Geometrical layout of the designed radiator

Figure 1 presents the structural layout of the designed antenna with different views. The substrate utilized to structure the dual-port antenna and PRS is Rogers RT 5880 substrate$\left( {{\varepsilon _{\text{r}}} = 2.2;\tan \delta = 0.0009} \right)$. The thickness of the substrate is 0.254 mm. Fan-shaped aperture has been etched from the substrate to excite the ceramic material. The ceramic material used as substrate is alumina having permittivity of 9.8 and loss tangent of 0.002. The optimized diameter (D) of cylindrical ceramic is taken as 8.0 mm. Table 1 lists the optimized dimension of various parameters of the proposed radiator. Figure 2 shows the pictures of fabricated radiator.

Figure 1. Geometrical layout of the designed radiator: (a) feeding structure, (b) expanded view of aperture, (c) partial reflecting surface, (d) side view of antenna.

Figure 2. Pictures of fabricated antenna: (a) designed aperture, (b) dual-port ceramic antenna, (c) partially reflecting surface, (d) 3D view of the proposed antenna.

Table 1. Optimized dimension of the designed mm-wave dual-port antenna

Analysis of the designed radiator

In this section, the detailed analysis of the designed radiator has been carried out using high-frequency structure simulator electromagnetic simulator. The analysis of the proposed antenna starts with single-port structure. Figure 3 shows the reflection coefficient variation of the designed antenna in the presence and absence of ceramic material. From Fig. 3, it is observed that the complete frequency band is due to the cylindrical ceramic material. Spectrum shown in Fig. 3 has two significant resonant peaks at 26.2 GHz and 32.1 GHz, respectively. In order to find the accountability of these resonances, Fig. 4 shows the E-field variation on cylindrical ceramic at 26.2 GHz and 32.1 GHz, respectively. From Fig. 4, it is confirmed that the resonant peaks at 26.2 GHz and 32.1 GHz are due to HEM_11δ and HEM_12δ modes, respectively [Reference Kajfez, Glisson and James16]. These resonant peaks can also be calculated as follows [Reference Mongia and Bhartia17]:

(1)\begin{align}{f_{{\text{r}},{\text{HE}}{{\text{M}}_{11\delta }}}} & = \frac{{6.321c}}{{\pi {\text{D}}\sqrt {{\varepsilon _{{\text{r}},{\text{eff}}}} + 2} }}\left[ 0.27 + 0.36\left( {\frac{D}{{4{H_{\text{D}}}}}} \right)\right.\nonumber \\ & \qquad \left. + 0.02{{\left( {\frac{D}{{4{H_{\text{D}}}}}} \right)}^2} \right]\end{align}

Figure 3. |S₁₁| variation in the presence and absence of DRA for single-port antenna.

Figure 4. E-field on cylindrical ceramic: (a) top view at 26.2 GHz, (b) side view at 26.2 GHz, (c) top view at 32.1 GHz, (d) side view at 32.1 GHz.

In eqn. (1), D and H _D denote the diameter and height of the alumina ceramic, respectively. From eqn. (1), the resonant frequency is found to be 25.95 GHz. In literature, no empirical formula is available to calculate the resonant frequency of HEM_12δ mode. However, it can be predicted on the basis aspect ratio of cylindrical ceramic by using the following formulation [Reference Sharma, Das, Gupta and Gangwar18]:

(2)\begin{equation}{{\text{f}}_{{\text{r}},{\text{HE}}{{\text{M}}_{12{{\delta }}}}}} \geq 1.25 \times {{\text{f}}_{{\text{r}},{\text{HE}}{{\text{M}}_{11{{\delta }}}}}}{\text{ }}\end{equation}

From eqn. (2), the resonant frequency of HEM_12δ mode is ≥32.43 GHz, which is closer to the simulated outcome.

Figure 5 shows the reflection coefficient (|S₁₁|) curve with various modifications in the shape of the aperture. Four different shapes are taken into the account: (i) square-shaped aperture, (ii) plus-shaped aperture, (iii) equal fan along with plus-shaped aperture, and (iv) unequal fan-shaped with plus-shaped aperture (proposed). From Fig. 5, it can be perceived that both the frequency bands are obtained with all shapes of aperture. However, only the impedance matching of each band is changing with aperture shape. The best reflection coefficient is achieved in the proposed case. One more thing is observed from Fig. 5 that square-/plus-shaped aperture is capable of creating both the hybrid modes, i.e. HEM_11δ and HEM_12δ. Aperture acts as a magnetic dipole, so it will be able to generate HEM_11δ. It is a very well known fact that HEM_12δ mode is the orthogonal mode of HEM_11δ. Square-shaped aperture excites the x-polarized and y-polarized waves with equal strength [Reference Guha, Gupta and Kumar19]. Due to which, HEM_12δ mode is also produced by square-/plus-shaped aperture at 32.1 GHz. Figure 6 shows the axial ratio variation with the aforementioned changes in the shape of aperture. It can be perceived from Fig. 6 that in case of asymmetrical fan-shaped aperture-loaded ceramic, the CP waves are obtained in both the operating bands. In the proposed aperture, fan blades are aligned orthogonally. These blades are able to create the orthogonal modes with equal amplitude. Now, change in the size of blades (make it asymmetrical) creates the path delay between the field line. This will result in the phase difference. Parametric analysis has been done to create the phase difference of 90°, which is shown in Fig. 7. It can be observed from Fig. 7(a) and (b) that as the fan blade size becomes nonuniform, the axial ratio moves toward the 3-dB down. In this wave, the necessary condition to create CP waves has been fulfilled by the proposed antenna [Reference Balanis20].

