Modeling of a Negative Refractive Index Metamaterial Unit-Cell and Array for Aircraft Surveillance Applications

In this paper, a unique negative refractive index-metamaterial (NRI-MTM) structure is proposed for traffic alert and collision avoidance system (TCAS) for aircraft application. The NRI-MTM unit-cell structure is a combination of square and circular split ring structures. The proposed square and circular split ring metamaterial (SCM) is designed on the top of a 1.6 mm thick FR-4 substrate, with a single circular split ring on the bottom of the substrate. The total dimension of the SCM unit-cell is 23 <inline-formula> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> 23mm2 (<inline-formula> <tex-math notation="LaTeX">$0.08\lambda \,\,\times \,\,0.08 \lambda$ </tex-math></inline-formula>), resonating at 1.06 GHz. This unit-cell is further used to create a 5 <inline-formula> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> 4 array structure with a dimension of 130 <inline-formula> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> 111mm2 (<inline-formula> <tex-math notation="LaTeX">$0.46\lambda \,\,\times \,\,0.4 \lambda$ </tex-math></inline-formula>). A High-Frequency Structure Simulator (HFSS) is used to assess the electromagnetic properties of the proposed structures. Furthermore, in order to analyse the double negative (DNG) behaviour of the designed SCM structure, the unit-cell is rotated from 0 to 360 degrees with a step of 60 degrees along the azimuthal plane. On doing so, it is observed that, the proposed SCM structures, including unit-cell and array structures, exhibit DNG and NRI properties over the frequency band of 0.5 to 1.2 GHz, thus making it optimum for aircraft surveillance application. The SCM unit-cell and 5 <inline-formula> <tex-math notation="LaTeX">$\times $ </tex-math></inline-formula> 4 SCM array structure were fabricated, and the measured results are in well agreement with the simulated results. Further, the array structure is used as a superstrate on the patch antenna to analyze the impact of the MTM on the antenna performance. In comparison to the antenna without the MTM structure, the <inline-formula> <tex-math notation="LaTeX">$5\times $ </tex-math></inline-formula> 4 SCM array loaded antenna has a gain improvement of 2.8dB and a bandwidth enhancement of 31.5MHz.

an overall dimension less than the operating wavelength 26 and exhibits exotic properties like negative permittivity (ε), 27 The associate editor coordinating the review of this manuscript and approving it for publication was Guido Valerio . negative permeability (µ), and negative refractive index (n) 28 is well known as metamaterial (MTM). 29 The conception of MTM inspired many researchers to 30 construct a distinctive contribution in the research field. The 31 exceptional behavior of MTM is not determined by the base 32 materials; rather, its properties are influenced by the pre-33 cise shape, size, geometry, and orientations. From a prag-34 matic perspective, MTM offers some special electromagnetic 35 (EM) features that are not typically available in conventional 36 materials [1]. 37 In contrast to natural materials, the MTM interacts with 38 both light and sound waves in unanticipated patterns. 39 exhibits the negative 'ε' [20]. Later, D.R Smith et al proposed 79 the new MTM structure with negative 'ε' and negative 'µ' 80 in the year 2000, and it was experimentally tested to obtain 81 its exceptional properties [21]. The MTM slab having two 82 dimensional (2D) array of uniformly reiterated unit-cells of 83 SRR and copper strips are experimentally validated for neg-84 ative refraction by Shelby et al in 2001 [22]. 85 A novel MTM unit-cell for satellite applications was pro-86 posed in [23]. The proposed MTM structure was developed 87 on a 1.6mm thick FR-4 dielectric substrate with an overall 88 dimension of 30 × 30mm 2 . For the parametric analysis and 89 comparative study, different substrate materials such as lossy 90 Polymide, Aluminum Nitride, and Rogers RT 6010 were 91 utilized. The combination of SRR and a capacitive loaded 92 strip (CLS) MTM was presented for L and S frequency band 93 operation in [24]. The primary objective of the proposed 94 MTM is to improve the gain of the patch antenna. The 95 inclusion of CLS has added additional capacitance to the 96 structure, resulting in a lower stop-band. 97 A novel three different triangular-shaped SRR with CLS 98 was proposed in [25]. The two coupled triangular SRR 99 was combined with CLS to generate the DNG character 100 for the frequency spectrum ranging from 2.5GHz to 6GHz. 101 All three triangle MTM configurations were simulated on 102 2 separate substrates, FR-4 and RT/Duroid 5880, in order to 103 analyze and compare the parameters. A novel MTM unit-cell 104 comprising a circular-shaped single-loop resonator (SLR) 105 and a pair of short capacitor-loaded strips was proposed 106 in [26]. The proposed MTM structure was developed on a 107 0.8mm thick Rogers 6006 substrate with an overall dimension 108 of 15 × 15 mm 2 . This novel structure exhibits DNG charac-109 teristics in the frequency range of 0.5GHz to 3GHz, making 110 it appropriate for L and S-band applications.

