Antennas with bidirectional radiation patterns are useful in mobile wireless communication systems such as microcellular base stations, high-speed wireless local area networks, microwave sensor networks, coal mines, street microcells, and indoor wireless access venues [1–3]. Further, circularly polarized (CP) antennas are commonly employed to enhance channel stability and reliability because they provide a stable connection between the transmitting and receiving terminals (regardless of antenna orientation). Conventionally, CP antennas are produced by exciting two orthogonal linearly polarized (LP) antennas with the same amplitude and 90° phase difference.

Recently, CP antennas with bidirectional radiation patterns have gained popularity because they provide large signal coverage and can mitigate multi-path distortion and polarization mismatches [4, 5]. Accordingly, several bidirectional antennas have been developed with LP and CP radiation patterns. Bidirectional antennas with CP radiation mostly produce different senses of polarization unless specifically designed for the same polarization sense in both hemispheres. This means that when a right-handed CP (RHCP) antenna radiates from one side, the sense of rotation at the opposite side will inevitably be left-handed CP (LHCP) [6–9].

Many studies have suggested the goal of attaining CP radiation of the same sense in opposite directions. For example, back-to-back, slot coupled patches were used to achieve same-sense CP radiation from the front and back in [10]. Further, a bidirectional microstrip antenna fed by a coplanar waveguide (CPW) to obtain CP radiation was proposed in [11]. The authors in [12] proposed a novel bidirectional waveguide antenna design with CP of the same sense in opposite radiating directions. Similarly, a single-layer single-feed antenna with bidirectional CP radiation of the same sense was presented in [13]. A composite, right-left-handed transmission line was used to feed two orthogonal wire dipole arrays spaced by λ_o/4 (where λ_o is the wavelength of the operating frequency in free space) to achieve bidirectional CP radiation of the same sense in [14]. The authors in [15] presented a bidirectional same-sense CP slot antenna using a polarization-converting surface, while the authors in [16] presented a bidirectional same-sense LHCP antenna using a 4 × 4 metasurface arranged in a back-to-back configuration. While they achieved wide impedance and axial ratio (AR) bandwidths, the gain was low (with different values in the top and bottom directions) and the antenna dimensions were quite large. In addition, the antenna had poor mechanical properties due to the microstrip feedline and an air gap between the radiator and the metasurface.

This paper presents the design of a bidirectional, same-sense CP antenna using metasurfaces, which has attracted extensive interest in antenna design over recent years [17–24]. The antenna consists of two identical 2 × 4 truncated patches (metasurfaces) placed back-to-back—one at the top and one at the bottom. Same-sense CP is obtained by truncating the metasurface patches at the top and bottom sides in opposite orientations. Through the use of 2 × 4 metasurfaces, a small compact size wideband bidirectional CP antenna is achieved without significant loss in performance. Moreover, the resulting bidirectional antenna demonstrates acceptable performance characteristics: an impedance bandwidth of 5.21–6.26 GHz (18.4%), an AR bandwidth of 5.36–6 GHz (11.2%), a peak gain of 5.29 dBic, and a radiation efficiency of >96% within the AR bandwidth. The antenna’s dimensions are 48 mm × 24 mm × 3.048 mm (0.91λ_o × 0.45λ_o × 0.05λ_o at 5.7 GHz).

The geometry and dimensions of the proposed bidirectional antenna are shown in Fig. 1. It can be observed in Fig. 1(a) that the antenna consists of two layers: the front metasurface printed on the top of Substrate 1 and the back metasurface printed on the bottom of Substrate 2. A ground plane with an etched slot is sandwiched between the two layers without an air gap. Substrates 1 and 2 are RO4003 (ɛ_r = 3.38, tanδ = 0.0027) with thicknesses of h₁ = h₂ = 1.524 mm. The metasurface is a periodic structure, with eight square metal plates arranged in a 2 × 4 layout with periodicity P. Further, a truncation (dimension L_C) is made on the patches of size W_P × W_P, as shown in Fig. 1(b). The primary radiating element of the antenna is a slot of length L_S and width W_S etched on the ground plane [17–19], as shown in Fig. 1(c). This is sandwiched between the dielectric substrates, as shown in Fig. 1(d). The 50 Ω coplanar waveguide feed line is printed in the middle of the antenna structure, as shown in Fig. 1(d). The feedline characteristic impedance can be controlled by varying its width (W_f), while the stub (L_O) is used to match the antenna impedance.

Each patch on the back-to-back metasurfaces is shaped as a truncated corner square patch [22–28] to produce CP radiation. To obtain same-sense CP, the truncated patches on the bottom of the metasurfaces are rotated by 90° with respect to the truncated patches at the tops of the metasurfaces (to ensure they can be in opposite orientations). If the orientation of the truncation at the top and bottom sides was the same, different senses of CP radiation would be observed in opposite directions. The optimized design parameters of the antenna for wide impedance, wide AR bandwidths, and bidirectional same sense CP radiation patterns are displayed in Table 1.

