High-index dielectrics are widely used in microwave antennas to control the radiation characteristics. Liquid water, with a high dielectric index at microwave frequency, is an interesting material to achieving tunable functionalities. Here, we demonstrate a water-loaded microwave antenna system that has high loss-tolerance and wideband tunability enabled by fluidity. Our simulation and experimental results show that the resonance frequency can be effectively tuned by the size of loading water. Furthermore, the antenna systems with water loading can achieve high radiation efficiency (>90%) in the ultra-high-frequency (0.3–3 GHz) band. This work brings about opportunities in realistic tunable microwave antenna designs enabled by liquid.

Microwave antennas play an integral part of modern communication technologies, including Wi-Fi, cell phones, and radar systems. Low-loss dielectric materials are widely applied in dielectric resonator antennas (DRAs) to control the radiation characteristics.1,2 However, the challenge is to find materials that have high refractive indexes, low loss, and low cost simultaneously. For conventional solid dielectrics (e.g., ceramics), high-index materials (i.e., with a relative permittivity over 70) usually have larger loss tangent (tanδ>104). Some dielectric materials with low loss require fine grains and/or homogeneous microstructures, which increases the cost.3 On the other hand, a DRA that utilizes a solid dielectric material to control the radiation characteristics is typically fixed, and the antenna geometry cannot be changed conveniently, which limits the tunability of the devices.

In hopes of developing an alternative to the existing systems, a lot of research attention has been paid to microwave antennas with water recently. Water is an interesting yet promising candidate for many antenna applications: it is low-cost, environmental friendly, fluidic, and has a high dielectric index of around 80 in or below the ultra-high-frequency (UHF, 0.3–3 GHz) band.4 Some of those works studied the prototypes of seawater antennas, in which metallic electrodes were immersed in the seawater to drive the system.5–7 Some other works used pure water to form tunable microwave resonators, taking advantage of water's high dielectric constant and fluidity;5,8–11 they all achieved reconfigurable antennas by adjusting the water level/volume. However, seawater is a poor conductor at microwave frequencies, and pure water has a loss tangent at the order of 0.1 in the UHF band (Ref. 4), about 2–3 orders of magnitude higher than conventional dielectric materials used in antenna systems (e.g., BaTi4O9 family3). In the previous attempts, achieving high efficiency with water remains to be a major challenge, and the working frequencies were typically limited below the UHF band (supplementary material).

To make a water-loaded microwave antenna with high efficiency and enhanced bandwidths, we propose three major design principles: (1) avoid using seawater as the conducting element because seawater's Ohmic resistance is very high especially at microwave frequency; (2) avoid strong field enhancement in the dielectrics, such as patch antenna-like designs where dielectrics are sandwiched between metal electrodes; and (3) the Ohmic resistance should not scale with the volume of water, so that tuning via adjusting the water level/volume will not cost extra energy loss.

Following these design principles, we have designed, fabricated, tested a water-loaded microwave antenna and demonstrate its performance in the UHF band. Figure 1 illustrates the configuration of our antenna design and the photograph of the experimental setup. The main radiating element is a dipole, with the feed point (i.e., with 50 Ω impedance) in the center. The dipole is loaded with a bowtie shape water cell placed close to the dipole in the near field (detailed geometry parameters can be found in Table S1 in the supplementary material). Without loss of generality, we set the height and width of one bowtie arm to be equal [Fig. 1(b)] since the apex angle only changes the effective loading size.12 The bowtie shape is used to suppress dielectric loss for two reasons. First, the electric field inside the water cell is not enhanced, i.e., the water is not sandwiched between metallic electrodes and the gap in the middle of the bowtie helps avoid the strong electric field at the feed of the dipole. Second, its Ohmic resistance will not increase with volume because the ratio of the height over effective cross-section remains constant for a set of similar triangles.

FIG. 1.

Schematic of the water-loaded antenna system and experimental configuration. (a) Rendering of an antenna system. The blue bowtie is the water cell, and the yellow wire is the metallic dipole. (b) The geometry of two side-views. The gap and thickness of water bowties are constants. (c) The photograph of the device-under-test (DUT) in the microwave anechoic chamber (the inset shows a closer look).

FIG. 1.

