A Novel-Structure LC Resonant Passive Wireless Sensor for NO₂ Sensing

Abstract

This work presents an LC resonant passive wireless gas sensor with a novel structure designed to mitigate the negative impact of substrate. The LC sensor antenna in the new structure, and the reader antenna, were designed and optimized utilizing HFSS software to improve the transfer efficiency. The superiority of the designed structure compared with general examples is highlighted and verified. The change in the substrate capacitance essentially makes no interference with the parameters of the LC sensor to be measured. The sensor for the new structure was prepared by combining etching and sputtering methods. The ZnO nanowires (NWs) were characterized to confirm their high purity and wurtzite crystal structure. The LC gas sensors demonstrated excellent wireless sensing performance, including a low detection limit of 0.5 ppm NO₂, high response of 1.051 and outstanding stability at 180 °C. The newly developed sensor structure not only prevented interference from the substrate during gas sensing testing, but also expanded the choice of sensor substrates, playing a critical role in the development of sensors based on the LC resonance principle.

1. Introduction

Nitrogen dioxide (NO₂) is a significant atmospheric contaminant primarily generated via the combustion of diverse fuels such as coal, chemicals, and natural gas in power generation facilities [1,2]. Large amounts of NO₂ exhaust have severe detrimental effects on the environment and human health, ranging from acid rain, photochemical smog and global warming to respiratory problems [3,4,5]. Due to the safety hazards and excessive size of conventional gas detection devices, it is very difficult to detect NO₂ in some special application scenarios, such as flammable and explosive, rotating, and confined spaces. As such, there is a great demand for a wireless and passive gas sensor to enable effective monitoring. To date, various wireless sensing devices have been created to monitor harmful gases, which use different measuring mechanisms, such as surface acoustic wave (SAW) [6,7,8], wireless passive LC resonance [9,10], and microwave evanescent mode [11,12], etc.

The wireless passive inductor-capacitor (LC) resonant approach is optimal for these special application scenarios because it facilitates energy conversion and data reading due to its lower operating frequency and stable signal transmission capability [13]. Gas sensors based on the LC resonance principle are increasingly popular due to their ease of readout for measured concentration and low installation costs [14]. Such sensors consist of LC antennas and gas sensing materials. When the gas parameter to be measured alters, the electrical properties of the gas sensing material will be changed, which in turn causes a variation in the input impedance of the sensor. Non-contact testing of the sensor signal can be achieved via an external signal reading system connecting a reader antenna to the sensor in a near-field coupling [15].

The use of LC sensors for gas detection has been extensively researched in numerous studies. For instance, Azzarelli et al. combined gas sensing materials with near field communication (NFC) tags and converted the gas sensing analog signal into a digital signal to recognize NH₃ in confined environments via smartphones [16]. Ma et al. utilized low-temperature cofired ceramic (LTCC) technology to fabricate LC gas sensors for NO₂ gas sensing [17]. Ma et al. developed a high-performance nanostructured conductive polymer as a switch material for NFC tags to enable the detection of NH₃ [18]. Overall, current research efforts primarily concentrate on innovations in application [19,20], the preparation process of sensor and substrate materials [21], and synthesis of gas sensing materials [22,23]. The structure of LC gas sensors is relatively homogeneous, i.e., gas sensing materials are coated on the interdigital capacitors (IDC) directly, or coated after adding a dielectric layer. This is inherently deficient due to the direct contact between the underside of the IDC and the substrate. Changes in the physical properties of the substrate can significantly affect the measured parameters. Our previous work had explored the use of high-performance ZnO nanowires (NWs) with wurtzite structure [24,25]. However, it is still a major challenge to optimize the sensor structure so that the existing high-performance gas sensing materials can be effectively combined with the LC resonant circuit to achieve low-cost, accurate, efficient, and stable wireless NO₂ gas monitoring [26].

In this work, a novel LC sensor structure was developed in which the IDC of the sensor no longer contacted the substrate. The advantages of this structure are discussed and verified, and the sensor structure implemented using a combination of etching and sputtering methods. The LC antenna and reader antenna are designed and optimized by utilizing HFSS, and good signal transmission was obtained. This work provides a new sensor structure based on the mutually inductive principle.

