Dual-band integrated antenna design

A quarter-wavelength square monopole antenna was initially designed to operate at 2.45 GHz, which was then modified into an inverted triangular-shaped monopole to achieve dual-band resonance. Subsequently, two opposite L-shaped slots were etched into the radiator to introduce a second resonance at 5.2 GHz by effectively increasing the electrical path length of the radiating element. Figure 1(a) shows the 2D layout of the proposed dual-band antenna, which was designed using the 3D full-wave Computer Simulation Technology Microwave Studio® simulation tool. The corresponding design dimensions are listed in Table 1.

Figure 1.

The top view of: (a) the proposed dual-band antenna; (b) the fabricated layout size compared to an Egyptian pound; (c) highly bendable with a gentle finger press.

Table 1.

The dimensions of the proposed dual-band flexible antenna

The antenna is excited using coplanar waveguide feeding, with a feedline width of 3.4 mm, a length of 10 mm, and a gap of 0.2 mm, achieving an input impedance of 50 Ω (matched to a standard 50 Ω sub-miniature version A [SMA] connector). The substrate used is Rogers ULTRALAM 3850, composed of highly flexible liquid crystalline polymer, with a very thin thickness of 0.1 mm, a dielectric constant of 2.9, and a loss tangent of 0.0025.

The design of the proposed antenna is based on fundamental principles of monopole radiation, slot-loaded resonators, and artificial magnetic conductor (AMC)-backed patch behavior. The initial structure is derived from a triangular monopole, known for its broadband characteristics due to the gradual current tapering. The triangle monopole itself is designed as a λ/4 resonator (with a virtual λ/4 image in the ground plane, making it equivalent to a λ/2 dipole). Therefore, the initial dimensions of the triangle were calculated so that the triangle monopole length is a quarter-wavelength at f₁ = 2.4 GHz as in (1):

(1)\begin{equation}{f_1} = \frac{c}{{4\,{L_{eff}}\sqrt {{\varepsilon _{eff}}} }}\end{equation}

where ${L_{eff}}$ is the effective length from the feed point to the apex, and ${\varepsilon _{eff}}$ is the effective permittivity of the substrate.

To achieve dual-band functionality, two inverted L-shaped slots (the large length [L_c] is four times the small length) were introduced. These two slots introduce additional current paths, thus generating a second resonant mode. Due to the finite limit of the monopole metal structure, the slot was designed as a parallel resonator at 5.2 GHz. Therefore, the initial slots length was designed to be a quarter wavelength length at 5.2 GHz as in (2):

(2)\begin{equation}{f_2} = \frac{c}{{4\,(1.25*\,{L_c})\,\sqrt {{\varepsilon _{eff}}} }}\end{equation}

Finally, the position and length of each slot were optimized using parametric analysis to shift the secondary resonance to 5.2 GHz, while maintaining the primary mode at 2.45 GHz.

Next, the antenna was fabricated using delicate lithographic technology and the SMA connector was soldered to the prototype using silver conductive epoxy adhesive. The antenna is highly compact in comparison to one Egyptian pound, as depicted in Fig. 1(b). As a test of how flexible it is, illustrated in Fig. 1(c) is the antenna bent by a slight finger press.

The antenna reflection coefficient was measured by the means of Rohde & Schwarz ZVB20 vector network analyzer, where the testing setup is displayable in the inset of Fig. 2. The measured result (dotted-red) is compared against the simulated one (solid-black) in Fig. 2. As observed, the measured outcome verifies and agrees with the simulated one concerning dual-band resonance at 2.45 and 5.2 GHz. The −10 dB impedance matching is similar at 2.45 GHz at a level of −22 dB. On the contrary, the measured −10 dB impedance matching at 5.2 GHz is at a level of −16 dB, while that of the simulated is −22 dB. The measured lower band extends from 2.3 to 2.9 GHz with the center frequency at 2.45 GHz; thus, a fractional bandwidth of 24.5%. The simulated one is slightly wider and extends from 2.2 to 2.9 GHz with the center frequency at 2.45 GHz; thus, a fractional bandwidth of 28.6%. Coming to the measured higher band, it ranges from 4.9 to 5.4 GHz with the center frequency at 5.2 GHz; hence, a fractional bandwidth of 9.6%. Whereas, the simulated higher band exhibits a slightly wider bandwidth, which ranges from 4.9 to 5.5 GHz with the center frequency at 5.2 GHz; hence, a fractional bandwidth of 11.5%. The slight discrepancies in both bands are due to the non‐clean measuring environment associated with the presence of the different subjects and objects that affect the outcomes.

