Three-Dimensionally Printed Dual-Slot-Fed Dielectric Resonator Antenna with Rectangular and Irregular Elements for 5G Applications

Abstract

In this paper, a novel dual-slot-fed dielectric resonator antenna (DRA) with rectangular and irregular elements, designed for 5G wireless applications, is presented. The DRA achieves wideband capability by combining the resonant modes of the rectangular and irregular DRA elements, which is a less common feature in conventional designs. A frequency ratio adjustment technique, based on the concept of inductive de-loading, is uniquely proposed for the independent frequency adjustment of the irregular DRA. Unlike traditional methods, an equivalent circuit presentation was developed to interpret the impedance characteristics of single-element DRAs, and to provide new insights into the presence of inductive de-loading from a circuit perspective. For verification, a dual-slot-fed prototype was fabricated through digital light processing (DLP)-based 3D printing technology, with the aim of customizable design and low-cost fabrication. The measured and simulated results of reflection coefficients and radiation patterns showed good agreements, with a measured bandwidth of 51.6% (2.96–5.02 GHz), effectively covering the desired 5G n77–n79 (3.3–5.0 GHz) frequency bands.

1. Introduction

Over the last three decades, there has been a growing interest in the study of dielectric resonator antennas (DRAs) for wireless applications. DRAs provide several advantages over traditional conductive antennas, such as higher efficiency, greater design flexibility, and benefits of radiating through the 3D structure [1]. The demand for wideband antennas has prompted exploration into various bandwidth enhancement methods, including stacked DRAs with varied permittivity [2,3,4], and modified DR shapes, such as a bowtie shape [5], stair shape [6,7], T shape [8], and trapezoidal configuration [9]. On the other hand, multi-band antenna design techniques with adjustable frequency ratios have shown significant potential for bandwidth enhancement by merging adjacent modes. Researchers have explored methods like using anisotropic materials [10], additional resonators [11,12,13], and loading elements on DRAs [14] for this aim. Shape deformation of DRAs [5], with the utilization of differences in field patterns between modes [15], is a more straightforward method to adjust frequency ratios.

Traditional manufacturing of DRAs with complex feeding methods and irregular shapes faces great challenges, particularly a high cost due to precision machining [16,17]. Recently, 3D printing, or additive manufacturing (AF), has revolutionized DRA fabrication by enabling the realization of complex structures with controllable permittivity. Solid-based techniques such as Fused Deposition Modelling (FDM), and slurry-based techniques like stereolithography (SLA) and digital light processing (DLP), are popular in ceramic and DRA fabrications [18]. The FDM technique has gained widespread popularity for DRA fabrications. Permittivity manipulation can be realized through subtractive dielectric tuning methods in FDM, such as changing the infill pattern and percentage of the filaments [19,20], or through additive dielectric tuning, which involves mixing high-permittivity dielectric materials into the filament [21].

Nevertheless, only limited studies have been conducted on slurry-based 3D-printed DRAs, due to the relatively low permittivity of the resin used. In [22], a 3D-printed super-shaped DRA operating at the 3.5 GHz band was proposed. However, the measured dielectric constant was approximately 2.7 because of the use of photosensitive resin with a low permittivity during the printing, without mixing in any high-permittivity materials.

The integration of UV-curable resin with high-permittivity materials like zirconia [23], MaTiO₃-xCaTiO₃ [24], and BaTiO₃ [25] has provided a revolutionary approach for DRA fabrication. These studies explore the process of slurry-based ceramic fabrication, enabling the creation of ceramics with desired material properties. One remarkable application of slurry-based 3D printing is the proposal of a 3D-printed Luneburg lens [26]. By using a 75 wt% 40LCBS3/60Al₂O₃ ceramic powder, the dielectric constant of the lens reaches 6.64 after sintering. Although dielectric materials with high permittivity are desired for their ability to enable size reduction, they often require sintering processes or higher concentrations of dielectric materials, and therefore significantly increase fabrication costs and printing complexity, along with the cost of a narrower bandwidth and limited applicability in wideband designs.

