Dual circular complementary split ring resonator based metamaterial sensor with high sensitivity and quality factor for textile material detection

Metamaterial (MTM) is a synthetic structure with peculiar characteristics not found in nature.¹ Negative permeability or permittivity characterizes single negative (SNG) structures, while negative permittivity and permeability characterize double negative (DNG) or left-handed (LH) structures. In recent years, metamaterials have received considerable attention and are promising for the construction of high-sensitivity sensors. Due to its unusual characteristics, MTM is often used in other applications such as perfect absorbers,^2–5 super-resolution,^6,7 biosensors,^8,9 antennas,^10–12 clocking,¹³ and energy harvesting.^14,15 The use of MTM structures in sensing opens up new degrees of freedom, resulting in the advancement of compact and highly sensitive devices such as strain sensors,¹⁶ displacement sensors,¹⁷ angular displacement sensors,¹⁸ temperature sensors,¹⁹ mass flow sensors,²⁰ and different material characterization sensors.^21,22 The dimensions of these sensors are large, the effective medium ratio (EMR) is medium, and the sensitivity and Q-factor are poor. A metamaterial configuration was designed and verified for the measurement of clothing fabrics in Ref. 23. The average absorptivity measured when this metamaterial is covered with a polyester fabric sample is only 41.5%. The shift in resonance frequency is low for the change in permittivity value. In Ref. 24, metamaterial grating was used for chemical sensing; the Q-factor is medium and the sensitivity is low since the small variation in chemical permittivity does not affect the visible change in resonance frequency. A metamaterial sensor was presented in Ref. 25 for oil and fuel adulteration detection. The sensitivity and Q-factor are good, but the measurement process is complex. In Ref. 26, a microwave resonant sensor for monitoring noninvasive blood glucose concentration fluctuations was introduced. The size of the sensor is high, and the Q-factor and EMR are low. A pesticide trace was detected using a graphene-based MTM sensor in Ref. 27. The performance of the sensor is desirable, but the Q-factor is not so good.

Textile fabric and moisture of fabric were identified by using the DNG metamaterial structure.²⁸ The Q-factor of the structure is high, but the other parameters are not satisfactory. ɛ-negative metamaterial was presented in Ref. 29 for quad-band microwave applications. The analysis of the presented work is good, but the EMR is not high; its value is medium. Some metamaterial structures were described in Refs. 30–32 for different wireless applications. The size of these structures is large, and the EMR is low. Another group of researchers used serial reduction rate split ring resonator (SRR)-based sensors ³³ and complementary split ring resonator (CSRR)-based sensors ³⁴ to calculate the complex permittivity of liquid samples, and they published their findings in two journal papers. The suggested sensors verified the meta-resonator-based sensors’ capacity to discriminate between various microfluidics with various complicated permittivity. A dual-band microwave sensor was presented in Ref. 35 for the characterization of the dielectric substrate. The sensor size is large, sensitivity is low, and the Q-factor is only 267.5. Edible oil quality assessment by using a microwave sensor, which was presented in Ref. 36. The problem with this study is that the sensor size is very large and the Q factor is low. In Ref. 37, it was suggested to detect complex permeability and complex permittivity using a single meta-resonator-based sensor. The modified double CSRR loaded into mutual transmission line (MTL) was used to design the sensor. Each material under test (MUT) underwent two tests, the first of which measured its permittivity and the second of which assessed its permeability. The Q-factor is 37.36, the sensitivity is average, and the size of the sensor is very large, 70 × 70 mm². Another sensor was demonstrated in Ref. 38 for liquid permittivity sensing. The outcomes of the sensor are average. In addition, several MTM sensors have been employed for a variety of applications in the S-, C-, and X-band frequency spectrum, including microfluidic sensors,^39,40 liquid sensors, ^41,42 permittivity sensors,⁴³ thickness sensors,^44,45 and temperature sensors.^46,47 These sensors perform well, but the sensitivity is low and the Q-factors are moderate. A planar microwave sensor was discussed in Ref. 48 for substrate material sensing. The size and Q-factor of the sensor are 40 × 50 and 240.

