ELO-STRIP BASED IMPEDANCE TUNING FOR DUAL-BAND SIW CAVITY-BACKED ANTENNAS FOR MMWAVE APPLICATIONS Authors: E Jeba Saral Darling, Velmathi Guruviah Published: 2026 ABSTRACT AbstractDual-band impedance matching at millimeter-wave frequencies remains challenging due to elevated dielectric losses and limited control over multiple resonances beyond 20 GHz. This work introduces the ELOngated strip (ELO-strip), a compact, substrate-less copper strip galvanically connected to a substrate-integrated waveguide (SIW) cavity-backed slot antenna for controlled dual-band operation at 27 GHz and 39.8 GHz. It perturbs the cavity current distribution by extending the effective electrical path and introducing distributed inductive–capacitive loading within the SIW resonator. This electromagnetic redistribution enhances impedance coupling and significantly broadens the achievable impedance bandwidth while maintaining compact cavity dimensions. The integration of two shorting pins modifies the electromagnetic boundary conditions to enable controlled excitation of the second resonant mode without requiring additional substrate layers. An asymmetric U-slot provides selective fine-tuning of inter-band frequency spacing. The conventional SIW configuration, achieves 9.83 dBi and 7.93 dBi with bandwidths of 5.85% and 1.28% at 27 GHz and 39.8 GHz. While the ELO-strip improves the reflection coefficient by 7 dB and 9 dB, maintaining gains of 9.03 dBi and 7.81 dBi, bandwidths of 5.74% and 1.97%, with radiation efficiency ranging from 83 to 94% across both bands. This demonstrates a controlled gain–impedance trade-off enabling application-specific designs.Keywords: Asymmetric U-slot, Dual-band, mmWave 5G, SIW cavity-backed slot antenna, Impedance matching, ELO-stripSubject terms: Engineering, Physics Introduction Millimeter-wave antennas enable high-capacity data transmission in 5G/6G wireless systems, immersive virtual reality, satellite-IoT connectivity, and advanced sensing platforms by supporting wide bandwidth, beam steering, and low-latency communication1–8. The commercial rollout of fifth-generation (5G) cellular networks has created strong demand for antenna elements operating across the 27 GHz (n257/n258) and 39.8 GHz (n260) millimeter-wave (mmWave) bands designated by the federal communications commission for high-frequency 5G services9. These time-division duplex (TDD) FR2 spectrum windows are architecturally significant because they must support spatially independent transmit and receive chains within a single radiating aperture, demanding a high gain–bandwidth product alongside minimal cross-band interference2,3. Simultaneous dual-band coverage of these two windows, separated by approximately 12 GHz, is particularly attractive for multi-beam base-station front-ends and dense MIMO arrays, where antenna compactness and spectral efficiency are equally constrained5,6. Beyond communications, mmWave antennas are increasingly integral to automotive radar, aerospace sensing, and internet-of-things (IoT) infrastructure, all of which impose stringent co-design requirements on bandwidth, directivity, and fabrication repeatability4–8. The convergence of these requirements motivates the development of a compact, single-layer, dual-band architecture that can be manufactured using standard printed-circuit-board (PCB) processes without recourse to specialised multilayer fabrication. Cavity-backed slot antennas implemented using substrate-integrated waveguide (SIW) technology have emerged as an effective and widely adopted platform for millimeter-wave antenna design10–12. By synthesising a planar metallic cavity from rows of via-post walls within a standard PCB laminate, SIW structures confine the dominant transverse-electric (TE) mode, thereby significantly reducing radiation leakage and transmission losses compared to conventional open microstrip structures above 30 GHz13. Moreover, the enclosed cavity configuration naturally suppresses surface-wave propagation and substrate radiation losses, thereby improving radiation efficiency and pattern stability at mmWave frequencies. However, the associated high unloaded Q-factor inherently restricts the achievable impedance bandwidth of conventional single-slot SIW radiators, making broadband and dual-band operation intrinsically difficult14. Achieving simultaneous − 10 dB impedance matching at both 28 GHz and 39.8 GHz further compounds this challenge, as a matching structure optimised for one resonant band often perturbs the impedance behaviour of the other unless the resonant mechanisms are sufficiently decoupled within the cavity15. Consequently, the trade-off between cavity-Q preservation and dual-band impedance bandwidth constitutes the central design challenge addressed in this work. A focused review of recently reported dual-band mmWave antennas reveals persistent limitations in impedance tuning within compact, single-layer SIW platforms. In16, a compact SIW vertically integrated filtering antenna operating at 25/37 GHz achieved cross-polarization discrimination (XPD) exceeding 35 dB; however, the fractional bandwidths remained limited to 4.4% and 3.2% because of the high-Q behaviour of the unperturbed SIW cavity. Gap-coupled antenna-in-package architectures reported in17 achieved dual-band impedance coverage at 25/39 GHz with wide fractional bandwidths of 19.5% and 7.8%, respectively, but required multilayer parasitic stacking and precise coupling-gap control that becomes increasingly fabrication-sensitive at K, Ka-band frequencies. In18, integrated filtering resonators within a 3D-printed antenna structure operating at 25/33 GHz achieved gains of 8.8 and 6.4 dBi with XPD greater than 18 dB, demonstrating that in-band reactive loading can simultaneously improve impedance characteristics and suppress spurious radiation; nevertheless, additive manufacturing introduces dielectric-constant repeatability issues at mmWave frequencies. The DGS-based MIMO array antenna reported in5 achieved gains of 9.8 and 8.8 dBi at 25/29 GHz by redistributing surface currents and introducing additional resonant paths; however, the large footprint