Figure 5. |S₁₁| variation with different changes in the aperture of single-port antenna.

Figure 6. Axial ratio variation with different changes in the aperture of single-port antenna.

Figure 7. Axial ratio variation with variation in size of fan blade: (a) lower operating band and (b) upper operating band.

After that, single-port antenna is converted into the dual port. There are two possibilities with dual-port antenna: (i) parallel orientation and (ii) perpendicular orientation. Figure 8 shows the reflection coefficient comparison of single-port and dual-port (with parallel and perpendicular orientation) antennas. From Fig. 8, it is confirmed that the reflection coefficient curve is approximately same in all three cases. Figure 9 shows the mutual coupling curve variation in parallel and perpendicular orientation of dual-port antenna. From Fig. 9, it is confirmed that mutual coupling reduced to −40 dB in case of perpendicular orientation. For this reason, perpendicular orientation has been chosen in the proposed one. Figure 10 shows the axial ratio variation with single-port and dual-port (with parallel and perpendicular orientation) antennas. It is confirmed that in all three cases, the axial ratio variation is approximately the same. It is required for MIMO antennas.

Figure 8. |S₁₁| variation of single-port and dual-port (parallel and perpendicular placement) antennas.

Figure 9. |S₁₂| variation of parallel and perpendicular placement of dual-port antenna.

Figure 10. Axial ratio variation of single-port and dual-port (parallel and perpendicular placement) antennas.

In the next step, a PRS is suspended over the radiator for tilts the radiation pattern obtained from different ports in different direction. In the proposed antenna, a superstrate that comprises of the parasitic square patches with continuous change in dimension, i.e. L ₁, L ₂, and L ₃ is placed over the dual-port antenna. These patches are called as capacitive grids. Beam titling can be achieved by continuous phase variation. This can be done through constant change in dimension of parasitic square patches (Fig. 1(c)). Therefore, the proposed superstrate acts as the PRS [Reference Qin, Gao, Mao, Wei, Xu and Li21]. Figure 11 shows the S-parameter variation in the absence and presence of PRS. From Fig. 11, it is confirmed that the S-parameter is approximately the same in both the cases. Figure 12 shows the axial ratio plot in the absence and presence of PRS. From Fig. 12, there is minute change in the axial ratio in the absence and presence of PRS. Figure 13 shows the 3D radiation pattern in the presence of PRS with port-1 and port-2 at 26.2 GHz and 32.1 GHz, respectively. From Fig. 13, it can be perceived that the radiation pattern is tilted to −35° with port-1 at 26.2 GHz, while beam is tilted to +35° with port-2 at 26.2 GHz. Similarly, in upper frequency band (i.e. 32.1 GHz), the beam is tilted to −35° and +35° with port-1 and port-2, respectively. The 3D far-field pattern shown in Fig. 13 is obtained after placing the PRS. Therefore, it shows tilting from broadside direction. Without PRS, the radiation pattern is broadsided (i.e. follows the operating mode). In the designed PRS, the dimension of unit cell is changing from edge to middle. Due to this uneven distribution of unit cell, phase gradient has been changed. It will tilt the radiation beam in the opposite direction [Reference Yu, Genevet, Kats, Aieta, Tetienne, Capasso and Gaburro22].

Figure 11. S-parameter variation of dual-port antenna in the presence/absence of PRS.

Figure 12. Axial ratio variation of dual-port antenna in the presence/absence of PRS.

Figure 13. 3D polar plot of dual-port antenna with PRS: (a) Pprt-1 at 26.2 GHz, (b) port-2 at 26.2 GHz, (c) port-1 at 32.1 GHz, (d) port-2 at 32.1 GHz.

