Historically, stealth technology, in the form of applying paint on the surface of aircrafts to reduce their radar cross-section (RCS), has long been used to reduce incident energy. However, since the paint has to be applied frequently between flights, it poses a disadvantage in terms of cost. To address this, a semi-permanent reflector structure, which acts as an artificial electromagnetic surface, has been proposed. Recently, in-depth studies on reflectors have been conducted using various metasurfaces. Among them, checkerboard metasurfaces [1–7] have been found to be effective for monostatic RCSs because they apply destructive interference, which can generate a null at boresight. To scatter the RCS pattern, other structures have also been presented, such as metasurfaces on uneven layers [4], blended artificial magnetic conductor (AMC) surfaces [5], and diffusion metasurfaces [6, 7]. In addition, an electromagnetic gradient surface has been proposed to concentrate the main beam in the desired direction by using a gradient phase distribution [8, 9]. However, these structures [1–9] have been implemented using passive unit cells. Moreover, while structures exhibiting fixed RCS patterns are effective for monostatic radars, this can be a disadvantage in multistatic radar environments.

In [10], metasurfaces utilizing liquid crystals, switch diodes, varactor diodes, and programmable coding were reviewed. While liquid crystals offer a simple structure for state change, they are expensive. Meanwhile, metasurfaces using switch diodes offer only two states: on/off. Furthermore, programmable coding metasurfaces use field-programmable gate array to adjust each unit cell, making the structure more complicated. In contrast, metasurfaces using varactor diodes offer the following advantages: low price, multistate provision, easy design, and easy fabrication. To control the RCS pattern, some researchers have implemented tunable AMCs using varactor diodes [11–13]. For instance, Sievenpiper et al. [11] established a tunable impedance surface with active devices located between each unit cell. Additionally, Costa et al. [12] and Hong et al. [13] reduced the number of active devices by developing a grouping of 2 × 2 elements using only one varactor diode. However, the surface is bound to expand rapidly based on the surface area of the aircraft, resulting in the need for a large number of varactor diodes. For example, if the structure in [11] has 100 × 100 elements, the total number of varactor diodes required would be 19,800. In this case, the grouping technique [12, 13] would reduce to total to 1,250. Therefore, grouping more elements is an effective way to reduce the number of active devices. However, it poses a weakness with regard to the beam steering angle, which decreases as the period of the unit cell increases. Therefore, a novel structure is necessary to reduce the number of active devices while maintaining the beam-steering angle.

In this paper, a novel metasurface called the hybrid checkerboard metasurface is proposed. It combines passive and active metasurfaces to control the RCS using a small number of active devices. The passive metasurface is designed as a conventional checkerboard structure to achieve low RCS at boresight using the dispersed beam. Meanwhile, the active metasurface consists of active AMCs for beam steering to actively control the RCS. Furthermore, the RCS pattern is analyzed with regard to change in phase distribution.

To reduce the number of active devices, passive and active AMCs were made to co-exist in the proposed structure, as shown in Fig. 1(a). As presented in Fig. 1(b), the hybrid checkerboard metasurface is composed of 2 × 2 active and passive metasurfaces, with each metasurface containing 3 × 3 AMCs [14]. Its total size is 170.4 mm × 170.4 mm. Fig. 1(c) lists the phase distribution of the hybrid checkerboard metasurface. The passive metasurface maintains a 180° phase difference between neighboring unit cells, similar to a conventional checkerboard metasurface. Notably, since the interaction of passive AMCs causes destructive interference with each other, the monostatic RCS level is expected to be reduced. Meanwhile, the active metasurface is composed of 3 × 3 active AMCs, with the continuous phase distribution pertaining to progressive phase α considered for RCS pattern control. The expected total RCS pattern of the hybrid checkerboard is shown in Fig. 1(d). Notably, the proposed structure was investigated for transverse magnetic (TM) polarization.

The active and passive AMCs were designed on Taconic TLY-5 substrate (ɛ_r = 2.2, tanδ = 0.0009, thickness = 0.762 mm). Their detailed structures are depicted in Fig. 2.