and open-ground radiation leakage reduced integration density and degraded front-to-back radiation characteristics. The HMSIW-on-through-glass-via (TGV) antenna presented in19 achieved XPD exceeding 20 dB at 27/38 GHz, although the design relied on specialized glass-substrate foundry infrastructure. Characteristic mode analysis (CMA)-based dual-key slot arrays reported in20 achieved gains of 10.5 and 11 dBi at 28/38 GHz by exciting orthogonal modal current distributions, but CMA primarily serves as a modal analysis methodology and offers limited post-fabrication impedance tunability. Moreover, the extremely thin 0.254 mm substrate constrained the available reactive-loading volume. Similarly, the work in21,22 reported a dual-circularly polarized array and SIW cavity-backed antennas at 28/38 GHz, in which cavity backing partially controlled the impedance environment through via-wall Q-loading; however, their large footprint and design complexity failed to improve the overall performance of the antennas. Collectively, these studies demonstrate that existing impedance-tuning approaches either rely on multilayer gap coupling, specialized fabrication technologies, or large radiating structures, while efficient reactive loading through compact techniques inside the SIW cavity remains largely unexplored. In parallel, dual-band and notch-bandwidth engineering techniques developed in ultra-wideband (UWB) monopole antennas have demonstrated that controlled reactive loading, defected-ground structures, and slot-based perturbation can effectively generate additional resonant current paths within compact radiating apertures23–29. These approaches provide important physical insight into impedance tuning and resonance manipulation through distributed reactive loading. Inspired by these mechanisms, the present work extends the concept of controlled reactive perturbation into an SIW cavity environment to realize stable dual-band impedance behaviour while preserving the compactness and manufacturability of the SIW platform. Fig. 1 presents a comparative taxonomy of impedance-tuning techniques30–33 employed in SIW cavity-backed slot antennas. Five principal impedance-tuning strategies have been reported for SIW cavity-backed slot antennas targeting millimeter-wave dual-band operation, each presenting a distinct trade-off between bandwidth, fabrication complexity, and integration density. The microstrip tapered-feed approach30 achieves a wide impedance bandwidth through a gradual width transition that progressively transforms the feed-line characteristic impedance to match the radiating aperture; the large physical footprint required by the taper conflicts directly with the compactness demanded by 5G FR2 front-end integration, and its inherently single-band nature precludes dual-band coverage without a secondary matching network. Reactive-stub loading31, implemented through open or short-circuit transmission-line stubs, offers a compact impedance correction mechanism that introduces conjugate reactances to cancel the imaginary part of the cavity input impedance at a target frequency; however, the achievable bandwidth remains narrow because the stub susceptance varies sharply with frequency, and multi-stub configurations required for dual-band coverage introduce additional parasitic resonances that perturb both bands simultaneously. Gap-coupled parasitic patch loading32 exploits fringing-field coupling across a controlled air gap to introduce a secondary resonant path, enabling dual-band operation with moderate bandwidth enhancement; the critical limitation is the extreme sensitivity of the inter-element coupling coefficient to the gap dimension and an additional substrate layer is invariably required to support the floating parasitic patch. The H-slot and shorting-pin configuration15 perturbs the cavity eigenmode field distribution by introducing inductive via-posts and an H-shaped aperture, generating a second resonant mode near the upper target frequency and enabling true single-layer dual-band realisation; however, shorting-pin placement is highly sensitive to drill-position tolerance, potentially introducing unintended modal coupling and pattern asymmetry, and the multi-via drilling process significantly increases fabrication complexity at 28–39 GHz. Defected ground structure techniques33 redistribute aperture surface currents by etching periodic or aperiodic apertures in the ground metallisation, widening the impedance bandwidth through additional resonant current paths; while effective for broadband matching, DGS apertures inherently increase back-radiation by allowing energy leakage through the ground plane, compromising the front-to-back ratio that the SIW cavity is specifically designed to maximise. The extra substrate layers, controlled gap dimensions, ground-plane perturbation, and establishing a measurably superior combination of fabrication accessibility, XPD preservation are found to be the limiting factors for effective and controlled impedance tuning. Comparative taxonomy of impedance matching/tuning in SIW cavity-backed slot antennas. To address the limitations identified above, this work proposes a physically distinct impedance-tuning mechanism designated as the Elongated Strip (ELO-strip). A dielectric-less copper strip of dimensions is soldered within the plane of the primary radiating patch, soldered orthogonally to the cavity. The strip behaves as a distributed reactive-loading structure that modifies the local cavity-current distribution and introduces additional inductive–capacitive loading inside the SIW aperture34–38. Because the strip extends into the cavity aperture without dielectric filling between the strip and cavity boundary, it introduces an air-loaded section that locally reduces the effective permittivity while increasing the effective electrical path length without requiring multilayer substrates or specialized conductive materials. The compound ELO-strip–primary-radiator structure, therefore, functions as a distributed impedance-transformation element capable of converting weak cavity resonances near 27 GHz and 39.8 GHz into