Highly Sensitive and Selective MEMS Gas Sensor Based on WO₃/Al₂O₃/Graphite for 2-Chloroethyl Ethyl Sulfide (2-CEES) Detection

Abstract

The detection and monitoring of toxic and harmful gases play a vital role in environmental protection, human health, and industrial and agricultural production. However, it is still challenging to develop gas sensors for the detection of toxic and harmful gases with high sensitivity, good recovery and excellent selectivity. In this study, WO₃/Al₂O₃/graphite composite materials were used for an MEMS 2-CEES gas sensor (dichlorodiethyl sulfide simulation), and the corresponding sensing properties were explored. The experimental results show that when the working temperature is 340 °C, the response of the sensor to 2-CEES gas with a concentration of 5.70 ppm is 69%, the response time is 5 s and the recovery time is 42 s. The sensor also has the advantages of long-term stability and high selectivity. Furthermore, the MEMS gas sensor array based on WO₃/Al₂O₃/graphite composite materials has been achieved and also exhibits excellent sensing performance. Overall, this study provides a strategy for realizing high-performance dichlorodiethyl sulfide gas sensors.

1. Introduction

According to the mechanism of toxicity in humans, toxic and harmful gases are classified into blister agents, nerve agents, blood agents, and pulmonary agents [1]. As the common blistering agent, dichlorodiethyl sulfide is notorious for harming human health, such as the erythema, edema, blistering, erosion, etc. [2]. In order to prevent the potential harm to human health, the sensitive detection of toxic and harmful gases is required. As the testing of dichlorodiethyl sulfide at laboratory levels is dangerous, experimental studies are usually performed with less toxic chemicals, among which 2-chloroethyl ethyl sulfide (2-CEES) is used as a simulant for dichlorodiethyl sulfide [3,4].

So far, there have been some well-developed technologies for the detection of toxic and harmful gases. These technologies include gas chromatography-mass spectrometry (GC-MS) [5], proton-transfer-reaction mass spectrometry (PTR-MS) [5], ion mobility spectrometry (IMS) [6], Raman spectroscopy [7], surface acoustic wave sensors (SAWs) [8], quartz-crystal microbalance sensors (QCMs) [9] and colorimetric tubes [10]. However, they have certain defects such as valuableness, low response and sensitivity, poor reversibility, etc. By comparison, gas sensors based on metal oxide semiconductors have advantages of low cost, portability, fast response time and easy fabrication [11,12,13].

There are a few semiconductor metal oxide sensors that have also been reported in the previous literature for the detection of 2-CEES gas. For example, Zheng et al. reported the detection of 2-chloroethyl ethyl sulfide (2-CEES) by crosslinked WO₃ nanonet. But the recovery time of the gas sensor to 2-CEES gas with a concentration of 5 ppm is up to 162 s [14]. Li et al. reported that hierarchical Fe₂O₃ nanotube arrays were synthesized by an in situ solvothermal method for the detection of 2-CEES gas. However, the recovery time of the gas sensor is 190 s [15]. Yang et al. reported that a hierarchical core–shell ZnFe₂O₄ microspheres gas sensor was fabricated for the detection of 2-CEES. However, the recovery time of the gas sensor for 1 ppm 2-CEES is as long as 546 s [13]. As is well known, the selectivity and response/recovery time of the sensors are important indicators for their practical application. Therefore, the goal of this paper is to improve the sensing properties of the above semiconductor metal oxide gas sensors, such as improving the selectivity to 2-CEES gas and shortening the response and recovery time.

