Frequency Reconfıgurable Lıbra Shape Antenna for mmW 5 g Communıcatıons

Due to the rapid advancements in electronic and wireless communication systems, there has been an increasing demand for portable electronic devices that operate across different frequency bands [1]. Consequently, multi-band and frequency reconfigurable antennas have become essential for modern communication devices [2‐4]. Frequency reconfigurable antennas can operate in multiple frequency bands without the need for physical alterations, utilizing electronic switching components such as PIN diodes [5, 6], micro-electromechanical systems (MEMS) switches [7‐9], varactor diodes [10‐12], liquid metals [13], field-effect transistors (FETs) [14], optical (photo-conductive) switches [15] and electronic radio frequency (RF) switching devices [16]. These antennas offer numerous advantages, including support for multiple frequency bands, adaptable performance, spectrum efficiency, cost-effective manufacturing, lightweight and compact design, ease of integration, scalability, energy efficiency, the ability to accommodate various application scenarios and flexibility for advanced communication systems.

Specifically, the 28 and 38 GHz frequency bands are crucial in 5G technology [17]. Antennas operating at these frequencies play a critical role in enhancing the performance and efficiency of 5G networks by enabling high data rates and low-latency communication [18]. The development of dual-band antennas that can operate at 28 and 38 GHz has become a focal point for the research community, aiming to provide seamless connectivity and coverage in these important frequency bands [19]. Furthermore, regulatory bodies such as the Federal Communications Commission (FCC) have allocated the 28 and 38 GHz bands for the development of the 5G ecosystem, underscoring the strategic importance of these frequencies in the deployment of next-generation wireless systems [20]. Looking at recent studies on frequency reconfigurable antennas operating between 28 and 38 GHz, Marzouk et al. introduced novel dual-band 28/38 GHz MIMO antennas designed for 5G mobile applications and emphasised their suitability for 5G mobile communications [21]. A notable work by Refaat et al. introduced a 28/38 GHz tuned reconfigurable antenna adapted for 5G mobile communications. The antenna was designed to operate at 28 GHz and 38 GHz frequencies and demonstrated the potential for frequency ability in the millimeter wave spectrum. This research contributes to developing optimised antennas for next-generation wireless systems [22]. Another related study by Dacosta et al. focused on 2 × 2 dual-band 28/38 GHz MIMO antenna design for 5G applications [23].

This study presents several advantages, primarily through developing a four-mode frequency reconfigurable antenna designed to operate between 26 and 40 GHz, targeting 5G communication applications. One significant advantage is the antenna’s ability to dynamically switch between frequency bands, which enhance flexibility and spectrum efficiency. The steps taken during this study include first discussing the design principles and simulation models to optimise the antenna’s performance, followed by the fabrication process and finally conducting experimental measurements to verify the antenna’s reconfigurability and performance metrics. This structured approach ensures that the developed antenna meets the stringent requirements of 5G applications while providing a versatile solution for future wireless communication needs.

The geometry of the proposed frequency-reconfigurable Libra shape antenna is illustrated in Fig. 1(a). The main radiating part of the antenna is obtained by combining a 50 Ω RF feed line with two patch antennas of different sizes through PIN diodes. A gap of 1 mm was kept between the patches to install the PIN diodes. In design processes, copper is used as a conductor with a conductivity of 5.8 × \({10}^{7}\) S/m. The proposed frequency-reconfigurable Libra antenna was printed on a 0.508-mm thickness, commercially-available substrate of Taconic TLY-5. This material was chosen due to its low dielectric constant and loss tangent (\({\varepsilon }_{r}\) = 2.2 and tanδ = 9 × \({ 10}^{-4}\)), which enable high gain, low transmission loss and wide bandwidth performance in the 27–40-GHz millimeter-wave range. Additionally, TLY-5 offers cost-effectiveness and ease of fabrication, making it suitable for prototyping applications.

The overall geometry of the proposed antenna is 20 × 20 × 0.578 mm³, and all geometric parameters are summarized in Table 1. While W₁ and L₁ do not directly contribute to radiation, they were optimised due to their significant impact on impedance matching and resonance behaviour. Furthermore, all geometric variables both rectangular (W₁ through L₃) and curved (r₁, r₂, θ₁ and θ₂) were defined as parametric in a full-wave simulation environment based on the finite integration technique and systematically swept. This parametric approach allowed for efficient tuning of the antenna’s performance across the desired frequency bands, depending on the PIN diode switching states.