well-matched radiating modes satisfying the -10 dB reflection criterion across both operating bands. An asymmetric U-slot further acts as a parasitic slot resonator that enables controlled tuning of the inter-band frequency spacing, while two shorting vias establish the dual-resonant cavity response. The principal contributions of this work are summarized as follows: The first demonstration of a galvanically integrated sub-wavelength reactive inclusion within an SIW cavity for controlled dual-band impedance tuning without requiring multilayer cavity configurations or external matching circuits. The first demonstration of a galvanically integrated sub-wavelength reactive inclusion within an SIW cavity for controlled dual-band impedance tuning without requiring multilayer cavity configurations or external matching circuits. The proposed ELO-strip improves the resonance depth by approximately 7 dB and 9 dB, at the lower and upper operating bands, respectively thereby confirming its role as an effective impedance-transformation mechanism. The proposed ELO-strip improves the resonance depth by approximately 7 dB and 9 dB, at the lower and upper operating bands, respectively thereby confirming its role as an effective impedance-transformation mechanism. High realized gain (9.83 dBi) using a compact single-layer SIW cavity-backed configuration suitable for mmWave 5G front-end integration is achieved. High realized gain (9.83 dBi) using a compact single-layer SIW cavity-backed configuration suitable for mmWave 5G front-end integration is achieved. A comprehensive benchmark against recently reported SIW, filtering-based, array-based, and non-SIW dual-band mmWave antennas to establish the proposed design trade-off between bandwidth controllability, fabrication simplicity, and radiation performance. A comprehensive benchmark against recently reported SIW, filtering-based, array-based, and non-SIW dual-band mmWave antennas to establish the proposed design trade-off between bandwidth controllability, fabrication simplicity, and radiation performance. The remainder of this article is organized as follows. Section “Configuration of the antenna element” presents the proposed antenna configuration together with its design methodology and operating principle. Section “Results and discussion” discusses the parametric analysis and experimental validation of the proposed structure, including the influence of the ELO-strip on impedance behaviour and radiation performance. Finally, section “conclusions and future work” summarizes the principal conclusions of this work. Configuration of the antenna element Rectangular patch antenna The proposed antenna consists of a rectangular patch antenna built on a dielectric substrate, Rogers 5880, with a relative permittivity of 2.2, loss tangent = 0.0009, and substrate height of 0.787 mm. Rogers 5880 is selected over FR-4 and similar PTFE-based laminates because its combination of low loss tangent, stable dielectric constant across the mmWave frequency range (25–42 GHz), and smooth copper-clad surface finish (Ra < 1.0 µm) minimises ohmic losses at the cavity wall interfaces, a critical requirement for maintaining high radiation efficiency above 30 GHz. The radiating patch dimensions and U-slot resonant length are determined through standard microstrip synthesis and subsequently refined through full-wave parametric optimisation in ANSYS Electronics Desktop 2023.R2. The rectangular patch antenna follows the design methodology of39, and is subsequently optimised using full-wave simulation. Dimensions of the U slot can be approximately calculated through40,41 from Eq. (1). Due to cavity loading and mutual coupling between the two arms, the effective resonant frequency shifts upwards. The resonating frequency can be calculated by (1). where = resonating frequency (GHz); = U-slot right arm length; = U-slot width; = U-slot thickness; = Effective permittivity. SIW cavity-backed antenna The SIW cavity-backed antenna is shown in Fig. 2. It consists of an SIW cavity, radiating slots, and a tapered microstrip feed. A single printed circuit board (PCB) is utilized for the implementation of the SIW cavity with a substrate height of h, and the cavity sidewalls are realized by etching metallic via posts along three edges of the slot. The effective and physical widths of the SIW cavity can be calculated using the following Eqs. (2–4). The SIW cavity occupies the upper metallic surface of the Rogers 5880 board and is bounded on three edges by rows of metallised via posts, with the radiating aperture slot forming the fourth boundary. The key parameters of the SIW cavity-backed antenna are the pitch S (distance between the centers of adjacent vias) and the diameter D of the vias. Geometry of the SIW cavity-backed antenna with ELO-strip (13 optimised parameters; Integrated feed/cavity/ELO section); fabricated on a single 0.787 mm Rogers RT5880 dielectric layer). The effective width of the SIW cavity42 can be approximated as The width and length of the SIW cavity are estimated using43 where are the higher and lower cutoff frequencies in GHz. The above equations for the dimensions and have been derived from the resonant-frequency equation for the rectangular cavity. The final optimized dimensions of the SIW cavity-backed antenna, along with the ELO-strip, are tabulated in Table 1. With S = 0.6 mm and D = 0.4 mm, the via leakage criteria are verified: S/D = 0.6/0.4 = 1.5 < 2 and D = 0.4 mm < λg/5 = 1.50 mm at 27 GHz. A broader tapered microstrip feed is employed to excite the antenna and improve impedance matching44. The SIW cavity is excited through a microstrip-to-SIW tapered transition in which the conductor width is progressively reduced from the 50 Ω feed reference plane to the via-wall aperture, converting the quasi-TEM microstrip mode into the dominant TE₁₀ SIW mode by adiabatically rotating the transverse electric field polarisation across the taper length. Table 1Optimised dimensions of the proposed dual-band SIW antenna with ELO-strip.ParameterNotationDimensions (mm)Ground lengthL_g18Ground/patch widthW_g14SIW effective widthA_s9Feed line