Frequency Agile Slot Antenna Using Contactless Capacitive Loading

In this paper, a reconfigurable slot antenna capable of tuning a full octave in frequency from 2 GHz to 4 GHz is presented. The design is based on changing the capacitive loading of a slot antenna by physically displacing a metal plate connected to a piezoelectric linear actuator that changes the apparent electrical length of the antenna. The antenna was simulated, fabricated, and measured to tune across the 2–4 GHz frequency range while maintaining a maximum return loss of greater than 9.0 dB. Simulations also show a 65.74% to 95.75% radiation efficiency across S-band. A broadside realized gain of greater than 1.73 dB was measured over the entire tuning range. Moreover, the tuning mechanism was proven robust above other methods by withstanding up to 100 W of power incident on the feed at 3 GHz.

. While these systems 27 offer many benefits when it comes to reliability, incorpo-28 ration into current PCB design processes, size, and weight, 29 they suffer in one key region: power handling. Their phys-30 ical properties cause serious limitations due to breakdown 31 and undesirable non-linear effects [12]. These effects have 32 limited the adoption of these highly adaptive antennas in 33 high power applications like radar. Next generation systems 34 need to operate in narrow frequency bands without risking 35 spectrum pollution due to non-linearities. These problems 36 can be solved with mechanical actuation of metal as a tuning 37 mechanism. 38 Tuning using movement of metal has been around since 39 the earliest days of radio [13]. It has been reliably used in 40 systems such as old radios, radars, and telecommunications 41 equipment. Static mechanical tuning is currently common-42 place in microwave components, such as using screws for 43 fine tuning of cavity filters [1]. Active, electrically controlled, 44 FIGURE 1. Model of tuning mechanism on 22 × 1 mm slot. The distance, d , is varied over the entire plate to adjust the capacitive loading of the slot.
concerns with this antenna is the ability to handle high power. 100 To maximize power handling, the minimum capacitive gap 101 needs to be as large as possible to limit dielectric breakdown, 102 but this comes with a few drawbacks. As one increases the 103 minimum capacitive gap, the area of the capacitive plate must 104 increase to give an equivalent amount of loading. A large 105 metal plate above a slot antenna causes undesirable inter-106 ference in the radiation pattern. The large plate also forms 107 a less sensitive capacitor as a larger portion of the plate is 108 farther from the center, which is the location of strongest 109 E-field.

110
The second main factor is tunability. A highly loaded slot 111 antenna is much easier to tune than a lightly loaded one, 112 as comparative changes in capacitance are much more easily 113 achieved with small distances of movement. However, more 114 loading makes the antenna electrically smaller as compared 115 to its operational frequency, which consequently results in 116 more loss [22]. Therefore, a balance between amount of load-117 ing, efficiency, tunable range, and power handling must be 118 met.

119
To achieve a full octave of frequency tuning, the loading 120 capacitance needs to change by a factor of four [23]. The 121 approximate relationships for capacitance and their relation-122 ship to antenna tuning range was explored using ANSYS 123 HFSS with the geometry seen in Figure 1. The capaci-124 tive gap was tuned from 10 µm to 40 µm, resulting in a 125 resonant frequency change from 1.87 GHz to 3.02 GHz. 126 The range is less than the octave predicted by above-127 mentioned assumptions, but the discrepancy can be attributed 128 to fringing capacitance as well as the metal plate adding 129 inductive loading that lowers the resonant frequency. Con-130 veniently, given a starting distance of 10 µm, the piezo-131 electric actuator can quadruple the distance four times, 132 making an octave of tuning well within the possibility of the 133 actuator.

134
The main body of the piezoelectric actuator is a large metal 135 box. There were concerns on what effect putting a large metal 136 box above an antenna would have on its radiation pattern, 137 loading, and efficiency. To see what effect these structures 138 have, a simulated model with a mounting structure for the 139 piezoelectric actuator, a simplified piezoelectric actuator, and 140 feeding structure was created in HFSS. A Dyson balun [24] 141 was chosen to feed the slot due to its reliable and simple 142 manufacture. It was necessary to put the feeding point close 143 to the end of the slot to ensure a good impedance match. The 144 resulting model is seen in Figure 2. 145 The actuator mechanism reduced the tuning range of the 146 antenna. To compensate for this reduction in tuning range, 147 VOLUME 10, 2022 FIGURE 2. HFSS model of the slot antenna with tuning mechanism included, showing how the physical model was approximated in simulation. Grey is aluminum, green/brown is plastic ( r = 2.5, tan(δ) = 0.1), white is PVC Plastic (simulating the nylon screws), and orange is copper. The copper is on a 125-mil-thick TMM3 dielectric substrate. The coaxial cable is modeled according to the specifications of a RG402/U Coaxial Cable.  remained above 1.81 dBi for all frequencies measured. The 164 minimum broadside gain decreased 1 dBi compared to the 165 minimum of 2.70 dBi with no structure. The gain shown 166 does not include mismatch loss. In summary, the structure 167 has some effects on the performance of the antenna, but the 168 effects were small or manageable. The design still functions 169 acceptably for its intended use.   9 dB, which is a 3 dB reduction compared to the simulated 215 results.

216
The realized gain results of the fabricated slot antenna 217 including the mount, actuator and balun are shown in 218 Figures 9-11 for both the co-polarization and cross-219 polarization at 2 GHz, 3 GHz, and 4 GHz. The fabricated 220 results closely match those of the simulated structure for the 221 co-polarized case, with a small decrease in gain. The max-222 imum gain is reduced from 5.09 dBi at 3 GHz (θ = 0 • ) 223 to 4.77 dBi. An increase in cross-polarization pollution is 224 also observed. The decrease in gain and the increase in 225 cross-polarization can be attributed to the feeding coaxial 226 line, the control lines, and the fixture holding the antenna 227 in place. Overall, the radiation patterns of the fabricated slot 228 antenna are not largely impacted by the introduction of the 229 tuning structure as indicated by simulation results, which 230 indicates that this type on an antenna can maintain expected 231 VOLUME 10, 2022 FIGURE 9. Realized gain pattern and polarization comparison at 2 GHz between simulated and measured filtenna results with mount, actuator, and balun.  126 W at 3 GHz according to calculations using the break-239 down voltage of the air gap between the tuning plates [23]. 240 To verify the validity of this analysis, a high power, RF front 241 end was used as illustrated in Figure 12. An amplifier capable 242 of 100 W was employed, and was protected from reflec-243 tion with a circulator and attenuator. This was followed by 244 a stepped-impedance low pass filter to condition the inter-245 harmonic distortion of the amplifier. The signal was then 246 passed to a bi-directional coupler for assessing the incident 247 and reflected voltage of the antenna and then finally to the 248 antenna itself. Using this setup, the amplifier was operated at 249 the full 100 W with no adverse effects.

251
A tunable slot antenna allows efficient operation in a noisy 252 spectrum. The tunable antenna introduced in this paper 253 has been simulated to tune over all of S-band, and been 254 demonstrated to have a tuning ratio of 2:1. It does so while 255 SHAHROKH SAEEDI (Senior Member, IEEE)