In this investigation, we developed an MTM-based sensor structure that can detect textile materials. The structure can be used in various textile industries due to its high sensitivity. The research is split into two parts: first, the metamaterial structure is built and studied, and then the structure is used as a sensor. Both numerically and experimentally, the sensor structure has been validated. A different parametric study has been done to select the proper design structure; surface current, E and H fields, equivalent circuit, EMR, Q-factor, and sensitivity are analyzed. Overall resonance frequency changes provide information to precisely estimate the material’s permittivity; textile materials are detected in real-time, altering the response of the intended model to the electromagnetic (EM) characteristics of the MUTs. The structure is sensitive to the material’s dielectric characteristics and responds to changes in resonance frequency. The suggested sensor may be utilized in a range of garment sectors to identify textile materials since it is compact and cost-effective, as well as having high sensitivity and an excellent quality factor. The developed sensor may be used to distinguish various textile materials by using the microwave frequency.

The readily available and reasonably priced material Rogers RT5880 is used to create the compact suggested MTM structure. The material used for the substrate has a dielectric constant of 2.2, a 1.575 mm thickness, and features a front side with a copper thickness of 35 µm, while the back side has no copper. Annealed copper is used as a resonator whose conductivity is 5.8 × 10⁷ S/m. There are two circular SRRs on the front side of the structure. Figure 1 depicts the front view of the suggested circular SRR-based MTM sensor, which is 10 × 10 mm² in dimensions. Numerical unit cells are simulated by the finite integration technique (FIT) of the CST microwave studio software. Table I displays the size in mm scale of the developed dual circular SRR-based metamaterial sensor.

The simulation arrangement for the suggested circular SRR-based MTM sensor is shown in Fig. 2. An EM wave is incident on the Z-axis, while perfect electric conductors (PECs) and perfect magnetic conductors (PMCs) are consecutively positioned along the X-axis and Y-axis.

The proposed design was selected through a trial-and-error process that involved studying the various designs of the square, rectangular, and circular SRR-based MTM structures. After obtaining the results, it is seen that the circular SRR-based MTM structure performs better than the square, rectangular, and other SRR-based MTM structures. So, the circular-shaped resonators are used to make the presented MTM structure. Figures 3(a) and 3(b) display the circular-shaped resonator-based structure and their S₂₁ graphs for the various designs. During the first step, one outer circular ring resonator (RR) is utilized, resulting in the identification of one resonance frequency. In step 2, one smaller circular RR is used, and there is no resonance frequency (f_r) found within the selected frequency range. For the third step, one outer and one inner circular RR are utilized and found to have one resonance frequency. In step 4, one circular SRR with one inside circular RR are used to find one resonance frequency. One outer and one inner circular SRR are used in the proposed design, and two resonance frequencies that span the maximum frequency bands are discovered. Table II lists the overall outcomes for the various designs. The frequency range of the S band is 2–4 GHz, the C band is 4–8 GHz, the X band is 8–12 GHz, and the Ku band is 12–18 GHz. It is seen that the performance of the intended design is superior to the other steps, so it is selected as final.

Three important EM properties of MTM: permittivity (ɛ_r), permeability (μ_r), and refraction index (n_r) were determined using the Nicolson–Ross–Weir (NRW) technique.^30–32 In addition, permittivity (ɛ_r), permeability (μ_r), and refraction index (n_r) were extracted using MATLAB code,⁴⁹ and the obtained results were then compared to those obtained from the CST simulations.

A popular methodology for EM and MTM description that relies on measuring the S₁₁ and S₂₁ parameters of the sample being tested is the NRW procedure. Measuring the properties of metamaterials requires the use of terms such as impedance and wave velocity. Using the NRW technique, effective parameters were extracted from the data on the normal incidence of spreading parameters.⁵⁰ 