Schematic of the water-loaded antenna system and experimental configuration. (a) Rendering of an antenna system. The blue bowtie is the water cell, and the yellow wire is the metallic dipole. (b) The geometry of two side-views. The gap and thickness of water bowties are constants. (c) The photograph of the device-under-test (DUT) in the microwave anechoic chamber (the inset shows a closer look).

Close modal

This antenna system takes full advantage of unique properties of a liquid loaded antenna for the enhanced frequency of operation. By changing the water level/volume, the operating frequency and hence the radiation characteristic can be tuned effectively: the shift of resonance frequency enables operation over a wider range of frequencies than provided by the metallic element itself. In principle, the geometry of water can be conveniently controlled via pressure control techniques13 for the frequency range that we are concerned here. With a hollow container, water can also be dynamically filled up to any desired level. For higher frequency applications such as 5G cellular networks,14 the microfluid techniques would be ready to enable adaptive antennas with liquids.15 We believe that this work paves the way for a realizable water-loaded microwave antenna, which can be a flexible and economical supplement to the current microwave antenna systems.

We used CST Microwave StudioTM to implement the initial design and full-wave electromagnetic simulations. The configuration follows Fig. 1(b), and the material property data were directly adopted from the CST material library (Fig. S3) which is consistent with published water property data at the UHF band.4 We evaluate three parameters of our antenna system to show its performance:16 reflection coefficient (S11), radiation efficiency, and realized gain. The impedance bandwidth is usually defined as the frequency band where S11<10dB. Hence, the change in the reflection coefficient manifests the tuning of working frequency. The radiation efficiency shows energy dissipation in the antenna system, which is essential in antenna systems with lossy dielectrics. The realized gain provides information of the radiation pattern and intensity.

Figure 2 shows how the radiation characteristics can be tuned by varying the size of distilled water bowties. In Figs. 2(a)–2(c), the performance of the baseline, 100 mm dipole alone is illustrated in solid grey. The solid red traces with various shades illustrate the performance of the water-loaded antenna (i.e., dipole plus water-loaded bowtie of varying sizes). In Fig. 2(a), the S11 illustrates how, as the size of the water bowtie is increased, the resonance frequency of the water-loaded antenna systems is decreased. These results are consistent with the knowledge that dielectric loading on metallic antennas has a universal feature of lowering the resonance frequency, which is the foundation of antenna miniaturization.17 Researchers have been engineering the geometry of dielectrics for decades in DRA,1,2,18–20 but liquids' unique properties enable dynamic tuning in a highly effective way.

FIG. 2.

CST simulations illustrating the performance of the water-loaded antenna. (a) When a water bowtie is loaded to a dipole antenna, the resonance frequency can be tuned by the size of the water bowtie. (b) The envelope of the multiple red curves in (a), demonstrating the tuning potential. (c) The change in realized gain at boresight follows the trend in (a) with a high value. (d)–(f) The radiation efficiency of the water bowtie (distilled water and seawater) loaded on different dipole antennas. (g) Electric field intensity of the water bowtie loading on a 100 mm dipole. Upper and lower panels in (d) correspond to circular and triangular dots in (a), respectively.

FIG. 2.

CST simulations illustrating the performance of the water-loaded antenna. (a) When a water bowtie is loaded to a dipole antenna, the resonance frequency can be tuned by the size of the water bowtie. (b) The envelope of the multiple red curves in (a), demonstrating the tuning potential. (c) The change in realized gain at boresight follows the trend in (a) with a high value. (d)–(f) The radiation efficiency of the water bowtie (distilled water and seawater) loaded on different dipole antennas. (g) Electric field intensity of the water bowtie loading on a 100 mm dipole. Upper and lower panels in (d) correspond to circular and triangular dots in (a), respectively.

Close modal

It is worth mentioning that loading of dielectrics could help match the antenna to the transmission line without matching circuits, and DRA uses this principle to achieve broadband as well.21 We can observe a similar phenomenon here: in Fig. 2(a), the loaded antenna typically has better impedance matching (lower S11) than dipole antenna itself until the size of the water bowtie becomes too big to have efficient coupling. However, the bandwidth increase of our design is mainly due to the shift of resonance frequency rather than just better impedance matching. The mechanism has a lot more flexibility because the system is reconfigurable.

Although the investigation shown in Fig. 2(a) can be regarded as “sampling” in a continuous tuning process realized by the liquidity of water, we can safely anticipate that the envelope of the red curves in Fig. 2(a) represents the dynamic tuning of distilled water loadings. We plot such an envelope in Fig. 2(b) to show the potential of tuning with a 100 mm dipole.

