Performance Analysis of a Low Profile Hybrid Antenna for Broadband Applications

Abstract

In this paper, a low profile dielectric resonator antenna (DRA) is proposed and investigated. To achieve the broad impedance bandwidth the proposed antenna geometry combines the dielectric resonator antenna and an underlying microstrip-fed slot with a narrow rectangular notch, which effectively broadens the impedance bandwidth by merging the resonances of slot and DRA. The physical insight gained by the detailed parametric study has led to find out a set of guidelines for designing the antennas for any particular frequency band. The design guidelines have been verified by simulating a set of antennas designed for different frequency bands. For validation, a prototype antenna is fabricated and tested experimentally. The measured results show that the proposed DRA offers an impedance bandwidth of about $125.34\%$ from 1.17 to 5.1 GHz with reasonable gain between 3.5 and 5.7 dBi. The volume of the proposed DRA is $0.16\lambda _{dr}^{3}$, where $\lambda _{dr}$ is the wavelength at center operating frequency of the DR. A comprehensive study on bandwidth shows that the proposed DRA provides maximum bandwidth in terms of the DR volume ($\hbox {BW}/V_{dr}$) and the DR height ($\hbox {BW}/h_{dr}$) than the other similar reported work on hybrid wideband DRA designs.

1 Introduction

Since its beginning in 1983 [1], dielectric resonator antennas (DRA) attracted huge attention from research community due to its interesting features such as small size, high radiation efficiency and compatibility with different feeding mechanisms. However, for single DRA the typical bandwidth is often below $10\%$ [2], which may not be sufficient for various broadband applications. The current trend of research in DRA is focusing predominately on bandwidth enhancement especially with a low profile design [3]. Previously, numerous techniques have been proposed to improve the bandwidth of the DRA. For instance, stacked DRA or DRA with multiple layers of different size and dielectric constants were used to increase the bandwidth [4, 5]. Besides, different geometries of DRA such as E-shape [6], U-shape [7], T-shape [8] and L-shape [9] were proposed for bandwidth enhancement. However, the above techniques [4,5,6,7,8,9] increase the complexity and volume of DRAs and require some additional process to modify the DRA geometry, which may increase the implementation cost of the DRAs. Moreover, special feeding mechanisms were also proposed for bandwidth enhancement e.g. a U-shaped DRA was fed by an elliptical microstrip patch in [7], whereas in [9], a T-shaped DRA was fed by an inverted-trapezoidal conformal line. Although, the reported bandwidth was above $70\%$, but at the expense of increased antenna complexity.

Recently, multiple resonance techniques, where DRA is integrated with microstrip patches, slot antennas or any other radiator often known as hybrid DRA can ensure wide bandwidth by combining the resonances of individual antennas. The advantage of this technique is to offer the size miniaturization and preserves the antenna polarization and radiation pattern within the working frequency band [10,11,12,13,14,15]. In [11], a hybrid antenna design is achieved by embedding the resonances of DRA and the coplanar waveguide (CPW) inductive slot. Similarly, in [12] DRoP (dielectric resonator on the patch) configuration is used to merge the resonances of microstrip patch with DRA. Although, in these reported works [11, 12] the antenna volume was not increased, but the bandwidth was limited to only $28.8\%$. On the other hand in [13], $118\%$ bandwidth is achieved but at the expense of increased volume and complexity.

In this paper, a simple low profile broadband hybrid DRA based on multiple resonance technique is proposed and investigated. The proposed antenna structure consists of a rectangular DRA (RDRA) backed by a rectangular slot in the ground plane. An inverted T-shaped feedline is proposed to excite both antennas, simultaneously. It effectively couples the energy for a wider range of frequencies between the feedline and radiating structures as compared to strip [6] or probe excitation mechanisms [5]. A comprehensive parametric study has been performed to observe the effect of DRA, slot and feedline dimensions on the input impedance of the DRA. Based on this study, a set of guidelines for designing the antennas for any particular frequency band are presented. These design guidelines have been verified by simulating a set of antennas designed for different frequency bands.

