I. Introduction
High data rates, large channel capacities, and frequency spectrum sharing capabilities require modern communication systems to have wide frequency bandwidths [1, 2]. In several countries, a sub-6 GHz (0.45–6 GHz) band has been designated as the primary frequency band for 5G communication, and the swift evolution of 5G communication means that base station antennas must have diverse capabilities, including wide bandwidth, high gain, and low weight.
The use of dipole antennas is a prevalent practice to achieve the required performance in terms of wide bandwidth and high gain. However, dipole antennas are problematic due to their intricate structures and feeding mechanisms. Moreover, achieving unidirectional radiation patterns in dipole antennas involves incorporating reflectors at a distance of around 0.25λ which results in high-profile structures [3–5]. Owing to their uncomplicated structures, low profiles, and ease of fabrication, microstrip patch antennas (MPAs) and shorted microstrip patch antennas (SMPAs) are systems [6, 7].
However, due to their single resonance operation with a high Q-factor, the impedance bandwidth of conventional MPAs and SMPAs is limited to 2%–3% [8, 9]. In recent decades, many researchers have tried to resolve this issue and extend the frequency bandwidth of these antennas. Reshaping the feeding structure to excite an additional nonradiative resonant mode close to the fundamental radiative mode of MPAs and SMPAs is the usual approach for enhancing frequency bandwidths [10, 11]. Two studies [12, 13] employed reformed feeds to achieve frequency bandwidths of 14% and 42%, respectively. However, this method usually demands multi-layer substrates, which, in conjunction with meandering probe feeds, makes implementation complicated and destroys the intrinsically lightweight and low-cost properties of planar antennas.
Inserting shorting posts and cutting slots on MPAs and SMPAs to relocate radiative resonant modes in proximity to each other is another technique for broadening frequency bandwidths. This method was applied in studies [14], [15], [16], and [17] to achieve frequency bandwidths greater than 6%, 11%, 15%, and 26%, respectively. However, these frequency bandwidths are not suitable for many wireless applications, and the excitation of radiative modes close to each other usually requires the antennas to be longer or wider than conventional ones.
A promising approach for enlarging the frequency bandwidths of planar antennas is to mount parasitic elements near the main radiators for multi-resonance operation [18–21]. In studies [18] and [19], an extra patch was placed vertically above the original radiating patch to generate a frequency bandwidth of around 20%. For this method, the parasitic patch is usually electrically disconnected from the probe feed and electromagnetically coupled to the core radiator. Therefore, original and parasitic patches must be printed on different substrates, leading to heavy structures, high costs, and alignment complications. In addition, to avoid dielectric loss and boost resistance to environmental conditions, all-metal antennas are preferred [22, 23]. In study [24], a stacked SMPA was proposed in which the parasitic patch was directly shorted to the main patch via a metallic wall to obtain a frequency bandwidth of up to 10% with a maximum gain of 4.9 dB. This structure was reformed in study [25] to combine a conventional SMPA with a conventional MPA to realize an all-metal antenna, improve the maximum gain, and achieve frequency bandwidths of around 9.6 dB and 26%. However, despite the full-metallic property of this antenna, the frequency bandwidth makes this structure impractical for modern communication systems.
In this study, we developed a novel wide-band patch antenna. All parts of the antenna were made of metal to counteract severe environmental conditions. In addition to enhanced robustness, the avoidance of dielectric power loss improved total efficiency across the operating frequency bandwidth by more than 90%. The maximum achievable boresight gain was 10.5 dB. The proposed all-metal antenna, which combines an MPA and three SMPAs, has a simple structure that can be easily fabricated by folding a single metallic patch. Furthermore, a rectangular metallic strip is inserted between the probe feed and the patch radiator to improve the impedance frequency bandwidth. A full-wave simulation was conducted to validate antenna performance, and a prototype was also fabricated and evaluated. The measurement and simulation results agreed, demonstrating an impedance frequency bandwidth of 67% (1.6–3.2 GHz) with a normal radiation pattern across the operating frequency band. The wide frequency bandwidth, ease of fabrication, and simple structure make this all-metal antenna suitable for wideband communication systems.
The rest of this paper is organized as follows. In Section II, the simulation results for the proposed all-metal patch antenna are discussed, as well as a parametric study to clarify the working mechanism of the proposed structure. The fabrication and testing of a prototype are explained in Section III, and the conclusion is presented in Section IV.