The realized gain at boresight shown in Fig. 2(c) follows the trend of Fig. 2(a). This provides some indication that the coupling of the water bowtie and dipole has little influence of the radiation pattern. Furthermore, it is anticipated from the symmetry of the antenna system that the radiation characteristics will not deviate much from original dipole antennas.16 

In Figs. 2(d)–2(f), we compare radiation efficiencies of devices covering a wide frequency range in the UHF band (reflection coefficient, total efficiency, and realized gain plots can be found in the supplementary material). The frequency ranges in each panel are within their respective impedance bandwidths. These efficiency plots assume continuous tuning as in Fig. 2(b), where each point corresponds to a resonance state. We investigated how the material loss would affect the performance by using both distilled water and seawater as loading dielectrics. In the frequency range of interest, the imaginary part of the relative permittivity (εi) of seawater is in the range of 105.1–39.6 (Ref. 4), much larger than that of distilled water (2.4–6.2). The seawater-loaded antenna is introduced as an extreme case in terms of dielectric loss of available water sources. Note that seawater here is not acting as a conducting element as in Refs. 5–7.

As shown in Figs. 2(d)–2(f), distilled water loading in general has radiation efficiencies around or above 95%. Even with lossy seawater, the radiation efficiency can reach 80% at the high frequency end. These results prove the suppression of energy dissipation in our antenna system and the practicality of antennas with water. The working frequency covers the majority of the UHF band with the configurations that we investigated. In principle, even a wider frequency band can be achieved if we use proper dipole antennas.

The main reasons for the high efficiency are as follows: (1) no confinement of the electric field by the sandwich structure; and (2) leaving a gap in the middle, where the field intensity is the strongest, directly limits the field penetration. Figure 2(g) shows the electric field intensities of two configurations of the water-loaded antenna, both with the fixed 100 mm dipole but loaded with 90-mm-long and 60-mm-long water bowties, operating at 1.1 GHz and 1.28 GHz, respectively. The field intensity inside water is minimized (note that the field is plotted in logarithm). Consequently, the displacement current inside water is small, and energy dissipation is suppressed.

Overall, these results presented in Fig. 2 indicate that for a given length constraint, the water-loaded antenna can operate efficiently at lower frequency than a dipole-only element, with minimum size extension. This tuning process with water loading is different from conventional LC tuning by an electrical matching circuit. A reactive matching circuit cannot compensate for the reduction of the radiation resistance of the dipole when tuned to lower/higher resonance frequency. The water tuning, on the contrary, will maintain the radiation resistance as the resonance is shifted. As a result, the realized radiation efficiency (radiated power divided by incident power) remains high in our tuning method.

To complement the computer simulation results, we fabricated and tested the dipole only and the water-loaded antenna in a shielded antenna anechoic chamber. Figure 1(c) shows a photograph of the experimental setup. The bowtie samples are hollow acrylic structures with minimum footprint, where water can be injected/drained via a small hole on the top. The distance between the bowtie and the metallic dipole was kept being 1.5 mm in all the measurements (the thickness of the acrylic case). A set of bowties with various sizes were fabricated following the simulation to demonstrate how the radiation is modulated. We measured the reflection coefficient (S11) and far field realized gain through a vector network analyzer (VNA). The realized gain was calculated using a gain substitution method.24 Details about experimental procedures and sample preparations can be found in the supplementary material.

The measured S11 results are shown in Figs. 3(a) and 3(b), for two different dipole lengths (100 mm and 150 mm). The results from dipole antennas without loading coincide with simulation results. The results from the distilled-water-loaded antenna exhibit a similar trend whereby the resonance frequency decreases as the amount of water loading increases (i.e., move to the left). Although the measured S11 is not as good as the simulated case because of practical implementation (i.e., cable mismatch, non-perfect calibration, and configuration variations), the trend illustrates that changing the size of the water bowtie is an effective way to tune the antenna.

FIG. 3.

(a) and (b) Reflection coefficient (S11) measured with water bowties loaded on a 150 mm dipole and a 100 mm dipole, respectively. (c) Realized gain at boresight measured with water bowties loaded on a 100 mm dipole. (d) and (e) Two characteristic radiation patterns of the E-plane and the H-plane, respectively. The corresponding data points are marked in (c) as well.

