Wideband Slot Array Antenna Fed by Gap Waveguide with Right-Handed Circular Polarization and High Gain

A wideband and high-gain circularly polarized (CP) 16×16 array antenna based on gap waveguide technology is presented for millimeter-wave applications at 28GHz frequency range. Four cavity-backed slots with linear polarized (LP) radiation are used as the subarray. CP is obtained by a 4× 4 sequential feeding network which is also expanded to achieve high gain. +e feeding network of the final array antenna consists of two layers based on the ridge gap waveguide (RGW), and it has four unconnected metal layers. It is shown by simulation that the proposed antenna has 20.5% impedance bandwidth over 25.8–31.7GHz and 3 dB axial ratio bandwidth near 10% over 27.2–30GHz. In addition, the maximum gain value for this antenna is 31.6 dBi at a frequency of 29GHz, which shows good performance compared to other structures.


Introduction
In recent years, high-frequency and millimeter-wave frequency bands have gained much attention due to the increasing demand for high-rate data transformation. Lowcost and high-gain antennas are required for point-to-point communication in the millimeter-wave frequency range. Furthermore, planar antennas are the most common because of the compact system requirements.
To overcome the aforementioned problems, a new waveguiding technology called gap waveguide has been introduced in [18,19]. Gap waveguides can be realized without the need for good electrical contact between the metal layers [20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. erefore, this technology makes the manufacturing process easier and much cheaper than hollow waveguides. e advantages of gap waveguides compared to substrate-based transmission lines, such as SIW and microstrip lines, are that these structures can be realized totally with metal, which makes them have low losses and high power handling. e gap waveguide structure is based on two perfect electrical conductor (PEC) and perfect magnetic conductor (PMC) boundary conditions. e PMC boundary condition acts like a high impedance surface, and it can be realized by using a periodic texture of metal pins called bed of nails. If the gap between two plates is smaller than the quarter of wavelength, then no waves can propagate, and all parallel-plate modes will be in cutoff frequency. By using a ridge, groove, or microstrip line between the bed of nails, the field will propagate in the desired direction. ese techniques are known as ridge gap waveguide (RGW) [31], groove gap waveguide (GGW) [32], and microstrip gap waveguide [33], respectively.
Antennas with circular polarization are widely used in satellite communication in the move (SOTM) and point-topoint moving links due to their resistance to polarization mismatch and multipath effects. Applications of these links require CP antennas with proper radiation characteristics, such as high gain, broadband bandwidth, and broadband axial ratio (AR) bandwidth. ere are also other desirable features such as simple structure, low profile, and low cost for mounting in cars, airplanes, trains, and other moving vehicles. ere are two ways to obtain circular polarization in planar array antennas. e first method is to use CP antenna elements in an array. However, it suffers from narrow bandwidth [35][36][37][38][39][40][41]. e second method is to employ the antenna element and subarray with a feeding system, which results in circular polarization. e sequential rotation technique (SRT) is a suitable choice to achieve wideband AR where antenna elements are fed by the same magnitude and progressive phase shifts of 90° [42]. Several studies have proposed that the sequential feeding network is employed in conventional planar array antennas to improve AR bandwidth [43][44][45]. Recently, a two-level sequential feeding network for an 8 × 8 CP array antenna is presented, in which the antenna gain reached to 23.5 dBi [20]. In this paper, a 4-way sequential-phase feeding network based on the RGW is proposed. en, it is expanded to feed a 16 × 16 CP array antenna, which achieves the highest gain. e feeding network of the antenna consists of two layers of 64 subarrays fed with a 2 × 2 cavity-backed slot antenna. e geometry of the antenna and the simulation results of the subarray are presented in Section 2. In Section 3, the sequential feeding network for a 4 × 4 CP array is investigated. e simulation results of the final antenna are presented in Section 4, and finally, the paper is concluded in Section 5. Figure 1 shows the schematic diagram of a 2 × 2 cavitybacked slot subarray, with linear polarization that is similar to the one which has been used in [24]. It consists of three unconnected metal layers as shown in Figure 1(a). In the radiation layer, each slot is surrounded by a cylindrical cavity to suppress surface waves and mutual coupling between other adjacent slots. As explained in [24], separating E-and H-planes of the antenna from the principle plane of the array reduces the sidelobe level of the antenna. erefore, each slot is tilted by 10°as shown in Figure 1(b). e cavity layer is formed by the metal pins and feeds four radiation slots in the same phase and amplitude as shown in Figure 1(c). Figure 1(d) depicts the feeding layer of the subarray that is based on RGW technology and the quasi-TEM mode, which is coupled to the cavity layer by a T-shaped ridge via a coupling slot. e dispersion diagram of the pin unit cell used in the subarray is shown in Figure 2. It is calculated by CST Eigenmode Solver for a pin unit cell with periodic boundary conditions on two sides. e pin dimensions are chosen to have a stopband with covering working frequency in 28 GHz. As can be seen, the stopband of the pin for parallel-plate propagation modes starts from 17 to 48 GHz, and it is suitable for using in the subarray structure. e dimensions of the subarray are presented in Table 1. Figure 3 shows the simulated reflection coefficient of the subarray with the periodic boundary condition in CST Microwave Studio. e impedance bandwidth of the subarray is from 25.7 to 30.3 GHz. Figure 4 shows the configuration of the 4 × 4 CP array. e structure consists of four 2 × 2 subarrays with sequential rotation placement. As shown in Figure 4, the bottom layer is the sequential-phase network with four outputs that is similar to the one in [20]. To achieve a 90°phase shift between ports 2 and 3, the length difference between L 2 and L 1 is chosen as 5λ/4, where λ is the wavelength of the center frequency of the antenna. e 90°phase difference between ports 4 and 5 is obtained by the same method. In addition, a 180°phase delay between two branches of the first power divider is achieved by L L − l R � λ/2. e dimensions of the sequential-phase feeding network are optimized by CST Microwave Studio. e simulated reflection and transmission of the sequential feeding network are shown in Figure 5. It is clear that the reflection coefficient from 24 to 35 GHz is below −27 dB, and the imbalance transmission coefficient of the outputs is less than 0.4 dB over the entire bandwidth. Figure 6 shows the phase difference of the sequential-phase feeding network. As can be seen, a right-handed 90°phase shifting between the outputs is evident at 27.7 GHz.