lengthL_f6.97Feed line widthW_f1.56U-slot thickness (arm width)U_th0.35U-slot right arm lengthU_ht1.80U-slot left arm lengthU_ht_l1.35U-slot opening widthU_Wt1.10Via pitch (centre-to-centre)S0.6Via diameterD0.4ELO-strip lengthEp_l2.0ELO-strip widthEp_w1.0 Optimised dimensions of the proposed dual-band SIW antenna with ELO-strip. At 27 GHz, the taper geometry preferentially excites the fundamental TE₁₀ cavity mode, concentrating the surface current density along the feed edge and the inner walls of the U-slot. At 39.8 GHz, the higher-order modes, concurrently supported within the same cavity volume, and the two shorting pins act as inductive posts that perturb the cavity boundary condition, splitting the degenerate modal frequencies and spatially constraining the upper-band current maximum to the U-slot arm regions rather than the feed throat. This modal selectivity ensures that each resonant band draws energy from a spatially distinct current path. The following section elaborates on the concept of ELO-strip and its mechanism. ELO-strip: placement and physical mechanism In this work, a novel impedance-matching technique is introduced in which an Elongated strip, referred to as the ELO-strip, a small conductive strip made of copper, is integrated orthogonally (at 90°) onto the surface of the radiating patch through reflow soldering. This concept is inspired by dielectric resonator antennas, where the dielectric extends beyond the ground plane to enhance coupling and impedance control45. The ELO-strip acts as a galvanically bonded conductive extension to the radiating patch, thereby forming a galvanic junction while introducing controlled reactive loading. Since it is placed in the direction of the main current, and not a separate parasitic element, there is a change in the effective current path and local inductive/capacitive loading as shown in Fig. 3. Because the strip is electrically continuous to the radiating patch, the two metals together form a single compound radiator with a modified shape.​ Therefore, stronger coupling is experienced. Since the ELO-strip is a dielectric-less strip, it shows higher radiation efficiency and low dielectric loss. Since the ELO-strip is partially air-loaded (Schneider formula gives εr,eff ≈ 1.78 for W/h = 1.27, lower than the substrate value of 2.2 but above the air value of 1.0), the design equations are simplified46 as follows (5–7): where is the width of the radiating edges (along the y-axis), is the length of the non-radiating edges (along the x-axis), εr,eff is the effective dielectric constant, and ΔL represents the correction to the patch length due to fringing-field effects. Note that Eq. (6) as shown, is a simplified form; the complete Hammerstad–Jensen (1975) expression includes εeff terms. Substituting fr = 27 GHz and εᵢ,eff = 1.78 into Eq. (5) yields = 4.17 mm and = 4.97 mm. The final HFSS-optimised dimensions differ significantly because the ELO-strip operates as a sub-wavelength inductive loading element, not as a resonant half-wave radiator. The 90° galvanic orientation and compound boundary conditions fundamentally alter the resonant behaviour relative to standard patch theory. These equations therefore serve only as a preliminary order-of-magnitude approximation; the final geometry is refined through full-wave electromagnetic simulation and optimization using HFSS. The electric field discontinuity introduced by the ELO-strip. The equivalent circuit in Fig. 4 was developed to represent the dominant electromagnetic behaviour of the proposed ELO-strip-enabled SIW cavity-backed slot antenna using physically identifiable resonant sub-sections. The extraction methodology follows the standard admittance-based equivalent-circuit modelling approach commonly adopted for SIW resonators, cavity-backed antennas, and slot-loaded microwave structures. Equivalent circuit of the cavity-backed SIW with ELO-strip. Initially, the complete antenna structure was simulated in HFSS using a tapered microstrip feed excitation over the operating frequency range. The simulated reflection coefficient was subsequently converted into the corresponding driving-point admittance , from which the dominant resonant characteristics were identified. The resonant frequencies were determined using the resonance condition: Following the Foster immittance extraction method reported in31 and47, the equivalent resistance Rcav, Lcav, Ccav, were estimated from the conductance component of the admittance through Eqs. (9–11): while the equivalent capacitance was extracted from the susceptance slope near resonance using the first-order Foster reactance expansion near parallel resonance: where . The corresponding equivalent inductance was then obtained using the resonance relation: where . The shorting-via section was represented using the branch , where the via inductance arises from the vertical current flowing through the metallic via post. Following the partial-inductance approximation for metallic via posts reported in10 and38, the via inductance was initially approximated as (12): where h is the substrate thickness and D is the via diameter. The extracted values , and represent the localized reactive loading and radiation leakage introduced by the shorting-via configuration. Similarly, the proposed ELO-strip was represented by the shunt resonant branch coupled through the fringing capacitances and . Following the distributed-inductive approximations reported in35 and36, the strip inductance and capacitance were initially estimated as Eqs. (13–14): where and denote the strip length and strip width, t is the strip thickness, h is the substrate height, respectively. Finally, the complete equivalent circuit was implemented in ADS for linear circuit optimization. The initially extracted values were iteratively refined until the circuit model closely matched the full-wave HFSS response across the operating band. The close agreement between the equivalent-circuit and full-wave simulated responses in Fig. 5 confirms that the proposed RLC network accurately captures the dominant electromagnetic mechanisms of the ELO-strip-loaded SIW cavity-backed