Tungsten oxide (WO₃) is a typical n-type metal oxide, and its nanostructures are widely used as gas-sensing materials because of its superior physical and chemical properties [16]. In view of this, WO₃/Al₂O₃/graphite composite materials were prepared by high-temperature sintering and mechanical mixing, and the WO₃/Al₂O₃/graphite gas sensor for the detection of 2-CEES was achieved in this study. The addition of graphite and Al₂O₃ can achieve the following advantages. Graphite not only has a very large theoretical specific surface area (2630 m²/g) but also has a high electrical conductivity so that the added graphite can provide a very large gas induction area per unit volume and effectively adsorb a large number of gas molecules [17]. In addition, the added Al₂O₃ can generate more gas interaction centers on the surface of WO₃ and improve adhesion and stability [18,19]. The experimental results show that the gas sensor exhibits excellent sensing performance. More importantly, the response to 5.70 ppm 2-CEES was much higher than that of NH₃, acetone and acetonitrile, showing good selectivity. In order to demonstrate the practical application, a gas sensor array composed of WO₃/Al₂O₃/graphite was prepared based on MEMS microprocessing technology. This sensor array also has a high response to 2-CEES. Therefore, this work may provide an effective strategy for detecting dichlorodiethyl sulfide due to the excellent gas-sensing performance of the sensor for 2-CEES gas.

2. Experimental Procedure

2.1. Chemical Reagent

All chemicals for this study were used as received without further purification. Acetone (C₃H₆O), ammonia (NH₃(aq)) and ethanol (C₂H₅OH) were purchased from Sinopharm Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. 2-Chloroethy ethyl sulfide (C₄H₉ClS), acetonitrile (C₂H₃N) and tungsten oxide (WO₃, average particle size < 100 nm) were purchased from Shanghai Macklin Biochemical Co., Ltd., Shanghai, China. Aluminum oxide (20 nm) was purchased from Shanghai D&B Biological Science and Technology Co., Ltd., Shanghai, China.

2.2. Materials Synthesis

First, WO₃ and Al₂O₃ were ground and mixed uniformly in a mortar. Then, the above mixed materials were put into a muffle furnace and calcined at 110 °C, 600 °C and 700 °C for 1 h in sequence. Then, the above-prepared materials and graphite were stirred in ethanol/deionized water (volume fraction 5:1) for 24 h and finally dried in an oven at 60 °C for 24 h. The samples with Al₂O₃ mass accounting for 2, 4, 8 and 12 wt. % of the total mass of Al₂O₃ and WO₃ were named S1, S2, S3 and S4, respectively.

2.3. Characterization

The morphology and structure of the sensing materials were characterized by transmission electron microscope (TEM, JEM-2100, JEOL, Tokyo, Japan), scanning electron microscopy (SEM, S-4800, Hitachi, Hitachi City, Japan), and X-ray diffraction (Japan; XRD, SmartLab, 9 KW, Rigaku, Tokyo, Japan).

2.4. Gas-Sensor Fabrication and Gas-Sensing System

Grind the sample in deionized (DI) water into a uniform slurry in an agate mortar, then brush it onto a ceramic tube to form a uniform film, and dry it at 180 °C for 40 min. Then, weld the ceramic tube onto the hexagonal base. The nickel chromium alloy spiral heating wire passing through the ceramic tube is used to adjust the working temperature. The gas-sensing performance of the sensor was tested by the static volume method. The sensor is tested in a sealed chamber of 18 L, and when the resistance of the sensor is stable, the calculated target gas or liquid is injected into the chamber through a micro-syringe, which is quickly and evenly distributed in the test chamber under the drive of a fan. This is a simple and easy method, and it is widely used in the configuration of standard gases. A commercial gas sensor tester (WS-30B) is used to test the gas-sensing performance of the sensor. The whole test system was placed in the fume hood equipped with a dehumidifier. The relative humidity of the experimental environment was controlled at 35% RH. The measurement circuits contain a sensor resistance (R_S), an adjustable standard resistance (R_L), and DC voltage source (V_C). The working temperature of the sensor is regulated by the heating voltage (V_H), and the formula R_S = (V_C − V_output)R_L/V_output is used for calculating the sensor resistance. The following formula $\mathrm{C}=\frac{22.4\times \mathsf{\phi }\times \mathsf{\rho }\times {\mathrm{V}}_1}{\mathrm{M}\times {\mathrm{V}}_2}\times 1000$ is used to calculate the volume of injected liquid organic compounds, where M (g/mol), ρ (g/mL) and V₁ (μL) are, respectively, the molecular weight, density and volume of the liquid organic compounds, V₂ (L) is the volume of the gas chamber, C (ppm) is the target gas concentration, and φ is the target gas volume fraction. The response of the sensor is defined as: response (%) = (R_a – R_g)/R_a × 100, where R_a and R_g are the resistance in air and target gas. Response time is defined as the time required for the resistance to reach 90% of the stable resistance after the sensor contacts the gas under test. Recovery time is defined as the time required for the resistance value to decrease to 90% of the stable resistance after the sensor is exposed to air.