2. Experimental

2.1. Design and Optimization of Antennas

The frequency design of sensor antennas not only requires a consideration of the application scenario, but also demands an evaluation of the material of the substrate. In this work, FR4 was used as the substrate, which had a very low dielectric constant of 4.4 and showed a very fast signal transmission rate. Nonetheless, an excessively high operating frequency may lead to substantial losses. Thus, careful consideration should be given to ensure that the operating frequency remains below 0.5 GHz [27].

The relationship between the operating frequency ($f_o$) of an LC antenna and its inductance ($L_S$) and capacitance ($C_S$) values can be expressed as follows [17].

$$f_o=\frac{1}{2\pi \sqrt{L_S{C}_S}}$$

(1)

Here, the values of $L_S$ and $C_S$ are reliant upon the structural dimensions of the circuit [17].

$$L_S=1.39\times {10}^{-6}\left(d_0{ + d}_i\right)N_i^{5/3}\mathit{log}\left(4\frac{d_o+d_i}{d_o-d_i}\right)$$

(2)

$$C_S=l_c\left(N_c-1\right){\epsilon }_0\frac{1+{\epsilon }_{r}K\left[{\left(1-{\left(d_s/d_c\right)}^2\right)}^{1/2}\right]}{2K\left(d_s/d_c\right)}$$

(3)

where $N_i$ is the number of loops, and $d_o$ and $d_i$ denote external and internal coil widths of the inductor, respectively. $l_c$ represents the length of the IDC fingers, $N_c$ refers to the count of contact points on each IDC terminal. ${\epsilon }_0$ denotes the air permittivity. ${\epsilon }_r$ is the dielectric constant of the FR4 substrate. $d_s$ refers to the width of the orifice between two neighboring fingers, while $d_c$ represents the total spacing between two adjacent fingers. $K$ is a mathematical function known as the complete elliptic integral of the first kind.

A three-dimensional model was established and simulated via HFSS. To fulfill the operating frequency specifications, it may be necessary to tune the size of the LC antenna. A flat helical inductor and IDC are connected through vias on the substrate to form a resonant circuit with ZnO NWs employed as the gas sensing material (Figure 1). Upon exposure to NO₂ molecules, the charge transfer takes place and then alters the $S_{11}$ of the LC resonant circuit. The designed parameters of LC antenna are displayed (

Table 1. Parameters of the designed LC antenna.

lc	dc	ds	Ni	do	di	Nc
10.5 mm	1.5 mm	0.5 mm	4.75	20.5 mm	12.5 mm	5

).

HFSS software can be utilized to determine both the size and height of the reader antenna. The S₁₁ parameters of the LC antenna can also be optimized. The smaller the $S_{11}$ parameter, the higher the signal or energy transfer between the reader antenna and the LC antenna. To obtain better signal transmission efficiency, the S₁₁ parameter is required to be less than −10 dB. After optimization, the LC antenna is prepared and tested practically to examine the resonant frequency and Voltage Standing Wave Ratio (VSWR) of the LC antenna.

2.2. Synthesis and Characterization of ZnO NWs

ZnO NWs were produced using a solvothermal method. Specifically, combine 2.5 g of NaOH, 0.1 g of sodium dodecyl sulfate (C₁₂H₂₅SO₄Na, SDS), and 0.44 g of Zn (CH₃CO₂)₂·2H₂O in 117 mL of ethanol and mix well until the components are fully dissolved. Then, add 27 mL of polyethylene glycol 400 (PEG400) to the mixture and stir thoroughly. The resulting solution was heated to 140 °C for 16 h. Following this step, the white sediment was gathered by centrifugation and washed several times to completely rid it of residual organic and inorganic ions, and eventually dried at 60 °C. The dried samples were then annealed in a tubular furnace for 3 h at a temperature of 500 °C. Subsequently, natural cooling of the samples to room temperature was carried out to achieve ZnO NWs for LC gas sensors. NaOH, Zn (CH3CO₂)₂·2H₂O, andPEG400 were purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China; SDS was purchased from Shanghai Macklin Biochemical Co., Ltd, Shanghai, China. All the reagents were analytical grade and deionized water was used in all experiments.