Developing further upon the work mentioned above, this paper investigates a wideband dual-slot-fed DRA with rectangular and irregular elements for 5G applications. A composite dielectric material with a moderate dielectric constant was selected and fabricated to balance trade-offs between achieving effective bandwidth and maintaining a low-cost and practical fabrication process. The wideband characteristic was realized by exciting the (quasi-)TE^y₁₁₁ and (quasi-)TE^y₂₁₁ modes of both the rectangular DRA (RDRA) and the irregular DRA, through adopting a parallel circuit design principle. Also, frequency ratio adjustment in the irregular DRA was achieved based on the concept of inductive de-loading for the TE^y₁₁₁ mode.

Additionally, equivalent circuits were developed to represent the resonances of the regular RDRA and irregular DRA. Equivalent circuit analysis is often used for metal-based structures, such as metasurfaces [27], filters [28] and metallic antennas [29]. However, the exploration of equivalent circuit presentation in DRAs is limited, especially in dual/multi-band DRAs with different feeding network configurations, like probes [30], probe and microstrip lines [31], and slots [32,33]. The equivalent circuit presentation for the slot-fed DRA in this paper helps to understand the impedance characteristics of DRAs and the concept of inductive de-loading.

After the design and simulation study, a prototype was fabricated using DLP-based 3D printing. The simulated and measured results of the reflection coefficients and radiation patterns will be compared and discussed. Further details are given in the following sections.

2. Antenna Configuration and Simulations

2.1. Configuration of Slot-Fed Single-Element DRAs

The principle of adjusting the frequency ratio in a DRA can be demonstrated as shown in the following diagram. Figure 1a and Figure 1b compare the y-directed magnetic fields of the TE^y₁₁₁ mode and the TE^y₂₁₁ mode within an RDRA. The H-fields of the TE^y₁₁₁ mode predominantly concentrate within the central region of the DRA, while the stronger portion of the TE^y₂₁₁ mode is confined to the two side semi-elliptical areas. Therefore, any modifications made to the internal fields of the distinct modes can directly influence and alter their respective resonant frequencies, due to the de-loading effect.

This observation presents the potential for mode frequency manipulation by introducing notches at the high-magnetic-field-density section of the first mode. By employing this technique, known as inductive de-loading [34], the effective inductance of the resonator is reduced by modifying the dielectric resonator’s geometry in regions of high magnetic field density. By using this technique, the resonant frequency of the first mode can be effectively increased with minor impact on the second mode. Consequently, the frequency ratio of the DRA can be precisely adjusted by deforming the DRA’s shape.

Figure 1c shows the basic slot-fed RDRA configuration. The RDRA is designed with a size of a = 30 mm, b = 14 mm, and d = 12 mm, and a dielectric constant ε_r = 8. The dielectric waveguide model (DWM) [35] can be applied to estimate the frequency of the RDRA modes. The resonant frequency, f_mnl, of the TE^y_mnl mode for an RDRA with length a, width b, height d, and dielectric constant ε_r is calculated by solving the following equations:

$$f_{mnl}={\displaystyle \frac{c}{2\pi \sqrt{{\epsilon }_r}}}\sqrt{{k_x}^2+{k_y}^2+{k_z}^2},$$

(1)

$$k_x={\displaystyle \frac{m\pi }{a}},$$

(2)

$$k_z={\displaystyle \frac{l\pi }{2d}},$$

(3)

$$k_y\tan \left({\displaystyle \frac{k_{y}b}{2}}\right)=\sqrt{{(\epsilon }_r-1){k_0}^2-{k_y}^2}$$

(4)

The subscripts m, n, and l correspond to the numbers of field variation along the x-, y- and z-directions, respectively. The parameter k₀ represents the free space wavenumber. k_x, k_y, and k_z indicate the wavenumbers in these three directions inside the RDRA. According to the DWM, the resonant frequencies of the TE^y₁₁₁ and TE^y₂₁₁ modes are estimated at 3.53 GHz and 4.82 GHz, respectively.