111
A circular, square, triangular, and hybrid single-loop res-112 onator (SLR) MTMs was proposed in [27]. All four MTM 113 structures were simulated on a 1mm thick Rogers 3003 sub-114 strate. Initially, circular, square, and triangular MTMs were 115 developed, with all of the designs exhibiting multi-band oper-116 ation with a relatively narrow bandwidth. To improve band-117 width and expand the number of frequency bands, a hybrid 118 SLR was developed by combining square and triangle-shaped 119 resonators. The new hybrid SLR design has increased the 120 number of frequency bands from three to four, consider-121 ably improving bandwidth. The proposed MTM structure is 122 suitable for S and L bands with a frequency ranging from 123 2.5 GHz to 5.5 GHz. Similarly, for L and S frequency band 124 applications, a significant number of DNG MTM structures 125 have been proposed, including S-shape [ The article is organized as follows: Section II details 160 the theoretical analysis of the proposed SCM unit-cell, 161 Section III describes design and the characteristic analysis 162 of the SCM unit-cell, Section IV discusses the results and 163 discussion of the proposed SCM structures, and Section V 164 concludes the work with potential future directions.

166
The SCM is a hybrid MTM structure. The standard expres-167 sions are used to construct each square and circular SRRs. 168 Initially, the theoretical procedure for constructing the outer 169 square SRR with a single gap is described. Later, the proce-170 dure adopted to construct the circular SRRs is discussed.

171
The structure and cross-section of the single outer square 172 SRR are depicted in Figure 2(a). At the resonating frequency, 173 the square SRR will be represented as a parallel LC circuit as 174 shown in Figure 2(b).

175
The overall capacitance (C 1 ) is the summation of the 176 surface capacitance (C Surf ) and the gap capacitance (C gap ).

177
When a square SRR structure is exposed to a magnetic field 178 imposed across the y-axis, an electromotive force (EMF) 179 is induced across the square SRR, making the structure to 180 function like an LC network with a resonating frequency (f 01 ) 181 expressed as [34] 182 where, L 1 is the inductance which represents the total length 184 of the square ring, and C 1 is the total capacitance. The inductance L 1 and capacitance C 1 are expressed as where a m denotes the mean length of the square ring, 188 a m = a + w 2 , h is height, and w is thickness.

189
The total capacitance (C 1 ), which is the combination of gap 190 and the surface capacitance is given by where the expressions for C gap and C Surf is given by where g denotes split gap width of square SRR.

196
By substituting the values µ 0 = 4π × 10 −7 N/A 2 , 197 a = 11.25mm, h = 1.6mm, w = 1mm, g = 0.5mm, and 198 ε 0 = 8.85 × 10 −12 F/m, an inductance (L 1 ) value of 71.32 nH 199 and capacitance (C 1 ) value of 685.97 pF were theoretically 200 recorded for the outer square SRR structure. Hence, the f 01 201 for the outermost ring without coupling effect is 0.72 GHz. 202 To find the f 01 with coupling effect, the separation distance 203 between outermost SRR and the consecutive SRR needs to 204 be included in the calculation. In our design, the distance 205 between two SRR is 5.6mm, hence, the f 01 for the out-206 ermost ring is 1.61 GHz which is close to the simulated 207 value (1.7GHz).

208
The structure and cross-section of a single split inner 209 circular SRR is depicted in Figure 3  structure, the resonating frequency, f 02 is given by [35], [36] 212 where, L 2 is the inductance which represents the diameter of 214 the circular ring, and C 2 is the total capacitance.

