Higher Order Mode Attenuation in Microstrip Patch Antenna with DGS H Filter Specification from 5 to 10 GHz Range

— This work proposes the application of a slot in the ground plane, Defected Ground Structure (DGS), of a microstrip patch antenna. The application of this technique aims to attenuate higher-order excited modes without deteriorating the antenna parameters in the fundamental mode. The methodology for the design of the slot in H-format is presented, based on the physical dimensions of the microstrip patch antenna. The positioning of the DGS concerning the feed line was determined from a detailed parametric analysis. The antenna was designed for the fundamental mode at 2.45 GHz frequency. Two antenna prototypes were built, with and without DGS, and measurements of reflection coefficient, gain, and radiation pattern were performed. The measured results show that the application of DGS considerably attenuated higher-order excited modes. A good agreement between simulated and measured results is achieved.

as increasing the size, weight, and cost of the wireless device. In this way, some alternative solutions that do not increase cost or complexity for the device can be attractive. Thus, solutions that use the antenna itself to filter these higher-order modes become interesting [4], [5].
Several studies are being developed for design designing antennas able to attenuate or suppress unwanted frequencies. Among the techniques used, we can highlight the use of periodic structures and DGS insertion. Periodic structures can increase the complexity and fabrication cost of the antenna, reducing some advantages of this antenna type, whereas the DGS generally uses simple geometries and is capable of providing rejection in a wide bandwidth [6], [7]. This work proposes the use of DGS in H-format for attenuation of higher-order excited modes in a microstrip patch antenna for WLAN application. The reference antenna radiates higher-order modes over a frequency range of 3 GHz to 10 GHz, which can cause interference in systems using licensed frequencies. The idea is to attenuate as much as possible all higher-order modes within this frequency range of interest. The methodology, discussions, and results will be demonstrated in the following sessions.

II. PREVIOUS WORKS
In [2], Chang  Biswas et al. [4] investigated harmonic suppression in rectangular patch antenna for S-band application around 3.11 GHz, using a stub, to improve impedance matching. Harmonic suppression was achieved by the addition of two truncated circular rings aligned with the microstrip feed line. The project allowed the suppression of 6 higher-order modes. However, the final set had a relatively large occupied area, which may be inappropriate for some applications.
Pattapu et al. [5] proposed a circular patch antenna applied for power transmission. The design operating frequency was 2.45 GHz. It was proposed to insert two DGS on a ground plane, one was a rectangular DGS and the other one was a loop DGS for suppression of higher-order harmonics. The results show a good attenuation in the rejection range, but the area occupied on the ground plane was relatively large. This situation decreased the total gain, as well as the front-to-back ratio.
Sena et al. [7] proposed the use of genetic algorithms to optimize the design of hexagonal slots to suppress higher-order modes. The antenna was designed for 5.8 GHz. An optimization using AG was performed to adjust the number and diameter of the slots, distance between it and positions in the ground plane. The proposal allowed the attenuation of a relatively small number of higher-order modes, as well as causing an increase in back radiation, reducing gain and efficiency.
Rahim et al. [8] proposed a circular patch antenna for the frequency of 2.45 GHz. For suppression of higher-order harmonics, the authors inserted 5 non-uniform rectangular slots as DGS. In addition, to improve impedance matching, the authors inserted a stub in a microstrip feed line. The proposal modifications allowed an improvement in impedance matching, but the attenuation of higher-order modes was limited. In addition, the authors made no comparison with the performance parameters between the reference antenna and the proposed antenna.
In [9], the authors proposed a rectangular patch microstrip antenna that incorporates fractal slots at 2.45 GHz with coupling feed. For attenuation of higher-order modes, an H-format DGS was added below the power line. The authors note that the variety of techniques has added improvements to impedance matching, radiation efficiency, and size reduction. However, the proposed antenna is complex and the performance is below when compared to other antennas presented in the state of the art of this work.