Full size image

Libra shape frequency reconfigurable antenna can dynamically adjust its operating frequencies according to the on/off state of the PIN diode, allowing it to cover multiple frequency bands without needing physical modification.

The equivalent circuit model of the PIN diode in on/off states is illustrated in Fig. 1(b). According to the literature [1], the PIN diode in the forward-biased (on) state is typically modelled as a series connection of a resistor (\({R}_{on}\)) and an inductor (\({L}_{on}\)) (RL circuit) with low component values, exhibiting a low-impedance short-circuit behaviour that facilitates current flow along the radiating section of the structure. In contrast, in the reverse-biased (off) state, the PIN diode is generally represented by an RLC circuit, where a series-connected resistor \(({R}_{off}\)) and capacitor (\({C}_{off}\)) are placed in parallel with an inductor (\({L}_{off}\)). This configuration exhibits high-impedance open-circuit behaviour, effectively preventing current from propagating along the radiator.

In this design, the MA4AGFCP910 PIN diode manufactured by MACOM is employed. According to the manufacturer’s datasheet [24], the equivalent circuit parameters are defined as \({R}_{on}\) = 5.2 Ω and \({L}_{on}\)=0.15 nH in the forward-biased (on) state, and \({R}_{off}\) = 1 MΩ, \({C}_{off}\)= 0.018 pF and \({L}_{off}\)=0.15 nH in the reverse-biased (off) state. This diode provides stable performance up to 50 GHz, with a typical power handling capacity of approximately 250 mW, while exhibiting low distortion and good linearity, making it well-suited for the low-power experimental conditions applied in this study.

Figure 1 (c) illustrates the DC bias circuit used to switch the PIN diodes between on/off states. A single DC voltage line is employed, branched through metallised vias to supply both diodes. During measurements, + 1.33 V is applied to activate the diode and 0 V to deactivate it. By adjusting the voltage distribution across the diodes, three reconfiguration states are obtained, allowing a compact and practical implementation without requiring separate bias lines.

Although Fig. 1 (c) schematically illustrates DC bias components (L_b and C_b) on both terminals of each PIN diode, in the actual simulation model and fabricated prototype, these components are implemented only on one side of each diode. The opposite terminal is directly connected to the patch. The bias network is modelled using a lumped RLC circuit in a full-wave electromagnetic simulation environment based on the finite integration technique, and positions are optimised to minimise their influence on RF performance.

In the antenna design, a microwave simulation software based on the finite integration technique was used to ensure high accuracy, and the structure excited via a waveguide port was optimised through parametric analysis by adjusting the dimensions and relative positions of the two radiating patch elements, the gap between them and the placement of the PIN diodes, to achieve resonance at the target frequencies and to generate distinct surface current distributions for each switching state.

As a result of the two PIN diodes used, a total of çevfour switch positions were created; these are (0/0), (0/1), (1/0) and (1/1). Analysing the designed antenna structure conditions under these 4 different conditions, the reflection coefficient (\({S}_{11}\)) graphs are obtained, as shown in Fig. 2. In Mode 1 state, no resonant frequency is generated since no current flows through the antenna structure. In Mode 2, two resonant frequencies, 27 and 37 GHz, occur, and the return loss values at these resonant frequencies are − 14 dB and − 38.59 dB, respectively. For PIN diodes in Mode 3, energy is transferred to the larger patch antenna and the return loss is below − 10 dB at frequencies between 36 and 40 GHz. Finally, for the Mode 4 case, return losses of − 39.25 dB, − 19 dB and − 15 dB are obtained at 28 GHz, 30.7 GHz and 33 GHz resonant frequencies, respectively.

Full size image

The reconfigurable nature of the antenna using PIN diodes allows for dynamic changes in the antenna’s operating characteristics by altering the surface current distribution. Each mode corresponds to different conditions of the diodes, leading to changes in the current paths and subsequently the radiation patterns and impedance characteristics of the antenna.