3. Results and Discussion

3.1. Structure and Measurement Circuit of the Sensor

The ceramic tube shell packaging is used to study the sensing characteristics of sensitive materials and the gas sensor. Figure 1a shows the photograph of the gas sensor that consists of a ceramic tube coated with sensing materials and a base with six probes. The measurement circuit of the gas sensor tester containing a sensor resistance (R_S), an adjustable standard resistance (R_L) and a DC voltage source (V_C) is shown in Figure 1b. The rated measurement voltage was added to both ends of the circuit, and then the voltage V_C at both ends of R_L can be measured, and the resistance of R_S can be calculated. To obtain sensor responses under different gases by real-time recording of R_S, the working temperature is crucial for the gas-sensing performance of metal oxides sensor. Thus, the working temperature of the sensor is regulated by the provided heating voltage (V_H) applying to the nickel chromium alloy. Figure 1c shows the structural diagram of the gas sensor that is welded to the base by platinum wires protruding from the gold electrodes at the ends of the ceramic tube.

3.2. Structure and Morphology

The microstructure of tungsten oxide, graphite and alumina oxide and the as-prepared composite S2 were examined by XRD, as shown in Figure 2a. The characteristic diffraction peaks at 23.55°, 34.11° and 49.88° can correspond well to the (020), (202) and (400) planes of WO₃ nanoparticles, respectively, and they can be indexed to the structure of WO₃ (JCPDS No.83-0950). No other secondary phases or impurity peaks were observed. This indicates the high purity of the synthetic WO₃/Al₂O₃/graphite composite materials. To further understand the surface morphologies of the synthesized samples, their surface morphologies were characterized by scanning electron microscopy (SEM), as shown in Figure 2b and Figure S1. As can be seen from the figures, the WO₃ and Al₂O₃ nanoparticles are tightly attached to the graphite sheets. But with the increase in the Al₂O₃ nanoparticle content, Al₂O₃ nanoparticles will also attach to the surface of WO₃ and agglomerate together. The phenomenon of agglomeration is not conducive to the gas-sensing performance, which indirectly indicates that the content of Al₂O₃ nanoparticles in the composites will have an optimal value.

The shape and size of the raw materials and synthesized sample were probed using a transmission electron microscope (TEM), as shown in Figure 3. As shown in Figure 3a, WO₃ is in the shape of nanoparticles with diameters ranging from tens to hundreds of nanometers. Figure 3b is a TEM image of Al₂O₃. It can be seen from the figure that the Al₂O₃ nanoparticles are tiny and aggregated together. Figure 3c is a TEM image of graphite. It can be seen from the figure that graphite is a multi-layer flake structure. The inset of Figure 3c shows the selected area electron diffraction (SAED) of the graphite, showing the polycrystalline diffraction ring. Figure 3d is the TEM image of the S2 sample. The figure reflects that the WO₃ nanoparticles and Al₂O₃ nanoparticles are attached to the surface of the graphite. To further confirm the distribution of WO₃, Al₂O₃, and graphite in the S2 sample, the energy-dispersive spectrometer (EDS) was used, and the results are shown in Figure 3e–i. It can be seen from the figures that the elements of Al, O, W and C are dispersed uniformly, which indicates that WO₃ nanoparticles and Al₂O₃ nanoparticles are evenly mixed and evenly distributed on the surface of graphite.