In [10], the authors proposed a rectangular patch antenna to operate at 2.45 GHz. To improve the impedance matching a L-shaped stub was added to the microstrip feed line. For suppression of higherorder harmonics, two I-format DGS were added under the microstrip feed line. The proposal allowed the suppression of a significant number of higher-order modes. However, the DGS occupied a significant area which led to a reduction in performance parameters such as gain and radiation diagram.
In this work, we propose a rectangular microstrip patch antenna, to operate at 2.45 GHz. For higherorder modes suppression, we propose to insert an H-shaped DGS. We used the equivalent circuit to design the initial dimensions of DGS. Besides, to improve impedance matching we use the ANSYS HFSS optimization tool, to indicate the best position of DGS concerning the inset fed. Thus, our proposal achieves significant higher-order mode suppression without causing changes in antenna performance parameters due to the minimum ground area occupancy. The performances of cited works in this section are compared in TABLE I. The results are compared in terms of the fundamental frequency, the number of higher-order modes suppressed, the proposed technique, and the area occupied by the DGS. The dielectric considered in the design was the FR4, with a thickness of 1.6 mm, an electrical permittivity of 4.4 and a loss tangent of 0.02. Fig. 1 illustrates the physical characteristics of the reference patch antenna, which is designed to operate in the dominant mode (TM10) at 2.45 GHz.
The antenna design follows the classic approach in [11]. To obtain an input impedance of 50Ω, the length and width of the feed line were calculated appropriately. To achieve an effective impedance matching, the inset-fed technique was used. The feed line width value (Wf) is related to the thickness (h) and effective dielectric constant (εef) of the substrate. The inset-fed (Y0) of the line is related to input resistance (Rin) and patch length [11], [12].  After the design, these specifications were considered for modeling and simulation within the ANSYS HFSS software. Simulations were performed for reflection coefficient, gain, radiation pattern and Smith chart. Fig. 2 shows the frequency response of |S11| in dB. It can be observed that in addition to the fundamental mode (TM10) the antenna radiates TM20, TM30 and TM40 harmonic modes (2 nd , 3 rd , and 4 th harmonics, respectively), TM02 and TM22 higher-order modes, and two spurious modes. The excited modes are analyzed by the electric field distribution on the patch antenna. According to [11], to determine the resonant modes, TMmn, we must evaluate the number of variations maximum and minimum of the electric field potential. Thus, the evaluation of the modes is given by the analysis of the electric field distribution on the radiating patch element regarding Lp (length) and Wp (width).
where c is the light speed in free space and εr is the relative permittivity of the substrate.
The following investigations are intended to examine the radiated electric fields for the modes. It is possible to identify higher-order modes through the study of modal fields. It can be observed the excited modes of the reference antenna in Fig. 3. So, modes: TM20, TM30 and TM40 are the 2 nd , 3 rd and 4 th harmonics of the fundamental mode frequency. Other modes are presented but are not characterized as harmonic modes. Spurious modes arise from the approximation of harmonic modes.
In the case of microstrip antennas, spurious modes are also excited by surface waves. It was not the object of study in this work. The excited modes have the characteristic of concentrating the electric field in the patch of the antenna, mainly in its extremities. In Fig. 3, the region in darker red represents a greater concentration of electric field, while the region in darker blue represents a lower concentration.   Through these, adjustments become possible that will allow allocating the rejection band as desired.