The RDRA was positioned on top of a ground plane, with a rectangular slot (W_slot = 2.6 mm, L_slot = 22 mm) centrally etched below the RDRA. Rogers RO4350B laminate (ε_r = 3.66, thickness H_sub = 0.508 mm) was used as the substrate (W_sub = 100 mm, L_sub = 100 mm). A 50 Ω microstrip line (W_f = 1.12 mm) was printed on the back side of the substrate, with an open stub length of s = 14 mm, to excite the TE^y₁₁₁ and TE^y₂₁₁ modes of the RDRA. The proposed RDRA was then simulated in CST Microwave Studio. Figure 2b presents the simulated reflection coefficient of the RDRA in a solid line noted as DRA-00, where 00 indicates no shape variations in both the x and y directions. The reflection coefficient response shows that the TE^y₁₁₁ and TE^y₂₁₁ modes were excited at 3.43 GHz and 4.94 GHz, respectively, which closely match the calculated resonant frequencies.

Similarly, Figure 1d shows the structure of an irregular DRA derived from that of a regular RDRA, which is noted as DRA-20, where 20 indicates two sine-like shape variations in the x direction, but no shape variation in the y direction. The sine-like cylinders have an elliptical radius R_x in the x-direction and an elliptical radius R_z in the z-direction, and a width L_y. R_z is chosen to be 12 mm, and L_y to be 14 mm, and the length of the sine-like cylinders coincides with the y-directed length of the cuboid, to ensure the overall size of the irregular DRA-20 is confined to that of DRA-00. Also, the DRA-20 has the same feeding network as the DRA-00. R_x may be varied to create different degrees of irregularity.

Figure 2a shows the variation in the irregular DRA-20 when R_x changes from 20 mm to 7.5 mm. At R_x = 20 mm, the shape approaches a cubic structure. As R_x decreases, the central notch enlarges. Finally, when R_x is reduced to 7.5 mm, the DRA-20 becomes two split sine-like cylinders, where any inappropriate assembly between the split resonator elements will lead to air perturbation and frequency shift. For R_x = 20, 12, 10, 8, 7.8, and 7.5 mm, the simulated results of the reflection coefficients are shown in Figure 2b and compared with that of the DRA-00. The resonant frequencies of the (quasi-)TE^y₁₁₁ and (quasi-)TE^y₂₁₁ modes, along with their frequency ratios, are listed in

Table 1. Simulated resonant frequencies and frequency ratios of the DRA-00 and DRA-20.

DRA	Frequency of Mode		Frequency Ratio
	(quasi-)TEy111	(quasi-)TEy211
DRA-00	3.43 GHz	4.94 GHz	1.440
DRA-20 Rx = 20 mm	3.46 GHz	4.955 GHz	1.432
DRA-20 Rx = 12 mm	3.54 GHz	4.98 GHz	1.407
DRA-20 Rx = 10 mm	3.615 GHz	5.005 GHz	1.385
DRA-20 Rx = 8 mm	3.845 GHz	5.01 GHz	1.303
DRA-20 Rx = 7.8 mm	3.9 GHz	5.005 GHz	1.283
DRA-20 Rx = 7.5 mm	4.02 GHz	5 GHz	1.244

for both DRAs. The results show a significant reduction in the frequency ratio (1.440 to 1.244) of the (quasi-)TE^y₁₁₁ and (quasi-)TE^y₂₁₁ modes through the inductive de-loading approach in the irregular DRA-20. Therefore, based on the parametric study, the irregular DRA-20 with R_x = 7.8 mm was selected as the DRA element, after making a trade-off between its relatively small frequency ratio (1.283) and ease of assembly.

For the mode identification, Figure 3 shows the simulated H_y field plots of the DRA-00 and DRA-20 with R_x = 7.8 mm in y = 0 and z = 0 planes, and the two modes can be clearly identified as (quasi-)TE^y₁₁₁ and (quasi-)TE^y₂₁₁ modes for both DRAs. It should be noted that the field plots of the two modes for the DRA-20 shown in Figure 3c,d are perturbated due to the shape deformation.