With regard to the expected RCS pattern shown in Fig. 1(d), Fig. 3 presents the phase distribution and planar array factor considered in this study to investigate the basic characteristics of the hybrid checkerboard metasurface, when the progressive phase α is 0°. In Fig. 3(b), the solid black circle at the center represents a lobe generated by the in-phase active AMCs. This lobe can be steered based on phase distribution. Furthermore, the four black dotted circles represent the lobes dispersed by the passive AMCs, which is also observed in the case of a conventional checkerboard metasurface. Fig. 3(c) and 3(d) show the RCS pattern of the hybrid checkerboard metasurface and a perfect electric conductor (PEC) at φ = 45° and 135° planes, respectively. The maximum RCS level of the PEC is 10.6 dBsm at boresight. Meanwhile, when all active AMCs are in the in-phase state, the main lobe occurs at boresight, with the maximum RCS level being 4.4 dBsm. This result indicates a reduction of 6.2 dBsm achieved by the passive AMCs. In addition, while the RCS pattern is scattered for the passive AMCs, the RCS level is higher than the PEC in directions other than boresight.

As mentioned earlier, the RCS pattern can be controlled by changing the phase distribution of the active AMCs. Fig. 4(a)–4(d) show the 3D RCS patterns obtained upon changing α. With an increase in α, the main lobe steered along the x-axis. To visualize the beam steering clearly, the polar plot in the φ = 0° plane is shown in Fig. 4(e). The results show that by increasing α to 60°, the main lobe changes by about 10°. Furthermore, the RCS level at boresight varies by 4.4 dBsm to reach −20.3 dBsm, while the maximum RCS changes from 4.4 dBsm to 1.3 dBsm. Furthermore, when α = 180°, the main lobe divides in both directions, resulting in a 3-dB reduction. In conclusion, the main lobe of the RCS pattern can be steered by up to 31° in the φ = 0° plane, while the RCS level at boresight can be reduced by 24.7 dB.

To compare the performance of the proposed hybrid checkerboard metasurface with that of a fully active metasurface, both were investigated in the φ = 0° plane, as shown in Fig. 5. The fully active metasurface was designed to have the same dimensions as the hybrid checkerboard metasurface, thus using 6 × 6 active AMCs. In this design, each column maintained the same phase, while adjacent columns had a progressive phase shift of α. Fig. 5 depicts that the beam steering angle is almost the same for both structures, with the maximum RCS level of the proposed metasurface being 5.1–6.3 dB less than the fully active metasurface.

Similar to the investigation involving the fully active metasurface, the performance of the proposed metasurface was compared to that of a fully passive metasurface in the φ = 135° plane, as shown in Fig. 6. The fully passive metasurface was also designed using 6 × 6 passive AMCs, with each adjacent AMC having a 180° phase difference, to ensure fair comparison. Fig. 6 shows that the RCS level of the proposed metasurface at θ = ±46° is 5.8–7.5 dB less compared to that of the fully passive metasurface. Furthermore, when the phases of the active AMCs in the hybrid checkerboard metasurface were not the same—i.e., α = 60°, 120°, and 180°—its monostatic RCS levels were found to be 7.9–18 dB less than that of the fully passive metasurface.

This study also analyzed the RCS properties in terms of tilt angles and to investigate the bandwidth of the hybrid checkerboard metasurface. For this purpose, the bistatic RCS patterns from 9.8 GHz to 10.3 GHz are presented in Fig. 7. In Fig. 7, the RCS patterns appear to be distorted because the phase of each active AMC did not remain constant at frequencies other than 10 GHz. As shown in Fig. 7(a), the bistatic RCS level at −10° was found to be 3 dBsm at 10 GHz. Moreover, the level decreased as the frequency deviated, reducing to 7.6 dB at 9.8 GHz and 3.6 dB at 10.3 GHz. In contrast, the monostatic RCS increased by 15.6 dB at 10.1 GHz and by 23 dB at 9.8 GHz compared to −20.3 dBsm at 10 GHz. Furthermore, as presented in Fig. 7(b), the bistatic RCS level at −20° decreased by 3.2 dB at 9.8 GHz compared to 4.2 dBsm at 10 GHz, while the monostatic RCS increased by 7.9 dB at 9.9 GHz and by 13 dB at 10.3 GHz. As per Fig. 7(c), the bistatic RCS level at −31° decreased by 7.3 dB at 9.8 GHz and by 7.6 dB at 10.3 GHz compared to 1.3 dBsm at 10 GHz, while the monostatic RCS increased by more than 14.3 dB at 9.8 GHz, 10.2 GHz, and 10.3 GHz.

