1. Introduction
In the realm of mmWave applications, the demand for enhanced communication capabilities has fueled the emergence of reconfigurable antennas as a pivotal solution. These antennas offer a crucial degree of flexibility and adaptability, specifically tailored to address the challenges inherent in mmWave frequencies. Given that mmWave frequencies pose unique hurdles in signal propagation and susceptibility to environmental obstacles, reconfigurable antennas play a vital role in mitigating these challenges [
1,
2,
3]. Their ability to adjust radiation patterns, polarization, frequency, and bandwidth characteristics in real-time provides a dynamic response to the intricacies of mmWave communication environments.
Moreover, beyond their technical advantages, reconfigurable antennas significantly contribute to the overall efficiency and cost-effectiveness of mmWave systems. By offering adaptability and versatility, these antennas facilitate streamlined design approaches, diminishing the reliance on complex and specialized hardware components. This, in turn, promotes the development of compact and lightweight mmWave devices, aligning with the escalating demand for miniaturization in contemporary communication systems.
In a parallel development, circularly polarized antennas have garnered substantial interest in the realm of mmWave communication systems. This interest is rooted in well-established advantages, such as their resistance to multipath interference and tolerance to misalignment between transmitting and receiving antennas [
4]. As such, the incorporation of circularly polarized antennas into mmWave designs represents a strategic choice to enhance the robustness and reliability of communication systems operating in these high-frequency bands. The synergy between reconfigurable antennas and circularly polarized antennas presents a holistic approach to addressing the multifaceted challenges and optimizing performance in the rapidly evolving landscape of mmWave applications. However, circular polarization can be achieved in one of two senses: LHCP or RHCP. As a result, it is important to design an antenna that supports the two polarization senses to sustain communications in challenging and demanding environments. Therefore, studies have been published proposing novel mmWave circularly polarized antennas [
5,
6,
7,
8]. However, these approaches share a limitation of providing fixed circular polarization senses, i.e., either LHCP or RHCP, thereby restricting their usage to a single polarization type. In contrast, an array that utilizes LHCP and RHCP has been proposed [
9]. Besides, other designs have been reported that incorporate multiple feeding ports to generate the desired circular polarization sense, albeit at the cost of added complexity [
10,
11]. In addition, a cost-effective design of a polarization reconfigurable mmWave antenna that is capable of radiating CP and LP waves provides another needed degree of flexibility to any communication system.
Prototypes of polarization reconfigurable mmWave antennas have been reported in several studies [
12,
13,
14,
15,
16,
17,
18]. For example, a K-band polarization reconfigurable patch antenna has been proposed in a layered structure using RF-MEMS to switch between two polarization modes, LHCP and LP, over impedance bandwidths of 11.8% and 3%, respectively, with a peak gain of ~3.9 dBic [
12]. Another polarization reconfigurable patch antenna that operates at 29 GHz has been proposed with LP, LHCP, and RHCP radiation modes [
13]. The findings demonstrated respective impedance bandwidths of 5.1% and 3.1% for the CP and LP modes in combination with an AR bandwidth of 1.7% and a maximum gain of ~8.5 dBi in the LP mode. However, the reconfigurability relies on external stimuli, using UV laser pulses, to control the phase change of the Germanium Telluride (GeTe) material, which limits the practicality and poses manufacturing challenges, thereby detracting from its potential implementation. In a more recent study, 4 PIN diodes were utilized in the design of a reconfigurable patch antenna with three polarization modes RCHP, LHCP, and LP operating at 29 GHz with an impedance bandwidth of ~5.4% in all cases [
14]. The measured respective gains are 3 dBic and 4 dBi in the CP and LP radiations albeit with no data for the AR bandwidths.
On the other hand, a 2 × 2 mmWave polarization reconfigurable patch antenna array has been reported by utilizing two PIN-diode pairs to switch between dual CP modes with respective impedance and AR bandwidths of 11% and 4% with an efficiency of 51% [
15]. Besides, linear, and square arrays with 10 and 2 × 2 T-shaped slot elements have been reported with respective impedance bandwidths of 3.3% and 10%, where two mechanically switchable CP senses have been achieved over an AR bandwidth of 3% for both arrays [
16]. Similarly, full polarization reconfigurability has been achieved using a phased array of 8 × 12 Butterfly elements with 4 feeding ports that switch the polarization by utilizing the required excitation amplitude and phase [
17]. The AR and impedance are presented over a frequency range of 27–29 GHz with a simulated total efficiency of 56–60%. Furthermore, a double-folded, polarization-reconfigurable, dual-antenna array has been proposed with a CP switching over a bandwidth of 7% through the activation of a single-pole-double-throw (SPDT) switch when 10 elements are used [
18]. In addition, a figure of merit was introduced and used to compare the performance to those in earlier studies. However, in [
15,
16,
17,
18], polarization reconfigurability has been achieved by utilizing a substantial number of elements in conjunction with customized feeding networks that naturally increase the cost and complexity compared to a single antenna configuration.