FIG. 3.

(a) and (b) Reflection coefficient (S11) measured with water bowties loaded on a 150 mm dipole and a 100 mm dipole, respectively. (c) Realized gain at boresight measured with water bowties loaded on a 100 mm dipole. (d) and (e) Two characteristic radiation patterns of the E-plane and the H-plane, respectively. The corresponding data points are marked in (c) as well.

Close modal

Figure 3(c) shows the realized gain at boresight for the 100 mm dipole antenna and the water-loaded antenna, along with the simulated results plotted directly from the CST simulations.25 The result from dipole antennas without loading coincides with the simulation result. When the amount of water loading increases, the resonance frequency decreases without compromising the performance: similar energy density can be delivered to (received from) boresight rather than dissipated in the water body.

To investigate the sensitivity of the water-loaded antenna system against dielectric loss, we loaded the bowtie cell with tap water as well. In the frequency range of interest (0.6–2.6 GHz), the εi of tap water is typical in the range of 3–15 (Refs. 22 and 23). The dielectric loss of tap water would be larger than that of distilled water, but it is much cheaper and more available. As can be seen within Fig. 3(c), the difference is negligible between the measured realized gain of distilled water loading and tap water loading. This result further proves the high tolerance against dielectric loss in our antenna system.

For the radiation pattern characterization of the various antenna configurations, E-plane and H-plane cuts were performed. Figure 3(d) (E-plane) and Fig. 3(e) (H-plane) illustrate the measured patterns for two water-loaded antenna configurations with 50-mm-long and 70-mm-long bowtie cells at their respective resonance frequencies (as determined from the minimum reflection coefficient). The data shown in Figs. 3(d) and 3(e) demonstrate that the measured radiation performance of the water-loaded antenna is in good agreement with the idealized dipole performance and exhibits the typical “donut” type radiation characteristics.

The radiation efficiency can be estimated by looking at the radiation pattern. The realized gain [Fig. 3(c)], E-plane [Fig. 3(d)], and H-plane [Fig. 3(e)] show that the water-loaded antenna system shares the “donut” shape and radiation intensity as dipole antennas. From the fact that the radiation patterns and realized gain coincide, when comparing the anechoic chamber measurements and CST simulations, we can deduce that the radiation efficiency of the water-loaded antenna systems should follow the CST simulation results (Fig. 2), which is above 90% in the UHF band. This conclusion applies to devices with distilled water and tap water. It also illustrates that the water-loading tuning mechanism has advantages over an electronic matching circuit tuning in terms of total energy efficiency because no energy is dissipated from the matching resistance, and the Ohmic loss in water cell is minimized.

This paper puts forward a water-loaded microwave antenna design with a unique ability to tune its frequency of operation with high efficiency. The water-loaded antenna consists of a metallic dipole element and water bowtie cells. The antenna design and results are supported with high-fidelity simulations and measured performance in the anechoic chamber. Those results illustrate high efficiency and enhanced bandwidth in the UHF band. The unique ability to tune the water-loaded antenna with a liquid material at the electromagnetic wave level (as compared to an electronic component level) provides unique opportunities for microwave antennas. Although our investigation is within the UHF band, the antenna system with water should work at other frequency bands as well where water permittivity remains high. The sizes of the water body and the emitter should simply scale with the working wavelength. We believe that this work brings about a practical design to include water in microwave antenna systems and a new degree of freedom of antenna tuning in general.

We hope that this work would encourage more explorations on antenna systems with low-cost liquid materials. In the case of water loading, there is still room for design optimization: better geometry, dynamically controlled loading cell, reduction of plastic usage via mold injection, etc., will be future interesting topics. Specifically, water might cause an additional weight to the antenna system and compromise portability. Nevertheless, our design is highly suitable for locations with sufficient water supplies. Another practical scenario is in systems where cooling water is used.

See supplementary material for device geometry, more plots of antenna performance, permittivity of water, sample fabrication procedure, and details of experimental setups.

This work was supported by the National Science Foundation under Grant No. 1555336. The authors would like to thank Kshitij Avinash Lele, Oho University graduate student, for his assistance in the shielded antenna anechoic measurements performed at Ohio University. The author would also like to thank Professor Eli Yablonovitch for the insightful discussions.

