Antenna design

As shown in Fig. 1, the proposed MIMO antenna is fabricated on a FR4 substrate with ε _r of 4.4, tanδ of 0.02, and the thickness of 1.59 mm. The proposed antenna consists of two skew symmetry open-loop square radiating elements printed on the top side of the substrate, and two rectangle ground planes with rectangle slot are fabricated on the bottom side of the substrate. Each radiating element contains a T-shaped slot and a small L-shaped slit. The small L-shaped slit mainly resonates at 3.5 GHz band, while the T-shaped slot inspires 4.5 GHz band. An I-shaped isolation structure with six small rectangle slots (0.5 mm × 0.3 mm) is applied to decrease the coupling between two elements. And the dimension of each antenna element is only 11.75 mm × 11.75 mm (0.1754λ ₀ × 0.1754λ ₀, λ ₀ is the free-space wavelength at the frequency of 4.5 GHz).

Fig. 1. Configuration of the proposed antenna. (a) top view. (b) bottom view.

Parameter analysis

To further understand the principle of the proposed MIMO antenna, the electric field distribution at two resonant frequencies is illustrated in Fig. 2. In Fig. 2(a), the simulated strongest electric field distribution at 3.5 GHz is mainly concentrated at the lower corner of the radiation element and the isolation structure. The lower intensity of electric field is scattered among the inner upper and right portion of DGS. As shown in Fig. 2(b), the larger intensity of electric field at 4.5 GHz is distributed at the two corners of the radiation element. Therefore, the geometry parameters affecting the proposed MMIO antenna performance can be optimized according to the electric field distribution.

Fig. 2. Simulated E-field distribution at two resonant frequencies. (a) 3.5 GHz. (b) 4.5 GHz.

Figure 3 illustrates the simulated S-parameters of the proposed antenna with different value of W8. As can be seen from Fig. 3(a), both resonant frequencies shift toward higher frequencies when the value of W8 becomes larger. As shown in Fig. 3(b), the isolation of the proposed antenna at the higher frequency band is affected by the value of W8. The optimized value of W8 is 0.6 mm.

Fig. 3. Simulated S-parameters of the proposed antenna with different value of W8.

The optimization of the value of W10 is exhibited in Fig. 4. S ₁₁ and S ₂₁ parameters almost remain unchanged for the lower frequency band with different value of W10. Therefore, the value of W10 can be effectively used to optimize S ₁₁ and S ₂₁ parameters of the higher frequency band. The most appropriate value of W10 is 1.0 mm.

Fig. 4. Simulated S-parameters of the proposed antenna with different value of W10.

Simulation and measurement result

The optimized configuration of the proposed antenna is shown in Table 1. The photograph of the fabricated antenna is presented in Fig. 5.

Table 1. Parameters of the proposed antenna (in millimetres).

Fig. 5. Photograph of the fabricated antenna. (a) top view. (b) bottom view.

Figure 6 shows the simulated and measured reflection coefficient. The measured impedance bandwidths of −10 dB reflection coefficients are 500 MHz (3.3–3.8 GHz) and 1500 MHz (4.3–5.8 GHz), respectively. Some deviation of S ₁₁ between simulation and measurement is resulted from the soldering process of SMA connectors at the higher frequency band. Therefore, the operating frequency bands of the proposed antenna can cover 5 G New Radio n78 (3.3–3.8 GHz), n79 (4.4–5 GHz) and WLAN (5.15−5.35 GHz) frequency bands. As shown in Fig. 7, the measured isolations at the lower frequency band and the higher frequency band are better than 15 and 17 dB, respectively. Some divergence between the measured and simulated results occurs due to the fabrication errors and losses.

Fig. 6. Simulated and measured reflection coefficients of the proposed antenna.

Fig. 7. Simulated and measured isolation of the proposed antenna.

To further demonstrate the performance of the proposed antenna, Fig. 8 presents the normalized measured antenna radiating patterns at both 3.5 and 4.5 GHz. Since the presented antenna has a symmetrical structure, port 1 is excited while port 2 is terminated by a 50 Ω load during the radiating pattern measurement process. As depicted in Fig. 8, the black solid and red dash lines represent co-polarization and cross-polarization, respectively.