215
The inductance (L 2 ) and capacitance (C 2 ) can be expressed 216 as 217 where R m is the circular ring mean radius, R m = R + w 2 .
The following expressions are used to calculate the σ and V , where θ is the angle depicted in Figure 3  The total capacitance (C 2 ) is given by, 8.85 × 10 −12 F/m, an inductance (L 2 ) value of 45.4297 nH 231 and capacitance (C 2 ) value of 473.70 pF were theoretically 232 recorded. Hence, the f 02 for the outer circular SRR structure 233 is 1.08 GHz. Since, this is the innermost SRR without any 234 following SRR structure, the resonating frequency remains 235 unchanged for with and without coupling.

245
In this section, the proposed SCM unit-cell is designed 246 and experimentally analysed inside the waveguide structure. 247 Further, the unit-cell is experimentally validated for DNG 248 material behaviour.  capacitance. An annealed copper having a conductivity (σ ) 268 of 5.8 × 10 7 S/m was used to construct the SCM structure. the unit-cell normal to E-field, and PMC (perfect magnetic 278 conductor) to the sides normal to the H-field is applied. All 279 SCM structures (unit-cell and array structures) presented in 280 this paper will be confined by the same boundary conditions. 281 For the simulation, a frequency range of 0.5 -2 GHz was 282 used. The radiation box is the simulation chamber, and its 283 dimension is 23 × 23 × 70mm 3 .

284
The proposed SCM unit-cell operates at TE 10 mode. The 285 electric field distribution, and the propagation of EM wave 286 inside the waveguide structure is illustrated in Figure 6.

287
The waveguide setup will provide the S-parameter data of 288 the designed structure, and the same data will be utilized to 289 retrieve the effective parameters such as 'ε', 'µ', and 'n'. The 290 'ε' and 'µ' are two important material qualities that determine 291 how materials polarize in electric and magnetic fields.
307 Figure 7 depicts a full five-step parametric simulation 308 used to create the final unit-cell structure, with the last stage 309 exhibiting significant transmission at 1.06 GHz. Figure 6    of an open-circuit stub on the coupled SCSRR elevated the 323 resonating frequency from 1.02GHz to 1.04GHz, as shown in 324 the S 11 and S 21 plots of step 3, and another stub on the ground 325 plane increased the resonating frequency from 0.99GHz to 326 1.06GHz, as shown in the S 11 and S 21 plots of step 5.

327
The proposed SCM unit-cell is rotated along the azimuthal 328 plane from 0 to 360 degrees with a 60-degree interval, 329 as shown in Figure 8, to further verify for DNG behav-330 ior. During these six stages, the values of effective char-331 acteristics such as 'ε', 'µ', and 'n' are recorded. From 332 Figure 8(a) to (f), it is evident that in all six stages the 333 proposed SCM exhibit negative characteristics throughout a 334 frequency band of approximately 0.5 -1.2 GHz and yield a 335 DNG response at the designed frequency, that is at 1.06 GHz. 336 As a result, the proposed SCM is considered as the double 337 negative MTM, and it is very well suited for L-band applica-338 tions. The performance of the SCM unit-cell is evaluated at 339 various azimuthal angles, and the results are summarized in 340 TABLE 2.

342
This section discusses the simulated and measured results 343 of SCM unit-cells and the number of SCM array structures. 344 Furthermore, the results of parametric analysis for various 345 split widths, distance between rings, metal width of rings, and 346 use of different substrate materials are thoroughly discussed. 347 A. FABRICATION SETUP 348 Figure 9 (a) to (e) depicts the fabricated SCM structures as 349 well as the experimental setup for evaluating MTM character-350 istics. Both the SCM unit-cell and the 5 × 4 SCM array were 351 measured using the same experimental setup. To determine 352 the ω m (magnetic resonance frequency) of the SCM struc-353 ture, the EM wave transmission through the SCM unit-cell 354 and the array structure must be measured. To measure the 355 wave transmission through the unit-cell, it is placed between 356 the two horn antennas, as shown in Figure 9(d).