The crystalline phases, microscopic morphology, and surface chemistry of the ZnO NWs were characterized by using an X-ray diffractometer (XRD-7000, Shimadzu Co., Ltd., Tykyo, Japan), scanning electron microscope (TESCAN MIRA LMS, Tescan Co., Ltd., Brno, Czech), transmission electron microscope (FEI Tecnai G2 F20, FEI Co., Ltd., Hillsboro, United States of America) and X-ray photoelectron spectrometer (Thermo Scientific K-Alpha X, Thermo Scientific Co., Ltd., Waltham, United States of America). Additionally, the selected area electron diffraction (SAED) patterns were recorded for further analysis.

2.3. LC Sensor Fabrication and Gas Sensing Measurement

The schematic diagram of the manufacturing process of the LC gas sensor is shown (Figure 2). The inductor was processed on an FR4 substrate with the etching method to obtain the LC antenna after ultrasonic cleaning for 10 min. Subsequently, the as-prepared ZnO NWs were homogeneously dispersed in a proper amount of absolute ethanol in an ultrasonic cleaner for 5 min to form the sensing slurry, which was then coated on the FR4 substrate to form a sensing layer with a thickness of approximately 100 μm. Then, the substrate with the ZnO NWs was placed in an oven for drying at 60 °C. After natural cooling, a plasma sputtering device was used to sputter platinum electrodes on its surface. Finally, the LC sensor was aged at 210 °C for 6 h.

A schematic diagram of the wireless gas sensing measurement setup is shown (Figure 3a). The LC gas sensor was positioned on a temperature-controlled heating table within a sealed gas chamber, while an external reader antenna was connected to a VNA (AV3656D) and placed parallel to the sensor at around 5 mm. This configuration enabled precise measurement and analysis of the sensor’s response to the presence of NO₂ gas. Subsequently, the network analyzer is used to analyze the information collected by the reader antenna. Measurements of the $S_{11}$ and corresponding frequency ($f_o$) were recorded at regular 10-s intervals. The response of the LC sensor was determined as $S$_11g/$S$_11a, in which $S$_11a and $S$_11g represent the $S_{11}$ of the LC sensor in air and in NO₂, respectively (Figure 3b). The determination of the response and recovery times of the sensor involves measuring the duration needed for a 90% alteration in the fully normalized $S_{11}$ amplitude. To compare responses, all measured $S_{11}$ were divided by the initial $S_{11}$ magnitude, resulting in a “normalized $S_{11}$” that started from −1.0. This approach ensured that all baselines originated at the same value, irrespective of any added mass attributable to the sensitive layer [28].

In the equivalent circuit of the radio signal transmission (Figure 4b), $R_a$ and $R_s$ represent the series resistances of the reader antenna and the sensor, respectively. $R_p$ is the leakage resistance of the IDC, which is equivalent to the parallel resistance of the $C_s$. $L_a$ is the inductance of the reader antenna, while $L_s$ is the inductance of the sensor. $k$ represents the degree of coupling between the reader antenna and the sensor, which is related to the relative position of the two inductors, was assumed to be constant during the measurement process, and therefore could be considered as a fixed value. The relationship between the input impedance ($Z_a$) and the above parameters can be expressed by Equations (4)–(6) [29], indicating that changes in both $C_s$ and $R_p$ affect the change in $Z_a$ in the circuit. The change of $Z_a$ in this work depends mainly on the changes in $R_p$, which are brought about by the interaction between NO₂ molecules and ZnO NWs. It can be inferred from Equation (7) that $S_{11}$ is equivalent to $Z_a$. As such, monitoring changes in the $S_{11}$ parameter may provide valuable insights into variations in gas properties.

$$Z_a=Z^{\prime }+j{Z}^{″}=\frac{{\omega }^2{k}^2{L}_s{L}_{a}A}{A^2+B^2}+j\left(\omega L_a-\frac{{\omega }^2{k}^2{L}_s{L}_{a}B}{A^2+B^2}\right)$$

(4)

$$A=R_s+\frac{R_p}{1+{\omega }^2{C}_s^2{R}_p^2}$$

(5)

$$B=\omega L_s-\frac{\omega C_s{R}_p^2}{1+{\omega }^2{C}_s^2{R}_p^2}$$

(6)

$${\mathrm{S}}_{11}=\left({\mathrm{Z}}_{\mathrm{a}}-{\mathrm{Z}}_0\right)/\left({\mathrm{Z}}_{\mathrm{a}}+{\mathrm{Z}}_0\right)|{\mathrm{Z}}_0=50\mathsf{\Omega }$$