Figure 2.

Comparison between the proposed dual-band antenna simulated and measured S11, in free space, with inset of testing setup.

Benchmarked in Fig. 3 is the simulated (solid-black) and tested (dotted-red) gain radiation patterns in normalized polar plot representation in the X–Z and Y–Z planes. The testing setup inside the anechoic chamber is perceived in the inset of Fig. 3. At 2.45 GHz, the antenna exhibits an isotropic pattern in the H-plane (Y–Z plane), and a figure-of-eight pattern in the E-plane (X–Z plane), where both the simulated and measured outcomes agree, as shown in Fig. 3(a) and Fig. 3(b), respectively. Both patterns establish the omni-directional radiation pattern nature of the monopole antenna. The same can be said at 5.2 GHz, as highlighted in Fig. 3(c) and Fig. 3(d) for the H-plane and E-plane, correspondingly. Finally, the antenna exhibits simulated realized gains of 2.1 and 4.5 dBi at 2.45 and 5.2 GHz, respectively, which are quite good considering the compactness of the antenna. Furthermore, it possesses simulated total efficiencies of 99%, at 2.45 GHz, and 97%, at 5.2 GHz. It is worth noting that the achieved measured total efficiencies were 90% at both resonant frequencies.

Figure 3.

Comparison between the proposed antenna simulated and measured radiation pattern, in free space, with inset of testing setup at: (a) 2.45 GHz (H-plane); (b) 2.45 GHz (E-plane); (c) 5.2 GHz (H-plane); (d) 5.2 GHz (E-plane).

Since the antenna is intended to operate within the vicinity of the human body, radiations towards the human body should be kept at a minimum. To convert the antenna’s omni-directional radiation pattern type to a unidirectional one opposite to the human body, an AMC array reflector is placed in between the antenna and the human body. This also aids in enhancing the antenna realized gain.

Demonstrated in Fig. 4(a) is the AMC unit-cell 2D layout, where Rogers RO3003 was incorporated for the substrate material with a thickness of 1.52 mm and a loss tangent of 0.001. Since the antenna resonates at two distinct frequencies, the AMC structure has to display dual-band operation. This was realized by etching a diamond-shaped inner ring of the patch and loading the corners with slits. The design can be summarized by achieving a 0° reflection phase, which was achieved at 2.45 and 5.2 GHz. To reach the optimum AMC array size, a study was conducted in terms of antenna gain and total efficiency taking into consideration overall size, which led to the fact that the 3 × 3 array size was the most favorable one. Illustrated in Fig. 4(b) is the fabricated integrated prototype comprising the dual-band antenna and dual-band 3 × 3 AMC array reflector separated by an 8 mm thick foam.

Figure 4.

The top view of: (a) the AMC unit-cell top view with dimensions of L = 25 mm, L_p = 24 mm, R _out = 10 mm, R _in = 9.5 mm, L_c = 5 mm; (b) The fabricated integrated design layout with a foam separation.

Subsequently, the fabricated integrated design reflection coefficient was tested (dotted-red) and benchmarked against its simulated (solid-black) counterpart, as highlighted in Fig. 5. This was applied in free space, where both sets of results highly match, as dual-band resonance is preserved at the desired center frequencies. At 2.45 GHz, the measured −10 dB impedance matching is −20 dB, while that of the simulated is −27 dB. Whereas, at 5.2 GHz, the measured −10 dB impedance matching is −15 dB, while that of the simulated is −22 dB.

Figure 5.

Comparison between the integrated design simulated and measured S11 in free space.

In free space (off-body), benchmarked in Fig. 6 are the simulated (solid-black) and tested (dotted-red) integrated design gain radiation patterns in normalized polar plot representation in the X–Z and Y–Z planes. The testing setup inside the anechoic chamber is observed in the inset of Fig. 6. As anticipated, the antenna exhibits a unidirectional radiation pattern type, courtesy of the AMC array reflector at both resonant frequencies and in both planes. The measured outcomes highly agree with the simulated ones.

Figure 6.

Comparison between the integrated design simulated and measured radiation pattern, in free space, with inset of testing setup inside the anechoic chamber at: (a) 2.45 GHz (E-plane); (b) 2.45 GHz (H-plane); (c) 5.2 GHz (E-plane); (d) 5.2 GHz (H-plane).