II. Antenna Design and Analysis
1. Antenna Geometry
The three-dimensional (3D) geometry and schematic of the proposed antenna, consisting mainly of four radiators, are depicted in Fig. 1, and the optimized dimensions are listed in Table 1. The first radiator is a suspended rectangular patch (A) with length LA, width w, and bearing height hA above the ground plane with a dimension of Lg × Wg. This component, which can be considered an MPA, is designed to resonate at a frequency of 2.45 GHz. For excitation, a vertical rectangular metallic sheet, marked as the “matching part” in Fig. 1, with width WF and length LF, is connected to an SMA connector at the bottom. The upper patches, marked as B, C, and D, are approximately half the length of A and are shorted by vertical connecting walls E, F, and G, respectively. These three radiators function as SMPAs and provide three additional attenuation poles. Since all units have equal width and approximately equal resonant length, their bearing heights (hE, hF, and hG) can be adjusted to tune resonance frequencies. Increasing the height of a patch antenna yields a larger fringing area and, subsequently, a lower resonant frequency. Therefore, according to the bearing height of the radiators, four in-band attenuation poles are sequentially attributed to B, A, C, and D.
2. Evolution of the Proposed Structure
To clarify the principle of broadening the frequency bandwidth, we studied the evolution of the proposed antenna. The evolution and corresponding simulation results using the Computer Simulation Technology (CST), developed by Dassault Systèmes Simulia company and based on the finite integration technique, are presented in Figs. 2 and 3, respectively. As previously stated, a single MPA provides a narrow frequency bandwidth. However, increasing the height of the antenna and using air as a substrate lowers the resonant frequency and extends the frequency bandwidth of MPAs [26]. Fig. 2(a) illustrates a probe feed rectangular patch with a length of 54 mm and a width of 50 mm suspended at a height of 7.2 mm above a ground plane with a dimension of 115 mm × 160 mm. The antenna resonated at 2.45 GHz, and an 8% frequency bandwidth was obtained (Fig. 3). To improve bandwidth performance, a parasitic square patch with a length of 48 mm (half the bearing height of the main patch) was placed vertically above the main patch at a height of 3.6 mm (Fig. 2(b)). The parasitic patch, which was electromagnetically coupled to the main radiator, added another radiative resonance at 2.75 GHz, and the frequency bandwidth widened to around 17%. The parasitic patch was then split roughly in the middle and shortened to the main patch at both ends with different heights (Fig. 2(c)). Therefore, the parasitic patch was transformed into two parasitic SMPAs to generate two resonances at 1.73 GHz and 3.09 GHz. It should be noted that these two parasitic radiators used the main patch as a shared ground plane. As described earlier, a greater height yields a larger fringing area and a consequently lower resonant frequency for the patch antenna. Hence, resonances at 1.73 GHz and 3.09 GHz were due to the right and left parasitic SMPA, respectively. The reflection coefficient for the antenna was higher than −10 dB for most of the operating frequencies. To mitigate this issue, a rectangular strip with a dimension of 6.7 mm × 15 mm was inserted between the probe feed and the antenna (Fig. 2(d)) as a matching network. Fig. 3 illustrates that an impedance frequency bandwidth of 63% was obtained for S11 ≤ −10 dB from 1.62–3.11 GHz. It should be noted that radiators are loaded with additional current paths to shift resonant frequencies. Finally, another patch was shorted to the left parasitic SMPA with a bearing height of 0.8 mm to realize another resonance at 3.17 GHz (Fig. 2(e)). It is worth mentioning that four resonances were in close proximity, and a wide frequency bandwidth of about 70% was obtained, with a range of 1.56–3.24 GHz. The observation that lower resonant frequencies are attributed to the patches with greater heights, can be further explored by analyzing the surface current distribution on the antenna at its resonant frequencies. The surface current distribution at four resonances, as depicted in Fig. 4, revealed concentrated currents on B, A, C, and D at frequencies of 1.73 GHz, 2.2 GHz, 3.09 GHz, and 3.17 GHz, respectively. These findings validate the origins of these resonances. The simulated peak gain, directivity, and total efficiency, which varied with frequency, are depicted in Fig. 5. The antenna’s gain value at each resonance exceeded that at the preceding resonance, reaching a peak of 10.5 dB at the fourth resonance. All resonators had approximately the same physical area but exhibited distinct resonant frequencies. Consequently, the electrical area of the resonators was larger for higher resonant frequencies, leading to a larger effective area and a higher gain. Due to the all-metal property of the proposed structure, power loss is reduced compared with a conventional patch antenna with a dielectric substrate and consequently, total efficiency remain higher than 90% over the operating frequency band. To provide further elucidation, the simulated 3D gain patterns of the proposed antenna at four resonant frequencies are depicted in Fig. 6. Clearly, the antenna exhibited stable unidirectional radiation patterns across the frequency band.