The S₁₁ and S₂₁ outcomes are derived through the CST simulator, as depicted in Fig. 4(a). Dual resonance frequencies at 3.56 and 8.49 GHz with magnitudes of −35.19 and −26.25 dB, respectively, are shown in the unit cell S₁₁ results. Dual-band resonance frequencies at 3.35 and 6.85 GHz with magnitudes of −26.29 and −40.03 dB, respectively, are also shown in the unit cell S₂₁ results. S₂₁ has a bandwidth of −10 dB at resonance frequencies of 60 and 655 MHz, respectively. Effective parameters are computed from the S₁₁ and S₂₁ results using Eqs. (6)–(8), as illustrated in Figs. 4(b)–4(d). Figure 4(b) indicates the relative permittivity vs frequency graph. The value of ɛ_r (imaginary) is all positive throughout the range of frequency, while the real value of ɛ_r displays two negative frequency ranges: 3.35–3.53 GHz and 6.85–8.23 GHz. Figure 4(c) displays the relative permeability vs frequency graph. The imaginary value of μ_r displays two negative frequency ranges: 3.35–3.60 and 6.85–8.85 GHz, while the real value of μ_r displays all positives throughout the entire frequency range. Figure 4(d) shows the refractive index vs frequency graph. The real value of n_r exhibits two negative frequency ranges: 3.35–3.50 and 6.87–8.01 GHz, while the imaginary value of n_r too indicates two negative frequency ranges: 3.35–3.50 and 6.87–8.01 GHz. The ionization effect on resonance frequency response typically occurs for MEMS or piezoelectric-based substrate materials. In general, rectangular flat plate resonators, if exposed to S- or C-band frequency operation, then gamma-rays affect the radiation response with respect to the fundamental width extensional mode (WEM). In addition, as the wave propagation continues, longitudinal standing waves affect the modes of propagation and effective parameter responses. For the proposed resonator, developed on the Rogers RT5880, which has the lowest electrical loss for reinforced Polytetrafluoroethylene (PTFE) material, and the distribution of dielectric properties using this substrate remains mostly uniform compared to MEMS or piezoelectric material. Therefore, the effect of ionization at the S- and C-bands would be minimal as per the theoretical point of view. Furthermore, experimental validation might be required further based on the proposed resonator if potential applications are exposed to the ionization field. It is noteworthy that, since the proposed resonator shows a high EMR value with a quality factor of 345% and 14.57% sensitivity, the narrowband operation might remain unchanged due to the radiation-induced field.

The inductances L1, L2, and capacitances C1, C2, C3, and C4 were utilized to create the equivalent circuit. The inductance and capacitance values are obtained from the ADS circuit simulator; these values are L1 = 0.674 nH, L2 = 1.17 nH, C1 = 0.576 pF, C2 = 1.16 pF, C3 = 0.92 pF, and C4 = 1.02 pF. The CST result is validated through analysis of the ADS results. The outer circular SRR contributes to the formation of L1 and C1 (LC circuit), while the inner circular SRR contributes to L2 and C2 (LC circuit). The coupling capacitor between the outer and inner circular SRR is formed by C3 and C4. Adjustments are made to the values of components in ADS to ensure that the resonance of S₂₁ matches those generated by CST. In order to adjust the first resonance frequency to 3.35 GHz, we adjusted capacitor C3. After addressing the second resonance at 6.85 GHz, the capacitor C4 has been modified. Finally, the simulated transmission coefficients from CST and ADS are presented in Fig. 6(b), and these two outcomes are nearly the same.

As design specifications, such as the split gap (g), the width of resonator $w$⁠, the substrate material, the gap between resonators (g₁), and the resonator material change, the transmission coefficient (S₂₁), and the resonance frequency (f_r) also change. These analyses are explained in the sections below.

The transmission resonance frequency and its magnitude are changed for the change in split gap, which is depicted in Fig. 7. Five split gaps of 0.30, 0.40, 0.50, 0.60, and 0.70 mm are utilized to verify the overall result. The intended structure is simulated, and data are accumulated for every split gap. The graph was generated once the simulation was finished. Creating a split gap in a MTM resonator patch mostly changes the f_r of the unit cell. A capacitance is introduced into the system by the resonator patch’s split gap, and this capacitance directly impacts the f_r of the metamaterial. The capacitance decreases with the increase in the split gap, and f_r is inversely proportional to capacitance. As a result, f_r increases with the increase in the split gap. The overall performance due to the variation in the split gap is shown in Table III. In the table, it is noticeable that the result is better for the gap of 0.5 mm than others. So, the split gap of 0.50 mm in the suggested structure is considered in this research.

The influence of resonator width on transmission resonance and magnitude is depicted in Fig. 8. The entire result is checked using five different resonator widths: 0.30, 0.40, 0.50, 0.60, and 0.70 mm. A simulation of the intended structure is run, and data are accumulated for every resonator width. The graph is created once the simulation is finished. The inductance increases with the increase in the resonator width, and f_r is inversely proportional to inductance. As a result, f_r decreases with the increase in the resonator width. Table IV presents an overall performance summary. The whole result due to the variation in the resonator width is shown in Table III. It is evident from the table that the resonator width of 0.5 mm yields better results than the other widths. Thus, in this study, the 0.50 mm resonator width in the proposed structure is considered.