In this study, the bandwidth of the hybrid checkerboard metasurface is defined as the frequency range in which the monostatic RCS level at θ = 0° reduces by more than 10 dB compared to that of a PEC (10.5–10.9 dBsm), and the bistatic RCS level at each steering angle maintains a variation of not more than 3 dB. To facilitate a comparison of the RCS levels at each steering angle (θ = −10°, θ = −20°, θ = −31°), the monostatic and bistatic RCS levels of the hybrid checkerboard metasurface and PEC are listed in Table 1. Based on the results, the bandwidth of the hybrid checkerboard metasurface is determined to be 9.9–10.1 GHz.

For measurement, two double-ridged horn antennas for Tx and Rx operating at 2–18 GHz were employed. To measure the reflection phase of the active AMCs, 4 × 4 active AMCs were fabricated, as shown in Fig. 8(a). Subsequently, all active AMCs were excited by the same reverse DC voltage. Drawing on the Skyworks datasheet of SMV-1231, the voltage was excited from 0 V to 8 V to realize a capacitance of 2.35 pF to 0.56 pF. The reflection phase was measured using an Anritsu 37247D vector network analyzer, the results of which are shown in Fig. 8(b). The reflection phase is achieved from −128° to 120°, with the phase range being 248°. At a capacitance of 2.35 pF (0 V), the reflection phase was −128°, which is considered the reference phase of 0°. Furthermore, at capacitances of 0.97 pF (3 V), 0.794 pF (4 V), and 0.613 pF (6 V), the reflection phase was observed to be −14°, 56°, and 114°, respectively.

To measure the RCS pattern of the 6 × 6 hybrid checkerboard metasurface, it was fabricated as shown in Fig. 8(c). Considering the phase results obtained for the active AMCs, RCS pattern at phase distribution of α = 0°, 120°, and 180° were measured. Fig. 8(d) and 8(e) show the measurement environment. The position of Tx is fixed, while that of Rx is moved along θ. Notably, the time-gating function was used in this measurement to reduce noise and, thereby, justify the accuracy of the measurement results [16]. Measurements were conducted from −55° to 55° at 5° intervals. The results obtained at measured phase distributions of α = 0°, 120°, and 180° are shown in Fig. 9, indicating that the measured results are well matched with the simulation results. By changing the phase distribution, a 14.7 dBsm variation in RCS at boresight was achieved, while the main lobe changed up to about 30°. Overall, these measurement results highlight that RCS patterns can be actively controlled by different phase distributions.

Table 2 compares the findings of the current study with those of previous research in terms of the number of active devices employed, the beam steering angle, and the maximum monostatic RCS reduction. Considering that a structure has 100 × 100 elements, the hybrid checkerboard metasurface would require only 1,250 active devices. Moreover, despite the small number of active devices involved, the beam steering angle achieved by the proposed metasurface was 31°, and its maximum monostatic RCS reduction was 31 dB. Thus, not only did the proposed structure achieve a performance similar to the others in the literature, but it did so with a lesser number of active devices compared to previous research.

In this paper, a hybrid checkerboard metasurface that uses a small number of active devices to control the RCS at 10 GHz is proposed. The proposed structure is composed of passive and active AMCs as its two halves, thus decreasing the number of active devices. The results show a reduction in the monostatic RCS by 31 dB, while the main lobe of the RCS pattern changed up to 31°. A comparison of the number of required active devices, beam steering angle, and maximum RCS reduction between the proposed metasurface and those used in other works was conducted, showing the former achieving similar RCS properties as the latter, even when using a small number of active devices. Therefore, this low-cost and effective RCS pattern control structure can be considered an alternative to conventional active metasurfaces, since it employs fewer varactor diodes.