slot antenna. And for further clarity on the structure of the ELO-strip and its soldering and mounting technique, the front, rear and side views have been indicated in Fig. 6. Reflection coefficients of the equivalent circuit model in ADS. Fabricated ELO-strip SIW cavity-backed antenna showing front, rear, and side views. Evolution of the dual-band antenna The design evolution of the proposed antenna progresses through five stages, as illustrated in Fig. 7, to systematically achieve the desired dual-band impedance response. Stage 1 represents the baseline SIW cavity-backed slot antenna, producing a single poorly matched resonance that does not satisfy the − 10 dB criterion at either 27 GHz or 39.8 GHz, confirming the fundamental bandwidth limitation imposed by the high unloaded Q-factor of the cavity. The introduction of two shorting pins in Stage 2 establishes localised inductive perturbations that shift the resonant modes and generate additional resonant paths enabling the first observable dual-resonance behaviour, though both resonances remain insufficiently matched. Stage 3 incorporates a symmetric U-slot as a parasitic slot resonator, simultaneously enabling coarse inter-band frequency spacing control between the two operating bands through elongation of the effective surface-current path within the cavity. In Stage 4, an asymmetric U-slot is introduced, and the desired dual-band resonances at 27 GHz and 39.8 GHz are attained through the combined effect of the shorting pins and the asymmetric U-slot geometry. Stage 5 introduces the galvanically connected ELO-strip, oriented orthogonally onto the primary radiating patch, which imposes distributed inductive–capacitive loading that elongates the effective current path, confirming that the ELO-strip operates exclusively as an impedance tuning that is functionally decoupled from frequency tuning. This functional orthogonality between the U-slot (frequency placement) and the ELO-strip (resonance enhancement) constitutes the core design principle of the proposed architecture, enabling independent optimisation of the two most critical dual-band matching parameters within a single layer. This validates the sequential design strategy and confirms that each structural addition addresses a distinct and non-overlapping impedance engineering objective. The comparative S11 plots in Fig. 8a) confirm a clear and consistent pattern: the U-slot (Stage 4 & Stage 5) controls resonance frequency placement and inter-band spacing, while the ELO-strip exerts a focused influence on the resonance depth. Specifically, the addition of the ELO-strip deepens the lower-band S₁₁ resonance by approximately 7 dB (from − 21 dB to − 28 dB) and the upper band by 9 dB (from − 15 dB to − 24 dB) without shifting the resonance frequencies perturbed by the U-slot. In Stage 4, as in Fig. 8b), the asymmetric U-slot concentrates surface currents strongly along the primary radiating aperture, thereby maximizing aperture efficiency and peak gain at the resonant frequencies. However, when the ELO-strip is introduced in Stage 5, part of this concentrated current is intentionally redistributed along the elongated strip path to generate additional inductive–capacitive loading required for stable dual-band impedance matching. This current redistribution improves the impedance bandwidth and resonance depth at both 27 GHz and 39.8 GHz, but slightly reduces the fraction of power contributing directly to broadside radiation, resulting in a modest gain reduction. The stages involved in the design of the proposed prototype: Stage 1—Stage 5 with the inclusion of shorting pins, asymmetric U-slot and ELO-strip. Performance variations of the proposed antenna (a) In terms of reflection characteristics and (b) In terms of gain. Operating mechanism-role of shorting pins, U-slot, and ELO-strip Shorting pins The proposed SIW cavity supports a discrete transverse electric eigenmode, whose resonant frequencies are governed by the effective cavity dimensions and dielectric constant. In the initial configuration, the cavity alone supports poorly matched higher-order resonant modes that inherently lie in the millimeter-wave region at 40 GHz. When two shorting pins are introduced near the microstrip tapered feed, inside the cavity, additional electrical paths are created between the top and bottom metallic layers, effectively perturbing the field distribution and introducing inductive loading. This loading modifies the boundary conditions of the cavity and produces multiple resonant paths, resulting in two distinct resonances at 27 GHz and 39.8 GHz. The lowest resonance is mainly associated with a shortened current loop formed by the shorting pins, while the two higher resonances correspond to modified cavity modes whose frequencies are shifted by the pin-induced perturbations. Asymmetric U-slot The asymmetric U-slot etched on the radiating patch behaves as a parasitic slot resonator electromagnetically coupled to the hybrid cavity modes. It introduces additional capacitive and inductive loading that reshapes the input impedance locus, thereby improving impedance matching across the two bands. The influence of the U-slots is shown as the surface current distributions on the proposed antenna at 27 and 39.8 GHz, respectively, in Fig. 9. Due to its asymmetrical arms, the U-slot interacts differently with the characteristic current directions of each mode, enabling the selective tuning of mid- and high-frequency resonances. This asymmetry ensures constructive radiation from different parts of the slot, thereby improving the effective aperture efficiency of the antenna. Moreover, the U-slot introduces an additional resonant path whose electrical length can be independently adjusted, enabling fine control of the frequency spacing between the 27 and 39.8-GHz bands. As a result, the uneven U-slot not only improves the radiation efficiency and gain but also ensures that the higher- order cavity modes are effectively coupled to free space. Current distribution around the U slot at the resonant frequencies of 27 and 39.8 GHz. ELO-strip When the ELO-strip is placed