Effect of Metamaterial Cells Array on a Microstrip Patch Antenna Design

— In this paper, a microstrip patch antenna (MPA) design was developed to verify the performance of a metamaterial cell array (MTM) Capacitive Loaded Loop (CLL) immersed in the substrate. The metamaterial was obtained from metal laminates designed inside a dielectric material, with dimensions carefully calculated to have the permeability, consequently the negative refractive index, in a certain frequency range. In this case, the geometric control of the unit cell parameters allowed its organization in a periodic arrangement to be immersed in the MPA substrate. Simulated and experimental results were obtained and compared for some parameters of the antenna, which showed a good relationship between MTM and MPA performance.


I. INTRODUCTION
The continuous search for new technologies has led researchers to study more and more the application of metamaterials (MTMs) in wireless communication system [1]. As it covers a wide range of materials, its properties can be exploited to control and improve the propagation characteristics, thus affecting the antennas functioning, in particular the microstrip patch antenna (MPA) [1]. MTMs are structures artificially obtained, usually periodic and can present electrical permittivity (ε), magnetic permeability (µ) and refractive index (n) simultaneously negative [2]- [4].
In literature, several geometries are characterized as MTMs structures that are applied in the patch, within the substrate and in the ground plane of MPAs, aiming to optimize their performance and its parameters [1], [3]- [15]. For projects with metamaterials in the patch, it was proposed in [1], to use a Capacitive Loaded Split Rectangular Loop (CLL) resonator to reduce and control the propagation of electromagnetic waves. In the project by [6], a wireless local area network (WLAN) antenna used resonators with MTM based on epsilon negative (ENG) transmission line (TL) with probe-fed. It obtained a greater bandwidth (BW) 2.3 times greater than the standard antenna and greater width of the axial ratio. In a ultra wideband (UWB) antenna [7], a circular SRR resonator was used with the closing of the inner ring. Optimized BW by 146.56% and gain of 8.57 dBi. For a miniaturized antenna to operate in five bands [12], the ring-shaped patch used an MTM resonator type Rectangular Complementary Split Ring Resonator (RCSRR) inside the ring. Thus, it reduced the dimensions by 46.8% and increased integrated with MTM cells, to control the beam direction. MTMs were inserted next to the antenna patch. This achieved a maximum gain of 7 dB to 4.9 GHz and increased BW by 40%. According to [17], an antenna with an MTM type Complementary Split Ring Resonators (CSRRs) in hexagonal form has been proposed. It achieved a 5.7% reduction in operating frequency, reflection coefficient (S11) from -34 dB at 2 GHz, greater gain and the radiation is unchanged. In [18], it was proposed a superstrate antenna using SRR type MTM cells in square rings. It achieved a gain of 9.991 dB at 4.521 THz and 1.11 dB at 3.846 THz, a decrease in half-power beam width (HPBW) compared to the conventional antenna. For [19], the antenna used Split Ring Resonator cells and thin wires (SRR/TW) over the antenna patch. At 303 GHz, it improved the beam focus, gain, BW and S11 (1.1 dB, 14.73 GHz and 12.63 dB, respectively), in addition to directivity and radiation. At 10 GHz it improved by 1.5 dB, 10.15 MHz and 2 dB for gain, BW and S11, respectively.
In projects with metamaterials within the substrate, MTMs SRR cells with square geometry helped to obtain frequency ranges (5.47 -5.6 GHz and 5.72 -5.82 GHz) for WLAN and increased gain [9]. In [13], the SRR labyrinth geometry obtained an increase in BW by 600%, better voltage standing wave ratio (VSWR) by 1.5% and a miniaturized antenna by 400%. And an antenna for THz [20], obtained a high gain using a rectangular MTR type SRR.
On the ground plane, an antenna for the frequencies global system for mobile communications (GSM), worldwide interoperability for microwave access (WiMAX), Bluetooth and WLAN [11], obtained a reduction in electromagnetic absorption with MTM SRR in hexagonal rings, by 66.47% and 66.52% for 1.8 GHz and 2.4 GHz. In a THz antenna [15], a circular shaped Split Ring Resonator (SRR) is used to construct the MTM surface. It obtained reduced antenna, S11 of -65 dB and the gain changed significantly in 1.02 THz.
