An X-Band High-Isolation MIMO Antenna with a Resonant Ring Metamaterial Decoupling Structure

Article information

J. Electromagn. Eng. Sci. 2026;26(2):193-199

Publication date (electronic) : 2026 March 31

doi : https://doi.org/10.26866/jees.2026.2.r.355

Xuemei Zheng^1,, Tongchao Zhang²

¹Key Laboratory of Modern Power System Simulation and Control and Renewable Energy Technology, Ministry of Education, Northeast Electric Power University, Jilin, China

²School of Electrical Engineering, Northeast Electric Power University, Jilin, China

^*Corresponding Author: Xuemei Zheng (e-mail: zhengxuemei@neepu.edu.cn)

Received 2024 November 26; Revised 2025 March 8; Accepted 2025 October 22.

Abstract

I. INTRODUCTION

The X-band (8–12 GHz), known for its short wavelength and high resolution, is fundamental to radar, satellite communication, and electronic warfare systems [1]. The growing demand for higher data throughput and more robust links in these applications has pushed conventional single-input single-output systems beyond their limits in this crowded spectrum. Multiple-input multiple-output (MIMO) technology, which uses spatial diversity to support parallel data streams, can significantly improve channel capacity without additional spectral allocation [2]. This is critical for modern X-band systems, where spectral efficiency is paramount. However, integrating multiple antenna elements into compact platforms reduces inter-element spacing, intensifying mutual coupling via surface waves and severely degrading isolation and overall MIMO performance. Thus, effective decoupling techniques are essential for developing compact, high-performance, broadband X-band MIMO arrays.

To address mutual coupling, various decoupling methods have been explored. Defected ground structures suppress coupling currents by etching patterns to form equivalent Inductor-Capacitor (LC) circuits, but the complexity of their design increases with bandwidth, and they have limited broadband adaptability [3, 4]. The neutralization line technique cancels mutual coupling by introducing a phase-inverted current path, but it is inherently narrowband and may distort radiation patterns [5, 6]. Recently, metamaterials such as electromagnetic bandgap (EBG) structures and frequency selective surfaces (FSS) have attracted attention due to their exceptional wave manipulation capabilities. While EBGs effectively suppress surface waves [7, 8], their typical unit size (around half a wavelength) is too large for dense X-band arrays. Similarly, FSS-based decouplers [9, 10] suffer from bandwidth limitations due to periodic resonance, making them unsuitable for ultra-wideband (UWB) X-band operation (8–12 GHz). There is thus a pressing need for a decoupling solution that is compact, simple, and effective across the entire X-band.

This paper proposes a symmetrical metamaterial decoupling structure to overcome these limitations. When placed above a two-element X-band MIMO antenna, the metamaterial layer effectively mitigates mutual coupling, achieving isolation levels higher than −22.7 dB over the ultra-wide bandwidth of 8.1–12.4 GHz, thereby meeting the requirements of modern wideband X-band applications.

II. Design of Proposed Antenna Array

The novel decoupling array antenna designed for this research consists of two layers. As shown in Fig. 1(a), the lower layer is a tree-shaped radiation patch, and the upper layer is a metamaterial layer. Fig. 1(b) presents a manufacturing photo of the prototype antenna. To validate the antenna, we evaluated its performance using a vector network analyzer and experimental simulations.

Fig. 1

(a) Schematic diagram of the proposed MIMO antenna with a metamaterial cover layer. (b) Physical prototype of the antenna with the metamaterial layer.

1. MIMO Antenna Structure

This study presents a metamaterial-based MIMO antenna, whose structure is illustrated in Fig. 2. The design employs a 1.6-mm-thick FR4 substrate (ɛ_r = 4.4, tanδ = 0.02) featuring a tree-shaped radiating patch on the top layer and a comb-shaped ground plane with an integrated stub on the bottom layer. The ground stub extends the −10 dB impedance bandwidth to 8–12 GHz (49.6% fractional bandwidth), covering the entire X-band. The key dimensions are detailed in Table 1.

Fig. 2

Geometry of the proposed antenna: (a) front view, (b) back view.

Table 1

Antenna array parameters (unit: mm)

2. Metamaterial Element

Electromagnetic properties are governed by the fundamental laws of electromagnetism. The permittivity (ɛ) and permeability (μ) of most natural materials are positive. In contrast, the negative ɛ and μ values in metamaterials arise from the unique interaction between their sub-wavelength unit structures and incident electromagnetic waves. Through careful design, metamaterials can satisfy the conditions for negative ɛ and μ within the target bandwidth. In this context, their electromagnetic behavior is critically dependent on the geometric precision of unit cells. Fig. 3 illustrates the monolayer metamaterial element we designed, which features two symmetrically arranged split rings with an internally thickened “C” ring for structural improvement, which excites an additional resonance mode. The final dimensions of the metamaterial unit, determined via parametric optimization, are summarized in Table 2.