Figure 4 shows the simulated 3D radiation patterns of the TE^y₁₁₁ mode at 3.43 GHz and the TE^y₂₁₁ mode at 5.005 GHz for DRA-00, as well as the simulated 3D radiation patterns of the DRA-20 with R_x = 7.8 mm, along with its perturbed quasi-TE^y₁₁₁ mode at 3.9 GHz, and its quasi-TE^y₂₁₁ mode at 5.005 GHz. Notably, radiation patterns on a linear scale have been included in Figure 4b,d for TE^y₂₁₁ mode identification, due to their bidirectional radiation patterns. Compared to the two modes of the DRA-00, the radiation patterns of the two quasi-TE modes of the DRA-S20 show the effect of perturbation after shape deformation.

2.2. Equivalent Circuits of Slot-Fed Single-Element DRAs

The frequency shift caused by shape modification can be interpreted through an equivalent circuit. The equivalent circuit of the DRA-00 shown in Figure 1c is presented in Figure 5. In this model, a transformer is employed immediately after a 50 Ω microstrip transmission line, to account for the coupling between the slot and the DRA. To accurately quantify the resonant frequency of each mode, two RLC circuits in series are utilized. The TE^y₁₁₁ and TE^y₂₁₁ modes of DRA-00 are modelled by the resonators SRLC1 and SRLC2, respectively. A similar equivalent circuit can be applied to represent the resonances of the DRA-20, but with some different circuit parameters.

Figure 6 shows the simulated reflection coefficients of the DRA-00 and DRA-20 with R_x = 7.8 mm, obtained using CST Microwave Studio. It indicates that the DRA-00 resonates at 3.43 GHz and 4.94 GHz. The mathematical expressions for determining the equivalent values of the RLC resonator components are provided in [36]. Therefore, by observing the simulated reflection coefficients presented in Figure 2b, we can extract the values of f₀ and S11(f₀) for each mode. Subsequently, the series resistance value R, along with the values of the series inductor and capacitor, can be calculated, as listed in

Table 2. Calculated parameters of the RLC resonator circuits and corresponding frequency resonances for the two single-element DRAs.

Parameters		DRA-00		DRA-20, Rx = 7.8 mm
		SRLC1	SRLC2	SRLC1	SRLC2
RLC Values	R (Ω)	41.2953	48.9513	39.3250	94.4101
	L (nH)	14.5306	17.8877	11.5217	23.7810
	C (pF)	0.1482	0.0580	0.1445	0.0425
Freq (GHz)	CST	3.43	4.94	3.9	5.005
	ADS	3.46	4.89	3.96	4.94

. Similarly, this method can be applied to obtain the equivalent circuit parameters for the DRA-20.

The simulated reflection coefficient of the single-element DRA-00 obtained from CST Microwave Studio, and that from the Advanced Design System (ADS) 2022 simulation software based on the equivalent circuit, are compared in Figure 6. The equivalent circuit presentation of the DRA-00 exhibits resonances at 3.46 GHz and 4.89 GHz, which agrees well with the results from CST Microwave Studio. Similarly, the simulated reflection coefficients of the DRA-20 from CST Microwave Studio and that from the ADS simulation software are compared in Figure 6, and they agree well with each other, with resonances at 3.9 GHz and 5.005 GHz from CST Microwave Studio, and resonances at 3.96 GHz and 4.94 GHz from the equivalent circuit in the ADS simulation software. The comparisons in Figure 6 show that the proposed equivalent circuit presentation can accurately describe the impedance characteristics of the single-element DRAs.

It can be observed that the inductor value of the resonator SRLC1 of the DRA-20 is smaller than that of the resonator SRLC1 of the DRA-00, while the capacitor value nearly remains the same, indicating reduced magnetic energy with little effect on electric stored energy. This confirms the de-loading principle explained from the magnetic field point of view that is shown in Figure 1a for the TE^y₁₁₁ mode, and it further validates the inductive de-loading effect for the TE^y₁₁₁ mode caused by the shape deformation on the DRA from a circuit perspective. On the other hand, the shape deformation caused the inductor value of SRLC2 to increase and the capacitance value to decrease from the DRA-00 to the DRA-20, indicating an increase in magnetic stored energy and a reduction in electric stored energy.