Simulation and optimization of the bidirectional CP antenna were performed with an Ansys High-Frequency Structure Simulator (HFSS), which is an electromagnetic wave simulator using a finite element method. In particular, the effects of key geometry on antenna characteristics were investigated. First, the response of the antenna was determined with all parameters fixed at their optimal values. Second, one design parameter was varied at a time during the parametric study. The simulated reflection coefficient of the antenna is shown in Fig. 2. The antenna exhibits an |S₁₁| < −10 dB impedance bandwidth of 18.4% covering 5.21–6.26 GHz. Further, the effect of varying the slot length (L_S) and slot width (W_S) is shown in Fig. 3(a) and (b), respectively. It can be observed that when the slot length increased from 15 to 18 mm, the resonance shifted to the lower frequencies. Similarly, when the slot width increased from 1.5 to 2.5 mm, the resonance also shifted to the lower frequencies. The stub length (L_O) was varied to achieve impedance matching, as shown in Fig. 4. When the stub length L_O = 5.3 mm, the impedance of the antenna was not fully matched. However, when L_O decreased to 4.3 and 3.3 mm, the resonance frequency reduced and good impedance matching was obtained. At L_O = 3.3 mm, the antenna was fully matched.

As shown in Fig. 5, an increase in the patch size of the antenna also shifted the resonance frequency downwards. Further, Fig. 6 indicates that the antenna produced almost the same AR characteristics for both its front (θ = 0°) and back (θ = 180°). Its AR bandwidth ranged from 5.36–6.0 GHz, equating to a fractional bandwidth of 11.2%. The effects of varying the metasurface patch and corner cut sizes on the AR are shown in Figs. 7 and 8, respectively. Further, an increase in the metasurface patch size W_P from 11.6 to 11.7 mm shifted the AR bandwidth to lower frequencies, while a decrease in patch size from 11.6 to 11.5 mm shifted the AR bandwidth to higher frequencies (Fig. 7). When the corner cut size L_C was increased from 2.7 to 2.9 mm, there was a considerable improvement in AR performance. Moreover, a further increase in the cut size from 2.9 to 3.1 mm narrowed the AR bandwidth, as shown in Fig. 8.

The gain of the antenna is presented in Fig. 9, where it can be observed that the gain was almost identical for both the front and back. The antenna also exhibited stable gain characteristics, with an average gain of 4.6 dBic within the AR bandwidth. The simulated radiation patterns in the x–z and y–z planes for both RHCP (co-polarization) and LHCP (cross-polarization) at 5.7 GHz are presented in Fig. 10(a) and (b), respectively. Here, good symmetrical RHCP bidirectional radiation and low cross-polarization can be observed. Further, by truncating the metasurface patches at the top and bottom sides of the antenna in opposite orientations, RHCP was achieved.

The proposed bidirectional CP antenna was fabricated and measured to verify our design concept, and a photograph of which is presented in Fig. 11. For S-parameter measurement, an Agilent N5230A network analyzer and a 3.5 mm coaxial calibration standard GCS35M were employed. Far-field measurements were conducted at the RFID/USN Center, Incheon, Republic of Korea. For the radiation pattern measurements, a full anechoic chamber and an Agilent E8362B network analyzer were employed. The proposed bidirectional antenna was used for reception, and a standard wideband horn antenna was used for transmission, with a transmission distance of 10 m established between them.

The proposed antenna was rotated from −180° to + 180° at 3°/s and a 1° scan angle while maintaining the position of the horn antenna. Overall, the proposed antenna achieved good agreement between the simulated and measured data. However, there were minimal disparities between the measurements and the HFSS simulations, which could be due to slight alignment errors during fabrication. The measured and simulated reflection coefficients for the fabricated antenna are displayed in Fig. 12. The measured impedance bandwidth for |S₁₁| < −10 dB was 5.21–6.26 GHz (18.4%), which is quite similar to the simulated impedance bandwidth of 5.21–6.25 GHz (18.2%). Further, Fig. 13 shows the measured and simulated gains for the front and back of the bidirectional antenna with a simulated average gain of 4.6 dBic and a measured average gain of 4.27 dBic. The AR and radiation patterns of the prototype at 5.7 GHz were measured and are displayed in Figs. 10 and 14, respectively. In addition, Fig. 14 shows the measured and simulated AR bandwidth for both the front and back. The simulated AR bandwidth was 5.44–6.10 GHz (11.5%) and the measured AR bandwidth was 5.36–6 GHz (11.2%) The measured and simulated radiation patterns of the antenna in the x–z and y–z planes for both RHCP and LHCP are depicted in Fig. 10(a) and (b), respectively. The measured and simulated radiation patterns demonstrate good symmetrical RHCP bidirectional radiation and low cross-polarization.

Further, a comparison of the performance of the proposed antenna to that of other bidirectional antenna designs described in the literature is presented in Table 2. The proposed antenna exhibited wide impedance bandwidth, wide 3 dB AR bandwidth, good symmetric bidirectional radiation, and high gain. Although the structure reported in [12] achieved a wider impedance bandwidth compared to our design, its low gain represents a significant drawback. Moreover, the designs reported in [10, 11], and [13–15] achieved lower gains and significantly smaller operating bandwidths compared to the proposed antenna.

A bidirectional antenna with same-sense CP was proposed in this paper. The bidirectional antenna consists of two 2 × 4 truncated metasurface patches, one at the top and one at the bottom of the antenna. It radiates same-sense RHCP waves in both the front and back directions. Further, the antenna produces good symmetric bidirectional RHCP radiation patterns, an impedance bandwidth of 5.21–6.26 GHz (18.4%), an AR bandwidth of 5.36–6 GHz (11.2%), a bidirectional gain of 3–5.29 dBic (within the AR bandwidth), and a high radiation efficiency of >96%. These characteristics render the proposed antenna suitable for wireless communication in environments such as tunnels, long streets, and coal mines.