I. INTRODUCTION
A microstrip patch antenna consists of a thin metal layer etched onto the top surface of a dielectric material with a grounded bottom [1–3]. Microstrip antennas offer advantages such as easy production, low cost, high gain, compact size, low radar cross-section, easy compatibility with integrated circuits, and the ability to be engraved onto a surface with both active and passive circuit elements [4, 5]. Meanwhile, their disadvantages include narrow bandwidth (usually less than 5%), low efficiency, poor polarization purity, limited power capacity, and tolerance issues [6, 7]. These inherent disadvantages often limit their application in a wide range of mobile and wireless communication domains. Several enhancement methods have been proposed in the literature to improve the bandwidth and gain properties of microstrip antennas [8]. The use of slotted structures is one such method [9]. When etched on the microstrip antenna surface, slot structures play an important role in increasing the antenna bandwidth [10, 11]. The bandwidth of an antenna can also be increased by changing its feeding technique [12]. Some examples of feeding techniques include proximity-coupled feed [13], aperture-coupled feed [14], and probe feed [15, 16], among others. Another common method employed to increase bandwidth and gain is the use of defected ground structures (DGS) in microstrip antenna design [17–19]. DGS structures are realized by introducing symmetrical and asymmetrical defects into the ground plane of the microstrip antenna structure, as a result of which mutual coupling and unwanted responses are reduced, thereby improving antenna radiation parameters [20–23]. The use of parasitic layers [24] and partially reflective surfaces [25–27] are other alternative methods reported in the literature to improve antenna parameters. In these methods, an additional layer containing metal elements is placed under or on top of the additional dielectric substrate to improve the antenna parameters [28–30]. Moreover, increasing the thickness of dielectric material is one of the simplest and most effective methods of increasing bandwidth and gain, as frequently reported in the literature [31]. Notably, the thickness of dielectric material can be increased by either changing the thickness of a single dielectric material [32, 33], using more than one dielectric material [34–36], or using an air gap between layers [37–41]. Along these lines, [42] adopted a multilayer structure, using a parasitic element layer to generate a notched band and tooth-shaped slots in the feed line to achieve impedance matching. However, while the parasitic structure employed in the study increased the gain at high frequencies, the opposite was observed at low frequencies. Meanwhile, Feng et al. [43] proposed a three-layer stacked microstrip antenna. In this multilayered structure design, two sets of 2 × 2 patch antenna arrays were printed on the second and top layer substrates to create lower and upper frequency bandwidths for 5G micro cell communication, respectively, thereby achieving high gain and good isolation. Furthermore, a dual-polarized frequency-reconfigurable antenna for 5G applications was proposed in [44], with the on–off function of PIN diodes utilized for frequency reconfiguration. In this multilayered dual-band antenna, good impedance matching was achieved in the bands using an improved U-shaped feed structure placed vertically between the layers, while an artificial magnetic conductor structure was employed to improve bandwidth performance. In the studies cited in [45] and [46], the antenna designs comprise a five-layer structure based on a substrate integrated suspended line (SISL). In these two studies, U-slot, L-strip, and SISL structures were employed to increase the number of bands. Meanwhile, a dual-band stacked patch antenna design was studied in [47] to achieve a wide E-plane beam width and stable gain. This antenna was first designed as a planar self-balanced magnetic dipole antenna with a semicircular patch stacked on it to create dual bands. Afterwards, another semicircular patch and symmetrical rectangular slits were added on top of the design to excite the high-order resonant mode for broadband operation. Notably, this design used an air substrate. Meanwhile, Zhou et al. [48] proposed a triband dual-polarized antenna comprising four layers—a lower band antenna, mid and upper band antennas, and a reflector layer. In [49], a dual-band dual-polarized slotted patch antenna with two layers was fabricated, with a 45° rotated square patch antenna used for the lower resonance and four trapezoidal slots employed for the upper resonance.