(7)

3. Results and Discussion

3.1. Simulation Result

Firstly, the S₁₁ parameters of the LC antenna are optimized by using the side length of the reader antenna. The side lengths of 0.8d_i, d_i, (d_i + d_o)/2, d_o, and 1.2d_o are selected for simulation using a square reader antenna. The simulation outcome of the $S_{11}$ parameter under various side lengths of the reader antenna (Figure 4a) demonstrates that, when the reader antenna is set to $\left(d_o+d_i\right)/2$, $S_{11}$ has the smallest value of −11.1 dB. Subsequently, the S₁₁ parameters were then optimized by means of optimizing the reading height of the reader antenna.

The simulation results for different reading heights with a side length of $\left(d_o+d_i\right)/2$ are shown (Figure 4b). At a reading height of 5 mm, the value of $S_{11}$ is a minimum of −19.2 dB. This value indicates a slight signal and energy loss between the LC antenna and the reader antenna.

The resistance variation of ZnO NWs was simulated to replicate the actual exposure to NO₂, whereas alterations in the $S_{11}$ of the LC sensor were monitored. It is discernible that the $S_{11}$ parameter of the designed LC antenna diminishes in a certain pattern with the rise in leakage resistance of the sensor (Figure 4c), demonstrating the design agreement with the requirements. Therefore, the LC antenna and the reader antenna were prepared according to the structural parameters.

A reading height of 5 mm was selected to test the LC antenna. The operating frequency of the antenna is 156.8 MHz, which satisfies the frequency requirements below 0.5 GHz. Deviations between the simulated and measured values are mainly attributed to parasitic effects [30]. At the operating frequency, the VSWR of the LC antenna is observed to be 1.238 (Figure 4d), which is less than 1.5, thus confirming that the designed antenna can effectively transfer energy and information via the reader antenna.

3.2. Novel Structure of the LC Sensor

The structure (Figure 5a) and physical diagram of the LC sensor (Figure 5b) are shown, respectively. Generally, the IDC is processed first, followed by the application of the gas sensing material. In contrast, this approach involves coating the gas sensing material onto the substrate and then sputtering the platinum on the IDC once the material has been fully adhered to the substrate. The most significant departure of this special structure from its conventional counterpart lies in the fact that the IDC no longer directly contacts the substrate.

Comparing the two sensor structures, the equivalent circuit diagrams are shown (Figure 6). In the general structure (Figure 6a), $R_1$ and $R_2$ represent the leakage resistances of the IDC, while $R_p$ denotes the parallel resistance of $R_1$ and $R_2$. Additionally, $C_S$ refers to the total capacitance of $C_1$ and $C_2$ arranged in parallel. In comparison, in this work (Figure 6b), $R_1$ and $R_3$ are the leakage resistances of the IDC, while $R_p$ represents the parallel resistance of $R_1$ and $R_3$. Furthermore, $C_S$ corresponds to the total capacitance of $C_1$ and $C_3$ arranged in parallel. Herein, $R_1$ and $C_1$, $R_2$ and $C_2$, and $R_3$ and $C_3$ represent the equivalent resistance and capacitance of gas sensing material, substrate, and free space, respectively. They are discussed as the main variation parameters. In practical application, $R_1$ and $C_1$ of the gas sensing material will alter with changes in gas concentration. As for $R_2$ and $C_2$, they will change due to the differences in the substrate, and $C_2$ will also change because of variations in gas adsorption by the substrate. However, $R_3$ and $C_3$ change as the gas atmosphere alters, but the change is relatively minor. Equations (4)–(7) reveal that, for the general coating method (Figure 6a), changes in the substrate and $C_2$ during the testing process will interfere with the alteration of $C_S$, thereby potentially affecting test results. In contrast, the structure utilized in this work (Figure 6b) involves the upper surface of the IDC contacting the surrounding air. As a result, $C_3$ and $R_3$ remain nearly unchanged, thus eliminating interference from the substrate.

Verification was conducted via HFSS software simulation. A fixed resistance ‘box’ was employed to simulate the gas sensing material. It was separately placed above and below the IDC to build the two different structures. Additionally, a variable capacitance ‘sheet’ was positioned above the substrate to replicate alterations in $C_2$, with a change range set between 3.5–4.5 pF (Figure 7). The thickness of the ‘box’ is 100 μm and the ‘sheet’ is thickness-free. Simulation was executed to observe both the $S_{11}$ parameter of the sensor antenna and the change in resonant frequency $f_0$ using the $f-S_{11}$ curve.