Fig. 3

Structural diagram of the metamaterial element.

Table 2

Metamaterial element parameters (unit: mm)

To characterize the electromagnetic behavior of the proposed metamaterial component, simulations were conducted using Ansys High-Frequency Structure Simulator (HFSS). First, we built a precise geometric model, defined the sizes of its components, and assigned the material parameters. We determined the proper port types and modes as well as the boundary conditions, such as perfect electric conductor and perfect magnetic conductor. We then selected a solver, set the parameters, and ran simulations. Finally, we analyzed the results and gathered key electromagnetic data to enable performance evaluation and optimization. Based on these simulations, we computed the S-parameters and used their real and imaginary components to determine the equivalent permittivity [11], as presented in Eqs. (1)–(4):

(1) z=±(1+S11)2-S212(1-S11)2-S212

(2) eink0d=S211-S11·z-1z+1

(3) n=1k0d{[Im(lneink0d)+2mπ]-i·Re(lneink0d)}

(4) μ=n·z, ɛ=nz

where k₀ is the wave number in free space, n is the refractive index, z is the wave impedance, ɛ is the equivalent dielectric constant, and μ is the equivalent magnetic permeability.

(5) zs=z·tanh (jk0tn)

The equivalent permittivity (ɛ) and permeability (μ) of the metamaterials were calculated using the S-parameter inversion method. Their negative properties are attributable to the strong resonant response of the element structure. Negative ɛ values can inhibit surface wave propagation and reduce near-field coupling between adjacent antenna elements. The negative μ value weakens the mutual coupling current by regulating the magnetic field distribution. Notably, the equivalent parameter method was used for the qualitative analysis of the decoupling trend. However, this accuracy of this method was limited by the thickness of the metasurface (t ≪ λ₀). Thus, a surface impedance model (Eq. 5) was developed to quantify the metasurface’s contribution to the isolation between the antenna elements.

Fig. 4 presents the equivalent permittivity and equivalent permeability values calculated using MATLAB based on the simulation results for the metamaterial unit in HFSS. Within the X-band, the real parts of the curves for permittivity and the equivalent permeability are negative.

Fig. 4

Permeability and equivalent permittivity.

III. Simulation and Performance Analysis

To validate the device’s decoupling efficacy, a 3 × 3 metamaterial array (unit periodicity: X = 7.6 mm, Y = 6.4 mm) was integrated as a metamaterial above the dual-element MIMO antenna, as shown in Fig. 1. The metamaterial decoupling layer was positioned H = 2 mm above the radiating patches.

When MIMO antenna arrays are miniaturized, they end up too close together, and the coupling current diminishes the system’s energy efficiency. When one of the antennas is excited and a side lobe propagating along a specific direction is generated, this side wave induces a current in the nearby antenna. This induced current is an indication of mutual coupling between the antennas and can have various impacts on the antennas’ performance. Due to the unique properties of the metamaterials, a region with negative permeability and negative permittivity— which is beneficial for enhancing isolation—forms in the area above the dual-element radiation patch.

Eq. (8) reveals that the wave vector k and the Poynting vector S exhibit opposing characteristics (S · k < 0), resulting in a reversed phase propagation relative to the energy flow. This anomalous propagation mechanism suppresses the lateral diffusion of surface waves along the antenna array’s transverse direction (the X-axis), redirecting energy into free-space radiation and thereby reducing mutual coupling currents between adjacent antenna elements [12]. Specifically, when Antenna I is excited, the edge-scattered waves are localized by the metamaterial structure instead of propagating along the X-axis toward Antenna II, as demonstrated by the field distribution comparisons in Fig. 5.

Fig. 5

Distribution and direction of coupled current in the MIMO antenna array: (a) Antenna array without the metamaterial decoupling layer, (b) Antenna array with the metamaterial decoupling layer.

Fig. 6(a) compares the S-parameters of the radiation patch before and after integrating the metamaterial layer. When the metamaterial layer is not introduced, the dual-element radiation patch covers the entire X-band frequency range. However, due to the strong mutual coupling, the maximum isolation is −12.5 dB, which is lower than the performance requirements of standard MIMO antennas [13]. Positioning the symmetrical split-ring metamaterial array assembly on top of the antenna array improved the isolation to −22.7 dB, achieving a 10.2 dB improvement and demonstrating effective decoupling. The experimentally obtained S-parameters of the metamaterial-integrated antenna are closely aligned with the simulation results shown in Fig. 6(b). Minor discrepancies were observed and were primarily attributed to inherent variations in fabrication during prototype realization.

Fig. 6

S-parameters: (a) simulated and (b) measured.

