1 Introduction

The fifth generation (5G) technology is the next step in mobile communication [1]. It enables immersive multimedia, audio, streaming video, and the internet. 5G technology is more advanced and enticing [2]. When the new 5G network was deployed, it brought new features such as faster mobile broadband and faster download and upload speeds for IoT. These new functionalities allow businesses to remotely control devices considerably more quickly and effectively than before [3]. Furthermore, Millimeter-wave offers viable solutions to all 5G needs [4]. Because the available bandwidth is substantially bigger than the occupied frequency of the radio spectrum, it will deliver much higher data rates and wider channels to many users [5]. The studies clarify that the optimum solution for 5G cellular communications is found to be in the mm-wave range between 28 and 38GHz [6].

In wireless communication systems, the antenna functions as a transceiver. So, the antenna should be small and lightweight, with optimal working bands and strong radiation performance [7]. However, the multipath problem is one of the most essential issues for any communication system. So, most research is towards to MIMO antenna. Any MIMO system should be able to increase the data rate without increasing power [8]. Another crucial parameter of any MIMO system is its size, which should be as small as feasible [9]. As a result, the radiating pieces should be placed as near together as feasible, taking into account the isolation factor [10]. Several mm-wave microstrip antenna topologies have been built, simulated, and analyzed for 5G applications in the 28 GHz and 38 GHz bands, or individually [11,12,13,14,15,16,17,18,19,20]. Ref [11] described a microstrip patch antenna for 28 GHz in millimeter wave. It features an impedance bandwidth of 4.6% and a high gain of 13.5 dB. This is accomplished by creating four four-element patch antenna arrays. Ref [12] investigated a 4 × 4 dual-band MIMO antenna for millimeter wave bands operating at 28 GHz and 38 GHz obtained a wide band of 14.3% and 5.26%, respectively and had an EBG construction that increased mutual coupling. Ref [13] presented a multi-band MIMO antenna constructed with a user-impact analysis for 4G and 5G mobile terminals. Ref [14] proposed antenna worked at a frequency of 5.3 GHz, which is suitable for future 5G services. Ref [15] presented a four-port T-shaped MIMO antenna that has a bandwidth from 22.43 to 31.66 GHz and is designed with a total dimension of 24 × 24 mm2. Ref [16] introduced an eight-element MIMO antenna for smart 5G devices that operates in the frequency spectrum 24.25 to 27.5 GHz. It has achieved isolation of 15 dB with a gain of 6.4 dBi. In [17], a six-element for 28/38 GHz with high gain is presented. It features impedance bandwidths of 27.7 to 28.1 GHz and 36.9 to 39.5 GHz, with realized gains of 13 and 10.01 dBi, respectively. Ref [18] proposed a quad-port dual-band MIMO antenna array for 5G communications that operates at 28 and 38 GHz. This setup provides a high gain of 7.9 and 13.7 dB at 28 and 37.3 GHz, respectively. Ref [19] introduced a 28/38 GHz patch antenna with an H-shaped slot for the MIMO applications. The antenna has an impedance bandwidth of 27.6—28.5 GHz and 36.9—38.9 GHz, with achieved gains of 9.0 dBi and 5.9 dBi. An inset-fed planar antenna array was operated at 28.44 /39.49 GHz with isolation ≥ 25 dB and gain values of 5.59 dBi and 5.70 dBi, as given in [20]. This study investigates the design and fabrication of a compact quad-port 28/38 GHz MIMO antenna. The cross-shape arrangement of parts is used to improve the port isolation while also achieving the requisite compactness. The distance between the radiating parts is short, so a parasitic strip is added to improve the port isolation. The MIMO parameters were computed and tested with a reasonable trend between the two outcomes.

The novelty of this design can be summarized as: (1) this design has a compact size (29 mm × 24.2 mm × 0.203 mm) which saves more space when increasing the number of ports. (2) It achieved a reasonable gain of 8.1/7.22 dB with the isolation ≥ 20 dB between ports at 28/38 GHz. (3) It achieved good levels of MIMO parameter values (The ECC of the structure is < 0.001, the DG ≥ 9.95 dB. and the CCL was ≤ 0.2 bit/s/Hz) which confirms the suitability of the antenna for 5G dual-band communication.