For validation, a prototype broadband antenna is fabricated and tested experimentally. Measured results show that the proposed antenna offers an impedance bandwidth of about $125.34\%$ from 1.17 to 5.1 GHz with reasonable gain between 3.5 and 5.7 dBi. Further, to quantify the low profile and wideband feature of the proposed antenna, two different bandwidths are calculated. First, is the ratio of bandwidth (BW) to the height of DR ($\hbox {BW}/h_{dr}$) as reported in [6] and the second is the ratio of bandwidth (BW) to the volume of the DR ($\hbox {BW}/V_{dr}$) as discussed in [8]. It is found that the proposed antenna offers highest bandwidth, highest ($\hbox {BW}/V_{dr}$) and highest ($\hbox {BW}/h_{dr}$) than the other reported conformal patch excited and the hybrid wideband DRA designs.

Fig. 1

Geometry of the proposed antenna, a cross sectional view, b bottom view and c photograph of the fabricated antenna

2 Antenna Design

Figure 1a, b illustrate the geometry of the proposed antenna. A prototype as shown in Fig. 1c, is fabricated on the FR-4 substrate with dielectric constant of 4.3 and a thickness of 1.5 mm. Due to easy fabrication and enhanced degrees of freedom than the other shapes, rectangular shape is chosen for DRA. The DRA is made of Rogers RO3010 of dielectric constant 10.2. As illustrated in Fig. 1b, the slot of size ($l_{g}\,\times \,w_{g}$) is created in the ground plane of the DRA. The distance from the edge of the slot to the edge of the DRA is “g”. A small rectangular notch right below the feedline is etched in the ground plane. It significantly improves the impedance match throughout the frequency band. The feeding mechanism adopts an inverted T-shaped feedline, which consists of three arms of different length. The input $50\,\Omega$ microstrip line is the first arm of the feed as shown in Fig. 1a, which is further extended by a length $l_{2}$ beneath the DRA and comprise the 2nd arm of the feed. At last, a metal strip of height $h_{2}$ has conformably adhered to the DRA’s wall, which is depicted by 3rd arm in the enclosed portion of Fig. 1a. The width of each arm is equal to 3 mm.

Fig. 2

Geometry of the reference antennas, a reference antenna-1 (conventional RDRA), b reference antenna-2 (slot antenna) and c proposed hybrid antenna

Fig. 3

Simulation reflection coefficient of the reference and proposed antennas

3 Design Consideration and Key Parameters

To ensure the wideband characteristic and miniaturized size, the proposed antenna structure incorporates two different antennas. First, the conventional rectangular DRA excited by the conformal microstrip feedline as denoted by reference antenna-1 in Fig. 2a and second is the microstrip-fed slot antenna shown as reference antenna-2 in Fig. 2b. The design procedure for obtaining the broad impedance bandwidth starts with the selection of operating frequency, which generally depends on dimension and the dielectric constant of the antenna. In this type of configuration, the DR should have a high dielectric constant to resonate, whereas slot substrate with low dielectric constant is needed to achieve a reasonable bandwidth. The DRA’s (reference antenna-1) fundamental resonance can be calculated by using the Eqs. 1–3 [16].

$$\begin{aligned} k_{x} \hbox {tan}\left( \frac{k_{x}l}{2}\right) =\sqrt{(\epsilon _{dr}-1)k_{0}^{2}-k_{x}^{2}} \end{aligned}$$

(1)

where

$$\begin{aligned} k_{0}=\frac{2\pi }{\lambda _{0}}, k_{y}=\frac{\pi }{w}, k_{z}=\frac{\pi }{2h} \end{aligned}$$

(2)

and

$$\begin{aligned} k_{x}^{2}+k_{y}^{2}+k_{z}^{2}=\epsilon _{dr}k_{0}^{2} \end{aligned}$$

(3)

