II. OPERATING PRINCIPLE OF THE PROPOSED ANTENNA

Figures 1(a) and 1(b) show the schematic of the CP printed dipole antenna in [Reference Luo, Chu and Zhu17] and the slot antenna proposed in this work, respectively. Their respective radiating elements are the four dipoles and four slots. For the CP dipole antenna [Reference Luo, Chu and Zhu17], the electric current flows along the four dipoles. For the proposed slot antenna, the electric field is distributed along the four radiating slots as the equivalent magnetic current. In our analysis, the slot antenna is placed in the XY-plane. One pair of parallel slots is arranged in the X-axis, whereas another pair of parallel slots is in the Y-axis. Besides, the length and width of four identical radiating slots are denoted by L and W, respectively, and the spacing between two parallel slots is expressed as D.

Fig. 1.

Schematic representation of wide-beamwidth CP antennas and their radiation patterns. (a) Printed dipole antenna. (b) Slot antenna. (c) Radiation patterns of the CP dipole. (d) Radiation patterns of the CP slot antenna.

Now, let us start with the analysis of radiation patterns for the proposed slot antenna on an infinite ground plane at first. When the width of the slot (W) is much narrower than the free-space wavelength (λ ₀), that is, W ≪ λ ₀, the electric field on each of these four slots can be assumed to be perpendicularly oriented across these narrow slots with no longitudinal electric field component, and its intensity varies as a cosine function along the slot. Based on the antenna theory in [Reference Garg, Bartia, Bahl and Ittipiboon19], the far-zone normalized electric field component in the XZ-plane, radiated by a pair of parallel slots along the X-axis, can be deduced as

(1) $${E_\varphi} (\theta ) = \displaystyle{{\cos \left( {kL \times \sin (\theta )/2} \right)} \over {{{\left( {kL \times \sin (\theta )/\pi} \right)}^2} - 1}} \times \cos (\theta ),$$

where k = 2π/λ ₀ is the wave number in free space.

Similarly, for a pair of parallel slots along the Y-axis, the far-zone normalized electric field component in the XZ-plane can be derived as

(2) $${E_\theta} (\theta ) = \cos \left( {kD \times \sin (\theta )/2} \right).$$

In order to obtain the perfect CP radiation with ∠E _θ –∠E _φ  = ±90° at the broadside, the vertical and horizontal paired-slots need to be excited in the same amplitude with 90° phase difference. With this arrangement, the AR varies as a function of the polar angle (θ) in the XZ-plane such that

(3) $${\rm AR}\left( \theta \right) = \left \vert {20 \times {{\log} _{10}}\left \vert {\displaystyle{{{E_\theta} (\theta )} \over {{E_\varphi} (\theta )}}} \right \vert} \right \vert. $$

The |E _θ | and |E _φ | radiation patterns of the printed CP dipoles are plotted in Fig. 1(c) based on [Reference Luo, Chu and Zhu17]. By using (1) and (2), the far-zone electric field components, |E _θ | and |E _φ |, of the slot antenna in the XZ-plane are calculated as shown in Fig. 1(d). Note that the |E _θ | radiation pattern is similar to the |E _φ | pattern of its dipole counterpart. Meanwhile, the |E _φ | pattern of the slot antenna and |E _θ | pattern of its dipole counterpart are also similar to each other.

In this work, each slot is designed with the length of L = 0.5λ ₀ as usual. As can be seen in Fig. 2, the |E _φ | radiation pattern (black-line) keeps almost constant regardless of varied spacing D, since the |E _φ | expression in (1) is not a function of D. In contrast, the |E _θ | radiation pattern (green-line) and AR (red-line) beamwidth are varied dramatically as D increases from 0.2λ ₀ to 0.6λ ₀. It can be understood from (2) and (3) that |E _θ | and AR are highly dependent on the spacing D.

Fig. 2.

Calculated |E _θ |, |E _φ | radiation patterns and 3 dB AR in the XZ-plane with varied spacing D. (a) D = 0.2λ ₀. (b) D = 0.4λ ₀. (c) D = 0.6λ ₀.

