Design of Metamaterial-Inspired High-Temperature Microwave Sensor on Alumina Ceramics

Abstract

The issues of high costs and sophisticated circuit structures limit applicability of traditional chipped sensors. This paper innovatively designs a chipless sensor based on the split-ring resonator combined with temperature-sensitive material as the substrate. The alumina ceramics are used as the sensitive material for reflecting the environment temperature in accordance with its characteristic that its dielectric constant increases monotonically with rising temperature. The simulation demonstrates that the resonant frequency of the sensor monotonically decreases from 8.58 GHz to 8.22 GHz with an offset of 0.36 GHz and a sensitivity of 0.9 MHz/°C for a variation from 500 °C and 900 °C. The sensor designed in this paper has good resonance characteristics, is wireless, passive, and low cost, has a planar structure, and is suitable for various harsh environments.

1. Introduction

Since its invention in the 1940s, Radio Frequency Identification (RFID) technology has evolved into a basic device for automatic identification in a wide range of sectors, which include aerospace, machinery production, rail transportation, and medical devices. Traditional temperature sensors must be calibrated or replaced on a regular basis due to their low stability in severe conditions. As a result, RFID technology is becoming increasingly widespread in today’s scientific and technology fields. It utilizes radio frequency to identify target objects under the influence of electromagnetic waves in space. This process does not include a human being and can be used to conduct non-contact two-way communication [1,2]. Since its inception to the present, RFID technology has advanced steadily in terms of efficiency, cost, and performance. More and more industries are integrating RFID technology, and system capabilities are being expanded.

Traditional RFID sensors are based on a built-in silicon chip, using additional electronic components, and detect changes in circuit voltage to characterize changes in physical parameters [3], but the shortcomings are that the design cost is too high and the silicon chip cannot work when the surrounding temperature is above 85 °C. Thus, the scope of use is very limited. Chipless RFID sensors do not need separate sensing circuits. In some unique circumstances or harsh environments (such as high temperature, strong vibration, strong electric fields, strong magnetic fields, etc.), the antenna’s electromagnetic characteristics can still perform their role correctly while removing some precision electronic components; chipless designs especially and dramatically cut expenses and increase the possibilities of use [4].

Numerous academics have presently focused their studies primarily on how RFID and wireless sensing technology can work together.

From the economic point of view, passive sensors are low cost and easy to operate [5]. The research of passive sensors is a popular trend nowadays.

At present, domestic and foreign research on chipless RFID sensors focuses on two aspects: one is the sensing measurement by the physical parameters of the sensitive material itself, and the other is the measurement by using the structural parameters of the sensor in combination with electronic components.

The sensor’s principle and coding method of a single split-ring resonator were investigated [6]. Alumina ceramics were used as both the substrate and the sensitive material. The ambient temperature rose from 28 °C to 1100 °C, and the resonant frequency of the sensor decreased from 2.417 GHz to 2.320 GHz. However, the sensing distance of the sensor was short and only 20 mm in the Maffei furnace temperature experiment.

A temperature sensor consisting of a miniature double piezoelectric wafer cantilever beam (aluminum–silicon) and a single split-ring resonator was designed [7]. Due to its special double-layer structure with different materials on the top and bottom layers, when the external temperature changes, the two materials will expand or shrink to different degrees; thus, the arm beam deflects and changes the frequency response of the split-ring resonator and the resonant frequency is shifted. According to the experimental results derived from the scaling model, the sensitivity of this sensor is as high as 2.5 MHz/°C, but the complexity of its double-layer cantilever structure and the fabrication process is extremely cumbersome. The measured results are less reliable and have poor hysteresis, which is not suitable for mass-production [8].

A passive electromagnetic temperature sensor with multiple input and output delay lines was designed based on the backscattering principle using split-ring resonator (SRR) resonator as well as a thermistor [9]. The thermistor was added to the magneto-inductive wave (MIW) to simulate the temperature increase by heating, and its frequency as well as impedance characteristics were used to determine the connection between its resistance value and resonant frequency.

