solder on the surface, there exist certain differences in the measured results. This problem can be solved by improving soldering technology. Meanwhile, the pattern distortion may be cursed by the quantization of the phase distribution. Since the measurement environment is not a microwave anechoic chamber and other factors interfere, additional backward radiation power is generated. There is a certain difference between the measurement backward radiation results and the simulation results.

Figure 15.

The measured results (a) at 15°, (b) at 30°, and (c) at 40°.

Table 1 shows the performance comparison of this proposed design with the relevant antennas [Reference Shuai, Liu, Yang and Liu32–Reference Zhang, Chen, Zheng, Hu, Qu, Zhao, Wang and Feng35]. It indicates that this proposed RIS-based antenna has a high peak gain, a small size for the miniaturization requirements, and a large beam scanning area. Especially, compared to the previous research [Reference Wang, Zeng, Zhao, Guo and Liu28], the phase compensation mechanism is optimized in this design, and the optimization of the 2 × 2 meta-atoms as a unit cell is proposed to simplify the bias network. In addition, the frequency selection is to fit the 2023 World Radiocommunication Conference (WRC-23) application background and provide data support for the discussion of WRC2023. In conclusion, the proposed RIS-based antenna has high gain and directivity and flexible reconfigurability, which can be utilized in wireless communication systems and other 5G applications.

Table 1.

Performance comparison of the proposed antenna and similar works

RIS of multi-beam

Beam imaging, as an important technology in current and even future wireless communications, helps to achieve dynamic target communication. The realization of multi-beam distribution is helpful for image information transmission in communication. Compared with traditional beam imaging methods, this metamaterial imaging method has the advantages of high speed and high accuracy. Some researchers put forward the metamaterial aperture for calculating imaging. This metamaterial aperture is excited by the waveguide and has frequency diversity, which makes it possible to realize stray beams with a low correlation coefficient in the frequency domain [Reference Li, Cui, Ji, Liu, Ding, Wan, Li, Jiang, Qiu and Zhang36]. The use of this aperture imaging not only makes full use of frequency domain resources but also reduces the complexity of beam imaging and can also obtain high-resolution target images. Combining FPGA technology, RIS can also realize automatic beam adjustment through the preset coding in the chip. Moreover, multi-beam distribution is vital for user target tracking and channel reallocation in wireless communication networks.

As shown in Fig. 16, by changing the unit distribution of the two states of the RIS, a variety of beam distributions can be generated, indicating that the RIS can quickly generate a variety of beams with different modes, which greatly improves the imaging speed and the quality of multi-receiver mobile communication. Due to the limitations of the current experimental conditions, the multi-beam radiation is only simulated. Experimental tests and further attempts in this direction will be considered in the future.

Figure 16.

Beamforming: (a) four-beam, (b) two-beam, (c) broad beam, and (d) multi-beam.