on the cavity aperture, it introduces an additional distributed sub-wavelength reactive loading element that strongly couples with the cavity–pin modes, forming hybrid patch–cavity resonances. The elongated and partially air-loaded structure of the ELO-strip increases the effective electrical path, leading to a modified current distribution that simultaneously introduces controlled reactive loading that shifts the hybrid-mode resonant frequencies to coincide with the 27 GHz and 39 GHz operating bands. This hybridization also broadens the impedance bandwidth by overlapping the strip and cavity resonances. Results and discussion Parametric analysis The SIW antenna parameters have been parameterised to achieve the desired dual-band operation and obtain matching characteristics across the upper and lower passbands. In each parametric sweep, one design variable was altered while keeping all remaining parameters unchanged to isolate its individual influence. This study focuses on key parameters, specifically the SIW cavity width (defined as the centre-to-centre spacing between adjacent via rows) and the arm length of the U-shaped slot. All parametric simulations were conducted using the Optimetrics module in ANSYS Electronics Desktop version 2023.R2. Effect of the right and left arms of the U-slot The shorter arm of the asymmetric U-slot was taken as the left arm and subjected to a parametric investigation. This study revealed that a preliminary optimum was found near 1.5 mm; however, through HFSS optimisation, 1.35 mm was considered as the final left arm length (U_ht_l = 1.35 mm, Table 1). Subsequently, a parametric sweep was carried out on the right arm (U_ht), which showed that although a length of 1.5 mm caused a noticeable shift in the resonance, it also led to deterioration of the upper-band response. Therefore, a length of 1.8 mm was chosen for the right arm of the asymmetric U-slot to obtain balanced multiband performance, as illustrated in Fig. 10a and b. The parametric analysis of the U-slot (a) Right arm with left arm length fixed at U_ht_l = 1.35 mm (b) Left arm. Parametric analysis of single versus dual ELO-strip The number of ELO-strips to be employed was parametrically validated with single Vs dual ELO-strips. The diagrammatic representation of the dual ELO-strips showing the exact configuration with which the analysis was done is shown in Fig. 11. The parametric analysis of the length and width of the ELO-strip was carried out as in Fig. 12. From the results derived from the parametric analysis and gain simulations, the following are the five conclusions regarding the ELO-strip and the inclusion of more than one ELO-strip. The dimensions of the strip do not affect the resonance unless they change the location of the strip. This is because the ELO-strip in the proposed design functions as a sub-wavelength reactive loading element, not as a resonant half-wave radiator. Its electrical length (≈ 0.18λg at 27 GHz) is well below the λg/4 threshold at which dimensional changes would shift resonance frequencies. Instead, the strip’s primary action is to alter the local current distribution at the cavity aperture, modifying the effective input reactance (primarily inductive loading) seen at the feed. Because this inductive perturbation is governed by the strip’s position and galvanic junction rather than its absolute dimensions, which remain sub-wavelength across the swept range, the S₁₁ resonance location is insensitive to the ± 50% dimensional variations shown. The gain difference (Fig. 13), however, does respond to strip dimensions, confirming that the strip influences the radiation mechanism even where its impedance effect saturates. The dimensions of the strip do not affect the resonance unless they change the location of the strip. This is because the ELO-strip in the proposed design functions as a sub-wavelength reactive loading element, not as a resonant half-wave radiator. Its electrical length (≈ 0.18λg at 27 GHz) is well below the λg/4 threshold at which dimensional changes would shift resonance frequencies. Instead, the strip’s primary action is to alter the local current distribution at the cavity aperture, modifying the effective input reactance (primarily inductive loading) seen at the feed. Because this inductive perturbation is governed by the strip’s position and galvanic junction rather than its absolute dimensions, which remain sub-wavelength across the swept range, the S₁₁ resonance location is insensitive to the ± 50% dimensional variations shown. The gain difference (Fig. 13), however, does respond to strip dimensions, confirming that the strip influences the radiation mechanism even where its impedance effect saturates. When two strips of the same dimensions are placed asymmetrically, the impedance matching deteriorates. When two strips of the same dimensions are placed asymmetrically, the impedance matching deteriorates. When the two strips are placed on the same line, they produce the same effect as that of the single strip. When the two strips are placed on the same line, they produce the same effect as that of the single strip. If the strips are placed back-to-back at a distance of separation (e.g., 0.5 mm away), they do not show any effect on the impedance matching as in Fig. 13. If the strips are placed back-to-back at a distance of separation (e.g., 0.5 mm away), they do not show any effect on the impedance matching as in Fig. 13. Among all evaluated topologies, the single ELO-strip configuration yields the highest gain while maintaining comparable impedance matching, as evident from Fig. 13. Therefore, single ELO-strip has been used in the proposed SIW cavity-backed antenna. The ELO-strip works in the above set of rules, as per the analysis done as simulations in HFSS. After concluding the dimensions and simulations, the proposed antenna was fabricated as shown in Fig. 14a) and measured the reflection characteristics with the Keysight P9377B Vector Network Analyzer (VNA) and in the anechoic chamber, as shown in Fig. 14). Among all evaluated topologies, the single ELO-strip configuration yields the highest gain