The research issues mentioned in the majority of studies include the significant cost of delicate materials, bulky tag antenna structures, and challenging fabrication processes. The sensors are preferred to be small and low-cost in the aerospace and manufacturing industries. Since the furnace wall, pipe wall, combustion chamber, and turbine engine have little space, the sensors are mounted on the them and rotated at high speed. Furthermore, the small sensors take some time to react to the temperature varies, preventing real-time measurements. Therefore, the research of a small and low-cost tag antenna has become the main problem of RFID research.

This study introduces a chipless passive sensor that can measure high temperatures based on the split-ring resonator and the operation of radar cross-section. The sensor label can produce a resonant response in a certain frequency range by varying the size and other aspects of the sensor built using HFSS simulation software, allowing the parameters of the sensor to be identified. It is evaluated and discussed how well the sensor performs at high temperatures, and it is explained how to use the resonator to monitor the temperature of the environment.

2. Principle Analysis

Figure 1 illustrates an RFID system for temperature monitoring. The two main components of the RFID measurement system which play a vital role in the generation and reception of a signal from the sensor are the vector network analyzer (VNA) and antenna subsystem. The function of a VNA is to generate the transmitted signals impinged on the sensor and detect the backscattered signals reflected towards the reader. LabVIEW software is used to record the measurement samples of the backscattered signals. The post-processing analysis is performed in MATLAB. The antenna subsystem is utilized to transmit and receive signals. Incident-plane EM waves are used to energize the RFID sensor when placed in the vicinity of the reader. The chipless RFID sensor absorbs electromagnetic waves transmitted by the reader. These EM waves, when impinged on the sensor, stimulate the current on the metallic surface of the resonating structure. In response to this, the modulated backscattered signals which encode data from the sensor containing the exclusive temperature of the attached device under test (DUT) are returned towards the reader.

The proposed passive chipless RFID sensor has the capability of tracking the temperature of an attached DUT. The temperature changes, causing the permittivity of the sensor substrate to change, which makes the operating frequency of the sensor shift considerably. The sensor converts the temperature parameter into a frequency parameter. The shift in frequency of the sensor is observed for variations in temperature. Subsequently, the RFID system can be used to monitor the real-time temperature from the resonant frequency.

The single split-ring resonator was innovatively improved and screen-printed on alumina ceramic to form the sensor [10], which is a simple sensor with low profile and small size. Given that the split-ring resonator on the antenna label is detuned with temperature and that there is a positive linear relationship between temperature and dielectric constant for alumina ceramics, the capacitance of the split-ring resonator changes as a result, which changes the resonant frequency and, ultimately, results in temperature-sensing.

2.1. Antenna Shape

Figure 2a–c show the structure, top view, and side view of the sensor, respectively.

Table 1. Meaning of antenna parameters.

Variables	Parameters
a	Substrate length
b	Split length
c	Distance from substrate edge to resonator split
d	Middle resonator width
t	Substrate thickness

shows the representative meanings of each parameter.

The sensor is composed of the copper metal resonator and the alumina ceramic substrate. The sensor is fabricated by high-temperature co-fired ceramic (HTCC) technology which exhibits merits of corrosion resistance and high thermal conductivity. The sensor signal strength is highly affected by the thickness and surface roughness of the metal film. To reduce unnecessary loss, the HTCC should be smooth. The alumina ceramic is used as the substrate for good mechanical strength. The copper paste with the resonator layout is printed onto the ceramic substrate and then sintered in the furnace with timed temperature control. The sintering process consists of three stages: heating, insulation, and cooling stage. By sintering at a high temperature, the organic material in the copper conductor can be well combined with the ceramic matrix and finally form the sensor.

When compared to the LC helical structure, the SRR structure’s resonant frequency response is more noticeable and its amplitude is significantly larger during transmission spectroscopy. Additionally, the electric field density at the gap increases as a result of charge accumulation [11], providing the SRR structure with a higher quality factor and significantly raising the sensor’s sensitivity.