357
The horn antenna and the unit-cell are separated by a 358 distance of 10 cm. Initially, the transmission spectra are 359 measured in free space and their values are recorded in 360 the absence of SCM unit-cell. Next, the recorded data was 361 VOLUME 10, 2022   To reduce the sample edge effect, the distance between the 374 horn antenna and the sample (SCM unit-cell or Complete 375 slab) must be carefully calibrated to project the majority of 376 the incident wave within the region covered by the sample. 377 Furthermore, a precise separation distance is essential for 378 measuring parameters such as transmission and reflection 379 coefficients with a high SNR (Signal to noise ratio). in Figure 10

389
A small discrepancy in magnitude and the frequency 390 of the measured and simulated results is observed in the 391 Figure 10(a). The resonances around 0.6 to 0.7GHz and 392 1.4 to 1.5GHz exhibit a frequency discrepancy between 393 the simulated and measured results, which may be due to 394 the addition of manufacturing and calibration errors with the 395 mutual coupling effect between two waveguide ports. Despite 396 these discrepancies, the fabricated result closely resembles 397 the simulated outcome. The difference between simulated 398 and measured results is also affected by the type of substrate 399 material, with approximately 5% experimental error being 400 acceptable for lossy substrate materials [38], [39].

401
In general, the difference between simulated and measured 402 data could be caused by one of the following factors: iii. Effect of the mutual resonance across the transmitting 407 and receiving waveguide ports will always modify the inter-408 pretations and cause minor variation in both data. 409 iv. The dielectric constant of the substrate material has a 410 considerable impact on the measurement data. In general, 411 the substrate permittivity will always influence the resonat-412 ing frequency. In general, the permittivity of the substrate 413 will always have an effect on the resonating frequency. 414 The increase in permittivity will slightly drag the peak 415 resonance points towards the lower frequencies. Further-416 more, the permittivity variation will also affect the capac-417 itance values between the ground plane and the radiating 418 patch.
419 v. The difference between the transmission and the reflec-420 tion coefficient, along with frequency shift and the break 421 in resonating frequency, is primarily caused by the fact 422 that not all of the incoming power will be reflected or 423 transmitted.

424
The real permittivity value reaches negative at 0.68 GHz 425 as illustrated in Figure 10 Figure 10 (c), and the 430 negative zone has been extended up to 1.25 GHz. The peak 431 permeability value of −28.9 dB is recorded at 0.71 GHz, 432 whereas at the resonating frequency, the permeability value is 433 −5.7 dB. Furthermore, the proposed SCM unit-cell exhibits 434 LHM characteristics in the frequency range of 0.67 GHz 435 to 1.21 GHz, as shown in Figure 10 (d). Hence, the pro-436 posed SCM structure is well suited for aircraft surveillance 437 applications.

438
The effective medium ratio (EMR) defines the compact-439 ness of any MTM structure [40]. The EMR is defined as 440 the ratio of the wavelength to the length of the MTM unit-441 cell, as given in Eq (16). To accomplish the double negative 442  This section discusses the various inter-unit cell coupling 448 effects for 1 × 2 and 2 × 2 SCM array structures. 449 Figure 11 depicts four identical 1 × 2 SCM array structures 450 with varying mutual coupling distances. The coupling dis-451 tance between SCM unit-cell is varied from 0.25mm to 1mm, 452 VOLUME 10, 2022 FIGURE 12. Coupling distance analysis for 2 × 2 SCM array structure. In this section, various array configurations of the proposed 475 SCM structure are discussed. Since a single MTM unit-cell 476 cannot manifest the rational exotic electromagnetic behavior, 477 several sets of SCM array structures, such as 2 × 1, 2 × 2, 478 and 5 × 4, are designed and analyzed for both S-parameters 479 and effective parameters. The same simulation and analysis 480 approach that was utilized for the SCM unit-cell was adopted 481 for all of the proposed SCM array structures. are separated by a distance of 1mm. Figure 14