The variation of $S_{11}$ and $f_0$ of the sensor with two different structures is shown (Figure 8). For clarity, $S_{11}$ and $f_0$ values extracted at a capacitance of 3.5 are selected as initial values ($S_I$ and $f_I$). Equations (8) and (9) express the changes in $S_{11}$ and $f_0$, denoted as $S_V$ and $f_V$, respectively. It is noteworthy that the $S_{11}$ (Figure 8a) and $f_0$ (Figure 8b) values of the LC sensor, utilizing the proposed sensor structure, exhibit minimal sensitivity to changes in substrate capacitance of $C_2$.

$$S_V=\frac{S_{11}-S_I}{S_I}$$

(8)

$$f_V=\frac{f_0-f_I}{f_I}$$

(9)

The simulation results provide clear evidence that the proposed sensor structure outperforms the conventional one, thereby effectively mitigating the detrimental effects of substrate gas adsorption on gas sensing testing. Furthermore, such a new structure can extend the range of substrate options for LC sensor.

3.3. Characterizations of ZnO NWs

The XRD pattern of the as-synthesized sample (Figure 9) indicated that the obtained samples are high purity ZnO with a hexagonal wurtzite structure (ICDD #99-0111) [31]. The as-prepared sample was observed by SEM and TEM (Figure 10). The ZnO samples exhibit a uniform and wire shape with smooth and non-branched surfaces and excellent dispersion (Figure 10a–c). The ZnO products display a narrow diameter distribution, ranging from 20–50 nm. Additionally, for a single ZnO NW, the diameter displays hardly any variation along the long axis. The length distribution of the ZnO NWs is slightly wider, ranging from 500 nm to a few micrometers. The excellent crystallinity of the synthesized ZnO NWs is evident from the clear diffraction fringe observed (Figure 10e), where the interplanar spacing between adjacent fringe is 2.6 Å. Based on the XRD results, it is determined that this spacing corresponds to the (002) crystal plane of the hexagonal fibrillated ZnO structure. This result indicates that the ZnO NWs are formed by the growth of ZnO grains along the c-axis direction (Figure 10d) [32]. Furthermore, the SAED pattern of the ZnO NWs (Figure 10f) displays a distinct diffraction dot pattern, providing evidence for the single-crystal nature of the as-prepared ZnO NWs.

The XPS analysis results of the ZnO NWs is shown (Figure 11). The full spectrum scan (Figure 11a), indicating the presence of spectral peaks for Zn, O, and C elements only, further confirms the high purity of the ZnO NWs. The high-resolution spectrum of Zn 2p (Figure 11b) reveals two distinct characteristic peaks. The peak at a binding energy of approximately 1021.27 eV corresponds to Zn 2p_3/2, while the peak at a binding energy of about 1044.33 eV corresponds to Zn 2p_1/2, with an energy interval of approximately 23 eV between them. This indicates that the element Zn in the ZnO NWs exists in the +2 valence form [33]. In the high-resolution spectrum of O 1s (Figure 11c), the observed peak at a binding energy of approximately 530.02 eV is caused by lattice oxygen in ZnO crystals. Additionally, the peaks at binding energies of around 531.04 eV and 536.04 eV are primarily associated with adsorbed oxygen on the material surface, as well as oxygen vacancies [34,35].

3.4. NO₂ Sensing Properties of the LC Sensor

According to the electrical model presented in previous reports, by analyzing the $S_{11}$ and $f_o$ data acquired from the reader antenna, it is possible to infer changes in the surrounding environment. The $f-S_{11}$ curves of the substrate alone, the LC antenna, and the LC gas sensor in both air and NO₂ environments at an operating temperature of 180 °C are measured (Figure 12). A pronounced decline in $f_o$ and $S_{11}$ amplitude can be observed when comparing the LC gas sensor to the LC antenna, both in air. This is attributable to the fact that ZnO possesses higher dielectric constant and conductivity values compared to those of air. The coating of ZnO on the IDC leads to an increase in its capacitance and a decrease in its leakage resistance. In comparison to the sensor in air, the $f_o$ and $S_{11}$ magnitudes of the sensor increase following exposure to NO₂. This increase is primarily attributed to the adsorption of NO₂ on the ZnO NWs, which subsequently changes the input impedance of the circuit and thereby modifies the $f_o$ and $S_{11}$. Moreover, it is evident that the $S_{11}$ of the LC sensor exhibits a measurement value of less than −10 dB both before and after exposure to NO₂. This observation further confirms the high energy and information transfer efficiency of the LC gas sensor with the reader antenna.

