Low-Cost 3D-Printed Coupling-Fed Frequency Agile Fluidic Monopole Antenna System

A low-cost 3D-printed frequency agile fluidic monopole antenna system is demonstrated to respond to the increasing demand for reconfigurable antennas, which can operate in a dynamic environment, in this paper. Antennas that can be reconfigured for different operating frequencies, polarizations, or radiation patterns are attracting attention. Traditional reconfigurable antennas using a metallic radiating element with electronic switches are limited by their pre-defined physical geometries. As conductive fluid, either liquid metal or ionized fluid has no defined shape, so it is possible to create the desired shape of a fluidic antenna to support different wireless environments. The fabrication of the leakage-free containers for fluidic antennas needs special consideration, and stereo-lithography-based 3D-printing technology is a possible option to support the fabrication. Moreover, researchers will have higher design freedom and accuracy to create new container shapes for fluidic antennas. The fluidic monopole antenna proposed is coupling-fed by a ring geometry for separating the electrical and mechanical structures; such an approach enables individual optimization and minimizes mutual disturbances in the system. A parametric study of the proposed coupling-feed geometry and the experimental verification of the antenna prototypes have been performed. Reasonable frequency agility from 3.2 to 5 GHz has been demonstrated, and the peak efficiency is about 80%. A maximum gain of 3.8 dBi is obtained. The radiation patterns of the antenna are stable across the operating bandwidth. The proposed antenna could be useful for the applications in the recent 5G mid-bands operations.


I. INTRODUCTION
Signal quality plays an important role in wireless communication systems, particularly the latest Internet of things (IoT) and the 5th generation mobile network (5G) technologies. An antenna is the first and also the last component in a wireless communication system; its performance would determine the data speed and reliability of the system. The radiating elements of the conventional antennas are usually made of metals. These materials usually have high conductivity and provide strong mechanical stability. However, the shapes of these metallic radiating elements are pre-defined in the design stage, so the agility the antennas is limited. In particular, modern wireless communications The associate editor coordinating the review of this manuscript and approving it for publication was Abhishek Kandwal.
requires the antennas to be multi-functional and adaptive to the dynamic environment in both far and near proximity. Many attempts have been reported to improve the reconfigurability of metallic antennas in the areas of operating frequency [1]- [3], bandwidth [4], [5] polarization [6], [7], and radiation pattern [8], [9], this is usually achieved by using electronic devices, such as PIN diodes or Micro-Electro-Mechanical System (MEMS) switches.
Low-cost radio frequency (RF) switches are usually lossy and the metal strip biasing circuits could be difficult to be embedded with the antennas [10]. These issues may affect the antenna performance significantly [11]. On the other hand, high performance RF switches, which can operate in the microwave or millimeter-wave frequency range for modern wireless systems, could be expensive [12]. Fluid has no defined shape; theoretically it is possible to create the desirable shape of an antenna for different diversities. Recently various highly conductive liquid metals, such as Galinstan, Eutectic Gallium-Indium (EGaIn) (σ = 3.4-3.46 × 10 6 Sm −1 ) or mercury (σ = 1.0 × 10 6 Sm −1 ), based fluidic antennas had been reported [13]- [19]. In [13], as the antenna was not properly matched, total efficiency of about 45% was obtained, while the maximum measured gain of the dual-band slot antenna in [14] was 3.25 dBi. An annular slot antenna with surface integrated fluidic channels was presented in [15], the frequency tuning was achieved by using different fluids, the gain fluctuation incurred in changing the fluid may not be desirable for practical applications. In [16], the fluid was only used for disturbing the E-field in the metallic microstrip monopole antenna, this design presented difficulties for complete drainage of the distilled water from the fluidic channels to change the operating frequency of the antenna. Additionally, the proposed design does not generate a significant frequency tuning in the impedance bandwidth of the antenna. A maximum frequency shift of 25% is observed when loaded with water. Other designs proposed to use the fluid as the radiating structures of the antennas. A Yagi-Uda formation using liquid metal filled monopoles was presented in [17]. Galinstan was used to reconfigure a 5-element Yagi-Uda using 6 fluidic monopoles. By injecting the appropriate amount of fluid, the antenna could be tuned in frequency and gain, and its radiation pattern could also be steered. However, the design was actuated manually using syringes, the repeatability of antenna performance cannot be guaranteed. Furthermore, it was not presented as a continuously tuning system. The two prototypes presented operated at 2.4 GHz and 3.87 GHz. Reference [18] presented a tunable mercury monopole, the measured gain obtained was between 0.5 dBi to 2.3 dBi with a maximum simulated efficiency of 54%. This design had a multiband operation in the lower band configuration which the frequency tuning was achieved by using two micro-pumps. These liquid metal designs achieved comparable performances with the ionized solution counterparts. However, as mercury is toxic and may raise safety concerns, while Galinstan and EGaIn are highly destructive to many other metals [22]- [24], such as aluminum and copper, by dissolving them; these issues set a barrier to widely use such liquid metals in general wireless systems.
On the other hand, ionized solutions, such as sodium chloride (NaCl) and potassium chloride (KCl) solution, and distilled water are easily accessible. Most importantly they are eco-friendly, so are the good alternatives for low-cost systems [20]- [22]. The fluid can be pumped into or out from a fluid container, such as a tube, to achieve the desirable dimensions for the particular operating frequency band.
In term of the tuning mechanism, the fluid in [16], [17] was manually controlled using syringes. Other designs used micro-pumps to inject fluid into the system and makes the tuning more automated [18]- [21]. To the authors' best knowledge, an automatic control for the frequency tuning with a closed-loop system is rare. In this paper, we propose a new design of fluidic antenna system to overcome the drawbacks of the traditional fluidic monopole antennas. The newly proposed coupling feed-ring geometry has the advantage that it does not directly contact with the ionized fluid, therefore the mechanical and electrical structures are separated, it gives the researchers more flexibility in creating new shapes for fluidic antennas and more importantly water leakage can be avoided. In addition, the optimization of the mechanical and electrical structures can be performed independently. High precision stereolithography (SLA) 3D-printing technology [26] is used for fabricating the leakage proof container. Finally, the closed-loop system which can automatically adjust the operating frequency in respond to the change of the communication channel condition is novel. CST microwave studio 2016 [27] was used in the simulation of the antenna performance.   2) a 3D-printed circular Polydimethylsiloxane (PDMS) tube for containing the ionized fluid, 3) a coupling feed geometry which consists of a 50 microstrip feed line, a vertical strip, y strip = 4.5 mm and a ring wrapped around circular tube, y ring = 1 mm, and 4) A peristaltic pump, its digital control board and a Raspberry R Pi 2 [29]. The container tube for the ionized fluid is built using the standard clear resin with a Form 1+ Formlabs 3D printer. The resin is made with a mixture of methacrylic acid esters and photo-initiators that have dielectric property close to acrylic plastic. Experiments shown in Fig. 2 were set up to measure the actual dielectric parameters of the cured clear resin and the ionized solutions. An Agilent 85070E Dielectric Probe Kit was used to evaluate the dielectric constants and loss tangents of the PDMS tube and the two types of ionized solutions investigated in the paper. NaCl and KCl solutions of different concentrations from 0.1 mol/liter, 0.5 mol/liter, 1 mol/liter to 2 mol/liter were measured in the frequency range from 0.5 GHz to 8.5 GHz. The weight of the NaCl and KCl salts are tabulated in Table 1. The measured dielectric constants and loss tangents were fed back to the CST models for EM co-simulation to provide more accurate calculations for choosing the better ionized solution.
Weight of the NaCl and KCl salts for different concentrations of the solutions used in the antenna design.

