Nitrogen dioxide (NO₂) is toxic and irritating gas, which comes mainly from motor vehicle exhaust and boiler exhaust emissions. The nervous system and respiratory tract will be visibly impaired when the concentration of NO₂ exceeds 3 ppm. In addition, high concentrations of NO₂ also have cancer-promoting and carcinogenic effect [1]. In the past decade, the demands for low cost, portable, high sensitivity and low power NO₂ sensors have greatly stimulated the research enthusiasm on novel sensing materials [2].

Two-dimensional (2D) materials such as graphene, transition metal disulfide and transition metal carbide (MXene) are considered as the most promising alternative candidate materials for the next generation of gas sensors due to the large specific surface area, unique electronic band structure and low noise for application in devices [3-5]. Molybdenum disulfide (MoS₂), as a transition metal disulfide, has not only layered structure similar to graphene, but also adjustable band gap [6, 7], excellent optical and electrical properties [8], which has great potential in the fields of energy storage [9], photocatalysis [10], nano devices [11], gas sensors [12, 13], humidity sensor [14] and so on. Lee et al. prepared twodimensional MoS₂ nanosheets by hydrothermal method and found that the sensitivity to NO₂ gas was related to the S/Mo precursor ratio [15]. However, they only studied the gas sensitivity of the material to 500 ppm NO₂. Kumar et al. developed a vertical arrangement MoS₂ flake synthesized by chemical vapor deposition (CVD) and the gas sensor showed high response but low recovery speed to NO₂ at room temperature [16]. These results fully demonstrated the potential of MoS₂ for NO₂ gas detection. However, there are still some problems to be solved for NO₂ sensor application: (1) The prepared MoS₂ films must be uniform, reliable, and easy to integrate; (2) The MoS₂ gas sensors should has excellent selectivity and response/recovery speed to NO₂ at room temperature.

Although there are many methods for the preparation of two-dimensional MoS₂ [14-18], the existing methods such as hydrothermal and micromechanical exfoliation have poor controllability in growth quality and area [17]. More important, the growth area is small, which cannot guarantee the uniformity of thickness and the consistency of layers in the region [18]. Therefore, there is an urgent need for a simple and effective method to grow two-dimensional MoS₂, as well as an economic process for manufacturing highly reliable sensor devices.

In this work, two-dimensional MoS₂ film was prepared by sulfuring molybdenum atomic layer on SiO₂/Si substrate. The important growth factors such as reaction temperature, heating rate, holding time and carrier gas velocity were regulated in the synthesis process and the quality of MoS₂ film was evaluated by optical microscope, atomic force microscope, Raman spectrum and photoluminescence spectrum. In addition, two-dimensional MoS₂ gas sensor was fabricated and its gas sensing performance was studied. The results showed that the MoS₂ sensor had good gas sensitive response only to ammonia and NO₂. In the last, the sensitivity mechanism of the material was explored initially.

Preparation of two-dimensional MoS₂ film: The MoS₂ film was prepared by a CVD system as shown in the Fig. S1b (Supporting information), which includes gas cylinder, gas path, flow control meter and a single temperature zone tubular furnace. First of all, molybdenum films with the thickness of 1 nm were prepared by electron beam evaporation on 300 nm SiO₂/Si substrates. Then, 0.8 g high purity sulfur (99.99%, Aladdin) was ground into powder and laid flat in the porcelain boat as shown in the Fig. S1d (Supporting information). Then, the sulfur was putted in the uptake of the tubular furnace heating zone so that the temperature of the position can be sublimation during the reaction. The Mo/SiO₂/Si substrate was placed in the constant temperature area of the tubular furnace. In the preparation, the temperature of tubular furnace during the whole reaction process as shown in the Fig. S1c (Supporting information) was set as follows: In the first stage, the temperature raised from room temperature to 200 ℃; in the second stage, the temperature raised from 200 ℃ to different reaction temperatures; in the third stage, the temperature was kept for heat preservation, the temperature was naturally lowered to room temperature. In the last, MoS₂ thin films were prepared. The effects of reaction temperature, heating rate, holding time and carrier gas velocity were studied. The name of samples were defined as "sample a-i" for different preparation conditions as shown in the Table S1 (Supporting information).

Fabrication and measurement of MoS₂ gas sensor: the schematic diagram of the MoS₂ gas sensor is shown in Fig. 1a. The Au electrodes were first vacuum evaporation on MoS₂ film. Then, the non-conductive AB adhesive coated on Au electrodes to increase its strength and protect electrode. The picture of real products of MoS₂ gas sensor could be seen in the Fig. 1b. The resistance of MoS₂ has obvious response to light through the previous report, so it was tested to exclude the influence of light by shading aluminum foil out of test chamber. Using static gas testing method, methanol, ethanol, acetone, formaldehyde, ammonia, and nitrogen dioxide were selected test. The gas concentration in the test was 100 ppm.

