As presented in Fig. 2, the meta-atoms of the TA consist of a three-layer metal patch printed on two layers of substrates. Notably, the TA was designed and optimized using a co-simulation of CST and MATLAB. It is characterized by a receive-transmit patch structure with metallic vias connecting the bottom (receiving patch) and top (transmitting patch) patches. The patches are separated by a metal plane in the middle to ensure effective electromagnetic wave control and transformation, as well as high transmission efficiency [7]. The substrate material was Rogers 4350B, with a relative dielectric constant (ɛ_r) of 3.66 and a loss tangent (tanδ) of 0.0037. The thickness of both substrates is 1.524 mm. A bonding film of Rogers 4450 with ɛ_r = 3.5 and a thickness of 0.1 mm was applied between the two substrate layers to bond them securely. The bottom patches are a set of square metal patches etched with U-shaped slots designed to receive the LP waves transmitted from the RA and the feed antenna and reflect orthogonal LP waves toward the RA. Furthermore, the middle metallic layer is a fully covered metal plane with circular holes to facilitate the passing of the metallic vias. The top patches (Top patches 1 and 2) are circular, with C-shaped slots etched at their centers and strip-shaped notches chambered in their counteractive edges. Notably, these top patches serve as radiating patches. By rotating them at different angles, the phase and amplitude of electromagnetic waves can be adjusted to realize the LP-to-CP conversion of vortex waves. Additionally, the C-shaped slots contribute to the CP OAM beam focusing over the dual operating bands. To achieve dual-band radiation, the smaller patch pairs, including Top patch 1 and the smaller bottom square patch, and larger patch pairs, including Top patch 2 and the larger bottom square patch, were interlaced and arranged at the bottom and top surfaces of the TA to generate higher and lower operating frequency bands, respectively. A 27 × 27 array of meta-atoms constitutes the entire TA.

To investigate the electromagnetic wave transmission performance of the TA, simulations were conducted by observing the effect of the rotation of the radiating patches (Top patches 1 or 2) on the amplitude and phase of the transmission coefficients when x- or y-polarized waves incident on the bottom surface of the TA. The results are presented in Fig. 3. T_{x_RHCP} and T_{y_LHCP} represent the transmission coefficients of the x- and y-polarized waves converted into RHCP and LHCP waves, respectively, while φ₁ and φ₂ are the rotation angles of Top patches 1 and 2, respectively. It was observed that when rotating the radiating patches designed for the lower (higher) band, the phase of T_{x_RHCP} (T_{y_LHCP}) varied with the rotation angle in the lower (higher) band, while the phase in the higher (lower) band remained nearly unchanged. Meanwhile, the amplitude curves in of T_{x_RHCP} and T_{y_LHCP} Fig. 3(a) indicate that the resonant frequency bands vary at around 9.0–11.0 GHz for the lower and higher bands, respectively. These demonstrate that by rotating the patches of the meta-atoms, the radiation in the lower and higher bands can be independently controlled. Additionally, the conversion of LP-to-CP waves can be validated using TA.

The RA is primarily employed in the design to convert LP waves into LP vortex waves. Its configuration is depicted in Fig. 4, showing that its meta-atoms feature an I-shaped structure. They were printed on a dielectric substrate of F4BM265 with a relative dielectric constant of 2.65, a loss tangent of 0.001, and a thickness of 3 mm. A metal ground plane was attached to the back of the substrate.

A 25 × 25 array of meta-atoms constitutes the RA, characterized by different rotations and transformations of its I-shaped structure at different positions. The meta-atoms were arranged symmetrically at the center of the substrate. Notably, these strategies work together to tune the phase and amplitude of electromagnetic waves to convert LP waves into LP vortex waves.