3.3. Gas-Sensing Properties

To select the optimum additive amount of Al₂O₃ and operating temperature, the gas-sensing properties of the sensors with different additive amounts of Al₂O₃ were tested at different operating temperatures. As shown in Figure 4a and Figure S2, the sensor based on sample S2 under 5.70 ppm 2-CEES exhibits the greatest response at each operating temperature. It can also be seen from the figure that the response of the sensor based on sample S2 increases with the increase in the working temperature at the beginning, and when the response reaches the maximum value, it decreases with the further increase in the working temperature. This phenomenon may be explained as follows. When the working temperature is low, water molecules in the air will be adsorbed to the surface of the sensing material, which hinders the interaction between oxygen ions on the surface active sites of the sensing material and 2-CEES gas molecules [15]. When the operating temperature increases, the oxidation rate and gas diffusion coefficient will be increased, so the response will increase. The response of the sensor is maximized when the interaction rate of the 2-CEES gas is equal to the diffusion rate of the 2-CEES gas [20]. However, as the operating temperature further increases, insufficient atmospheric oxygen adsorbed on the sensor surface results in insufficient sensing sites on the sensing material surface, so the response begins to degrade [21].

Next, the response time and recovery time of the sensor based on sample S2 were studied, and the response time and recovery time to 5.70 ppm of 2-CEES at different temperatures are shown in

Table 1. The sensor response, response time and recovery time of the S2 sensor to 5.70 ppm of 2-CEES gas at different temperatures.

Temperature	Response	Response Time	Recovery Time
260 °C	58%	17 s	170 s
300 °C	70%	12 s	142 s
340 °C	69%	5 s	42 s
440 °C	65%	4 s	12 s

. It can be seen that the higher the operating temperature, the shorter the response time and recovery time. This is because the higher the operating temperature, the stronger the adsorption and desorption ability of the gas on the sensor surface. Next, the selectivity of the sensor based on sample S2 was investigated. Figure 4b and Figure S3 exhibit the sensing responses of the S2 sensor to 5.70 ppm of 2-CEES and interfering gases (ammonia, ethanol, acetone and acetonitrile) at different temperatures. As can be seen from the figure, when the operating temperature is 260 °C, 300 °C, 340 °C and 440 °C, the response to ammonia, ethanol and acetonitrile is relatively low and changes little, and the response to 5.70 ppm 2-CEES was much higher. However, as the working temperature increases, the response to acetone is improved. Therefore, to sum up, the optimal operating temperature of the sensor based on sample S2 is 340 °C. At this time, the response to 2-CEES gas with a concentration of 5.70 ppm was 69%. The response of 2-CEES is almost 6 times higher than that of ammonia and ethanol, 3 times higher than that of acetone and 70 times higher than that of acetonitrile. The response time and recovery time of the sensor based on sample S2 to 5.70 ppm 2-CEES gas were 5 s and 42 s, respectively, as shown in Figure 4c. The sensing response of the sensor to 0.10 ppm 2-CEES at the temperature of 340 °C is shown in Figure 4d. When 0.10 ppm 2-CEES is injected, the S2 sensor has 27% response, indicating that the sensor can detect 2-CEES gas with low concentration. This helps its application in real life to detect 2-CEES gas and can detect the leakage of 2-CEES gas as soon as possible.

The response–recovery curves of the sensor based on sample S2 toward different 2-CEES gas concentrations in a range of 0.10–11.40 ppm are shown in Figure 5a. It can be seen from the figure that the response increased with the increment of 2-CEES gas concentration. This indicates that the sensor has good response recovery characteristics. The linear dependence curve of the response on the gas concentration was also studied (Figure 5b). The initial response increased with increasing 2-CEES gas concentration, and the response tended to saturate when the 2-CEES gas concentration exceeded 3.10 ppm. This shows that the sensor has a relatively wide detection range. A dynamic response diagram of the reproducibility and stability of the sensor is shown in Figure 5c,d. At the end of each cycle, the resistance of the sensor tends to return to its original appearance. Figure 5c shows that the sensor exhibits good repeatability for 5.70 ppm 2-CEES at a heating temperature of 340 °C. The gas-sensitive performance of the sensor was tested for 6 consecutive days, and the results showed excellent stability. The response of the sensor to 2-CEES gas was maintained near 69% despite some fluctuations, as shown in Figure 5d. The sensor based on the S2 sample has good repeatability and long-term stability, which shows the application value of detecting dichlorodiethyl sulfide gas. To demonstrate the superiority of the developed sensor, the gas-sensing performances to 2-CEES gas between this work and those reported in previous publications were compared (

Table 2. Comparison of gas-sensing characteristics of the S2 sensor with those reported in the literature.

