A dual-band MIMO PIFA antenna with a printed rectangular open-ended ring for 5G/WLAN mobile terminal

The design procedure of the proposed antenna element is divided into four steps. Figure 2(a) visualizes the design steps of this antenna, while figure 2(b) depicts their S₁₁ responses. First, the typical PIFA antenna is designed to operate at 3.6 GHz to cover the 3.4-3.8 GHz band (Ant. a). Here, the resonant frequency is determined by the length and width of the radiating plate, which can be approximated by the formula (1) [20]:

$$\begin{eqnarray}&&{f}_{0}=\displaystyle \frac{c}{4({w}_{p}+{l}_{p})\sqrt{{\epsilon }_{r}}}\end{eqnarray} \tag{ 1 }$$

where c is the velocity of light, l_p and w_p are the length and width of the radiating plate, respectively, _r is the relative dielectric constant of the substrate, and f₀ is the operating frequency of the PIFA. From the result of the Ant. a, which is depicted in figure 2(b) (Ant. a), it can be clearly observed that the PIFA provides a resonance at 3.6 GHz with a −6 dB working bandwidth of 500 MHz (3.37-3.87 GHz), covering the 3.6 GHz 5G band. It is worth noting that the achieved −10 dB bandwidth is 280 MHz (3.47-3.75 GHz). Figure 3 depicts the current distribution of the Ant. a at 3.6 GHz. As shown, the current distributes on the surface of the strip and on some edges of the radiating plate of the PIFA.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Then, the next step in our design consists of achieving dual-band characteristics to cover the entire WLAN band (5.15-5.85 GHz). For this, an L-shaped parasitic element and an L-shaped slot are added to the design (Ant. b and Ant. c). Here, instead of using the traditional method of etching slots on the radiating plate to excite the second resonance frequency of the PIFA and then adding a parasitic element to enhance the upper-band bandwidth, we propose a new method consisting of using a parasitic element to achieve dual-band operation and slots on the radiating plate to extend the upper-band bandwidth. The reason behind selecting the parasitic element to excite the first resonance frequency of the upper band rather than the slots is that etching slots on the radiating plate creates a narrow bandwidth that decreases with increasing the length of the slot, which influences the total bandwidth of the upper band. To prove that, the reflection coefficients S₁₁ of the L-slotted PIFA with different values of L₂ are displayed in figure 4(a). It can be clearly seen that the upper band bandwidth decreases when the length L₂ increases: when L₂ increases from 9 to 11 mm, the upper resonance frequency moves from 5.7 to 5.1 GHz, and the upper bandwidth shifts from 24 to 13 MHz. Further, figure 4(b) demonstrates that the upper bandwidth of the PIFA with only the L-shaped parasitic element is about twice as large as that of the L-slotted PIFA. It is worth mentioning that the size of the L-slot and the L-shape parasitic element are selected to have a resonance at 5.5 GHz. Finally, a comparison between the traditional and the proposed methods is depicted in figure 5 which shows that the proposed method has good impedance matching and a wider bandwidth.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The reflection coefficients of Ant. b and Ant. c are displayed in figure 2(b). The −6 dB and −10 dB working bandwidths of the upper band of Ant. b are 350 MHz and 190 MHz, respectively. For Ant. c, the −6 dB working bandwidth of the upper band is 730 MHz, while for the −10 dB impedance condition, Ant. b has two bands in the upper band with a bandwidth of 240 MHz and 150 MHz, respectively.

To further explain the dual-band behavior of Ants. b and c, current distributions at the resonant frequencies of each antenna are shown in figure 6. For Ant. b and at 3.6 GHz, there is a nearly identical current distribution to the one of the PIFA without a L-shaped parasitic element (figure 3) with weak current flowing through the parasitic element. Thus, the first resonance is basically dependent on the PIFA and is slightly affected by the L-shaped parasitic element. At 5.37 GHz, the current is strongly distributed on the L-shape parasitic element and relatively weak at the edges of the PIFA next to the L-shaped parasitic element, which may be due to the coupling between them. This indicates that the L-shaped parasitic element is responsible for the second resonant frequency. The same behavior is observed for Ant. c at 3.6 and 5.37 GHz. At 5.75 GHz, we can observe that the L-shaped slot is responsible for creating another resonance in the upper band.