The broadband response of the negative equivalent parameters across the X-band arises from the coupling between the low-frequency electric resonance (8 GHz) of the thickened outer C-shaped ring (UL1 = 6.2 mm, UW1 = 6.4 mm) and the high-frequency magnetic resonance (12 GHz) of the inner narrow ring (UL2 = 4.6 mm, UW2 = 0.5 mm), as shown in Fig. 4. Structural optimization via a feed structure (UL3 = 0.4 mm, UW3 = 2 mm) introduces an additional localized resonance near 10.5 GHz (indicated by the arrows in Fig. 6), bridging the mid-band negative parameter gap and enabling full X-band coverage.

Fig. 7 illustrates the antenna’s current distribution across a trio of designated frequencies: 8.2 GHz (low band), 10.2 GHz (mid-band), and 11.8 GHz (high band). When port 1 is excited, port 2 is terminated with a 50 Ω matched load. As shown in the left panels of Fig. 7 (without the metamaterial layer), when the metamaterial layer is not present, the current generated by the excited antenna spreads to the adjacent antennas. The right side of Fig. 7 represents the surface current of the radiation patch with the loaded material layer. This demonstrates that the metamaterial layer considerably reduces coupling currents.

Fig. 7

Antenna surface current diagram: (a) 8.2 GHz, (b) 10.2 GHz, and (c) 11.8 GHz.

Fig. 8 shows the two-dimensional radiation pattern of the far-field E-plane and H-plane of the MIMO antenna at 8.2 GHz, 10.2 GHz, and 11.8 GHz. The E-plane radiation pattern exhibits a classical dipole-like “figure 8” shape, while the H-plane radiation pattern remains nearly omnidirectional, with a maximum gain variation of ±1.2 dB in azimuth—the level the antenna was designed to reach. Fig. 8(a) shows the actual radiation pattern measurement in a microwave anechoic chamber. The dashed curves in Fig. 8(b)–8(d) represent the radiation patterns of the metamaterial-integrated antenna. These results are in strong agreement with the patterns from experimental tests under single-port excitation, confirming that the decoupling design’s performance matches the simulation predictions.

Fig. 8

Measured and simulated radiation patterns: (a) experimental setup, (b) 8.2 GHz, (c) 10.2 GHz, and (d) 11.8 GHz.

The envelope correlation coefficient (ECC) is an important parameter that indicates the signal correlation between antenna elements in a MIMO system. The correlation is calculated based on the envelopes of the signals received at two antenna ports. Diversity gain (DG) refers to the improvement in anti-fading performance that the system achieves when using multiple antennas or signal paths to transmit and receive signals. ECC and DG are critical metrics for evaluating the diversity performance and mutual coupling characteristics of MIMO antennas:

while DG evaluates fading resistance improvement through spatial diversity:

The ECC and DG of the proposed MIMO antenna are shown in Fig. 9. The proposed antenna achieved an ECC below 0.002 and a DG exceeding 9.99 dB across the entire X-band. These results surpass the standard thresholds (ECC < 0.5, DG ≤10 dB). According to the ECC and DG diagrams, the antenna exhibited excellent MIMO diversity and coupling performance within the working frequency range.

Fig. 9

ECC and diversity gain.

The peak gain of an antenna, defined as the maximum gain value in a specific direction of its radiation pattern, quantifies its ability to focus input power into a directional beam through electromagnetic field interference. As shown in Fig. 10, without the metamaterial layer, the antenna’s peak gain remains below 3.8 dB (2.1–3.7 dB) across the X-band. In contrast, after integrating the metamaterial array, the peak gain increases to 5.2–6.1 dB, marking a 38% enhancement. This improvement was attributed to the metamaterial’s ability to suppress surface waves and optimize radiation efficiency. The experimentally measured peak gain pattern exhibited close agreement with the numerical simulations across the operational bandwidth, with observed deviations attributable to inherent fabrication tolerances.

Fig. 10

Peak gain of the antenna.

IV. Performance Comparison

The performance of the decoupling antenna designed in this paper was compared with results from previous literature [14–20]. Table 3 demonstrates that the proposed antenna achieved a wider bandwidth (8–12 GHz) than prior designs and stabler impedance matching (S₁₁ = < −10 dB). The improved isolation and low ECC achieved by this metamaterial structure make the antenna highly suitable for MIMO systems operating in the X-band, such as automotive radars and satellite communication systems.

Table 3

Metamaterial element parameters

V. Conclusion

This paper proposes a single-layer metamaterial decoupling layer to suppress mutual coupling in broadband MIMO omni-directional antenna arrays. Through iterative electromagnetic simulations and parametric optimization, the design achieved a reduction in mutual coupling from −12.5 dB to −22.7 dB across the X-band, a peak gain enhancement to 5.2–6.1 dB (a 38% improvement over the baseline), and a radiation efficiency exceeding 85% with stable omnidirectional radiation patterns (front-to-back ratio >15 dB). Comprehensive evaluations of ECC (<0.002), DG (>9.99 dB), and impedance matching (S₁₁ < −10 dB) validate the antenna’s suitability for X-band wireless systems, particularly in satellite communications and phased array radars requiring high isolation and wideband operation.

Article information Continued