A Miniaturized Eight-Port MIMO Antenna for 5G Ultra-Slim Smartphones

Abstract

This paper presents the design of a miniaturized eight-port multiple-input multiple-output (MIMO) antenna. Each proposed antenna element occupies only 6.2 × 7 × 0.8 mm³, making it ideal for integration into ultra-slim 5G smartphones. The proposed MIMO antenna offers a −10 dB impedance bandwidth of 1.7 GHz (3.3 to 5.0 GHz), covering the N77 (3.3–4.2 GHz), N78 (3.3–3.8 GHz), and N79 (4.8–5.0 GHz) bands. The isolation between the antenna elements is larger than 16 dB. The envelope correlation coefficient (ECC) of the designed antenna is lower than 0.09. Meanwhile, the peak gain exceeds 2.2 dBi, the efficiency surpasses 70%, and the diversity gain (DG) is higher than 9.95. This innovative design addresses the demands of ultra-slim mobile devices while maintaining high performance.

1. Introduction

Recently, the evolution of 5G communication technologies has brought about a significant transformation in wireless communications. The World Radio Communication Conference 2015 (WRC-15) allocated critical frequency bands spanning 3.3–3.4 GHz, 3.6–3.7 GHz, and 4.8–5 GHz. Subsequently, China approved the frequency bands of 3.3–3.6 GHz and 4.8–5 GHz for 5G communication [1]. Although there is growing anticipation for 6G communications, 5G continues to be a cornerstone in the field of wireless communication. It is confronted with various challenges associated with both sub-6 GHz and millimeter-wave bands [2]. For instance, in the sub-6 GHz band, issues like limited bandwidth and interference need to be addressed. To tackle these challenges and improve the performance of 5G networks, multi-input multi-output (MIMO) antenna technology has emerged as a key solution [3]. The utilization of multiple spatial dimensions within MIMO antennas allows for enhanced signaling and spectral efficiency, making them essential for the high-speed, low-latency demands of 5G communication [4,5].

The design of MIMO antennas in confined spaces, especially within the constraints of mobile devices, is a topic of intense discussion. Achieving high performance in such environments is a critical challenge for 5G antenna technology. To accommodate the 5G frequency bands without enlarging the device, antennas are often integrated on metal frames [6]. Research on 5G MIMO antennas primarily focuses on the 3.3 GHz to 5 GHz frequency range [7,8,9]. In [10], an MIMO antenna with a co-coupled fed IFA (inverted-F antenna) loop structure was proposed, covering the frequency bands of 3.3 GHz to 6.4 GHz. An MIMO antenna comprising an inverted L-shaped feed branch was presented in [11], operating within 3.3–3.6 GHz and 4.8–5 GHz. Furthermore, reference [12] proposed an MIMO antenna system composed of a dual-arm tortuous monopole radiating element spanning the 3.2 to 6 GHz bands. An element slotted octagon-shaped antenna was proposed in [13], which enabled the antenna to operate in the sub-6 GHz band (3.1–4.5 GHz).