In the presented study three modes of polarization reconfigurability are achieved using a singly fed open-loop antenna that incorporates two PIN diodes only. In addition, the antenna offers impedance and CP bandwidths of 12.9% and 7%, respectively, in combination with a gain of ~8.5 dBic and an estimated total efficiency of 79%. Moreover, the proposed configuration avoids the need to utilize UV laser pulses, multi-layer PCB structures, or large arrays with complex feeding networks.
It should be noted that the RHCP and LHCP modes have been achieved when a single PIN diode is forward-biased while the other PIN diode is reverse-biased. On the other hand, when both PIN diodes are under zero-bias conditions, the LP mode is achieved with dual narrow impedance bandwidths since the total impedance presented by the two unbiased diodes at the input of the antenna is different from that presented using one forward, and one reverse, biased PIN diodes. Therefore, the proposed configuration offers the distinct advantage of offering hybrid reconfigurability by varying polarization and bandwidth without any compromise on the performance. To the best of the authors’ knowledge, this is the first attempt to design a mmWave antenna that offers hybrid reconfiguration. It may be worth pointing out that bandwidth reconfigurability received increased interest in recent years since an antenna with bandwidth control could partially perform a filtering task as well as the radiation [
19,
20,
21].
The effectiveness of the proposed antenna design is demonstrated through comprehensive simulation and measurement results. These results showcase promising performance in terms of impedance and axial ratio bandwidths. The outcome of this study highlights the potential usefulness of the proposed CP reconfigurable antenna for mmWave applications, as well as practical implementations in satellite and wireless communications systems.
The paper is organized as follows:
Section 2 summarizes the key contributions and
Section 3 introduces the utilized configuration and design principles of the polarization reconfigurable loop antenna.
Section 4 describes the utilized PIN diodes, and DC biasing network while
Section 5 presents the switching mechanism of the polarization modes. It should be noted that a single biasing network has been utilized for the two PIN diodes, which provides further simplicity in the design. Moving on to
Section 6, details are provided concerning the prototypes, fabrication processes, and measurements of the proposed antenna. Subsequently,
Section 7 provides a comparison between the performance of the proposed antenna against those reported in the literature.
Section 8 presents the conclusion comments.
3. Antenna Configuration
Figure 1 presents the configurations of a reconfigurable antenna designed to operate at 28 GHz. The proposed antenna consists of two concentric loops [
22,
23,
24] in which the outer and inner loops represent the active and parasitic elements, respectively. The proposed antenna is printed on the upper surface of the Rogers RO4003C dielectric substrate, with an equal-sized ground plane positioned beneath the substrate. Besides, the outer loop’s radius needs to be selected so that the circumference is approximately one effective wavelength,
. Two gaps have been created on the outer loop to house two PIN diodes that facilitate reconfigurability. The gaps have been optimized to create an optimal traveling wave current distribution along the loop, which is crucial for achieving circular polarization.
Figure 2 illustrates the parametric analysis, showcasing the variation in the axial ratio for various values of Δ
φ2 with the aim of identifying the most favorable angular gap that yields the widest axial ratio bandwidth. As a result, the optimum Δ
φ2 has been determined as 20°. Besides,
Figure 3 presents the simulated axial ratio, comparing two scenarios: with and without the parasitic loop. Upon utilizing the inner loop, the AR ≤ 3 dB bandwidth experiences a notable increase from 2.5% to 8.1%. As anticipated, the parasitic loop significantly enhances the AR bandwidth, given that each loop generates a single AR minimum point. The merging of these two minimum AR points results in the observed bandwidth improvement. The circumference of the outer loop can be calculated as
where
is the effective relative permittivity, and
λ0 is the free-space wavelength.