Based on the above advantages, this paper designs a new single split-ring resonator by rotating two opposite 45° tangent square split-ring 45° counterclockwise and connecting its left and right two ring right angles with metal to increase the current and improve the resonance characteristics; its structure is shown in Figure 2.

There are two splits in the ring of the spiral inductor, so the SRR-equivalent lumped circuit model is approximately estimated by an inductor, as shown in Figure 3. Two side parts, inductors and capacitance, are denoted as L and C, respectively. L and C represent the self-inductor produced by a sheet with a split and self-capacitor produced by the electric charges accumulate at the split. Researchers have shown that the equivalent LC circuit model of a sensor shows the relation between the resonant frequency and capacitance [12,13]. The equation gives the relation between capacitance and permittivity of substrate.

The capacitor, inductor, and excitation power supply are crucial components of the traditional LRC resonant circuit. The current distribution and strength on the resonator’s surface is significantly influenced by their shape and size, respectively. The sensing parameter is the capacitance component, and the peak shift of the resonant response is directly impacted by and reflects changes in capacitance size. The change in temperature alters the dielectric constant which can alter the capacitance parameters [14]. The three states of a circuit are capacitive, inductive, and resistive. The circuit will have a resonant response and produce a resonant frequency when the equivalent impedance is pure resistance. Whether capacitive or inductive, the loop state is constantly in a state of distuning [15].

An LRC resonant circuit can be compared to the SRR structure. The SRR structure produces an induced current to create an induced electric field by positioning the SRR sensor in a changing magnetic field that is absolutely perpendicular to its surface. The resonance effect will alter whenever there is a change in the induced electric field. In the SRR structure, the split ring is equivalent to adding two equal capacitors on either side of the ring, and a weak capacitance will be produced [16]. In contrast, the opposite splits are in opposite directions, resulting in the concentration of the electric field and the opposite accumulation of charge, which produces the capacitance effect. The amount of the SRR structure’s loading capacitance directly relates to the length of the split. A negative linear relationship exists between the size of the SRR structure’s split and the amount of its loading capacitance; the smaller the split, the lower the capacitance.

The temperature-sensing of the sensor is based on the principle that the dielectric constant of alumina ceramic changes in response to temperature changes resulting in the resonant frequency shifting in response to the change in dielectric constant [17]. Therefore, based on the designed antenna shape, this paper verifies the feasibility of the principle through simulation. The substrate length a, split length b, distance between the split and the edge of the substrate c, the middle resonator width d, and substrate thickness t can be adjusted to change the resonant frequency response of a given SRR structure. The variation of inductive and capacitive properties distinctly affects the resonant frequency of the SRR since their fundamental resonance character can be modeled by an LC circuit. The inductive properties are mainly influenced by the dimensions of the metallic component. On the other hand, the capacitive properties are primarily impacted by the structure of the SRR and substrate permittivity. The permittivity change corresponds with the temperature change. The variation of permittivity results in a resonant frequency shift. Thus, this method can also be used to change the capacitance of the sensor monitoring the temperature.

2.2. Sensitive Material

One of the essential components of a sensor label is sensitive material, which has a built-in sensitivity to particular environmental factors and may be used to track changes in the environment. The sensitive material used in this study is an alumina ceramic, which has coherent chemical and physical properties and a temperature-dependent dissipation factor. As the temperature rises, the alumina ceramic’s thermal motion of the ions and dipoles becomes stronger and its dielectric constant soars, which directly impacts the change in resonant frequency [18].