Another form of design is to insert cells on both sides of a dielectric plate, or in the patch and ground plane. In a microstrip dipole antenna [4], cells with divided circular geometry, such as Split Ring Resonator (SRR), obtained an increase in BW, gain and decrease in the value of the S11. For a high gain antenna in THz [21], it was used a unit cell of the axially symmetric fishnet with a low index MTM (LIM). Optimized the gain by 4 dB, directivity and better radiation efficiency. In a UWB antenna [22], the authors used a Symmetric Split Ring Resonator (SSRR) MTM cell in divided circular rings. It was obtained a BW of 128.3% and a gain of 6.12 dBi. Finally, a combination of square resonant rings (SRR) and complementary split ring resonator (CSRR) [23], was performed on a 2.4 GHz antenna. The SRR was inserted next to the feed line, while the CSRR was below the feed line, in ground plane. Achieved 4.55% BW without affecting coupling and directivity.
In adittion, another characteristic of projects is the use of the transmission line as an MTM resonator [8], in which a miniaturized antenna is based on the active MTM resonator used in transmission line.
With this, it was obtained a greater gain than the traditional antenna.
On the other hand, a multiple input multiple output (MIMO) antenna inserted the CLL type MTM cells [10], next to the feed line of each antenna to reduce electromagnetic interference and increase the isolation between the two antennas. As a result, good isolation with less than 25 dB and a wide bandwidth of 65.5% was achieved.
Some studies also report other types of resonator designs, such as MTM designs reconfigurable for multiple frequencies in THz with Electric Split Ring Resonator (eSRR) cells [24], based on microelectromechanical systems in the resonators. In [25], a sensor was proposed that uses a microwave technique with coupled antenna. It used a rectangular split ring resonator (SRR) MTM cell. It obtained a high Q factor, reduced size, good sensitivity and high repeatability for detection applications. Finally, a resonator based on an RLC circuit modified to 8.84 GHz [26], was based on a conventional rectangular SRR, obtaining a greater BW at 3.4 GHz.
Based on these studies, the present work has as its main contribution the analysis of a MPA composed This paper is structured as follows: Section 2 presents the metamaterial design. Section 3 describes the design of the MPA with and without MTM. Section 4 presents the discussions and results. Finally, the conclusions are presented in Section 5.

II. METAMATERIAL DESIGN
To obtain MTMs used in MPAs project proposed in this work, initially the unit cell presented in [1], [3], [5] was structurally modified to operate in the project's frequency range of interest (5.8 GHz). Next, an array of unit cells was organized to obtain the metamaterial substrate used in the MPAs. The characterization of the effective permittivity (εeff) and the effective permeability (µeff) of the unit cell that originated the arrangement was made by adjusting the copper laminate dimensions that constitutes it, so that electromagnetic interactions occur that make them negative in MPA operating frequency. Important to highlight that for a unit cell to perform well as a metamaterial, its geometry must be much smaller than the operating wavelength. This characteristic is essential to obtain the negative ε, µ and n parameters [12], [16], [27], [28].
In some cases, cell dimensions do not meet the desired results, so it becomes necessary to gradually adjust their dimension so that the computational results meet the conditions to become metamaterial.
The excitation of a unit cell, for the case of simulation analysis, requires the oblique incidence of a plane electromagnetic wave, in addition to the application of suitable periodic boundary conditions, to obtain the scattering matrix (S) parameters correctly, i.e. the reflection coefficient (S11) and the transmission coefficient (S12) [27], [29]. These parameters are calculated according to (1) and (2) [16], [28], [29]. In this context, once the parameters S11 and S12 are obtained by simulation using ANSYS HFSS ® software, they are exported and mathematically manipulated through an algorithm developed in Scilab software to obtain εeff, µeff and n [16], [28], [29]. Numerical simulations used a tool with a mathematical model based on the finite element method.
Where l1 is the maximum length of the unit cell, and k0 is the wave number in free space.
From Lp and Wp, the results obtained for the length (Lg) and width (Wg) of the ground plane, can be approximated by the following equations [30]: The MPAs were excited from a quarter wavelength feed line integrated into the patch to have a maximum impedance matching between the 50  line and the radiant element [31], [32].  Table 1. The MPA simulation used the ANSYS HFSS ® software.  In addition, for the MPAs construction, two FR4 slabs with dimensions of Lg x Wg and thicknesses (t' and t") were used. For MPAs with MTM, on the first slab with thickness (t'), the metamaterial cell array was designed on one side, while on the other side the copper laminate was maintained. On the second slab with thickness (t"), on one side the antenna patch was designed, and on the other side the copper laminate was removed. Then the two slabs were overlapped, such that the unit cell array was positioned between the two slabs, equidistant from the ground plane and the antenna patch, as seen in Fig. 3 and Fig. 4. Thus, the FR4 slabs used formed a substrate with a total thickness of 3.1 mm using 0.03 mm thick copper laminate on its sides.     7 shows the simulated and measured results for the reflection coefficient parameter (S11) for MPA without MTM and MPAs with substrates based on MTMs, considering the distance (0) between the cell centers along the direction ± ŷ (0) and dimension Wg for 4 mm, 5 mm and 6 mm. It is observed that small variations in 0 can cause significant changes in BW and impedance matching between the feed line and the patch. Therefore, in the variation of 0 = 6 mm (Fig. 7d) it presented the lowest reflection coefficient (S11), while the largest bandwidth was obtained by 0 = 5 mm (Fig. 7c).