491
The simulated data is then utilized to extract the effective 492 parameters of the SCM array structure.

493
As illustrated in Figure 14 Figure 15 (b) shows the 507 simulation setup for retrieving the S-parameters. A frequency 508 spectrum of 0.5 -2 GHz was used for the simulation.   setup for retrieving the S-parameters. A frequency spectrum 530 of 0.5 -2 GHz was used for the simulation. As depicted 531 in Figure 16 (d), the simulated 5 × 4 SCM array resonates 532 at 1.06 GHz with a magnitude of −24.2 dB, whereas the 533 measured SCM resonates at 1.06 GHz with a magnitude of 534 −22.8 dB. The simulated data is then utilized to find the 535 effective parameters of the SCM array structure. As shown 536 in Figure 16 (e), the real 'ε', 'µ', and 'n' values for a 537 5 × 4 SCM array become negative at 0.83 GHz, with ampli-538 tudes of −4.98 dB, −21.83 dB, and −9.81 dB, respectively. 539 The amplitude values of 'ε', 'µ', and 'n' for the designed fre-540 quency are −7.12 dB, −4.99 dB, and −5.90 dB, respectively. 541 The imaginary plot of the effective parameters is illustrated in 542 Figure 16 (f). This structure has a double negative region for 543 the frequency band of 0.83 -1.19 GHz. 544 Figure 17 depicts the surface current, electric (E) and 545 magnetic (H) field distributions, and vector representations 546 of the E and H field distributions of the proposed NRI-MTM 547 at 1.06GHz. The surface current distribution depicts the scat-548 tering of electrical current caused by the applied electro-549 magnetic (EM) fields. The density of the surface current is 550 maximum at the outer SRR and the upper surface of the 551 subsequent circular SRRs, as shown in Figure 17(a). Since 552 the two ring (Outer SRR and Inner SRR) currents are parallel, 553 it tries to strengthen the H-field formed by them, resulting in 554 a strong coupled H-field encircling these two rings, as seen 555 in Figure 17(d). In addition to H-field, the density of E-field 556 is particularly high in the region of the two outer SRR. The 557 strong E and H-field around the outer SRR causes E and H 558 resonance, which leads to resonances in scattering variables 559 at the operating frequency.

560
The dependence of the H-field on the current density may 561 be evaluated by comparing the distribution of the H-field 562 along with the current density. By carefully inspecting the 563 E-field distribution as illustrated in Figure 17(b), it is clear 564 that the E-field distribution is strongly connected to the 565 rate of change of the H-field, implying that a high E-field 566 VOLUME 10, 2022 H-field is high. As a result, it satisfies the E and H-field 568 relation stated by Maxwell's equation [41]. Hence, the contri-569 bution of various parts of the resonator to the EM properties 570 allows S 21 to resonate at the desired frequency. The moderate 571 current distribution is observed at the split of the circular SRR 572 placed on the ground plane. The E-field distribution and its 573 vector representation for SCM unit-cell, 2 × 1 SCM array, 574 and 2 × 2 SCM array at the resonant frequency are shown 575 in Figure 17(b). The E-field distribution is particularly dense     As shown in Figure 19, the spacing between the square and 626 circular rings was varied from 8 mm to 10 mm, and the corre-627 sponding simulated values of S 21 and resonating frequency 628 for each gap variation were recorded. Increased distance 629 between consecutive rings reduces overall capacitance and 630 inductance, resulting in a rise in resonating frequency. From 631 the simulated results, it is clear that the gap between the 632 consecutive rings has a major influence on the resonating 633 frequency but a negligible effect on the S 21 . For the proposed 634 structure, the consecutive rings are separated by a distance 635 of 9 mm, and an S 21 of −35.01 dB at 1.06 GHz has been 636 recorded for this width. In theory, increasing the split ring widths reduces the overall 642 mutual capacitance and inductance, resulting in a higher 643 resonating frequency. As illustrated in Figure 20, the ring 644 width was varied from 0.5 mm to 2.5 mm, and the corre-645 sponding simulated values of S 21 and resonating frequency 646 for each ring width variation were recorded. From the sim-647 ulated results, it is clear that the ring width has a significant 648 effect on both resonating frequency and S 21 . For the proposed 649 structure, the ring width of 1 mm was chosen, and an S 21 of 650 −34.98 dB at 1.06 GHz has been recorded for this width. 651       As discussed in Section III, the proposed SCM exhibits nega-682 tive 'ε', 'µ', and 'n'. The properties of DNG have vast appli-683 cations in the antenna field for enhancing the performance, 684 and in many research works, the DNG material is utilized 685 as a superstrate to enhance the antenna gain and bandwidth. 686 Hence, in order to assess the proposed NRI-MTM, an RMPA 687 resonating at 1.06GHz is designed using HFSS software as 688 illustrated in Figure 22.