To address the design challenges associated with maintaining adequate isolation in 5G MIMO antennas, various decoupling methods have been proposed [10,14,15,16]. These methods aim to enhance isolation and reduce mutual interference within the limited space of mobile terminals, particularly in the sub-6 GHz bands. Among these, the ground decoupling method has been widely applied to 5G MIMO antenna systems. Specifically, the use of Defected Ground Structure (DGS) has been shown to be effective in modifying the ground plane of the antenna to create specific defects that disrupt surface currents, thereby reducing coupling between adjacent antennas [17]. A variant of DGS, inserted L-shaped and rectangle slots in the ground plane, has been presented to contribute significantly to the reduction in coupling [18]. The offset-fed configuration is another approach that achieves high isolation between antenna elements by concentrating the surface current and near fields of each element on their respective sides, eliminating the need for additional isolation techniques [19]. This design is particularly effective for adjacent antennas with irregular structures that are completely symmetrical, enhancing isolation without the need for additional isolation designs [20]. However, the effectiveness of this approach is mainly dependent on the geometric shape of the antenna itself, with limited effectiveness for antenna designs with relatively uniform current distributions. Self-isolated elements, designed to minimize mutual coupling through specific network configurations, have also been proposed for decoupling [21]. Additionally, the use of extra parasitic patches as a decoupling method has been employed to enhance isolation [22]. These parasitic elements act as supplementary radiating sources, generating extra resonances that influence the current distribution and, consequently, the isolation between antennas. The introduction of parasitic elements has been shown to result in higher isolation between the excited antenna and those not excited [23]. Electromagnetic Bandgap (EBG) structures have emerged as a research hotspot for decoupling in 5G MIMO antennas. The cascading EBG structures have been introduced to minimize mutual coupling among antenna elements and enhance the performance of MIMO antennas [24]. A fractal H-shaped metamaterial EBG isolator has been presented, which increases isolation by over 10 dB [25]. Pattern diversity techniques have also been employed to reduce mutual coupling. These techniques enable the antenna to function with orthogonal modes, namely, in-phase and out-of-phase modes, which effectively minimize the strong coupling between antenna elements [26]. The excitation of one port induces an in-phase mode, whereas the other port’s excitation produces an out-of-phase mode, thereby achieving high isolation when the ports are activated individually and resulting in in-phase and out-of-phase current distributions on the two monopoles [27]. In summary, the various decoupling techniques for 5G MIMO antennas, including ground decoupling methods, offset-fed configurations, self-isolated elements, parasitic elements, EBG structures, and pattern diversity techniques, play a crucial role in enhancing isolation and reducing mutual interference, addressing the design challenges posed by the need for multiple antenna elements in limited spaces.

This paper presents a compact eight-element MIMO antenna designed for 5G smartphones. The antenna element’s −10 dB impedance bandwidth encompasses the N77 (3.3–4.2 GHz), N78 (3.3–3.8 GHz), and N79 (4.8–5.0 GHz) bands. Several decoupling structures have been proposed to ensure high isolation between the antenna elements. Ansys Electronics Desktop 2024 R1 is used as the electromagnetic simulator. The performance of the proposed eight-element MIMO antenna is validated through both simulation and measurement.

2. Antenna Geometry and Design

The geometry of the proposed eight-element MIMO antenna is depicted in Figure 1. The proposed MIMO antenna is printed on a main board, two long sideboards, and two short sideboards. The main board has a dimension of 150 mm × 75 mm × 0.8 mm. The short sideboard and the long sideboard measure 75 mm × 0.8 mm × 7 mm and 150 mm × 7 mm × 0.8 mm, respectively. The proposed eight-element MIMO antenna has an overall size of 150 mm × 75 mm × 7 mm. The radiators of eight antenna elements are printed on the sideboards, which are positioned perpendicularly to the main board and are fed by a microstrip line measuring 8 mm × 1.5 mm. The four side boards are bonded to the main board as well as to each other using metallic adhesive. Both the sideboards and the main board are made from an FR4 substrate with a dielectric constant (ε_r) of 4.4 and a loss tangent (tanδ) of 0.02.

2.1. Antenna Element Design

Figure 2a shows the radiator of type 1 antenna elements (including Ant. 2, Ant. 3, Ant. 6, and Ant. 7) which are printed on the short sideboards. Figure 2b shows the radiator of type 2 antenna elements (including Ant.1, Ant.4, Ant.5, and Ant.8) which are printed on the long sideboards. Figure 2c illustrates the defected ground plane of both type 1 and type 2 antenna elements. The overall size of two types of antenna elements measures 6.8 mm × 7 mm × 0.8 mm.

Table 1. Dimensions of the proposed antenna elements (unit: mm).