2 Antenna Structure

Figure 1 illustrates the single antenna layout. It consists of a rectangular patch with three triangle slots in its top edge to enhance the impedance matching. This design is etched on a Roger RO4003 with εr = 3.55, tan δ = 0.0027, and thickness = 0.203mm. The single antenna is investigated by the CST simulator to optimize the dimension. Table 1 presents the optimum dimensions. The antenna has been fabricated, and the model is illustrated in Fig. 2. The experimental data is tested by utilizing the ZVA67 VNA. The antenna operated at 28.2 GHz with S11 < -10 dB from 27.86 GHz to 28.77, and at 37.45 GHz with S11 < -10 dB from 37.07 GHz to 37.76 from the simulation results. It also worked at 28.1 GHz from 27.44 GHz to 28.83 and 37.81 GHz from 37.63 GHz to 38.08 GHz from the tested results.

Fig. 1
figure 1

The single antenna geometry

Table 1 The single element dimensions (Unit: mm)
Fig. 2
figure 2

The S11 of the single antenna

The best results are obtained by developing the antenna structure as shown in Fig. 3. Antenna #1 is a conventional patch antenna with a full ground to work at the fundamental mode at 28.5 GHz as shown in the dotted red line in Fig. 3. To control the second band, a single triangle slot is etched on the top of the patch as shown in Fig. 3 (Antenna #2) to make the antenna to be operated at 28.1GHz and reduced a little bit the second band as illustrated in the dashed blue line in Fig. 3. Finally to tune the second band to 37.5 GHz, three slots on the top of the patch are developed (Antenna #3) as shown in the black solid line in Fig. 3. Figure 4 illustrates the single antenna's current distribution at two frequencies 28 /37.5 GHz. The current is collected around the patch at 28 GHz and around the triangular slots at 37.5 GHz which illustrates the contribution of each part in the radiation.

Fig. 3
figure 3

The developments of the single element

Fig. 4
figure 4

The current distribution of the single element

The radiation characteristics in different planes ( x–z and y–z) for the structure at 28/37.5GHz are shown in Fig. 5. The antenna has a broadside pattern at ϴ = 0° because of the presence of the full ground plane. As well, it is observed that the difference between Co / Cross polarization ≥ 20 dB in both planes confirms its diversity.

Fig. 5
figure 5

The radiation patterns of the antenna: (a) @28GHz, and (b) @37.5GHz

To investigate the antenna performance in terms of reflection coefficient, parametric studies at different antenna dimensions are conducted. The effect of the patch width (Wp) on the reflection coefficient is illustrated in Fig. 6(a). it is seen that by increasing the Wp, the two resonance frequencies are affected. As well, the effect of the slot width (T) on the reflection coefficient is illustrated in Fig. 6(b). it is clear that by increasing the T, a small shift is observed at the 28 GHz band while the 38 GHz band is highly affected.

Fig. 6
figure 6

The parametric study at different antenna dimensions (a) changing Wp, (b) changing T

The effect of the changing x and y on the reflection coefficient is illustrated in Fig. 7. it is seen that by increasing the x and y, the two resonance frequencies are affected as illustrated in Fig. 7. It is seen that the antenna should designed carefully to achieve the desired two 28/38 GHz bands.

Fig. 7
figure 7

The parametric study at different antenna dimensions (a) changing x (b) changing y

3 Quad-Port Mimo Configuration

The quad MIMO antenna is examined in this section. Figure 8 shows the development of the recommended MIMO antenna. The antenna elements in the three stages are aligned face-to-face and orthogonally to the other elements. MIMO#1 is a quad-MIMO antenna with a full substrate and ground. It operates at 29 GHz and 38.5 GHz, as depicted by the dashed red line in Fig. 9. The substrate and ground plane are then sliced along the four edges in MIMO#2. It is organized in a crisscross pattern to make the design more compact. It runs at 28.4 / 38 GHz, as shown in the dashed blue line in Fig. 9. Finally, MIMO#3 is developed by adding a parasitic strip between the elements to enhance its isolation. It works at 28.1 GHz and 37.8 GHz as shown in the black solid line in Fig. 9. The differences between the resonance frequencies and the single antenna in Fig. 2 are due to the coupling effects between ports. The isolation between ports in the three stages is shown in Fig. 10. The isolation is ≥ 20 dB in the three stages while we choose the third one which achieves the desired frequency 28/38 GHz bands.