while maintaining comparable impedance matching, as evident from Fig. 13. Therefore, single ELO-strip has been used in the proposed SIW cavity-backed antenna. The ELO-strip works in the above set of rules, as per the analysis done as simulations in HFSS. After concluding the dimensions and simulations, the proposed antenna was fabricated as shown in Fig. 14a) and measured the reflection characteristics with the Keysight P9377B Vector Network Analyzer (VNA) and in the anechoic chamber, as shown in Fig. 14). Two ELO-strips adjacent and one after the other. (a) and (b) The parametric analysis of the length and width of the ELO-strip, respectively. Analysis of the reflection coefficients and gain on the different topologies of ELO-strips. (a) The fabricated prototype and (b) measurement of the prototype in an anechoic chamber with the reference antenna. Reflection coefficient The measured and simulated S₁₁ characteristics of the proposed SIW cavity-backed antenna with and without the ELO-strip demonstrate the strong influence of the strip on impedance matching and multiband behaviour. In terms of the reflection coefficients, without the ELO-strip, the SIW cavity with shorting vias exhibits only weak resonances near 27 and 39.8 GHz, with poor impedance matching outside these bands due to inefficient radiation coupling as seen in Fig. 15. When the ELO-strip is introduced, two well-defined,deeper resonances appear at approximately 27 and 39.8 GHz, all satisfying the − 10 dB criterion. The lower and upper bands are mainly associated with the fundamental and higher-order SIW cavity modes. The measured S₁₁ response closely follows the simulated result, with minor frequency shifts (0.8 dB at the upper band) attributed to fabrication tolerances and material uncertainties. The ELO-strip enhances the reflection coefficient by 7 dB and 9 dB at 27 and 39.8 GHz, respectively. To emphasize the significance of the proposed work, a comprehensive comparison with recently reported dual-band MIMO, array, wearable, filtering, HMSIW antenna designs48–50 has been presented in Table 2. As observed, several state-of-the-art MIMO antennas have been developed for 5G millimeter-wave applications operating at the widely adopted 28/38 GHz bands. Measured and simulated results for the parameter S11. Table 2Comparison of the proposed design with different state-of-the-art designs.RefDimensions (mm3)TechnologyFreq (GHz)XPD (dB)BW (%)Gain (dBi)1618 × 14 × 0.6SIW vertical integrated25/37> 354.4/3.25.1/7.71710 × 10 × 1.09 †AIP-stacked patch25/39NM19.5/7.87/1018NMPrinted filtering antenna25/33> 184.4/6.28.8/6.4537.7 × 37.7 × 0.508DGS with MIMO25/29NM12.56/10.719.8/8.8193.2 × 3.25 × 0.45HMSIW on through glass vias (TGV)27/38> 2010/5.56.1/5.62019.75 × 26 × 0.254†Slot array antenna28/38> 1612.5/21.0510.5/112114 × 32 × 0.882Parasitic patch28/38> 157.3/7.58/7.9 dBic4829 × 24.2 × 0.203Quad port printed antenna28/38> 156.3/3.198.1/7.224918 × 8.5 × 0.25Wearable antenna28/38> 202.5/1.324.14/7.735014 × 14 × 0.8Dual-band MIMO antenna28/38NM10/8.446.92/6.8This work18 × 14 × 0.786SIW cavity-backed without ELO-strip27/39.8> 205.85/1.289.83/7.93With ELO-strip27/39.8> 185.74/1.979.04/7.81†Array gain. NM, - not mentioned; AIP, - antenna-in-package. Comparison of the proposed design with different state-of-the-art designs. †Array gain. NM, - not mentioned; AIP, - antenna-in-package. Radiation pattern The three-dimensional radiation pattern does not exhibit a closed-form shape due to the hybrid excitation of cavity, slot, and feed-induced modes within the SIW structure. Unlike conventional canonical antennas, the proposed design involves strong near-field coupling, non-uniform current distribution, and finite cavity boundaries, which collectively preclude an analytical closed-form solution. Consequently, the radiation characteristics are best interpreted through full-wave electromagnetic simulations, revealing a dominant broadside radiation with controlled asymmetry and suppressed back radiation. Figure 16 illustrates the three-dimensional realized gain patterns of the antenna, (a) without and (b) with the ELO-strip, at 27 GHz and 39.8 GHz. 3D view of gain plots of the proposed antenna, without ELO-strip (a) & (b) and with ELO-strip (c) & (d) at 27 GHz and 39.8 GHz, respectively. Without the ELO-strip at 27 GHz, a dominant main lobe orthogonal to the aperture is observed, exhibiting effective coupling through the radiating slot and efficient excitation of the mode. The absence of significant back radiation proves the strong field confinement provided by the SIW cavity and the enhanced front-to-back ratio achieved by the cavity-backed configuration. At 39.8 GHz, the radiation pattern becomes slightly more directive due to the increased electrical size. The gain plot with the ELO-strip is in a front view so that the ELO-strip is visible. The ELO-strip hybridizes the cavity–pin resonant modes, significantly deepening the impedance match (ΔS11 ≈ − 5 dB) at the cost of a moderate efficiency reduction (4% at 27 GHz, 9.7% at 39.8 GHz) due to additional conduction loss in the strip. The proposed dual-band SIW cavity-backed antenna demonstrates stable broadside radiation at both 27 GHz and 39.8 GHz, with and without the ELO-strip configuration in Fig. 17. Radiation patterns of the proposed antenna with and without the ELO-strip (a) at 27 GHz and (b) 39.8 GHz. At 27 GHz, the co-polarized component exhibits a smooth, symmetric main beam attributed to efficient excitation of the fundamental cavity mode and uniform aperture field distribution. Cross-polarization remains substantially suppressed across the entire angular range in both E- and H-planes, confirming strong polarization discrimination and negligible modal interference. At 39.8 GHz, the radiation pattern becomes more directive due to higher-order mode excitation and increased electrical aperture size, yet cross-polarization purity is well maintained. Without the ELO-strip, the antenna retains acceptable broadside performance, although minor asymmetries and ripples observed at 39.8 GHz are attributed to fabrication sensitivities and elevated modal