More specifically, as D = 0.2λ ₀, the beamwidth of the |E _θ | radiation pattern in Fig. 2(a) is wider than that of |E _φ | radiation pattern. Using equation (3), the calculated 3 dB AR beamwidth is found to be extremely narrow in a range of θ from −43 to 43°. When the spacing D increases to 0.4λ ₀, the beamwidth of the |E _θ | radiation pattern decreases dramatically as illustrated in Fig. 2(b), and it becomes almost the same as the |E _φ | pattern in a wide range of θ from −60 to +60°. Hence, the 3 dB AR can be further enlarged to cover a wide range of the polar angle from −70 to 70°. In Fig. 2(c), as D reaches to 0.6λ ₀, the beamwidth of the |E _θ | pattern continues to decrease and becomes narrower than that of |E _φ | pattern. Thus, the 3 dB AR beamwidth becomes narrow in the polar axis from −30 to 30°. To summarize these findings, Table 2 is tabulated to quantitatively depict the value of 3 dB AR beamwidth as the spacing D varies from 0λ ₀ to 0.9λ ₀. In final, it can be found that a maximum 3 dB AR beamwidth of about 140° is obtained at D = 0.4λ ₀.

Table 2.

The 3 dB AR beamwidth versus varied spacing D.

III. ANTENNA DESIGN AND IMPLEMENTATION

Based on the above analysis, a slot antenna with the wide AR beamwidth is designed on a dielectric substrate with the relative permittivity of 2.2 and the thickness of h = 3.18 mm as shown in Fig. 3. A feeding network is formed on top of the substrate to properly feed the proposed CP antenna with four radiating slots, and a ground plane with the size of G × G is selected at the bottom surface of the substrate. Besides, two pairs of slots with the same length of L and width of W are etched symmetrically onto the ground plane for radiation. In this design, the slot is slightly shorter than 0.5λ ₀ in length in order to take into account the inductive effect of the slot line short-circuited ends. With reference to the results depicted in Fig. 2 and Table 1, the spacing D between two parallel slots is chosen as about 0.4λ ₀ towards the widest AR beamwidth.

Fig. 3.

Configuration of the proposed CP antenna with feeding network.

In the following, a 1–4 probe-to-microstrip feeding network with equal amplitudes and quadrature phase difference between output ports is designed to achieve the CP radiation at the broadside direction, i.e. Z-axis in Fig. 3. Firstly, an input impedance of 200 Ω for a slot is obtained at the resonant frequency of 3.4 GHz by selecting the strip width (W ₂), feeding point (D ₁), and slot width (W) properly. These impedances (200 Ω) of the slots are then considered as the load impedances at the four output ports (E, F, H, and I) of a 1–4 microstrip feeding network.

With reference to antenna structure in Fig. 3, it is required the input signal at Port A (coaxial probe) to be divided into four-way output signals at Ports E, F, H, and I with equal amplitude and 90° phase difference. Based on the network theorem in microwave engineering [Reference Pozar20] and the above constraints, the 1–4 microstrip feeding network can be designed as depicted in Fig. 4. It can be seen that the input signal from Port A is split into two ways, i.e. A to B and A to C, in equal amplitude and 180° phase difference. The characteristic impedances of microstrip line in these two ways are set as 100 Ω, aiming to match well the input Port A. Besides, the path difference between A to B and A to C is about λ _g/2 (λ _g is the guided-wavelength), thus allow output signals along these two paths to have the same amplitude and 180° phase difference. The impedances along the four output paths (B to E, B to F, C to H, and C to I) are set as 200 Ω in order to match well with the input impedance of the radiating slot at its resonance, and the path difference about λ _g/4 is employed to produce 90° phase difference. Finally, this designed feed network is utilized to provide two pairs of equal-amplitude signals with 90° phase difference at the output ports of E, F, H, and I, respectively. Compared to the feeding network in [Reference Luo, Chu and Zhu17], the feeding network proposed herein is simple in geometrical structure and straightforward in the design procedure.