4 × 4 CP Array Design
Simulated AR and gain of both 2 × 2 and 4 ×4 subarrays are presented in Figure 7. e AR magnitude for the 2 × 2 subarray is above 20 dB over the frequency range of 24-33 GHz due to its linear polarization. e 3 dB AR bandwidth of the 4 × 4 array is from 27.2 GHz to 30 GHz, which indicates that circular radiation has been achieved in this frequency range. e gain of the 2 × 2 subarray is below 15 dBi. By expanding this subarray to a 4 × 4 array, the gain has been increased above 17 dBi with the maximum gain of 19 dBi at 28.7 GHz. e schematic diagram of the presented feeding network is illustrated in Figure 8. ere is no enough space to place ridge lines and power dividers between 4-way power dividers for a compact integration due to the space limitation. erefore, the feeding network of the antenna cannot be placed in one single layer, and it is required to add another layer at the bottom of the structure. Subsequently, the final array antenna consists of four layers. Figure 9 shows the two feeding network layers of a 4 × 8 array. One sequential feeding network is rotated by 180°, and since the 4way sequential feeding network is not rotationally symmetric, a phase difference of 180°occurs between the outputs of the subfeeds at the top and bottom half of the entire feeding network. If identical radiators are connected to the International Journal of Antennas and Propagation feeding network, this will lead to a null at the boresight direction.

16 × 16 CP Array Antenna Design
To overcome this problem, the phase difference must be removed. To achieve this, an E-probe transition is designed between feed-1 and feed-2 to match the ridge gap waveguide line in the bottom layer to sequential feeding networks in the top layer. Figure 9 shows the simulated structure of the Eprobe transition between feed-1 and feed-2 layers. e dimensions of the transition are shown in Figure 10. As can be seen in Figure 11, the input reflection coefficient of the transition is almost below −20 dB over the frequency range of 24-33 GHz, and the output reflection coefficient is −3 dB. e simulated output phase differences of the proposed transition are shown in Figure 12. It can be seen that the         International Journal of Antennas and Propagation outputs have 180°phase difference. erefore, this phase difference and the phase difference caused by the rotation of the sequential feeding network are canceled with each other, and two 4-way sequential-phase feeding networks are excited in the same phase and amplitude. e expanded feeding network for a 16 × 16 array antenna is shown in Figure 8. In the bottom layer (feed-2), a feeding network with 8 outputs is designed. is feeding network consists of 7 T-junction power dividers and excites

Port 2
Port 3 g e l f w e Port 1 Figure 10: e structure of the E-probe transition between feed-1 and feed-2 layers (l f � 7 mm, w f � 3.45 mm, w e � 0.6 mm, g e � 2.38 mm, l t � 3.28 mm, w t � 1.34 mm, l b � 3 mm, and w b � 1 mm). all coupling slots in the feed-2 layer with the same phase and amplitude. An E-probe transition is utilized to match the TE 10 mode of the WR-28 rectangular   International Journal of Antennas and Propagation waveguide to the quasi-TEM mode of the ridge gap waveguide. e geometry of the proposed transition is shown in Figure 13. Figure 14 shows the perspective view of the simulated 16 × 16 CP array antenna simulated with CST and HFSS software to compare results. e simulated reflection coefficient of the array is illustrated in Figure 15. e result shows that the simulated results of CST and HFSS are in quite good agreement, and the impedance bandwidth is 20% from 25.8 to 31.7 GHz. As shown in Figure 16, the simulated 3 dB AR bandwidth is 10% from 27.2 to 30 GHz. e simulated gain of the array antenna is plotted in Figure 17. According to this figure, the gain is above 28 dBi from 24.5 to 31.9 GHz. e maximum gain value is 31.6 dBi which occurred at 29 GHz. Figure 18 shows the simulated normalized LHCP and RHCP radiation pattern of the antenna in two orthogonal planes at 28 and 29 GHz. e maximum sidelobe level of the antenna is about −13.9 dB, and the cross-polarization level is below −34 dB. Table 2 presents the results of several planar CP array antennas for comparison. e proposed antenna has good impedance bandwidth and maximum gain comparable to other antennas with different feeding mechanisms. However, the AR bandwidth has decreased compared to the common RGW-based antenna in [20], which has a two-level sequential feeding network, while the feeding network of the proposed antenna has one sequential-phase rotation level.

Conclusion
In this paper, a wideband and high-gain RHCP array antenna has been presented. e CP array uses a linear polarized 2 × 2 subarray based on GGW technology and employs the sequential rotation RGW feeding network. In addition, the 4 × 4 CP array is expanded to a 16 × 16 array antenna.
e simulated results show that the impedance bandwidth and 3 dB AR bandwidth are 20% and 10%, respectively, with maximum gain.

Data Availability
e CST and HFSS simulation data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.