3. Parametric Study
We conducted a parametric study to confirm the origin of each resonance. The lengths of the four radiators (LA, LB, LC, and LD) were varied to evaluate the performance of the proposed structure. Fig. 7 shows the effect of these parameters on the reflection coefficients. It should be noted that other parameters were kept as optimum values while each parameter was studied. The resonant frequencies of these four resonances are referenced as f1, f2, f3, and f4.
Fig. 7(a) clearly shows that f2 shifted downward when the length of the main radiator LA was increased, and vice versa. Thus, the second resonance was related to the main radiator. Since A was a shared finite ground plane for B and C, f1 and f3 were also expected to be affected by this parameter. The frequency shift was more insignificant for f1, which can be attributed to the higher bearing height of B. Fig. 7(b) shows that a significant effect on the first resonance was observed when the length of the right parasitic SMPA was changed. Frequency f1 decreased with increasing LB, but other resonances were hardly affected. Therefore, it can be inferred that the first resonance arose from B. Although it was expected that C would resonate at the third resonance frequency, Fig. 7(c) reveals that LC also had a significant effect on the fourth resonance. Indeed, C was a finite ground plane on which D was shorted with a height of 0.8 mm. Therefore, increasing LC led to a downward shift of both f3 and f4. Finally, Fig. 7(d) displays the variation in the simulated reflection coefficient relative to the length of D. In contrast to f1, f2, and f3 which were influenced slightly, f4 was shifted significantly by LD, suggesting that fourth resonance is contributed by D.
In conclusion, a wideband characteristic was achieved due to the proximities of the four resonances sequentially provided by B, A, C, and D.
III Experimental Verification
To validate the working principle and the predicted wideband performance, we fabricated a handmade prototype of the proposed antenna. A thin 0.25-mm-thick copper strip placed above a 3-mm-thick aluminum ground plane was folded to produce the all-metal structure, as shown in Fig. 8. The input reflection coefficient was measured using a vector network analyzer manufactured by Agilent Technology Incorporation. The simulation and measurement of S11 as a function of frequency are illustrated in Fig. 9. There was a small discrepancy between the measurement and simulation results attributable to the tolerances arising from the handmade fabrication process. Nevertheless, the measurement results agreed well with the simulation results. The measured impedance frequency bandwidth for S11 ≤ −10 dB was about 67%, with a range of 1.6–3.2 GHz. Four in-band resonances obviously occurred across the frequency band, confirming the design principle.
An anechoic chamber developed at Khajeh Nasir Toosi University of Technology was employed to evaluate the radiation characteristics. The simulated and measured normalized radiation patterns near the four resonant frequencies are presented in Fig. 10. Good agreement between the measurement and simulation results was achieved.
The radiation patterns were stable and unidirectional across the operating band, although a little asymmetry was observed at the Eplane, especially at lower frequencies due to the asymmetric structure of this plane (x-direction), H-plane radiation patterns were completely symmetrical across the operating band, as expected.
Moreover, the achievable boresight E-plane and H-plane cross-polarizations were lower than −20 dB. Since the simulated E-plane cross-polarizations were lower than −45 dB, they were not incorporated into the curves. The measured boresight gains according to the simulation data are depicted in Fig. 11.
The measurement results agreed well with the simulation results, showing that a maximum boresight gain of 10 dB was achieved at 3.1 GHz. To confirm the usefulness of the proposed structure, it was compared with earlier variants of multilayered structures, as shown in Table 2.
Although the antenna proposed in [24] is compact, its narrow frequency bandwidth makes it impractical for wide-band applications. In addition, it uses a dielectric substrate, which results in high power loss and, consequently, low gain and low efficiency. The proposed structure has an approximately comparable size to the antennas presented in other works, but it is obviously superior in terms of frequency bandwidth and gain.
IV. Conclusion
We propose a novel wide-band all-metal antenna in this paper. The antenna has a simple structure that can be easily fabricated by folding a single metallic strip. The proposed antenna can be viewed as an integration of an MPA and three SMPAs leading to quad resonance operation. A metallic strip is placed between the antenna and probe excitation as a matching network to improve input S11. A prototype was fabricated and evaluated to validate the design procedure. The measured impedance frequency bandwidth was 67% for S11 ≤ −10 dB from 1.6 to 3.2 GHz. A maximum peak gain of up to 10 dB was achieved. The E- and H-plane radiation patterns were generally symmetrical and unidirectional across the frequency band. Due to its simple structure, light weight, low cost, and wide frequency bandwidth, the proposed antenna is suitable for applications to modern communication systems.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