1.
J.
Galejs
,
J. Res. Natl. Bur. Stand.
66
,
557
562
(
1962
).
2.
R. K.
Mongia
and
P.
Bhartia
,
Int. J. RF Microwave Comput.-Aided Eng.
4
,
230
247
(
1994
).
3.
M. T.
Sebastian
,
Dielectric Materials for Wireless Communication
(
Elsevier
,
Oxford
,
2010
).
4.
T.
Meissner
and
F.
Wentz
,
J. IEEE Trans. Geosci. Remote Sens.
42
,
1836
1849
(
2004
).
5.
Y. H.
Qian
and
Q. X.
Chu
,
IEEE Antennas Wireless Propag. Lett.
16
,
360
363
(
2017
).
6.
C.
Hua
,
Z.
Shen
, and
J.
Lu
,
IEEE Trans. Antennas Propag.
62
,
5968
5973
(
2014
).
7.
L.
Xing
,
Y.
Huang
,
S. S.
Alja'afreh
, and
S. J.
Boyes
, in
Antennas and Propagation Conference (LAPC)
(IEEE,
2012
), pp.
1
4
.
8.
C.
Lin
,
G.
Li
,
P.
Liu
,
Y.
Qin
, and
Y.
Huang
,
in 2016 International Symposium on Antennas and Propagation (ISAP)
(IEEE,
2016
), pp.
114
115
.
9.
M.
Zou
,
J.
Pan
, and
Z.
Shen
, in
2016 17th International Symposium on Antenna Technology and Applied Electromagnetics (ANTEM)
(IEEE,
2016
), pp.
1
4
.
10.
M.
Zou
,
Z.
Shen
, and
J.
Pan
,
Appl. Phys. Lett.
108
,
014102
(
2016
).
11.
L.
Xing
,
Y.
Huang
,
Q.
Xu
, and
S.
Alja'afreh
,
Electron. Lett.
51
,
1958
1960
(
2015
).
12.
L. Z.
Thamae
and
Z.
Wu
,
IEEE Trans. Antennas Propag.
58
,
3707
3710
(
2010
).
13.
L.
Yan
,
P.
Wu
,
Q.
Song
,
W.
Zhu
,
Z.
Wu
,
D.
Tsai
,
F.
Capasso
, and
A.
Liu
, in
CLEO: QELS Fundamental Science
(OSA,
2015
).
14.
M.
Tehrani
,
M.
Uysal
, and
H.
Yanikomeroglu
,
IEEE Commun. Mag.
52
,
86
92
(
2014
).
15.
G.
Whitesides
,
Nature
442
,
368
373
(
2006
).
16.
C. A.
Balanis
,
Antenna Theory: Analysis and Design
(
John Wiley & Sons, Inc
.,
Hoboken
,
2005
).
17.
R.
Chatterjee
, NASA STI/Recon Technical Report No. Report A (
1985
), p.
86
.
18.
S. R.
Best
, in
2011 International Workshop on Antenna Technology (iWAT)
(IEEE,
2011
), pp.
90
93
.
19.
D.
Guha
,
Y. M.
Antar
,
A.
Ittipiboon
,
A.
Petosa
, and
D.
Lee
,
IEEE Antennas Wireless Propag. Lett.
5
,
373
376
(
2006
).
20.
D.
Guha
,
B.
Gupta
, and
Y. M.
Antar
,
IEEE Trans. Antennas Propag.
60
,
393
398
(
2012
).
21.
M.
Lapierre
,
Y. M.
Antar
,
A.
Ittipiboon
, and
A.
Petosa
,
IEEE Microwave Wireless Compon. Lett.
15
,
7
9
(
2005
).
22.
R. K.
Crane
,
Propagation Handbook for Wireless Communication System Design
(
CRC Press
,
Boca Raton
,
2003
).
23.
V.
Komarov
,
S.
Wang
, and
J.
Tang
,
Encyclopedia of RF and Microwave Engineering
(
John Wiley & Sons, Inc
.,
Hoboken
,
2005
), Vol.
4
, pp.
3694
3711
.
24.
IEEE Standard 149-1979 (R2008)
,
IEEE Standard Test Procedures for Antennas
(
IEEE
,
1979
), pp.
84
85
.
25.
See http://www.cst.com for “CST - Computer Simulation Technology” (
2018
); accessed 30 December 2017.

Supplementary Material