The two PIN diodes function as switches capable of toggling between forward-bias and reverse-bias states to select the desired polarization sense. For the utilized PIN diodes, the forward-bias resistance is set to 5.2 Ω, while the reverse-bias resistance is 15 kΩ. By adjusting the gaps Δ
φ1 and Δ
φ2 as well as controlling the PIN diode states, the polarization sense can be electronically switched between RHCP and LHCP over a relatively wide frequency range. In addition, LP radiation can be achieved when the two diodes are unbiased. The antenna is connected via a microstrip line at ϕ = 0, featuring a width designated by the parameter
. The microstrip line has been designed to attain a 50 Ω impedance and has been connected to the coaxial cable using the structure shown in
Figure 1. In addition, two holes have been incorporated in the pads that are utilized for the SMA connector fixing. Furthermore, the integration of vias into the design yields several benefits, encompassing improved grounding, mitigation of surface waves, and the expansion of bandwidth, thereby enhancing overall antenna performance. Consequently, the presented antenna array is furnished with six vias, each possessing a radius of 1 mm and situated at 4 mm intervals, center to center. Detailed configuration parameters are summarized in
Table 1.
4. DC Biasing Network
The PIN diode possesses distinctive electrical characteristics that render it an excellent choice for various switching applications. Its primary advantage lies in its capacity to alter its resistance based on the applied bias voltages, allowing it to seamlessly transition between conducting and non-conducting states. When incorporated in antennas, the PIN diode functions as a switch, facilitating the antenna’s ability to switch between different signal paths. In the proposed configuration, the MA4AGFCP910 PIN diode has been utilized [
25]. It should be noted that the PIN diodes have been modeled in CST using lumped elements in the equivalent RLC circuits of the forward and reverse-biased diodes using the parameters presented in
Table 2.
As mentioned earlier, the PIN diodes have been placed in two strategically placed gaps incorporated within the outer loop. The PIN diodes are then subjected to a biasing process facilitated by a designated bias line.
Figure 4 illustrates the proposed loop antenna that incorporates a single CPS to supply the DC bias current to the two PIN diodes. The surface current distribution of the antenna with a CPS bias line is presented in
Figure 4b. A periodic current distribution is observed along the bias line due to the RF current leakage from the antenna. Such current leakage results in unwanted electromagnetic radiation and causes the bias line to function as a radiating element with fields that interfere with and distort those radiated by the antenna, particularly the CP radiation. This is in addition to altering the antenna’s key characteristics, including resonance frequency, input impedance, AR, gain, and radiation pattern. To mitigate this issue, the development of a novel bias line structure is imperative one that effectively carries the requisite DC bias current while preventing the unwanted flow of RF current. To curtail RF current propagation within the coplanar strip line, a PBG structure with a period of
λeff/4 has been integrated into the bias line, as illustrated in
Figure 5 [
26], to present a high impedance that minimizes the leakage of the RF current along the CPS line. Therefore, the PBG structure effectively acts as an RF choke that suppresses the flow of RF current signal along the coplanar strip line, thereby ensuring optimal performance and efficiency. The PBG section consists of a sequence of high and low quarter wavelength impedance transformer sections. The gap between the lines of the high-impedance section is defined as
A, the PBG cell’s length is
B, the gap between the thicker lines of the low-impedance PBG cells is
C, and the widths of the CPS line are
D and
E. The design parameters for the CPS bias line and PBG are summarized in
Table 3.
Employing the CPS-PBG facilitates the biasing of the PIN diodes with no need for lumped chokes or capacitors, which eliminates additional losses introduced by these components, and hence results in higher efficiency. Furthermore, to achieve a design that optimizes cost-effectiveness and efficiency while eliminating the need for an additional feeding network to bias the second PIN diode, a modification was introduced by aligning both diodes in the same direction. Such an arrangement ensures that a single switch and a single biasing circuit are sufficient to achieve reconfigurability, which simplifies the design, improves efficiency, and reduces the required DC power. It should be noted that the biasing CPS-PBG line has been tilted by 45° to avoid sharp corners that exist if vertical and horizontal sections are utilized for biasing.
Figure 6 showcases the reflection and transmission coefficients of the RF choke through the utilization of two ports positioned at each end of the biasing line. Notably, the stopband of the proposed RF choke encompasses the antenna’s operating bandwidth. On the other hand,
Figure 7 presents the antenna’s surface current distribution along the CPS-PBG bias line at 28 GHz, where it can be noted that adding the PBG sections effectively cuts off the RF current from flowing across the CPS line, which means the CPS-PBG line serves the purpose as a distributed RF choke. Furthermore, as can be observed from
Figure 7, the biasing network has been connected to the circular loop using strip line sections with lengths of
Wa and
Wb that have been adjusted for optimum matching.