Alumina ceramics are chosen for two reasons: in terms of performance, they have a greater thermal conductivity and can sense the external temperature more sensitively with less error, while they have good chemical properties and higher mechanical strength, making them less susceptible to wear and imparting a longer service life [6]. In terms of cost-effectiveness, alumina ceramics have high temperature resistance and can be used at temperatures up to 1400 °C. The melting temperature of copper is 1080 °C, so the sensor measurement temperature ranges from 500 °C to 900 °C. The backside substrate of the alumina ceramic is pasted on the object under measurement, while the copper side does not contact the object. They have a larger measurement range and are less constrained by environmental factors, which makes them more cost-effective than other temperature-sensitive materials for temperature detection in harsh environments.

3. Simulation Analysis

The dielectric properties of the component materials control the resonant frequency of the resonator, and the dielectric constant of the alumina ceramics continues to climb asymptotic with temperature. As a result, HFSS application is also used to design the structure, maximize the performance, and simulate the temperature change in the monitored target environment by changing the dielectric constant of the sensitive material in order to make the frequency response of the sensor more intelligible.

In this design, the sensor’s sensing function is accomplished by embedding it in a vacuum waveguide that measures 24 mm in length, 24 mm in width, and 16.66 mm in height. The simulation assumes that the designed sensor has a center frequency of 8.5 GHz; that the two surfaces parallel to the YOZ plane and the two surfaces parallel to the XOZ plane are ideal electric conductor surfaces and ideal magnetic conductor surfaces, respectively; and that the electromagnetic waves are excited on the two surfaces parallel to the XOY plane and are excited in the negative direction along the X axis. By setting these special boundary conditions, the alternating electromagnetic field is simulated.

One of the crucial factors that affects the resonator’s physical characteristics is the sensitive material’s dielectric constant. The resonant response of the sensor is controlled by the following settings on the resonator’s dielectric constant.

Alumina ceramic has a coefficient of thermal expansion of 7.2 × 10⁻⁶/°C. Since its antenna volume is 288 mm³, alumina ceramic has a relatively minor dimensional change of 1.603 mm³ in the temperature range of 500 °C to 900 °C. The resonant frequency is primarily determined by the specifications of the resonator construction because the thermal expansion characteristics of the alumina ceramic have less of an effect. The simulation’s standard means of simulating variations in ambient temperature is via varying the magnitude of the alumina ceramic’s dielectric constant. There is a distinct minimum value for the transmission response curve at each temperature, as shown by the observed transmission response graph for the identified sensor size. The resonant frequency of the resonator is represented by this minimum value, also known as the resonant point. By comparing the level of the received signal in the frequency domain, the sensor’s interrogation unit will be able to determine this minimum [19].

The substrate length, split length, distance from substrate edge to resonator split, middle resonator width, and substrate thickness are separately studied to optimize the resonant frequency of reflection coefficient S₂₁, as shown in Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8. Meanwhile, the relationship between the different dimensions with varying dielectric constant and reflection coefficient S₂₁ are simulated, as shown in Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8. The dielectric constant values of the substrate are changed from 10.17 to 10.91. In addition to this, the parameters of a, b, c, d, e, and t set as 5–7 mm with step 0.5 mm, 1.5–3.5 mm with step 0.5 mm, 2.5–4.5 mm with step 0.5 mm, 0.1–4.1 mm with step 1 mm and 0.3–0.7 mm with step 0.1 mm, respectively, are optimized in HFSS.

For a given dimension, one of these parameters is optimized by fixing the other four parameters in the simulation. Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8 indicate that the physical dimensions of the sensor slightly alter the result of the resonant frequency S₂₁ dramatically deteriorating due to the electrical properties of the sensor changing. Figure 4, Figure 5, Figure 6 and Figure 7 show that all dimensions of substrate length, split length, distance from substrate edge to resonator split, and middle resonator width are resonant with variation of dielectric constant. Minority dimensions of substrate thickness are shown to be resonant in Figure 8. It illustrates that substrate thickness is the most sensitive to S₂₁. This is because substrate thickness directly affects the sensitive material’s electrical characteristics. The substrate length, distance from substrate edge to resonator split, and middle resonator width impact the inductance of an equivalent model; on the other hand, the split length affects the capacitance of an equivalent model. Therefore, the dielectric constant changing can be clearly detected through the offset of the resonant frequency, with the largest offset being 0.36 GHz. Finally, the optimized values are a = 5.5 mm, b = 3.5 mm, c = 3.5 mm, d = 2.1 mm, and t = 0.5 mm, since they perform a monotone resonant frequency with increasing dielectric constant.