The simulated amplitude and phase of the reflection coefficients for different structural transformations of the meta-atoms in the lower band are detailed in Fig. 5. Notably, since the meta-atoms exhibited similar characteristics in the higher band, those observations are not presented here for brevity. Fig. 5 shows that the phase variation curves are nearly parallel over the lower band, implying a capability for independent phase adjustment. Meanwhile, the amplitude curves are almost stable and greater than −1 dB, which demonstrates high reflection efficiency. When reshaping the meta-atom by controlling the flare angle α, as shown in Fig. 5(a), and varying it from 10° to 130°, the phase variation of the reflected electromagnetic waves reached approximately 180°, indicating a wide phase-adjustment range and a wide operating bandwidth. Furthermore, when the whole meta-atom was rotated by 90°, as shown in Fig. 5(b), another continuous 180° range of phase adjustment was achieved for the reflected electromagnetic waves. These two separate phase adjustments were combined to achieve a 360° range of phase adjustment using the RA. Overall, this comprehensive phase adjustment enables the generation of LP vortex waves by strategically configuring and arranging meta-atoms.

The developed components were assembled together to construct the FTA antenna. The configuration of the proposed dual-port dual-band dual-CP OAM-generating FTA antenna is illustrated in Fig. 6. The antenna primarily consists of an RA, a dual-port dual-LP rectangular feed antenna embedded at the center of the RA, and a TA. The TA is stacked above the RA, with an air layer inserted between them. In this context, it should be mentioned that the feed antenna has a thin substrate composed of the dielectric material F4BM265 (ɛ_r = 2.65 and tanδ = 0.001), featuring a thickness of 1.5 mm. The feed antenna is fed by two SMP ports. Two sets of smaller and larger meta-atoms in the TA are interlaced with each other for the lower band with RHCP radiation and the higher band with LHCP radiation, respectively. Furthermore, the profile height of the antenna was optimized to 46 mm—approximately 1/5.4 of the focal length f, calculated using the formula f = D/(2 × tan(27.5°)) ≈ 246 mm, where D refers to the dimension of the metasurfaces [11].

When feed port 1 of the proposed antenna is excited, a lower band with y-polarized vortex waves will be produced between the TA and RA and then converted into RHCP vortex waves for radiation. When feed port 2 is excited, a higher band with x-polarized vortex waves will be produced from the RA and then converted into LHCP vortex waves for radiation. The operating OAM mode was l = −1. Notably, to generate an OAM mode of l = +1, the structure of the RA should be reversed in terms of the flare angle α of the meta-atom, the rotation angle of the entire meta-atom structure, and the position of the meta-atoms. Moreover when optimizing the antenna dimensions at a fixed operating frequency, the rotation angle and flare angle of the TA and RA should be adjusted correspondingly to achieve good impedance matching and other desired performances.

The simulated and measured S-parameters (|S₁₁|, |S₂₂|, and |S₂₁|) of the proposed antenna are presented in Fig. 7, along with a photograph of the fabricated prototype. It is observed that the measured impedance bandwidth (|S₁₁| or |S₂₂| < −10 dB) of the proposed antenna is 9.1–9.5 GHz (4.1%) and 11.1–12.1 GHz (8.6%) for the lower and higher bands, respectively. Meanwhile, the simulated impedance bandwidth (|S₁₁| or |S₂₂| < −10 dB) is 9.1–9.4 GHz (3.1%) and 10.9–12.1 GHz (10.4%) for the lower and higher bands, respectively. Furthermore, the measured isolation (|S₂₁|) is lower than −16 dB across the lower and higher bands. Notably, the lower and higher bands were primarily determined by the smaller and larger patches applied in the TA, respectively. The difference between the simulation and measurement results can primarily be attributed to manufacturing errors and measurement inaccuracies.

To quantitatively assess the characteristics of the vortex beam produced and the polarization conversion of the vortex waves, simulated near-field distributions of the proposed antenna at 9.2 and 11.2 GHz with OAM mode l = −1 were investigated, the results of which are depicted in Figs. 8 and 9, respectively. For each investigated frequency, the observed amplitude and phase planes were selected at 27 mm (0.83λ₀) above the top surface of the RA and 296 mm (9.1λ₀) above the top surface of the TA. The size of the observed plane was 320 mm × 320 mm (9.8λ₀ × 9.8λ₀). Additionally, OAM mode purities were calculated, defined as the ratio of the power of the OAM mode to the total power, as expressed by Eq. (4) [11]:

where l denotes the considered OAM mode, ω_l refers to the OAM mode spectrum, and ω_k signifies the existing OAM mode spectrum. Both the near-field distributions and mode purities were numerically calculated and visualized in MATLAB. As shown in Figs. 8 and 9, the simulated mode purities were greater than 0.8.