2.3. Configuration of Dual-Slot-Fed Two-Element DRA

The configuration of the proposed dual-slot-fed DRA with rectangular and irregular elements, connected in parallel with a fork-like microstrip line feed structure, is depicted in Figure 7. The rectangular DRA-00 and the irregular DRA-20 with R_x = 7.8 mm have been described in Section 2.1. Detailed parameters are presented in

Table 3. Parameters of the wideband dual-slot-fed DRA.

Parameters	Values	Parameters	Values	Parameters	Values
a1	30 mm	Ly	14 mm	Wf1	1.12 mm
b1	22 mm	Wsub	100 mm	Wf2	0.3 mm
d	12 mm	Lsub	100 mm	L1	41 mm
εr	8	Hsub	0.508 mm	L2	15 mm
Rx	7.8 mm	Wslot	2.6 mm	Ls	9 mm
Rz	12 mm	Lslot	21 mm	s	13 mm

.

Figure 8a shows the simulated reflection coefficients of the proposed dual-slot-fed DRA. This DRA resonates at 3.43 GHz, 3.93 GHz, and 4.91 GHz, corresponding to the TE^y₁₁₁ mode of the DRA-00, the quasi-TE^y₁₁₁ mode of the DRA-20, and the (quasi-)TE^y₂₁₁ mode of both the DRA-00 and the DRA-20, respectively. The simulation results show a −10 dB bandwidth of 46.0% (3.213–5.132 GHz), covering the 3.3–5.0 GHz bands. Furthermore, the simulated reflection coefficients of the DRA-00 and DRA-20 are shown in Figure 8 for comparison. It is evident from the results that the resonances observed in the dual-slot-fed DRA are attributed to the resonances of the single-element DRAs.

To investigate the effect of R_x, the simulated reflection coefficients of the dual-slot-fed DRA for different values of R_x are plotted in Figure 8b. As R_x increases from 7.5 mm to 9 mm, the resonant frequency of the second mode gradually moves downward from 4.06 GHz to 3.74 GHz. This validates the effectiveness of the frequency ratio adjustment technique employed for multi-band DRAs.

Moreover, the simulated H_y field distributions of the proposed dual-slot DRA in the z = 0 plane, for the three modes presented in Figure 8a, are shown in Figure 9. These distributions demonstrate that the first mode is mainly from the TE^y₁₁₁ mode of the DRA-00, the second mode from the quasi-TE^y₁₁₁ mode of the DRA-20, and the third mode from the TE^y₂₁₁ mode of the DRA-00 and the quasi-TE^y₂₁₁ mode of the DRA-20.

3. Experimental Verification: Fabrication and Measurements

3.1. DLP-Based 3D Printing Process

Figure 10 shows the DLP-based 3D printing process, from ceramic slurry in Figure 10a to 3D-printed ceramic blocks in Figure 10e. The slurry consisted of 55 wt% Sigma-Aldrich BaTiO₃ (Barium Titanate) ceramic powder (Sigma Aldrich, St. Louis, MA, USA) and 45 wt% UV-curing resin, mixed at 1500 RPM for 30 min. An ANYCUBIC PHOTON D2 DLP 3D printer (ANYCUBIC, Shenzhen, China) with a 405 nm UV-LED light source was employed for printing, as shown in Figure 10b.

To effectively print the DRA elements, a transition film of 0.02 mm was first printed using a slurry of 20 wt% BaTiO₃ powder and 80 wt% UV-curing resin, as shown in Figure 10c, which exhibits remarkable strength and adhesion, to ensure secured attachments of the printed DRA element.

In contrast to the commonly used industrial top-down printers in other studies, this research employed a bottom-up desktop printer. This setup presents specific challenges related to gravity and adhesion when using the slurry-based printing material. One critical issue affecting printability that was encountered during the DLP printing process was the adhesion failure between the printed structures and the metal build platform, particularly when the concentration of the BaTiO₃ ceramic powder in the slurry exceeded 30 wt%.