Fig. 4.

Equivalent transmission-line topology of the feeding network.

In Section II, the analysis of our proposed slot antenna was executed under the ideal assumption of an infinite ground plane. In practical implementation, the overall size of the ground plane is definitely finite, so the induced equivalent magnetic currents on the finite ground plane edges will affect the radiation pattern of an antenna [Reference Namiki, Murayama and Ito21, Reference Bokhari, Mosig and Gardiol22]. Figure 5 depicts the simulated 3 dB AR beamwidth of the proposed antenna as a function of the finite ground size G varying from 100 to 200 mm. When G is enlarged from 100 to 160 mm, the beamwidth under AR < 3 dB is increased from 82 to 126°. Beyond G = 160 mm, the 3 dB AR beamwidth falls down slightly. If G is large enough, the 3 dB AR beamwidth tends to approach 140°, which corresponds to the case of an infinite ground plane.

Fig. 5.

The 3 dB AR beamwidth of the proposed CP antenna in the XZ-plane with varied ground plane sizes.

To explain the behavior of AR beamwidth in detail, the AR beamwidth at G = 120, 160, and 200 mm against the θ is further studied and the relevant results are plotted in Fig. 6. At G = 120 mm, the AR value smoothly varies with respect to θ, and the 3 dB AR beamwidth becomes relatively narrow of about 86°. When G is increased to 160 mm, the beamwidth (AR < 3 dB) reaches to 126°. However, the AR curve has the ripple due to the unexpected edge diffraction from finite ground plane. At G = 200 mm, the 3 dB AR beamwidth slightly decreases to about 109°. Thereafter, of the three listed cases, G = 160 mm is the most suitable value to achieve a wide AR beamwidth. Till now, the proposed slot antenna in Fig. 3 with finite ground plane can be designed for practical realization and Table 3 tabulates all its dimensional parameters.

Fig. 6.

The 3 dB AR in the XZ-plane with varied ground plane size G. (a) G = 120 mm. (b) G = 160 mm. (c) G = 200 mm.

Table 3.

Dimensions of the proposed antenna in Fig. 3.

IV. RESULTS AND DISCUSSION

To confirm the design performance, the proposed antenna was simulated with the EM simulator IE3D, and was fabricated following these design parameters listed in Table 3. Figure 7 shows the top- and bottom-view photographs of the fabricated CP antenna prototype. The return loss is measured using the Agilent N5230A Vector Network Analyzer. The whole radiation pattern, including the AR and gain, are experimentally evaluated with the near-field SATIMO antenna test system.

Fig. 7.

Photographs of the fabricated antenna. (a) Top view. (b) Bottom view.

Figure 8 depicts the simulated and measured AR as a function of θ at the central frequency of 3.4 GHz in the XZ-plane. Under the 3 dB AR definition, the measured AR beamwidth has reached 126°, which matches well with the simulated one. The slot antenna radiates the left- and right-hand circular polarizations, in short, LHCP and RHCP, below and above the horizontal plane as shown in Fig. 3, respectively. The simulated and measured radiation patterns in the XZ and YZ planes at 3.4 GHz are depicted in Fig. 9, the bidirectional CP radiation patterns with the low cross-polarization are obtained for the proposed antenna, which can be well used for applications such as microwave sensor networks and broadcasting base stations [Reference Shen, Lu and Cao23, Reference Lu and Wang24].

Fig. 8.

Simulated and measured AR of the proposed antenna in Fig. 7 with respect to the θ at 3.4 GHz.

Fig. 9.

Simulated and measured radiation patterns at 3.4 GHz.

The simulated and measured results in Figs 10(a)–10(c) show the peak gain, AR, and return loss against frequency. In the operating band of 3.33–3.47 GHz, the antenna has achieved the expected CP radiation performance such as AR < 3 dB, |S ₁₁| < −16 dB, and gain > 4 dBic.

Fig. 10.

Simulated and measured antenna parameters with respect to frequency. (a) Gain. (b) AR. (c) |S ₁₁|.