In this paper, a compact multilayer multiband microstrip U-slot patch antenna design that ensures high gain is presented. High gain is especially important for outdoor Internet-of-Things (IoT) applications—it not only facilitates long-distance outdoor IoT communication but also improves resilience against interference by unwanted signals from unintended directions. The multiband structure of the proposed antenna enables it to operate in multiple frequency bands, thereby supporting diverse application requirements. The multiband feature is realized using a DGS structure designed on the ground surface, while the multilayer structure helps achieve high gain. PETG and FR4 are selected as antenna substrate materials due to their low cost and wide availability. PETG is also chosen for its ability to be easily fabricated as a dielectric substrate with the desired filling ratio and thickness using a 3D printer. In addition, PETG is more durable than conventional polylactic acid (PLA) printing materials. Furthermore, while U-slot configurations, DGS, and parasitic elements have individually been studied in detail in the antenna literature, the current work presents a novel integration of these techniques using 3D printing technology within a multilayered, multiband microstrip patch antenna structure, resulting in enhanced radiation and gain features. A prototype is fabricated to examine the functionality of the antenna design presented in this study. Measurement results are obtained for the prototype and then compared with the simulated results to find that both are in harmony. Overall, the designed and fabricated antenna is both compact and has the ability to operate within the frequency ranges for C/A (±1.023 MHz), L1C (±1.023 MHz), and MCode (±5.115 MHz) services of the GPS L1 band (1,575.42 MHz), the 3GPP non-terrestrial network (NTN) uplink satellite band (1,980–2,010 MHz), and the shipborne interrogator-transponder (SIT) system band (2,930–2,950 MHz). Moreover, since the presented design also covers a substantial segment of the 2.4 GHz ISM band, it is suitable for use in many real-world scenarios as well as in IoT applications.
II. ANTENNA DESIGN
The proposed antenna was first designed as a U-slot rectangular microstrip patch antenna etched on an FR4 substrate and embedded into a PETG layer to operate in the ISM (2.4–2.5 GHz) band. Notably, it did not feature a parasitic layer, reflector layer, or DGS structure. The parasitic and reflector layers were subsequently added, and geometrical optimizations were conducted by implementing the parameter sweeping method. The RF module of COMSOL Multiphysics software was employed for the design, simulation, and optimization processes. During these processes, the gain, bandwidth, and resonant frequency parameters were especially considered. The designed multilayer U-slot microstrip patch antenna without a parasitic layer, reflector layer, and DGS structure is shown in Fig. 1.
Fig. 1 shows that the designed multilayer U-slot microstrip patch antenna consists of three layers. The first and third layers are made of FR4 material with a dielectric constant of εr = 4.15 and a height of h = 1.6 mm, while the second layer—integral to improving the basic antenna parameters—is composed of PETG material with a dielectric constant of εr = 2.62 and a height of h = 3 mm. The multilayer U-slot antenna design in Fig. 1 is fed coaxially from its middle point, with the distance of the feeding point from the lower edge being xf = 16.1 mm. The design parameters of the antenna are provided in Table 1, while the S11 plot of the multilayer U-slot microstrip patch antenna design without the parasitic layer, reflector layer, and DGS structure (Ant1) is illustrated in Fig. 2.
The S11 graph in Fig. 2 clarifies that the antenna design without the parasitic and reflective layers (Ant1) achieved a center frequency of 2,454 MHz and a 5.54% impedance bandwidth (2,386–2,522 MHz). The simulated 2D and 3D radiation patterns of the design are depicted in Fig. 3, confirming that Ant1 achieved a simulated gain of 6.19 dBi at the center frequency.
After successfully realizing the antenna design shown in Fig. 1, a parasitic layer and a reflector layer were added to the design to improve the main antenna parameters, as shown in Fig. 4.