Figure 8 and
Figure 9 present the variations of reflection coefficient and axial ratio, respectively, for various antenna configurations; without any biasing, with the CPS line alone, and after the additions of the PBG to the CPS lines. It can be noted from these results that adding the CPS line without the PBG has deteriorated the performance significantly due to the undesired flow of the unbalanced RF current. On the other hand, the performance of a standing-alone antenna has been preserved by adding the PBG to the CPS line to achieve the equivalent of an RF choke that stops any RF current flowing along the biasing network over the desired frequency range. In addition, the antenna performance has been preserved across the operating bandwidth, which demonstrates that the CPS-PBG stopped the RF current leakage over a relatively wide frequency range compared to that of a traditional filter.
5. Polarization Switching by Utilizing Two PIN Diodes
Figure 10 presents the mechanism of switching the polarization state with the corresponding equivalent circuits, where three cases have been considered.
Case A: During this biasing, the resistance of PIN diode 1 decreases significantly, enabling easy flow of current across the gap. As a result, the current passes through the outer loop and reaches the reverse biased PIN diode 2, which is in the “OFF” state in which the diode resistance increases considerably. This behavior resembles that of an open switch, preventing current from flowing through the PIN diode 2. In this case, LHCP is achieved since the current flows clockwise throughout the antenna.
Case B: Offers RHCP by switching the bias voltage in a manner where PIN diode 2 is forward-biased, while PIN diode 1 is reverse-biased, which changes the current’s direction to counterclockwise resulting in RHCP wave.
Case C: Offers LP radiation by keeping the two PIN diodes under zero-biased conditions.
It should be noted that the chosen alignment of the two PIN diodes simplifies the design and eliminates the need for an additional biasing network, thereby optimizing cost-effectiveness while achieving the desired polarization modes. In addition, with this alignment, it is not possible to have the two PIN diodes, neither forward, nor reverse, biased simultaneously. Hence the unbiased condition of the two PIN diodes has been utilized to achieve the LP polarization.
Table 4 summarizes the states of the reconfigurable loop antenna, categorized according to various polarization states.
Figure 11 illustrates the current distribution along the loop’s circumference at 28 GHz. In contrast to linearly polarized antennas, where the current’s amplitude fluctuates to form a standing wave along the length of the antenna, circularly polarized waves require a traveling wave current distribution to be radiated, where the current’s amplitude fluctuates along the loop, gradually decreasing as it reaches the other gap on the loop.
7. Performance Comparison
Table 5 presents a comparison between the performance of the proposed antenna compared to polarization reconfigurable antenna prototypes that are reported in the literature. From the
Table 5, it is evident that the presented antenna outperforms the single antenna prototypes reported in [
12,
13,
14] in terms of the reported impedance and AR bandwidths as well as gain. In addition, the proposed antenna offers three polarization modes compared to two in [
12], employs a simple switching mechanism compared to that utilized in [
13], and needs only two PIN diodes compared to four in [
14]. On the other hand, the studies in [
15,
16,
17,
18] are focused on polarization reconfigurable mmWave arrays, which naturally require a more complex design in terms of feed network and switching mechanisms for many antennas. However, the presented single-loop antenna offers wider impedance and CP bandwidths than those reported in [
15,
16,
17] and a slightly narrower impedance bandwidth compared to [
18]. In terms of the polarization modes, the arrays reported in [
15,
16] offer two polarization states compared to three in the present study. Although full polarization reconfigurability has been achieved in [
17], four excitation ports were used, which increases the cost and complexity compared to the case of a singly fed antenna. The total efficiency of the presented design is higher than those achieved in [
15,
17,
18] and close to that reported in [
16]. On the other hand, higher gains have been reported in [
16,
17,
18] due to the considerable number of utilized elements. The overall footprint of the proposed configuration has increased owing to the required length of the CPS-PBG section, which brings benefits to the proposed configuration. However, the size can be reduced if needed by truncating the left-hand side, as well as the top, of the PCB with no impact on performance. According to the figure of merit reported (FOM) in [
18], which considers the achieved CP bandwidth, gain, polarization states, and number of antennas. The proposed antenna in this article offers a FOM of 50 compared to 9.75/5.1, 1.44, and 10 in [
16,
17,
18], respectively, which confirms the potential and cost-effectiveness of the proposed antenna. It should be noted that the proposed configuration offers another distinct advantage of hybrid configurations since the bandwidth is also reconfigured from a wider single band in the CP modes to dual-narrower bands in the case of LP radiation.