Fig. 8. Measured 2D radiating patterns of the proposed antenna.(a) 3.5 GHz XOY-plane. (b) 3.5 GHz YOZ-plane. (c) 4.5 GHz XOY-plane. (d) 4.5 GHz YOZ-plane.

To better demonstrate the MIMO potentials of the presented antenna, ECC was calculated. As one critical standard to assess the performance of MIMO antenna, ECC is a measurement indicator of the correlation level between two radiating elements. The lower ECC is, the better the MIMO performance can be obtained. It is a consensus that the value of ECC must be lower than 0.5 to provide an excellent diversity gain [Reference Vaughan and Andersen20]. The ECC between antenna element i and antenna element j can be calculated by the far field pattern as shown in equation (1). The E_{θ /ϕ},_i/j (θ,ϕ) in equation (2) is the θ (or ϕ)-polarized electric far-field patterns of any two antenna elements in a spherical coordinate.

(1)$$\rho _{{\rm ij}}{\rm} = \displaystyle{{\iint_{4\pi } {A_{{\rm ij}}( \theta , \;\phi ) {\rm sin}(\! \theta ) d\theta d\phi } } \over {\sqrt {\iint_{4\pi } {A_{{\rm ii}}( \theta , \;\phi ) {\rm sin}(\! \theta ) d\theta d\phi \iint_{ 4\pi } {{\rm A}_{{\rm ij}}( \theta , \;\phi ) {\rm sin}(\! \theta ) d\theta d\phi } } } }}$$

where:

(2)$$A_{{\rm ij}}( \theta , \;\phi ) = E_{{\rm \theta , i}} ( \theta , \;\phi ) E^\ast _{{\rm \theta , j}}( \theta , \;\phi ) + E_{{\rm \phi }{\rm , i}} ( \theta , \;\phi ) E^\ast _{{\rm \phi }{\rm , j}} ( \theta , \;\phi ) $$

Moreover, ECC can also be estimated by the following equation (3):

(3)$$\rho _{{\rm ij}}{\rm} = \displaystyle{{{\vert {S_{{\rm ii}}^{\rm \ast } S_{{\rm ij}} + S_{{\rm ji}}^{\rm \ast } S_{{\rm jj}}} \vert }^ 2} \over {( {1-( {{\vert {S_{{\rm ii}}} \vert }^ 2 + {\vert {S_{{\rm ji}}} \vert }^ 2} ) } ) ( {1-( {{\vert {S_{{\rm jj}}} \vert }^ 2 + {\vert {S_{{\rm ij}}} \vert }^ 2} ) } ) }}$$

where S _ii is the reflection coefficient of antenna element i, and S _ij (i ≠ j) represents the transmission coefficient between two antenna elements. Equation (3) can be used only under some assumptions such as a lossless antenna in a rich isotropic multipath (RIMP) scenario. Equation (3) is selected to make a rough estimation of ECC. The real and imaginary part of both S ₁₁ and S ₂₁ were measured by the Vector Network Analyser (VNA) N5224A. Since the proposed antenna has a quiet symmetrical structure, only Port 1 is measured. The calculated ECC of the antenna is depicted in Fig. 9. The calculated ECCs of the antenna are below 0.02 in both operating bands, which is far smaller than 0.5.

Fig. 9. Simulated and measured ECC of the proposed antenna.

In Fig. 10, the peak gain and radiating efficiency were illustrated. The measured peak gain of the proposed antenna is nearly 4.5 and 5 dBi at 3.3–3.8 GHz and 4.3–5.8 GHz frequency bands, respectively. The measured efficiency is better than 55% in both operating bands. A performance comparison between the proposed antenna with state-of-the-art 5G/WLAN MIMO antennas is depicted in Table 2. The main contribution of our work is that we propose a dual-band 5G/WLAN MIMO antenna with lower ECC, smaller size and wider frequency bandwidths.

Fig. 10. Gain and efficiency of the proposed antenna.

Table 2. State-of-the-art of MIMO antenna.