Fig. 8
figure 8

The development geometries of the suggested MIMO antenna

Fig. 9
figure 9

The suggested MIMO developments

Fig. 10
figure 10

The suggested MIMO developments: (a) S21, (b) S31, and (c) S41

Figure 11 shows the current distributions of the MIMO antenna at 28 /38 GHz. It's observed that the isolation is achieved by focusing the current around the ports with a small amount transferred to other ports and the most current is concentrated around the patch and the triangular slots.

Fig. 11
figure 11

The proposed crossed MIMO antenna's current distribution in the different four ports: (a) @ 28 GHz, and (b) @38 GHz

4 Results and Discussion

As shown in Fig. 12, the antenna area is reduced to 24.2 × 29 mm2 by arranging the four elements adjacent to each other. The radiators are separated by 1.2 mm. The antenna has been fabricated and its model is illustrated in Fig. 12(b) and its measuring setup is shown in Fig. 12(c). The antenna is operated at 28.1 GHz with S11 < -10 dB beginning from 27.75 GHz to 28.65 and at 37.8 GHz with S11 < -10 dB beginning from 37.1 GHz to 38.24 from the simulation outcomes, and at 28.1 GHz with S11 < -10 dB beginning from 27.5 GHz to 29.3 and at 37.81 GHz with S11 < -10 dB beginning from 37.5 GHz to 38.2 GHz from the tested outcomes as shown in Fig. 13. Additionally, the isolation at port 1 is ≥ 20 dB at 28/38 GHz, confirming the strong isolation between ports seen in Fig. 14. It is observed that there is a good match between the two outcomes. However, a slight shift is introduced due to the fabrication and testing arrangement, as well as human error in the fabrication process.

Fig. 12
figure 12

MIMO antenna configurations: (a) 2D view (b) Fabricated prototype (c) Measuring Setup with VNA

Fig. 13
figure 13

The return loss outcomes of the proposed crossed MIMO antenna

Fig. 14
figure 14

The proposed crossed MIMO antenna: (a) S21, (b) S31, and (c) S41

Figure 15 depicts the experimental setup for testing the MIMO antenna radiation patterns and gain. The horn reference antenna is meant to operate between 26 and 40 GHz and serves as a transmitter antenna, while the MIMO antenna under test is placed at 50 cm as a receiving antenna. The MIMO antenna rotates in both the azimuth and elevation planes. The reference antenna is connected to VNA port 1 and the MIMO antenna is attached to port 2 of the VNA, and the antenna gain is estimated from the S21 as [21, 22]. The tested normalized radiation patterns in comparison with the simulated one at 28 GHz and 38 GHz in both two x–z and y–z planes are illustrated in Fig. 16. The antenna has a broadside pattern at ϴ = 0° because of the presence of the full ground plane with a good matching between the two outcomes. The simulated gain of the antenna at 28 GHz and 38 GHz is 8.5 dBi and 7.8 dBi while the testes gain at the same frequency bands is 8.1dBi and 7.22 dBi.

Fig. 15
figure 15

The MIMO antenna under testing inside the anechoic chamber

Fig. 16
figure 16

The radiation pattern outcomes for the Proposed crossed MIMO antenna at port#1 in both (x–z) and (y–z) planes: (a) @28 GHz, and (b) @38 GHz

Finally, the MIMO performance is investigated by calculation of the ECC, DG, CCL, Mean Effective Gain (MEG), and the Total Active Reflection Coefficient MIMO parameters. There are several equations utilized to extract these parameters [23,24,25,26,27]. These parameters are calculated at port 1. The ECC between ports is calculated using S-parameters and radiation patterns as Eqs. (1), (2) [26, 27] and its outcomes are displayed in Fig. 17.