The antenna design in Fig. 4 was achieved by adding two additional layers to the antenna design presented in Fig. 1, with the aim of improving the antenna parameters. As shown in Fig. 4, these additional layers are composed of FR4 material and are supported by rods made of PETG material. A patch antenna with a width of Wpp = 42 mm and length of Lpp = 37 mm is etched on the top layer of the parasitic layer, which is supported by PETG rods of length hrod1 = 24.4 mm. While the U-slot patch width (Wp) and patch length (Lp) are increased to 42 mm and 37 mm, respectively, the slot dimensions remained the same as in the previous design. The reflective layer at the bottom is also made of FR4 material, with its width and length being Wref = 65 mm and Lref = 65 mm, respectively, supported by PETG rods of length hrod2 = 7 mm. The reflector layer has a circular slot with a radius of 4 mm at its center point to allow the coaxial antenna feed to pass through it. The S11 plot of the proposed multilayer U-slot microstrip patch antenna design with a parasitic layer and a reflector layer (Ant2) is depicted in Fig. 5.
The S11 graph in Fig. 5 shows that the antenna design consisting of parasitic and reflective layers (Ant2) attained a center frequency of 2,434 MHz and a 5.42% impedance bandwidth (2,358–2,490 MHz). The simulated 2D and 3D radiation patterns of the design are illustrated in Fig. 6, showing that Ant2 achieved a simulated gain of 8.99 dBi at the center frequency.
Finally, to increase the number of operational bands of the antenna design illustrated in Fig. 4, a dumbbell DGS structure was implemented on the ground surface of the multilayer antenna. Notably, the dumbbell DGS form was chosen over other DGS designs for its simplicity and ease of analysis [50]. The classic dumbbell served as the reference for the initial DGS design, which was then customized to achieve the final hybrid form: a spiral-head DGS structure [51]. The designed hybrid dumbbell DGS structure etched on the ground surface is shown in Fig. 7.
In the DGS structure depicted in Fig. 7, two dumbbell structures are placed symmetrically on the vertical axis so as not to affect the radiation pattern of the microstrip antenna. The width of each dumbbell is Wdb = 10 mm, its length is Ldb = 11 mm, and its thickness is ddb = 1 mm. The design parameters of the proposed DGS are provided in Table 2.
The S11 plot of the multilayer U-slot microstrip patch antenna design with parasitic and reflector layers using the DGS structure (Ant3) is traced in Fig. 8.
An examination of the S11 graph in Fig. 8 confirms that the antenna design with parasitic and reflective layers using the DGS structure (Ant3) has multiple bands. In particular, Ant3 has four operating bands (<−10 dB): 1,532–1,562 MHz, 1,970–2,000 MHz, 2,392–2,492 MHz, and 2,878–2,902 MHz.
To further investigate the effect of DGS on antenna design, we conducted various parametric studies by changing dumbbell line thickness, width, length, position, and the number of DGS elements. The dumbbell thickness (ddb) was increased iteratively from 0. 5 mm to 1.5 mm in steps of 0.25 mm, with the width (Wdb) and length (Ldb) of the dumbbell kept constant.
Fig. 9 presents the S11, realized far-field gain, and 2D radiation pattern (2,450 MHz) plots obtained for the first parametric study. Fig. 9(a) shows that the resonance frequencies of the bands, except for the third band, shifts to higher frequencies upon increasing the dumbbell thickness in steps of 0.25 mm. Furthermore, while the ISM band did not change significantly, its bandwidth varies slightly. Moreover, although the widest band is achieved at 0.75 mm, a thickness of 1 mm is preferred for the final design since the gain value is higher for 1 mm of thickness, as shown in Fig. 9(b). An additional resonance occurs at 2,330 MHz for 0.5 mm thickness, but the gain value at this thickness is lower than that at other thicknesses at this frequency. Additionally, the second resonance is barely below the −10 dB level. For these two reasons, 0.5 mm thickness is ignored for the final design.
Fig. 9(c) depicts that the lowest backlobe level is achieved when the thickness is 1 mm. Notably, although 0.5 mm thickness leads to a better backlobe level than 1 mm, it was not chosen for the final design for the reasons mentioned above.