The sensitive performance was studied on the CGS-1TP intelligent gas sensitivity analysis system (Beijing Elite Tech Co., Ltd., China). Gas concentration was obtained by static distribution method [19]. In this work, the sensitivity of NO₂ is defined as ΔR/R_a×100%, where ΔR = R_g – R_a, R_g and R_a are the resistances of materials in sensitive gas and air. However, the sensitivity of reduction gases such as NH₃ is defined as (–ΔR)/R_a×100%. Response/recovery time is defined as the time required when the resistance change reaches 90% of the total change.

The effect of preparation temperature on MoS₂ film growth was studied firstly. The temperatures of tubular furnaces in the preparation were set at 650 ℃, 750 ℃ and 850 ℃ for samples a, b and c, respectively. Figs. S2a-c (Supporting information) were optical microscope (OM) photographs of samples a–c, which could reflect the uniformity of the MoS₂ film (Supporting information). In the OM images, the MoS₂ film on the SiO₂/Si substrate appeared to be blue, and the area without MoS₂ appeared purple (Fig. 2a). Obviously, the samples grown at 750 ℃ and 850 ℃ were more uniform in the whole area than the sample grown at 650 ℃. In addition, the surface uniformity of MoS₂ film was characterized by atomic force microscopy in Fig. 2b. Obviously, the surface of MoS₂ film is very flat and continuous, which facilitates electron transport when used as sensitive material. The quality and layer number of the MoS₂ film could be studied by Raman spectra (Fig. 2c and Fig. S3 in Supporting information). Generally, two typical Raman peak, E_2g¹ in-plane vibration mode and A_1g out plane vibration mode were investigated to reflect the crystal structure of MoS₂. For the "sample a" as shown in Fig. S3, the E_2g¹ and A_1g were 383.5 cm^-1 and 408.1 cm^-1, with the peak distance of 24.6 cm^-1. As shown in Fig. 2c, the E_2g¹ and A_1g of "sample b" were located at 384.0 cm^-1 and 406.6 cm-1 [20], with the peak distance of 22.6 cm^-1. As reported, the peak frequency difference between E_2g¹ and A_1g decreased with layer number. Therefore, the MoS₂ film prepared at 750 ℃ had more layers then prepared at 650 ℃. In addition, the intensity of "sample b" stronger than that of "sample a", which indicated its higher quality. For the "sample c" prepared at 850 ℃, no Raman peaks were observed, which might be due to the evaporation of sulfur powder too quickly to reaction with Mo on the SiO₂/Si. In addition to Raman characteristics, we also investigated the PL properties of the MoS₂ films, which could reflect the number of layers, film quality and energy level structure of MoS₂ film. As shown in Fig. 2d, the obvious PL peak could be observed in "sample b", but there is no obvious PL peak in other samples (Fig. S4 in Supporting information). Therefore, the 750 ℃ was the best reaction temperature in the preparation of MoS₂ film.

The effect of holding time on film growth was further investigated with the reaction temperature set as 750 ℃. The holding time of samples d, b and e were 0 min, 30 min and 60 min, respectively. As shown in Figs. S2d, b and e (Supporting information), the MoS₂ film was only obtained at the holding time of 30 min. Short holding time means short contact time between S and Mo. However, the long holding time might be led to the excess S deposition on the substrate surface. As shown in Fig. S3, no Raman or PL peaks was observed due to the absence of MoS₂. Therefore, 30 min was a suitable holding time.

Then, the effect of heating rate on film growth was also researched. Three different heating rates (3 ℃/min, 25 ℃/min and 40 ℃/min) were set to prepare the MoS₂ film. As shown in Figs. S2b, f and g (Supporting information), almost no MoS₂ was formed when the heating rate was 3 ℃/min, which due to the sulfur evaporate prematurely and the reaction cannot be carried out. When the heating rate was 40 ℃/min, a rough surface on substrate was observed in the Fig. S2g. It was worth noting that both the Raman and PL peaks were observed. The rough surface on substrate might be because of the violent chemical reaction between sulfur and molybdenum layer. Therefore, 25 ℃/min was a suitable heating rate.

In the last, the carrier gas velocity on MoS₂ film growth was investigated. Figs. S2h, i and b are the OM images of MoS₂ film prepared under three different carrier gas velocities of 3 sccm, 50 sccm and 100 sccm, respectively. Obviously, all three samples shown similar blue surface. Both Raman and PL spectra proved that these MoS₂ films were successfully prepared. The intensities of both Raman and PL peaks increased with the carrier gas velocity. In addition, the Raman peak distance of "sample i" was 23.1, which smaller than that of "sample b". The smaller Raman peak distance might be from the small layer of MoS₂ film. Therefore, the faster carrier gas velocity would help to the growth of MoS₂ film.