Zoom In Zoom Out Reset image size

Lastly, a rectangular open-ended ring parasitic structure is printed on the top layer of the substrate next to the L-shaped parasitic element in order to enhance the impedance matching of the antenna element at the upper band. This generates a new resonance at the center frequency of the WLAN band (5.5 GHz), as illustrated in figure 2 (Ant. d). The simulated reflection coefficient of this structure and the other previous configurations is shown in the same figure. Comparing the reflection coefficient of the final configuration of the antenna element to the previous starting configurations, it is evident that the bandwidth and impedance matching at the upper band are significantly improved.

Figure 7 shows the total efficiency of Ant. a to Ant. c. The total efficiency ranges from 75% to 97 % in the lower band for all antennas. However, in the upper band, it ranges from 19% to 97% for Ant. b, 69% to 97% for Ant. c, and 58% to 94 % for Ant. d. Due to using a printed rectangular open-ended ring parasitic structure, the total efficiency of the final configuration decreases from 91% to 64% compared with the configuration of the Ant. c. The reason for this decrease is that the printed rectangular open-ended ring parasitic structure introduces more losses at its resonant frequency (5.5 GHz).

Zoom In Zoom Out Reset image size

In the final configuration of the antenna element, there is a slight change in the PIFA first resonance frequency (f₀ = 3.6 GHz) when slots and parasitic elements are added to the antenna. Thus, additional optimization would be necessary. The final dimensions of the proposed antenna element are listed in table 1 and its reflection coefficient under CST and HFSS software is displayed in figure 8. The −6 dB working bandwidth in the lower and upper bands is 510 MHz (3.39-3.9 GHz) and 710 MHz (5.14-5.85 GHz). While for the -10 dB condition, the bandwidth in the lower and upper bands is 280 MHz (3.47-3.75 GHz) and 610 MHz (5.2-5.81 GHz) in both software, respectively.

Zoom In Zoom Out Reset image size

Figure 9 illustrates the dual-band antenna element's current distributions at resonant frequencies, demonstrating the impact of all the previously discussed steps. Since the current is distributed on the strip's surface as well as on the PIFA antenna's radiation plate at 3.6 GHz, it is evident that the radiating plate of the PIFA is what causes the resonance frequency, whereas at 5.75 GHz, the current is distributed on the L-shaped slot. While the majority of the currents at 5.29 GHz are concentrated in the L-shaped parasitic element. At 5.5 GHz, most of the currents are dispersed around the rectangular open-ended ring.

Zoom In Zoom Out Reset image size

Figure 10 plots the radiation and total antenna efficiencies of the suggested antenna element. As can be observed, the CST radiation efficiency across the lower and upper bands varies roughly between 96%-98% and 61%-96%, while in HFSS it varies between 97%-99% and 60%-95%, respectively. Plus, the total efficiency in CST varies between 63.7%-96.7% in the lower band (55.5%-88.5% in HFSS) and between 58.6%-93.8% (55%-95 in HFSS) in the upper band.

Zoom In Zoom Out Reset image size

The 2D radiation patterns of the antenna element in the xz and yz planes in CST and HFSS are plotted in figure 11. The maximum radiation pattern on the xz plane at 3.6 GHz, 5.29 GHz, and 5.5 GHz occurs at angles of 315°, 311°, 6°, while at 5.74 GHz, the maximum radiation is quite uniform. On the yz plane, the maximum radiation takes place at θ = 300° at 3.6 GHz, 5.29 GHz, and 5.5 GHz, while at 5.74 GHz, the maximum radiation pattern occurs at θ = 377°.

Zoom In Zoom Out Reset image size

The peak realized gain of the proposed antenna element is shown in figure 12. In CST, the minimum peak realized gain obtained in the lower band is 4.28 dB (3.6 dB in HFSS), while in the upper band it is 5.1 dB (5 dB in HFSS). However, the maximum peak realized gain obtained in the lower and upper bands in CST is 5.71 dB and 6.1 dB (5.97 dB and 7.18 dB in HFSS), respectively.