III. ANTENNA OPERATION
The vertical cut of the simulated H-field of the proposed fluidic antenna is presented in Fig. 3. On the left-hand side, there is the microstrip feeding line. The radiation pattern agrees to that of a conventional monopole antenna. Moreover, it can be observed that the fluidic antenna operates like a monopole antenna with its quarter wavelength approximately equals to the height of the fluid (y fluid ). Also, as shown in Fig. 3, the source of the radiation comes through from the coupling feed-ring to the ionized solution.
A microstrip line is connected to a coupling feed-ring delivering the signal to the antenna. The top of the feeding-ring is 5 mm above the microstrip line surface. The feeding ring design allows the container to be completely watertight, as it does not require a direct contact with the ionized fluid in the tube. Also, the coupling feed provides a mechanism for wideband frequency tuning. Fig. 4a shows the fluidic antenna system, it consists of a fluidic monopole antenna, a R&S R FSH8 vector network analyzers (VNA) [29], a Raspberry Pi (RPi), a pump controller, a peristaltic pump, a plastic beaker with ionized solution serving as a reservoir, a silicone pipe connecting the antenna to the pump, a RF coaxial cables connecting the VNA to the antenna feed port, a USB cable connecting the VNA and the RPi. These components assemble the closed-loop system.

IV. FLUIDIC ANTENNA SYSTEM
To monitor and control the whole system, a user interface was developed. Fig. 4b presents the workflow of the system. Once the operating frequency is determined according to the wireless channel condition, the pre-defined height of the ionized solution, which based on the previous experimental results, stored in the SD card of RPi will be set as the initial height of the fluidic antenna. The RPi will instruct the pump controller to pump in the ionized solution from the reservoir into the 3D-printed tube. At the same time, the S 11 value of the fluidic antenna will be also monitored by the VNA and the results are feedback to the RPi for controlling the peristaltic pump operation. The pump is driven by a PWM signal using the micro-step method, which provides the high precision of flow rate and smooth rotation. A computer program is written for retrieving the S 11 values from VNA, analyzing and comparing the real time data, then sending the corresponding commands to the VNAs and pumps. Based on the measured S 11 , the RPi will decide whether to pump in or out the ionized solution to achieved the desirable operating frequency. A Satimo StarLab was used to measure the efficiency of the antennas in the operating frequency range.