The S₂₁ curve for different substrate materials is displayed in Fig. 9. This study also looked into the effects of altering the dielectric material on metamaterial performance. Three different types of Rogers with FR-4 were analyzed in this study: RO4350B, RT5870, and RT5880. In the initial trial, Rogers RT5870 with a thickness (th) of 1.57 mm, an (ɛ) of 2.33, and a (δ) of 0.0012 had been utilized. The second test utilized RO4350B with a th = 1.524 mm, an ɛ = 3.66, and a δ = 0.0037. In the third test, RT5880 was chosen, featuring a th = 1.575 mm, an ɛ = 2.2, and a δ = 0.0009. The fourth and final trial involved FR-4 with a (th) of 1.5 mm, an (ɛ) of 4.3, and a (δ) of 0.025. Table V shows the overall findings for the substrate material that was employed. In the table, we can see that the resonance frequency and its magnitude values are more suitable for the Rogers RT5880 substrate material than the others. So substrate Rogers RT5880 was used in the final analysis.

The impact of the resonator gap on the S₂₁ resonance and magnitude are depicted in Fig. 10. To verify the overall outcome, five distinct resonator gaps are used: 0.30, 0.40, 0.50, 0.60, and 0.70 mm. The formulated structure is simulated, and data are gathered for each resonator gap. The graph is created once the simulation is finished. The resonance frequency changes with the change in inductance and capacitance. Table VI presents an overall performance summary. In this analysis, the proposed structure’s 0.50 mm resonator gap is chosen for the better performance of the sensor than others.

The S₂₁ curve for different resonator materials is exposed in Fig. 11. In this study, nickel, silver, gold, and copper were used as resonator materials. We simulate the defined structure and collect data for every resonator material. After the simulation is complete, the graph is produced. Table VII presents an overall performance summary. In this analysis, copper was used as a resonator material for better performance than the others.

Figure 12 shows the S₂₁ response for cross polarization. In the figure, it is seen that the cross polarization is insensitive. The co polarization result is also analyzed and seen as sensitive, and the result is not expected based on frequency and magnitude. So, the structure works as a cross polarizer.

The findings of the simulation have been verified by constructing and measuring the proposed MTM cell. The vector network analyzer is the N5227A, made by Agilent, and is equipped with waveguide ports. Figure 13 depicts the MTM cell’s constructed prototype. Figure 14 shows the unit cell’s experimental configuration. The S₂₁ of the circular SRR unit cell is depicted in Fig. 14, along with the measured and simulated values. 3.35 and 6.85 GHz are the simulated resonances, and 3.41 and 7.01 GHz are the experimental resonances. The S- and C-bands are covered by both of these resonances. Whereas the measured notches are −15.65 and −23.77 dB, the simulated notches are −26.29 and −40.03 dB. It is evident in Fig. 15 that the results of the simulation and measurements differ slightly. This occurrence could have some reasons. First, a calibration error of the Agilent N5227A vector network analyzer (VNA) using the Agilent N4694, 60 001 Ecal, produced a discrepancy in results between those that were measured and simulated. In addition, it could be the result of extremely small errors created when producing the proposed substrate layer. The mutual effect between the sending and receiving ends of two waveguide ports consistently affects the readings, leading to minute differences in both data. Furthermore, this effect will keep having an impact. Finally, but just as importantly, an essential factor in the results is the substrate material’s permittivity.

The designed structure’s EMR is 8.96, since λ = 89.55 mm at 3.85 GHz and L = 10 mm.

Recently, metamaterial sensing has been studied in great detail. Despite the great interest of researchers in this topic, the proposed MTM sensor has been used to investigate the sensing use of various bands. Compared to traditional sensors, microwave sensors offer several advantages, such as rapid measurement, nondestructive, precise measurement, complete automation, and the capability to be manufactured in a lab or online. The simulation setup for the circular CSRR-based MTM sensor model intended for textile material characterizations is displayed in Fig. 16(a). The sensor’s layer arrangement is shown in Fig. 16(b). The arrangement consists of five layers: substrate, copper patch, substrate, and material under test (MUT), i.e., a layer of textile material. Layers of material under test are put between two waveguide ports and positioned between two MTM structures. To analyze the sensing performance, the dielectric constant of the sensing MUTs was varied. This alteration significantly affects the capacitance of the microstrip line, causing shifting at the resonance frequency. The entire MUT sensing process, utilizing the intended structure, is depicted in Fig. 17. Initially, a prototype of the circular CSRR-based metamaterial (MTM) sensor was designed and fabricated. Subsequently, the sample for the material under test (MUT) was prepared. Following that, the MUT was placed between the layers of the MTM sensor. Finally, the sandwiched arrangement was positioned in two waveguide ports and measured using a VNA.