With above equations, the DRA resonance is calculated at 3.6 GHz for $w = 22$, $l = 20$, $h = 10$ and $\epsilon _{dr} = 10.2$ which is close to the simulated frequency response, covering $5\%$ bandwidth from 3.2 to 3.37 GHz as shown in Fig. 3. The slot (reference antenna-2) resonance frequency can be approximated by the slot length and effective dielectric constant of the substrate. It can be given as:

$$\begin{aligned} f= \frac{1.8c}{l_{g}\sqrt{\epsilon _{eff}}} \end{aligned}$$

(4)

For length $(l_{g}) = 60\,\hbox {mm}$ and dielectric constant $(\epsilon _{dr})= 4.3$, the slot resonates at 4.8 GHz, which is considerably matched with the reflection coefficient curve shown in Fig. 3. It can be noticed that both slot and DRA resonance frequencies are far away from each other and are difficult to merge. However, when the DRA is coupled to the slot in the ground plane as shown in Fig. 2c, the effective permittivity ($\epsilon _{eff}$) seen by the DRA is varied which lowers the slot resonant frequency [14]. Figure 3 illustrates the reflection coefficient curve of the hybrid antenna structure (shown in Fig. 2c). As can be seen, both antennas (reference antenna 1 and 2) collectively provide wide impedance bandwidth of $125.3\%$ (1.17–5.1 GHz), which $96\%$ more than the reference antenna-1 and $80\%$ more than the reference antenna-2. The reason behind the enhanced bandwidth is that both structures, when stacked together, forms a kind of monopole antenna thus renders wide band response.

With reference to Fig. 1, there are various parameters, which may influence the antenna performance; these include the dimensions of DR, slot and feedline. For more insight, an extensive parametric study is carried out using a full wave electromagnetic simulator CST Microwave Studio [17] and discussed in next subsections.

Fig. 4

a Reflection coefficient and b input impedance variation as a function of slot length ($l_{g}$)

3.1 Slot Length (l _p )

Slot length has a crucial role in hybrid antenna design. Its length affects the coupling of the electromagnetic energy from feedline to the DR and excites the different resonant mode of the hybrid structure. An appropriate slot length for DR excitation depends on the DR resonance frequency [18]. Therefore, slot length should be selected in such a manner that it could appropriately combine the resonating mode of the DRA and slot. Figure 4a, b shows the effect of the slot length ($l_{g}$) variation on reflection coefficient and the input impedance of the antenna, respectively. All other parameters such as slot width, DRA size and position are kept constant. It can be seen that by increasing the slot length ($l_{g}$) from 30 to 60 mm, upper two resonances are largely affected and shift downwards as expected, while a slight variation is observed in lower resonance frequency. This variation confirms that upper two modes are due to the slot and lower mode belongs to DRA. The simulation result, shown in Fig. 4b indicates that by changing the slot length the impedances of all three modes could be matched to the source impedance.

Fig. 5

Reflection coefficient as a function of a DR width and b feedline dimension $l_{2}$ and $h_{2}$

3.2 DR Width (w)

Figure 5a shows the simulated reflection coefficient with different values of DR width (w) while keeping all other antenna dimensions constant. Decreasing the DR’s width (w) from 22 to 10 mm, the first resonance shifted towards the higher frequency, whereas the slot resonance frequencies remain almost unaffected. However, the intermediate resonance is slightly affected by changing the DR width (w). It is merged with the first resonance for $\hbox {w} = 10\,\hbox {mm}$, which reveals the impact of coupling between the slot and the DR. This analysis confirms that the DR controls the lower resonance frequency of the design.