Increasing the exposure time of the UV light was identified as a potential solution to improve the adhesion of the bottom layer of the model to the platform. However, this approach comes with trade-offs, as the long exposure time can negatively affect the overall printing accuracy. Additionally, there is a threshold beyond which further increments in exposure time reach their limit and no longer help adhesion enhancement.

To effectively solve this interfacial bonding problem, a ceramic–resin transition film was employed. This transition film was printed by using a slurry containing 20 wt% BaTiO₃ ceramic powder and 80 wt% UV-curing resin. Also, ensuring an appropriate thickness of this film (0.02 mm, as shown in Figure 10c) and proper levelling of the build platform was critical to avoiding further printing issues. Despite the fact that the transition film was extremely thin, it exhibited remarkable strength and adhesion, and finally enabled the further exploration of complex DRA geometries beyond regular shapes.

The DRA elements were built layer-by-layer, with each layer set to 0.02 mm for optimal accuracy in the z-direction. Figure 10d shows the final printed DRA elements attached onto the transition film. The printed DRA elements were then cured in a curing machine for better hardness, as shown in Figure 10e. The final assembled dual-slot-fed DRA is shown in Figure 10f, with the DRA elements glued to the ground plane to ensure stable attachment. In addition, the microstrip feed structure on the back side of the substrate is presented in Figure 10g.

3.2. Measured Dielectric Property and Antenna Properties

In this study, the dielectric constant of the UV-curing resin and printed DRA elements were measured experimentally using a Keysight 85070E dielectric probe kit (Keysight, Santa Rosa, CA, USA), as shown in Figure 11a. The obtained results are shown in Figure 12a, which indicates that the dielectric constant of resin is around 3 in the band of interest, 3.3–5 GHz, and the dielectric constant of the printed DRA elements shows a small fluctuation around 8.

The reflection coefficients of both the regular and irregular DRA elements were measured using an Agilent E5071B Network Analyser (Agilent, Santa Clara, CA, USA), whereas the radiation patterns of the proposed dual-slot-fed DRA were measured in an anechoic chamber, as shown in Figure 11b. Figure 12b shows the measured reflection coefficient of the printed dual-slot-fed DRA compared with the simulation. The measured −10 dB impedance bandwidth is 51.6% (2.96–5.02 GHz), agreeing well with the simulated result (46.0%). The bandwidth fully covers the desired 3.3–5.0 GHz frequency band, making it suitable for 5G applications. Additionally, the printed dual-slot-fed DRA has three resonance modes at 3.43 GHz, 3.93 GHz, and 4.83 GHz, respectively, and these three modes merge together to form a wideband response. The first two modes closely match the simulation results, while the third mode shows a slight 1.6% difference. Due to a small air gap between the DRAs and the ground plane, introduced during the assembly process, the effective permittivity and electromagnetic boundary conditions were affected by this air gap, particularly for higher-order modes. Combined with minor dimensional inaccuracies and material property variations during the printing process, these factors account for the observed 1.6% frequency deviation. The measured and simulated radiation patterns shown in Figure 13 exhibit a high degree of agreement as well. The slight difference observed can be attributed to fabrication tolerance, difference in material properties, and test environment. On the whole, the measured and simulated responses agree well.

Recent advancements in DRA designs have explored various shapes, dielectric constants, and techniques to achieve wideband performance.

Table 4. Comparison between the proposed and existing wideband DRAs.

Ref.	Freq (GHz)	εr	Antenna Type	Bandwidth
[6]	4.1–6.81	12.8	Rectangular and half-cylindrical DRA with stair-shaped slot	49.67%
[37]	3.08–4.16	9.5	Wheel-shaped DRA	30%
[38]	4.88–6.88	6.85	Cylindrical DRA with arced aperture	34%
[39]	3.18–4.91	9.8	Droplet-shaped bowtie DRA with balun-fed aperture	42.7%
[40]	8.5–12	20	Cylindrical DRA	34%
This paper	2.96–5.02	8	RDRA and irregular DRA with dual slots	51.6%