Parameter	Value	Parameter	Value	Parameter	Value
W1	6.2	W8	6.2	L4	1.5
W2	1.0	W9	0.6	L5	0.7
W3	1.5	W10	0.7	L6	1.0
W4	1.0	W11	7.0	L7	0.3
W5	1.5	L1	1.0	L8	1.0
W6	2.0	L2	0.8	L9	4.0
W7	2.2	L3	1.7

shows the optimized geometric parameters of the proposed antenna elements.

2.2. Evolution of the Radiator of Antenna Elements

To enhance the understanding of the design evolution for antenna elements, four kinds of radiators have been developed (Stage 1 to Stage 4), as depicted in Figure 3a. Stage 3 is the radiator of the proposed type 1 antenna element, while stage 4 is the radiator of the proposed type 2 antenna element. The simulated reflection coefficients for each stage are shown in Figure 3b.

2.3. Decoupling Structure

Numerous decoupling techniques have been explored to enhance the isolation between adjacent antenna elements. As shown in Figure 4, the proposed decoupling structures (DS) DS1 and DS2 are decoupling slots etched on the mainboard’s top surface, while DS3 and DS4 are etched on the outer surfaces of the short and long sideboards, respectively. DS1 is introduced to improve the isolation between adjacent type 1 and type 2 antenna elements. DS2 is employed to mitigate coupling between adjacent type 1 antenna elements and also between adjacent type 2 antenna elements. DS3 is used to reduce coupling both between Ant. 1 and Ant. 8, as well as between Ant. 4 and Ant. 5. DS4 is applied to mitigate coupling between Ant. 2 and Ant. 3, as well as between Ant. 6 and Ant. 7.

To validate the operational principle of the decoupling structures, the surface current distributions with and without the decoupling structures are compared, as shown in Figure 5 and Figure 6. This indicates that these structures play a certain role in influencing the current distribution in contrast to the absence of them. As can be seen from Figure 5, when Ant. 1 is excited at 3.5 GHz, the current density on the outer surfaces of the sideboard is concentrated on the right side of DS3 (the side closer to Ant.1), which significantly reduces the coupling between Ant. 1 and Ant. 8. On the ground plane, the current density is distributed over the inner edges of the slots of DS1 and DS2. This decreases the coupling between Ant. 1 and Ant. 2 and further enhances the decoupling of Ant. 1 and Ant. 8. When Ant. 1 is excited at 4.8 GHz, as shown in Figure 6, the effect of DS3 weakens, and DS1 and DS2 play a crucial role in providing decoupling in this frequency band.

As shown in Figure 7a, the simulated transmission coefficients without decoupling structures do not meet the requirements in the desired frequency bands, especially the transmission coefficients of S₁₃, S₂₃, and S₄₅. Figure 7b reveals that the introduction of the decoupling slots significantly enhances the isolation. Compared with the results in Figure 7, the isolation significantly improves across multiple ports: S₁₃ increases from 5 dB to 30 dB, S₂₃ from 13 dB to 30 dB, and S₄₅ from 5 dB to 15 dB.

3. Antenna Analysis

This section focuses on two main aspects, including the impact of the lengths of rectangular microstrips and the influence of a hand phantom on the antenna’s performance. In the Single Hand Mode (SHM), the reflection coefficients and transmission coefficients are presented. Moreover, the electric field distribution and 3D radiation patterns are introduced in the SHM.

3.1. Parameter Analysis

The impact of the lengths of the rectangular microstrips, denoted as W₃ and W₁, on the reflection coefficients has been analyzed to achieve optimal performance. Figure 8a presents the reflection coefficients as a function of W₁. It is observed that an increase in the W₁ value results in negligible effects on the lower frequency band, while it induces a shift in the higher frequency band. This observation is pivotal for fine-tuning the antenna’s impedance matching within the targeted frequency bands. Figure 8b depicts the influence of W₃ on the reflection coefficients, demonstrating that a reduction in W₃ leads to minimal alterations in the higher frequency range, while the lower frequency range experiences an upward shift towards higher frequencies. Specifically, setting W₃ to 1.5 mm ensures that the −10 dB impedance bandwidths for both the higher and lower frequency bands align with the specified requirements. This optimization is crucial for ensuring the antenna’s operational efficiency across the desired frequency bands.