In addition, with the graphs of parameter S11 showing values within the range of -10 dB, it is clear that increasing 0 between cells results in decreased resonance frequency while increasing bandwidth.
It is also evident that in the desired region of antenna operation (Fig. 1b), the CLL cell has a magnetic characteristic, different from the electrical one. These characteristics induce the εeff of the antenna substrate to increase, which causes the resonance frequencies to decrease, allowing a miniaturization of the antennas, since lower frequencies can be achieved with the smaller dimensions for the antennas [31].  With the measurements, it is clear that there is a good agreement between the simulated and the measured results. It can be seen from Fig. 7b that the increase in BW occurred due to the emergence of an artificial mode (f1) generated by the presence of the unit cell array introduced into the substrate.  Fig. 8 and Fig. 9 show the electrical field distribution in the MPA without MTM patch and ground plane (Fig. 8) and the MPA with MTM (Fig. 10), where it is found that insertion of the MTM cells into the substrate causes an expansion in electrical widening of the physical structure of the patch. According to the electric field distribution in the ground plane of Fig. 9, the appearance of electric fields along the ± ŷ direction is observed, giving rise to the emergence of artificial modes (AM). As the physical dimensions are inversely proportional to the frequency [31], one can identify the artificial and natural modes in the S11 curves. Therefore, it is evident that the artificial resonant mode predicted in simulations was validated in the measurements through S11 parameter.     Fig. 11 shows the simulated results for the voltage standing wave ratio (VSWR), with a desirable value equal to or less than 2 [36], for MPAs design. When modifying the distance between the cells, the MPA with 0 of 6 mm obtains the lowest VSWR value, as well as obtaining a better impedance matching, which caused a lower reflection coefficient, as shown in Fig. 7d.  important aspect of analysis is the influence of MTM on the rear lobe of MPAs, that is, it is possible to analyze that the antennas with 0 = 5 mm had a decrease in their module, on the other hand the antennas with 0 = 4 mm and 0 = 6 mm presented the rear lobe radiation diagram similar to that of the standard antenna. These effects are due to the scattering created by the MTM cells strongly influenced by the distance between the cells and it is noticed that depending on the configuration, distance between the cells and periodic disposition, the rear lobe may decrease or increase.   It is observed that the MPA with MTM of 0 = 5 mm (Fig. 13c), has a greater gain in relation to the others that have MTM, while that with 0 of 4 mm (Fig. 13b) has the lowest gain. In addition, the greatest gain is achieved when theta equal to -15° for MPA without MTM, and as the grating of the metamaterial cells increases, the direction of maximum propagation rotates to theta equal to -40° in the case with of 6 mm between cells (Fig. 13d). From the results obtained in this work, Table II presents a comparative summary of the parameters, for the simulated and measured results of MPAs with and without MTM. To validate the results obtained in this work, Table III shows a comparison between several works in the literature, in which it is possible to analyze the frequency at which the antenna operates, the type of resonator used in the making of the metamaterial, the bandwidth achieved and the gain of these antennas. V. CONCLUSION In this paper, an MTM geometry is characterized and inserted into the MPAs substrate with 9 x 7 element arrays. The main proposal is the construction and applicability of MTMs inserted in the substrate, with the possibility of manufacturing printed circuit boards that use the antenna structure. The distribution of the cells and their variations among them were chosen to verify the influence of the distance variation (0) on the antenna performance. In this context, it was found that small variations in the ± ŷ direction of 0 cause large variations in some antenna standardization parameters. Initially, the projected structures were analyzed through simulation using ANSYS HFSS ® software version 19.2 and software Scilab, from which prototypes were fabricated, which allowed the experimental results to be obtained. The introduction of metamaterial in the MPA enabled a better performance, in the reflection coefficient (S11), bandwidth (BW), impedance matching and electric field distribution, besides provoking the emergence of artificial modes that influenced the enlargement of the bandwidth.
In adittion, it was verified that the proposed antenna has miniaturization capacity, according to the responses presented. However, there was a little decrease in a gain and radiation efficiency. The measured and simulated results were compared and showed good agreement.