For the second parametric study, the width (Wdb) and length (Ldb) of the dumbbell were changed for different values of dumbbell thickness (ddb). The dumbbell line thickness, width, and length were changed in steps of 0.25 mm, 1.25 mm, and 1 mm, respectively. Notably, the initial values were 0.5 mm, 7.5 mm, and 9 mm for the line thickness, width, and length of the dumbbell, respectively, while their final values were 1.5 mm, 12.5 mm, and 13 mm, respectively. Fig. 10 shows the S11, realized far-field gain, and 2D radiation pattern (2,450 MHz) plots for the second parametric study. From Fig. 10(a), it is evident that the resonance frequencies of the bands shift to lower frequencies as a result of the iterative increases. Overall, the S11 plot presented in Fig. 10(a) shows that ddb of 1 mm and 1.25 mm are the most suitable for the final design, since the four bands are properly formed at these specifications. However, as shown in Fig. 10(b), the realized gain value at 2,450 MHz for a thickness of 1.25 mm, width of 11.25 mm, and length of 12 mm is lower than those obtained for the other steps. Additionally, Fig. 10(c) clarifies that the backlobe level remains consistent across all dumbbell thicknesses, except for the 1.25 mm ddb thickness. Although the backlobe level is even lower at the 1.25 mm thickness, the main lobe level is also degraded; therefore, this thickness was not selected. Therefore, we concluded that a thickness of 1 mm, width of 10 mm, and length of 11 mm would be appropriate for the final design.
As a third parametric study, the S11, gain, and 2D radiation patterns of the antenna design were investigated by gradually changing the number of dumbbell elements. First, a design without any dumbbell elements was considered. Then, the effects of including only the upper DGS pair and only the lower DGS pair were evaluated. Finally, simulations were conducted by including both pairs of DGS elements, indicating the completed version of the design. The S11, realized gain, and 2D radiation patterns obtained for the simulations are presented in Fig. 11(a), 11(b), and 11(c), respectively. Fig. 11(a) clearly shows that the first, second, and fourth bands do not form when no DGS element is included in the antenna design.
This observation also holds true when only the upper DGS pair is present. In the case where only the lower DGS element pair is added to the design, only the first and third bands are formed. Finally, when the two DGS element pairs are added to the antenna design, as proposed in the completed design, all four bands are properly formed. The gain plots in Fig. 11(b) show that the structure comprising the two DGS elements achieved the highest gain values across all bands. Furthermore, it is evident from the 2D radiation pattern plots at 2,450 MHz in Fig. 11(c) that almost the same mainlobe levels were achieved for all stages. Although the lowest backlobe was obtained when using only the lower DGS pair, since the second and fourth bands could not be achieved in this setup, the design chosen for this work included both DGS pairs.
As the final parametric study, the S11, realized gain, and 2D radiation patterns of the antenna design were analyzed by varying the horizontal and vertical distances between the dumbbell elements step by step. In Fig. 12, the horizontal distance between the DGS elements is denoted by ddx1, while the vertical distance between the DGS elements and the feed point is expressed as ddyf. To analyze the effect of position shifts of the DGS elements, +2 mm, +4 mm, −2 mm, and −4 mm were added to the original values of ddx1 and ddyf one by one. The corresponding S11, realized gain, and 2D radiation results were then compared with those of the design proposed in this study. Fig. 12(a) shows that the first, second, and fourth bands are not formed when the horizontal distance between the DGS elements are gradually changed. Similarly, when the distance of the DGS elements from the feed point is varied step by step, the four desired bands are not properly achieved.
Fig. 12(b) presents the realized far-field gain plots, indicating that the proposed design exhibits the highest gain values at all resonance frequencies. Finally, it is observed from Fig. 12(c) that when the vertical distance is −4 mm, the backlobe is smaller than at the other positions. However, under this condition, the four desired bands are not properly formed. As a result, it was concluded that the original DGS element positions of the proposed design were optimum in terms of the number of bands, gain, and mainlobe level.
III. RESULTS AND COMPARISONS
1. Simulated and Measured Results
The designs of the multilayer U-slot microstrip antenna without the parasitic and reflective layers (Ant1), upon the addition of the parasitic and reflective layers (Ant2), and when using the DGS structure in addition to the parasitic and reflective layers (Ant3) were simulated. In the simulations, the range between 1,000 MHz and 3,000 MHz was scanned in steps of 2 MHz. S11 plots of the antenna designs are shown in Fig. 13.