3.3 Feedline Dimension

The feeding mechanism plays an important role in DRA bandwidth enhancement. Its significance is shown in Fig. 5b; where reflection coefficients for different combinations of feedline length $l_{2}$ and $h_{2}$ are compared with the conventional microstrip feed, which can be obtained by assuming the $l_{2}$ and $h_{2}$ equals 0 mm in Fig 1a. It can be seen that the antenna offers an impedance BW of $19.2\%$ when fed by microstrip line (i.e. $l_{2} = h_{2} = 0\, \hbox {mm}$). In this case, DRA is not excited but slot loaded with DRA hence the result is similar as microstrip-fed slot antenna shown in Fig. 2b. Whereas, by applying inverted T-shaped feedline both resonating structures get excited appropriately, resulting in an improved impedance matching and bandwidth. It can be seen that an optimum antenna excitation is achieved at $l_{2} = 6$ and $h_{2} = 8\, \hbox {mm}$, where the maximum bandwidth is approximately $125\%$ or 6.5 times of the BW obtained for $l_{2} = h_{2} = 0\, \hbox {mm}$.

Fig. 6

a Reflection coefficient and b input impedance variation as a function of notch length (d)

3.4 Depth of Small Rectangular Notch d

The small rectangular notch right below the feedline has a large impact on impedance matching. To show its importance, firstly it is removed from the ground plane by assuming the parameter d equals 0 mm in Fig. 1b, keeping all other antenna dimensions constant. Figure 6a, b shows the reflection coefficient and input impedance for various notch length (d) conditions, respectively. As can be seen, when there is no notch behind the feedline (i.e. $d=0\, \hbox {mm}$), the antenna resonates at three frequencies, but with poor impedance matching. However, by introducing a rectangular notch and varying its length from 0 to 11 mm, both input reactance and resistance corresponds to each resonance frequency are decreased and tend to 0 and $50\,\Omega$, respectively, resulting in an optimal impedance matching.

4 Design Principle and Validation

A comprehensive study of the proposed antenna developed in the preceding section can be used as to formulate some design guidelines to scale down/up the proposed design for desired frequency band. These design guidelines for realizing the hybrid slot-DRA are presented as below.

4.1 Frequency Band

First, we define the lowest and the highest resonant frequency points as $f_{l}$ and $f_{h}$, respectively and find the ratio $f_{h}/f_{l}= 5/1.17 = 4.27$.

4.2 Slot Dimensions

The slot length controls the coupling between the feedline and the DRA. An optimized slot length accounts to merge the multi resonant modes of the hybrid structure. The approximate slot dimensions that can produce broadband performance are: length $(l_{g})= \lambda _{l}/4.16$, width $(w_{g})= \lambda _{l}/6.5$, where $\lambda _{l}$ calculated corresponds to lowest frequency point $f_{l}$.

4.3 DRA Dimensions

length $(l)= \lambda _{l}/12.5$, width $(w)= \lambda _{l}/11.4$, height $(h)= \lambda _{l}/25$

4.4 Rectangular Notch Depth (d)

Adjusting the depth of the rectangular notch significantly improves the impedance match with no appreciable change in resonance frequencies. It is observed to achieve an optimum impedance match when slot depth $(d)= \lambda _{l}/22.72$ and $\hbox {width} = \lambda _{l}/62.5$.

4.5 Ground Plane Dimensions

length $l_{1}= \lambda _{l}/2.5$, width $w_{1}= \lambda _{l}/3.125$.

4.6 Feedline Dimensions

Conformal strip $(h_{2})= \lambda _{l}/31.25$.

Fig. 7

Simulated reflection coefficients of four antennas designed for four different frequency bands as given in Table 1

Fig. 8

Simulated and measured reflection coefficient

Table 1 The dimensions of four different frequency band antennas determined using the proposed design method