Fig. 17
figure 17

The port 1MIMO antenna's ECC

$$ECC={\rho }_{e}(i,j,N)=\frac{{\left|\sum\limits_{n=1}^{N}{S}_{i,n}^{*}{S}_{n,j}\right|}^{2}}{\prod\limits_{k=i,j}\left[1-\sum\limits_{n=1}^{N}{S}_{k,n}^{*}{S}_{n,k}\right]}$$
(1)
$$ECC={\rho }_{e}=\frac{{\left|{\int \int }_{4\Pi }\left[{F}_{1}\left(\theta ,\varphi \right)\bullet {F}_{2}\left(\theta ,\varphi \right)d\Omega \right]\right|}^{2}}{{\int \int }_{4\Pi }{\left|{F}_{1}\left(\theta ,\varphi \right)\right|}^{2}d\Omega {\int \int }_{4\Pi }{\left|{F}_{2}\left(\theta ,\varphi \right)\right|}^{2}d\Omega }$$
(2)

The value of the ECC is ≤ 0.001 within the desired dualbands. The DG is calculated and related to ECC as Eq. (3) [26] and its outcome is illustrated in Fig. 18.

Fig. 18
figure 18

The port 1 MIMO antenna's DG

$$\text{DG}=10\times \sqrt{1-{|\text{ECC}|}^{2}}$$
(3)

The DG had values ≥ 9.95 dB within the desired dualbands.

Also, the CCL is calculated and extracted using Eqs. (4), (5) [27] and its outcome is shown in Fig. 19.

Fig. 19
figure 19

The port 1 MIMO antenna's CCL

$$C(Loss)=-{\text{log}}_{2}\text{det}({\psi }^{R})$$
(4)
$$\begin{array}{c}{\psi }^{R}=\left[\begin{array}{cccc}{\rho }_{11}& {\rho }_{12}& {\rho }_{13}& {\rho }_{14}\\ {\rho }_{21}& {\rho }_{22}& {\rho }_{23}& {\rho }_{24}\\ {\rho }_{31}& {\rho }_{32}& {\rho }_{33}& {\rho }_{34}\\ {\rho }_{41}& {\rho }_{42}& {\rho }_{43}& {\rho }_{44}\end{array}\right],{\rho }_{ii}=1-\sum\limits_{n=1}^{4}{\left|{S}_{in}\right|}^{2}\\ and\\ {\rho }_{ij}=-\left|\sum\limits_{n=1}^{4}{S}_{in}^{*}{S}_{nj}\right|,for\begin{array}{cc}& \end{array}i,j=\text{1,2},3or4\end{array}$$
(5)

The CCL achieved a value ≤ 0.2 bit/s/Hz within the desired dual bands.

Further, the MEG and the TARC are calculated and extracted using equations in [26, 27]. The TARC has achieved -20dB and -18 dB at both 28/38 GHz with around the same results when the phases are changed as shown in Fig. 20 while the MEG achieved -3dB at 28/38 GHz frequency bands.

Fig. 20
figure 20

The port 1 MIMO antenna's TRAC

The tested outcomes are compared to the simulated ones with a good trend between them. All of these parameters are lower than the acceptable level which confirms the system isolation and diversity.

Table 2 compares the literature review designs [28,29,30,31,32,33,34] to the proposed antenna to assess the suggested antenna performance. It is noted that the proposed compact MIMO antenna has supported the 5G applications with a smaller size with a greater level of isolation, gain, BW, and MIMO characteristics.

Table 2 The comparison between the previous mm-wave antenna and the proposed antenna

5 Conclusion

This study designs and fabricates a compact quad-port cross-shaped antenna with improved port isolation. The single antenna is made up of a rectangular patch radiator with three triangle slots along the top edge for good matching. The proposed MIMO antenna model was tested to validate the simulation results. The reported MIMO antenna operates in two bands at 28/38 GHz with isolation ≥ 20 dB. The results of the simulation and experiment are consistent. The MIMO antenna radiation patterns and gain were validated by testing the E and H planes within the anechoic chamber. A reasonable gain of 8.1 dBi and 7.22 dBi at 28 /38 GHz, respectively have been introduced. Also, the antenna has achieved a good MIMO performance between ports. The ECC, DG, CCL, MEG, and TARC have achieved values ≤ 0.001 ≥ 9.95 dB, ≤ 0.2 bit/s/Hz, -3 dB, and -20/-18 dB over the two working bands. All these features would support the reported work to be more convenient in the 5G network applications.