Figs. 8 and 9 exhibit one spiral period of phase variation at both 9.2 and 11.2 GHz. A doughnut-shaped counterclockwise rotation is observed in the phase distributions, indicating the vortex wave characteristics of OAM mode l = −1. Additionally, as observed from the amplitude distributions, a central hollow is present at the plane located at a height of 296 mm. This hollow cannot be observed in the amplitude distribution of the plane located at a height of 27 mm. This phenomenon can mainly be attributed to the positioning of the observed plane between the TA and RA, where the electromagnetic effects of both the TA and RA strongly interact with the near-field amplitude. Furthermore, an LP vortex beam with OAM mode l = −1 was produced between the TA and RA in the lower and higher bands. This can be observed in Fig. 8(a), 8(b), and 8(e), as well as in Fig. 9(a), 9(b), and 9(e). When passing through the TA, the LP vortex waves were converted into CP vortex waves, as can be observed in Fig. 8(c), 8(d), and 8(f), as well as in Fig. 9(c), 9(d), and 9(f). Notably, the phase distributions for the CP vortex waves at both 9.2 and 11.2 GHz do not show good characteristics for OAM mode l = −1, primarily due to the limited phase and amplitude control abilities of the metasurfaces applied in an antenna with a small volume.

The measured near-field distributions at 9.2 and 11.2 GHz are depicted in Fig. 10. The corresponding mode purities are also presented. The measurement process was performed using a near-field probe, which collected amplitude data of the x- and y-polarized components at multiple planes above the fabricated antenna. The calculated measured plane was 296 mm above the TA surface at 9.2 and 11.2 GHz. It is evident that the measured near-field amplitude exhibits a decline compared to the simulated one, while the measured phase distributions are consistent with the simulated results. Furthermore, the measured mode purity at 9.2 GHz is approximately 0.7. The discrepancy between the simulation and measurement can mainly be attributed to fabrication and measurement errors.

The radiation performance of the proposed antenna was measured in an anechoic chamber. The measurement setup is depicted in Fig. 11(a), while the simulated and measured axial ratios (AR) of the proposed antenna are presented in Fig. 11(b). The measured AR bandwidth (AR < 3 dB) ranges from 8.9 to 9.3 GHz (4.4%) and from 10.8 to 11.5 GHz (6.1%) for the lower and higher bands, respectively. The simulated AR bandwidth (AR < 3 dB) ranges from 8.9 to 9.4 GHz (5.5%) and from 10.7 to 11.5 GHz (7.2%) for the lower and higher bands, respectively. Furthermore, the simulated and measured two-dimensional (2D) normalized radiation patterns of the proposed antenna in the xoz plane are depicted in Fig. 11(c) and 11(d). An obvious hollow is observed at the center of the radiation patterns, demonstrating a distinctive characteristic of the vortex wave. Additionally, it is evident that LHCP and RHCP radiation patterns are generated in the lower and higher bands, respectively. Furthermore, the simulated peak gain of the proposed antenna reached 12.4 and 15.8 dBi at 9.2 and 11.2 GHz, respectively, while the measured peak gain reached 12.1 and 15.3 dBi at 9.2 GHz and 11.2 GHz, respectively. The simulated and measured cross-polarizations reached up to −3.6 and −2.3 dBi at 9.2 GHz, and 1.3 and 4.5 dBi at 11.2 GHz, respectively. The simulated radiation efficiency of the proposed antenna was found to be approximately 77% and 83% at 9.2 and 11.2 GHz, respectively. Notably, the discrepancies between the simulation and measurement results can mainly be attributed to fabrication errors and measurement inaccuracies.

Table 1 provides a comparison of the proposed antenna and other similar antennas reported in the open literature. This demonstrates that the proposed antenna has a distinctive characteristic—it employs the polarization conversion of the LP vortex wave into a CP vortex wave to achieve a dual CP vortex beam. In contrast, current approaches typically use CP or LP waves that are converted into CP vortex waves. Additionally, the proposed antenna offers the advantage of a decrease in the antenna profile-to-focal length ratio by up to approximately 1/5.4 for FTA and FRA configurations.