To verify the design guidelines, four antennas of different frequency bands (1.17–5) GHz, (2–8.33) GHz, (3–12.5) GHz and (4–16.6) GHz have been simulated using the CST simulator. The corresponding simulated reflection coefficient curves are demonstrated in Fig. 7. It is found that all four antennas exhibit impedance bandwidth $\ge 125\%$, thus validate the accuracy of the proposed design method. Table 1 shows the design values of the all four antennas. In each case, the length ratio between ground, slot and DR ($l_{1}{:}l_{g}{:} l$) is 5:3:1 whereas, the width ratio ($w_{1}{:} w_{g}{:}w$) is around 4:2:1. For verifying the method, the first antenna design is fabricated and experimentally examined. Figure 8 shows the simulated and measured reflection coefficient curves of the designed antenna-1. The measured result shows fair agreement with the simulated one. For $|S_{11}|\le -10\, \hbox {dB}$, the measured frequency band is 1.17–5.1 GHz, which is in close agreement with simulated result from 1.1 to 5 GHz. This slight difference might be due to a small air gap between DR and substrate that was left during placement of DRA over the slot antenna as fabrication was done manually. This air gap will reduce the effective dielectric constant as well Q-factor of the proposed antenna, which may result in enhanced bandwidth in the measurement. Besides, the impedance bandwidth is also sensitive to the relative position of DRA over the slot. Hence a little misalignment between DRA and slot may affect the bandwidth.

Fig. 9

Gain and efficiency plot of the proposed antenna

Fig. 10

Measured and simulated radiation pattern at a 1.2 GHz b 2.4 GHz and c 4.8 GHz

Figure 9 shows the simulated and measured peak gain of the antenna here gain is varied in the range of 3.5–5.7 dBi. As seen, the measured gain is in well agreement with the simulated gain at lower frequencies while a noticeable difference is observed at higher frequencies which is justified with the measured reflection coefficient in this range. From the inset graph of Fig. 9, it can be noticed that radiation efficiency is more than $90\%$ from 1.17 to 4.5 GHz and above $80\%$ in the remaining band. As shown in Fig. 10, radiation patterns in H-plane (yz-plane) and E-plane (xz-plane) were also measured in far-field condition and compared with the simulated patterns. Three most useful commercial frequencies 1.2, 2.4 and 4.8 GHz were chosen for comparison. It is seen that in yz-plane the measured patterns are bidirectional and show close agreement with simulated results at all frequencies. In xz-plane, the measured radiation patterns show a good match with simulated in broadside direction ($0^{\circ }$ and $180^{\circ }$), while a slight difference is observed in endfire direction ($90^{\circ }$ and $270^{\circ }$) at 1.2 and 2.4 GHz. At 4.8 GHz, a slight tilt is observed in the xz-plane pattern that is mainly due to the higher order mode. Further, it is observed that the proposed antenna is linearly polarized throughout the frequency band. Table 2 shows the comparison of impedance bandwidth of the proposed DRA with the similar works reported previously. Only measured results have been compared in Table 2.

Table 2 Comparison of the proposed antenna with recently published papers

To demonstrate the low profile feature of the antenna, the impedance bandwidth is calculated in terms of DRA size; ratio of bandwidth to DRA volume ($\hbox {BW}/V_{dr}$) [8], and the ratio of bandwidth to DRA height ($\hbox {BW}/h_{dr}$) [6]. The parameter $V_{dr}$ shows the volume of the DRA. For calculating the $\hbox {BW}/V_{dr}$ and $\hbox {BW}/h_{dr}$ each dimension of DRA is normalized by $\lambda _{dr}$, where $\lambda _{dr}$ is wavelength in DR at center frequency of the coverage band. From Table 2, it can be seen that the proposed antenna offers the highest percentage bandwidth, which validates its wideband characteristic. In addition, the proposed antenna tenders highest BW per volume and highest BW per height that confirm its low profile feature.

5 Conclusion

In this paper, a novel low profile hybrid (DRA + slot) broadband antenna is proposed and studied. A detailed analysis for input impedance has been done using parameters variation. Further, a set of guidelines has been presented for designing the antennas for the different frequency band. These guidelines have been verified using the simulated reflection coefficients for a set of antennas operating at different frequency bands. For validation, a prototype antenna is fabricated and tested experimentally. The measured result shows that the antenna offers an impedance bandwidth of $125.34\%$ (1.17–5.12 GHz). In addition, it achieves a stable radiation pattern with measured gain of 3.5–5.7 dBi, in the desired frequency band. It has examined that the proposed DRA exhibits highest fractional BW, highest BW per volume and highest BW per height compared to the earlier reported works.