Materials	Temperature	Response	Response Time	Recovery Time	Ref.
WO3/Al2O3/gra-phite	340 °C	69% #	5 s	42 s	This work
Pt-CdSnO3 thin film	300 °C	33.46 *	8 s	125 s	[22]
Al-doped ZnO NPs	500 °C	954 #	2 s	127 s	[23]
Sm2O3 doped SnO2 NPs	200 °C	540 #	50 s	1200 s	[24]
Ru-CdSnO3 thin film	350 °C	62.12 ^	5 s	185 s	[25]
crosslinked WO3 nanonets	217 °C	58 ^	1 s	181 s	[14]

The response calculation formula is as follows: R^#(%) = (R_a − R_g)/R_a × 100; S_# = R_a/R_g; S* = (G_g − G_a)/G_a; S^{^} = (I_g − I_a)/I_a. R_a and R_g are, respectively, the resistances in the air and in the test gas atmosphere. G_a and G_g are, respectively, the conductance in the air and in a target simulant medium. I_a and I_g are, respectively, the currents in the air and in the presence of simulant.

), and the corresponding comparison results demonstrate that the WO₃/Al₂O₃/graphite sensor exhibits high response, low working temperature, and it also shows fast response and recovery time. Therefore, all these results confirm that the WO₃/Al₂O₃/graphite are promising gas-sensing materials for dichlorodiethyl sulfide gas detection in practical applications.

In order to better apply a WO₃/Al₂O₃/graphite gas sensor to detect dichlorodiethyl sulfide gas in real life, the MEMS chip array was applied. The structure diagram of the MEMS gas sensor chip is shown in Figure 6a. It can be seen that the gas-sensitive materials are loaded in the middle of the MEMS chip, and four Pt electrodes are connected to the ends. Some platinum electrodes are used for heating, while others are used for information collection. Figure 6b,c clearly show the planar SEM image of the MEMS micro-heating chip. After the gas-sensitive materials are loaded in the heating plates and then connected with external circuits by wire bonding, the MEMS gas sensor array is achieved. As shown in Figure 6d, the MEMS gas sensor array testing system with ten test channels has been developed, and the gas sensors in each channel of the system can be tested under various types of gases. Next, to verify the consistency of the MEMS gas sensor array, we selected a sensor randomly for gas detection, and the response of the gas sensor to 2-CEES gas at a concentration of 5.70 ppm was measured at an optimal operating temperature of 340 °C. The response to various interfering gases, such as ammonia, ethanol and acetone, at the same concentration of 5.70 ppm under an operating temperature of 340 °C, are shown in Figure 7. As can be seen from Figure 7a, the single channel MEMS gas sensor has a high response to 2-CEES gas with a concentration of 5.70 ppm, and it has a rapid response and recovery. At the same time, it can be seen from Figure 7b, c and d that there was almost no response to ammonia and ethanol, and the response to acetone was much smaller than that to 2-CEES, which shows that the randomly selected MEMS gas sensor has good selectivity. These above experimental results indicate that sample S2’s gas-sensing material can be well applied in MEMS chips for detecting dichlorodiethyl sulfide gas.

3.4. Sensing Mechanism

The adsorption and desorption properties of gases near the surface of sensing materials influence the resistance of sensing materials [26,27,28,29]. A depletion layer theory can be used for the explanation of the gas-sensing mechanism. When the sensor is placed in a clean air atmosphere without 2-CEES, a lot of oxygen is adsorbed on the sensor surface, and then it forms ${\mathrm{O}}_2^{-}$, ${\mathrm{O}}^{-}$ and ${\mathrm{O}}^{2-}$ by reacting with the electrons transferred from the surface of the sensing materials [30,31]. The concentration of conduction band electrons in the sensing material decreases, an electron depletion layer is formed on the surface, and the resistance of the sensor increases. The reactions are described by Equations (1)–(4) [32].