Figure 4b, Figure 5e, Figure 6c, Figure 7c and Figure 8c show the same and best resonant frequency response among each parameter. When the dielectric constant rises from 10.17 to 10.91 corresponding to the temperature rising from 500 °C to 900 °C with step 100 °C, the resonant frequency falls from 8.58 GHz to 8.22 GHz with a 0.36 GHz offset. Therefore, the proposed sensor implements the temperature-sensing function and detects variation in the ambient temperature by adjusting the sensor’s physical parameters.

The link between the resonant frequency and the experimentally determined dielectric constant of alumina ceramic is shown in Figure 9. The resonant frequency of the sensor label gradually drops as the dielectric constant rises. The resonant frequency drops down pretty gradually between dielectric constant values of 10.17 and 10.80. The velocity increases and the resonant frequency lowers as the dielectric constant rises from 10.80 to 10.91.

The literature [2] provides a link between the temperature of alumina ceramics and the dielectric constant. Figure 10 illustrates the relationship between the dielectric constant of alumina ceramics between 50 °C and 1000 °C. The fitting relationship between the alumina ceramics’ dielectric constant and the ambient temperature is discovered during nonlinear fitting. In the simulation temperature range, the alumina ceramics’ dielectric constant rises as the temperature rises, and the growth rate rises gradually.

$$\epsilon =8E-0.7T^2+0.0007T+9.6499$$

(1)

where ε is the dielectric constant of alumina ceramic and T is the ambient temperature.

The main cause of the two outcomes mentioned above is that, as temperature rises, the thermal motion of the ions and dipoles in alumina ceramics is enhanced, and as a result the dielectric constant rises. The capacitor effect is strengthened, the charge in the resonator rises, the induced current rises, and the resonant frequency falls all at the same time.

The link between the sensor’s resonant frequency and temperature characteristics is depicted in Figure 11 in order to clearly illustrate the aforementioned impact. Figure 11 is created by integrating the relationship between “temperature–dielectric constant” of alumina ceramics in reference [2] and Figure 9 by integrating the relationship of “dielectric constant–resonant frequency” obtained through simulation. The one-to-one relationship between “temperature–resonant frequency” is easier to see in Figure 11: as temperature rises, the resonant frequency displays a falling trend, and the speed of the decrease is steady.

According to Formula (2), the sensitivity of the sensor is calculated to be 0.9 MHz/°C, that is, when the temperature changes by 1 °C, the resonant frequency of the sensor changes by 0.9 MHz.

$$\eta =\frac{|f_1-f_2|}{|T_1-T_2|}$$

(2)

where f₁ and f₂ are the measurement of the resonant frequency and T₁ and T₂ are the corresponding temperature of the resonant frequency.

Based on the above simulation data, it can be seen that the performance of the sensor is stable and the sensitivity is high.

4. Conclusions

In this research, a RFID chipless passive temperature sensor with a modified split-ring resonator and metal metamaterial substrate of alumina ceramic is designed. It has the advantages of a simple structure, low profile, and stable material. The stable temperature-sensitive material is designed using alumina ceramics. The equivalent circuit of the sensor is analyzed. The principle of a chipless RFID system describes that the permittivity of the sensing substrate changes with temperature, resulting in variation of the sensor’s resonant frequency. Simulation shows that the resonant frequency decreases approximately linearly with temperature. The designed sensor operates from 500 °C to 900 °C and responds with a shift in resonant frequency of 0.36 GHz and average sensitivity of 0.9 MHz/°C. Therefore, the designed sensor is suitable for real-time monitoring of temperature through observing the resonant frequency.