The operating temperature has the greatest effect on the gas sensing properties of the LC gas sensor based on the metal oxide semiconductor, therefore the response of the LC sensor at variable temperatures was initially evaluated [36]. The response variation of the LC gas sensor to 1 ppm NO₂ occurred across an operating temperature range of 60–210 °C (Figure 13a). As the operating temperature of the sensor increases, its response exhibits a gradual rise and reaches a peak value at 150 °C, after which it decreases gradually with further increases in the operating temperature. This phenomenon can primarily be attributed to the significant effect of operating temperature on both the number of surface-active sites of the ZnO NWs and the kinetics of gas adsorption and reaction.

The response/recovery times of the LC sensor were assessed at different operating temperatures for detecting 1 part per million (ppm) concentration of NO₂ (Figure 13b). The response/recovery times of the LC gas sensor decrease rapidly with increasing operating temperature. This phenomenon can primarily be attributed to the significant acceleration of NO₂ molecule adsorption and desorption processes on the ZnO NWs at higher operating temperatures, effectively shortening the response/recovery times. However, it should be noted that the sensor response decreases at excessively high temperatures. Therefore, a suitable combination of the LC sensor response and response/recovery times is necessary when determining the suitable operating temperature of the sensor [37]. In this work, the LC gas sensor exhibits its maximum response to 1 ppm NO₂ at 150 °C with a value of 1.058, with only a minor difference in response observed at 180 °C with a value of 1.051. However, notable discrepancies are observed in the response/recovery times at 180 °C compared to that recorded at 150 °C. Therefore, it is determined that the optimal operating temperature of the designed LC sensor is 180 °C.

To thoroughly examine the gas sensing capabilities of the LC sensor at its optimal operating temperature, response and recovery curves were obtained for a range of NO₂ concentrations (0.5, 1, 2, 3, and 4 ppm) (Figure 13c). These measurements provided a comprehensive understanding of the sensor’s performance under varying NO₂ concentrations. The LC gas sensor exhibits excellent response–recovery characteristics across a range of NO₂ concentrations. Additionally, the response of the LC sensor to different concentrations of NO₂ at 180 °C was demonstrated (Figure 13d). Notably, a strong linear relationship between the response and NO₂ concentration is observed within the concentration range of 0.5–4 ppm. This observation suggests that the designed sensor may be well-suited for quantitative detection and analysis of NO₂. It is also noteworthy that the LC gas sensor is capable of effectively detecting NO₂ at extremely low concentrations of 0.5 ppm. This underscores the potential value of the LC gas sensor in various industrial and environmental applications.

Given their passive property, the LC gas sensors are highly suitable for use in scenarios where battery replacement is impractical or difficult. However, this feature also presents a challenge to the stability of such sensors, as their reliability and long-term performance must be maintained without the benefit of regular power source maintenance. The response–recovery curves of the LC sensor to 1 ppm NO₂ at 180 °C over five consecutive tests are shown (Figure 14a). The sensor exhibits consistent response–recovery characteristics across each test cycle. Furthermore, the S₁₁ parameter of the sensor can revert to its initial value after each test, indicating an excellent reversibility, repeatability, and stability. To further assess the stability of the sensor over a longer time, the changes in sensor response to 1 ppm NO₂ over a 60-day timeframe were recorded (Figure 14b). Notably, only minor fluctuations are observed in the response of the sensor over this extended period, underscoring its exceptional stability for detecting NO₂ at 180 °C.

To further investigate the selectivity of LC sensors for different kinds of gases, the responses of the sensor towards different gases are evaluated at the optimal operating temperature. The response of the LC sensor towards NO₂ at 180 °C is significantly higher than that of other tested gases (Figure 15), indicating exceptional NO₂ selectivity of the developed LC gas sensor.