The analysis and design part of the DGS were presented in [10], [18]. Our design should allow suppression of higher-order modes while maintaining the S11 greater than -8 dB for all modes but without affecting antenna performance. The intention is, if possible, to improve some performance parameters. Thus, we designed the DGS to have a cut-off frequency equal to twice the antenna frequency (4.9 GHz) and a rejection bandwidth sufficient to suppress higher-order modes, of a resonance frequency of 9 GHz. To find the optimal position of the DGS in the ground plane the ANSYS HFSS optimization tool was used. Thus, this position allowed the best antenna impedance matching, occupying the smallest possible area, maintaining important characteristics such as front-toback ratio, gain, and unchanged radiation diagram. The projected dimensions are listed in TABLE III.  modes. After parameterization using the ANSYS HFSS optimization tool, the DGS was positioned on the ground plane as shown in Fig. 7. In our case, the DGS was not centralized concerning the feed line. It was displaced from 1.0 mm relative to the center of the feed line to the right and 0.2 mm in direction of the lower edge of the patch.  The reactive components shown in Fig. 5, C and L can be calculated as function of fc and f0 as [18]: where Z0 is the characteristic impedance of 50 Ω.
After the design and analysis of the DGS influence on fundamental mode, simulations were performed in order to obtain some parameters of the antenna. Fig. 8 shows a comparison between the S11 of the reference antenna and the proposed antenna. It can be observed that the proposed antenna presents better results, because besides completely suppressing most of the higher-order modes and, in the frequency range between 2.5 GHz to 10 GHz, also improved the impedance matching at 2.45 GHz, by reducing the S11 from -20 dB to -24 dB.
In fact, that occurs because the impedance of the reference antenna is 47.48 + j16.98 Ω and with the parametrization of the position of the DGS the impedance has a significant decrease of the reactive part and an increase of the resistive part. So, the proposed antenna obtains an impedance of 51.83 + j0.08 Ω.

V. MEASURED RESULTS
For purposes of proving the effects addressed in the previous sections, two prototypes were constructed, one of the reference antenna and the other of the proposed antenna. These prototypes are illustrated in Fig. 10. Measurements were performed with an Agilent Technologies brand vector network analyzer: model E5071C with a measurement frequency range from 1 to 10 GHz, for visualization of higher-order modes, within this range of interest. Fig. 11 shows the measurement setup, where AUT is the antenna under test. Fig. 11. Measurement setup.
The first measurement was performed for the reference antenna. We measure the S11 to prove the existence of higher-order modes. Fig. 12 illustrates this measurement. It can be noticed that all modes confirmed in simulation appear. A good agreement between the results can be observed. Then, measurements for the proposed antenna with the DGS in H-format were performed. As in simulation, the DGS altered the antenna impedance, significantly attenuating the higher modes, and most of them were suppressed. Once again, a good agreement between the results can be observed. Fig. 13 illustrates this proof. Fig. 14 (a) shows the measured Smith chart for the reference antenna. We can see that there is good agreement with the simulated results. As in the simulation, the inductive reactance is significant, with values closer to zero being desirable.
The measured Smith chart, for the proposed antenna, is shown in Fig. 14 (b). As for the simulated values, it can be seen that a significant reduction of the inductive reactive part of this parameter occurred, as well as, it can be observed that improvement in impedance matching. Fig. 13. Comparison between simulated and measured results for the S11 for the proposed antenna.   The simulated co-and cross-polarized radiation patterns and measured co-polarized radiation pattern of the reference and proposed antenna on E-plane, at the resonant frequency 2.44 GHz are plotted in Fig. 15(a). It shows that the antenna has a directional radiation pattern in E-plane. The cross-polarization level is more than 40 dB below than the co-polarization level. Fig. 15 VI. CONCLUSIONS This work aimed on suppression of radiated higher-order modes. To this end, a DGS with H-format was added to the ground plane of the antenna. The DGS was designed to reduce the large inductive reactance of the reference antenna and its insertion produced a reduction of 50 % in that value. There was a suppression of almost all higher-order modes within the frequency range of interest, considering excitations below S11 of 10 dB. The antenna parameters at the frequency of interest, 2.45 GHz, remained unchanged, with an improvement of others. It was possible because we optimize the position of the DGS, using the ANSYS HFSS optimization tool. The proposed antenna preserved the frequency fundamentally and limited interference problems without the need to couple filters to the communication system. It is suitable for applications in compact and low-cost communication systems.