V. PARAMETRIC STUDY OF THE COUPLING FEED GEOMETRY
As shown in Fig. 1a, the proposed coupling feed geometry consists of a 50 microstrip line, a vertical strip and the feed-ring. The parametric studies of the height of the vertical strip (y strip ) and the width of the feed-ring (y ring ) have been performed to investigate the S 11 of the antenna and provide the design guidelines of the proposed coupling feed geometry. In the parametric studies, the height of fluid length (y fluid ) is kept at 15 mm and the same PDMS tube model is used. The initial model is based on the optimum model described in Section II.
The S 11 of the proposed antenna at different vertical strip heights (y strip ) is shown in Fig. 5. It can be observed that the minimum height should be higher than 3 mm for the proper  monopole operation of the antenna. When the height further increases beyond 3 mm, the operating frequency decreases. This feature can be utilized for deciding the center frequency of the frequency tuning range. For example, if an antenna operating at lower frequency band is required, a higher vertical height should be selected.
The effect caused by width of the feed-ring (y ring ) has also been investigated. The models of five different feed-ring widths in the range from 0.5 mm to 1.5 mm have been simulated. It can be observed from Fig. 6 that the width of the feed-ring can be used for fine tuning the impedance matching. When the width deviates from the optimum value, i.e. 1 mm in this case, the impedance matching is worse. Fig. 7 shows the measured S 11 , in dB, from 2 to 6 GHz at different operating frequency bands of the optimized model. The simulated result for fluid height (y fluid ) equals 15 mm is also plotted in the figure for comparison. After considering the balance between conductivity and dielectric loss, ionized solution of 0.1 mol/liter NaCl was selected in the simulation and measurement. As expected the fluidic antenna system presents different operating frequency bands depending on the height of the ionized fluid that is pumped into VOLUME 7, 2019   the 3D-printed monopole antenna. The measured result also agrees reasonably with the simulated result at y fluid = 15 mm. Fig. 8 shows the measured efficiency of the antenna over frequency for the different heights of the fluid. The maximum efficiency is about 80%. Fig. 9 presents the gain in dBi from 3 to 5 GHz for different operating frequencies of the AUT. With a maximum gain of 3.8 dBi at the 4.5 GHz when the fluid height is 5 mm. The peak of the gain corresponds to the operating frequency bands shown in S 11 result. The radiation patterns of the AUT at different fluid heights are also measured to verify the performance of the antennas. Fig. 10 shows the measured radiation patterns of the AUT for different fluid heights at the frequency point of its minimum S 11 parameter presented in Fig. 7. The measurements were performed in an anechoic chamber. Two cuts in E-plane (Phi = 0 • and Phi = 90 • ) at the AUT's minimum S 11 frequency points are presented. The measured radiation patterns are stable in all operating bands and are in reasonable agreement with a conventional monopole antenna and the simulated radiation patterns considering the presence of the RF cables and adaptors in the measurement.

VII. DISCUSSION
The design process of a fluidic antenna system has been demonstrated through three parts: driving the pump to produce an accurate fluid height control, sending the command to Raspberry Pi from PC to control the pump indirectly, and obtaining the S-parameters data from VNA by PC in real time, which corresponds to the three main elements in a closed-loop system: actuator, controller and feedback.
The measurement of S 11   impedance bandwidth, providing the opportunity to control and adjust the operating frequency for different applications with a low-loss and low-cost method. The coupling feed geometry supports a wide impedance matching at the center frequency, i.e. 4 GHz, while the impedance becomes worse as the height changes. The proposed antenna can be dynamically controlled to cover different 5G mid-bands, particularly the 3.3 -3.8 GHz, 4.5 -5 GHz and 3.8 -4.2 GHz bands for mobile applications in different countries [30]. The proposed system could be suitable for multiple communications applications that require a good mixture of coverage and capacity, which is expected to form the basis of many initial 5G services. Table 2 presents the specifications of a selection of the most recently reported monopole fluidic antenna designs for the comparisons with the proposed design in this paper. The tuning frequency of the antennas are all in the GHz range. Difference gain values are obtained in different operating bands. And the method indicates the tuning method used, either manually using a syringe, with a micro-pump or with a closed-loop system. Most designs trade off gain in order to achieve wider frequency tuning range [16], [18]. Although other designs may present wider tuning frequency range, the design we proposed presents more stable gain in its operating bands with appropriate tuning frequency range for 5G mid-bands services. Efficiency in the design presented here is about 80%, which is comparable to other proposed designs (90.4% in [16], 85% in [17]) and greater than 54% in [18].

VIII. CONCLUSION
A novel closed-loop fluidic antenna system, which is able to reconfigure itself in real time to operate at a desirable working frequency for different communication environment and applications with stable transmission of signal, has been proposed. The frequency agility of fluidic monopole antenna is achieved by automatically adjusting the height of ionized solution inside the watertight 3D-printed tube. A parametric study has also been provided for references. The proposed antenna could be useful for the applications in the recently released 5G mid-bands operations.