Experimental outcomes

In this section, experimentally measured antenna parameters are compared with optimized simulated results. Figure 14 shows the measured and simulated S-parameter for the proposed radiator. It is measured by using keysight-based E8363C PNA. From Fig. 14, it is observed that there is good agreement between measured and simulated S-parameter. The proposed antenna works in between dual frequency range (i.e. 26.1–27.5 GHz and 31.7–33.6 GHz, respectively). Isolation level is more than 35 dB between the ports. Figure 15(a) and (b) shows the measured and simulated axial ratio variation in lower and upper working frequency band, respectively. It is measured inside the anechoic chamber. From Fig. 15, it can be said that there is good agreement between measured and simulated axial ratio. The designed radiator supports the CP waves from 26.5 to 27.1 GHz and 31.9 to 33.1 GHz in lower and upper working frequency band, respectively. Figure 16 shows the measured and simulated 2D left- and right-handed circular polarization (RHCP) radiation patterns in XZ plane at 26.75 GHz and 31.5 GHz with port-1 and port-2, respectively. From Fig. 16, it is observed that the patterns obtained from port-1 and port-2 have been tilted in different direction by ±35°. Another observation obtained from Fig. 16 was that the designed radiator acts as RHCP with both the antenna ports. After seeing the radiation pattern, it can be observed that cross-pol level is high in the designed radiator. It is due to presence of PRS. In the presence of PRS, cross-pol level increases with increment in beam tilting. In the designed antenna, the cross-pol level is approximately 15 dB down in the direction of maximum radiation. It is a good value of efficient radiator. For the proposed radiator, the value of FBR is 5.7 dB and 6.4 dB at 26.1 GHz and 32.5 GHz. respectively. It is low because of the presence of aperture. One can improve the value of FBR by placing the reflector on the lower side of the substrate.

Figure 14. Measured and simulated S-parameter of the proposed radiator.

Figure 15. Measured and simulated axial ratio variation of the proposed antenna: (a) lower frequency band and (b) upper frequency band.

Figure 16. Measured and simulated LHCP and RHCP radiation pattern: (a) port-1 at 26.75 GHz, (b) port-2 at 26.75 GHz, (c) port-1 at 31.5 GHz, (d) port-2 at 31.5 GHz.

Figure 17 shows the gain and radiation efficiency variation of the designed radiator. Gain is calculated with the help of two antenna techniques [Reference Yu, Genevet, Kats, Aieta, Tetienne, Capasso and Gaburro22]. From Fig. 17, it can be perceived that the radiation efficiency is more than 90% in the working band. The gain is approximately 7.0 dBi in both frequency ranges. Table 2 shows the performance comparison of the proposed dual-port ceramic-based mm-wave radiator with other existing antennas on the basis of impedance bandwidth, axial ratio bandwidth, gain, and pattern diversity. Data from Table 2 confirms the overall performance of the designed antenna is better in comparison to other existing ceramic-based mm-wave antenna.

Figure 17. Gain (measured/simulated) and radiation efficiency (simulated) curve of the designed mm-wave antenna.

Table 2. Performance comparison of the designed antenna with other existing ceramic-based mm-wave antenna

Diversity factors such as envelop correlation coefficient (ECC) and diversity gain (DG) are highly significant in case of multi-port radiator. ECC tells about the correlation (either in terms of scattering or far-field parameter) between the ports [Reference Sharawi1]. For competent multi-port radiator, the ECC should be as low as possible. DG expresses the gain of diversity aerial in the fading situation [Reference Sharawi1]. Generally, the value of ECC of an efficient MIMO is taken as <0.3, while the DG is approximately 10 dB. There are two methods to measure the ECC and DG, i.e. by S-parameter and by far-field parameter. The following formulation is used for this purpose [Reference Sharawi1]:

(3)\begin{align}* & {\text{EC}}{{\text{C}}_{{\text{S}} - {\text{parameter}}}}\nonumber \\ & = \frac{{{{\left| {{\text{S}}_{11}^{\text{*}}{{\text{S}}_{12}} + {\text{S}}_{21}^{\text{*}}{{\text{S}}_{22}}} \right|}^2}}}{{\left( {\left( {1 - \left( {{{\left| {{{\text{S}}_{11}}} \right|}^2} + {{\left| {{{\text{S}}_{21}}} \right|}^2}} \right)} \right)\left( {1 - \left( {{{\left| {{{\text{S}}_{22}}} \right|}^2} + {{\left| {{{\text{S}}_{12}}} \right|}^2}} \right)} \right)} \right)}}{\text{ }}\end{align}

(4)\begin{equation}{\text{EC}}{{\text{C}}_{\text{F}}} = \frac{{{{\left| {\iint\limits_{4\pi } {\left[ {{E_i}\left( {\theta ,\phi } \right)*{E_j}\left( {\theta ,\phi } \right)} \right]\operatorname{d} \Omega }} \right|}^2}}}{{\iint\limits_{4\pi } {{{\left| {{E_i}\left( {\theta ,\phi } \right)} \right|}^2}\operatorname{d} \Omega \iint\limits_{4\pi } {{{\left| {{E_j}\left( {\theta ,\phi } \right)} \right|}^2}\operatorname{d} \Omega }}}}\end{equation}

(5)\begin{equation}DG = \sqrt {1 - ECC} \end{equation}

In eqn. (3–5), the symbols have their usual meaning. Figure 18 shows the DG and ECC variation of the designed aerial using S-parameter. From Fig. 18, it is perceived that the ECC is <0.15 and DG is around 10 dB inside the operating band. Table 3 provides the value of ECC and DG of the proposed antenna using far-field. From Table 3, the ECC and DG of the designed aerial are in the standard limit.

Figure 18. ECC and diversity gain curve of the designed mm-wave antenna.

Table 3. Measured value of ECC and DG using far-field