Upon examining the S11 graphs in Fig. 13, a comparison of the antenna design without a parasitic and reflective layer (Ant1; center frequency fant1 = 2,454 MHz) and the antenna design with the parasitic and reflective layers (Ant2; center frequency fant2 = 2,432 MHz) show that the center frequency shifts toward the left, but the bandwidth does not change. Furthermore, when the S11 output of the antenna design with the parasitic and reflective layers and the DGS structure (Ant3; center frequency fant3 = 2,448 MHz) is compared to that of the other designs, it is clear that the number of operating bands of the antenna increases to four in the case of the former. To further demonstrate the efficiency of the designs, the simulated realized far-field gain graphs of Ant1, Ant2, and Ant3 are presented in Fig. 14.
Fig. 14 confirms that the gain value of the final antenna design (Ant3), which consists of the parasitic layer, reflective layer, and DGS structure, increased compared to the first design (Ant1). The simulated radiation efficiency plot of the final antenna design (Ant3) is depicted in Fig. 15.
In Fig. 15, it is observed that the multilayer U-slot microstrip patch antenna design with a parasitic layer, reflective layer, and DGS structure achieved a radiation efficiency of 0.968, implying 96.8% radiation efficiency at the center frequency of 2,448 MHz. Notably, the radiation efficiency of a technically perfect radiating antenna should be 100% under ideal conditions. A radiation efficiency of 100% means that the antenna gain is equal to its directivity—that is, all the power transmitted to the antenna will be radiated into space, and no loss will occur. As a matter of fact, such an antenna design and fabrication is impossible to achieve in real life, since real-life materials always involve some loss. Therefore, considering that 100% radiation efficiency cannot be achieved, the 96.8% achieved by the designed antenna is considered to be good radiation efficiency.
Next, the designed multilayer U-slot microstrip antenna was fabricated, as presented in Fig. 16. We used the Rohde & Schwarz ZNLE3 vector network analyzer (100 kHz–3 GHz) to measure the S11 parameters of the fabricated antenna. We also employed a radiation measurement system consisting of a transmitter that included an ADF4351-based wideband signal synthesizer and a receiver unit using an AD8318 logarithmic RF detector both of which were managed by a microcontroller. A photo of the radiation pattern measuremesnt system is presented in Fig. 17, while the simulated and measured S11 plots are depicted in Fig. 18.
As shown in Fig. 18, the fabricated antenna attained four bands, with the center frequencies at 1,588 MHz, 1,990 MHz, 2,469 MHz, and 2,948 MHz (<−10 dB), respectively. The designed and fabricated antenna is compact and exhibits the ability to operate within frequency ranges pertaining to different standards. For instance, the first band (1,571–1,602 MHz) can be used for the C/A (±1.023 MHz), L1C (±1.023 MHz), and MCode (±5.115 MHz) services of the GPS L1 band at 1,575.42 MHz center frequency. The second band (1,959–2,027 MHz) supports the NTN uplink satellite band in 3GPP (1,980–2,010 MHz). Meanwhile, the fourth band of the proposed antenna is suitable for SIT systems (2,930–2,950 MHz). The simulation results also show that the proposed antenna covers the 2.4 GHz ISM band (2,400–2,483.5 MHz). However, the measurement results differ slightly—the simulated bandwidth ranges from 2,396–2,492 MHz (3.92%), while the measured bandwidth covers 2,421–2,506 MHz (3.44%). This difference can be attributed to imperfections in the manufacturing and measurement processes. Better results might have been achieved under more precise fabrication conditions. The simulated and measured 2D radiation pattern diagrams are illustrated in Fig. 19.
Fig. 19 clarifies that the measurement and simulation results agree well. Notably, imperfect fabrication and measurement conditions might have caused the differences between the simulation and measurement results at 150°–240° in the normalized antenna gain plot in Fig. 19. Some examples of imperfect fabrication include soldering errors on the antenna surface and unavoidable air filling between layers. A simulated maximum gain of 9.27 dBi was achieved.