3.5. Gas Sensing Mechanism of the LC Sensor

The gas sensing mechanism of the LC gas sensor to NO₂ is briefly described (Figure 16). When the sensor is in air, a significant number of oxygen molecules are adsorbed on the surface of ZnO NWs in the form of chemisorbed oxygen of ${\mathrm{O}}_2^{-}$, ${\mathrm{O}}^{-}$, or ${\mathrm{O}}^{2-}$ [38,39]. These adsorbed oxygen species trap electrons from the conduction band of the ZnO NWs, resulting in the formation of an electron depletion layer on the ZnO surface and a subsequent decrease in carrier concentration. As a result of these changes, the resistance of the material increases, with the corresponding chemical reaction equations shown in Equations (10)–(13) [40,41].

$${\mathrm{O}}_2\left(\mathrm{g}\right)\to {\mathrm{O}}_2\left(\mathrm{ads}\right)$$

(10)

$${\mathrm{O}}_2\left(\mathrm{ads}\right)+e^{-}\to {\mathrm{O}}_2^{-}\left(\mathrm{ads}\right)$$

(11)

$${\mathrm{O}}_2\left(\mathrm{ads}\right)+{2\mathrm{e}}^{-}\to 2{\mathrm{O}}^{-}\left(\mathrm{ads}\right)$$

(12)

$${\mathrm{O}}_2\left(\mathrm{ads}\right)+{4\mathrm{e}}^{-}\to 2{\mathrm{O}}^{2-}\left(\mathrm{ads}\right)$$

(13)

When the sensor is exposed to NO₂, the surface of ZnO NWs becomes chemisorbed with $\mathrm{N}{\mathrm{O}}_2$. This chemisorbed state of $\mathrm{N}{\mathrm{O}}_2^{-}$ further traps electrons in the conduction band of ZnO and reacts with the surface adsorbed oxygen. Both reactions lead to the capture of extra electrons from the ZnO conduction band and expansion of the depletion layer, causing a reduction in carrier concentration. Consequently, this results in an increase in ZnO NWs resistance and ultimately an increase in $S_{11}$. The corresponding reaction equations are provided as follows.

$$\mathrm{N}{\mathrm{O}}_2\left(\mathrm{g}\right)\to \mathrm{N}{\mathrm{O}}_2\left(\mathrm{ads}\right)$$

(14)

$$\mathrm{N}{\mathrm{O}}_2\left(\mathrm{ads}\right)+{\mathrm{O}}^{-}\left(\mathrm{ads}\right)+2{\mathrm{e}}^{-}\to \mathrm{N}{\mathrm{O}}_2^{-}\left(\mathrm{ads}\right)+{\mathrm{O}}^{2-}\left(\mathrm{ads}\right)$$

(15)

$$\mathrm{N}{\mathrm{O}}_2\left(\mathrm{ads}\right)+{\mathrm{e}}^{-}\to \mathrm{N}{\mathrm{O}}_2^{-}\left(\mathrm{ads}\right)$$

(16)

When the sensor exposed to air again, NO₂ molecules are desorbed from the surface of ZnO NWs, and the reaction proceeds in reverse. Electrons seized by NO₂ are released back into the conduction band of the material, causing an increase in carrier concentration and a decrease in the electron depletion layer. These changes eventually result in a reduction in resistance of the sensor and a corresponding decrease in $S_{11}$ magnitude value, which gradually returns to its initial value.

4. Conclusions

A new LC gas sensor with a novel sensor structure was successfully developed, in which the IDC of the LC sensor no longer contacted the substrate, reducing the adverse effect of the substrate. Following measurement, it is found that the VSWR of the designed sensor antenna was 1.238, which is less than 1.5, and the $S_{11}$ parameter of the LC sensor remained below −10 dB both before and after exposure to NO₂, indicating an effective improvement in transfer efficiency between the reader antenna and the LC antenna. A comprehensive study of the gas sensing property for the prepared LC sensor to NO₂ was fully conducted, and the results revealed that the optimal operating temperature was 180 °C. At this temperature, the sensor showed high response of 1.051 towards NO₂ and a low detection limit of 0.5 ppm NO₂, while maintaining stability during the 60-day testing period. Moreover, this work provides a novel sensor structure based on the inductive coupling principle, broadening the range of substrate options, and significantly reducing interference from the sensor substrate during testing.