Zoom In Zoom Out Reset image size

The simulated curves of the S-parameters for Ant 1, 2, 3, and 7 are plotted in figure 14. The 6-dB bandwidths in the lower and upper bands of Ant1, Ant2, Ant3, and Ant 7 are 440 MHz (3.41-3.85 GHz) and 700 MHz (5.14-5.86 GHz), 490 MHz (3.42-3.91 GHz) and 710 MHz (5.15-5.86 GHz), 400 MHz (3.43-3.83 GHz) and 710 MHz (5.15-5.86 GHz), and 550 MHz (3.39-3.94 GHz) and 730 MHz (5.14-5.87 GHz) in CST. Whereas in HFSS are 440 MHz (3.41-3.85 GHz) and 720 MHz (5.13-5.85 GHz), 490 MHz (3.42-3.91 GHz) and 690 MHz (5.16-5.85 GHz), 380 MHz (3.44-3.82 GHz) and 700 MHz (5.15-5.85 GHz), and 540 MHz (3.39-3.93 GHz) and 710 MHz (5.14-5.85 GHz), respectively. The minimum isolation between adjacent antenna elements in CST is 12.2 dB in the lower band and 12 dB in HFSS, while in the upper band it is 16.4 dB in CST and 15.9 dB in HFSS. Plus, the maximum isolation in the lower band is between Ant3 and Ant7: 37.5 dB in CST and 33.7 dB in HFSS, while in the upper band it is between Ant1 and Ant6: 37.3 dB in CST and 30.6 dB in HFSS.

Zoom In Zoom Out Reset image size

Further, the surface currents of all eight antenna elements at 3.6 GHz, 5.29 GHz, 5.5 GHz, and 5.75 GHz are displayed in figure 15. At 3.6 GHz, there is strong current flowing on the ground plane around and between antenna elements, which increases the mutual coupling between antenna elements. However, at the other resonant frequencies, we can observe that the surface current is distributed only around antenna elements, which justifies the high isolation value in the upper band.

Zoom In Zoom Out Reset image size

The radiation and antenna efficiencies of Ant1 to Ant3 and Ant7 are plotted in figure 16. The radiation efficiency in the lower and upper bands ranges from 95% to 97% and 57.2% to 96.2% in CST, while in HFSS it ranges from 96.9% to 98.1% and 58.4% to 95.5%, respectively. In addition, the antenna efficiency in CST of all antennas is between 61.1% and 91.2% in the lower band (65.7% and 97.4% in HFSS), while in the upper band it ranges from 53% and 95.4% (52.9% and 95.1% in HFSS).

Zoom In Zoom Out Reset image size

The 2D radiation patterns in the xz and yz planes at the center frequency of the 3.6 GHz 5G band and 5.5 GHz WLAN band for Ant1 to Ant3 and Ant 7 on both software are depicted in figures 17 and 18. We can see that each antenna has a different radiation pattern, which lowers the ECC.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Peak realized gains of Ant1 to 3 and Ant7 are plotted in figure 19. The minimum realized gain for Ant1 to 3 and Ant7 in the lower band is 4.16 dB, 4.6 dB, 4.87 dB, and 4.37 dB in CST, while in HFSS it is 4.33 dB, 4.66 dB, 4.86 dB, and 4.37 dB. Whereas in the upper band it is 3.7 dB, 4 dB, 4.3 dB and 4.2 dB in CST, and 4.7 dB, 4.5 dB, 5.4 dB and 4.5 dB in HFSS, respectively. The maximum realized gain in the lower band is 5.5 dB, 7.12 dB, 6.15 dB, and 5.1 dB for Ant1, Ant2, Ant3, and Ant7 in CST, and it is 5.51 dB, 7.21 dB, 5.68 dB, and 5.6 dB in HFSS. While in the upper band it is 6.7 dB, 5.9 dB, 5.8 dB and 6.3 dB in CST, and 6.8 dB, 6.3 dB, 5.8 dB and 6.4 dB in HFSS, respectively.