In a short, the optimum synthesis parameters for MoS₂ film were heating rate of 25 ℃/min, reaction temperature of 750 ℃, holding time of 30 min and carrier gas velocity of 100 sccm. The surface of the sample obtained under this condition is even, which can be seen from the AFM characterization of "sample b" in Fig. 2b.

The gas sensing performance of MoS₂ gas sensor was tested. As shown in Fig. 3a, 100 ppm of gases which included reducing gases such as ethanol, methanol, acetone, formaldehyde, NH₃ and oxidizing gas such as NO₂ were tested at room temperature. Interestingly, the MoS₂ gas sensor showed good response to NH₃(Fig. S5 in Supporting information) and NO₂ and almost no response to other gases, which indicated the good selectivity of the MoS₂ gas sensor. The relationship between sensitivity and NO₂ concentration was shown in Fig. 3b. The sensitivity of MoS₂ gas sensor to 1, 5, 10, 20, 50 and 100 ppm NO₂ is 1.2%, 4.2%, 7.0%, 13.0%, 20.9% and 31.3%, respectively. Fig. 3c is a fitting curve of sensitivity and NO₂ concentration for MoS₂ gas sensor. Obviously, the logarithm of concentration was linear with logarithm of sensitivity. It can be expressed as lg(S) = 0.16 + 0.69lg(C), where S and C represent sensitivity and gas concentration respectively. Furthermore, the dynamic responses of the MoS₂ gas sensor was shown in Fig. 3d, which demonstrated the stable sensing recyclability toward NO₂ gas. In addition, the response recovery curve of MoS₂ gas sensor to NH₃ reaction is shown in Fig. S3, the response is about 20, but the recovery time is longer. Compared with other NO₂ sensors reported in Table 1 [21-25], the MoS₂ gas sensor in this work has great advantages in high NO₂ sensitivity, low limit of detection, rapid response/recovery speed and can operation at room temperature.

The gas-sensitive mechanism of MoS₂ gas sensor could be understood from the structure of MoS₂ materials. As shown in Fig. 4, the sulfur atoms are mainly on the outer layer of MoS₂. These sulfur atoms are easily to polarize because of the large atom. For the NO₂ gas, the electronegativity of O (3.52) is larger than N (3.08), which causes some of positive charge on the nitrogen atom [26]. When NO₂ is in contact with MoS₂ film, the NO₂ will absorbs electrons and becomes NO₂^- (ads) [27]. As we known, MoS₂ is a N-type semiconductor, in which the electrons are the main carriers [12]. The formation of NO₂^- (ads) will increases the resistance of the MoS₂ film, which was very consisted with the result of Fig. 3b. NO₂ and adsorbed electrons will react as follows [28].

On the other hand, the electronegativity of H (2.20) is smaller than N (3.08), which causes some of negative charge on the nitrogen atom. The electron of NH₃ will transfer to MoS₂ film and decrease the resistance. In addition, the MoS₂ film synthesized by indirect CVD method is polycrystalline. There are a large number of suspended bonds on the grain boundary surface. These suspended bonds are relatively easy to be absorbed and utilized by NO₂, thus making gas sensing properties [29].

In this work, MoS₂ films were successfully prepared on SiO₂/Si substrates by chemical vapor deposition. The quality of the film was observed by optical microscope, Raman spectra and PL spectra. By studying the control variables, the optimum conditions for the synthesis of MoS₂ film were determined. The characterization results showed that the optimum synthesis parameters were heating rate of 25 ℃/min, reaction temperature of 750 ℃, holding time of 30 min and carrier gas velocity of 100 sccm. The MoS₂ gas sensor was fabricated and its gas sensing performance was tested. It has good sensitive performance to NO₂ gas detection, which include the good selectivity, fast response/ recovery speed and low limit of detection. The sensitivity to 100 ppm of NO₂ was 31.3%, and the response/recovery times were 4 s and 5 s, respectively. In addition, the limit of detection could be as low as 1 ppm. The sensitive mechanism is attributed to the structure of the MoS₂ material and the electrophilic ability of the NO₂ gas.

We greatly thank the financial supports from the National Natural Science Foundation of China (Nos. 51572173, 51602197, 51771121 and 51702212), Shanghai Municipal Science and Technology Commission (Nos.19ZR1435200, 18511110600 and 19JC1410402), Innovation Program of Shanghai Municipal Education Commission (No. 2019- 01-07-00-07-E00015), Shanghai Academic/Technology Research Leader Program (No. 19XD1422900).

Supplementary material related to this article can befound, in the online version, at doi:https://doi.org/10.1016/j.cclet.2019.11.032.