complexity. The close correspondence between measured and simulated patterns at both frequencies validates the cavity-slot excitation mechanism and confirms structural integrity of the design. The With/Without ELO co-pol curves overlay almost identically across all four subplots, which is a key technical merit. The XPD is visually maintained at 20 dB at broadside in all four cuts, supporting the claimed polarisation purity. as in Fig. 18. Normalised radiation patterns of the proposed antenna with and without the ELO-strip at (a) 27 GHz and (b) 39.8 GHz. Gain and radiation efficiency The visual separation of material/conduction losses from matching and pattern effects is summarised through the gain curves as in Fig. 19. Here, the gains simulated and measured values of the proposed SIW antenna, both with and without the ELO-strip, are plotted. The proposed SIW cavity-backed antenna exhibits peak gains of 9.83 dBi and 7.93 dBi at 27 and 39.8 GHz, respectively, in its baseline (without ELO-strip) configuration. Upon integration of the ELO-strip, the gain reduces to 9.03 dBi and 7.81 dBi at the respective bands. This reduction of approximately 0.8 dBi and 0.12 dBi is consistent with the fundamental relationship between antenna gain and radiation efficiency as expressed by Eq. (15): where G is the realised gain, η_rad is the radiation efficiency, and D is the directivity34. The ELO-strip introduces additional copper conduction loss due to its finite conductivity (σ_Cu = 5.8 × 10⁷ S/m), which, though high, contributes non-negligibly at millimeter-wave frequencies where the skin depth δ = (πfμσ)^{− 1/2} ≈ 0.43 µm at 27 GHz and 0.35 µm at 39.8 GHz, making surface roughness effects on loss non-trivial. The simulated radiation efficiency confirms this interpretation: the efficiency decreases from 94 to 92% at 27 GHz, and from 92 to 83% at 39.8 GHz, as depicted in Fig. 20. These efficiency changes precisely match the gain reductions when directivity is held approximately constant. At 39.8 GHz, the ELO-strip exhibits stronger localized surface-current concentration and enhanced near-field coupling with the surrounding SIW via boundaries. This additional reactive-field localization redistributes part of the aperture field away from the dominant broadside radiating mode, causing a portion of the stored electromagnetic energy to remain confined within the cavity instead of contributing effectively to far-field radiation, causing a reduction in the radiation efficiency at both the operating bands. The efficiency is 94% and 92% for the proposed antenna without ELO-strip, and 92% and 83% with ELO-strip. Simulated radiation efficiency exceeds 80% at both operating frequencies, confirming low ohmic and dielectric loss within the SIW cavity. With the dual peaks at 27 GHz and 39.8 GHz, the plot in Fig. 21 substantiates the simulated and measured received power of the antenna (with ELO-strip). The consistent variation of the received power expresses the compliance with the Friis transmission relation when the path and input power are fixed. The transmitted power is 0 dBm, while the distance between the transmitter and receiver is 1 m. Simulated and measured results ofgains of the proposed antenna with and without the ELO-strip. Simulated results of the radiation efficiency of the proposed antenna with and without ELO-strip. Simulated and measured received power of the proposed antenna (with ELO-strip). The SIW cavity-backed antenna without the ELO-strip also operates at the same resonant frequencies with comparable impedance bandwidths and higher gains. The radiation performance, particularly the antenna gain, can be further improved through the implementation of array configurations or by incorporating advanced structures such as metasurfaces or frequency-selective surfaces (FSS). These structures may be integrated with the proposed antenna designs either as superstrates or as reflecting layers. Conclusions and future work Unlike parasitic patches that interact through capacitive gap coupling or dielectric resonator extensions that require specialised substrates, this work introduces a new impedance-tuning approach, referred to as the ELO-strip, in which a galvanically integrated, sub-wavelength reactive element is included within the SIW cavity to achieve improved impedance characteristics. A comparative analysis between antennas, with and without the ELO-strip, is presented, and the corresponding results are examined in detail. Parametric investigations were performed to determine the optimal geometrical parameters of the proposed structure. The two proposed SIW antennas exhibit dual resonances at 27 GHz and 39.8 GHz, respectively, thereby covering the 5G FR2 n257 and n260 millimeter-wave allocations. A comprehensive performance comparison in terms of impedance bandwidth, return loss, radiation efficiency, and gain confirms that the ELO-strip serves as an effective impedance-matching technique with only a marginal reduction (8% and 1.5%) in gain at both bands. Further investigation is required to refine the ELO-strip, particularly with respect to fabrication considerations and practical implementation. Moreover, by extending the proposed designs based on established design principles, both antennas can be adapted for multiple-input multiple-output (MIMO) system implementations. Author contributions V.G. conceived the research idea, supervised the work, and contributed to methodology development and manuscript revision. E.J.S.D. performed the antenna design, simulations, experimental validation, data analysis, and wrote the main manuscript text. E.J.S.D. prepared all figures. V.G. and E.J.S.D. reviewed and approved the manuscript. Funding Open access funding provided by Vellore Institute of Technology. No funding was received for conducting this study. All authors certify that they have no entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript. Data availability All data supporting the findings of this study are available within the paper. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data supporting the findings of this study are available within the paper.