Zoom In Zoom Out Reset image size

The envelope correlation coefficient (ECC) is an important parameter to evaluate diversity performance. It measures how well the received signals from the transmitting antennas are independent. The ECC of Ant1 to Ant3 and Ant7 is calculated using the formula in [21] using the radiation pattern approach and are displayed in figure 20. As shown, the ECC in CST is below 0.05 and 0.03 in the lower and upper bands, while in HFSS it is below 0.18 and 0.04, respectively, satisfying the ECC requirements for MIMO antenna systems, which state that they must be smaller than 0.5.

Zoom In Zoom Out Reset image size

Diversity combats multipath fading by transmitting multiple copies of the same signals to the receiver via different channel paths. Diversity gain measures the improvement in the communication system due to the use of diversity and can be calculated using the following formula (2):

$$\begin{eqnarray}&&{DG}=10\sqrt{1-{\left({ECC}\right)}^{2}}\end{eqnarray} \tag{ 2 }$$

From figure 21, it can be seen that the value of DG in the lower band is <9.9 dB in CST, and in HFSS it is <9.5 dB. Whereas in the upper band, it is <9.9 dB in both software.

Zoom In Zoom Out Reset image size

The mean effective gain (MEG) of an antenna is defined as the ratio of the received power in a fading environment to the power that would be received by an isotropic antenna and can be calculated using the formula below:

$$\begin{eqnarray}&&{{MEG}}_{i}=0.5[\ 1-\displaystyle \sum _{j=1}^{M}| {S}_{{ij}}{| }^{2}]\ \end{eqnarray} \tag{ 3 }$$

iwhere M is the total number of antenna elements. In order to achieve optimal power balance and diversity performance, the absolute value of the ratio ∣MEG_i ∣/∣MEG_j ∣ should be below 3, where i and j are the number of antenna elements.

The calculated MEG of Ant1 to Ant3 and Ant7 is presented in table 2. The MEGs of all antennas are below -4 dBi, and it can be easily seen that the ratio of MEGs is below 3.

The multiplexing efficiency (ME) η_Mux is the ratio of the signal-to-noise ratio of an ideal MIMO antenna to the signal-to-noise ratio of a given MIMO antenna and is expressed as [22]:

$$\begin{eqnarray}&&{\eta }_{{{Mux}}_{{i}{j}}}=\sqrt{(1-| {\rho }_{c}{| }^{2}){\eta }_{i}{\eta }_{j})}\end{eqnarray} \tag{ 4 }$$

where ρ_c is the complex correlation coefficient between the antenna elements (∣ρ_c ∣² ≈ ECC). η_i and η_j are the total efficiency of the ith and jth antenna elements. Figure 22 plots the MEs of Ant1 to Ant3 and Ant7. It can be observed that the value of the ME of all antennas is above -3 dB.

Zoom In Zoom Out Reset image size

The ratio of the square root of the total reflected power to the square root of the total incident power is known as the total active reflection coefficient (TARC) and can be computed using the following expression [23]:

$$\begin{eqnarray}&&{TARC}=\sqrt{(| {S}_{{mm}}+{S}_{{mn}}{| }^{2})+(| {S}_{{nm}}+{S}_{{nn}}{| }^{2})}/\sqrt{2}\end{eqnarray} \tag{ 5 }$$

The channel's capacity loss resulting from the MIMO links' correlation is called channel capacity loss (CCL) and is given by [24]:

$$\begin{eqnarray}&&{CCL}=-{{log}}_{10}{\det }({a}_{k})\end{eqnarray} \tag{ 6 }$$

where a_k is the correlation matrix and its element can be calculated using the expressions below:

$$\begin{eqnarray}&&{a}_{{ii}}=1-[| {S}_{{ii}}{| }^{2}+| {S}_{{ij}}{| }^{2}]\end{eqnarray} \tag{ 7 }$$

$$\begin{eqnarray}&&{a}_{{ij}}=-[{S}_{{ii}}^{* }{S}_{{ij}}+{S}_{{ji}}^{* }{S}_{{ij}}]\end{eqnarray} \tag{ 8 }$$

Both TARC and CCL are displayed in figures 23 and 24. From this figure, we can observe that the bandwidth of the adjacent antennas in the lower band is reduced and the CCL is more than 0.4 dB due to the coupling between them compared with the upper band, and that because in the upper band the current distribution is mainly concentrated in the vicinity of the L-shaped parasitic element, which increases the isolation.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size