compares the performance of the proposed 3D-printed dual-slot-fed DRA with that of other wideband DRAs available in the literature. In [37], a wheel-shaped DRA with a dielectric constant of 9.5 was employed, and multi-mode resonance was excited, to obtain a 30% bandwidth (3.08–4.16 GHz). Varshney et al. designed a stair-shaped slot-excited DRA, combining rectangular and two half-cylindrical shapes with a dielectric constant of 12.8, for C-band applications, and an impedance bandwidth of 49.67% was achieved [6]. In [38], a low-profile cylindrical DRA with an arced-aperture feed and a dielectric constant of 9.8 was designed, and an impedance bandwidth of 31.8% was achieved. Furthermore, Dwivedy proposed a droplet-shaped bow-tie DRA with a dielectric constant of 9.8, achieving a bandwidth of 42.7% (3.18–4.91 GHz) for sub-6G and IoT applications [39]. To provide a wide bandwidth of 34% and circular polarization for X-band radar applications, Rad et al. designed a stair-shaped DRA with a high dielectric constant of 20 [40]. Compared to these designs, this work introduces a novel dual-slot-fed DRA with a moderate permittivity of 8 that combines resonant modes from rectangular and irregular elements, which achieved the widest impedance bandwidth of 51.6% (2.96–5.02 GHz). The proposed design uniquely employs the frequency ratio adjustment technique via inductive de-loading, and its significant enhancement demonstrates the effectiveness of using 3D printing to fabricate innovative DRA geometries with flexible dielectric materials.

4. Conclusions

In conclusion, this research presents a novel wideband 3D-printed dual-slot-fed DRA, formed using a rectangular DRA element and an irregular DRA element. The performance of the proposed DRA has been thoroughly analyzed through simulation and measurement. The DRA employs a microstrip feed line with two branches in parallel and two rectangular slots placed beneath the DRA elements. Notably, the first mode of the irregular DRA element could be independently adjusted through the frequency ratio adjustment technique, based on the concept of inductive de-loading. The inductor value of the first mode was reduced when the resonator shape was deformed from DRA-00 to DRA-20, with its capacitor value remaining almost the same, while the shape deformations led to an increase in inductor value and a decrease in capacitance value for the second mode, but its resonance frequency was nearly un-changed.

Through shape deformation by decreasing the x-directed elliptical radius R_x, a significant reduction in the frequency ratio could be observed from 1.440 to 1.244 in simulation, exhibiting the great potential of shape deformation for mode interaction and bandwidth enhancement. By merging the modes of the DRA-00 and DRA-20, a wide bandwidth was achieved, and the simulation results showed a −10 dB impedance bandwidth of 46.0% (3.213–5.132 GHz), which could effectively cover the n77–n79 (3.3–5.0 GHz) bands.

The fabricated prototype of the proposed dual-slot-fed DRA using a DLP-based 3D printing method shows the potential of the additive manufacturing technique for realizing complex and irregular antenna shapes. By using the chosen ceramic slurry, consisting of 55 wt% BaTiO₃ ceramic powder and 45 wt% UV-curing resin, DRA elements with a specific dielectric constant and dimensions can be printed. Furthermore, the measured and simulated reflection coefficients and radiation patterns show good agreements. The measured dual-slot-fed DRA presented resonant modes at 3.43 GHz, 3.93 GHz, and 4.83 GHz, and these three modes merged together to realize a wide impedance bandwidth. The same was observed in simulation. The measured −10 dB impedance bandwidth of 51.6% (2.96–5.02 GHz) can effectively cover the desired n77–n79 (3.3–5.0 GHz) bands, which means that the proposed DRA is highly suitable for a variety of 5G wireless applications at these frequencies. Its wideband capabilities, customizable design, and low cost make it ideal for small-cell networks, Internet of Things (IoT) devices, vehicular networks supporting vehicle-to-everything (V2X) communications, and 5G industrial automation. Overall, this work not only contributes to the current knowledge in antenna technology, but also sets a foundation for future exploration and innovation in the design and manufacture of DRAs.