689
TCAS is a surveillance system mandated in all aircrafts 690 to prevent air accidents. It utilizes a directional antenna to 691 monitor any nearby aircrafts. The frequency band of 960-692 1215MHz is allotted for the TCAS/aircraft surveillance appli-693 cations. TCAS employs 1.03GHz radio frequency to send the 694 inquiry signal for examining the range and location of all 695 neighboring aircraft, and 1.09GHz radio frequency to receive 696 the response of nearby aircraft. Hence, the RMPA is designed 697 for 1.06GHz, which is the central frequency of the TCAS 698 antenna.

699
As shown in Figure 22 (a), the microstrip feedline with 700 inset-cut feeding method is utilized to feed the antenna, 701 and the ground plane is a complete copper as shown 702 in Figure 22(b). The RT/Duroid 5880 substrate material 703 with a loss tangent (tanδ) = 0.0009, relative permittivity 704 (ε r ) = 2.2 is used to simulate the RMPA. The overall 705 dimension of the RMPA is 114 × 91 × 3mm 3 . The opti-706 mized antenna dimension is given in TABLE 7. The RMPA 707 resonates at 1.06 GHz with −19.52 dB reflection coeffi-708 cient, and 13.10MHz (1.0609-1.074GHz) bandwidth, which 709 is depicted in Figure 23. The antenna gain plot is depicted 710 in Figure 24, and the proposed RMPA has a maximum gain 711 of 4.60 dB. 712 VOLUME 10, 2022   multiple times to find the optimum distance between the 719 RMPA and the MTM slab. The better result has been noticed 720 when the NRI-MTM slab is kept at 50mm distance from the 721 RMPA. The effect of distance between the RMPA and the 722 NRI-MTM slab on the reflection coefficient and antenna gain 723 is depicted in Figure 26 and 27, respectively.

724
The distance between the radiating patch and the metama-725 terial superstrate is generally called as resonant distance (h), 726 and it is given by [44], [45]    where ϕ 1 is reflection phase of the ground plane, ϕ 2 is 729 reflection phase of the metamaterial, and λ is operating conductor, its reflection phase is close to π(ϕ 1 = π), hence 733 the Eq. can be reduced to For 1.06GHz resonating frequency, the operating wavelength 736 becomes 283mm, and the reflection phase of the radiating 737 patch can be varied from 0 to π. Hence, the maximum 738 distance between radiator and the superstrate layer is λ/2.

739
In our work, the maximum gain of 7.4 dB is obtained when 740 the distance between radiating patch and the superstrate 741 layer is λ/6.

742
As the separation distance between the RMPA and the 743 MTM slab increased from 10mm to 60mm, a considerable 744 improvement in gain and fluctuation in reflection coefficient 745 has been observed. When the separation distance is 50mm, 746 a reflection coefficient of −21.82dB, a bandwidth (BW) 747 of 34.50MHz (1.0785GHz -1.0440GHz), and a gain of 748 7.4dB has been recorded at 1.06GHz resonating frequency. 749 In comparison to the proposed RMPA, which has a reflection 750 coefficient of −19.52dB, a BW of 13.10MHz (1.0740GHz 751 -1.0609GHz), and a gain of 4.6dB, the deployment of an 752 NRI MTM slab with 5 × 4 SCM unit-cell on the RMPA has 753 improved the gain and BW of the antenna up to 2.8dB, and 754 21.40MHz, respectively. The role of the NRI MTM slab on the field distribution 756 and antenna gain improvement can be analyzed with the 757 aid of Figure 28. In the proposed RMPA, as depicted in 758 Figure 28 (a), the propagation of EM waves along the sides 759 is perpendicular to the YOX plane. The use of MTM as 760 a superstrate will distract EM wave propagation, and the 761 NRI properties of MTM will divert the flow of EM waves 762 in the horizontal direction, as depicted in Figure 28 (b). 763 As a result, the proposed NRI MTM slab as a superstrate 764 increases the density of EM wave propagation in the hori-765 zontal direction parallel to the YOX plane. The increase in 766 horizontal wave propagation caused by the MTM slab mod-767 ifies the primary beam direction, enhancing the horizontal 768 gain [39].

774
In this research, the combination of square and circular split