3.2. Assessment of Simulated Operational Scenarios

In this section, the influence of a hand phantom on the antenna’s performance is investigated. Figure 9 illustrates a typical application scenario, specifically the Single Hand Mode (SHM), and presents the corresponding simulated S-parameters and radiation patterns. The reflection coefficients of these antenna elements exhibit a significant degradation across two frequency bands, as illustrated in Figure 10a. Notably, the transmission coefficients exhibit minor fluctuations in the SHM, as observed in Figure 10b.

The three-dimensional (3D) radiation patterns and electric field distributions under the SHM are depicted in Figure 11 and Figure 12, respectively. Under SHM, the influence of the human hand on the antenna’s radiation such as absorption or reflection may affect the reflection and transmission coefficients. However, it is also important to note that this effect is considered a typical scenario in smartphone communication. In the design of smartphone antennas, understanding the electric field distribution is crucial for evaluating the antenna’s radiation properties and operational efficiency. An optimal distribution of the electric field can mitigate the radiation impact on the human body while also enhancing signal transmission and reception. An ideal antenna design aims to achieve a uniform electric field distribution, concentrated in the desired direction, thereby improving communication quality and minimizing energy dissipation. By employing meticulous simulation and testing, engineers can optimize the structure and positioning of the antenna, refining the electric field distribution patterns to accommodate diverse usage scenarios and performance specifications. While complete mitigation may not be feasible, the goal is to minimize the impact and maintain reliable communication.

In Figure 13, the radiation patterns of antenna type 1 and type 2 at 3.5 GHz and 4.8 GHz are observed to assess the influence of the hand. Under the SHM, there is degradation in the radiation patterns of the antenna elements. Specifically, for both the type 1 and type 2 antenna elements, the radiation patterns deviate from their ideal patterns with the hand’s presence. This degradation may be attributed to the hand acting as a scattering and absorbing medium. It disrupts the normal propagation of electromagnetic waves emitted by the antenna elements, thus affecting the radiation performance and resulting in the observed changes in the radiation patterns.

4. Experiment Results and Discussions

Figure 14 shows the photographs of the fabricated antenna prototype along with the testing equipment, including a Keysight Vector Network Analyzer (VNA) N5224A (Keysight Technologies, Santa Rosa, CA, USA) and an anechoic chamber.

Figure 15 shows the simulated and measured reflection coefficients and transmission coefficients. The measured −10 dB impedance bandwidth is 1.7 GHz, ranging from 3.3 to 5 GHz. The minimum observed isolation is between Ant. 1 and Ant. 2, at 16 dB. Figure 16 displays the acquired gain and radiating efficiency for Ant. 1 and Ant. 2. Within the operational bands N77 and N78, the proposed MIMO antenna achieves a peak gain and radiation efficiency of 4.66 dBi and 94.99%, respectively. In the N79 band, the peak gain and radiation efficiency are 5.24 dBi and 86.20%, respectively. The differences between the simulated and measured results are mainly due to manufacturing tolerances and measurement error.

Figure 17 presents the simulated and measured normalized 2D radiation patterns of the type 1 antenna element at 3.5 GHz and 4.8 GHz, while Figure 18 represents those of the type 2 antenna element. The radiation patterns of type 1 and type 2 antenna elements are nearly quasi-omnidirectional in the xoy plane and omnidirectional in the yoz plane. The disparities between the simulated and measured outcomes can be primarily attributed to antenna manufacturing tolerances, the influence of the measurement environment, and the inherent limitations in the accuracy of measurement equipment.