We have used three different textile materials—wool, fleece, and denim—to assess the sensing performance. The permittivity (ɛ_m) and loss tangent (tan δ_m) of wool are 1.7 and 0.031, respectively, and the thickness (t_m) is 0.75 mm. The ɛ_m and tan δ_m of fleece is 1.35 and 0.022 and t_m is 0.6 mm. In the case of denim, ɛ_m = 1.75, tan δ_m = 0.095, and t_m = 0.55 mm. Following the completion of the experimental setup, each textile material’s scattering parameters are the measured, and measured data are used to create graphs. Figures 18(a) and 18(b) illustrate how variations in the dielectric constant affect the behavior of the sensor. These alterations were brought on by an alteration in the patch’s coupling and mutual capacitance. There is a linear relationship between frequency and dielectric constant, with the resonance frequency moving toward the lower frequency as the sensor layer’s dielectric constant rises. The performance of the sensor for various textile material characterizations is registered in Table VIII.

In Table VIII, due to the change in permittivity, it is observed that the f_r is shifted 530 MHz between the air and wool, 420 MHz between the air and fleece, and 640 MHz between the air and denim.

The Q-factor is the ratio of the f_r to the +3 dB bandwidth. It is an important factor to consider while examining a metamaterial sensor’s sensing capacity. It is highly problematic because most MTM sensors have poor quality factors and large errors due to measurements, which restricts their use for a variety of applications. The quality-factor formula is Q = f_r/Δf,⁵⁴ where f_r = resonance frequency and Δf = +3 dB bandwidth. When the dependent variable is equivalent to 70.7% of the minimal transmission value, the difference between the two independent variable values is known as the +3 dB bandwidth. The wool material has been shown to have the highest quality factor. The Q-factor has been calculated using the above formula, and the result is 345. The structure can be used in various textile industries due to its high sensitivity.

This article describes a microwave sensing application that uses a dual circular SRR-based metamaterial sensor with high quality factors and sensitivity. The proposed sensor’s originality is its high sensitivity, high-quality factor, unique structure, and ease of use. The change in resonance frequency is visible due to the small change in the textile material’s dielectric properties. The optimized cell’s dimension is 10 × 10 mm², and the obtained resonances are 3.35 and 6.85 with notches of −26.29 and −40.03 dB, respectively. Different parametric analyses have been done to select the optimum design. The fabricated sensor’s sensing performance is investigated using three distinct textile materials. When the permittivity value changes by 0.40, the sensor’s EMR, quality factor, and sensitivity change by 8.96%, 345%, and 14.57%, respectively. The change in permittivity of the textile materials causes the resonances to shift. There are resonance shifts of 530 MHz between the wool and air, 420 MHz between the fleece and air, and 640 MHz between the denim and air. The outcomes of the simulation are identical to the laboratory results. The suggested sensor can be used to detect textile materials in a variety of garment industries because it is small and affordable, its sensitivity is high, and its quality factor is good.

This research was funded by the Ministry of Higher Education (MOHE) through the Fundamental Reseasrch Grant Scheme (FRGS) under the grant number FRGS/1/2022/TK07/UKM/02/23.

The authors have no conflicts to disclose.

Haitham Alsaif: Conceptualization (equal); Formal analysis (equal); Investigation (equal); Writing – original draft (equal). Md. Rashedul Islam: Conceptualization (equal); Data curation (equal); Investigation (equal); Writing – original draft (equal). Ahasanul Hoque: Data curation (equal); Methodology (equal); Writing – review & editing (equal). Mohamed S. Soliman: Formal analysis (equal); Validation (equal); Visualization (equal). Md. Shabiul Islam: Funding acquisition (equal); Validation (equal); Writing – review & editing (equal). Mohammad Tariqul Islam: Funding acquisition (equal); Methodology (equal); Project administration (equal); Software (equal).

The data that support the findings of this study are available within the article.