Tabulated in Table 2 is the gain enhancement feature associated with the AMC array reflector compared to the antenna without it. At 2.45 GHz, the gain increased from 2.1 to 8.1 dBi, and at 5.2 GHz, it improved from 4.5 to 9.4 GHz. It is worth mentioning that the integrated design total efficiency is higher than 90% at both resonant frequencies.

Table 2.

Gain enhancement feature associated with the AMC array structure

Loaded human body scenario integrated antenna

Since the antenna should operate within the vicinity of the human body, then its performance, in terms of reflection coefficient and SAR levels, should be evaluated in human loading. The possible operation position is plotted in Fig. 7(a) and the body-loading testing setup can be witnessed in Fig. 7(b).

Figure 7.

Body-loading reflection coefficient evaluation: (a) Computer Simulation Technology environment and (b) testing setup.

Starting with the reflection coefficient parameter, many case studies were practiced that involved evaluating the integrated antenna on the human arm and human chest, as well as, in the absence of the AMC reflector. These outcomes were compared with the simulated outcome, as portrayed in Fig. 8. Inspecting Fig. 8, dual-band resonances were attained and preserved at the desired center frequencies of 2.45 and 5.2 GHz for the on arm case (dashed green curve) and on chest case (dashed red curve). However, without the AMC reflector and despite the 8 mm separation, dual-band resonances were lost (dotted brown curve). This highlights the significance of the AMC array structure in decoupling and isolating the antenna from the lossy nature of the human body.

Figure 8.

Comparison between the integrated design simulated and measured S11 for body-loading scenarios, with and without the AMC.

To evaluate the SAR values, averaged over 1 g of tissue, a three-layered human tissue model was built, where the characteristics of each tissue layer at each resonant frequency are tabulated in Table 3 [20]. With an input power of 100 mW, the SAR was studied in two cases. The first was the scenario of the integrated antenna with a separation distance of 1 mm from the human phantom, while the second was the antenna, in the absence of the AMC reflector, at a distance of 8 mm from the human phantom. Demonstrated in Fig. 9 and tabulated in Table 4 are the peak SAR outcomes at each resonant frequency. Due to the incorporation of the AMC reflector, the SAR levels dropped drastically from 11.35 to 0.39 W/kg at 2.45 GHz, and from 7.52 to 0.074 W/kg at 5.2 GHz. As such, the proposed dual-band and highly flexible integrated antenna design can be used for off-body wireless communication and low SAR medical applications, since it is not harmful to the human body.

Figure 9.

The integrated design peak SAR, averaged over 1 g of tissue, at: (a) 2.45 GHz and (b) 5.2 GHz.

Table 3.

Human tissue layers’ properties at 2.45 and 5.2 GHz [20]

Table 4.

Comparison between the peak SAR values, averaged over 1 g of tissue, of the proposed antenna with and without the proposed AMC

Compared to existing designs, the proposed antenna offers significant improvements in terms of SAR performance, compactness, and mechanical flexibility. It achieves ultra-low SAR values of 0.39 and 0.07 W/kg at 2.45 and 5.2 GHz, respectively. Well below the safety thresholds recommended by ICNIRP and IEEE. Unlike previous designs, it incorporates a highly flexible substrate and a dual-band AMC to enhance gain while maintaining low body exposure. These combined features make it one of the most suitable candidates for safe, reliable, and compact wearable biomedical applications.

Finally, tabulated in Table 5 is the assessment of the proposed work against similar published work in the literature. The comparison parameters include the antenna electrical size, the decoupling structure physical size, the integrated design realized gain, the integrated design total efficiency, and peak SAR levels, averaged over 1 g of tissue. As tabulated, the proposed antenna electrical size is smaller than [Reference Khan, Sethi, Malik, Jabbar, Khalid, Almuhlafi and Himdi2] and [Reference Ahmad, Ghaffar, Husain and Kim8], but slightly bigger than [Reference Ramasamy, Sapna and Jayasheela6] and [Reference Ahmad, Paracha, Sheikh and Ghaffar10] at 2.45 GHz. However, a larger array size was used for the suggested decoupling structure in comparison with [Reference Ramasamy, Sapna and Jayasheela6, Reference Ahmad, Ghaffar, Husain and Kim8] and [Reference Ahmad, Paracha, Sheikh and Ghaffar10]. Yet, this aided in achieving higher gain values and lower SAR outcomes, at closer distances from the human body, than the compared against references. Finally, the efficiency of the integrated design is much better than [Reference Ahmad, Paracha, Sheikh and Ghaffar10].

Table 5.

Comparison between the proposed work and recently published work in the literature