It can be concluded from Fig. 1 that when the first diode is off, it creates a high impedance state at that location, effectively open-circuiting that part of the antenna. The second diode being open makes a low impedance path, allowing current to flow through that part of the antenna. The surface current is likely to be concentrated around the area where the second diode is located, with higher density near the open-state diode and lower density near the closed-state diode. In Mode 3, the surface current is expected to be higher around the first diode area where the current has a low impedance path. Due to the high impedance of the second diode, the current distribution of the antenna will change, potentially resulting in a different radiation pattern and changes in the resonant frequency of the antenna. Having both diodes open means that both regions have low impedance paths, allowing surface currents to flow freely through both parts of the antenna. Since there are no high-impedance blocks in the current path, the surface current density is expected to be more uniformly distributed, compared to Modes 2 and 3. This mode is expected to provide the widest bandwidth or the most efficient radiation pattern as it is fully utilised without any interference path between the antenna elements.

Furthermore, the surface current distributions given in Fig. 3 are obtained for the operating frequencies. For the (1/0) switching state, the maximum current density is around 420 A/m, which triggers the larger patch element, as shown in Fig. 3(c). In contrast, for the (0/1) switching state, the maximum current density is around 536 A/m, which triggers the smaller patch element, as shown in Fig. 3(b). Finally, for the (1/1) switching state, the maximum current density is around 396 A/m, which triggers both patch elements, as demonstrated in Fig. 3(f).

Full size image

The radiation patterns of the Libra shape reconfigurable antenna for all resonant frequencies are given in Fig. 4. For mode 2, radiation patterns are obtained at 27 and 37 GHz. Gain values at these frequencies are 8.52 and 9.03 dBi, respectively. For mode 3, gains of 8.76, 7.37 and 7.08 dBi are obtained at 36 GHz, 38 GHz and 39.7 GHz, respectively. Finally, for mode 4, gains of 7.48, 8.44, and 9.26 dBi are achieved at 28, 30.7 and 33 GHz, respectively.

Full size image

The real and imaginary parts of the impedances are presented concerning the frequency in Fig. 5. The real part of the impedance shows peaks and troughs over the frequency range from 20 to 40 GHz, indicating resonance at multiple frequencies. The oscillations in the imaginary part of the impedance indicate the presence of reactive contribution due to inductive and capacitive elements in the antenna structure. The resonance points (where the imaginary part crosses zero) are very important for understanding the efficiency of the antenna at certain frequencies. At these points, the antenna is fully resistive, which typically corresponds to optimum power transfer.

Full size image

The antenna structure is fabricated and measured for four different modes in the microwave laboratory. The measurement setup is shown in Fig. 6. Antenna fabrication is carried out using an LPKF-E33 printed circuit device. With this device, which has CNC-based software, resonator structures are formed by scraping the copper surface with fine tips. The reflection coefficient of the reconfigurable scale antenna is obtained using the Agilent Technologies PNA-L Vector Network Analyser (VNA).

Full size image

The measured variations of each mode concerning frequency are given in Fig. 7. For Mode 1, there is no remarkable antenna resonance. For Mode 2, a resonance occurs at 25 GHz at the centre with a value of − 14 dB for a bandwidth of 500 MHz. At the same time, a resonance occurred at 37 GHz centre with a value of − 40 dB, bandwidth of 3.6 GHz. For Mode 3, a broadband resonance centred at 27 GHz was obtained, with an S₁₁ value of − 21 dB and a bandwidth of 1.1 GHz. In addition, two other resonances were observed at 36 and 39.9 GHz, with S₁₁ values of − 16 dB and − 10 dB, respectively. For Mode 4 (1/1), the first resonance occurred at 25 GHz, with a bandwidth of 1.4 GHz. Another resonance was detected at 28 GHz with an S₁₁ value of − 14 dB and a bandwidth of 500 MHz. The resonance at 33 GHz showed an S₁₁ value of − 24 dB and a bandwidth of 2.3 GHz. Additionally, further resonances were recorded at 30.7 GHz and 34.7 GHz, with S₁₁ values of − 18 dB and − 30 dB, respectively. Minor discrepancies observed between the simulated and measured resonance frequencies are attributed to practical factors such as fabrication tolerances, PIN diode soldering inaccuracies, measurement setup variations and slight deviations in the actual material properties, compared to those assumed in the simulation (Table 2).