$${\mathrm{O}}_2\left(\mathrm{g}\mathrm{a}\mathrm{s}\right)\to {\mathrm{O}}_2\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)$$

(1)

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

(2)

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

(3)

$${\mathrm{O}}^{-}\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)+{\mathrm{e}}^{-}\to {\mathrm{O}}^{2-}\left(\mathrm{a}\mathrm{d}\mathrm{s}\right)$$

(4)

Subsequently, at a certain temperature, the sensor is transferred to the 2-CEES atmosphere, and 2-CEES is rapidly decomposed into ClCH₂CH₂· and·SCH₂CH₃. Due to the strong electrophilic properties of Cl and S, when they are adsorbed on the surface of the sensing material, they will react with ${\mathrm{O}}^{-}$, are quickly oxidized and release a large number of electrons. The reactions are described by Equations (5) and (6). These free electrons will be transferred to the conduction band of the sensing material. The concentration of free electrons in the conduction band increases, resulting in the decrease in the depletion layer and the corresponding decrease in the resistance of the sensing material [33], as shown in Figure 8.

$$2\mathrm{C}\mathrm{l}\mathrm{C}{\mathrm{H}}_2\mathrm{C}{\mathrm{H}}_2·+8{\mathrm{O}}^{-}\to 2\mathrm{C}{\mathrm{O}}_2+\mathrm{C}{\mathrm{l}}_2+4{\mathrm{H}}_2\mathrm{O}+8{\mathrm{e}}^{-}$$

(5)

$$2\mathrm{C}{\mathrm{H}}_3\mathrm{C}{\mathrm{H}}_2\mathrm{S}·+13{\mathrm{O}}^{-}\to 2\mathrm{S}{\mathrm{O}}_2+5{\mathrm{H}}_2\mathrm{O}+2\mathrm{C}{\mathrm{O}}_2+13{\mathrm{e}}^{-}$$

(6)

The process of Al₂O₃ distribution in WO₃ increases the specific surface area of WO₃, generates more gas interaction centers on the surface of WO₃, and improves adhesion and stability. WO₃ distributed with Al₂O₃ still retains the n-type semiconductor properties of undistributed WO₃. The incorporation of Al₂O₃ can introduce more gas adsorption active sites for WO₃. When the content of Al₂O₃ particles is low, the distribution has only a slight effect on WO₃. With the increase in the content of Al₂O₃ particles, the adsorption of 2-CEES increases. However, with the continued incorporation of Al₂O₃ particles, the aggregation of Al₂O₃ particles will reduce the number of active sites (as shown in Figure S1b,c), and the adsorption capacity of the material for 2-CEES will decrease. Therefore, S2 can be used to obtain the maximum gas-sensitive property. The addition of graphite can increase the specific surface area of the sensing material, provide more interface contact and thus lead to more gas adsorption and a higher carrier mobility of sensing materials [34].

4. Conclusions

In this study, WO₃/Al₂O₃/graphite nanocomposites were obtained with a simple high-temperature sintering and mechanical mixing. Compared with the sensors with different Al₂O₃ contents, the sensor with Al₂O₃ content of 4 wt. % exhibited the best gas-sensing characteristics. Experimental results show that when the working temperature is 340 °C, the response time to 2-CEES gas with a concentration of 5.70 ppm is 5 s, the recovery time is 42 s and the response is 69%. The response of 2-CEES is almost 6 times higher than that of ammonia and ethanol, 3 times higher than that of acetone and 70 times higher than that of acetonitrile, which shows that this sensor exhibits an excellent selectivity to several possible interferents. The sensor also has the advantages of good repeatability and stability, and it can detect low-concentration 2-CEES gas. The MEMS gas sensor array based on WO₃/Al₂O₃/graphite composite materials also exhibits excellent sensing performance. Considering the simple synthesis method and excellent gas-sensing performance, there is no doubt that WO₃/Al₂O₃/graphite prepared in this study is a promising sensing material and will be widely used in the detection of toxic and harmful gas.