Overall, the antenna design proposed in this study combines the complementary advantages of the U-slot with those of multilayer and parasitic structures to produce a collective effect that enhances overall antenna performance. This integration leads to a relatively compact antenna structure that achieves multiple resonant bands, along with improved gain and radiation efficiency. Additionally, the use of the 3D printing technique for design and manufacturing enables custom control over the width, length, thickness, and infill ratio of the substrate, allowing for optimized performance. 3D printing also offers a more cost-effective and efficient method of substrate fabrication. Furthermore, since the antenna design presented in this study operates within the ISM band and features a compact structure, it is well suited for IoT applications. Its compact size allows easy integration into various IoT devices.
2. Comparisons
Table 3 presents a comparison of the antenna performance parameters of the proposed antenna design with those of previous designs reported in the literature. Notably, the electrical dimensions mentioned in the comparison table were calculated based on the ground plane, which represents the largest dimension of the antennas proposed in the considered works, and height to ensure accurate comparison. Furthermore, λ0 was used to define the electrical size, denoting the free-space wavelength in the starting frequency.
As presented in Table 3, despite the impedance bandwidth of the antennas proposed in [42, 43] being greater than that of the proposed antenna, their number of bands is lower. Moreover, our work achieved better peak gain than [42], although that realized in [43] is slightly larger. The proposed antenna design also exhibits a lower ground plane electrical size compared to the two previous works. Furthermore, the maximum radiation efficiency of the proposed antenna is slightly better than that of [43], whereas the radiation efficiency of [42] was not given. The study presented in [44] has a wider bandwidth than our design, but its gain is lower than ours. Moreover, our work achieved four bands and a more compact ground plane electrical size than [44], whose antenna is also practically more difficult to realize than the one proposed in our study. The antenna designed in this study is also better than those in [45] and [46] in terms of both ground plane electrical size and the number of bands. The maximum gain obtained in [45] is smaller than that achieved in our study, while that obtained in [46] is slightly larger than in our study. Nonetheless, the structures of both these designs [45, 46] are difficult to fabricate. Upon comparing the proposed antenna with that reported in [47], it is observed that the gain value achieved in [47] is lower than that of our design. Additionally, our work has a smaller ground plane electrical size and a higher number of bands than those in [47]. Finally, although the antennas presented in [48] and [49] have wider impedance bandwidths, our work achieved a larger number of bands and a smaller ground plane electrical size. Lastly, while the antenna reported in [48] has a slightly larger peak gain value, our work attained a higher peak gain value than [49].
In this work, we employed different techniques to enhance gain and ensure that the antenna supported multiple bands. The results confirmed that these techniques fit the purpose of the study. However, we also acknowledge that these approaches may increase the height of the overall antenna design and reduce the bandwidth of the primary design.
IV. CONCLUSION
In this paper, the design of a compact multiband multilayer high-gain microstrip patch antenna for IoT applications is proposed. We established a clear analytical framework for the selection and optimization of the design parameters, ensuring that the interaction between the U-slot, DGS, and parasitic elements was fully utilized. The performance enhancements resulting from this methodical approach were validated through simulations. The proposed antenna design successfully achieved four operating bands: 1,588 MHz, 1,990 MHz, 2,469 MHz, and 2,948 MHz. After the proposed design was fabricated, its simulation and measurement results were compared. A comparison of the S11 measurement and simulation results showed slight shifts in frequency between the two, which can be attributed to possible fabrication errors. A simulated maximum gain of 9.27 dBi and a maximum radiation efficiency of 96.8% were achieved using the antenna design. Overall, the proposed antenna design demonstrated several improvements, including higher gain, better radiation efficiency, and a larger number of bands compared to conventional designs. The proposed antenna was also compared with other designs reported in the literature to verify its competitiveness. According to the measurement results, the proposed antenna can operate in the ISM band and, therefore, can be used in IoT applications.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