To assess the performance of the proposed design, the envelope correlation coefficients (ECCs) and diversity gains (DGs) were computed as key parameters for evaluating MIMO antenna capability. ECC was calculated using Equation (1) to precisely determine the level of independence between pairs of antenna elements [28]. Here, $\stackrel{\rightharpoonup }{F_1}\left(\theta ,\varphi \right)$ and $\stackrel{\rightharpoonup }{F_2}\left(\theta ,\varphi \right)$ represent the three-dimensional far-field radiation pattern, and $\Omega$ stands for the solid angle. As depicted in Figure 19, all the measured ECC values are less than 0.09 within the operating frequency bands. Figure 19 also illustrates the calculated DGs using formula (2) [29], achieving a 10 dB diversity gain within the targeted operational band, substantiating the proposed antenna’s suitability for integration into smartphones.

$$ECC={\displaystyle \frac{\left|{{\displaystyle \iint }}_{4\pi }\left[\stackrel{\rightharpoonup }{F_1}\left(\theta ,\varphi \right)\ast \stackrel{\rightharpoonup }{F_2}\left(\theta ,\varphi \right)\right]d\Omega \right|}{{{\displaystyle \iint }}_{4\pi }{\left|\stackrel{\rightharpoonup }{F_1}\left(\theta ,\varphi \right)\right|}^{2}d\Omega {{\displaystyle \iint }}_{4\pi }{\left|\stackrel{\rightharpoonup }{F_2}\left(\theta ,\varphi \right)\right|}^{2}d\Omega }}$$

(1)

$$DG=10\times \sqrt{1-EC{C}^2}$$

(2)

Table 2. Comparison of the Designed MIMO Antenna with Reported MIMO Antennas.

Ref	Operating Bands (GHz)	Total Size (mm2)	Dimension of Single Antenna Element (mm3)	Isolation (dB)	ECC	Max Efficiency
[28]	3.445–4.2 (−10 dB)	150 × 75 × 0.8	11.57 × 17.95 × 0.8	>10	<0.3	84
[30]	3.4–3.6 and 4.8–5.1 (−10 dB)	150 × 75 × 7	15 × 3.1 × 0.8	>11.5	<0.08	79
[31]	3.4–4.7 (−10 dB)	37 × 30 × 0.8	17 × 8 × 0.8	>15	<0.005	95
[32]	3.3–5 (−10 dB)	150 × 75 × 0.8	6.8 × 7 × 1	>12	<0.12	82
[33]	3.3–5 (−10 dB)	150 × 75 × 0.8	15.5 × 17 × 0.8	>10	<0.1	72
[34]	3.3–6 (−6 dB)	150 × 75 × 7	15 × 13 × 0.8	>10	<0.12	71
This work	3.3–5 (−10 dB)	150 × 75 × 7	6.8 × 7 × 0.8	>16	<0.09	95

compares the designed antenna with some reported MIMO antennas for the 5G frequency bands. The proposed MIMO antenna has a smaller size than those mentioned. Significantly, it achieves higher isolation levels than all of them. In terms of ECC, it is lower than that of most, with exceptions in the cases of references [30,31]. This clearly demonstrates the superiority of the designed antenna in size—compactness, isolation, and ECC performance.

5. Conclusions

An eight-port MIMO antenna intended for future ultra-slim smartphones was designed and fabricated. It operates across the 5G bands N77, N78, and N79. The design boasts a compact footprint (7 × 6.2 mm²) and excellent isolation performance (16 dB). The measured −10 dB impedance bandwidth is 1.7 GHz (3.3–5 GHz). All the ECC values are below 0.016, which are small enough to be considered as uncorrelated for any two radiated waves in a MIMO system. The measured peak gain is 5 dB. A high radiation efficiency of 70% was achieved for Ant. 1, Ant. 4, Ant. 5, and Ant. 8, while Ant. 2, Ant. 3, Ant. 6, and Ant. 7 reached an efficiency of 78%. The investigated MIMO antenna is a good candidate for integration into ultra-slim smartphones.