Full size image

The measured radiation patterns of the antenna are shown in Fig. 8 together with simulated patterns. These radiation patterns are generally in agreement with each other. The observed discrepancies in the measured radiation patterns are primarily attributed to laboratory-specific factors such as residual reflections from surrounding surfaces, measurement setup misalignments and the lack of an anechoic environment, which introduce uncertainty despite calibration efforts. Note that the measurement setup does not include calibrated reference antennas or absolute power detection; thus, the radiation patterns represent relative behaviour rather than absolute gain or beamwidth values. For this reason, the E-plane was selected as the sole measurement plane, as it provides the most representative radiation behaviour of the antenna while conforming to the limitations of the measurement setup. The deviation observed specifically at 39.7 GHz may be related to increased alignment sensitivity and more pronounced parasitic effects from the biasing circuitry at this frequency. Since the bias network is modelled using ideal lumped RLC elements in the simulation, such physical parasitic influences may not be fully captured.

Full size image

In the literature, the average gain level for a conventional single-element patch antenna is between 2 and 6 dBi. The frequency reconfigurable antenna literature comparison table is given in Table 3. Using 3 PIN diodes at 28 and 38 GHz, the maximum gain level is 10.2 dBi, but it has only 2 resonance frequencies [27]. The main disadvantages of frequency reconfigurable antennas operating at a limited number of resonant frequencies include limited frequency coverage, limitations in broadband applications, low data rate, efficiency issues and difficulties in adapting to future technologies. In another study, the maximum gain level with a resonance frequency of 28 and 38 GHz was 9.01, but 18 switching elements PIN diodes were used [29]. Having a large number of electronic switching elements has disadvantages such as power loss, creation of parasitic capacitance and inductance, electromagnetic compatibility (EMC), cost and complexity, temperature and safety problems.

In the study carried out, using only 2 PIN diodes as switching elements, the \({S}_{11}\) value is below − 10 dB at almost all frequencies in the 26–40 GHz range. In this design, the maximum gain value of 9.26 dBi is observed, and this level indicates a good gain value. The fact that the study has a wide frequency band, the use of a small number of electronic switching elements, high gain and high bandwidth, etc., all in a single study, distinguishes this study from other studies in the literature. These performance metrics, particularly high gain and wide bandwidth achieved with a low number of switching elements, are directly linked to enhanced 5G performance such as increased coverage range, improved signal quality, higher data rates and reduced energy consumption. Therefore, the proposed antenna design offers practical advantages for compact, low-power and high-speed 5G communication systems.

In this study, a frequency-reconfigurable Libra shape antenna, which has made significant progress in the field of wireless communication, has been designed, analysed in a simulation environment and tested. The antenna caters to the growing demand for portable electronic devices operating across various frequency bands, adapting to dynamic communication environments. It ensures a seamless transition between multiple (26–40) frequency bands to meet this need, enhancing coverage, reliability and data transfer speeds.

Between 26 and 40 GHz, the S₁₁ value of the antenna is below − 10 dB, indicating that the antenna operates almost continuously in this frequency range. The frequency range of 26 to 40 GHz, commonly referred to as the millimeter wave spectrum, is used for various applications in telecommunications (5G), satellite communications, radar systems, scientific research and industrial and medical applications.

The experimental results supported the simulation results and confirmed that the antenna can achieve optimum performance in the specified frequency ranges. In particular, the maximum observed gain is 9.26 dBi, which is significantly higher than the average gain levels reported for conventional single-element patch antennas and further emphasizes the effectiveness of our design.

In conclusion, the frequency-reconfigurable Libra shape antenna represents a significant step forward in antenna technology and offers a practical solution to the challenges posed by the ever-evolving wireless communication landscape. Future work could focus on further optimising the design of the antenna for additional frequency bands, exploring alternative reconfigurability techniques and integrating advanced materials to improve performance. This research lays the groundwork for the continued development of adaptive antennas that can meet the growing demands of next-generation wireless applications.